vc4: Use NEON to speed up utile stores on Pi2+.

Improves 1024x1024 TexSubImage2D by 41.2371% +/- 3.52799% (n=10).

.travis.yml

@@ -1,11 +1,26 @@
 language: c
 
-sudo: false
+sudo: true
 dist: trusty
 
 cache:
-  apt: true
-  ccache: true
+  directories:
+    - $HOME/.ccache
+
+addons:
+  apt:
+    packages:
+      - libdrm-dev
+      - x11proto-xf86vidmode-dev
+      - libexpat1-dev
+      - libxcb-dri2-0-dev
+      - libx11-xcb-dev
+      - llvm-3.5-dev
+      # llvm-config is not in the dev package?
+      - llvm-3.5
+      # LLVM packaging is broken and misses this dep.
+      - libedit-dev
+      - scons
 
 env:
   global:
@@ -21,266 +36,18 @@ env:
     - XCBPROTO_VERSION=xcb-proto-1.11
     - LIBXCB_VERSION=libxcb-1.11
     - LIBXSHMFENCE_VERSION=libxshmfence-1.2
-    - LIBTXC_DXTN_VERSION=libtxc_dxtn-1.0.1
-    - LIBVDPAU_VERSION=libvdpau-1.1
-    - LIBVA_VERSION=libva-1.6.2
-    - LIBWAYLAND_VERSION=wayland-1.11.1
     - PKG_CONFIG_PATH=$HOME/prefix/lib/pkgconfig
-    - LD_LIBRARY_PATH="$HOME/prefix/lib:$LD_LIBRARY_PATH"
-
-matrix:
-  include:
-    - env:
-        - LABEL="make loaders/classic DRI"
-        - BUILD=make
-        - MAKEFLAGS="-j4"
-        - MAKE_CHECK_COMMAND="make check"
-        - DRI_LOADERS="--enable-glx --enable-gbm --enable-egl --with-egl-platforms=x11,drm,surfaceless,wayland --enable-osmesa"
-        - DRI_DRIVERS="i915,i965,radeon,r200,swrast,nouveau"
-        - GALLIUM_ST="--enable-dri --disable-opencl --disable-xa --disable-nine --disable-xvmc --disable-vdpau --disable-va --disable-omx --disable-gallium-osmesa"
-        - GALLIUM_DRIVERS=""
-        - VULKAN_DRIVERS=""
-      addons:
-        apt:
-          packages:
-            - xz-utils
-            - x11proto-xf86vidmode-dev
-            - libexpat1-dev
-            - libx11-xcb-dev
-    - env:
-        # NOTE: Building SWR is 2x (yes two) times slower than all the other
-        # gallium drivers combined.
-        # Start this early so that it doesn't hunder the run time.
-        - LABEL="make Gallium Drivers SWR"
-        - BUILD=make
-        - MAKEFLAGS="-j4"
-        - MAKE_CHECK_COMMAND="true"
-        - LLVM_VERSION=3.9
-        - LLVM_CONFIG="llvm-config-${LLVM_VERSION}"
-        - OVERRIDE_CC="gcc-5"
-        - OVERRIDE_CXX="g++-5"
-        - DRI_LOADERS="--disable-glx --disable-gbm --disable-egl"
-        - DRI_DRIVERS=""
-        - GALLIUM_ST="--enable-dri --disable-opencl --disable-xa --disable-nine --disable-xvmc --disable-vdpau --disable-va --disable-omx --disable-gallium-osmesa"
-        - GALLIUM_DRIVERS="swr"
-        - VULKAN_DRIVERS=""
-      addons:
-        apt:
-          sources:
-            - ubuntu-toolchain-r-test
-            - llvm-toolchain-trusty-3.9
-          packages:
-            # LLVM packaging is broken and misses these dependencies
-            - libedit-dev
-            # From sources above
-            - g++-5
-            - llvm-3.9-dev
-            # Common
-            - xz-utils
-            - x11proto-xf86vidmode-dev
-            - libexpat1-dev
-            - libx11-xcb-dev
-            - libelf-dev
-    - env:
-        - LABEL="make Gallium Drivers Other"
-        - BUILD=make
-        - MAKEFLAGS="-j4"
-        - MAKE_CHECK_COMMAND="true"
-        - LLVM_VERSION=3.9
-        - LLVM_CONFIG="llvm-config-${LLVM_VERSION}"
-        - DRI_LOADERS="--disable-glx --disable-gbm --disable-egl"
-        - DRI_DRIVERS=""
-        - GALLIUM_ST="--enable-dri --disable-opencl --disable-xa --disable-nine --disable-xvmc --disable-vdpau --disable-va --disable-omx --disable-gallium-osmesa"
-        - GALLIUM_DRIVERS="i915,nouveau,r300,r600,radeonsi,freedreno,svga,swrast,vc4,virgl,etnaviv,imx"
-        - VULKAN_DRIVERS=""
-      addons:
-        apt:
-          sources:
-            - llvm-toolchain-trusty-3.9
-          packages:
-            # LLVM packaging is broken and misses these dependencies
-            - libedit-dev
-            # From sources above
-            - llvm-3.9-dev
-            # Common
-            - xz-utils
-            - x11proto-xf86vidmode-dev
-            - libexpat1-dev
-            - libx11-xcb-dev
-            - libelf-dev
-    - env:
-        # NOTE: Analogous to SWR above, building Clover is quite slow.
-        - LABEL="make Gallium ST Clover"
-        - BUILD=make
-        - MAKEFLAGS="-j4"
-        - MAKE_CHECK_COMMAND="true"
-        - LLVM_VERSION=3.6
-        - LLVM_CONFIG="llvm-config-${LLVM_VERSION}"
-        - OVERRIDE_CC=gcc-4.7
-        - OVERRIDE_CXX=g++-4.7
-        - DRI_LOADERS="--disable-glx --disable-gbm --disable-egl"
-        - DRI_DRIVERS=""
-        - GALLIUM_ST="--disable-dri --enable-opencl --enable-opencl-icd --enable-llvm --disable-xa --disable-nine --disable-xvmc --disable-vdpau --disable-va --disable-omx --disable-gallium-osmesa"
-        # i915 most likely doesn't work with OpenCL.
-        # Regardless - we're doing a quick build test here.
-        - GALLIUM_DRIVERS="i915"
-        - VULKAN_DRIVERS=""
-      addons:
-        apt:
-          sources:
-            - llvm-toolchain-trusty-3.6
-          packages:
-            - libclc-dev
-            # LLVM packaging is broken and misses these dependencies
-            - libedit-dev
-            - g++-4.7
-            # From sources above
-            - llvm-3.6-dev
-            - clang-3.6
-            - libclang-3.6-dev
-            # Common
-            - xz-utils
-            - x11proto-xf86vidmode-dev
-            - libexpat1-dev
-            - libx11-xcb-dev
-            - libelf-dev
-    - env:
-        - LABEL="make Gallium ST Other"
-        - BUILD=make
-        - MAKEFLAGS="-j4"
-        - MAKE_CHECK_COMMAND="true"
-        - DRI_LOADERS="--disable-glx --disable-gbm --disable-egl"
-        - DRI_DRIVERS=""
-        - GALLIUM_ST="--enable-dri --disable-opencl --enable-xa --enable-nine --enable-xvmc --enable-vdpau --enable-va --enable-omx --enable-gallium-osmesa"
-        # We need swrast for osmesa and nine.
-        # i915 most likely doesn't work with most ST.
-        # Regardless - we're doing a quick build test here.
-        - GALLIUM_DRIVERS="i915,swrast"
-        - VULKAN_DRIVERS=""
-      addons:
-        apt:
-          packages:
-            # Nine requires gcc 4.6... which is the one we have right ?
-            - libxvmc-dev
-            # Build locally, for now.
-            #- libvdpau-dev
-            #- libva-dev
-            - libomxil-bellagio-dev
-            # LLVM packaging is broken and misses these dependencies
-            - libedit-dev
-            # Common
-            - xz-utils
-            - x11proto-xf86vidmode-dev
-            - libexpat1-dev
-            - libx11-xcb-dev
-            - libelf-dev
-    - env:
-        - LABEL="make Vulkan"
-        - BUILD=make
-        - MAKEFLAGS="-j4"
-        - MAKE_CHECK_COMMAND="make -C src/gtest check && make -C src/intel check"
-        - LLVM_VERSION=3.9
-        - LLVM_CONFIG="llvm-config-${LLVM_VERSION}"
-        # XXX: we want to test the WSI, but those are enabled via the EGL toggles
-        # XXX: Platform X11 dependencies are checked when --enable-glx is set
-        - DRI_LOADERS="--enable-glx --disable-gbm --enable-egl --with-egl-platforms=x11,wayland"
-        # i965 is needed for Intel's Vulkan driver (anv)
-        - DRI_DRIVERS="i965"
-        # XXX: enable DRI for EGL above
-        - GALLIUM_ST="--enable-dri --disable-opencl --disable-xa --disable-nine --disable-xvmc --disable-vdpau --disable-va --disable-omx --disable-gallium-osmesa"
-        - GALLIUM_DRIVERS=""
-        - VULKAN_DRIVERS="intel,radeon"
-      addons:
-        apt:
-          sources:
-            - llvm-toolchain-trusty-3.9
-          packages:
-            # LLVM packaging is broken and misses these dependencies
-            - libedit-dev
-            # From sources above
-            - llvm-3.9-dev
-            # Common
-            - xz-utils
-            - x11proto-xf86vidmode-dev
-            - libexpat1-dev
-            - libx11-xcb-dev
-            - libelf-dev
-    - env:
-        - LABEL="scons"
-        - BUILD=scons
-        - SCONSFLAGS="-j4"
-        # Explicitly disable.
-        - SCONS_TARGET="llvm=0"
-        # Keep it symmetrical to the make build.
-        - SCONS_CHECK_COMMAND="scons llvm=0 check"
-      addons:
-        apt:
-          packages:
-            - scons
-            # Common
-            - xz-utils
-            - x11proto-xf86vidmode-dev
-            - libexpat1-dev
-            - libx11-xcb-dev
-            - libelf-dev
-    - env:
-        - LABEL="scons LLVM"
-        - BUILD=scons
-        - SCONSFLAGS="-j4"
-        - SCONS_TARGET="llvm=1"
-        # Keep it symmetrical to the make build.
-        - SCONS_CHECK_COMMAND="scons llvm=1 check"
-        - LLVM_VERSION=3.3
-        - LLVM_CONFIG="llvm-config-${LLVM_VERSION}"
-      addons:
-        apt:
-          packages:
-            - scons
-            # LLVM packaging is broken and misses these dependencies
-            - libedit-dev
-            - llvm-3.3-dev
-            # Common
-            - xz-utils
-            - x11proto-xf86vidmode-dev
-            - libexpat1-dev
-            - libx11-xcb-dev
-            - libelf-dev
-    - env:
-        - LABEL="scons SWR"
-        - BUILD=scons
-        - SCONSFLAGS="-j4"
-        - SCONS_TARGET="swr=1"
-        - LLVM_VERSION=3.9
-        - LLVM_CONFIG="llvm-config-${LLVM_VERSION}"
-        # Keep it symmetrical to the make build. There's no actual SWR, yet.
-        - SCONS_CHECK_COMMAND="true"
-        - OVERRIDE_CC="gcc-5"
-        - OVERRIDE_CXX="g++-5"
-      addons:
-        apt:
-          sources:
-            - ubuntu-toolchain-r-test
-            - llvm-toolchain-trusty-3.9
-          packages:
-            - scons
-            # LLVM packaging is broken and misses these dependencies
-            - libedit-dev
-            # From sources above
-            - g++-5
-            - llvm-3.9-dev
-            # Common
-            - xz-utils
-            - x11proto-xf86vidmode-dev
-            - libexpat1-dev
-            - libx11-xcb-dev
-            - libelf-dev
+  matrix:
+    - BUILD=make
+    - BUILD=scons
 
 install:
+  - export PATH="/usr/lib/ccache:$PATH"
   - pip install --user mako
 
   # Since libdrm gets updated in configure.ac regularly, try to pick up the
   # latest version from there.
-  - for line in `grep "^LIBDRM.*_REQUIRED=" configure.ac`; do
+  - for line in `grep "^LIBDRM_.*_REQUIRED=" configure.ac`; do
       old_ver=`echo $LIBDRM_VERSION | sed 's/libdrm-//'`;
       new_ver=`echo $line | sed 's/.*REQUIRED=//'`;
       if `echo "$old_ver,$new_ver" | tr ',' '\n' | sort -Vc 2> /dev/null`; then
@@ -325,70 +92,21 @@ install:
 
   - wget http://dri.freedesktop.org/libdrm/$LIBDRM_VERSION.tar.bz2
   - tar -jxvf $LIBDRM_VERSION.tar.bz2
-  - (cd $LIBDRM_VERSION && ./configure --prefix=$HOME/prefix --enable-vc4 --enable-freedreno --enable-etnaviv-experimental-api && make install)
+  - (cd $LIBDRM_VERSION && ./configure --prefix=$HOME/prefix --enable-vc4 --enable-etnaviv-experimental-api && make install)
 
   - wget $XORG_RELEASES/lib/$LIBXSHMFENCE_VERSION.tar.bz2
   - tar -jxvf $LIBXSHMFENCE_VERSION.tar.bz2
   - (cd $LIBXSHMFENCE_VERSION && ./configure --prefix=$HOME/prefix && make install)
 
-  # libtxc-dxtn uses the patented S3 Texture Compression
-  # algorithm. Therefore, we don't want to use this library but it is
-  # still possible through setting the USE_TXC_DXTN variable to yes in
-  # the travis web UI.
-  #
-  # According to Wikipedia, the patent expires on October 2, 2017:
-  # https://en.wikipedia.org/wiki/S3_Texture_Compression#Patent
-  - if test "x$USE_TXC_DXTN" = xyes; then
-      wget https://people.freedesktop.org/~cbrill/libtxc_dxtn/$LIBTXC_DXTN_VERSION.tar.bz2;
-      tar -jxvf $LIBTXC_DXTN_VERSION.tar.bz2;
-      (cd $LIBTXC_DXTN_VERSION && ./configure --prefix=$HOME/prefix && make install);
-    fi
-
-  - wget http://people.freedesktop.org/~aplattner/vdpau/$LIBVDPAU_VERSION.tar.bz2
-  - tar -jxvf $LIBVDPAU_VERSION.tar.bz2
-  - (cd $LIBVDPAU_VERSION && ./configure --prefix=$HOME/prefix && make install)
-
-  - wget http://www.freedesktop.org/software/vaapi/releases/libva/$LIBVA_VERSION.tar.bz2
-  - tar -jxvf $LIBVA_VERSION.tar.bz2
-  - (cd $LIBVA_VERSION && ./configure --prefix=$HOME/prefix --disable-wayland --disable-dummy-driver && make install)
-
-  - wget http://wayland.freedesktop.org/releases/$LIBWAYLAND_VERSION.tar.xz
-  - tar -axvf $LIBWAYLAND_VERSION.tar.xz
-  - (cd $LIBWAYLAND_VERSION && ./configure --prefix=$HOME/prefix --enable-libraries --without-host-scanner --disable-documentation --disable-dtd-validation && make install)
-
-  # Generate the header since one is missing on the Travis instance
-  - mkdir -p linux
-  - printf "%s\n" \
-           "#ifndef _LINUX_MEMFD_H" \
-           "#define _LINUX_MEMFD_H" \
-           "" \
-           "#define __NR_memfd_create 319" \
-           "#define SYS_memfd_create __NR_memfd_create" \
-           "" \
-           "#define MFD_CLOEXEC             0x0001U" \
-           "#define MFD_ALLOW_SEALING       0x0002U" \
-           "" \
-           "#endif /* _LINUX_MEMFD_H */" > linux/memfd.h
-
 script:
   - if test "x$BUILD" = xmake; then
-      test -n "$OVERRIDE_CC" && export CC="$OVERRIDE_CC";
-      test -n "$OVERRIDE_CXX" && export CXX="$OVERRIDE_CXX";
-      export CC="$CC -isystem`pwd`";
-
       ./autogen.sh --enable-debug
-        $DRI_LOADERS
-        --with-dri-drivers=$DRI_DRIVERS
-        $GALLIUM_ST
-        --with-gallium-drivers=$GALLIUM_DRIVERS
-        --with-vulkan-drivers=$VULKAN_DRIVERS
+        --with-egl-platforms=x11,drm
+        --with-dri-drivers=i915,i965,radeon,r200,swrast,nouveau
+        --with-gallium-drivers=svga,swrast,vc4,virgl,r300,r600,etnaviv,imx
         --disable-llvm-shared-libs
-        &&
-      make && eval $MAKE_CHECK_COMMAND;
-    fi
-
-  - if test "x$BUILD" = xscons; then
-      test -n "$OVERRIDE_CC" && export CC="$OVERRIDE_CC";
-      test -n "$OVERRIDE_CXX" && export CXX="$OVERRIDE_CXX";
-      scons $SCONS_TARGET && eval $SCONS_CHECK_COMMAND;
+        ;
+      make && make check;
+    elif test x$BUILD = xscons; then
+      scons;
     fi

Android.common.mk

@@ -78,22 +78,10 @@ endif
 
 ifeq ($(MESA_ENABLE_LLVM),true)
 LOCAL_CFLAGS += \
+	-DHAVE_LLVM=0x0305 -DMESA_LLVM_VERSION_PATCH=2 \
 	-D__STDC_CONSTANT_MACROS \
 	-D__STDC_FORMAT_MACROS \
 	-D__STDC_LIMIT_MACROS
-
-  ifeq ($(MESA_ANDROID_MAJOR_VERSION),5)
-    LOCAL_CFLAGS += -DHAVE_LLVM=0x0305 -DMESA_LLVM_VERSION_PATCH=2
-    ELF_INCLUDES := external/elfutils/0.153/libelf
-  endif
-  ifeq ($(MESA_ANDROID_MAJOR_VERSION),6)
-    LOCAL_CFLAGS += -DHAVE_LLVM=0x0307 -DMESA_LLVM_VERSION_PATCH=0
-    ELF_INCLUDES := external/elfutils/src/libelf
-  endif
-  ifeq ($(MESA_ANDROID_MAJOR_VERSION),7)
-    LOCAL_CFLAGS += -DHAVE_LLVM=0x0308 -DMESA_LLVM_VERSION_PATCH=0
-    ELF_INCLUDES := external/elfutils/libelf
-  endif
 endif
 
 ifneq ($(LOCAL_IS_HOST_MODULE),true)

VERSION

@@ -1 +1 @@
-17.0.7
+17.1.0-devel

bin/.cherry-ignore

@@ -1,107 +0,0 @@
-# The commit addressed an earlier commit ccdd5b3738e which did not land in branch.
-d49d275c415d60ae08dc3e52d8db11f19a44010f anv/blorp: Don't sanitize the swizzle for blorp_clear
-# The commit addressed an earlier commit 0567ab0407e which did not land in branch.
-# A stable specific backported patch will follow later ...
-bc5d587a80b64fb3e0a5ea8067e6317fbca2bbc5 radv: Invalidate L2 for TRANSFER_WRITE barriers
-# stable: the commit depends on earlier commit 656e30b6860 which did
-# 	  not land in branch, and some other dependencies.
-d2d6cf6c8387ac06c8559027f57683a61b48671b anv: Add the pci_id into the shader cache UUID
-# stable: 17.1 nomination only. Fixes earlier commit 0b613c20 which
-# 	  did not land in branch.
-d4601b0efc7f5e24e3f39fefa8e29e79560245ce freedreno: fix crash if ctx torn down with no rendering
-# stable: 17.1 nomination only. the commit depends on earlier commit
-# 	  7748c3f5eb1 which did not land in branch and the part that
-# 	  could be merged is not really providing anything relevant to
-# 	  stable.
-b81d85f1754928139f9f01474495e024946aa1b4 configure.ac: Fix typos.
-# extra: this reverts a commit in stable only because a new feature in
-# 	 master replaces it.
-e75001811e3b66986b4ede165a0fdde703d4f05b Revert "etnaviv: Cannot render to rb-swapped formats"
-# extra: the commit depends on earlier commits which did not land in
-#	 branch.
-dee31311eb024a636466e359b43d3a67b0135f32 Revert "i965/fs: Don't emit SEL instructions for type-converting MOVs."
-# fixes: finds 2 commit references but it is really only one, and only
-#        in master.
-2b3c490e23a6cd7f4f6bb489267c2b9b4fde5b20 radv: fix typo in a2b10g10r10 fast clear calculation
-# fixes: removes unused symbols but it doesn't really add nor fix
-#        anything relevant.
-9807e9dea67a60ab8ee04741c75b748eade582ee anv: remove unused anv_dispatch_table dtable
-# fixes: removes unused symbols but it doesn't really add nor fix
-#        anything relevant.
-320561bd831b81843e732d38d2b254aa3b25f2f6 radv: remove unused radv_dispatch_table dtable
-# fixes: it just makes a local function static.
-8b79f0ed0812a3a80dd91e47112fd722bd2d6978 radv: make radv_resolve_entrypoint static
-# fixes: reported only because referes to fixes in 2 commits, but they
-#        are not in stable.
-3f2cb699cfe0481f214c709b5a1375caf0a9c7fe android: vulkan: add support for libmesa_vulkan_util
-# fixes: reported only because referes to fixes in 2 commits, but they
-#        are not in stable.
-61c38d14b7b1d0ccacc440340ee0b1bc4454eca2 android: r600: fix libmesa_amd_common dependency
-# fixes: removes unused symbols but it doesn't really add nor fix
-#        anything relevant.
-b69a03e12aa38ed5d8b60a056630ed445f6097dc i965: remove dead brw_new_shader() declaration
-# fixes: it just cleans .gitignore.
-f282ace67862c0633d0a8135e4808867740d0d39 i965: remove i965_symbols_test reference from .gitignore
-# fixes: reported only because referes to fixes in 2 commits, but they
-#        are not in stable.
-15603055fb36a630cf3023219e9dd5c4a49c0596 anv: automake: ensure that the destination directory is created
-# fixes: reported only because referes to fixes in 2 commits, but they
-#        are not in stable.
-a6840efc096bd26e92bcee1a50515d845988a2f9 anv: provide required gem stubs for the tests
-# stable: 17.1 nomination only.
-65e0c3fba74ee98cacadbba4bd005b930609b65e radeonsi/gfx9: fix texture buffer objects and image buffers with IDXEN==0
-5c94779585e24e8bd1bd41707521584af4251de3 radeonsi/gfx9: fix most things wrong with shader images
-482e6b07cc6ce4b2ceac8188be19dbf252eaecde radeonsi/gfx9: fix 1D array shader images
-8e8570a9e8bae7f4d3ad623475dfadc715a828d7 radeonsi/gfx9: add a workaround for viewing a slice of 3D as a 2D image
-60a20e6879e4ce0911b12848ffd9e372f096590e radeonsi/gfx9: set MAX_PRIMGRP_IN_WAVE in the correct register
-80814819c28353a38c03d4cdba39983b8cf260ac radeonsi/gfx9: don't set deprecated field PARTIAL_ES_WAVE_ON
-f466683cb07796fa89f96ef87a6f076218ae6db8 radeonsi/gfx9: fix gl_ViewportIndex
-283a1d1e27b5456cfda848a54b9d74be0993e038 radeonsi/gfx9: make some PA & DB registers match the closed Vulkan driver
-# stable: 17.1 nomination only. Fixes earlier commit 651ec926fc1 which
-# 	  did not land in branch.
-c43b4bc85eddba8bc31665cfee5928bed8343516 anv: Don't place scratch buffers above the 32-bit boundary
-# stable: 17.1 nomination only. Disk cache doesn't feature yet in branch.
-ce412371513c90bf9156f22c3567ee57750ef264 disk_cache: reduce default cache size to 5% of filesystem
-4e1f3afea9bdeddb0d21f00d25319bce580d80c3 disk_cache: use block size rather than file size
-22fa3d90a92c1628215d0f5fccbe1116d4f5147f util/disk_cache: remove percentage based max cache limit
-# stable: 17.1 nomination only. Fixes earlier commit 68d6d097f15 which
-# 	  did not land in branch.
-69e6eab6533ff48f72223cd21ef640242c52598b winsys/amdgpu: fix Polaris12 (RX 550) breakage
-# stable: 17.1 nomination only. Contributes to earlier commit
-# 	  3f0d69af20e which did not land in branch.
-2add79a73291e40621081b9a12938ac1931b9e96 radv: apply the tess+GS hang workaround to Polaris12 as well
-# stable: 17.1 nomination only. Reverts earlier commit
-#         c5bf7cb52942cb7df9f5d73746ffbf3c102d12cc which did not land
-#         in branch.
-1456da91c8a14f77dd347981e2bc95e0644e0fd2 Revert "mesa: Require mipmap completeness for glCopyImageSubData(), sometimes."
-# stable: 17.1 nomination only. Fixes a crash triggered by earlier
-#         commit 7c3b8ed87859bfdfb985d21685115a729f9cd138 which did
-#         not land in branch.
-bc074a45180eddf30ea723bbdf89895e2c7684ca i965: Don't try to unmap NULL program cache BO.
-# stable: 17.1 nomination only. Fixes a regression caused by commit
-#         bdd644976952 which did not land in branch.
-854ed47f3e1501e4cc87bf9f19c6d4a1ad2bab08 radeonsi: mark fast-cleared textures as compressed when dirtying
-# stable: 17.1 nomination only. MSAA was enabled by 1832ef6cd9b which
-#         did not land in branch.
-f52e63069a3fad23e03d42306a42bd20f0159da3 swr: move msaa resolve to generalized StoreTile
-# stable: 17.1 nomination only. 17.0 doesn't support
-#         VK_KHR_push_descriptor, so it doesn't need this patch.
-5ff48581119258214801de24a327bdd6a29e0ccc radv/meta: fix restoring a push descriptor set
-# fixes:  Fixes earlier commit 126d5ad which did not land in branch.
-9da104593386f6e8ddec8f0d9d288aceb8908fe1 radv: fix regression in descriptor set freeing.
-# stable: Fixes earlier commit 4d30024238e which did not land in
-#         branch.
-d884d1a6540ec0f60768c30df47f0228a37ea61c vc4: Only build the NEON code on arm32.
-# stable: Fixes earlier commit 6403e376511 which did not land in
-#         branch.
-6f21b5601cc1260eac53f65c8941b3aa66d0f5e9 i965: Solve Android native fence fd double close
-# stable: Depends on earlier commit 5f99c490089 which did not land in
-#         branch.
-0549ea15ec380f3ca6df76ce53ff4c30bfc21dbf radeonsi: fix primitive ID in fragment shader when using tessellation
-# stable: Depends on earlier commit 9fd9a7d0ba3 which did not land in
-#         branch.
-b84b631c6381d9b36bca5d0e7cc67dd23af188c1 radeonsi: load patch_id for TES-as-ES when exporting for PS
-# stable: The bug that they fixed is probably more benign than a
-#         possible assertion in the X server.
-19b61799e3d06795d783b34fdbbf8474ef1e9a7c st/mesa: don't cast the incomplete framebufer to st_framebuffer
-2f0450c627e5158d49aa1320eed9a5f6cb184838 radeonsi: add new vega10 pci ids

bin/get-extra-pick-list.sh

@@ -10,28 +10,26 @@
 # $ bin/get-extra-pick-list.sh | tee picklist
 
 # Use the last branchpoint as our limit for the search
-latest_branchpoint=`git merge-base origin/master HEAD`
+# XXX: there should be a better way for this
+latest_branchpoint=`git branch | grep \* | cut -c 3-`-branchpoint
 
 # Grep for commits with "cherry picked from commit" in the commit message.
 git log --reverse --grep="cherry picked from commit" $latest_branchpoint..HEAD |\
 	grep "cherry picked from commit" |\
-	sed -e 's/^[[:space:]]*(cherry picked from commit[[:space:]]*//' -e 's/)//'  > already_picked
-
-# For each cherry-picked commit...
-cat already_picked | cut -c -8 |\
+	sed -e 's/^[[:space:]]*(cherry picked from commit[[:space:]]*//' -e 's/)//' |\
+	cut -c -8 |\
 while read sha
 do
-	# ... check if it's referenced (fixed by another) patch
+	# Check if the original commit is referenced in master
 	git log -n1 --pretty=oneline --grep=$sha $latest_branchpoint..origin/master |\
 		cut -c -8 |\
 	while read candidate
 	do
-		# And flag up if it hasn't landed in branch yet.
-		if grep -q ^$candidate already_picked ; then
-			continue
+		# Check if the potential fix, hasn't landed in branch yet.
+		found=`git log -n1 --pretty=oneline --reverse --grep=$candidate $latest_branchpoint..HEAD |wc -l`
+		if test $found = 0
+		then
+			echo Commit $candidate might need to be picked, as it references $sha
 		fi
-		echo Commit $candidate references $sha
 	done
 done
-
-rm -f already_picked

bin/get-fixes-pick-list.sh

@@ -1,61 +0,0 @@
-#!/bin/bash
-
-# Script for generating a list of candidates [referenced by a Fixes tag] for
-# cherry-picking to a stable branch
-#
-# Usage examples:
-#
-# $ bin/get-fixes-pick-list.sh
-# $ bin/get-fixes-pick-list.sh > picklist
-# $ bin/get-fixes-pick-list.sh | tee picklist
-
-# Use the last branchpoint as our limit for the search
-latest_branchpoint=`git merge-base origin/master HEAD`
-
-# List all the commits between day 1 and the branch point...
-git log --reverse --pretty=%H $latest_branchpoint > already_landed
-
-# ... and the ones cherry-picked.
-git log --reverse --grep="cherry picked from commit" $latest_branchpoint..HEAD |\
-	grep "cherry picked from commit" |\
-	sed -e 's/^[[:space:]]*(cherry picked from commit[[:space:]]*//' -e 's/)//'  > already_picked
-
-# Grep for commits with Fixes tag
-git log --reverse --pretty=%H -i --grep="fixes:" $latest_branchpoint..origin/master |\
-while read sha
-do
-	# For each one try to extract the tag
-	fixes_count=`git show $sha | grep -i "fixes:" | wc -l`
-	if [ "x$fixes_count" != x1 ] ; then
-		echo WARNING: Commit $sha has nore than one Fixes tag
-	fi
-	fixes=`git show $sha | grep -i "fixes:" | head -n 1`
-	# The following sed/cut combination is borrowed from GregKH
-	id=`echo ${fixes} | sed -e 's/^[ \t]*//' | cut -f 2 -d ':' | sed -e 's/^[ \t]*//' | cut -f 1 -d ' '`
-
-	# Bail out if we cannot find suitable id.
-	# Any specific validation the $id is valid and not some junk, is
-	# implied with the follow up code
-	if [ "x$id" = x ] ; then
-		continue
-	fi
-
-	# Check if the offending commit is in branch.
-
-	# Be that cherry-picked ...
-	# ... or landed before the branchpoint.
-	if grep -q ^$id already_picked ||
-	   grep -q ^$id already_landed ; then
-
-		# Finally nominate the fix if it hasn't landed yet.
-		if grep -q ^$sha already_picked ; then
-			continue
-		fi
-
-		echo Commit $sha fixes $id
-	fi
-
-done
-
-rm -f already_picked
-rm -f already_landed

bin/get-pick-list.sh

@@ -8,16 +8,13 @@
 # $ bin/get-pick-list.sh > picklist
 # $ bin/get-pick-list.sh | tee picklist
 
-# Use the last branchpoint as our limit for the search
-latest_branchpoint=`git merge-base origin/master HEAD`
-
 # Grep for commits with "cherry picked from commit" in the commit message.
-git log --reverse --grep="cherry picked from commit" $latest_branchpoint..HEAD |\
+git log --reverse --grep="cherry picked from commit" origin/master..HEAD |\
 	grep "cherry picked from commit" |\
 	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='^CC:.*mesa-stable' $latest_branchpoint..origin/master |\
+git log --reverse --pretty=%H -i --grep='^\([[:space:]]*NOTE: .*[Cc]andidate\|CC:.*mesa-stable\)' HEAD..origin/master |\
 while read sha
 do
 	# Check to see whether the patch is on the ignore list.

bin/get-typod-pick-list.sh

@@ -12,16 +12,13 @@
 # This script intentionally _never_ checks for specific version tag
 # Should we consider folding it with the original get-pick-list.sh
 
-# Use the last branchpoint as our limit for the search
-latest_branchpoint=`git merge-base origin/master HEAD`
-
 # Grep for commits with "cherry picked from commit" in the commit message.
-git log --reverse --grep="cherry picked from commit" $latest_branchpoint..HEAD |\
+git log --reverse --grep="cherry picked from commit" origin/master..HEAD |\
 	grep "cherry picked from commit" |\
 	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='^CC:.*mesa-dev' $latest_branchpoint..origin/master |\
+git log --reverse --pretty=%H -i --grep='^CC:.*mesa-dev' HEAD..origin/master |\
 while read sha
 do
 	# Check to see whether the patch is on the ignore list.

configure.ac

@@ -943,31 +943,43 @@ llvm_add_target() {
 # $1 is the llvm-config command with arguments.
 strip_unwanted_llvm_flags() {
     # Use \> (marks the end of the word)
-    echo " `$1`" | sed \
-        -e 's/\s\+-m\S*//g' \
-        -e 's/\s\+-DNDEBUG\>//g' \
-        -e 's/\s\+-D_GNU_SOURCE\>//g' \
-        -e 's/\s\+-pedantic\>//g' \
-        -e 's/\s\+-W\S*//g' \
-        -e 's/\s\+-O\S*//g' \
-        -e 's/\s\+-g\S*//g' \
-        -e 's/-fno-rtti\>/-Fno-rtti/g' \
-        -e 's/\s\+-f\S*//g' \
-        -e 's/-Fno-rtti\>/-fno-rtti/g' \
-        -e 's/^ //'
+    echo `$1` | sed \
+        -e 's/-march=\S*//g' \
+        -e 's/-mtune=\S*//g' \
+        -e 's/-mcpu=\S*//g' \
+        -e 's/-DNDEBUG\>//g' \
+        -e 's/-D_GNU_SOURCE\>//g' \
+        -e 's/-pedantic\>//g' \
+        -e 's/-Wcovered-switch-default\>//g' \
+        -e 's/-O.\>//g' \
+        -e 's/-g\>//g' \
+        -e 's/-Wall\>//g' \
+        -e 's/-Wcast-qual\>//g' \
+        -e 's/-Woverloaded-virtual\>//g' \
+        -e 's/-fcolor-diagnostics\>//g' \
+        -e 's/-fdata-sections\>//g' \
+        -e 's/-ffunction-sections\>//g' \
+        -e 's/-fno-exceptions\>//g' \
+        -e 's/-fomit-frame-pointer\>//g' \
+        -e 's/-fvisibility-inlines-hidden\>//g' \
+        -e 's/-fPIC\>//g' \
+        -e 's/-fstack-protector-strong\>//g'
 }
 
 llvm_set_environment_variables() {
     if test "x$LLVM_CONFIG" != xno; then
         LLVM_VERSION=`$LLVM_CONFIG --version | egrep -o '^[[0-9.]]+'`
+        LLVM_LDFLAGS=`$LLVM_CONFIG --ldflags`
         LLVM_BINDIR=`$LLVM_CONFIG --bindir`
         LLVM_CPPFLAGS=`strip_unwanted_llvm_flags "$LLVM_CONFIG --cppflags"`
+        LLVM_CFLAGS=$LLVM_CPPFLAGS   # CPPFLAGS seem to be sufficient
+        LLVM_CXXFLAGS=`strip_unwanted_llvm_flags "$LLVM_CONFIG --cxxflags"`
         LLVM_INCLUDEDIR=`$LLVM_CONFIG --includedir`
         LLVM_LIBDIR=`$LLVM_CONFIG --libdir`
 
-        # We need to respect LLVM_CPPFLAGS when compiling LLVM headers.
+        # We need to respect LLVM_CFLAGS when compiling LLVM headers.
         save_CFLAGS="$CFLAGS"
-        CFLAGS="$CFLAGS $LLVM_CPPFLAGS"
+        CFLAGS="$CFLAGS $LLVM_CFLAGS"
 
         AC_COMPUTE_INT([LLVM_VERSION_MAJOR], [LLVM_VERSION_MAJOR],
             [#include "${LLVM_INCLUDEDIR}/llvm/Config/llvm-config.h"])
@@ -989,21 +1001,19 @@ llvm_set_environment_variables() {
             LLVM_VERSION_INT="${LLVM_VERSION_MAJOR}${LLVM_VERSION_MINOR}"
         fi
 
-        FOUND_LLVM=yes
+        DEFINES="${DEFINES} -DHAVE_LLVM=0x0$LLVM_VERSION_INT -DMESA_LLVM_VERSION_PATCH=$LLVM_VERSION_PATCH"
+        MESA_LLVM=1
     else
-        FOUND_LLVM=no
+        MESA_LLVM=0
         LLVM_VERSION_INT=0
     fi
 }
 
-require_llvm() {
-    if test "x$enable_gallium_llvm" != "xyes"; then
-        AC_MSG_ERROR([--enable-gallium-llvm is required when building $1])
+llvm_check_version_for() {
+    if test "x$MESA_LLVM" = x0; then
+        AC_MSG_ERROR([LLVM $1 or newer is required for $2])
+        return
     fi
-}
-
-llvm_require_version() {
-    require_llvm $2
 
     llvm_target_version_major=`echo $1 | cut -d. -f1 | egrep -o '^[[0-9]]+'`
     llvm_target_version_minor=`echo $1 | cut -d. -f2 | egrep -o '^[[0-9]]+'`
@@ -1044,7 +1054,7 @@ radeon_llvm_check() {
         amdgpu_llvm_target_name='amdgpu'
     fi
 
-    llvm_require_version $*
+    llvm_check_version_for $*
 
     llvm_add_target $amdgpu_llvm_target_name $2
 
@@ -1721,34 +1731,6 @@ if test -n "$with_dri_drivers"; then
 fi
 
 
-dnl
-dnl Gallium LLVM
-dnl
-dnl With follow-up commits we'll rework --enable-gallium-llvm to --enable-llvm
-dnl Since that is too invasive to stable, do the more conservative thing for now
-dnl and consider it as a global LLVM toggle.
-dnl
-AC_ARG_ENABLE([gallium-llvm],
-    [AS_HELP_STRING([--enable-gallium-llvm],
-        [build gallium LLVM support @<:@default=enabled on x86/x86_64@:>@])],
-    [enable_gallium_llvm="$enableval"],
-    [enable_gallium_llvm=auto])
-
-if test "x$enable_gallium_llvm" = xauto; then
-    if test "x$FOUND_LLVM" = xyes; then
-        case "$host_cpu" in
-        i*86|x86_64|amd64) enable_gallium_llvm=yes;;
-        *) enable_gallium_llvm=no;;
-        esac
-    else
-        enable_gallium_llvm=no
-    fi
-fi
-
-if test "x$enable_gallium_llvm" = xyes -a "x$FOUND_LLVM" = xno; then
-    AC_MSG_ERROR([--enable-gallium-llvm selected but llvm-config is not found])
-fi
-
 #
 # Vulkan driver configuration
 #
@@ -1756,7 +1738,7 @@ fi
 AC_ARG_WITH([vulkan-drivers],
     [AS_HELP_STRING([--with-vulkan-drivers@<:@=DIRS...@:>@],
         [comma delimited Vulkan drivers list, e.g.
-        "intel,radeon"
+        "intel"
         @<:@default=no@:>@])],
     [with_vulkan_drivers="$withval"],
     [with_vulkan_drivers="no"])
@@ -2030,21 +2012,7 @@ if test "x$enable_opencl" = xyes; then
        AC_MSG_ERROR([Clover requires libelf])
     fi
 
-    if test "x${ac_cv_cxx_compiler_gnu}" = xyes; then
-        altivec_enabled=no
-        AC_COMPILE_IFELSE([AC_LANG_SOURCE([
-        #if !defined(__VEC__) || !defined(__ALTIVEC__)
-        #error "AltiVec not enabled"
-        #endif
-        ])], altivec_enabled=yes)
-
-        if test "$altivec_enabled" = yes; then
-            CLOVER_STD_OVERRIDE="-std=gnu++11"
-        fi
-        AC_SUBST([CLOVER_STD_OVERRIDE])
-    fi
-
-    llvm_require_version $LLVM_REQUIRED_OPENCL "opencl"
+    llvm_check_version_for $LLVM_REQUIRED_OPENCL "opencl"
 
     llvm_add_default_components "opencl"
     llvm_add_component "all-targets" "opencl"
@@ -2169,6 +2137,24 @@ if ! echo "$egl_platforms" | grep -q 'x11'; then
     GL_PC_CFLAGS="$GL_PC_CFLAGS -DMESA_EGL_NO_X11_HEADERS"
 fi
 
+dnl
+dnl Gallium LLVM
+dnl
+AC_ARG_ENABLE([gallium-llvm],
+    [AS_HELP_STRING([--enable-gallium-llvm],
+        [build gallium LLVM support @<:@default=enabled on x86/x86_64@:>@])],
+    [enable_gallium_llvm="$enableval"],
+    [enable_gallium_llvm=auto])
+
+if test -z "$with_gallium_drivers"; then
+    enable_gallium_llvm=no
+fi
+if test "x$enable_gallium_llvm" = xauto; then
+    case "$host_cpu" in
+    i*86|x86_64|amd64) enable_gallium_llvm=yes;;
+    esac
+fi
+
 dnl Directory for XVMC libs
 AC_ARG_WITH([xvmc-libdir],
     [AS_HELP_STRING([--with-xvmc-libdir=DIR],
@@ -2216,16 +2202,14 @@ AC_ARG_WITH([d3d-libdir],
 AC_SUBST([D3D_DRIVER_INSTALL_DIR])
 
 dnl
-dnl r300 doesn't strictly require LLVM, but for performance reasons we
-dnl highly recommend LLVM usage. So require it at least on x86 and x86_64
-dnl architectures.
+dnl Gallium helper functions
 dnl
-r300_require_llvm() {
-    case "$host" in *gnux32) return;; esac
-    case "$host_cpu" in
-    i*86|x86_64|amd64) require_llvm $1
-        ;;
-    esac
+gallium_require_llvm() {
+    if test "x$enable_gallium_llvm" = "xyes"; then
+        llvm_check_version_for $LLVM_REQUIRED_GALLIUM "gallium"
+    else
+        AC_MSG_ERROR([--enable-gallium-llvm is required when building $1])
+    fi
 }
 
 dnl
@@ -2314,7 +2298,7 @@ if test -n "$with_gallium_drivers"; then
             HAVE_GALLIUM_R300=yes
             PKG_CHECK_MODULES([RADEON], [libdrm_radeon >= $LIBDRM_RADEON_REQUIRED])
             require_libdrm "r300"
-            r300_require_llvm "r300"
+            gallium_require_llvm "r300"
             ;;
         xr600)
             HAVE_GALLIUM_R600=yes
@@ -2333,9 +2317,7 @@ if test -n "$with_gallium_drivers"; then
             PKG_CHECK_MODULES([AMDGPU], [libdrm_amdgpu >= $LIBDRM_AMDGPU_REQUIRED])
             require_libdrm "radeonsi"
             radeon_gallium_llvm_check $LLVM_REQUIRED_RADEONSI "radeonsi"
-            if test "x$enable_egl" = xyes; then
-                require_basic_egl "radeonsi"
-            fi
+            require_basic_egl "radeonsi"
             ;;
         xnouveau)
             HAVE_GALLIUM_NOUVEAU=yes
@@ -2357,12 +2339,13 @@ if test -n "$with_gallium_drivers"; then
             ;;
         xswrast)
             HAVE_GALLIUM_SOFTPIPE=yes
-            if test "x$enable_gallium_llvm" = xyes; then
+            if test "x$MESA_LLVM" = x1 && test "x$enable_gallium_llvm" == "xyes";  then
                 HAVE_GALLIUM_LLVMPIPE=yes
             fi
             ;;
         xswr)
-            llvm_require_version $LLVM_REQUIRED_SWR "swr"
+            llvm_check_version_for $LLVM_REQUIRED_SWR "swr"
+            gallium_require_llvm "swr"
 
             swr_require_cxx_feature_flags "C++11" "__cplusplus >= 201103L" \
                 ",-std=c++11" \
@@ -2394,9 +2377,7 @@ if test -n "$with_gallium_drivers"; then
         xvirgl)
             HAVE_GALLIUM_VIRGL=yes
             require_libdrm "virgl"
-            if test "x$enable_egl" = xyes; then
-                require_basic_egl "virgl"
-            fi
+            require_basic_egl "virgl"
             ;;
         *)
             AC_MSG_ERROR([Unknown Gallium driver: $driver])
@@ -2405,8 +2386,8 @@ if test -n "$with_gallium_drivers"; then
     done
 fi
 
-if test "x$enable_gallium_llvm" == "xyes" -a "$with_gallium_drivers"; then
-    llvm_require_version $LLVM_REQUIRED_GALLIUM "gallium"
+if test "x$enable_gallium_llvm" == "xyes"; then
+    llvm_check_version_for $LLVM_REQUIRED_GALLIUM "gallium"
     llvm_add_default_components "gallium"
 fi
 
@@ -2416,22 +2397,14 @@ if test "x$HAVE_GALLIUM_ETNAVIV" != xyes -a "x$HAVE_GALLIUM_IMX" == xyes  ; then
     AC_ERROR([Building with imx requires etnaviv])
 fi
 
-dnl
-dnl Set defines and buildtime variables only when using LLVM.
-dnl
-if test "x$enable_gallium_llvm" = xyes; then
-    DEFINES="${DEFINES} -DHAVE_LLVM=0x0$LLVM_VERSION_INT -DMESA_LLVM_VERSION_PATCH=$LLVM_VERSION_PATCH"
+dnl Set LLVM_LIBS - This is done after the driver configuration so
+dnl that drivers can add additional components to LLVM_COMPONENTS.
+dnl Previously, gallium drivers were updating LLVM_LIBS directly
+dnl by calling llvm-config --libs ${DRIVER_LLVM_COMPONENTS}, but
+dnl this was causing the same libraries to be appear multiple times
+dnl in LLVM_LIBS.
 
-    LLVM_LDFLAGS=`$LLVM_CONFIG --ldflags`
-    LLVM_CFLAGS=$LLVM_CPPFLAGS   # CPPFLAGS seem to be sufficient
-    LLVM_CXXFLAGS=`strip_unwanted_llvm_flags "$LLVM_CONFIG --cxxflags"`
-
-    dnl Set LLVM_LIBS - This is done after the driver configuration so
-    dnl that drivers can add additional components to LLVM_COMPONENTS.
-    dnl Previously, gallium drivers were updating LLVM_LIBS directly
-    dnl by calling llvm-config --libs ${DRIVER_LLVM_COMPONENTS}, but
-    dnl this was causing the same libraries to be appear multiple times
-    dnl in LLVM_LIBS.
+if test "x$MESA_LLVM" != x0; then
 
     if ! $LLVM_CONFIG --libs ${LLVM_COMPONENTS} >/dev/null; then
        AC_MSG_ERROR([Calling ${LLVM_CONFIG} failed])
@@ -2533,7 +2506,8 @@ AM_CONDITIONAL(NEED_RADEON_DRM_WINSYS, test "x$HAVE_GALLIUM_R300" = xyes -o \
 AM_CONDITIONAL(NEED_WINSYS_XLIB, test "x$enable_glx" = xgallium-xlib)
 AM_CONDITIONAL(NEED_RADEON_LLVM, test x$NEED_RADEON_LLVM = xyes)
 AM_CONDITIONAL(HAVE_GALLIUM_COMPUTE, test x$enable_opencl = xyes)
-AM_CONDITIONAL(HAVE_GALLIUM_LLVM, test "x$enable_gallium_llvm" = xyes)
+AM_CONDITIONAL(HAVE_GALLIUM_LLVM, test "x$MESA_LLVM" = x1 -a \
+                                       "x$enable_gallium_llvm" = xyes)
 AM_CONDITIONAL(USE_VC4_SIMULATOR, test x$USE_VC4_SIMULATOR = xyes)
 if test "x$USE_VC4_SIMULATOR" = xyes -a "x$HAVE_GALLIUM_ILO" = xyes; then
     AC_MSG_ERROR([VC4 simulator on x86 replaces i965 driver build, so ilo must be disabled.])
@@ -2818,7 +2792,7 @@ else
 fi
 
 echo ""
-if test "x$enable_gallium_llvm" = xyes; then
+if test "x$MESA_LLVM" = x1; then
     echo "        llvm:            yes"
     echo "        llvm-config:     $LLVM_CONFIG"
     echo "        llvm-version:    $LLVM_VERSION"
@@ -2865,7 +2839,7 @@ echo "        CFLAGS:          $cflags"
 echo "        CXXFLAGS:        $cxxflags"
 echo "        Macros:          $defines"
 echo ""
-if test "x$enable_gallium_llvm" = xyes; then
+if test "x$MESA_LLVM" = x1; then
     echo "        LLVM_CFLAGS:     $LLVM_CFLAGS"
     echo "        LLVM_CXXFLAGS:   $LLVM_CXXFLAGS"
     echo "        LLVM_CPPFLAGS:   $LLVM_CPPFLAGS"

docs/index.html

@@ -16,6 +16,17 @@
 
 <h1>News</h1>
 
+<h2>January 23, 2017</h2>
+<p>
+<a href="relnotes/12.0.6.html">Mesa 12.0.6</a> is released.
+This is a bug-fix release.
+<br>
+NOTE: This is an extra release for the 12.0 stable branch, as per developers'
+feedback. It is anticipated that 12.0.6 will be the final release in the 12.0
+series. Users of 12.0 are encouraged to migrate to the 13.0 series in order
+to obtain future fixes.
+</p>
+
 <h2>January 5, 2017</h2>
 <p>
 <a href="relnotes/13.0.3.html">Mesa 13.0.3</a> is released.

docs/releasing.html

@@ -158,6 +158,9 @@ To setup the branchpoint:
 	git checkout master # make sure we're in master first
 	git tag -s X.Y-branchpoint -m "Mesa X.Y branchpoint"
 	git checkout -b X.Y
+	git checkout master
+	$EDITOR VERSION # bump the version number
+	git commit -as
 	git push origin X.Y-branchpoint X.Y
 </pre>
 
@@ -342,7 +345,7 @@ Here is one solution that I've been using.
 	tar -xaf mesa-$__version.tar.xz &amp;&amp; cd mesa-$__version &amp;&amp; scons &amp;&amp; cd ..
 
 	# Test the automake binaries
-	rm -rf cd mesa-$__version
+	rm -rf mesa-$__version
 	tar -xaf mesa-$__version.tar.xz &amp;&amp; cd mesa-$__version
 	./configure \
 		--with-dri-drivers=i965,swrast \

docs/relnotes.html

@@ -21,6 +21,7 @@ The release notes summarize what's new or changed in each Mesa release.
 </p>
 
 <ul>
+<li><a href="relnotes/12.0.6.html">12.0.6 release notes</a>
 <li><a href="relnotes/13.0.3.html">13.0.3 release notes</a>
 <li><a href="relnotes/12.0.5.html">12.0.5 release notes</a>
 <li><a href="relnotes/13.0.2.html">13.0.2 release notes</a>

docs/relnotes/12.0.6.html

@@ -0,0 +1,148 @@
+<!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 12.0.6 Release Notes / January 23, 2017</h1>
+
+<p>
+Mesa 12.0.6 is a bug fix release which fixes bugs found since the 12.0.5 release.
+</p>
+<p>
+Mesa 12.0.6 implements the OpenGL 4.3 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 4.3.  OpenGL
+4.3 is <strong>only</strong> available if requested at context creation
+because compatibility contexts are not supported.
+</p>
+
+
+<h2>SHA256 checksums</h2>
+<pre>
+65339ba5d76a45225b8b56f9a1da9db15c569e1d163760faa2921da0a8461741  mesa-12.0.6.tar.gz
+7d6da9744c1022a4c2ab6ad01a206984d00443fb691568011d01b3dd97e36448  mesa-12.0.6.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=92234">Bug 92234</a> - [BDW] GPU hang in Shogun2</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=95130">Bug 95130</a> - Derivatives of gl_Color wrong when helper pixels used</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98329">Bug 98329</a> - [dEQP, EGL, SKL, BDW, BSW] dEQP-EGL.functional.image.render_multiple_contexts.gles2_renderbuffer_depth16_depth_buffer</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99030">Bug 99030</a> - [HSW, regression] transform feedback fails on Linux 4.8</li>
+
+<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99354">Bug 99354</a> - [G71] &quot;Assertion `bkref' failed&quot; reproducible with glmark2</li>
+
+</ul>
+
+<h2>Changes</h2>
+
+<p>Chad Versace (3):</p>
+<ul>
+  <li>i965/mt: Disable aux surfaces after making miptree shareable</li>
+  <li>i965/mt: Disable HiZ when sharing depth buffer externally (v2)</li>
+  <li>anv: Handle vkGetPhysicalDeviceQueueFamilyProperties with count == 0</li>
+</ul>
+
+<p>Emil Velikov (5):</p>
+<ul>
+  <li>docs: add sha256 checksums for 12.0.5</li>
+  <li>get-typod-pick-list.sh: add new script</li>
+  <li>automake: use shared llvm libs for make distcheck</li>
+  <li>egl/wayland: use the destroy_window_callback for swrast</li>
+  <li>Update version to 12.0.6</li>
+</ul>
+
+<p>Fredrik Höglund (1):</p>
+<ul>
+  <li>dri3: Fix MakeCurrent without a default framebuffer</li>
+</ul>
+
+<p>Ilia Mirkin (1):</p>
+<ul>
+  <li>nouveau: take extra push space into account for pushbuf_space calls</li>
+</ul>
+
+<p>Jason Ekstrand (19):</p>
+<ul>
+  <li>spirv/nir: Fix some texture opcode asserts</li>
+  <li>spirv/nir: Add support for shadow samplers that return vec4</li>
+  <li>spirv/nir: Properly handle gather components</li>
+  <li>anv/pipeline: Set binding_table.gather_texture_start</li>
+  <li>nir: Add a helper for determining the type of a texture source</li>
+  <li>nir/lower_tex: Add some helpers for working with tex sources</li>
+  <li>nir/lower_tex: Add support for lowering coordinate offsets</li>
+  <li>i965/nir: Enable NIR lowering of txf and rect offsets</li>
+  <li>i965: Get rid of the do_lower_unnormalized_offsets pass</li>
+  <li>spirv/nir: Don't increment coord_components for array lod queries</li>
+  <li>anv/image: Assert that the image format is actually supported</li>
+  <li>spirv/nir: Move opcode selection higher up in handle_texture</li>
+  <li>spirv/nir: Refactor type handling in handle_texture</li>
+  <li>nir/spirv: Refactor coordinate handling in handle_texture</li>
+  <li>spirv/nir: Handle texture projectors</li>
+  <li>spirv/nir: Add support for ImageQuerySamples</li>
+  <li>anv/device: Return the right error for failed maps</li>
+  <li>anv/device: Implicitly unmap memory objects in FreeMemory</li>
+  <li>anv/descriptor_set: Write the state offset in the surface state free list.</li>
+</ul>
+
+<p>Kenneth Graunke (2):</p>
+<ul>
+  <li>spirv: Move cursor before calling vtn_ssa_value() in phi 2nd pass.</li>
+  <li>i965: Properly flush in hsw_pause_transform_feedback().</li>
+</ul>
+
+<p>Marek Olšák (6):</p>
+<ul>
+  <li>cso: don't release sampler states that are bound</li>
+  <li>radeonsi: always restore sampler states when unbinding sampler views</li>
+  <li>radeonsi: fix incorrect FMASK checking in bind_sampler_states</li>
+  <li>radeonsi: disable CE on SI + AMDGPU</li>
+  <li>radeonsi: disable the constant engine (CE) on Carrizo and Stoney</li>
+  <li>gallium/radeon: fix the draw-calls HUD query</li>
+</ul>
+
+<p>Matt Turner (3):</p>
+<ul>
+  <li>i965/fs: Rename opt_copy_propagate -&gt; opt_copy_propagation.</li>
+  <li>i965/fs: Add unit tests for copy propagation pass.</li>
+  <li>i965/fs: Reject copy propagation into SEL if not min/max.</li>
+</ul>
+
+<p>Michel Dänzer (1):</p>
+<ul>
+  <li>cso: Don't restore nr_samplers in cso_restore_fragment_samplers</li>
+</ul>
+
+<p>Nicolai Hähnle (1):</p>
+<ul>
+  <li>radeonsi: enable WQM in PS prolog when needed</li>
+</ul>
+
+
+</div>
+</body>
+</html>

docs/relnotes/17.0.0.html

@@ -14,7 +14,7 @@
 <iframe src="../contents.html"></iframe>
 <div class="content">
 
-<h1>Mesa 17.0.0 Release Notes / February 13, 2017</h1>
+<h1>Mesa 17.0.0 Release Notes / TBD</h1>
 
 <p>
 Mesa 17.0.0 is a new development release.
@@ -33,8 +33,7 @@ because compatibility contexts are not supported.
 
 <h2>SHA256 checksums</h2>
 <pre>
-696578f0b83796470511a88a95fff15a2a25fa201a9e487716f2ca20c177c3ab  mesa-17.0.0.tar.gz
-39db3d59700159add7f977307d12a7dfe016363e760ad82280ac4168ea668481  mesa-17.0.0.tar.xz
+TBD.
 </pre>
 
 
@@ -63,222 +62,13 @@ Note: some of the new features are only available with certain drivers.
 
 <ul>
 
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=70623">Bug 70623</a> - libglx.so: undefined symbol: _glapi_tls_Context</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=72902">Bug 72902</a> - [IVB/HSW/BDW] DOTA2 segfaults unless Mesa is configured with (non-default) --enable-glx-tls</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=73778">Bug 73778</a> - _glapi_tls_Dispatch undefined</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=77662">Bug 77662</a> - Fail to render to different faces of depth-stencil cube map</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=89043">Bug 89043</a> - undefined symbol: _glapi_tls_Dispatch</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=91281">Bug 91281</a> - Tonga VCE 2160p encode fails with  BO to small for addr</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=92234">Bug 92234</a> - [BDW] GPU hang in Shogun2</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=92634">Bug 92634</a> - gallium's vl_mpeg12_decoder does not work with st/va</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=92760">Bug 92760</a> - Add FP64 support to the i965 shader backends</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=92925">Bug 92925</a> - Incorrect GEN for ASTC in Surface Format Table</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=93551">Bug 93551</a> - Divinity: Original Sin Enhanced Edition(Native) crash on start</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=94512">Bug 94512</a> - X segfaults with glx-tls enabled in a x32 environment</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=94900">Bug 94900</a> - HD6950 GPU lockup loop with various steam games (octodad[always], saints row 4[always], dead island[always], grid autosport[sometimes])</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=94904">Bug 94904</a> - [vulkan, BSW] dEQP-VK.api.object_management.multithreaded_per_thread_device intermittent crash</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=95460">Bug 95460</a> - Please add more drivers (freedreno, virgl) to features.txt status document</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=96959">Bug 96959</a> - nop.sat generated by pow workaround?</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=97102">Bug 97102</a> - [dri][swr] stack overflow / infinite loop with GALLIUM_DRIVER=swr</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=97232">Bug 97232</a> - Line rendering broken in Dolphin when using gl_ClipDistance</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=97287">Bug 97287</a> - GL45-CTS.vertex_attrib_binding.basic-inputL-case1 fails</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=97321">Bug 97321</a> - Query INFO_LOG_LENGTH for empty info log should return 0</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=97420">Bug 97420</a> - &quot;#version 0&quot; crashes glsl_compiler</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=97422">Bug 97422</a> - trying to call a number as a function results into a crash</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=97447">Bug 97447</a> - GL 3.0 compatibility context exposes GL_ARB_compute_shader</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=97473">Bug 97473</a> - Memory corruption when uploading DXT5 cubemap faces</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=97715">Bug 97715</a> - [ILK,G45,G965] piglit.spec.arb_separate_shader_objects.misc api error checks</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=97779">Bug 97779</a> - [regression, bisected][BDW, GPU hang] stuck on render ring, always reproducible</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=97804">Bug 97804</a> - Later precision statement isn't overriding earlier one</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=97952">Bug 97952</a> - /usr/include/string.h:518:12: error: exception specification in declaration does not match previous declaration</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=97967">Bug 97967</a> - glsl/tests/cache-test regression</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98005">Bug 98005</a> - VCE dual instance encoding inconsistent since st/va: enable dual instances encode by sync surface</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98012">Bug 98012</a> - [IVB] Segfault when running Dolphin twice with Vulkan</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98134">Bug 98134</a> - dEQP-GLES31.functional.debug.negative_coverage.get_error.buffer.draw_buffers wants a different GL error code</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98172">Bug 98172</a> - Concurrent call to glClientWaitSync results in segfault in one of the waiters.</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98238">Bug 98238</a> - witcher 2: objects are black when changing lod</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98243">Bug 98243</a> - dEQP mismatched UBO precision qualifiers</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98245">Bug 98245</a> - GLES3.1 link negative dEQP &quot;expected linking to fail, but passed.&quot;</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98250">Bug 98250</a> - dEQP-GLES31.functional.debug.negative_coverage.get_error.texture.texparameterIiv/texparameterIuiv failure</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98263">Bug 98263</a> - [radv] The Talos Principle fails to launch with &quot;Fatal error: Cannot set display mode.&quot;</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98297">Bug 98297</a> - Can't configure a desktop with 3x4k monitors in one row</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98299">Bug 98299</a> - Compute shaders generate stupid divides</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98307">Bug 98307</a> - &quot;st/glsl_to_tgsi: explicitly track all input and output declaration&quot; broke flightgear colors on rs780</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98326">Bug 98326</a> - [dEQP, EGL] pbuffer depth/stencil tests fail</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98327">Bug 98327</a> - [dEQP, EGL] dEQP-EGL.functional.resize not supported</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98328">Bug 98328</a> - [dEQP, EGL] luminance tests fail</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98329">Bug 98329</a> - [dEQP, EGL, SKL, BDW, BSW] dEQP-EGL.functional.image.render_multiple_contexts.gles2_renderbuffer_depth16_depth_buffer</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98330">Bug 98330</a> - [dEQP, EGL] dEQP-EGL.functional.buffer_age.no_preserve fails</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98339">Bug 98339</a> - dEQP-EGL: Got EGL_BAD_MATCH: eglCreateSyncKHR()</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98343">Bug 98343</a> - dEQP-EGL: GL_INVALID_ENUM at teglCreateContextExtTests</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98415">Bug 98415</a> - Vulkan Driver JSON file contains incorrect field</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98421">Bug 98421</a> - src/loader/loader.c:111:40: error: unknown type name ‘drmDevicePtr’</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98431">Bug 98431</a> - UnrealEngine v4 demos startup fails to blorp blit assert</li>
-
 <li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98480">Bug 98480</a> - Support R8 image texture in ES 3.1</li>
 
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98512">Bug 98512</a> - radeon r600 vdpau: Invalid command stream: texture bo too small</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98518">Bug 98518</a> - [r600g, bisected] regression: NI/Turks MSAA texture corruption with FreeCAD and Wine games</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98526">Bug 98526</a> - glsl/tests/general-ir-test regression</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98595">Bug 98595</a> - glsl: ralloc assertion &quot;info-&gt;canary == CANARY&quot; failed</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98599">Bug 98599</a> - xterm menus corrupt since tgsi/scan: handle indirect image indexing correctly</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98632">Bug 98632</a> - Fix build on Hurd without PATH_MAX</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98681">Bug 98681</a> - ir_builder_print_visitor.cpp:401:67: error: expected ')' before 'PRIx64'</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98694">Bug 98694</a> - &quot;(5=2)?1:1&quot; as array size decleration crashes glsl_compiler</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98740">Bug 98740</a> - bitcode.cpp:102:8: error: ‘Error’ is not a member of ‘llvm’</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98767">Bug 98767</a> - [swrast] ralloc.c:84: get_header: Assertion `info-&gt;canary == CANARY' failed.</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98774">Bug 98774</a> - glsl/tests/warnings-test regression</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98815">Bug 98815</a> - [SKL/BDW GT2] large perf regression in TessMark</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98840">Bug 98840</a> - nir clone test fails</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98893">Bug 98893</a> - [SKL] piglit.spec.arb_shader_image_load_store.semantics intermittent</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98914">Bug 98914</a> - mesa-vdpau-drivers: breaks vdpau for mpeg2video</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98917">Bug 98917</a> - [BDW SKL BSW KBL] Tessellation CTS tests regression</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98975">Bug 98975</a> - Wasteland 2 Directors Cut: Hangs. GPU fault</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99010">Bug 99010</a> - --disable-gallium-llvm no longer recognized</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99013">Bug 99013</a> - [regression, bisected] radeonsi: commit 4c8c13b3  &quot;Use amdgcn intrinsics for fs interpolation&quot; makes system unusable</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99030">Bug 99030</a> - [HSW, regression] transform feedback fails on Linux 4.8</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99038">Bug 99038</a> - [dEQP, EGL, SKL, BDW, BSW] dEQP-EGL.functional.negative_api.create_pixmap_surface crashes</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99072">Bug 99072</a> - [byt,ivb,snb] ES3-CTS.gtf.GL3Tests.shadow regression</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99085">Bug 99085</a> - [EGL] dEQP-EGL.functional.sharing.gles2.multithread intermittent</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99097">Bug 99097</a> - [vulkancts] dEQP-VK.image.store regression</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99100">Bug 99100</a> - [SKL,BDW,BSW,KBL] dEQP-VK.glsl.return.return_in_dynamic_loop_dynamic_vertex regression</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99119">Bug 99119</a> - swr_fence_work.cpp(42): error: argument of type &quot;std::nullptr_t&quot; is incompatible with parameter of type &quot;unsigned long&quot;</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99144">Bug 99144</a> - Incorrect rendering using glDrawArraysInstancedBaseInstance and first != 0 on Skylake</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99154">Bug 99154</a> - Link time error when using multiple builtin functions</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99158">Bug 99158</a> - vdpau segfaults and gpu locks with kodi on R9285</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99185">Bug 99185</a> - dEQP-EGL.functional.image.modify.tex_rgb5_a1_tex_subimage_rgba8</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99188">Bug 99188</a> - dEQP-EGL.functional.create_context_ext.robust_gl_30.rgb565_no_depth_no_stencil</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99210">Bug 99210</a> - ES3-CTS.functional.texture.mipmap.cube.generate.rgba5551_*</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99214">Bug 99214</a> - Crash in library libswrAVX.so when assigning vertex buffer object pointers with elements of type GL_DOUBLE</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99219">Bug 99219</a> - The Stanley Parable GPU hang when starting a new game</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99229">Bug 99229</a> - [G33] thousands of tests crash</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99231">Bug 99231</a> - [HSW][i965] Crash in upload_3dstate_streamout()</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99287">Bug 99287</a> - piglit.spec.glsl-1_10.execution.vs-nested-return-sibling-loop regression</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99303">Bug 99303</a> - [REGRESSION][BISECTED] DMs are crashing on start with &quot;radeon&quot;</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99314">Bug 99314</a> - [g33] glsl regressions</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99339">Bug 99339</a> - Blender line rendering broken after removing XY clipping of lines</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99354">Bug 99354</a> - [G71] &quot;Assertion `bkref' failed&quot; reproducible with glmark2</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99389">Bug 99389</a> - Mesa build broken: sid_tables.h</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99391">Bug 99391</a> - [ILK,G45,G965] piglit regressions</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99401">Bug 99401</a> - [g33] regression: piglit.spec.!opengl 1_0.gl-1_0-beginend-coverage</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99419">Bug 99419</a> - Crash(Segmentation fault) si_shader_select in Master Of Orion</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99450">Bug 99450</a> - [amdgpu] Payday 2 visual glitches on some models</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99451">Bug 99451</a> - polygon offset use after free</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99456">Bug 99456</a> - Firefox crashing when opening about:support with WebGL2 enabled</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99631">Bug 99631</a> - segfault with OSVRTrackerView and openscenegraph git master</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99633">Bug 99633</a> - rasterizer/core/clip.h:279:49: error: ‘const struct API_STATE’ has no member named ‘linkageCount’</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99637">Bug 99637</a> - VLC video has corrupted colors when using VDPAU output on Radeon SI</li>
-
 </ul>
 
 <h2>Changes</h2>
 
-<ul>
-<li>Building RADV requires --enable-gallium-llvm</li>
-<li>The vulkan headers vk_platform.h and vulkan.h are no longer installed</li>
-<li>The configure options --with-sha1 and --disable-shader-cache are
-removed alongside their respective library requirements</li>
-</ul>
+TBD.
 
 </div>
 </body>

docs/relnotes/17.0.1.html

@@ -1,221 +0,0 @@
-<!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 17.0.1 Release Notes / March 4, 2017</h1>
-
-<p>
-Mesa 17.0.1 is a bug fix release which fixes bugs found since the 17.0.0 release.
-</p>
-<p>
-Mesa 17.0.1 implements the OpenGL 4.5 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 4.5.  OpenGL
-4.5 is <strong>only</strong> available if requested at context creation
-because compatibility contexts are not supported.
-</p>
-
-
-<h2>SHA256 checksums</h2>
-<pre>
-e819bd3e515dac26faf9836d8f27a4ddf05323b9b23afb6c06536d4ac82e2743  mesa-17.0.1.tar.gz
-96fd70ef5f31d276a17e424e7e1bb79447ccbbe822b56844213ef932e7ad1b0c  mesa-17.0.1.tar.xz
-</pre>
-
-
-<h2>New features</h2>
-<p>None</p>
-
-
-<h2>Bug fixes</h2>
-
-<ul>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98869">Bug 98869</a> - Electronic Super Joy graphic artefacts (regression,bisected)</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99532">Bug 99532</a> - Compute shader doesn't give right result under some circumstances</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99677">Bug 99677</a> - heap-use-after-free in glsl</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99692">Bug 99692</a> - [radv] Mostly broken on Hawaii PRO/CIK ASICs</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99850">Bug 99850</a> - Tessellation bug on Carrizo</li>
-
-</ul>
-
-
-<h2>Changes</h2>
-
-<p>Bas Nieuwenhuizen (4):</p>
-<ul>
-  <li>radv: Never try to create more than max_sets descriptor sets.</li>
-  <li>radv: Reset emitted compute pipeline when calling secondary cmd buffer.</li>
-  <li>radv: Only use PKT3_OCCLUSION_QUERY when it doesn't hang.</li>
-  <li>radv: Use correct size for availability flag.</li>
-</ul>
-
-<p>Ben Crocker (3):</p>
-<ul>
-  <li>gallivm: Reenable PPC VSX (v3)</li>
-  <li>gallivm: Improve debug output (V2)</li>
-  <li>gallivm: Override getHostCPUName() "generic" w/ "pwr8" (v4)</li>
-</ul>
-
-<p>Brendan King (1):</p>
-<ul>
-  <li>egl/dri3: implement query surface hook</li>
-</ul>
-
-<p>Christian Gmeiner (2):</p>
-<ul>
-  <li>etnaviv: move pctx initialisation to avoid a null dereference</li>
-  <li>etnaviv: remove number of pixel pipes validation</li>
-</ul>
-
-<p>Connor Abbott (1):</p>
-<ul>
-  <li>anv: fix Get*MemoryRequirements for !LLC</li>
-</ul>
-
-<p>Daniel Stone (1):</p>
-<ul>
-  <li>egl/wayland: Don't use DRM format codes for SHM</li>
-</ul>
-
-<p>Dave Airlie (6):</p>
-<ul>
-  <li>tgsi: fix memory leak in tgsi sanity check</li>
-  <li>radv: change base aligmment for allocated memory.</li>
-  <li>radv: fix cik macroModeIndex.</li>
-  <li>radv: adopt some init config workarounds from radeonsi.</li>
-  <li>radv: fix depth format in blit2d.</li>
-  <li>radv: fix txs for sampler buffers</li>
-</ul>
-
-<p>Emil Velikov (8):</p>
-<ul>
-  <li>docs: add sha256 checksums for 17.0.0</li>
-  <li>bin/get-extra-pick-list: use git merge-base to get the branchpoint</li>
-  <li>bin/get-extra-pick-list: rework to use already_picked list</li>
-  <li>bin/get-typod-pick-list.sh: limit `git grep ...' to only as needed</li>
-  <li>bin/get-pick-list.sh: limit `git grep ...' only as needed</li>
-  <li>bin/get-pick-list.sh: remove ancient way of nominating patches</li>
-  <li>bin/get-fixes-pick-list.sh: add new script</li>
-  <li>Update version to 17.0.1</li>
-</ul>
-
-<p>Eric Anholt (1):</p>
-<ul>
-  <li>vc4: Avoid emitting small immediates for UBO indirect load address guards.</li>
-</ul>
-
-<p>Grazvydas Ignotas (3):</p>
-<ul>
-  <li>r300g: only allow byteswapped formats on big endian</li>
-  <li>gallium/u_queue: fix a crash with atexit handlers</li>
-  <li>gallium/u_queue: set num_threads correctly if not all threads start</li>
-</ul>
-
-<p>Hans de Goede (1):</p>
-<ul>
-  <li>glx/glvnd: Fix GLXdispatchIndex sorting</li>
-</ul>
-
-<p>Ilia Mirkin (4):</p>
-<ul>
-  <li>gm107/ir: fix address offset bitfield for ATOMS</li>
-  <li>nvc0: set the render condition in the compute object</li>
-  <li>st/mesa: don't pass compare mode for stencil-sampled textures</li>
-  <li>nvc0: disable linked tsc mode in compute launch descriptor</li>
-</ul>
-
-<p>Jason Ekstrand (10):</p>
-<ul>
-  <li>i965/sampler_state: Clamp min/max LOD to 14 on gen7+</li>
-  <li>i965/sampler_state: Pass texObj into update_sampler_state</li>
-  <li>i965/sampler_state: Set the "Base Mip Level" field on Sandy Bridge</li>
-  <li>intel/blorp: Swizzle clear colors on the CPU</li>
-  <li>i965/fs: Fix the inline nir_op_pack_double optimization</li>
-  <li>anv: Add an invalidate_range helper</li>
-  <li>anv/query: clflush the bo map on non-LLC platforms</li>
-  <li>genxml: Make MI_STORE_DATA_IMM more consistent</li>
-  <li>anv/query: Perform CmdResetQueryPool on the GPU</li>
-  <li>intel/blorp: Explicitly flush all allocated state</li>
-</ul>
-
-<p>Jose Maria Casanova Crespo (1):</p>
-<ul>
-  <li>glsl: non-last member unsized array on SSBO must fail compilation on GLSL ES 3.1</li>
-</ul>
-
-<p>Kenneth Graunke (1):</p>
-<ul>
-  <li>mesa: Do (TCS &amp;&amp; !TES) draw time validation in ES as well.</li>
-</ul>
-
-<p>Leo Liu (1):</p>
-<ul>
-  <li>configure.ac: check require_basic_egl only if egl enabled</li>
-</ul>
-
-<p>Lionel Landwerlin (2):</p>
-<ul>
-  <li>anv: wsi: report presentation error per image request</li>
-  <li>i965/fs: fix uninitialized memory access</li>
-</ul>
-
-<p>Marek Olšák (6):</p>
-<ul>
-  <li>radeonsi: fix UNSIGNED_BYTE index buffer fallback with non-zero start (v2)</li>
-  <li>gallium/util: remove unused u_index_modify helpers</li>
-  <li>gallium/u_index_modify: don't add PIPE_TRANSFER_UNSYNCHRONIZED unconditionally</li>
-  <li>gallium/u_queue: fix random crashes when the app calls exit()</li>
-  <li>radeonsi: fix broken tessellation on Carrizo and Stoney</li>
-  <li>amd/common: fix ASICREV_IS_POLARIS11_M for Polaris12</li>
-</ul>
-
-<p>Mauro Rossi (2):</p>
-<ul>
-  <li>android: radeonsi: fix sid_table.h generated header include path</li>
-  <li>android: glsl: build shader cache sources</li>
-</ul>
-
-<p>Michel Dänzer (1):</p>
-<ul>
-  <li>configure.ac: Drop LLVM compiler flags more radically</li>
-</ul>
-
-<p>Nicolai Hähnle (3):</p>
-<ul>
-  <li>winsys/amdgpu: reduce max_alloc_size based on GTT limits</li>
-  <li>radeonsi: handle MultiDrawIndirect in si_get_draw_start_count</li>
-  <li>radeonsi: fix UINT/SINT clamping for 10-bit formats on &lt;= CIK</li>
-</ul>
-
-<p>Samuel Iglesias Gonsálvez (1):</p>
-<ul>
-  <li>glsl: fix heap-use-after-free in ast_declarator_list::hir()</li>
-</ul>
-
-<p>Tapani Pälli (1):</p>
-<ul>
-  <li>android: fix droid_create_image_from_prime_fd_yuv for YV12</li>
-</ul>
-
-
-</div>
-</body>
-</html>

docs/relnotes/17.0.2.html

@@ -1,185 +0,0 @@
-<!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 17.0.2 Release Notes / March 20, 2017</h1>
-
-<p>
-Mesa 17.0.2 is a bug fix release which fixes bugs found since the 17.0.1 release.
-</p>
-<p>
-Mesa 17.0.2 implements the OpenGL 4.5 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 4.5.  OpenGL
-4.5 is <strong>only</strong> available if requested at context creation
-because compatibility contexts are not supported.
-</p>
-
-
-<h2>SHA256 checksums</h2>
-<pre>
-2e0f41e7974ba7a36ca32bbeaf8ebcd65c8fd4d2dc9872f04d4becbd5e7a8cb5  mesa-17.0.2.tar.gz
-f8f191f909e01e65de38d5bdea5fb057f21649a3aed20948be02348e77a689d4  mesa-17.0.2.tar.xz
-</pre>
-
-
-<h2>New features</h2>
-<p>None</p>
-
-
-<h2>Bug fixes</h2>
-
-<ul>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=68504">Bug 68504</a> - 9.2-rc1 workaround for clover build failure on ppc/altivec: cannot convert 'bool' to '__vector(4) __bool int' in return</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=97988">Bug 97988</a> - [radeonsi] playing back videos with VDPAU exhibits deinterlacing/anti-aliasing issues not visible with VA-API</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99484">Bug 99484</a> - Crusader Kings 2 - Loading bars, siege bars, morale bars, etc. do not render correctly</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99715">Bug 99715</a> - Don't print: &quot;Note: Buggy applications may crash, if they do please report to vendor&quot;</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=100049">Bug 100049</a> - &quot;ralloc: Make sure ralloc() allocations match malloc()'s alignment.&quot; causes seg fault in 32bit build</li>
-
-</ul>
-
-
-<h2>Changes</h2>
-
-<p>Alex Smith (3):</p>
-<ul>
-  <li>radv: Emit pending flushes before executing a secondary command buffer</li>
-  <li>radv: Flush before copying with PKT3_WRITE_DATA in CmdUpdateBuffer</li>
-  <li>radv/ac: Fix shared memory offset calculation</li>
-</ul>
-
-<p>Bas Nieuwenhuizen (3):</p>
-<ul>
-  <li>radv: Disable HTILE for textures with multiple layers/levels.</li>
-  <li>radv: Emit cache flushes before CP DMA.</li>
-  <li>Revert "radv: Emit cache flushes before CP DMA."</li>
-</ul>
-
-<p>Dave Airlie (3):</p>
-<ul>
-  <li>radv: drop Z24 support.</li>
-  <li>radv: disable mip point pre clamping.</li>
-  <li>radv: setup llvm target data layout</li>
-</ul>
-
-<p>Emil Velikov (4):</p>
-<ul>
-  <li>docs: add sha256 checksums for 17.0.1</li>
-  <li>cherry-ignore: add the swizzle blorp_clear fix</li>
-  <li>i965: move brw_define.h ifndef guard to the top</li>
-  <li>Update version to 17.0.2</li>
-</ul>
-
-<p>Fredrik Höglund (2):</p>
-<ul>
-  <li>radv: fix the dynamic buffer index in vkCmdBindDescriptorSets</li>
-  <li>radv/ac: fix multiple descriptor sets with dynamic buffers</li>
-</ul>
-
-<p>Gregory Hainaut (1):</p>
-<ul>
-  <li>glapi: fix typo in count_scale</li>
-</ul>
-
-<p>Ilia Mirkin (2):</p>
-<ul>
-  <li>nvc0: take extra pushbuf space into account for pushbuf_space calls</li>
-  <li>nvc0: increase alignment to 256 for texture buffers on fermi</li>
-</ul>
-
-<p>Jacob Lifshay (1):</p>
-<ul>
-  <li>vulkan/wsi: Improve the DRI3 error message</li>
-</ul>
-
-<p>James Legg (1):</p>
-<ul>
-  <li>radv: Fix using more than 4 bound descriptor sets</li>
-</ul>
-
-<p>Jason Ekstrand (7):</p>
-<ul>
-  <li>anv/blorp/clear_subpass: Only set surface clear color for fast clears</li>
-  <li>anv: Accurately advertise dynamic descriptor limits</li>
-  <li>anv: Stall before fast-clear operations</li>
-  <li>anv: Properly handle destroying NULL devices and instances</li>
-  <li>anv/blorp: Turn off AUX after doing a CCS_D resolve</li>
-  <li>anv/blorp: Only set a clear color for resolves if fast-cleared</li>
-  <li>nir/intrinsics: Make load_barycentric_input take a 2-component coor</li>
-</ul>
-
-<p>Jonas Pfeil (1):</p>
-<ul>
-  <li>ralloc: Make sure ralloc() allocations match malloc()'s alignment.</li>
-</ul>
-
-<p>Kenneth Graunke (1):</p>
-<ul>
-  <li>egl: Ensure ResetNotificationStrategy matches for shared contexts.</li>
-</ul>
-
-<p>Marek Olšák (3):</p>
-<ul>
-  <li>st/mesa: reset sample_mask, min_sample, and render_condition for PBO ops</li>
-  <li>st/mesa: set blend state for PBO readbacks</li>
-  <li>radeonsi: mark all bound shader buffer ranges as initialized</li>
-</ul>
-
-<p>Matt Turner (1):</p>
-<ul>
-  <li>clover: Work around build failure with AltiVec.</li>
-</ul>
-
-<p>Nanley Chery (2):</p>
-<ul>
-  <li>anv/pass: Avoid accessing attachment array out of bounds</li>
-  <li>anv/image: Remove extra dependency on HiZ-specific variable</li>
-</ul>
-
-<p>Nicolai Hähnle (2):</p>
-<ul>
-  <li>st/glsl_to_tgsi: avoid iterating past the head of the instruction list</li>
-  <li>st/mesa: inform the driver of framebuffer changes before compute dispatches</li>
-</ul>
-
-<p>Robert Foss (1):</p>
-<ul>
-  <li>mesa: Avoid read of uninitialized variable</li>
-</ul>
-
-<p>Samuel Iglesias Gonsálvez (5):</p>
-<ul>
-  <li>i965/fs: mark last DF uniform array element as 64 bit live one</li>
-  <li>i965/fs: detect different bit size accesses to uniforms to push them in proper locations</li>
-  <li>i965/fs: fix indirect load DF uniforms on BSW/BXT</li>
-  <li>i965/fs: fix source type when emitting MOV_INDIRECT to read ICP handles</li>
-  <li>i965/fs: emit MOV_INDIRECT with the source with the right register type</li>
-</ul>
-
-<p>Samuel Pitoiset (1):</p>
-<ul>
-  <li>radeonsi: disable sinking common instructions down to the end block</li>
-</ul>
-
-
-</div>
-</body>
-</html>

docs/relnotes/17.0.3.html

@@ -1,189 +0,0 @@
-<!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 17.0.3 Release Notes / April 1, 2017</h1>
-
-<p>
-Mesa 17.0.3 is a bug fix release which fixes bugs found since the 17.0.2 release.
-</p>
-<p>
-Mesa 17.0.3 implements the OpenGL 4.5 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 4.5.  OpenGL
-4.5 is <strong>only</strong> available if requested at context creation
-because compatibility contexts are not supported.
-</p>
-
-
-<h2>SHA256 checksums</h2>
-<pre>
-8253edf1bdd7b14ab63d5982349143a5c9ac3767f39a63257cc9d7e7d92f60f1  mesa-17.0.3.tar.gz
-ca646f5075a002d60ef9123c8a4331cede155c01712ef945a65c59a5e69fe7ed  mesa-17.0.3.tar.xz
-</pre>
-
-
-<h2>New features</h2>
-<p>None</p>
-
-
-<h2>Bug fixes</h2>
-
-<ul>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=96743">Bug 96743</a> - [BYT, HSW, SKL, BXT, KBL] GPU hangs with GfxBench 4.0 CarChase</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99246">Bug 99246</a> - [d3dadapter+radeonsi &amp; bisect] EVE-Online : hang on wormhole sight</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=100061">Bug 100061</a> - LODQ instruction generated with invalid dst mask</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=100182">Bug 100182</a> - Flickering in The Talos Principle on Sky Lake GT4.</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=100201">Bug 100201</a> - Windows scons build with MSVC toolchain and LLVM 4.0 fails</li>
-
-</ul>
-
-
-<h2>Changes</h2>
-
-<p>Alex Deucher (1):</p>
-<ul>
-  <li>radeonsi: add new polaris12 pci id</li>
-</ul>
-
-<p>Andres Gomez (5):</p>
-<ul>
-  <li>glsl: on UBO/SSBOs link error reset the number of active blocks to 0</li>
-  <li>cherry-ignore: add the Invalidate L2 for TRANSFER_WRITE barriers fix</li>
-  <li>cherry-ignore: add the Flush after unmap in gbm/dri fix</li>
-  <li>cherry-ignore: corrected typo in the Flush after unmap in gbm/dri fix</li>
-  <li>Update version to 17.0.3</li>
-</ul>
-
-<p>Axel Davy (2):</p>
-<ul>
-  <li>st/nine: Resolve deadlock in surface/volume dtors when using csmt</li>
-  <li>st/nine: Use atomics for available_texture_mem</li>
-</ul>
-
-<p>Bas Nieuwenhuizen (1):</p>
-<ul>
-  <li>radv: flush DB cache before and after HTILE decompress.</li>
-</ul>
-
-<p>Dave Airlie (1):</p>
-<ul>
-  <li>radv: fix primitive reset index emission</li>
-</ul>
-
-<p>Emil Velikov (1):</p>
-<ul>
-  <li>docs: add sha256 checksums for 17.0.2</li>
-</ul>
-
-<p>Ilia Mirkin (1):</p>
-<ul>
-  <li>st/mesa: set result writemask based on ir type</li>
-</ul>
-
-<p>Jan Vesely (1):</p>
-<ul>
-  <li>clover: use pipe_resource references</li>
-</ul>
-
-<p>Jason Ekstrand (9):</p>
-<ul>
-  <li>anv/query: Invalidate the correct range</li>
-  <li>anv/GetQueryPoolResults: Actually implement the spec</li>
-  <li>anv/image: Return early when unbinding an image</li>
-  <li>anv/query: Fix the location of timestamp availability</li>
-  <li>anv: Make anv_get_layerCount a macro</li>
-  <li>anv/blorp: Use anv_get_layerCount everywhere</li>
-  <li>anv/cmd_buffer: Apply flush operations prior to executing secondaries</li>
-  <li>anv/cmd_buffer: Fix bad indentation</li>
-  <li>anv: Flush caches prior to PIPELINE_SELECT on all gens</li>
-</ul>
-
-<p>José Fonseca (1):</p>
-<ul>
-  <li>c11/threads: Include thr/xtimec.h for xtime definition when building with MSVC.</li>
-</ul>
-
-<p>Juan A. Suarez Romero (1):</p>
-<ul>
-  <li>tests/cache_test: allow crossing mount points</li>
-</ul>
-
-<p>Karol Herbst (1):</p>
-<ul>
-  <li>nvc0/ir: treat FMA like MAD for operand propagation</li>
-</ul>
-
-<p>Kenneth Graunke (1):</p>
-<ul>
-  <li>i965: Fall back to GL 4.2/4.3 on Haswell if the kernel isn't new enough.</li>
-</ul>
-
-<p>Marek Olšák (1):</p>
-<ul>
-  <li>radeonsi: don't hang on shader compile failure</li>
-</ul>
-
-<p>Matt Turner (1):</p>
-<ul>
-  <li>i965/fs: Don't emit SEL instructions for type-converting MOVs.</li>
-</ul>
-
-<p>Nanley Chery (1):</p>
-<ul>
-  <li>intel: Correct the BDW surface state size</li>
-</ul>
-
-<p>Nicolai Hähnle (1):</p>
-<ul>
-  <li>mesa/main: fix MultiDrawElements[BaseVertex] validation of primcount</li>
-</ul>
-
-<p>Rob Clark (1):</p>
-<ul>
-  <li>freedreno: fix memory leak</li>
-</ul>
-
-<p>Tim Rowley (1):</p>
-<ul>
-  <li>swr: [rasterizer jitter] fix llvm &gt;= 5.0 build break</li>
-</ul>
-
-<p>Timothy Arceri (2):</p>
-<ul>
-  <li>glsl: fix lower jumps for returns when loop is inside an if</li>
-  <li>mesa: update lower_jumps tests after bug fix</li>
-</ul>
-
-<p>Topi Pohjolainen (1):</p>
-<ul>
-  <li>i965/gen8+: Do full stall when switching pipeline</li>
-</ul>
-
-<p>Xu Randy (2):</p>
-<ul>
-  <li>anv/blorp: Fix a crash in CmdClearColorImage</li>
-  <li>anv/genX: Solve the vkCreateGraphicsPipelines crash</li>
-</ul>
-
-</div>
-</body>
-</html>

docs/relnotes/17.0.4.html

@@ -1,149 +0,0 @@
-<!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 17.0.4 Release Notes / April 17, 2017</h1>
-
-<p>
-Mesa 17.0.4 is a bug fix release which fixes bugs found since the 17.0.3 release.
-</p>
-<p>
-Mesa 17.0.4 implements the OpenGL 4.5 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 4.5.  OpenGL
-4.5 is <strong>only</strong> available if requested at context creation
-because compatibility contexts are not supported.
-</p>
-
-
-<h2>SHA256 checksums</h2>
-<pre>
-c4c34ba05d48f76b45bc05bc4b6e9242077f403d63c4f0c355c7b07786de233e  mesa-17.0.4.tar.gz
-1269dc8545a193932a0779b2db5bce9be4a5f6813b98c38b93b372be8362a346  mesa-17.0.4.tar.xz
-</pre>
-
-
-<h2>New features</h2>
-<p>None</p>
-
-
-<h2>Bug fixes</h2>
-
-<ul>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=99515">Bug 99515</a> - SIGSEGV MAPERR on Android nougat-x86 with mesa 17.0.0rc</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=100391">Bug 100391</a> - SachaWillems deferredmultisampling asserts</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=100452">Bug 100452</a> - push_constants host memory leak when resetting command buffer</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=100582">Bug 100582</a> - [GEN8+] piglit.spec.arb_stencil_texturing.glblitframebuffer corrupts state.gl_texture* assertions</li>
-
-</ul>
-
-
-<h2>Changes</h2>
-
-<p>Alex Deucher (1):</p>
-<ul>
-  <li>radeonsi: add new polaris10 pci id</li>
-</ul>
-
-<p>Alex Smith (1):</p>
-<ul>
-  <li>radv: Invalidate L2 for TRANSFER_WRITE barriers</li>
-</ul>
-
-<p>Andres Gomez (1):</p>
-<ul>
-  <li>docs: add sha256 checksums for 17.0.3</li>
-</ul>
-
-<p>Craig Stout (1):</p>
-<ul>
-  <li>anv/cmd_buffer: fix host memory leak</li>
-</ul>
-
-<p>Emil Velikov (3):</p>
-<ul>
-  <li>Revert "cherry-ignore: add the Flush after unmap in gbm/dri fix"</li>
-  <li>Revert "freedreno: fix memory leak"</li>
-  <li>Update version to 17.0.4</li>
-</ul>
-
-<p>Fabio Estevam (1):</p>
-<ul>
-  <li>loader: Move non-error message to debug level</li>
-</ul>
-
-<p>Ilia Mirkin (4):</p>
-<ul>
-  <li>nvc0/ir: fix LSB/BFE/BFI implementations</li>
-  <li>nvc0/ir: fix overwriting of offset register with interpolateAtOffset</li>
-  <li>nvc0: increase texture buffer object alignment to 256 for pre-GM107</li>
-  <li>nouveau: when mapping a persistent buffer, synchronize on former xfers</li>
-</ul>
-
-<p>Jason Ekstrand (5):</p>
-<ul>
-  <li>i965/fs: Always provide a default LOD of 0 for TXS and TXL</li>
-  <li>anv/pipeline: Properly handle unset gl_Layer and gl_ViewportIndex</li>
-  <li>anv/blorp: Align vertex buffers to 64B</li>
-  <li>i965/blorp: Align vertex buffers to 64B</li>
-  <li>i965/blorp: Bump the batch space estimate</li>
-</ul>
-
-<p>Jerome Duval (2):</p>
-<ul>
-  <li>haiku: build fixes around debug defines</li>
-  <li>haiku/winsys: fix dt prototype args</li>
-</ul>
-
-<p>Julien Isorce (4):</p>
-<ul>
-  <li>winsys/radeon: check null in radeon_cs_create_fence</li>
-  <li>winsys/radeon: check null return from radeon_cs_create_fence in cs_flush</li>
-  <li>radeon: initialize hole variable before calling container_of</li>
-  <li>radeon_drm_bo: explicitly check return value of drmCommandWriteRead</li>
-</ul>
-
-<p>Kenneth Graunke (4):</p>
-<ul>
-  <li>i965: Document the sad story of the kernel command parser.</li>
-  <li>i965: Set screen-&gt;cmd_parser_version to 0 if we can't write registers.</li>
-  <li>i965: Skip register write detection when possible.</li>
-  <li>i965: Set kernel features before computing max GL version.</li>
-</ul>
-
-<p>Marek Olšák (1):</p>
-<ul>
-  <li>targets: export radeon winsys_create functions to silence LLVM warning</li>
-</ul>
-
-<p>Michal Srb (1):</p>
-<ul>
-  <li>st: Add cubeMapFace parameter to st_finalize_texture.</li>
-</ul>
-
-<p>Thomas Hellstrom (1):</p>
-<ul>
-  <li>gbm/dri: Flush after unmap</li>
-</ul>
-
-
-</div>
-</body>
-</html>

docs/relnotes/17.0.5.html

@@ -1,144 +0,0 @@
-<!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 17.0.5 Release Notes / April 28, 2017</h1>
-
-<p>
-Mesa 17.0.5 is a bug fix release which fixes bugs found since the 17.0.4 release.
-</p>
-<p>
-Mesa 17.0.5 implements the OpenGL 4.5 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 4.5.  OpenGL
-4.5 is <strong>only</strong> available if requested at context creation
-because compatibility contexts are not supported.
-</p>
-
-
-<h2>SHA256 checksums</h2>
-<pre>
-7510eee0d0077860b250d30d73305048c2df4ba09ea8fc04e4f3eec7beece301  mesa-17.0.5.tar.gz
-668efa445d2f57a26e5c096b1965a685733a3b57d9c736f9d6460263847f9bfe  mesa-17.0.5.tar.xz
-</pre>
-
-
-<h2>New features</h2>
-<p>None</p>
-
-
-<h2>Bug fixes</h2>
-
-<ul>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=97524">Bug 97524</a> - Samplers referring to the same texture unit with different types should raise GL_INVALID_OPERATION</li>
-
-</ul>
-
-
-<h2>Changes</h2>
-
-<p>Andres Gomez (16):</p>
-<ul>
-  <li>cherry-ignore: Add the pci_id into the shader cache UUID</li>
-  <li>cherry-ignore: fix crash if ctx torn down with no rendering</li>
-  <li>cherry-ignore: Fix typos.</li>
-  <li>cherry-ignore: Revert "etnaviv: Cannot render to rb-swapped formats"</li>
-  <li>cherry-ignore: Revert "i965/fs: Don't emit SEL instructions for type-converting MOVs."</li>
-  <li>cherry-ignore: fix typo in a2b10g10r10 fast clear calculation</li>
-  <li>cherry-ignore: remove unused anv_dispatch_table dtable</li>
-  <li>cherry-ignore: remove unused radv_dispatch_table dtable</li>
-  <li>cherry-ignore: make radv_resolve_entrypoint static</li>
-  <li>cherry-ignore: vulkan: add support for libmesa_vulkan_util</li>
-  <li>cherry-ignore: r600: fix libmesa_amd_common dependency</li>
-  <li>cherry-ignore: remove dead brw_new_shader() declaration</li>
-  <li>cherry-ignore: remove i965_symbols_test reference from .gitignore</li>
-  <li>cherry-ignore: automake: ensure that the destination directory is created</li>
-  <li>cherry-ignore: provide required gem stubs for the tests</li>
-  <li>Update version to 17.0.5</li>
-</ul>
-
-<p>Boyan Ding (2):</p>
-<ul>
-  <li>nvc0/ir: Properly handle a "split form" of predicate destination</li>
-  <li>nir: Destination component count of shader_clock intrinsic is 2</li>
-</ul>
-
-<p>Emil Velikov (5):</p>
-<ul>
-  <li>docs: add sha256 checksums for 17.0.4</li>
-  <li>winsys/sw/dri: don't use GNU void pointer arithmetic</li>
-  <li>st/clover: add space between &lt; and ::</li>
-  <li>configure.ac: check require_basic_egl only if egl enabled</li>
-  <li>st/mesa: automake: honour the vdpau header install location</li>
-</ul>
-
-<p>Francisco Jerez (2):</p>
-<ul>
-  <li>intel/fs: Use regs_written() in spilling cost heuristic for improved accuracy.</li>
-  <li>intel/fs: Take into account amount of data read in spilling cost heuristic.</li>
-</ul>
-
-<p>Grazvydas Ignotas (1):</p>
-<ul>
-  <li>radv: report timestampPeriod correctly</li>
-</ul>
-
-<p>Jason Ekstrand (5):</p>
-<ul>
-  <li>anv/blorp: Flush the texture cache in UpdateBuffer</li>
-  <li>anv/cmd_buffer: Flush the VF cache at the top of all primaries</li>
-  <li>anv/cmd_buffer: Always set up a null surface state</li>
-  <li>anv/cmd_buffer: Use the null surface state for ATTACHMENT_UNUSED</li>
-  <li>anv/blorp: Properly handle VK_ATTACHMENT_UNUSED</li>
-</ul>
-
-<p>Kenneth Graunke (1):</p>
-<ul>
-  <li>i965/vec4: Avoid reswizzling MACH instructions in opt_register_coalesce().</li>
-</ul>
-
-<p>Marek Olšák (1):</p>
-<ul>
-  <li>st/mesa: invalidate the readpix cache in st_indirect_draw_vbo</li>
-</ul>
-
-<p>Nanley Chery (1):</p>
-<ul>
-  <li>anv/cmd_buffer: Disable CCS on BDW input attachments</li>
-</ul>
-
-<p>Nicolai Hähnle (4):</p>
-<ul>
-  <li>mesa: fix remaining xfb prims check for GLES with multiple instances</li>
-  <li>mesa: extract need_xfb_remaining_prims_check</li>
-  <li>mesa: move glMultiDrawArrays to vbo and fix error handling</li>
-  <li>vbo: fix gl_DrawID handling in glMultiDrawArrays</li>
-</ul>
-
-<p>Rob Clark (1):</p>
-<ul>
-  <li>util/queue: don't hang at exit</li>
-</ul>
-
-<p>Timothy Arceri (1):</p>
-<ul>
-  <li>mesa: validate sampler type across the whole program</li>
-</ul>
-
-</div>
-</body>
-</html>

docs/relnotes/17.0.6.html

@@ -1,186 +0,0 @@
-<!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 17.0.6 Release Notes / May 12, 2017</h1>
-
-<p>
-Mesa 17.0.6 is a bug fix release which fixes bugs found since the 17.0.5 release.
-</p>
-<p>
-Mesa 17.0.6 implements the OpenGL 4.5 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 4.5.  OpenGL
-4.5 is <strong>only</strong> available if requested at context creation
-because compatibility contexts are not supported.
-</p>
-
-
-<h2>SHA256 checksums</h2>
-<pre>
-f1b2497d553e9a584f0caa3a2d9d310e27ead15fb0af170da69f6e70fb5031cd  mesa-17.0.6.tar.gz
-89ecf3bcd0f18dcca5aaa42bf36bb52a2df33be89889f94aaaad91f7a504a69d  mesa-17.0.6.tar.xz
-</pre>
-
-
-<h2>New features</h2>
-<p>None</p>
-
-
-<h2>Bug fixes</h2>
-
-<ul>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98428">Bug 98428</a> - Undefined non-weak-symbol in dri-drivers</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=100854">Bug 100854</a> - YUV to RGB Color Space Conversion result is not precise</li>
-
-</ul>
-
-
-<h2>Changes</h2>
-
-<p>Adam Jackson (1):</p>
-<ul>
-  <li>egl/platform/drm: Don't take display ownership until gbm is initialized</li>
-</ul>
-
-<p>Andres Gomez (7):</p>
-<ul>
-  <li>docs: add sha256 checksums for 17.0.5</li>
-  <li>travis: replace Trusty-based LLVM toolchain apt-get with apt addon</li>
-  <li>travis: add the possibility of using the txc-dxtn library</li>
-  <li>cherry-ignore: 17.1 nominations only</li>
-  <li>cherry-ignore: fix regression in descriptor set freeing.</li>
-  <li>cherry-ignore: rejected commits</li>
-  <li>Update version to 17.0.6</li>
-</ul>
-
-<p>Ben Boeckel (1):</p>
-<ul>
-  <li>scons: update for LLVM 4.0</li>
-</ul>
-
-<p>Brian Paul (1):</p>
-<ul>
-  <li>st/mesa: move duplicated st_ws_framebuffer() function into header file</li>
-</ul>
-
-<p>Chad Versace (3):</p>
-<ul>
-  <li>egl: Emit error when EGLSurface is lost</li>
-  <li>egl/android: Cancel any outstanding ANativeBuffer in surface destructor</li>
-  <li>egl/android: Mark surface as lost when dequeueBuffer fails</li>
-</ul>
-
-<p>Christian Gmeiner (1):</p>
-<ul>
-  <li>etnaviv: add L8A8_UNORM texture format</li>
-</ul>
-
-<p>Dave Airlie (2):</p>
-<ul>
-  <li>radv/wsi: report presentation error per image request</li>
-  <li>radv: enable POLARIS12 support.</li>
-</ul>
-
-<p>Emil Velikov (21):</p>
-<ul>
-  <li>travis: correct libdrm required regex to also track libdrm itself</li>
-  <li>travis: add nearly all gallium drivers to the list</li>
-  <li>travis: use both cores for make/make check</li>
-  <li>travis: bring the scons build on par with AppVeyor</li>
-  <li>travis: explicitly LD_LIBRARY_PATH the local libraries</li>
-  <li>travis: enable apt cache</li>
-  <li>travis: automatically manage ccache caching</li>
-  <li>travis: remove unused -dev packages</li>
-  <li>travis: rework "if test" blocks in the script section</li>
-  <li>travis: split out matrix from env</li>
-  <li>travis: add separate "scons" and "scons llvm" targets</li>
-  <li>travis: add "scons swr" to the build matrix</li>
-  <li>travis: add "make swr" to the build matrix</li>
-  <li>travis: split the make target to three separate ones</li>
-  <li>travis: model scons check target like the make one</li>
-  <li>travis: add Gallium state-tracker targets</li>
-  <li>travis: enable wayland support</li>
-  <li>travis: bump MAKEFLAGS to -j4</li>
-  <li>gallium/dri: always link against shared glapi</li>
-  <li>mesa/dri: always link against shared glapi</li>
-  <li>glx: glX_proto_send.py: use correct compile guard GLX_INDIRECT_RENDERING</li>
-</ul>
-
-<p>Eric Anholt (1):</p>
-<ul>
-  <li>nir: Pick just the channels we want for bitmap and drawpixels lowering.</li>
-</ul>
-
-<p>Ilia Mirkin (1):</p>
-<ul>
-  <li>gallium/targets: fix bool setting on BE architectures</li>
-</ul>
-
-<p>Jason Ekstrand (1):</p>
-<ul>
-  <li>anv/cmd_buffer: Use the device allocator for QueueSubmit</li>
-</ul>
-
-<p>Johnson Lin (1):</p>
-<ul>
-  <li>nir/lower_tex: Fix minor error in YUV color conversion matrix</li>
-</ul>
-
-<p>Marek Olšák (2):</p>
-<ul>
-  <li>radeonsi: adjust ESGS ring buffer size computation on VI</li>
-  <li>radeonsi: apply the tess+GS hang workaround to Polaris12 as well</li>
-</ul>
-
-<p>Nicolai Hähnle (1):</p>
-<ul>
-  <li>radeonsi: fix gl_PrimitiveID in tessellation with instanced draws on SI</li>
-</ul>
-
-<p>Philipp Zabel (3):</p>
-<ul>
-  <li>renderonly: close transfer prime_fd</li>
-  <li>renderonly: drop resources on destroy</li>
-  <li>renderonly: use drmIoctl</li>
-</ul>
-
-<p>Rhys Kidd (3):</p>
-<ul>
-  <li>travis: Support LLVM 3.8+ on Trusty-based Travis-CI via apt-get not apt addon</li>
-  <li>travis: Add radv vulkan driver to continuous integration</li>
-  <li>travis: Add radeonsi to continuous integration</li>
-</ul>
-
-<p>Rob Clark (1):</p>
-<ul>
-  <li>freedreno/a3xx: fix hang w/ large render targets and small gmem</li>
-</ul>
-
-<p>Samuel Iglesias Gonsálvez (5):</p>
-<ul>
-  <li>i965/vec4: fix vertical stride to avoid breaking region parameter rule</li>
-  <li>i965/vec4: fix register width for DF VGRF and UNIFORM</li>
-  <li>i965/vec4: don't modify regioning parameters to the sources of DF align1 instructions</li>
-  <li>anv: anv_gem_mmap() returns MAP_FAILED as mapping error</li>
-  <li>anv: vkBindImageMemory() should return VK_ERROR_OUT_OF_{HOST,DEVICE}_MEMORY on failure</li>
-</ul>
-
-</div>
-</body>
-</html>

docs/relnotes/17.0.7.html

@@ -1,145 +0,0 @@
-<!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 17.0.7 Release Notes / June 1, 2017</h1>
-
-<p>
-Mesa 17.0.7 is a bug fix release which fixes bugs found since the 17.0.6 release.
-</p>
-<p>
-Mesa 17.0.7 implements the OpenGL 4.5 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 4.5.  OpenGL
-4.5 is <strong>only</strong> available if requested at context creation
-because compatibility contexts are not supported.
-</p>
-
-
-<h2>SHA256 checksums</h2>
-<pre>
-bc68d13c6b1a053b855ac453ebf7e62bd89511adf44bad6c613e09f7fa13390a  mesa-17.0.7.tar.gz
-f6d75304a229c8d10443e219d6b6c0c342567dbab5a879ebe7cfa3c9139c4492  mesa-17.0.7.tar.xz
-</pre>
-
-
-<h2>New features</h2>
-<p>None</p>
-
-
-<h2>Bug fixes</h2>
-
-<ul>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=98833">Bug 98833</a> - [REGRESSION, bisected] Wayland revert commit breaks non-Vsync fullscreen frame updates</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=100741">Bug 100741</a> - Chromium - Memory leak</li>
-
-<li><a href="https://bugs.freedesktop.org/show_bug.cgi?id=100925">Bug 100925</a> - [HSW/BSW/BDW/SKL] Google Earth is not resolving all the details in the map correctly</li>
-
-</ul>
-
-
-<h2>Changes</h2>
-
-<p>Andres Gomez (1):</p>
-<ul>
-  <li>docs: add sha256 checksums for 17.0.6</li>
-</ul>
-
-<p>Bartosz Tomczyk (1):</p>
-<ul>
-  <li>mesa: Avoid leaking surface in st_renderbuffer_delete</li>
-</ul>
-
-<p>Chad Versace (1):</p>
-<ul>
-  <li>egl: Partially revert 23c86c74, fix eglMakeCurrent</li>
-</ul>
-
-<p>Daniel Stone (7):</p>
-<ul>
-  <li>vulkan: Fix Wayland uninitialised registry</li>
-  <li>vulkan/wsi/wayland: Remove roundtrip when creating image</li>
-  <li>vulkan/wsi/wayland: Use per-display event queue</li>
-  <li>vulkan/wsi/wayland: Use proxy wrappers for swapchain</li>
-  <li>egl/wayland: Don't open-code roundtrip</li>
-  <li>egl/wayland: Use per-surface event queues</li>
-  <li>egl/wayland: Ensure we get a back buffer</li>
-</ul>
-
-<p>Emil Velikov (5):</p>
-<ul>
-  <li>st/va: fix misplaced closing bracket</li>
-  <li>anv: automake: list shared libraries after the static ones</li>
-  <li>radv: automake: list shared libraries after the static ones</li>
-  <li>egl/wayland: select the format based on the interface used</li>
-  <li>Update version to 17.0.7</li>
-</ul>
-
-<p>Eric Anholt (2):</p>
-<ul>
-  <li>renderonly: Initialize fields of struct winsys_handle.</li>
-  <li>vc4: Don't allocate new BOs to avoid synchronization when they're shared.</li>
-</ul>
-
-<p>Hans de Goede (1):</p>
-<ul>
-  <li>glxglvnddispatch: Add missing dispatch for GetDriverConfig</li>
-</ul>
-
-<p>Ilia Mirkin (1):</p>
-<ul>
-  <li>nvc0/ir: SHLADD's middle source must be an immediate</li>
-</ul>
-
-<p>Jason Ekstrand (2):</p>
-<ul>
-  <li>i965/blorp: Do and end-of-pipe sync on both sides of fast-clear ops</li>
-  <li>i965: Round copy size to the nearest block in intel_miptree_copy</li>
-</ul>
-
-<p>Lucas Stach (1):</p>
-<ul>
-  <li>etnaviv: stop oversizing buffer resources</li>
-</ul>
-
-<p>Nanley Chery (2):</p>
-<ul>
-  <li>anv/formats: Update the three-channel BC1 mappings</li>
-  <li>i965/formats: Update the three-channel DXT1 mappings</li>
-</ul>
-
-<p>Pohjolainen, Topi (1):</p>
-<ul>
-  <li>intel/isl/gen7: Use stencil vertical alignment of 8 instead of 4</li>
-</ul>
-
-<p>Samuel Iglesias Gonsálvez (3):</p>
-<ul>
-  <li>i965/vec4/gs: restore the uniform values which was overwritten by failed vec4_gs_visitor execution</li>
-  <li>i965/vec4: fix swizzle and writemask when loading an uniform with constant offset</li>
-  <li>i965/vec4: load dvec3/4 uniforms first in the push constant buffer</li>
-</ul>
-
-<p>Tom Stellard (1):</p>
-<ul>
-  <li>gallivm: Make sure module has the correct data layout when pass manager runs</li>
-</ul>
-
-</div>
-</body>
-</html>

include/GL/internal/dri_interface.h

@@ -1164,6 +1164,8 @@ struct __DRIdri2ExtensionRec {
 #define __DRI_IMAGE_FORMAT_ARGB2101010  0x100a
 #define __DRI_IMAGE_FORMAT_SARGB8       0x100b
 #define __DRI_IMAGE_FORMAT_ARGB1555     0x100c
+#define __DRI_IMAGE_FORMAT_R16          0x100d
+#define __DRI_IMAGE_FORMAT_GR1616       0x100e
 
 #define __DRI_IMAGE_USE_SHARE		0x0001
 #define __DRI_IMAGE_USE_SCANOUT		0x0002
@@ -1192,6 +1194,8 @@ struct __DRIdri2ExtensionRec {
 #define __DRI_IMAGE_FOURCC_R8		0x20203852
 #define __DRI_IMAGE_FOURCC_GR88		0x38385247
 #define __DRI_IMAGE_FOURCC_ARGB1555	0x35315241
+#define __DRI_IMAGE_FOURCC_R16		0x20363152
+#define __DRI_IMAGE_FOURCC_GR1616	0x32335247
 #define __DRI_IMAGE_FOURCC_RGB565	0x36314752
 #define __DRI_IMAGE_FOURCC_ARGB8888	0x34325241
 #define __DRI_IMAGE_FOURCC_XRGB8888	0x34325258

include/c11/threads.h

@@ -30,9 +30,6 @@
 #define EMULATED_THREADS_H_INCLUDED_
 
 #include <time.h>
-#ifdef _MSC_VER
-#include <thr/xtimec.h>  // for xtime
-#endif
 
 #ifndef TIME_UTC
 #define TIME_UTC 1
@@ -44,13 +41,11 @@
 typedef void (*tss_dtor_t)(void*);
 typedef int (*thrd_start_t)(void*);
 
-#ifndef _MSC_VER
 struct xtime {
     time_t sec;
     long nsec;
 };
 typedef struct xtime xtime;
-#endif
 
 
 /*-------------------- enumeration constants --------------------*/

include/pci_ids/radeonsi_pci_ids.h

@@ -202,7 +202,6 @@ CHIPSET(0x67C9, POLARIS10_, POLARIS10)
 CHIPSET(0x67CA, POLARIS10_, POLARIS10)
 CHIPSET(0x67CC, POLARIS10_, POLARIS10)
 CHIPSET(0x67CF, POLARIS10_, POLARIS10)
-CHIPSET(0x67D0, POLARIS10_, POLARIS10)
 CHIPSET(0x67DF, POLARIS10_, POLARIS10)
 
 CHIPSET(0x98E4, STONEY_, STONEY)
@@ -212,5 +211,4 @@ CHIPSET(0x6981, POLARIS12_, POLARIS12)
 CHIPSET(0x6985, POLARIS12_, POLARIS12)
 CHIPSET(0x6986, POLARIS12_, POLARIS12)
 CHIPSET(0x6987, POLARIS12_, POLARIS12)
-CHIPSET(0x6995, POLARIS12_, POLARIS12)
 CHIPSET(0x699F, POLARIS12_, POLARIS12)

include/vulkan/vulkan.h

@@ -6,7 +6,7 @@ extern "C" {
 #endif
 
 /*
-** Copyright (c) 2015-2016 The Khronos Group Inc.
+** Copyright (c) 2015-2017 The Khronos Group Inc.
 **
 ** Licensed under the Apache License, Version 2.0 (the "License");
 ** you may not use this file except in compliance with the License.
@@ -43,7 +43,7 @@ extern "C" {
 #define VK_VERSION_MINOR(version) (((uint32_t)(version) >> 12) & 0x3ff)
 #define VK_VERSION_PATCH(version) ((uint32_t)(version) & 0xfff)
 // Version of this file
-#define VK_HEADER_VERSION 38
+#define VK_HEADER_VERSION 39
 
 
 #define VK_NULL_HANDLE 0
@@ -145,6 +145,7 @@ typedef enum VkResult {
     VK_ERROR_INCOMPATIBLE_DISPLAY_KHR = -1000003001,
     VK_ERROR_VALIDATION_FAILED_EXT = -1000011001,
     VK_ERROR_INVALID_SHADER_NV = -1000012000,
+    VK_ERROR_OUT_OF_POOL_MEMORY_KHR = -1000069000,
     VK_RESULT_BEGIN_RANGE = VK_ERROR_FRAGMENTED_POOL,
     VK_RESULT_END_RANGE = VK_INCOMPLETE,
     VK_RESULT_RANGE_SIZE = (VK_INCOMPLETE - VK_ERROR_FRAGMENTED_POOL + 1),
@@ -225,13 +226,28 @@ typedef enum VkStructureType {
     VK_STRUCTURE_TYPE_IMPORT_MEMORY_WIN32_HANDLE_INFO_NV = 1000057000,
     VK_STRUCTURE_TYPE_EXPORT_MEMORY_WIN32_HANDLE_INFO_NV = 1000057001,
     VK_STRUCTURE_TYPE_WIN32_KEYED_MUTEX_ACQUIRE_RELEASE_INFO_NV = 1000058000,
+    VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_FEATURES_2_KHR = 1000059000,
+    VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_PROPERTIES_2_KHR = 1000059001,
+    VK_STRUCTURE_TYPE_FORMAT_PROPERTIES_2_KHR = 1000059002,
+    VK_STRUCTURE_TYPE_IMAGE_FORMAT_PROPERTIES_2_KHR = 1000059003,
+    VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_IMAGE_FORMAT_INFO_2_KHR = 1000059004,
+    VK_STRUCTURE_TYPE_QUEUE_FAMILY_PROPERTIES_2_KHR = 1000059005,
+    VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_MEMORY_PROPERTIES_2_KHR = 1000059006,
+    VK_STRUCTURE_TYPE_SPARSE_IMAGE_FORMAT_PROPERTIES_2_KHR = 1000059007,
+    VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_SPARSE_IMAGE_FORMAT_INFO_2_KHR = 1000059008,
     VK_STRUCTURE_TYPE_VALIDATION_FLAGS_EXT = 1000061000,
+    VK_STRUCTURE_TYPE_VI_SURFACE_CREATE_INFO_NN = 1000062000,
     VK_STRUCTURE_TYPE_OBJECT_TABLE_CREATE_INFO_NVX = 1000086000,
     VK_STRUCTURE_TYPE_INDIRECT_COMMANDS_LAYOUT_CREATE_INFO_NVX = 1000086001,
     VK_STRUCTURE_TYPE_CMD_PROCESS_COMMANDS_INFO_NVX = 1000086002,
     VK_STRUCTURE_TYPE_CMD_RESERVE_SPACE_FOR_COMMANDS_INFO_NVX = 1000086003,
     VK_STRUCTURE_TYPE_DEVICE_GENERATED_COMMANDS_LIMITS_NVX = 1000086004,
     VK_STRUCTURE_TYPE_DEVICE_GENERATED_COMMANDS_FEATURES_NVX = 1000086005,
+    VK_STRUCTURE_TYPE_SURFACE_CAPABILITIES2_EXT = 1000090000,
+    VK_STRUCTURE_TYPE_DISPLAY_POWER_INFO_EXT = 1000091000,
+    VK_STRUCTURE_TYPE_DEVICE_EVENT_INFO_EXT = 1000091001,
+    VK_STRUCTURE_TYPE_DISPLAY_EVENT_INFO_EXT = 1000091002,
+    VK_STRUCTURE_TYPE_SWAPCHAIN_COUNTER_CREATE_INFO_EXT = 1000091003,
     VK_STRUCTURE_TYPE_BEGIN_RANGE = VK_STRUCTURE_TYPE_APPLICATION_INFO,
     VK_STRUCTURE_TYPE_END_RANGE = VK_STRUCTURE_TYPE_LOADER_DEVICE_CREATE_INFO,
     VK_STRUCTURE_TYPE_RANGE_SIZE = (VK_STRUCTURE_TYPE_LOADER_DEVICE_CREATE_INFO - VK_STRUCTURE_TYPE_APPLICATION_INFO + 1),
@@ -840,6 +856,8 @@ typedef enum VkFormatFeatureFlagBits {
     VK_FORMAT_FEATURE_BLIT_DST_BIT = 0x00000800,
     VK_FORMAT_FEATURE_SAMPLED_IMAGE_FILTER_LINEAR_BIT = 0x00001000,
     VK_FORMAT_FEATURE_SAMPLED_IMAGE_FILTER_CUBIC_BIT_IMG = 0x00002000,
+    VK_FORMAT_FEATURE_TRANSFER_SRC_BIT_KHR = 0x00004000,
+    VK_FORMAT_FEATURE_TRANSFER_DST_BIT_KHR = 0x00008000,
     VK_FORMAT_FEATURE_FLAG_BITS_MAX_ENUM = 0x7FFFFFFF
 } VkFormatFeatureFlagBits;
 typedef VkFlags VkFormatFeatureFlags;
@@ -863,6 +881,7 @@ typedef enum VkImageCreateFlagBits {
     VK_IMAGE_CREATE_SPARSE_ALIASED_BIT = 0x00000004,
     VK_IMAGE_CREATE_MUTABLE_FORMAT_BIT = 0x00000008,
     VK_IMAGE_CREATE_CUBE_COMPATIBLE_BIT = 0x00000010,
+    VK_IMAGE_CREATE_2D_ARRAY_COMPATIBLE_BIT_KHR = 0x00000020,
     VK_IMAGE_CREATE_FLAG_BITS_MAX_ENUM = 0x7FFFFFFF
 } VkImageCreateFlagBits;
 typedef VkFlags VkImageCreateFlags;
@@ -3206,6 +3225,18 @@ VK_DEFINE_NON_DISPATCHABLE_HANDLE(VkSurfaceKHR)
 
 typedef enum VkColorSpaceKHR {
     VK_COLOR_SPACE_SRGB_NONLINEAR_KHR = 0,
+    VK_COLOR_SPACE_DISPLAY_P3_LINEAR_EXT = 1000104001,
+    VK_COLOR_SPACE_DISPLAY_P3_NONLINEAR_EXT = 1000104002,
+    VK_COLOR_SPACE_SCRGB_LINEAR_EXT = 1000104003,
+    VK_COLOR_SPACE_SCRGB_NONLINEAR_EXT = 1000104004,
+    VK_COLOR_SPACE_DCI_P3_LINEAR_EXT = 1000104005,
+    VK_COLOR_SPACE_DCI_P3_NONLINEAR_EXT = 1000104006,
+    VK_COLOR_SPACE_BT709_LINEAR_EXT = 1000104007,
+    VK_COLOR_SPACE_BT709_NONLINEAR_EXT = 1000104008,
+    VK_COLOR_SPACE_BT2020_LINEAR_EXT = 1000104009,
+    VK_COLOR_SPACE_BT2020_NONLINEAR_EXT = 1000104010,
+    VK_COLOR_SPACE_ADOBERGB_LINEAR_EXT = 1000104011,
+    VK_COLOR_SPACE_ADOBERGB_NONLINEAR_EXT = 1000104012,
     VK_COLOR_SPACE_BEGIN_RANGE_KHR = VK_COLOR_SPACE_SRGB_NONLINEAR_KHR,
     VK_COLOR_SPACE_END_RANGE_KHR = VK_COLOR_SPACE_SRGB_NONLINEAR_KHR,
     VK_COLOR_SPACE_RANGE_SIZE_KHR = (VK_COLOR_SPACE_SRGB_NONLINEAR_KHR - VK_COLOR_SPACE_SRGB_NONLINEAR_KHR + 1),
@@ -3741,6 +3772,136 @@ VKAPI_ATTR VkBool32 VKAPI_CALL vkGetPhysicalDeviceWin32PresentationSupportKHR(
 #define VK_KHR_SAMPLER_MIRROR_CLAMP_TO_EDGE_EXTENSION_NAME "VK_KHR_sampler_mirror_clamp_to_edge"
 
 
+#define VK_KHR_get_physical_device_properties2 1
+#define VK_KHR_GET_PHYSICAL_DEVICE_PROPERTIES_2_SPEC_VERSION 1
+#define VK_KHR_GET_PHYSICAL_DEVICE_PROPERTIES_2_EXTENSION_NAME "VK_KHR_get_physical_device_properties2"
+
+typedef struct VkPhysicalDeviceFeatures2KHR {
+    VkStructureType             sType;
+    void*                       pNext;
+    VkPhysicalDeviceFeatures    features;
+} VkPhysicalDeviceFeatures2KHR;
+
+typedef struct VkPhysicalDeviceProperties2KHR {
+    VkStructureType               sType;
+    void*                         pNext;
+    VkPhysicalDeviceProperties    properties;
+} VkPhysicalDeviceProperties2KHR;
+
+typedef struct VkFormatProperties2KHR {
+    VkStructureType       sType;
+    void*                 pNext;
+    VkFormatProperties    formatProperties;
+} VkFormatProperties2KHR;
+
+typedef struct VkImageFormatProperties2KHR {
+    VkStructureType            sType;
+    void*                      pNext;
+    VkImageFormatProperties    imageFormatProperties;
+} VkImageFormatProperties2KHR;
+
+typedef struct VkPhysicalDeviceImageFormatInfo2KHR {
+    VkStructureType       sType;
+    const void*           pNext;
+    VkFormat              format;
+    VkImageType           type;
+    VkImageTiling         tiling;
+    VkImageUsageFlags     usage;
+    VkImageCreateFlags    flags;
+} VkPhysicalDeviceImageFormatInfo2KHR;
+
+typedef struct VkQueueFamilyProperties2KHR {
+    VkStructureType            sType;
+    void*                      pNext;
+    VkQueueFamilyProperties    queueFamilyProperties;
+} VkQueueFamilyProperties2KHR;
+
+typedef struct VkPhysicalDeviceMemoryProperties2KHR {
+    VkStructureType                     sType;
+    void*                               pNext;
+    VkPhysicalDeviceMemoryProperties    memoryProperties;
+} VkPhysicalDeviceMemoryProperties2KHR;
+
+typedef struct VkSparseImageFormatProperties2KHR {
+    VkStructureType                  sType;
+    void*                            pNext;
+    VkSparseImageFormatProperties    properties;
+} VkSparseImageFormatProperties2KHR;
+
+typedef struct VkPhysicalDeviceSparseImageFormatInfo2KHR {
+    VkStructureType          sType;
+    const void*              pNext;
+    VkFormat                 format;
+    VkImageType              type;
+    VkSampleCountFlagBits    samples;
+    VkImageUsageFlags        usage;
+    VkImageTiling            tiling;
+} VkPhysicalDeviceSparseImageFormatInfo2KHR;
+
+
+typedef void (VKAPI_PTR *PFN_vkGetPhysicalDeviceFeatures2KHR)(VkPhysicalDevice physicalDevice, VkPhysicalDeviceFeatures2KHR* pFeatures);
+typedef void (VKAPI_PTR *PFN_vkGetPhysicalDeviceProperties2KHR)(VkPhysicalDevice physicalDevice, VkPhysicalDeviceProperties2KHR* pProperties);
+typedef void (VKAPI_PTR *PFN_vkGetPhysicalDeviceFormatProperties2KHR)(VkPhysicalDevice physicalDevice, VkFormat format, VkFormatProperties2KHR* pFormatProperties);
+typedef VkResult (VKAPI_PTR *PFN_vkGetPhysicalDeviceImageFormatProperties2KHR)(VkPhysicalDevice physicalDevice, const VkPhysicalDeviceImageFormatInfo2KHR* pImageFormatInfo, VkImageFormatProperties2KHR* pImageFormatProperties);
+typedef void (VKAPI_PTR *PFN_vkGetPhysicalDeviceQueueFamilyProperties2KHR)(VkPhysicalDevice physicalDevice, uint32_t* pQueueFamilyPropertyCount, VkQueueFamilyProperties2KHR* pQueueFamilyProperties);
+typedef void (VKAPI_PTR *PFN_vkGetPhysicalDeviceMemoryProperties2KHR)(VkPhysicalDevice physicalDevice, VkPhysicalDeviceMemoryProperties2KHR* pMemoryProperties);
+typedef void (VKAPI_PTR *PFN_vkGetPhysicalDeviceSparseImageFormatProperties2KHR)(VkPhysicalDevice physicalDevice, const VkPhysicalDeviceSparseImageFormatInfo2KHR* pFormatInfo, uint32_t* pPropertyCount, VkSparseImageFormatProperties2KHR* pProperties);
+
+#ifndef VK_NO_PROTOTYPES
+VKAPI_ATTR void VKAPI_CALL vkGetPhysicalDeviceFeatures2KHR(
+    VkPhysicalDevice                            physicalDevice,
+    VkPhysicalDeviceFeatures2KHR*               pFeatures);
+
+VKAPI_ATTR void VKAPI_CALL vkGetPhysicalDeviceProperties2KHR(
+    VkPhysicalDevice                            physicalDevice,
+    VkPhysicalDeviceProperties2KHR*             pProperties);
+
+VKAPI_ATTR void VKAPI_CALL vkGetPhysicalDeviceFormatProperties2KHR(
+    VkPhysicalDevice                            physicalDevice,
+    VkFormat                                    format,
+    VkFormatProperties2KHR*                     pFormatProperties);
+
+VKAPI_ATTR VkResult VKAPI_CALL vkGetPhysicalDeviceImageFormatProperties2KHR(
+    VkPhysicalDevice                            physicalDevice,
+    const VkPhysicalDeviceImageFormatInfo2KHR*  pImageFormatInfo,
+    VkImageFormatProperties2KHR*                pImageFormatProperties);
+
+VKAPI_ATTR void VKAPI_CALL vkGetPhysicalDeviceQueueFamilyProperties2KHR(
+    VkPhysicalDevice                            physicalDevice,
+    uint32_t*                                   pQueueFamilyPropertyCount,
+    VkQueueFamilyProperties2KHR*                pQueueFamilyProperties);
+
+VKAPI_ATTR void VKAPI_CALL vkGetPhysicalDeviceMemoryProperties2KHR(
+    VkPhysicalDevice                            physicalDevice,
+    VkPhysicalDeviceMemoryProperties2KHR*       pMemoryProperties);
+
+VKAPI_ATTR void VKAPI_CALL vkGetPhysicalDeviceSparseImageFormatProperties2KHR(
+    VkPhysicalDevice                            physicalDevice,
+    const VkPhysicalDeviceSparseImageFormatInfo2KHR* pFormatInfo,
+    uint32_t*                                   pPropertyCount,
+    VkSparseImageFormatProperties2KHR*          pProperties);
+#endif
+
+#define VK_KHR_shader_draw_parameters 1
+#define VK_KHR_SHADER_DRAW_PARAMETERS_SPEC_VERSION 1
+#define VK_KHR_SHADER_DRAW_PARAMETERS_EXTENSION_NAME "VK_KHR_shader_draw_parameters"
+
+
+#define VK_KHR_maintenance1 1
+#define VK_KHR_MAINTENANCE1_SPEC_VERSION  1
+#define VK_KHR_MAINTENANCE1_EXTENSION_NAME "VK_KHR_maintenance1"
+
+typedef VkFlags VkCommandPoolTrimFlagsKHR;
+
+typedef void (VKAPI_PTR *PFN_vkTrimCommandPoolKHR)(VkDevice device, VkCommandPool commandPool, VkCommandPoolTrimFlagsKHR flags);
+
+#ifndef VK_NO_PROTOTYPES
+VKAPI_ATTR void VKAPI_CALL vkTrimCommandPoolKHR(
+    VkDevice                                    device,
+    VkCommandPool                               commandPool,
+    VkCommandPoolTrimFlagsKHR                   flags);
+#endif
+
 #define VK_EXT_debug_report 1
 VK_DEFINE_NON_DISPATCHABLE_HANDLE(VkDebugReportCallbackEXT)
 
@@ -4167,6 +4328,42 @@ typedef struct VkValidationFlagsEXT {
 
 
 
+#ifdef VK_USE_PLATFORM_VI_NN
+#define VK_NN_vi_surface 1
+#define VK_NN_VI_SURFACE_SPEC_VERSION     1
+#define VK_NN_VI_SURFACE_EXTENSION_NAME   "VK_NN_vi_surface"
+
+typedef VkFlags VkViSurfaceCreateFlagsNN;
+
+typedef struct VkViSurfaceCreateInfoNN {
+    VkStructureType             sType;
+    const void*                 pNext;
+    VkViSurfaceCreateFlagsNN    flags;
+    void*                       window;
+} VkViSurfaceCreateInfoNN;
+
+
+typedef VkResult (VKAPI_PTR *PFN_vkCreateViSurfaceNN)(VkInstance instance, const VkViSurfaceCreateInfoNN* pCreateInfo, const VkAllocationCallbacks* pAllocator, VkSurfaceKHR* pSurface);
+
+#ifndef VK_NO_PROTOTYPES
+VKAPI_ATTR VkResult VKAPI_CALL vkCreateViSurfaceNN(
+    VkInstance                                  instance,
+    const VkViSurfaceCreateInfoNN*              pCreateInfo,
+    const VkAllocationCallbacks*                pAllocator,
+    VkSurfaceKHR*                               pSurface);
+#endif
+#endif /* VK_USE_PLATFORM_VI_NN */
+
+#define VK_EXT_shader_subgroup_ballot 1
+#define VK_EXT_SHADER_SUBGROUP_BALLOT_SPEC_VERSION 1
+#define VK_EXT_SHADER_SUBGROUP_BALLOT_EXTENSION_NAME "VK_EXT_shader_subgroup_ballot"
+
+
+#define VK_EXT_shader_subgroup_vote 1
+#define VK_EXT_SHADER_SUBGROUP_VOTE_SPEC_VERSION 1
+#define VK_EXT_SHADER_SUBGROUP_VOTE_EXTENSION_NAME "VK_EXT_shader_subgroup_vote"
+
+
 #define VK_NVX_device_generated_commands 1
 VK_DEFINE_NON_DISPATCHABLE_HANDLE(VkObjectTableNVX)
 VK_DEFINE_NON_DISPATCHABLE_HANDLE(VkIndirectCommandsLayoutNVX)
@@ -4322,6 +4519,7 @@ typedef struct VkObjectTableIndexBufferEntryNVX {
     VkObjectEntryTypeNVX          type;
     VkObjectEntryUsageFlagsNVX    flags;
     VkBuffer                      buffer;
+    VkIndexType                   indexType;
 } VkObjectTableIndexBufferEntryNVX;
 
 typedef struct VkObjectTablePushConstantEntryNVX {
@@ -4393,6 +4591,171 @@ VKAPI_ATTR void VKAPI_CALL vkGetPhysicalDeviceGeneratedCommandsPropertiesNVX(
     VkDeviceGeneratedCommandsLimitsNVX*         pLimits);
 #endif
 
+#define VK_EXT_direct_mode_display 1
+#define VK_EXT_DIRECT_MODE_DISPLAY_SPEC_VERSION 1
+#define VK_EXT_DIRECT_MODE_DISPLAY_EXTENSION_NAME "VK_EXT_direct_mode_display"
+
+typedef VkResult (VKAPI_PTR *PFN_vkReleaseDisplayEXT)(VkPhysicalDevice physicalDevice, VkDisplayKHR display);
+
+#ifndef VK_NO_PROTOTYPES
+VKAPI_ATTR VkResult VKAPI_CALL vkReleaseDisplayEXT(
+    VkPhysicalDevice                            physicalDevice,
+    VkDisplayKHR                                display);
+#endif
+
+#ifdef VK_USE_PLATFORM_XLIB_XRANDR_EXT
+#define VK_EXT_acquire_xlib_display 1
+#include <X11/extensions/Xrandr.h>
+
+#define VK_EXT_ACQUIRE_XLIB_DISPLAY_SPEC_VERSION 1
+#define VK_EXT_ACQUIRE_XLIB_DISPLAY_EXTENSION_NAME "VK_EXT_acquire_xlib_display"
+
+typedef VkResult (VKAPI_PTR *PFN_vkAcquireXlibDisplayEXT)(VkPhysicalDevice physicalDevice, Display* dpy, VkDisplayKHR display);
+typedef VkResult (VKAPI_PTR *PFN_vkGetRandROutputDisplayEXT)(VkPhysicalDevice physicalDevice, Display* dpy, RROutput rrOutput, VkDisplayKHR* pDisplay);
+
+#ifndef VK_NO_PROTOTYPES
+VKAPI_ATTR VkResult VKAPI_CALL vkAcquireXlibDisplayEXT(
+    VkPhysicalDevice                            physicalDevice,
+    Display*                                    dpy,
+    VkDisplayKHR                                display);
+
+VKAPI_ATTR VkResult VKAPI_CALL vkGetRandROutputDisplayEXT(
+    VkPhysicalDevice                            physicalDevice,
+    Display*                                    dpy,
+    RROutput                                    rrOutput,
+    VkDisplayKHR*                               pDisplay);
+#endif
+#endif /* VK_USE_PLATFORM_XLIB_XRANDR_EXT */
+
+#define VK_EXT_display_surface_counter 1
+#define VK_EXT_DISPLAY_SURFACE_COUNTER_SPEC_VERSION 1
+#define VK_EXT_DISPLAY_SURFACE_COUNTER_EXTENSION_NAME "VK_EXT_display_surface_counter"
+
+
+typedef enum VkSurfaceCounterFlagBitsEXT {
+    VK_SURFACE_COUNTER_VBLANK_EXT = 0x00000001,
+    VK_SURFACE_COUNTER_FLAG_BITS_MAX_ENUM_EXT = 0x7FFFFFFF
+} VkSurfaceCounterFlagBitsEXT;
+typedef VkFlags VkSurfaceCounterFlagsEXT;
+
+typedef struct VkSurfaceCapabilities2EXT {
+    VkStructureType                  sType;
+    void*                            pNext;
+    uint32_t                         minImageCount;
+    uint32_t                         maxImageCount;
+    VkExtent2D                       currentExtent;
+    VkExtent2D                       minImageExtent;
+    VkExtent2D                       maxImageExtent;
+    uint32_t                         maxImageArrayLayers;
+    VkSurfaceTransformFlagsKHR       supportedTransforms;
+    VkSurfaceTransformFlagBitsKHR    currentTransform;
+    VkCompositeAlphaFlagsKHR         supportedCompositeAlpha;
+    VkImageUsageFlags                supportedUsageFlags;
+    VkSurfaceCounterFlagsEXT         supportedSurfaceCounters;
+} VkSurfaceCapabilities2EXT;
+
+
+typedef VkResult (VKAPI_PTR *PFN_vkGetPhysicalDeviceSurfaceCapabilities2EXT)(VkPhysicalDevice physicalDevice, VkSurfaceKHR surface, VkSurfaceCapabilities2EXT* pSurfaceCapabilities);
+
+#ifndef VK_NO_PROTOTYPES
+VKAPI_ATTR VkResult VKAPI_CALL vkGetPhysicalDeviceSurfaceCapabilities2EXT(
+    VkPhysicalDevice                            physicalDevice,
+    VkSurfaceKHR                                surface,
+    VkSurfaceCapabilities2EXT*                  pSurfaceCapabilities);
+#endif
+
+#define VK_EXT_display_control 1
+#define VK_EXT_DISPLAY_CONTROL_SPEC_VERSION 1
+#define VK_EXT_DISPLAY_CONTROL_EXTENSION_NAME "VK_EXT_display_control"
+
+
+typedef enum VkDisplayPowerStateEXT {
+    VK_DISPLAY_POWER_STATE_OFF_EXT = 0,
+    VK_DISPLAY_POWER_STATE_SUSPEND_EXT = 1,
+    VK_DISPLAY_POWER_STATE_ON_EXT = 2,
+    VK_DISPLAY_POWER_STATE_BEGIN_RANGE_EXT = VK_DISPLAY_POWER_STATE_OFF_EXT,
+    VK_DISPLAY_POWER_STATE_END_RANGE_EXT = VK_DISPLAY_POWER_STATE_ON_EXT,
+    VK_DISPLAY_POWER_STATE_RANGE_SIZE_EXT = (VK_DISPLAY_POWER_STATE_ON_EXT - VK_DISPLAY_POWER_STATE_OFF_EXT + 1),
+    VK_DISPLAY_POWER_STATE_MAX_ENUM_EXT = 0x7FFFFFFF
+} VkDisplayPowerStateEXT;
+
+typedef enum VkDeviceEventTypeEXT {
+    VK_DEVICE_EVENT_TYPE_DISPLAY_HOTPLUG_EXT = 0,
+    VK_DEVICE_EVENT_TYPE_BEGIN_RANGE_EXT = VK_DEVICE_EVENT_TYPE_DISPLAY_HOTPLUG_EXT,
+    VK_DEVICE_EVENT_TYPE_END_RANGE_EXT = VK_DEVICE_EVENT_TYPE_DISPLAY_HOTPLUG_EXT,
+    VK_DEVICE_EVENT_TYPE_RANGE_SIZE_EXT = (VK_DEVICE_EVENT_TYPE_DISPLAY_HOTPLUG_EXT - VK_DEVICE_EVENT_TYPE_DISPLAY_HOTPLUG_EXT + 1),
+    VK_DEVICE_EVENT_TYPE_MAX_ENUM_EXT = 0x7FFFFFFF
+} VkDeviceEventTypeEXT;
+
+typedef enum VkDisplayEventTypeEXT {
+    VK_DISPLAY_EVENT_TYPE_FIRST_PIXEL_OUT_EXT = 0,
+    VK_DISPLAY_EVENT_TYPE_BEGIN_RANGE_EXT = VK_DISPLAY_EVENT_TYPE_FIRST_PIXEL_OUT_EXT,
+    VK_DISPLAY_EVENT_TYPE_END_RANGE_EXT = VK_DISPLAY_EVENT_TYPE_FIRST_PIXEL_OUT_EXT,
+    VK_DISPLAY_EVENT_TYPE_RANGE_SIZE_EXT = (VK_DISPLAY_EVENT_TYPE_FIRST_PIXEL_OUT_EXT - VK_DISPLAY_EVENT_TYPE_FIRST_PIXEL_OUT_EXT + 1),
+    VK_DISPLAY_EVENT_TYPE_MAX_ENUM_EXT = 0x7FFFFFFF
+} VkDisplayEventTypeEXT;
+
+typedef struct VkDisplayPowerInfoEXT {
+    VkStructureType           sType;
+    const void*               pNext;
+    VkDisplayPowerStateEXT    powerState;
+} VkDisplayPowerInfoEXT;
+
+typedef struct VkDeviceEventInfoEXT {
+    VkStructureType         sType;
+    const void*             pNext;
+    VkDeviceEventTypeEXT    deviceEvent;
+} VkDeviceEventInfoEXT;
+
+typedef struct VkDisplayEventInfoEXT {
+    VkStructureType          sType;
+    const void*              pNext;
+    VkDisplayEventTypeEXT    displayEvent;
+} VkDisplayEventInfoEXT;
+
+typedef struct VkSwapchainCounterCreateInfoEXT {
+    VkStructureType             sType;
+    const void*                 pNext;
+    VkSurfaceCounterFlagsEXT    surfaceCounters;
+} VkSwapchainCounterCreateInfoEXT;
+
+
+typedef VkResult (VKAPI_PTR *PFN_vkDisplayPowerControlEXT)(VkDevice device, VkDisplayKHR display, const VkDisplayPowerInfoEXT* pDisplayPowerInfo);
+typedef VkResult (VKAPI_PTR *PFN_vkRegisterDeviceEventEXT)(VkDevice device, const VkDeviceEventInfoEXT* pDeviceEventInfo, const VkAllocationCallbacks* pAllocator, VkFence* pFence);
+typedef VkResult (VKAPI_PTR *PFN_vkRegisterDisplayEventEXT)(VkDevice device, VkDisplayKHR display, const VkDisplayEventInfoEXT* pDisplayEventInfo, const VkAllocationCallbacks* pAllocator, VkFence* pFence);
+typedef VkResult (VKAPI_PTR *PFN_vkGetSwapchainCounterEXT)(VkDevice device, VkSwapchainKHR swapchain, VkSurfaceCounterFlagBitsEXT counter, uint64_t* pCounterValue);
+
+#ifndef VK_NO_PROTOTYPES
+VKAPI_ATTR VkResult VKAPI_CALL vkDisplayPowerControlEXT(
+    VkDevice                                    device,
+    VkDisplayKHR                                display,
+    const VkDisplayPowerInfoEXT*                pDisplayPowerInfo);
+
+VKAPI_ATTR VkResult VKAPI_CALL vkRegisterDeviceEventEXT(
+    VkDevice                                    device,
+    const VkDeviceEventInfoEXT*                 pDeviceEventInfo,
+    const VkAllocationCallbacks*                pAllocator,
+    VkFence*                                    pFence);
+
+VKAPI_ATTR VkResult VKAPI_CALL vkRegisterDisplayEventEXT(
+    VkDevice                                    device,
+    VkDisplayKHR                                display,
+    const VkDisplayEventInfoEXT*                pDisplayEventInfo,
+    const VkAllocationCallbacks*                pAllocator,
+    VkFence*                                    pFence);
+
+VKAPI_ATTR VkResult VKAPI_CALL vkGetSwapchainCounterEXT(
+    VkDevice                                    device,
+    VkSwapchainKHR                              swapchain,
+    VkSurfaceCounterFlagBitsEXT                 counter,
+    uint64_t*                                   pCounterValue);
+#endif
+
+#define VK_EXT_swapchain_colorspace 1
+#define VK_SWAPCHAIN_COLOR_SPACE_SPEC_VERSION 1
+#define VK_SWAPCHAIN_COLOR_SPACE_EXTENSION_NAME "VK_EXT_swapchain_colorspace"
+
+
 #ifdef __cplusplus
 }
 #endif

scons/llvm.py

@@ -105,26 +105,8 @@ def generate(env):
             'HAVE_STDINT_H',
         ])
         env.Prepend(LIBPATH = [os.path.join(llvm_dir, 'lib')])
-        # LIBS should match the output of `llvm-config --libs engine mcjit bitwriter x86asmprinter irreader`
-        if llvm_version >= distutils.version.LooseVersion('4.0'):
-            env.Prepend(LIBS = [
-                'LLVMX86Disassembler', 'LLVMX86AsmParser',
-                'LLVMX86CodeGen', 'LLVMSelectionDAG', 'LLVMAsmPrinter',
-                'LLVMDebugInfoCodeView', 'LLVMCodeGen',
-                'LLVMScalarOpts', 'LLVMInstCombine',
-                'LLVMTransformUtils',
-                'LLVMBitWriter', 'LLVMX86Desc',
-                'LLVMMCDisassembler', 'LLVMX86Info',
-                'LLVMX86AsmPrinter', 'LLVMX86Utils',
-                'LLVMMCJIT', 'LLVMExecutionEngine', 'LLVMTarget',
-                'LLVMAnalysis', 'LLVMProfileData',
-                'LLVMRuntimeDyld', 'LLVMObject', 'LLVMMCParser',
-                'LLVMBitReader', 'LLVMMC', 'LLVMCore',
-                'LLVMSupport',
-                'LLVMIRReader', 'LLVMAsmParser',
-                'LLVMDemangle', 'LLVMGlobalISel', 'LLVMDebugInfoMSF',
-            ])
-        elif llvm_version >= distutils.version.LooseVersion('3.9'):
+        # LIBS should match the output of `llvm-config --libs engine mcjit bitwriter x86asmprinter`
+        if llvm_version >= distutils.version.LooseVersion('3.9'):
             env.Prepend(LIBS = [
                 'LLVMX86Disassembler', 'LLVMX86AsmParser',
                 'LLVMX86CodeGen', 'LLVMSelectionDAG', 'LLVMAsmPrinter',

src/amd/Android.common.mk

@@ -55,7 +55,7 @@ LOCAL_C_INCLUDES := \
 	external/llvm/include \
 	external/llvm/device/include \
 	external/libcxx/include \
-	$(ELF_INCLUDES)
+	external/elfutils/$(if $(filter 5,$(MESA_ANDROID_MAJOR_VERSION)),0.153/,$(if $(filter 6,$(MESA_ANDROID_MAJOR_VERSION)),src/))libelf
 
 LOCAL_STATIC_LIBRARIES := libLLVMCore
 

src/amd/common/ac_llvm_util.c

@@ -119,7 +119,6 @@ static const char *ac_get_llvm_processor_name(enum radeon_family family)
 	case CHIP_POLARIS10:
 		return "polaris10";
 	case CHIP_POLARIS11:
-	case CHIP_POLARIS12:
 		return "polaris11";
 #endif
 	default:

src/amd/common/ac_nir_to_llvm.c

@@ -264,7 +264,7 @@ static LLVMValueRef get_shared_memory_ptr(struct nir_to_llvm_context *ctx,
 	LLVMValueRef ptr;
 	int addr_space;
 
-	offset = LLVMConstInt(ctx->i32, idx * 16, false);
+	offset = LLVMConstInt(ctx->i32, idx, false);
 
 	ptr = ctx->shared_memory;
 	ptr = LLVMBuildGEP(ctx->builder, ptr, &offset, 1, "");
@@ -1267,9 +1267,6 @@ static void visit_alu(struct nir_to_llvm_context *ctx, nir_alu_instr *instr)
 		src[1] = to_float(ctx, src[1]);
 		result = LLVMBuildFRem(ctx->builder, src[0], src[1], "");
 		break;
-	case nir_op_irem:
-		result = LLVMBuildSRem(ctx->builder, src[0], src[1], "");
-		break;
 	case nir_op_idiv:
 		result = LLVMBuildSDiv(ctx->builder, src[0], src[1], "");
 		break;
@@ -1721,17 +1718,15 @@ static LLVMValueRef visit_vulkan_resource_index(struct nir_to_llvm_context *ctx,
 	unsigned desc_set = nir_intrinsic_desc_set(instr);
 	unsigned binding = nir_intrinsic_binding(instr);
 	LLVMValueRef desc_ptr = ctx->descriptor_sets[desc_set];
-	struct radv_pipeline_layout *pipeline_layout = ctx->options->layout;
-	struct radv_descriptor_set_layout *layout = pipeline_layout->set[desc_set].layout;
+	struct radv_descriptor_set_layout *layout = ctx->options->layout->set[desc_set].layout;
 	unsigned base_offset = layout->binding[binding].offset;
 	LLVMValueRef offset, stride;
 
 	if (layout->binding[binding].type == VK_DESCRIPTOR_TYPE_UNIFORM_BUFFER_DYNAMIC ||
 	    layout->binding[binding].type == VK_DESCRIPTOR_TYPE_STORAGE_BUFFER_DYNAMIC) {
-		unsigned idx = pipeline_layout->set[desc_set].dynamic_offset_start +
-			layout->binding[binding].dynamic_offset_offset;
 		desc_ptr = ctx->push_constants;
-		base_offset = pipeline_layout->push_constant_size + 16 * idx;
+		base_offset = ctx->options->layout->push_constant_size;
+		base_offset +=  16 * layout->binding[binding].dynamic_offset_offset;
 		stride = LLVMConstInt(ctx->i32, 16, false);
 	} else
 		stride = LLVMConstInt(ctx->i32, layout->binding[binding].size, false);
@@ -1750,12 +1745,9 @@ static LLVMValueRef visit_vulkan_resource_index(struct nir_to_llvm_context *ctx,
 static LLVMValueRef visit_load_push_constant(struct nir_to_llvm_context *ctx,
                                              nir_intrinsic_instr *instr)
 {
-	LLVMValueRef ptr, addr;
+	LLVMValueRef ptr;
 
-	addr = LLVMConstInt(ctx->i32, nir_intrinsic_base(instr), 0);
-	addr = LLVMBuildAdd(ctx->builder, addr, get_src(ctx, instr->src[0]), "");
-
-	ptr = build_gep0(ctx, ctx->push_constants, addr);
+	ptr = build_gep0(ctx, ctx->push_constants, get_src(ctx, instr->src[0]));
 	ptr = cast_ptr(ctx, ptr, get_def_type(ctx, &instr->dest.ssa));
 
 	return LLVMBuildLoad(ctx->builder, ptr, "");
@@ -2088,9 +2080,6 @@ static LLVMValueRef visit_load_var(struct nir_to_llvm_context *ctx,
 		LLVMValueRef ptr = get_shared_memory_ptr(ctx, idx, ctx->i32);
 		LLVMValueRef derived_ptr;
 
-		if (indir_index)
-			indir_index = LLVMBuildMul(ctx->builder, indir_index, LLVMConstInt(ctx->i32, 4, false), "");
-
 		for (unsigned chan = 0; chan < ve; chan++) {
 			LLVMValueRef index = LLVMConstInt(ctx->i32, chan, false);
 			if (indir_index)
@@ -2195,9 +2184,6 @@ visit_store_var(struct nir_to_llvm_context *ctx,
 		ptr = get_shared_memory_ptr(ctx, idx, ctx->i32);
 		LLVMValueRef derived_ptr;
 
-		if (indir_index)
-			indir_index = LLVMBuildMul(ctx->builder, indir_index, LLVMConstInt(ctx->i32, 4, false), "");
-
 		for (unsigned chan = 0; chan < 4; chan++) {
 			if (!(writemask & (1 << chan)))
 				continue;
@@ -2252,7 +2238,7 @@ static int image_type_to_components_count(enum glsl_sampler_dim dim, bool array)
 }
 
 static LLVMValueRef get_image_coords(struct nir_to_llvm_context *ctx,
-				     nir_intrinsic_instr *instr)
+				     nir_intrinsic_instr *instr, bool add_frag_pos)
 {
 	const struct glsl_type *type = instr->variables[0]->var->type;
 	if(instr->variables[0]->deref.child)
@@ -2267,8 +2253,6 @@ static LLVMValueRef get_image_coords(struct nir_to_llvm_context *ctx,
 	LLVMValueRef res;
 	int count;
 	enum glsl_sampler_dim dim = glsl_get_sampler_dim(type);
-	bool add_frag_pos = (dim == GLSL_SAMPLER_DIM_SUBPASS ||
-			     dim == GLSL_SAMPLER_DIM_SUBPASS_MS);
 	bool is_ms = (dim == GLSL_SAMPLER_DIM_MS ||
 		      dim == GLSL_SAMPLER_DIM_SUBPASS_MS);
 
@@ -2394,11 +2378,12 @@ static LLVMValueRef visit_image_load(struct nir_to_llvm_context *ctx,
 	} else {
 		bool is_da = glsl_sampler_type_is_array(type) ||
 			     glsl_get_sampler_dim(type) == GLSL_SAMPLER_DIM_CUBE;
+		bool add_frag_pos = glsl_get_sampler_dim(type) == GLSL_SAMPLER_DIM_SUBPASS;
 		LLVMValueRef da = is_da ? ctx->i32one : ctx->i32zero;
 		LLVMValueRef glc = LLVMConstInt(ctx->i1, 0, false);
 		LLVMValueRef slc = LLVMConstInt(ctx->i1, 0, false);
 
-		params[0] = get_image_coords(ctx, instr);
+		params[0] = get_image_coords(ctx, instr, add_frag_pos);
 		params[1] = get_sampler_desc(ctx, instr->variables[0], DESC_IMAGE);
 		params[2] = LLVMConstInt(ctx->i32, 15, false); /* dmask */
 		if (HAVE_LLVM <= 0x0309) {
@@ -2457,7 +2442,7 @@ static void visit_image_store(struct nir_to_llvm_context *ctx,
 		LLVMValueRef slc = i1false;
 
 		params[0] = to_float(ctx, get_src(ctx, instr->src[2]));
-		params[1] = get_image_coords(ctx, instr); /* coords */
+		params[1] = get_image_coords(ctx, instr, false); /* coords */
 		params[2] = get_sampler_desc(ctx, instr->variables[0], DESC_IMAGE);
 		params[3] = LLVMConstInt(ctx->i32, 15, false); /* dmask */
 		if (HAVE_LLVM <= 0x0309) {
@@ -2517,7 +2502,7 @@ static LLVMValueRef visit_image_atomic(struct nir_to_llvm_context *ctx,
 		bool da = glsl_sampler_type_is_array(type) ||
 		          glsl_get_sampler_dim(type) == GLSL_SAMPLER_DIM_CUBE;
 
-		coords = params[param_count++] = get_image_coords(ctx, instr);
+		coords = params[param_count++] = get_image_coords(ctx, instr, false);
 		params[param_count++] = get_sampler_desc(ctx, instr->variables[0], DESC_IMAGE);
 		params[param_count++] = i1false; /* r128 */
 		params[param_count++] = da ? i1true : i1false;      /* da */
@@ -3169,15 +3154,6 @@ static void tex_fetch_ptrs(struct nir_to_llvm_context *ctx,
 		*fmask_ptr = get_sampler_desc(ctx, instr->texture, DESC_FMASK);
 }
 
-static LLVMValueRef apply_round_slice(struct nir_to_llvm_context *ctx,
-				      LLVMValueRef coord)
-{
-	coord = to_float(ctx, coord);
-	coord = ac_emit_llvm_intrinsic(&ctx->ac, "llvm.rint.f32", ctx->f32, &coord, 1, 0);
-	coord = to_integer(ctx, coord);
-	return coord;
-}
-
 static void visit_tex(struct nir_to_llvm_context *ctx, nir_tex_instr *instr)
 {
 	LLVMValueRef result = NULL;
@@ -3235,11 +3211,6 @@ static void visit_tex(struct nir_to_llvm_context *ctx, nir_tex_instr *instr)
 		}
 	}
 
-	if (instr->op == nir_texop_txs && instr->sampler_dim == GLSL_SAMPLER_DIM_BUF) {
-		result = get_buffer_size(ctx, res_ptr, true);
-		goto write_result;
-	}
-
 	if (instr->op == nir_texop_texture_samples) {
 		LLVMValueRef res, samples, is_msaa;
 		res = LLVMBuildBitCast(ctx->builder, res_ptr, ctx->v8i32, "");
@@ -3339,16 +3310,15 @@ static void visit_tex(struct nir_to_llvm_context *ctx, nir_tex_instr *instr)
 	/* Pack texture coordinates */
 	if (coord) {
 		address[count++] = coords[0];
-		if (instr->coord_components > 1) {
-			if (instr->sampler_dim == GLSL_SAMPLER_DIM_1D && instr->is_array && instr->op != nir_texop_txf) {
-				coords[1] = apply_round_slice(ctx, coords[1]);
-			}
+		if (instr->coord_components > 1)
 			address[count++] = coords[1];
-		}
 		if (instr->coord_components > 2) {
 			/* This seems like a bit of a hack - but it passes Vulkan CTS with it */
 			if (instr->sampler_dim != GLSL_SAMPLER_DIM_3D && instr->op != nir_texop_txf) {
-				coords[2] = apply_round_slice(ctx, coords[2]);
+				coords[2] = to_float(ctx, coords[2]);
+				coords[2] = ac_emit_llvm_intrinsic(&ctx->ac, "llvm.rint.f32", ctx->f32, &coords[2],
+								1, 0);
+				coords[2] = to_integer(ctx, coords[2]);
 			}
 			address[count++] = coords[2];
 		}
@@ -3407,8 +3377,7 @@ static void visit_tex(struct nir_to_llvm_context *ctx, nir_tex_instr *instr)
 	 * The sample index should be adjusted as follows:
 	 *   sample_index = (fmask >> (sample_index * 4)) & 0xF;
 	 */
-	if (instr->sampler_dim == GLSL_SAMPLER_DIM_MS &&
-	    instr->op != nir_texop_txs) {
+	if (instr->sampler_dim == GLSL_SAMPLER_DIM_MS) {
 		LLVMValueRef txf_address[4];
 		struct ac_tex_info txf_info = { 0 };
 		unsigned txf_count = count;
@@ -4461,13 +4430,6 @@ LLVMModuleRef ac_translate_nir_to_llvm(LLVMTargetMachineRef tm,
 	memset(shader_info, 0, sizeof(*shader_info));
 
 	LLVMSetTarget(ctx.module, "amdgcn--");
-
-	LLVMTargetDataRef data_layout = LLVMCreateTargetDataLayout(tm);
-	char *data_layout_str = LLVMCopyStringRepOfTargetData(data_layout);
-	LLVMSetDataLayout(ctx.module, data_layout_str);
-	LLVMDisposeTargetData(data_layout);
-	LLVMDisposeMessage(data_layout_str);
-
 	setup_types(&ctx);
 
 	ctx.builder = LLVMCreateBuilderInContext(ctx.context);
@@ -4495,7 +4457,7 @@ LLVMModuleRef ac_translate_nir_to_llvm(LLVMTargetMachineRef tm,
 				idx++;
 			}
 
-			shared_size *= 16;
+			shared_size *= 4;
 			var = LLVMAddGlobalInAddressSpace(ctx.module,
 							  LLVMArrayType(ctx.i8, shared_size),
 							  "compute_lds",

src/amd/common/ac_nir_to_llvm.h

@@ -27,7 +27,6 @@
 #include "llvm-c/Core.h"
 #include "llvm-c/TargetMachine.h"
 #include "amd_family.h"
-#include "../vulkan/radv_descriptor_set.h"
 
 struct ac_shader_binary;
 struct ac_shader_config;
@@ -77,8 +76,7 @@ enum ac_ud_index {
 	AC_UD_MAX_UD = AC_UD_VS_MAX_UD,
 };
 
-// Match MAX_SETS from radv_descriptor_set.h
-#define AC_UD_MAX_SETS MAX_SETS
+#define AC_UD_MAX_SETS 4
 
 struct ac_userdata_locations {
 	struct ac_userdata_info descriptor_sets[AC_UD_MAX_SETS];

src/amd/common/amdgpu_id.h

@@ -157,7 +157,7 @@ enum {
 #define ASICREV_IS_POLARIS10_P(eChipRev)\
 	((eChipRev >= VI_POLARIS10_P_A0) && (eChipRev < VI_POLARIS11_M_A0))
 #define ASICREV_IS_POLARIS11_M(eChipRev)   \
-	(eChipRev >= VI_POLARIS11_M_A0 && eChipRev < VI_POLARIS12_V_A0)
+	(eChipRev >= VI_POLARIS11_M_A0)
 #define ASICREV_IS_POLARIS12_V(eChipRev)\
 	(eChipRev >= VI_POLARIS12_V_A0)
 

src/amd/vulkan/Makefile.am

@@ -21,7 +21,9 @@
 
 include Makefile.sources
 
-noinst_HEADERS = \
+vulkan_includedir = $(includedir)/vulkan
+
+vulkan_include_HEADERS = \
 	$(top_srcdir)/include/vulkan/vk_platform.h \
 	$(top_srcdir)/include/vulkan/vulkan.h
 
@@ -57,21 +59,8 @@ VULKAN_SOURCES = \
 	$(VULKAN_GENERATED_FILES) \
 	$(VULKAN_FILES)
 
-VULKAN_LIB_DEPS = \
-	libvulkan_common.la \
-	$(top_builddir)/src/vulkan/wsi/libvulkan_wsi.la \
-	$(top_builddir)/src/amd/common/libamd_common.la \
-	$(top_builddir)/src/amd/addrlib/libamdgpu_addrlib.la \
-	$(top_builddir)/src/compiler/nir/libnir.la \
-	$(top_builddir)/src/util/libmesautil.la \
-	$(LLVM_LIBS) \
-	$(LIBELF_LIBS) \
-	$(PTHREAD_LIBS) \
-	$(AMDGPU_LIBS) \
-	$(LIBDRM_LIBS) \
-	$(PTHREAD_LIBS) \
-	$(DLOPEN_LIBS) \
-	-lm
+VULKAN_LIB_DEPS =
+
 
 if HAVE_PLATFORM_X11
 AM_CPPFLAGS += \
@@ -103,6 +92,22 @@ endif
 noinst_LTLIBRARIES = libvulkan_common.la
 libvulkan_common_la_SOURCES = $(VULKAN_SOURCES)
 
+VULKAN_LIB_DEPS += \
+	libvulkan_common.la \
+	$(top_builddir)/src/vulkan/wsi/libvulkan_wsi.la \
+	$(top_builddir)/src/amd/common/libamd_common.la \
+	$(top_builddir)/src/amd/addrlib/libamdgpu_addrlib.la \
+	$(top_builddir)/src/compiler/nir/libnir.la \
+	$(top_builddir)/src/util/libmesautil.la \
+	$(LLVM_LIBS) \
+	$(LIBELF_LIBS) \
+	$(PTHREAD_LIBS) \
+	$(AMDGPU_LIBS) \
+	$(LIBDRM_LIBS) \
+	$(PTHREAD_LIBS) \
+	$(DLOPEN_LIBS) \
+	-lm
+
 nodist_EXTRA_libvulkan_radeon_la_SOURCES = dummy.cpp
 libvulkan_radeon_la_SOURCES = $(VULKAN_GEM_FILES)
 

src/amd/vulkan/radv_cmd_buffer.c

@@ -1552,7 +1552,7 @@ void radv_CmdBindDescriptorSets(
 		radv_bind_descriptor_set(cmd_buffer, set, idx);
 
 		for(unsigned j = 0; j < set->layout->dynamic_offset_count; ++j, ++dyn_idx) {
-			unsigned idx = j + layout->set[i + firstSet].dynamic_offset_start;
+			unsigned idx = j + layout->set[i].dynamic_offset_start;
 			uint32_t *dst = cmd_buffer->dynamic_buffers + idx * 4;
 			assert(dyn_idx < dynamicOffsetCount);
 
@@ -1818,9 +1818,6 @@ void radv_CmdExecuteCommands(
 {
 	RADV_FROM_HANDLE(radv_cmd_buffer, primary, commandBuffer);
 
-	/* Emit pending flushes on primary prior to executing secondary */
-	si_emit_cache_flush(primary);
-
 	for (uint32_t i = 0; i < commandBufferCount; i++) {
 		RADV_FROM_HANDLE(radv_cmd_buffer, secondary, pCmdBuffers[i]);
 
@@ -1830,7 +1827,6 @@ void radv_CmdExecuteCommands(
 	/* if we execute secondary we need to re-emit out pipelines */
 	if (commandBufferCount) {
 		primary->state.emitted_pipeline = NULL;
-		primary->state.emitted_compute_pipeline = NULL;
 		primary->state.dirty |= RADV_CMD_DIRTY_PIPELINE;
 		primary->state.dirty |= RADV_CMD_DIRTY_DYNAMIC_ALL;
 	}
@@ -1974,7 +1970,7 @@ void radv_CmdDraw(
 
 static void radv_emit_primitive_reset_index(struct radv_cmd_buffer *cmd_buffer)
 {
-	uint32_t primitive_reset_index = cmd_buffer->state.index_type ? 0xffffffffu : 0xffffu;
+	uint32_t primitive_reset_index = cmd_buffer->state.last_primitive_reset_index ? 0xffffffffu : 0xffffu;
 
 	if (cmd_buffer->state.pipeline->graphics.prim_restart_enable &&
 	    primitive_reset_index != cmd_buffer->state.last_primitive_reset_index) {
@@ -2411,13 +2407,7 @@ static void radv_handle_depth_image_transition(struct radv_cmd_buffer *cmd_buffe
 		range.baseMipLevel = 0;
 		range.levelCount = 1;
 
-		cmd_buffer->state.flush_bits |= RADV_CMD_FLAG_FLUSH_AND_INV_DB |
-		                                RADV_CMD_FLAG_FLUSH_AND_INV_DB_META;
-
 		radv_decompress_depth_image_inplace(cmd_buffer, image, &range);
-
-		cmd_buffer->state.flush_bits |= RADV_CMD_FLAG_FLUSH_AND_INV_DB |
-		                                RADV_CMD_FLAG_FLUSH_AND_INV_DB_META;
 	}
 }
 
@@ -2580,8 +2570,7 @@ void radv_CmdPipelineBarrier(
 			flush_bits |= RADV_CMD_FLAG_FLUSH_AND_INV_DB;
 			break;
 		case VK_ACCESS_TRANSFER_WRITE_BIT:
-			flush_bits |= RADV_CMD_FLAG_FLUSH_AND_INV_CB |
-			              RADV_CMD_FLAG_INV_GLOBAL_L2;
+			flush_bits |= RADV_CMD_FLAG_FLUSH_AND_INV_CB;
 			break;
 		default:
 			break;
@@ -2617,7 +2606,6 @@ void radv_CmdPipelineBarrier(
 			break;
 		case VK_ACCESS_COLOR_ATTACHMENT_READ_BIT:
 		case VK_ACCESS_TRANSFER_READ_BIT:
-		case VK_ACCESS_TRANSFER_WRITE_BIT:
 		case VK_ACCESS_INPUT_ATTACHMENT_READ_BIT:
 			flush_bits |= RADV_CMD_FLUSH_AND_INV_FRAMEBUFFER | RADV_CMD_FLAG_INV_GLOBAL_L2;
 		default:

src/amd/vulkan/radv_descriptor_set.c

@@ -275,13 +275,12 @@ radv_descriptor_set_create(struct radv_device *device,
 		uint32_t layout_size = align_u32(layout->size, 32);
 		set->size = layout->size;
 		if (!cmd_buffer) {
-			if (pool->current_offset + layout_size <= pool->size &&
-			    pool->allocated_sets < pool->max_sets) {
+			if (pool->current_offset + layout_size <= pool->size) {
 				set->bo = pool->bo;
 				set->mapped_ptr = (uint32_t*)(pool->mapped_ptr + pool->current_offset);
 				set->va = device->ws->buffer_get_va(set->bo) + pool->current_offset;
 				pool->current_offset += layout_size;
-				++pool->allocated_sets;
+
 			} else {
 				int entry = pool->free_list, prev_entry = -1;
 				uint32_t offset;
@@ -418,7 +417,6 @@ VkResult radv_CreateDescriptorPool(
 	pool->full_list = 0;
 	pool->free_nodes[max_sets - 1].next = -1;
 	pool->max_sets = max_sets;
-	pool->allocated_sets = 0;
 
 	for (int i = 0; i  + 1 < max_sets; ++i)
 		pool->free_nodes[i].next = i + 1;
@@ -496,7 +494,6 @@ VkResult radv_ResetDescriptorPool(
 		radv_descriptor_set_destroy(device, pool, set, false);
 	}
 
-	pool->allocated_sets = 0;
 	pool->current_offset = 0;
 	pool->free_list = -1;
 	pool->full_list = 0;

src/amd/vulkan/radv_device.c

@@ -570,7 +570,7 @@ void radv_GetPhysicalDeviceProperties(
 		.storageImageSampleCounts                 = VK_SAMPLE_COUNT_1_BIT,
 		.maxSampleMaskWords                       = 1,
 		.timestampComputeAndGraphics              = false,
-		.timestampPeriod                          = 1000000.0 / pdevice->rad_info.clock_crystal_freq,
+		.timestampPeriod                          = 100000.0 / pdevice->rad_info.clock_crystal_freq,
 		.maxClipDistances                         = 8,
 		.maxCullDistances                         = 8,
 		.maxCombinedClipAndCullDistances          = 8,
@@ -1121,7 +1121,7 @@ VkResult radv_AllocateMemory(
 	if (pAllocateInfo->memoryTypeIndex == RADV_MEM_TYPE_GTT_WRITE_COMBINE)
 		flags |= RADEON_FLAG_GTT_WC;
 
-	mem->bo = device->ws->buffer_create(device->ws, alloc_size, 65536,
+	mem->bo = device->ws->buffer_create(device->ws, alloc_size, 32768,
 					       domain, flags);
 
 	if (!mem->bo) {
@@ -2024,7 +2024,7 @@ radv_init_sampler(struct radv_device *device,
 			     S_008F38_XY_MAG_FILTER(radv_tex_filter(pCreateInfo->magFilter, max_aniso)) |
 			     S_008F38_XY_MIN_FILTER(radv_tex_filter(pCreateInfo->minFilter, max_aniso)) |
 			     S_008F38_MIP_FILTER(radv_tex_mipfilter(pCreateInfo->mipmapMode)) |
-			     S_008F38_MIP_POINT_PRECLAMP(0) |
+			     S_008F38_MIP_POINT_PRECLAMP(1) |
 			     S_008F38_DISABLE_LSB_CEIL(1) |
 			     S_008F38_FILTER_PREC_FIX(1) |
 			     S_008F38_ANISO_OVERRIDE(is_vi));

src/amd/vulkan/radv_formats.c

@@ -730,6 +730,9 @@ uint32_t radv_translate_dbformat(VkFormat format)
 	case VK_FORMAT_D16_UNORM:
 	case VK_FORMAT_D16_UNORM_S8_UINT:
 		return V_028040_Z_16;
+	case VK_FORMAT_X8_D24_UNORM_PACK32:
+	case VK_FORMAT_D24_UNORM_S8_UINT:
+		return V_028040_Z_24; /* deprecated on SI */
 	case VK_FORMAT_D32_SFLOAT:
 	case VK_FORMAT_D32_SFLOAT_S8_UINT:
 		return V_028040_Z_32_FLOAT;

src/amd/vulkan/radv_image.c

@@ -510,7 +510,6 @@ radv_image_alloc_fmask(struct radv_device *device,
 
 	image->fmask.offset = align64(image->size, image->fmask.alignment);
 	image->size = image->fmask.offset + image->fmask.size;
-	image->alignment = MAX2(image->alignment, image->fmask.alignment);
 }
 
 static void
@@ -572,7 +571,6 @@ radv_image_alloc_cmask(struct radv_device *device,
 	/* + 8 for storing the clear values */
 	image->clear_value_offset = image->cmask.offset + image->cmask.size;
 	image->size = image->cmask.offset + image->cmask.size + 8;
-	image->alignment = MAX2(image->alignment, image->cmask.alignment);
 }
 
 static void
@@ -583,7 +581,6 @@ radv_image_alloc_dcc(struct radv_device *device,
 	/* + 8 for storing the clear values */
 	image->clear_value_offset = image->dcc_offset + image->surface.dcc_size;
 	image->size = image->dcc_offset + image->surface.dcc_size + 8;
-	image->alignment = MAX2(image->alignment, image->surface.dcc_alignment);
 }
 
 static unsigned
@@ -655,9 +652,6 @@ radv_image_alloc_htile(struct radv_device *device,
 	if (device->debug_flags & RADV_DEBUG_NO_HIZ)
 		return;
 
-	if (image->array_size > 1 || image->levels > 1)
-		return;
-
 	image->htile.size = radv_image_get_htile_size(device, image);
 
 	if (!image->htile.size)

src/amd/vulkan/radv_meta_blit2d.c

@@ -26,7 +26,6 @@
 
 #include "radv_meta.h"
 #include "nir/nir_builder.h"
-#include "vk_format.h"
 
 enum blit2d_dst_type {
 	/* We can bind this destination as a "normal" render target and render
@@ -284,10 +283,8 @@ radv_meta_blit2d_normal_dst(struct radv_cmd_buffer *cmd_buffer,
 
 	for (unsigned r = 0; r < num_rects; ++r) {
 		VkFormat depth_format = 0;
-		if (dst->aspect_mask == VK_IMAGE_ASPECT_STENCIL_BIT)
-			depth_format = vk_format_stencil_only(dst->image->vk_format);
-		else if (dst->aspect_mask == VK_IMAGE_ASPECT_DEPTH_BIT)
-			depth_format = vk_format_depth_only(dst->image->vk_format);
+		if (dst->aspect_mask != VK_IMAGE_ASPECT_COLOR_BIT)
+			depth_format = dst->image->vk_format;
 		struct blit2d_src_temps src_temps;
 		blit2d_bind_src(cmd_buffer, src_img, src_buf, &src_temps, src_type, depth_format);
 

src/amd/vulkan/radv_meta_buffer.c

@@ -524,8 +524,6 @@ void radv_CmdUpdateBuffer(
 	assert(!(va & 3));
 
 	if (dataSize < 4096) {
-		si_emit_cache_flush(cmd_buffer);
-
 		cmd_buffer->device->ws->cs_add_buffer(cmd_buffer->cs, dst_buffer->bo, 8);
 
 		radeon_check_space(cmd_buffer->device->ws, cmd_buffer->cs, words + 4);

src/amd/vulkan/radv_private.h

@@ -543,7 +543,6 @@ struct radv_descriptor_pool {
 	int free_list;
 	int full_list;
 	uint32_t max_sets;
-	uint32_t allocated_sets;
 	struct radv_descriptor_pool_free_node free_nodes[];
 };
 

src/amd/vulkan/radv_query.c

@@ -131,7 +131,6 @@ VkResult radv_GetQueryPoolResults(
 	VkDeviceSize                                stride,
 	VkQueryResultFlags                          flags)
 {
-	RADV_FROM_HANDLE(radv_device, device, _device);
 	RADV_FROM_HANDLE(radv_query_pool, pool, queryPool);
 	char *data = pData;
 	VkResult result = VK_SUCCESS;
@@ -142,20 +141,23 @@ VkResult radv_GetQueryPoolResults(
 		char *src = pool->ptr + query * pool->stride;
 		uint32_t available;
 
+		if (flags & VK_QUERY_RESULT_WAIT_BIT) {
+			while(!*(volatile uint32_t*)(pool->ptr + pool->availability_offset + 4 * query))
+				;
+		}
+
+		if (!*(uint32_t*)(pool->ptr + pool->availability_offset + 4 * query) &&
+		    !(flags & VK_QUERY_RESULT_PARTIAL_BIT)) {
+			if (flags & VK_QUERY_RESULT_WITH_AVAILABILITY_BIT)
+				*(uint32_t*)dest = 0;
+			result = VK_NOT_READY;
+			continue;
+
+		}
+
+		available = *(uint32_t*)(pool->ptr + pool->availability_offset + 4 * query);
 		switch (pool->type) {
-		case VK_QUERY_TYPE_TIMESTAMP: {
-			if (flags & VK_QUERY_RESULT_WAIT_BIT) {
-				while(!*(volatile uint32_t*)(pool->ptr + pool->availability_offset + 4 * query))
-					;
-			}
-
-			available = *(uint32_t*)(pool->ptr + pool->availability_offset + 4 * query);
-			if (!available && !(flags & VK_QUERY_RESULT_PARTIAL_BIT)) {
-				result = VK_NOT_READY;
-				break;
-
-			}
-
+		case VK_QUERY_TYPE_TIMESTAMP:
 			if (flags & VK_QUERY_RESULT_64_BIT) {
 				*(uint64_t*)dest = *(uint64_t*)src;
 				dest += 8;
@@ -164,32 +166,8 @@ VkResult radv_GetQueryPoolResults(
 				dest += 4;
 			}
 			break;
-		}
 		case VK_QUERY_TYPE_OCCLUSION: {
-			volatile uint64_t const *src64 = (volatile uint64_t const *)src;
-			uint64_t result = 0;
-			int db_count = get_max_db(device);
-			available = 1;
-
-			for (int i = 0; i < db_count; ++i) {
-				uint64_t start, end;
-				do {
-					start = src64[2 * i];
-					end = src64[2 * i + 1];
-				} while ((!(start & (1ull << 63)) || !(end & (1ull << 63))) && (flags & VK_QUERY_RESULT_WAIT_BIT));
-
-				if (!(start & (1ull << 63)) || !(end & (1ull << 63)))
-					available = 0;
-				else {
-					result += end - start;
-				}
-			}
-
-			if (!available && !(flags & VK_QUERY_RESULT_PARTIAL_BIT)) {
-				result = VK_NOT_READY;
-				break;
-
-			}
+			uint64_t result = *(uint64_t*)(src + pool->stride - 16);
 
 			if (flags & VK_QUERY_RESULT_64_BIT) {
 				*(uint64_t*)dest = result;
@@ -205,11 +183,8 @@ VkResult radv_GetQueryPoolResults(
 		}
 
 		if (flags & VK_QUERY_RESULT_WITH_AVAILABILITY_BIT) {
-			if (flags & VK_QUERY_RESULT_64_BIT) {
-				*(uint64_t*)dest = available;
-			} else {
-				*(uint32_t*)dest = available;
-			}
+			*(uint32_t*)dest = available;
+			dest += 4;
 		}
 	}
 
@@ -382,14 +357,11 @@ void radv_CmdEndQuery(
 		radeon_emit(cs, va + 8);
 		radeon_emit(cs, (va + 8) >> 32);
 
-		/* hangs for VK_COMMAND_BUFFER_LEVEL_SECONDARY. */
-		if (cmd_buffer->level == VK_COMMAND_BUFFER_LEVEL_PRIMARY) {
-			radeon_emit(cs, PKT3(PKT3_OCCLUSION_QUERY, 3, 0));
-			radeon_emit(cs, va);
-			radeon_emit(cs, va >> 32);
-			radeon_emit(cs, va + pool->stride - 16);
-			radeon_emit(cs, (va + pool->stride - 16) >> 32);
-		}
+		radeon_emit(cs, PKT3(PKT3_OCCLUSION_QUERY, 3, 0));
+		radeon_emit(cs, va);
+		radeon_emit(cs, va >> 32);
+		radeon_emit(cs, va + pool->stride - 16);
+		radeon_emit(cs, (va + pool->stride - 16) >> 32);
 
 		break;
 	default:

src/amd/vulkan/radv_radeon_winsys.h

@@ -94,7 +94,7 @@ struct radeon_info {
 	uint32_t                    compute_rings;
 	uint32_t                    vce_fw_version;
 	uint32_t                    vce_harvest_config;
-	uint32_t                    clock_crystal_freq; /* in kHz */
+	uint32_t                    clock_crystal_freq;
 
 	/* Kernel info. */
 	uint32_t                    drm_major; /* version */

src/amd/vulkan/radv_wsi.c

@@ -347,20 +347,16 @@ VkResult radv_QueuePresentKHR(
 
 	for (uint32_t i = 0; i < pPresentInfo->swapchainCount; i++) {
 		RADV_FROM_HANDLE(wsi_swapchain, swapchain, pPresentInfo->pSwapchains[i]);
-		VkResult item_result;
 
 		assert(radv_device_from_handle(swapchain->device) == queue->device);
 		if (swapchain->fences[0] == VK_NULL_HANDLE) {
-			item_result = radv_CreateFence(radv_device_to_handle(queue->device),
+			result = radv_CreateFence(radv_device_to_handle(queue->device),
 						  &(VkFenceCreateInfo) {
 							  .sType = VK_STRUCTURE_TYPE_FENCE_CREATE_INFO,
 								  .flags = 0,
 								  }, &swapchain->alloc, &swapchain->fences[0]);
-			if (pPresentInfo->pResults != NULL)
-				pPresentInfo->pResults[i] = item_result;
-			result = result == VK_SUCCESS ? item_result : result;
-			if (item_result != VK_SUCCESS)
-				continue;
+			if (result != VK_SUCCESS)
+				return result;
 		} else {
 			radv_ResetFences(radv_device_to_handle(queue->device),
 					 1, &swapchain->fences[0]);
@@ -376,14 +372,11 @@ VkResult radv_QueuePresentKHR(
 					     pPresentInfo->waitSemaphoreCount, NULL, 0, false, base_fence);
 		fence->submitted = true;
 
-		item_result = swapchain->queue_present(swapchain,
-						       pPresentInfo->pImageIndices[i]);
+		result = swapchain->queue_present(swapchain,
+						  pPresentInfo->pImageIndices[i]);
 		/* TODO: What if one of them returns OUT_OF_DATE? */
-		if (pPresentInfo->pResults != NULL)
-			pPresentInfo->pResults[i] = item_result;
-		result = result == VK_SUCCESS ? item_result : result;
-		if (item_result != VK_SUCCESS)
-			continue;
+		if (result != VK_SUCCESS)
+			return result;
 
 		VkFence last = swapchain->fences[2];
 		swapchain->fences[2] = swapchain->fences[1];

src/amd/vulkan/si_cmd_buffer.c

@@ -291,7 +291,6 @@ void si_init_config(struct radv_physical_device *physical_device,
 		raster_config_1 = 0x0000002a;
 		break;
 	case CHIP_POLARIS11:
-	case CHIP_POLARIS12:
 		raster_config = 0x16000012;
 		raster_config_1 = 0x00000000;
 		break;
@@ -373,15 +372,6 @@ void si_init_config(struct radv_physical_device *physical_device,
 	radeon_set_context_reg(cs, R_028408_VGT_INDX_OFFSET, 0);
 
 	if (physical_device->rad_info.chip_class >= CIK) {
-		/* If this is 0, Bonaire can hang even if GS isn't being used.
-		 * Other chips are unaffected. These are suboptimal values,
-		 * but we don't use on-chip GS.
-		 */
-		radeon_set_context_reg(cs, R_028A44_VGT_GS_ONCHIP_CNTL,
-				       S_028A44_ES_VERTS_PER_SUBGRP(64) |
-				       S_028A44_GS_PRIMS_PER_SUBGRP(4));
-
-		radeon_set_sh_reg(cs, R_00B51C_SPI_SHADER_PGM_RSRC3_LS, S_00B51C_CU_EN(0xffff));
 		radeon_set_sh_reg(cs, R_00B41C_SPI_SHADER_PGM_RSRC3_HS, 0);
 		radeon_set_sh_reg(cs, R_00B31C_SPI_SHADER_PGM_RSRC3_ES, S_00B31C_CU_EN(0xffff));
 		radeon_set_sh_reg(cs, R_00B21C_SPI_SHADER_PGM_RSRC3_GS, S_00B21C_CU_EN(0xffff));
@@ -394,6 +384,7 @@ void si_init_config(struct radv_physical_device *physical_device,
 			 *
 			 * LATE_ALLOC_VS = 2 is the highest safe number.
 			 */
+			radeon_set_sh_reg(cs, R_00B51C_SPI_SHADER_PGM_RSRC3_LS, S_00B51C_CU_EN(0xffff));
 			radeon_set_sh_reg(cs, R_00B118_SPI_SHADER_PGM_RSRC3_VS, S_00B118_CU_EN(0xffff));
 			radeon_set_sh_reg(cs, R_00B11C_SPI_SHADER_LATE_ALLOC_VS, S_00B11C_LIMIT(2));
 		} else {
@@ -402,6 +393,7 @@ void si_init_config(struct radv_physical_device *physical_device,
 			 * - VS can't execute on CU0.
 			 * - If HS writes outputs to LDS, LS can't execute on CU0.
 			 */
+			radeon_set_sh_reg(cs, R_00B51C_SPI_SHADER_PGM_RSRC3_LS, S_00B51C_CU_EN(0xfffe));
 			radeon_set_sh_reg(cs, R_00B118_SPI_SHADER_PGM_RSRC3_VS, S_00B118_CU_EN(0xfffe));
 			radeon_set_sh_reg(cs, R_00B11C_SPI_SHADER_LATE_ALLOC_VS, S_00B11C_LIMIT(31));
 		}

src/amd/vulkan/winsys/amdgpu/radv_amdgpu_surface.c

@@ -274,19 +274,6 @@ static void radv_set_micro_tile_mode(struct radeon_surf *surf,
 		surf->micro_tile_mode = G_009910_MICRO_TILE_MODE(tile_mode);
 }
 
-static unsigned cik_get_macro_tile_index(struct radeon_surf *surf)
-{
-	unsigned index, tileb;
-
-	tileb = 8 * 8 * surf->bpe;
-	tileb = MIN2(surf->tile_split, tileb);
-
-	for (index = 0; tileb > 64; index++)
-		tileb >>= 1;
-
-	assert(index < 16);
-	return index;
-}
 
 static int radv_amdgpu_winsys_surface_init(struct radeon_winsys *_ws,
 					   struct radeon_surf *surf)
@@ -448,7 +435,6 @@ static int radv_amdgpu_winsys_surface_init(struct radeon_winsys *_ws,
 				AddrSurfInfoIn.tileIndex = 10; /* 2D displayable */
 			else
 				AddrSurfInfoIn.tileIndex = 14; /* 2D non-displayable */
-			AddrSurfInfoOut.macroModeIndex = cik_get_macro_tile_index(surf);
 		}
 	}
 

src/amd/vulkan/winsys/amdgpu/radv_amdgpu_winsys.c

@@ -106,7 +106,6 @@ get_chip_name(enum radeon_family family)
 	case CHIP_FIJI: return "AMD RADV FIJI";
 	case CHIP_POLARIS10: return "AMD RADV POLARIS10";
 	case CHIP_POLARIS11: return "AMD RADV POLARIS11";
-	case CHIP_POLARIS12: return "AMD RADV POLARIS12";
 	case CHIP_STONEY: return "AMD RADV STONEY";
 	default: return "AMD RADV unknown";
 	}
@@ -271,10 +270,6 @@ do_winsys_init(struct radv_amdgpu_winsys *ws, int fd)
 		ws->family = FAMILY_VI;
 		ws->rev_id = VI_POLARIS11_M_A0;
 		break;
-	case CHIP_POLARIS12:
-		ws->family = FAMILY_VI;
-		ws->rev_id = VI_POLARIS12_V_A0;
-		break;
 	default:
 		fprintf(stderr, "amdgpu: Unknown family.\n");
 		goto fail;

src/compiler/Android.glsl.mk

@@ -36,7 +36,6 @@ include $(CLEAR_VARS)
 LOCAL_SRC_FILES := \
 	$(LIBGLCPP_FILES) \
 	$(LIBGLSL_FILES) \
-	$(LIBGLSL_SHADER_CACHE_FILES)
 
 LOCAL_C_INCLUDES := \
 	$(MESA_TOP)/src/mapi \

src/compiler/Makefile.glsl.am

@@ -73,6 +73,7 @@ glsl_tests_general_ir_test_SOURCES =			\
 	glsl/tests/builtin_variable_test.cpp		\
 	glsl/tests/invalidate_locations_test.cpp	\
 	glsl/tests/general_ir_test.cpp			\
+	glsl/tests/lower_int64_test.cpp			\
 	glsl/tests/opt_add_neg_to_sub_test.cpp		\
 	glsl/tests/varyings_test.cpp
 glsl_tests_general_ir_test_CFLAGS =			\

src/compiler/Makefile.sources

@@ -20,8 +20,11 @@ LIBGLSL_FILES = \
 	glsl/blob.c \
 	glsl/blob.h \
 	glsl/builtin_functions.cpp \
+	glsl/builtin_functions.h \
+	glsl/builtin_int64.h \
 	glsl/builtin_types.cpp \
 	glsl/builtin_variables.cpp \
+	glsl/generate_ir.cpp \
 	glsl/glsl_parser_extras.cpp \
 	glsl/glsl_parser_extras.h \
 	glsl/glsl_symbol_table.cpp \
@@ -88,6 +91,7 @@ LIBGLSL_FILES = \
 	glsl/lower_distance.cpp \
 	glsl/lower_if_to_cond_assign.cpp \
 	glsl/lower_instructions.cpp \
+	glsl/lower_int64.cpp \
 	glsl/lower_jumps.cpp \
 	glsl/lower_mat_op_to_vec.cpp \
 	glsl/lower_noise.cpp \

src/compiler/builtin_type_macros.h

@@ -78,6 +78,16 @@ DECL_TYPE(dmat3x4, GL_DOUBLE_MAT3x4, GLSL_TYPE_DOUBLE, 4, 3)
 DECL_TYPE(dmat4x2, GL_DOUBLE_MAT4x2, GLSL_TYPE_DOUBLE, 2, 4)
 DECL_TYPE(dmat4x3, GL_DOUBLE_MAT4x3, GLSL_TYPE_DOUBLE, 3, 4)
 
+DECL_TYPE(int64_t,  GL_INT64_ARB,          GLSL_TYPE_INT64,   1, 1)
+DECL_TYPE(i64vec2,  GL_INT64_VEC2_ARB,     GLSL_TYPE_INT64,   2, 1)
+DECL_TYPE(i64vec3,  GL_INT64_VEC3_ARB,     GLSL_TYPE_INT64,   3, 1)
+DECL_TYPE(i64vec4,  GL_INT64_VEC4_ARB,     GLSL_TYPE_INT64,   4, 1)
+
+DECL_TYPE(uint64_t, GL_UNSIGNED_INT64_ARB,      GLSL_TYPE_UINT64, 1, 1)
+DECL_TYPE(u64vec2,  GL_UNSIGNED_INT64_VEC2_ARB, GLSL_TYPE_UINT64, 2, 1)
+DECL_TYPE(u64vec3,  GL_UNSIGNED_INT64_VEC3_ARB, GLSL_TYPE_UINT64, 3, 1)
+DECL_TYPE(u64vec4,  GL_UNSIGNED_INT64_VEC4_ARB, GLSL_TYPE_UINT64, 4, 1)
+
 DECL_TYPE(sampler,           GL_SAMPLER_1D,                   GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_1D,   0, 0, GLSL_TYPE_VOID)
 DECL_TYPE(sampler1D,         GL_SAMPLER_1D,                   GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_1D,   0, 0, GLSL_TYPE_FLOAT)
 DECL_TYPE(sampler2D,         GL_SAMPLER_2D,                   GLSL_TYPE_SAMPLER, GLSL_SAMPLER_DIM_2D,   0, 0, GLSL_TYPE_FLOAT)

src/compiler/glsl/ast.h

@@ -195,6 +195,8 @@ enum ast_operators {
    ast_float_constant,
    ast_bool_constant,
    ast_double_constant,
+   ast_int64_constant,
+   ast_uint64_constant,
 
    ast_sequence,
    ast_aggregate
@@ -255,6 +257,8 @@ public:
       unsigned uint_constant;
       int bool_constant;
       double double_constant;
+      uint64_t uint64_constant;
+      int64_t int64_constant;
    } primary_expression;
 
 

src/compiler/glsl/ast_function.cpp

@@ -27,6 +27,7 @@
 #include "ir.h"
 #include "main/core.h" /* for MIN2 */
 #include "main/shaderobj.h"
+#include "builtin_functions.h"
 
 static ir_rvalue *
 convert_component(ir_rvalue *src, const glsl_type *desired_type);
@@ -760,6 +761,12 @@ convert_component(ir_rvalue *src, const glsl_type *desired_type)
       case GLSL_TYPE_DOUBLE:
          result = new(ctx) ir_expression(ir_unop_d2u, src);
          break;
+      case GLSL_TYPE_UINT64:
+         result = new(ctx) ir_expression(ir_unop_u642u, src);
+         break;
+      case GLSL_TYPE_INT64:
+         result = new(ctx) ir_expression(ir_unop_i642u, src);
+         break;
       }
       break;
    case GLSL_TYPE_INT:
@@ -776,6 +783,12 @@ convert_component(ir_rvalue *src, const glsl_type *desired_type)
       case GLSL_TYPE_DOUBLE:
          result = new(ctx) ir_expression(ir_unop_d2i, src);
          break;
+      case GLSL_TYPE_UINT64:
+         result = new(ctx) ir_expression(ir_unop_u642i, src);
+         break;
+      case GLSL_TYPE_INT64:
+         result = new(ctx) ir_expression(ir_unop_i642i, src);
+         break;
       }
       break;
    case GLSL_TYPE_FLOAT:
@@ -792,6 +805,12 @@ convert_component(ir_rvalue *src, const glsl_type *desired_type)
       case GLSL_TYPE_DOUBLE:
          result = new(ctx) ir_expression(ir_unop_d2f, desired_type, src, NULL);
          break;
+      case GLSL_TYPE_UINT64:
+         result = new(ctx) ir_expression(ir_unop_u642f, desired_type, src, NULL);
+         break;
+      case GLSL_TYPE_INT64:
+         result = new(ctx) ir_expression(ir_unop_i642f, desired_type, src, NULL);
+         break;
       }
       break;
    case GLSL_TYPE_BOOL:
@@ -810,6 +829,14 @@ convert_component(ir_rvalue *src, const glsl_type *desired_type)
       case GLSL_TYPE_DOUBLE:
          result = new(ctx) ir_expression(ir_unop_d2b, desired_type, src, NULL);
          break;
+      case GLSL_TYPE_UINT64:
+         result = new(ctx) ir_expression(ir_unop_i642b,
+                                         new(ctx) ir_expression(ir_unop_u642i64,
+                                                                src));
+         break;
+      case GLSL_TYPE_INT64:
+         result = new(ctx) ir_expression(ir_unop_i642b, desired_type, src, NULL);
+         break;
       }
       break;
    case GLSL_TYPE_DOUBLE:
@@ -828,7 +855,60 @@ convert_component(ir_rvalue *src, const glsl_type *desired_type)
       case GLSL_TYPE_FLOAT:
          result = new(ctx) ir_expression(ir_unop_f2d, desired_type, src, NULL);
          break;
+      case GLSL_TYPE_UINT64:
+         result = new(ctx) ir_expression(ir_unop_u642d, desired_type, src, NULL);
+         break;
+      case GLSL_TYPE_INT64:
+         result = new(ctx) ir_expression(ir_unop_i642d, desired_type, src, NULL);
+         break;
       }
+      break;
+   case GLSL_TYPE_UINT64:
+      switch (b) {
+      case GLSL_TYPE_INT:
+         result = new(ctx) ir_expression(ir_unop_i2u64, src);
+         break;
+      case GLSL_TYPE_UINT:
+         result = new(ctx) ir_expression(ir_unop_u2u64, src);
+         break;
+      case GLSL_TYPE_BOOL:
+         result = new(ctx) ir_expression(ir_unop_i642u64,
+                                         new(ctx) ir_expression(ir_unop_b2i64,
+                                                                src));
+         break;
+      case GLSL_TYPE_FLOAT:
+         result = new(ctx) ir_expression(ir_unop_f2u64, src);
+         break;
+      case GLSL_TYPE_DOUBLE:
+         result = new(ctx) ir_expression(ir_unop_d2u64, src);
+         break;
+      case GLSL_TYPE_INT64:
+         result = new(ctx) ir_expression(ir_unop_i642u64, src);
+         break;
+      }
+      break;
+   case GLSL_TYPE_INT64:
+      switch (b) {
+      case GLSL_TYPE_INT:
+         result = new(ctx) ir_expression(ir_unop_i2i64, src);
+         break;
+      case GLSL_TYPE_UINT:
+         result = new(ctx) ir_expression(ir_unop_u2i64, src);
+         break;
+      case GLSL_TYPE_BOOL:
+         result = new(ctx) ir_expression(ir_unop_b2i64, src);
+         break;
+      case GLSL_TYPE_FLOAT:
+         result = new(ctx) ir_expression(ir_unop_f2i64, src);
+         break;
+      case GLSL_TYPE_DOUBLE:
+         result = new(ctx) ir_expression(ir_unop_d2i64, src);
+         break;
+      case GLSL_TYPE_UINT64:
+         result = new(ctx) ir_expression(ir_unop_u642i64, src);
+         break;
+      }
+      break;
    }
 
    assert(result != NULL);
@@ -1260,6 +1340,12 @@ emit_inline_vector_constructor(const glsl_type *type,
                case GLSL_TYPE_BOOL:
                   data.b[i + base_component] = c->get_bool_component(i);
                   break;
+               case GLSL_TYPE_UINT64:
+                  data.u64[i + base_component] = c->get_uint64_component(i);
+                  break;
+               case GLSL_TYPE_INT64:
+                  data.i64[i + base_component] = c->get_int64_component(i);
+                  break;
                default:
                   assert(!"Should not get here.");
                   break;
@@ -1267,8 +1353,7 @@ emit_inline_vector_constructor(const glsl_type *type,
             }
 
             /* Mask of fields to be written in the assignment. */
-            constant_mask |=
-               ((1U << rhs_components) - 1) << base_lhs_component;
+            constant_mask |= ((1U << rhs_components) - 1) << base_lhs_component;
             constant_components += rhs_components;
 
             base_component += rhs_components;

src/compiler/glsl/ast_to_hir.cpp

@@ -58,6 +58,7 @@
 #include "main/shaderobj.h"
 #include "ir.h"
 #include "ir_builder.h"
+#include "builtin_functions.h"
 
 using namespace ir_builder;
 
@@ -259,6 +260,26 @@ get_implicit_conversion_operation(const glsl_type *to, const glsl_type *from,
       case GLSL_TYPE_INT: return ir_unop_i2d;
       case GLSL_TYPE_UINT: return ir_unop_u2d;
       case GLSL_TYPE_FLOAT: return ir_unop_f2d;
+      case GLSL_TYPE_INT64: return ir_unop_i642d;
+      case GLSL_TYPE_UINT64: return ir_unop_u642d;
+      default: return (ir_expression_operation)0;
+      }
+
+   case GLSL_TYPE_UINT64:
+      if (!state->has_int64())
+         return (ir_expression_operation)0;
+      switch (from->base_type) {
+      case GLSL_TYPE_INT: return ir_unop_i2u64;
+      case GLSL_TYPE_UINT: return ir_unop_u2u64;
+      case GLSL_TYPE_INT64: return ir_unop_i642u64;
+      default: return (ir_expression_operation)0;
+      }
+
+   case GLSL_TYPE_INT64:
+      if (!state->has_int64())
+         return (ir_expression_operation)0;
+      switch (from->base_type) {
+      case GLSL_TYPE_INT: return ir_unop_i2i64;
       default: return (ir_expression_operation)0;
       }
 
@@ -513,12 +534,12 @@ bit_logic_result_type(ir_rvalue * &value_a, ir_rvalue * &value_b,
     *     (|). The operands must be of type signed or unsigned integers or
     *     integer vectors."
     */
-   if (!type_a->is_integer()) {
+   if (!type_a->is_integer_32_64()) {
       _mesa_glsl_error(loc, state, "LHS of `%s' must be an integer",
                         ast_expression::operator_string(op));
       return glsl_type::error_type;
    }
-   if (!type_b->is_integer()) {
+   if (!type_b->is_integer_32_64()) {
       _mesa_glsl_error(loc, state, "RHS of `%s' must be an integer",
                        ast_expression::operator_string(op));
       return glsl_type::error_type;
@@ -599,11 +620,11 @@ modulus_result_type(ir_rvalue * &value_a, ir_rvalue * &value_b,
     *    "The operator modulus (%) operates on signed or unsigned integers or
     *    integer vectors."
     */
-   if (!type_a->is_integer()) {
+   if (!type_a->is_integer_32_64()) {
       _mesa_glsl_error(loc, state, "LHS of operator %% must be an integer");
       return glsl_type::error_type;
    }
-   if (!type_b->is_integer()) {
+   if (!type_b->is_integer_32_64()) {
       _mesa_glsl_error(loc, state, "RHS of operator %% must be an integer");
       return glsl_type::error_type;
    }
@@ -721,7 +742,7 @@ shift_result_type(const struct glsl_type *type_a,
     *     must be signed or unsigned integers or integer vectors. One operand
     *     can be signed while the other is unsigned."
     */
-   if (!type_a->is_integer()) {
+   if (!type_a->is_integer_32_64()) {
       _mesa_glsl_error(loc, state, "LHS of operator %s must be an integer or "
                        "integer vector", ast_expression::operator_string(op));
      return glsl_type::error_type;
@@ -1093,6 +1114,8 @@ do_comparison(void *mem_ctx, int operation, ir_rvalue *op0, ir_rvalue *op1)
    case GLSL_TYPE_INT:
    case GLSL_TYPE_BOOL:
    case GLSL_TYPE_DOUBLE:
+   case GLSL_TYPE_UINT64:
+   case GLSL_TYPE_INT64:
       return new(mem_ctx) ir_expression(operation, op0, op1);
 
    case GLSL_TYPE_ARRAY: {
@@ -1256,6 +1279,10 @@ constant_one_for_inc_dec(void *ctx, const glsl_type *type)
       return new(ctx) ir_constant((unsigned) 1);
    case GLSL_TYPE_INT:
       return new(ctx) ir_constant(1);
+   case GLSL_TYPE_UINT64:
+      return new(ctx) ir_constant((uint64_t) 1);
+   case GLSL_TYPE_INT64:
+      return new(ctx) ir_constant((int64_t) 1);
    default:
    case GLSL_TYPE_FLOAT:
       return new(ctx) ir_constant(1.0f);
@@ -1535,7 +1562,7 @@ ast_expression::do_hir(exec_list *instructions,
          error_emitted = true;
       }
 
-      if (!op[0]->type->is_integer()) {
+      if (!op[0]->type->is_integer_32_64()) {
          _mesa_glsl_error(&loc, state, "operand of `~' must be an integer");
          error_emitted = true;
       }
@@ -2006,6 +2033,14 @@ ast_expression::do_hir(exec_list *instructions,
       result = new(ctx) ir_constant(this->primary_expression.double_constant);
       break;
 
+   case ast_uint64_constant:
+      result = new(ctx) ir_constant(this->primary_expression.uint64_constant);
+      break;
+
+   case ast_int64_constant:
+      result = new(ctx) ir_constant(this->primary_expression.int64_constant);
+      break;
+
    case ast_sequence: {
       /* It should not be possible to generate a sequence in the AST without
        * any expressions in it.
@@ -2132,6 +2167,8 @@ ast_expression::has_sequence_subexpression() const
    case ast_float_constant:
    case ast_bool_constant:
    case ast_double_constant:
+   case ast_int64_constant:
+   case ast_uint64_constant:
       return false;
 
    case ast_aggregate:
@@ -3828,6 +3865,8 @@ apply_type_qualifier_to_variable(const struct ast_type_qualifier *qual,
                           "varying variables may not be of type struct");
          break;
       case GLSL_TYPE_DOUBLE:
+      case GLSL_TYPE_UINT64:
+      case GLSL_TYPE_INT64:
          break;
       default:
          _mesa_glsl_error(loc, state, "illegal type for a varying variable");
@@ -4886,6 +4925,9 @@ ast_declarator_list::hir(exec_list *instructions,
             switch (check_type->base_type) {
             case GLSL_TYPE_FLOAT:
             break;
+            case GLSL_TYPE_UINT64:
+            case GLSL_TYPE_INT64:
+               break;
             case GLSL_TYPE_UINT:
             case GLSL_TYPE_INT:
                if (state->is_version(120, 300))
@@ -5220,13 +5262,11 @@ ast_declarator_list::hir(exec_list *instructions,
           *     sized by an earlier input primitive layout qualifier, when
           *     present, as per the following table."
           */
-         const enum ir_variable_mode mode = (const enum ir_variable_mode)
-            (earlier == NULL ? var->data.mode : earlier->data.mode);
          const bool implicitly_sized =
-            (mode == ir_var_shader_in &&
+            (var->data.mode == ir_var_shader_in &&
              state->stage >= MESA_SHADER_TESS_CTRL &&
              state->stage <= MESA_SHADER_GEOMETRY) ||
-            (mode == ir_var_shader_out &&
+            (var->data.mode == ir_var_shader_out &&
              state->stage == MESA_SHADER_TESS_CTRL);
 
          if (t->is_unsized_array() && !implicitly_sized)
@@ -7863,9 +7903,10 @@ ast_interface_block::hir(exec_list *instructions,
          }
 
          if (var->type->is_unsized_array()) {
-            if (var->is_in_shader_storage_block() &&
-                is_unsized_array_last_element(var)) {
-               var->data.from_ssbo_unsized_array = true;
+            if (var->is_in_shader_storage_block()) {
+               if (is_unsized_array_last_element(var)) {
+                  var->data.from_ssbo_unsized_array = true;
+               }
             } else {
                /* From GLSL ES 3.10 spec, section 4.1.9 "Arrays":
                 *
@@ -7873,10 +7914,6 @@ ast_interface_block::hir(exec_list *instructions,
                 * block and the size is not specified at compile-time, it is
                 * sized at run-time. In all other cases, arrays are sized only
                 * at compile-time."
-                *
-                * In desktop GLSL it is allowed to have unsized-arrays that are
-                * not last, as long as we can determine that they are implicitly
-                * sized.
                 */
                if (state->es_shader) {
                   _mesa_glsl_error(&loc, state, "unsized array `%s' "

src/compiler/glsl/builtin_functions.cpp

@@ -61,6 +61,7 @@
 #include "glsl_parser_extras.h"
 #include "program/prog_instruction.h"
 #include <math.h>
+#include "builtin_functions.h"
 
 #define M_PIf   ((float) M_PI)
 #define M_PI_2f ((float) M_PI_2)
@@ -470,6 +471,13 @@ shader_clock(const _mesa_glsl_parse_state *state)
    return state->ARB_shader_clock_enable;
 }
 
+static bool
+shader_clock_int64(const _mesa_glsl_parse_state *state)
+{
+   return state->ARB_shader_clock_enable &&
+          state->ARB_gpu_shader_int64_enable;
+}
+
 static bool
 shader_storage_buffer_object(const _mesa_glsl_parse_state *state)
 {
@@ -532,15 +540,21 @@ fp64(const _mesa_glsl_parse_state *state)
 }
 
 static bool
-compute_shader(const _mesa_glsl_parse_state *state)
+int64(const _mesa_glsl_parse_state *state)
 {
-   return state->stage == MESA_SHADER_COMPUTE;
+   return state->has_int64();
 }
 
 static bool
-compute_shader_supported(const _mesa_glsl_parse_state *state)
+int64_fp64(const _mesa_glsl_parse_state *state)
 {
-   return state->has_compute_shader();
+   return state->has_int64() && state->has_double();
+}
+
+static bool
+compute_shader(const _mesa_glsl_parse_state *state)
+{
+   return state->stage == MESA_SHADER_COMPUTE;
 }
 
 static bool
@@ -562,6 +576,11 @@ vote(const _mesa_glsl_parse_state *state)
    return state->ARB_shader_group_vote_enable;
 }
 
+static bool
+integer_functions_supported(const _mesa_glsl_parse_state *state)
+{
+   return state->extensions->MESA_shader_integer_functions;
+}
 /** @} */
 
 /******************************************************************************/
@@ -729,6 +748,12 @@ private:
    B1(floatBitsToUint)
    B1(intBitsToFloat)
    B1(uintBitsToFloat)
+
+   BA1(doubleBitsToInt64)
+   BA1(doubleBitsToUint64)
+   BA1(int64BitsToDouble)
+   BA1(uint64BitsToDouble)
+
    ir_function_signature *_packUnorm2x16(builtin_available_predicate avail);
    ir_function_signature *_packSnorm2x16(builtin_available_predicate avail);
    ir_function_signature *_packUnorm4x8(builtin_available_predicate avail);
@@ -741,6 +766,10 @@ private:
    ir_function_signature *_unpackHalf2x16(builtin_available_predicate avail);
    ir_function_signature *_packDouble2x32(builtin_available_predicate avail);
    ir_function_signature *_unpackDouble2x32(builtin_available_predicate avail);
+   ir_function_signature *_packInt2x32(builtin_available_predicate avail);
+   ir_function_signature *_unpackInt2x32(builtin_available_predicate avail);
+   ir_function_signature *_packUint2x32(builtin_available_predicate avail);
+   ir_function_signature *_unpackUint2x32(builtin_available_predicate avail);
 
    BA1(length)
    BA1(distance);
@@ -1104,15 +1133,15 @@ builtin_builder::create_intrinsics()
                                           ir_intrinsic_group_memory_barrier),
                 NULL);
    add_function("__intrinsic_memory_barrier_atomic_counter",
-                _memory_barrier_intrinsic(compute_shader_supported,
+                _memory_barrier_intrinsic(compute_shader,
                                           ir_intrinsic_memory_barrier_atomic_counter),
                 NULL);
    add_function("__intrinsic_memory_barrier_buffer",
-                _memory_barrier_intrinsic(compute_shader_supported,
+                _memory_barrier_intrinsic(compute_shader,
                                           ir_intrinsic_memory_barrier_buffer),
                 NULL);
    add_function("__intrinsic_memory_barrier_image",
-                _memory_barrier_intrinsic(compute_shader_supported,
+                _memory_barrier_intrinsic(compute_shader,
                                           ir_intrinsic_memory_barrier_image),
                 NULL);
    add_function("__intrinsic_memory_barrier_shared",
@@ -1190,7 +1219,7 @@ builtin_builder::create_builtins()
                 _##NAME(glsl_type::ivec4_type), \
                 NULL);
 
-#define FID(NAME)                                \
+#define FI64(NAME)                                \
    add_function(#NAME,                          \
                 _##NAME(always_available, glsl_type::float_type), \
                 _##NAME(always_available, glsl_type::vec2_type),  \
@@ -1204,6 +1233,10 @@ builtin_builder::create_builtins()
                 _##NAME(fp64, glsl_type::dvec2_type),  \
                 _##NAME(fp64, glsl_type::dvec3_type),  \
                 _##NAME(fp64, glsl_type::dvec4_type),  \
+                _##NAME(int64, glsl_type::int64_t_type), \
+                _##NAME(int64, glsl_type::i64vec2_type),  \
+                _##NAME(int64, glsl_type::i64vec3_type),  \
+                _##NAME(int64, glsl_type::i64vec4_type),  \
                 NULL);
 
 #define FIUD_VEC(NAME)                                            \
@@ -1222,6 +1255,14 @@ builtin_builder::create_builtins()
                 _##NAME(fp64, glsl_type::dvec2_type),  \
                 _##NAME(fp64, glsl_type::dvec3_type),  \
                 _##NAME(fp64, glsl_type::dvec4_type),  \
+                _##NAME(int64, glsl_type::int64_t_type), \
+                _##NAME(int64, glsl_type::i64vec2_type),  \
+                _##NAME(int64, glsl_type::i64vec3_type),  \
+                _##NAME(int64, glsl_type::i64vec4_type),  \
+                _##NAME(int64, glsl_type::uint64_t_type), \
+                _##NAME(int64, glsl_type::u64vec2_type),  \
+                _##NAME(int64, glsl_type::u64vec3_type),  \
+                _##NAME(int64, glsl_type::u64vec4_type),  \
                 NULL);
 
 #define IU(NAME)                                \
@@ -1258,6 +1299,14 @@ builtin_builder::create_builtins()
                 _##NAME(fp64, glsl_type::dvec2_type), \
                 _##NAME(fp64, glsl_type::dvec3_type), \
                 _##NAME(fp64, glsl_type::dvec4_type), \
+                _##NAME(int64, glsl_type::int64_t_type), \
+                _##NAME(int64, glsl_type::i64vec2_type),  \
+                _##NAME(int64, glsl_type::i64vec3_type),  \
+                _##NAME(int64, glsl_type::i64vec4_type),  \
+                _##NAME(int64, glsl_type::uint64_t_type), \
+                _##NAME(int64, glsl_type::u64vec2_type),  \
+                _##NAME(int64, glsl_type::u64vec3_type),  \
+                _##NAME(int64, glsl_type::u64vec4_type),  \
                 NULL);
 
 #define FIUD2_MIXED(NAME)                                                                 \
@@ -1296,6 +1345,21 @@ builtin_builder::create_builtins()
                 _##NAME(fp64, glsl_type::dvec2_type, glsl_type::dvec2_type),           \
                 _##NAME(fp64, glsl_type::dvec3_type, glsl_type::dvec3_type),           \
                 _##NAME(fp64, glsl_type::dvec4_type, glsl_type::dvec4_type),           \
+                                                                        \
+                _##NAME(int64, glsl_type::int64_t_type, glsl_type::int64_t_type),           \
+                _##NAME(int64, glsl_type::i64vec2_type, glsl_type::int64_t_type),           \
+                _##NAME(int64, glsl_type::i64vec3_type, glsl_type::int64_t_type),           \
+                _##NAME(int64, glsl_type::i64vec4_type, glsl_type::int64_t_type),           \
+                _##NAME(int64, glsl_type::i64vec2_type, glsl_type::i64vec2_type),           \
+                _##NAME(int64, glsl_type::i64vec3_type, glsl_type::i64vec3_type),           \
+                _##NAME(int64, glsl_type::i64vec4_type, glsl_type::i64vec4_type),           \
+                _##NAME(int64, glsl_type::uint64_t_type, glsl_type::uint64_t_type),           \
+                _##NAME(int64, glsl_type::u64vec2_type, glsl_type::uint64_t_type),           \
+                _##NAME(int64, glsl_type::u64vec3_type, glsl_type::uint64_t_type),           \
+                _##NAME(int64, glsl_type::u64vec4_type, glsl_type::uint64_t_type),           \
+                _##NAME(int64, glsl_type::u64vec2_type, glsl_type::u64vec2_type),           \
+                _##NAME(int64, glsl_type::u64vec3_type, glsl_type::u64vec3_type),           \
+                _##NAME(int64, glsl_type::u64vec4_type, glsl_type::u64vec4_type),           \
                 NULL);
 
    F(radians)
@@ -1330,8 +1394,8 @@ builtin_builder::create_builtins()
    F(log2)
    FD(sqrt)
    FD(inversesqrt)
-   FID(abs)
-   FID(sign)
+   FI64(abs)
+   FI64(sign)
    FD(floor)
    FD(trunc)
    FD(round)
@@ -1408,6 +1472,16 @@ builtin_builder::create_builtins()
                 _mix_sel(shader_integer_mix, glsl_type::bvec2_type, glsl_type::bvec2_type),
                 _mix_sel(shader_integer_mix, glsl_type::bvec3_type, glsl_type::bvec3_type),
                 _mix_sel(shader_integer_mix, glsl_type::bvec4_type, glsl_type::bvec4_type),
+
+                _mix_sel(int64, glsl_type::int64_t_type, glsl_type::bool_type),
+                _mix_sel(int64, glsl_type::i64vec2_type, glsl_type::bvec2_type),
+                _mix_sel(int64, glsl_type::i64vec3_type, glsl_type::bvec3_type),
+                _mix_sel(int64, glsl_type::i64vec4_type, glsl_type::bvec4_type),
+
+                _mix_sel(int64, glsl_type::uint64_t_type,  glsl_type::bool_type),
+                _mix_sel(int64, glsl_type::u64vec2_type, glsl_type::bvec2_type),
+                _mix_sel(int64, glsl_type::u64vec3_type, glsl_type::bvec3_type),
+                _mix_sel(int64, glsl_type::u64vec4_type, glsl_type::bvec4_type),
                 NULL);
 
    add_function("step",
@@ -1466,6 +1540,34 @@ builtin_builder::create_builtins()
                 _uintBitsToFloat(glsl_type::uvec4_type),
                 NULL);
 
+   add_function("doubleBitsToInt64",
+                _doubleBitsToInt64(int64_fp64, glsl_type::double_type),
+                _doubleBitsToInt64(int64_fp64, glsl_type::dvec2_type),
+                _doubleBitsToInt64(int64_fp64, glsl_type::dvec3_type),
+                _doubleBitsToInt64(int64_fp64, glsl_type::dvec4_type),
+                NULL);
+
+   add_function("doubleBitsToUint64",
+                _doubleBitsToUint64(int64_fp64, glsl_type::double_type),
+                _doubleBitsToUint64(int64_fp64, glsl_type::dvec2_type),
+                _doubleBitsToUint64(int64_fp64, glsl_type::dvec3_type),
+                _doubleBitsToUint64(int64_fp64, glsl_type::dvec4_type),
+                NULL);
+
+   add_function("int64BitsToDouble",
+                _int64BitsToDouble(int64_fp64, glsl_type::int64_t_type),
+                _int64BitsToDouble(int64_fp64, glsl_type::i64vec2_type),
+                _int64BitsToDouble(int64_fp64, glsl_type::i64vec3_type),
+                _int64BitsToDouble(int64_fp64, glsl_type::i64vec4_type),
+                NULL);
+
+   add_function("uint64BitsToDouble",
+                _uint64BitsToDouble(int64_fp64, glsl_type::uint64_t_type),
+                _uint64BitsToDouble(int64_fp64, glsl_type::u64vec2_type),
+                _uint64BitsToDouble(int64_fp64, glsl_type::u64vec3_type),
+                _uint64BitsToDouble(int64_fp64, glsl_type::u64vec4_type),
+                NULL);
+
    add_function("packUnorm2x16",   _packUnorm2x16(shader_packing_or_es3_or_gpu_shader5),   NULL);
    add_function("packSnorm2x16",   _packSnorm2x16(shader_packing_or_es3),                  NULL);
    add_function("packUnorm4x8",    _packUnorm4x8(shader_packing_or_es31_or_gpu_shader5),   NULL);
@@ -1479,6 +1581,10 @@ builtin_builder::create_builtins()
    add_function("packDouble2x32",    _packDouble2x32(fp64),                   NULL);
    add_function("unpackDouble2x32",  _unpackDouble2x32(fp64),                 NULL);
 
+   add_function("packInt2x32",     _packInt2x32(int64),                    NULL);
+   add_function("unpackInt2x32",   _unpackInt2x32(int64),                  NULL);
+   add_function("packUint2x32",    _packUint2x32(int64),                   NULL);
+   add_function("unpackUint2x32",  _unpackUint2x32(int64),                 NULL);
 
    FD(length)
    FD(distance)
@@ -2964,15 +3070,15 @@ builtin_builder::create_builtins()
                 NULL);
    add_function("memoryBarrierAtomicCounter",
                 _memory_barrier("__intrinsic_memory_barrier_atomic_counter",
-                                compute_shader_supported),
+                                compute_shader),
                 NULL);
    add_function("memoryBarrierBuffer",
                 _memory_barrier("__intrinsic_memory_barrier_buffer",
-                                compute_shader_supported),
+                                compute_shader),
                 NULL);
    add_function("memoryBarrierImage",
                 _memory_barrier("__intrinsic_memory_barrier_image",
-                                compute_shader_supported),
+                                compute_shader),
                 NULL);
    add_function("memoryBarrierShared",
                 _memory_barrier("__intrinsic_memory_barrier_shared",
@@ -2984,10 +3090,39 @@ builtin_builder::create_builtins()
                               glsl_type::uvec2_type),
                 NULL);
 
+   add_function("clockARB",
+                _shader_clock(shader_clock_int64,
+                              glsl_type::uint64_t_type),
+                NULL);
+
    add_function("anyInvocationARB", _vote(ir_unop_vote_any), NULL);
    add_function("allInvocationsARB", _vote(ir_unop_vote_all), NULL);
    add_function("allInvocationsEqualARB", _vote(ir_unop_vote_eq), NULL);
 
+   add_function("__builtin_idiv64",
+                generate_ir::idiv64(mem_ctx, integer_functions_supported),
+                NULL);
+
+   add_function("__builtin_imod64",
+                generate_ir::imod64(mem_ctx, integer_functions_supported),
+                NULL);
+
+   add_function("__builtin_sign64",
+                generate_ir::sign64(mem_ctx, integer_functions_supported),
+                NULL);
+
+   add_function("__builtin_udiv64",
+                generate_ir::udiv64(mem_ctx, integer_functions_supported),
+                NULL);
+
+   add_function("__builtin_umod64",
+                generate_ir::umod64(mem_ctx, integer_functions_supported),
+                NULL);
+
+   add_function("__builtin_umul64",
+                generate_ir::umul64(mem_ctx, integer_functions_supported),
+                NULL);
+
 #undef F
 #undef FI
 #undef FIUD_VEC
@@ -3858,6 +3993,42 @@ builtin_builder::_uintBitsToFloat(const glsl_type *type)
    return sig;
 }
 
+ir_function_signature *
+builtin_builder::_doubleBitsToInt64(builtin_available_predicate avail, const glsl_type *type)
+{
+   ir_variable *x = in_var(type, "x");
+   MAKE_SIG(glsl_type::i64vec(type->vector_elements), avail, 1, x);
+   body.emit(ret(bitcast_d2i64(x)));
+   return sig;
+}
+
+ir_function_signature *
+builtin_builder::_doubleBitsToUint64(builtin_available_predicate avail, const glsl_type *type)
+{
+   ir_variable *x = in_var(type, "x");
+   MAKE_SIG(glsl_type::u64vec(type->vector_elements), avail, 1, x);
+   body.emit(ret(bitcast_d2u64(x)));
+   return sig;
+}
+
+ir_function_signature *
+builtin_builder::_int64BitsToDouble(builtin_available_predicate avail, const glsl_type *type)
+{
+   ir_variable *x = in_var(type, "x");
+   MAKE_SIG(glsl_type::dvec(type->vector_elements), avail, 1, x);
+   body.emit(ret(bitcast_i642d(x)));
+   return sig;
+}
+
+ir_function_signature *
+builtin_builder::_uint64BitsToDouble(builtin_available_predicate avail, const glsl_type *type)
+{
+   ir_variable *x = in_var(type, "x");
+   MAKE_SIG(glsl_type::dvec(type->vector_elements), avail, 1, x);
+   body.emit(ret(bitcast_u642d(x)));
+   return sig;
+}
+
 ir_function_signature *
 builtin_builder::_packUnorm2x16(builtin_available_predicate avail)
 {
@@ -3967,6 +4138,42 @@ builtin_builder::_unpackDouble2x32(builtin_available_predicate avail)
    return sig;
 }
 
+ir_function_signature *
+builtin_builder::_packInt2x32(builtin_available_predicate avail)
+{
+   ir_variable *v = in_var(glsl_type::ivec2_type, "v");
+   MAKE_SIG(glsl_type::int64_t_type, avail, 1, v);
+   body.emit(ret(expr(ir_unop_pack_int_2x32, v)));
+   return sig;
+}
+
+ir_function_signature *
+builtin_builder::_unpackInt2x32(builtin_available_predicate avail)
+{
+   ir_variable *p = in_var(glsl_type::int64_t_type, "p");
+   MAKE_SIG(glsl_type::ivec2_type, avail, 1, p);
+   body.emit(ret(expr(ir_unop_unpack_int_2x32, p)));
+   return sig;
+}
+
+ir_function_signature *
+builtin_builder::_packUint2x32(builtin_available_predicate avail)
+{
+   ir_variable *v = in_var(glsl_type::uvec2_type, "v");
+   MAKE_SIG(glsl_type::uint64_t_type, avail, 1, v);
+   body.emit(ret(expr(ir_unop_pack_uint_2x32, v)));
+   return sig;
+}
+
+ir_function_signature *
+builtin_builder::_unpackUint2x32(builtin_available_predicate avail)
+{
+   ir_variable *p = in_var(glsl_type::uint64_t_type, "p");
+   MAKE_SIG(glsl_type::uvec2_type, avail, 1, p);
+   body.emit(ret(expr(ir_unop_unpack_uint_2x32, p)));
+   return sig;
+}
+
 ir_function_signature *
 builtin_builder::_length(builtin_available_predicate avail, const glsl_type *type)
 {
@@ -5693,7 +5900,13 @@ builtin_builder::_shader_clock(builtin_available_predicate avail,
 
    body.emit(call(shader->symbols->get_function("__intrinsic_shader_clock"),
                   retval, sig->parameters));
-   body.emit(ret(retval));
+
+   if (type == glsl_type::uint64_t_type) {
+      body.emit(ret(expr(ir_unop_pack_uint_2x32, retval)));
+   } else {
+      body.emit(ret(retval));
+   }
+
    return sig;
 }
 

src/compiler/glsl/builtin_functions.h

@@ -0,0 +1,68 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+#ifndef BULITIN_FUNCTIONS_H
+#define BULITIN_FUNCTIONS_H
+
+extern void
+_mesa_glsl_initialize_builtin_functions();
+
+extern ir_function_signature *
+_mesa_glsl_find_builtin_function(_mesa_glsl_parse_state *state,
+                                 const char *name, exec_list *actual_parameters);
+
+extern ir_function *
+_mesa_glsl_find_builtin_function_by_name(const char *name);
+
+extern gl_shader *
+_mesa_glsl_get_builtin_function_shader(void);
+
+extern ir_function_signature *
+_mesa_get_main_function_signature(glsl_symbol_table *symbols);
+
+extern void
+_mesa_glsl_release_builtin_functions(void);
+
+namespace generate_ir {
+
+ir_function_signature *
+udiv64(void *mem_ctx, builtin_available_predicate avail);
+
+ir_function_signature *
+idiv64(void *mem_ctx, builtin_available_predicate avail);
+
+ir_function_signature *
+umod64(void *mem_ctx, builtin_available_predicate avail);
+
+ir_function_signature *
+imod64(void *mem_ctx, builtin_available_predicate avail);
+
+ir_function_signature *
+umul64(void *mem_ctx, builtin_available_predicate avail);
+
+ir_function_signature *
+sign64(void *mem_ctx, builtin_available_predicate avail);
+
+}
+
+#endif /* BULITIN_FUNCTIONS_H */

src/compiler/glsl/builtin_types.cpp

@@ -409,5 +409,17 @@ _mesa_glsl_initialize_types(struct _mesa_glsl_parse_state *state)
       add_type(symbols, glsl_type::dmat4x2_type);
       add_type(symbols, glsl_type::dmat4x3_type);
    }
+
+   if (state->ARB_gpu_shader_int64_enable) {
+      add_type(symbols, glsl_type::int64_t_type);
+      add_type(symbols, glsl_type::i64vec2_type);
+      add_type(symbols, glsl_type::i64vec3_type);
+      add_type(symbols, glsl_type::i64vec4_type);
+
+      add_type(symbols, glsl_type::uint64_t_type);
+      add_type(symbols, glsl_type::u64vec2_type);
+      add_type(symbols, glsl_type::u64vec3_type);
+      add_type(symbols, glsl_type::u64vec4_type);
+   }
 }
 /** @} */

src/compiler/glsl/builtin_variables.cpp

@@ -30,6 +30,7 @@
 #include "main/uniforms.h"
 #include "program/prog_statevars.h"
 #include "program/prog_instruction.h"
+#include "builtin_functions.h"
 
 using namespace ir_builder;
 

src/compiler/glsl/generate_ir.cpp

@@ -0,0 +1,33 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+#include "ir_builder.h"
+#include "builtin_functions.h"
+#include "program/prog_instruction.h" /* for SWIZZLE_X, &c. */
+
+using namespace ir_builder;
+
+namespace generate_ir {
+
+#include "builtin_int64.h"
+
+}

src/compiler/glsl/glcpp/glcpp-parse.y

@@ -1334,7 +1334,8 @@ add_builtin_define(glcpp_parser_t *parser, const char *name, int value)
 }
 
 glcpp_parser_t *
-glcpp_parser_create(glcpp_extension_iterator extensions, void *state, gl_api api)
+glcpp_parser_create(const struct gl_extensions *extension_list,
+                    glcpp_extension_iterator extensions, void *state, gl_api api)
 {
    glcpp_parser_t *parser;
 
@@ -1369,6 +1370,7 @@ glcpp_parser_create(glcpp_extension_iterator extensions, void *state, gl_api api
    parser->error = 0;
 
    parser->extensions = extensions;
+   parser->extension_list = extension_list;
    parser->state = state;
    parser->api = api;
    parser->version = 0;
@@ -2335,6 +2337,21 @@ _glcpp_parser_handle_version_declaration(glcpp_parser_t *parser, intmax_t versio
       parser->extensions(parser->state, add_builtin_define, parser,
                          version, parser->is_gles);
 
+   if (parser->extension_list) {
+      /* If MESA_shader_integer_functions is supported, then the building
+       * blocks required for the 64x64 => 64 multiply exist.  Add defines for
+       * those functions so that they can be tested.
+       */
+      if (parser->extension_list->MESA_shader_integer_functions) {
+         add_builtin_define(parser, "__have_builtin_builtin_sign64", 1);
+         add_builtin_define(parser, "__have_builtin_builtin_umul64", 1);
+         add_builtin_define(parser, "__have_builtin_builtin_udiv64", 1);
+         add_builtin_define(parser, "__have_builtin_builtin_umod64", 1);
+         add_builtin_define(parser, "__have_builtin_builtin_idiv64", 1);
+         add_builtin_define(parser, "__have_builtin_builtin_imod64", 1);
+      }
+   }
+
    if (explicitly_set) {
       ralloc_asprintf_rewrite_tail(&parser->output, &parser->output_length,
                                    "#version %" PRIiMAX "%s%s", version,

src/compiler/glsl/glcpp/glcpp.h

@@ -205,6 +205,7 @@ struct glcpp_parser {
 	size_t info_log_length;
 	int error;
 	glcpp_extension_iterator extensions;
+	const struct gl_extensions *extension_list;
 	void *state;
 	gl_api api;
 	unsigned version;
@@ -225,7 +226,8 @@ struct glcpp_parser {
 };
 
 glcpp_parser_t *
-glcpp_parser_create (glcpp_extension_iterator extensions, void *state, gl_api api);
+glcpp_parser_create(const struct gl_extensions *extension_list,
+                    glcpp_extension_iterator extensions, void *state, gl_api api);
 
 int
 glcpp_parser_parse (glcpp_parser_t *parser);

src/compiler/glsl/glcpp/pp.c

@@ -218,7 +218,7 @@ glcpp_preprocess(void *ralloc_ctx, const char **shader, char **info_log,
 {
 	int errors;
 	glcpp_parser_t *parser =
-		glcpp_parser_create(extensions, state, gl_ctx->API);
+		glcpp_parser_create(&gl_ctx->Extensions, extensions, state, gl_ctx->API);
 
 	if (! gl_ctx->Const.DisableGLSLLineContinuations)
 		*shader = remove_line_continuations(parser, *shader);

src/compiler/glsl/glsl_lexer.ll

@@ -107,17 +107,29 @@ literal_integer(char *text, int len, struct _mesa_glsl_parse_state *state,
 {
    bool is_uint = (text[len - 1] == 'u' ||
 		   text[len - 1] == 'U');
+   bool is_long = (text[len - 1] == 'l' || text[len - 1] == 'L');
    const char *digits = text;
 
+   if (is_long)
+      is_uint = (text[len - 2] == 'u' && text[len - 1] == 'l') ||
+                (text[len - 2] == 'U' && text[len - 1] == 'L');
    /* Skip "0x" */
    if (base == 16)
       digits += 2;
 
    unsigned long long value = strtoull(digits, NULL, base);
 
-   lval->n = (int)value;
+   if (is_long)
+      lval->n64 = (int64_t)value;
+   else
+      lval->n = (int)value;
 
-   if (value > UINT_MAX) {
+   if (is_long && !is_uint && base == 10 && value > (uint64_t)LLONG_MAX + 1) {
+      /* Tries to catch unintentionally providing a negative value. */
+      _mesa_glsl_warning(lloc, state,
+                         "signed literal value `%s' is interpreted as %lld",
+                         text, lval->n64);
+   } else if (!is_long && value > UINT_MAX) {
       /* Note that signed 0xffffffff is valid, not out of range! */
       if (state->is_version(130, 300)) {
 	 _mesa_glsl_error(lloc, state,
@@ -135,7 +147,10 @@ literal_integer(char *text, int len, struct _mesa_glsl_parse_state *state,
 			 "signed literal value `%s' is interpreted as %d",
 			 text, lval->n);
    }
-   return is_uint ? UINTCONSTANT : INTCONSTANT;
+   if (is_long)
+      return is_uint ? UINT64CONSTANT : INT64CONSTANT;
+   else
+      return is_uint ? UINTCONSTANT : INTCONSTANT;
 }
 
 #define LITERAL_INTEGER(base) \
@@ -462,13 +477,13 @@ layout		{
 \|=		return OR_ASSIGN;
 -=		return SUB_ASSIGN;
 
-[1-9][0-9]*[uU]?	{
+[1-9][0-9]*([uU]|[lL]|ul|UL)?	{
 			    return LITERAL_INTEGER(10);
 			}
-0[xX][0-9a-fA-F]+[uU]?	{
+0[xX][0-9a-fA-F]+([uU]|[lL]|ul|UL)?	{
 			    return LITERAL_INTEGER(16);
 			}
-0[0-7]*[uU]?		{
+0[0-7]*([uU]|[lL]|ul|UL)?		{
 			    return LITERAL_INTEGER(8);
 			}
 
@@ -591,6 +606,16 @@ resource	KEYWORD(420, 300, 0, 0, RESOURCE);
 sample		KEYWORD_WITH_ALT(400, 300, 400, 320, yyextra->ARB_gpu_shader5_enable || yyextra->OES_shader_multisample_interpolation_enable, SAMPLE);
 subroutine	KEYWORD_WITH_ALT(400, 300, 400, 0, yyextra->ARB_shader_subroutine_enable, SUBROUTINE);
 
+    /* Additional words for ARB_gpu_shader_int64 */
+int64_t		KEYWORD_WITH_ALT(0, 0, 0, 0, yyextra->ARB_gpu_shader_int64_enable, INT64_TOK);
+i64vec2		KEYWORD_WITH_ALT(0, 0, 0, 0, yyextra->ARB_gpu_shader_int64_enable, I64VEC2);
+i64vec3		KEYWORD_WITH_ALT(0, 0, 0, 0, yyextra->ARB_gpu_shader_int64_enable, I64VEC3);
+i64vec4		KEYWORD_WITH_ALT(0, 0, 0, 0, yyextra->ARB_gpu_shader_int64_enable, I64VEC4);
+
+uint64_t	KEYWORD_WITH_ALT(0, 0, 0, 0, yyextra->ARB_gpu_shader_int64_enable, UINT64_TOK);
+u64vec2		KEYWORD_WITH_ALT(0, 0, 0, 0, yyextra->ARB_gpu_shader_int64_enable, U64VEC2);
+u64vec3		KEYWORD_WITH_ALT(0, 0, 0, 0, yyextra->ARB_gpu_shader_int64_enable, U64VEC3);
+u64vec4		KEYWORD_WITH_ALT(0, 0, 0, 0, yyextra->ARB_gpu_shader_int64_enable, U64VEC4);
 
 [_a-zA-Z][_a-zA-Z0-9]*	{
 			    struct _mesa_glsl_parse_state *state = yyextra;

src/compiler/glsl/glsl_parser.yy

@@ -97,6 +97,7 @@ static bool match_layout_qualifier(const char *s1, const char *s2,
 
 %union {
    int n;
+   int64_t n64;
    float real;
    double dreal;
    const char *identifier;
@@ -136,6 +137,7 @@ static bool match_layout_qualifier(const char *s1, const char *s2,
 %token ATTRIBUTE CONST_TOK BOOL_TOK FLOAT_TOK INT_TOK UINT_TOK DOUBLE_TOK
 %token BREAK BUFFER CONTINUE DO ELSE FOR IF DISCARD RETURN SWITCH CASE DEFAULT
 %token BVEC2 BVEC3 BVEC4 IVEC2 IVEC3 IVEC4 UVEC2 UVEC3 UVEC4 VEC2 VEC3 VEC4 DVEC2 DVEC3 DVEC4
+%token INT64_TOK UINT64_TOK I64VEC2 I64VEC3 I64VEC4 U64VEC2 U64VEC3 U64VEC4
 %token CENTROID IN_TOK OUT_TOK INOUT_TOK UNIFORM VARYING SAMPLE
 %token NOPERSPECTIVE FLAT SMOOTH
 %token MAT2X2 MAT2X3 MAT2X4
@@ -173,6 +175,7 @@ static bool match_layout_qualifier(const char *s1, const char *s2,
 %token <real> FLOATCONSTANT
 %token <dreal> DOUBLECONSTANT
 %token <n> INTCONSTANT UINTCONSTANT BOOLCONSTANT
+%token <n64> INT64CONSTANT UINT64CONSTANT
 %token <identifier> FIELD_SELECTION
 %token LEFT_OP RIGHT_OP
 %token INC_OP DEC_OP LE_OP GE_OP EQ_OP NE_OP
@@ -451,6 +454,20 @@ primary_expression:
       $$->set_location(@1);
       $$->primary_expression.uint_constant = $1;
    }
+   | INT64CONSTANT
+   {
+      void *ctx = state->linalloc;
+      $$ = new(ctx) ast_expression(ast_int64_constant, NULL, NULL, NULL);
+      $$->set_location(@1);
+      $$->primary_expression.int64_constant = $1;
+   }
+   | UINT64CONSTANT
+   {
+      void *ctx = state->linalloc;
+      $$ = new(ctx) ast_expression(ast_uint64_constant, NULL, NULL, NULL);
+      $$->set_location(@1);
+      $$->primary_expression.uint64_constant = $1;
+   }
    | FLOATCONSTANT
    {
       void *ctx = state->linalloc;
@@ -2304,6 +2321,14 @@ basic_type_specifier_nonarray:
    | UIMAGE2DMS             { $$ = "uimage2DMS"; }
    | UIMAGE2DMSARRAY        { $$ = "uimage2DMSArray"; }
    | ATOMIC_UINT            { $$ = "atomic_uint"; }
+   | INT64_TOK              { $$ = "int64_t";	   }
+   | I64VEC2                { $$ = "i64vec2";	   }
+   | I64VEC3                { $$ = "i64vec3";	   }
+   | I64VEC4                { $$ = "i64vec4";	   }
+   | UINT64_TOK             { $$ = "uint64_t";	   }
+   | U64VEC2                { $$ = "u64vec2";	   }
+   | U64VEC3                { $$ = "u64vec3";	   }
+   | U64VEC4                { $$ = "u64vec4";	   }
    ;
 
 precision_qualifier:

src/compiler/glsl/glsl_parser_extras.cpp

@@ -20,6 +20,8 @@
  * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
  * DEALINGS IN THE SOFTWARE.
  */
+#define __STDC_FORMAT_MACROS 1
+#include <inttypes.h> /* for PRIx64 macro */
 #include <stdio.h>
 #include <stdarg.h>
 #include <string.h>
@@ -37,6 +39,7 @@
 #include "glsl_parser.h"
 #include "ir_optimization.h"
 #include "loop_analysis.h"
+#include "builtin_functions.h"
 
 /**
  * Format a short human-readable description of the given GLSL version.
@@ -608,6 +611,7 @@ static const _mesa_glsl_extension _mesa_glsl_supported_extensions[] = {
    EXT(ARB_fragment_layer_viewport),
    EXT(ARB_gpu_shader5),
    EXT(ARB_gpu_shader_fp64),
+   EXT(ARB_gpu_shader_int64),
    EXT(ARB_post_depth_coverage),
    EXT(ARB_sample_shading),
    EXT(ARB_separate_shader_objects),
@@ -1247,6 +1251,14 @@ ast_expression::print(void) const
       printf("%f ", primary_expression.double_constant);
       break;
 
+   case ast_int64_constant:
+      printf("%" PRId64 " ", primary_expression.int64_constant);
+      break;
+
+   case ast_uint64_constant:
+      printf("%" PRIu64 " ", primary_expression.uint64_constant);
+      break;
+
    case ast_bool_constant:
       printf("%s ",
 	     primary_expression.bool_constant
@@ -1704,7 +1716,7 @@ set_shader_inout_layout(struct gl_shader *shader,
          if (state->out_qualifier->out_xfb_stride[i]->
                 process_qualifier_constant(state, "xfb_stride", &xfb_stride,
                 true)) {
-            shader->info.TransformFeedback.BufferStride[i] = xfb_stride;
+            shader->TransformFeedbackBufferStride[i] = xfb_stride;
          }
       }
    }
@@ -1808,16 +1820,15 @@ set_shader_inout_layout(struct gl_shader *shader,
       break;
 
    case MESA_SHADER_FRAGMENT:
-      shader->info.redeclares_gl_fragcoord =
-         state->fs_redeclares_gl_fragcoord;
-      shader->info.uses_gl_fragcoord = state->fs_uses_gl_fragcoord;
-      shader->info.pixel_center_integer = state->fs_pixel_center_integer;
-      shader->info.origin_upper_left = state->fs_origin_upper_left;
-      shader->info.ARB_fragment_coord_conventions_enable =
+      shader->redeclares_gl_fragcoord = state->fs_redeclares_gl_fragcoord;
+      shader->uses_gl_fragcoord = state->fs_uses_gl_fragcoord;
+      shader->pixel_center_integer = state->fs_pixel_center_integer;
+      shader->origin_upper_left = state->fs_origin_upper_left;
+      shader->ARB_fragment_coord_conventions_enable =
          state->ARB_fragment_coord_conventions_enable;
-      shader->info.EarlyFragmentTests = state->fs_early_fragment_tests;
-      shader->info.InnerCoverage = state->fs_inner_coverage;
-      shader->info.PostDepthCoverage = state->fs_post_depth_coverage;
+      shader->EarlyFragmentTests = state->fs_early_fragment_tests;
+      shader->InnerCoverage = state->fs_inner_coverage;
+      shader->PostDepthCoverage = state->fs_post_depth_coverage;
       shader->BlendSupport = state->fs_blend_support;
       break;
 

src/compiler/glsl/glsl_parser_extras.h

@@ -250,6 +250,11 @@ struct _mesa_glsl_parse_state {
       return ARB_gpu_shader_fp64_enable || is_version(400, 0);
    }
 
+   bool has_int64() const
+   {
+      return ARB_gpu_shader_int64_enable;
+   }
+
    bool has_420pack() const
    {
       return ARB_shading_language_420pack_enable || is_version(420, 0);
@@ -610,6 +615,8 @@ struct _mesa_glsl_parse_state {
    bool ARB_gpu_shader5_warn;
    bool ARB_gpu_shader_fp64_enable;
    bool ARB_gpu_shader_fp64_warn;
+   bool ARB_gpu_shader_int64_enable;
+   bool ARB_gpu_shader_int64_warn;
    bool ARB_post_depth_coverage_enable;
    bool ARB_post_depth_coverage_warn;
    bool ARB_sample_shading_enable;

src/compiler/glsl/glsl_to_nir.cpp

@@ -169,8 +169,6 @@ glsl_to_nir(const struct gl_shader_program *shader_prog,
    shader->info->name = ralloc_asprintf(shader, "GLSL%d", shader_prog->Name);
    if (shader_prog->Label)
       shader->info->label = ralloc_strdup(shader, shader_prog->Label);
-   shader->info->clip_distance_array_size = sh->Program->ClipDistanceArraySize;
-   shader->info->cull_distance_array_size = sh->Program->CullDistanceArraySize;
    shader->info->has_transform_feedback_varyings =
       shader_prog->TransformFeedback.NumVarying > 0;
 
@@ -254,6 +252,22 @@ constant_copy(ir_constant *ir, void *mem_ctx)
       }
       break;
 
+   case GLSL_TYPE_UINT64:
+      /* Only float base types can be matrices. */
+      assert(cols == 1);
+
+      for (unsigned r = 0; r < rows; r++)
+         ret->values[0].u64[r] = ir->value.u64[r];
+      break;
+
+   case GLSL_TYPE_INT64:
+      /* Only float base types can be matrices. */
+      assert(cols == 1);
+
+      for (unsigned r = 0; r < rows; r++)
+         ret->values[0].i64[r] = ir->value.i64[r];
+      break;
+
    case GLSL_TYPE_BOOL:
       /* Only float base types can be matrices. */
       assert(cols == 1);
@@ -927,8 +941,7 @@ nir_visitor::visit(ir_call *ir)
          nir_builder_instr_insert(&b, &instr->instr);
          break;
       case nir_intrinsic_shader_clock:
-         nir_ssa_dest_init(&instr->instr, &instr->dest, 2, 32, NULL);
-         instr->num_components = 2;
+         nir_ssa_dest_init(&instr->instr, &instr->dest, 1, 32, NULL);
          nir_builder_instr_insert(&b, &instr->instr);
          break;
       case nir_intrinsic_store_ssbo: {
@@ -1310,6 +1323,12 @@ type_is_float(glsl_base_type type)
    return type == GLSL_TYPE_FLOAT || type == GLSL_TYPE_DOUBLE;
 }
 
+static bool
+type_is_signed(glsl_base_type type)
+{
+   return type == GLSL_TYPE_INT || type == GLSL_TYPE_INT64;
+}
+
 void
 nir_visitor::visit(ir_expression *ir)
 {
@@ -1465,6 +1484,7 @@ nir_visitor::visit(ir_expression *ir)
    case ir_unop_f2b:  result = nir_f2b(&b, srcs[0]);   break;
    case ir_unop_i2b:  result = nir_i2b(&b, srcs[0]);   break;
    case ir_unop_b2i:  result = nir_b2i(&b, srcs[0]);   break;
+   case ir_unop_b2i64:result = nir_b2i64(&b, srcs[0]); break;
    case ir_unop_d2f:  result = nir_d2f(&b, srcs[0]);   break;
    case ir_unop_f2d:  result = nir_f2d(&b, srcs[0]);   break;
    case ir_unop_d2i:  result = nir_d2i(&b, srcs[0]);   break;
@@ -1478,12 +1498,40 @@ nir_visitor::visit(ir_expression *ir)
       assert(supports_ints);
       result = nir_u2d(&b, srcs[0]);
       break;
+   case ir_unop_i642i: result = nir_i2i32(&b, srcs[0]);   break;
+   case ir_unop_i642u: result = nir_i2u32(&b, srcs[0]);   break;
+   case ir_unop_i642f: result = nir_i642f(&b, srcs[0]);   break;
+   case ir_unop_i642d: result = nir_i642d(&b, srcs[0]);   break;
+
+   case ir_unop_u642i: result = nir_u2i32(&b, srcs[0]);   break;
+   case ir_unop_u642u: result = nir_u2u32(&b, srcs[0]);   break;
+   case ir_unop_u642f: result = nir_u642f(&b, srcs[0]);   break;
+   case ir_unop_u642d: result = nir_u642d(&b, srcs[0]);   break;
+
+   case ir_unop_i2i64: result = nir_i2i64(&b, srcs[0]);   break;
+   case ir_unop_u2i64: result = nir_u2i64(&b, srcs[0]);   break;
+   case ir_unop_f2i64:
+   case ir_unop_d2i64:
+      result = nir_f2i64(&b, srcs[0]);
+      break;
+   case ir_unop_i2u64: result = nir_i2u64(&b, srcs[0]);   break;
+   case ir_unop_u2u64: result = nir_u2u64(&b, srcs[0]);   break;
+   case ir_unop_f2u64:
+   case ir_unop_d2u64:
+      result = nir_f2u64(&b, srcs[0]);
+      break;
    case ir_unop_i2u:
    case ir_unop_u2i:
+   case ir_unop_i642u64:
+   case ir_unop_u642i64:
    case ir_unop_bitcast_i2f:
    case ir_unop_bitcast_f2i:
    case ir_unop_bitcast_u2f:
    case ir_unop_bitcast_f2u:
+   case ir_unop_bitcast_i642d:
+   case ir_unop_bitcast_d2i64:
+   case ir_unop_bitcast_u642d:
+   case ir_unop_bitcast_d2u64:
    case ir_unop_subroutine_to_int:
       /* no-op */
       result = nir_imov(&b, srcs[0]);
@@ -1537,6 +1585,14 @@ nir_visitor::visit(ir_expression *ir)
    case ir_unop_unpack_double_2x32:
       result = nir_unpack_double_2x32(&b, srcs[0]);
       break;
+   case ir_unop_pack_int_2x32:
+   case ir_unop_pack_uint_2x32:
+      result = nir_pack_int_2x32(&b, srcs[0]);
+      break;
+   case ir_unop_unpack_int_2x32:
+   case ir_unop_unpack_uint_2x32:
+      result = nir_unpack_int_2x32(&b, srcs[0]);
+      break;
    case ir_unop_bitfield_reverse:
       result = nir_bitfield_reverse(&b, srcs[0]);
       break;
@@ -1627,7 +1683,7 @@ nir_visitor::visit(ir_expression *ir)
    case ir_binop_div:
       if (type_is_float(out_type))
          result = nir_fdiv(&b, srcs[0], srcs[1]);
-      else if (out_type == GLSL_TYPE_INT)
+      else if (type_is_signed(out_type))
          result = nir_idiv(&b, srcs[0], srcs[1]);
       else
          result = nir_udiv(&b, srcs[0], srcs[1]);
@@ -1639,7 +1695,7 @@ nir_visitor::visit(ir_expression *ir)
    case ir_binop_min:
       if (type_is_float(out_type))
          result = nir_fmin(&b, srcs[0], srcs[1]);
-      else if (out_type == GLSL_TYPE_INT)
+      else if (type_is_signed(out_type))
          result = nir_imin(&b, srcs[0], srcs[1]);
       else
          result = nir_umin(&b, srcs[0], srcs[1]);
@@ -1647,7 +1703,7 @@ nir_visitor::visit(ir_expression *ir)
    case ir_binop_max:
       if (type_is_float(out_type))
          result = nir_fmax(&b, srcs[0], srcs[1]);
-      else if (out_type == GLSL_TYPE_INT)
+      else if (type_is_signed(out_type))
          result = nir_imax(&b, srcs[0], srcs[1]);
       else
          result = nir_umax(&b, srcs[0], srcs[1]);
@@ -1670,8 +1726,8 @@ nir_visitor::visit(ir_expression *ir)
       break;
    case ir_binop_lshift: result = nir_ishl(&b, srcs[0], srcs[1]); break;
    case ir_binop_rshift:
-      result = (out_type == GLSL_TYPE_INT) ? nir_ishr(&b, srcs[0], srcs[1])
-                                           : nir_ushr(&b, srcs[0], srcs[1]);
+      result = (type_is_signed(out_type)) ? nir_ishr(&b, srcs[0], srcs[1])
+                                          : nir_ushr(&b, srcs[0], srcs[1]);
       break;
    case ir_binop_imul_high:
       result = (out_type == GLSL_TYPE_INT) ? nir_imul_high(&b, srcs[0], srcs[1])
@@ -1683,7 +1739,7 @@ nir_visitor::visit(ir_expression *ir)
       if (supports_ints) {
          if (type_is_float(types[0]))
             result = nir_flt(&b, srcs[0], srcs[1]);
-         else if (types[0] == GLSL_TYPE_INT)
+         else if (type_is_signed(types[0]))
             result = nir_ilt(&b, srcs[0], srcs[1]);
          else
             result = nir_ult(&b, srcs[0], srcs[1]);
@@ -1695,7 +1751,7 @@ nir_visitor::visit(ir_expression *ir)
       if (supports_ints) {
          if (type_is_float(types[0]))
             result = nir_flt(&b, srcs[1], srcs[0]);
-         else if (types[0] == GLSL_TYPE_INT)
+         else if (type_is_signed(types[0]))
             result = nir_ilt(&b, srcs[1], srcs[0]);
          else
             result = nir_ult(&b, srcs[1], srcs[0]);
@@ -1707,7 +1763,7 @@ nir_visitor::visit(ir_expression *ir)
       if (supports_ints) {
          if (type_is_float(types[0]))
             result = nir_fge(&b, srcs[1], srcs[0]);
-         else if (types[0] == GLSL_TYPE_INT)
+         else if (type_is_signed(types[0]))
             result = nir_ige(&b, srcs[1], srcs[0]);
          else
             result = nir_uge(&b, srcs[1], srcs[0]);
@@ -1719,7 +1775,7 @@ nir_visitor::visit(ir_expression *ir)
       if (supports_ints) {
          if (type_is_float(types[0]))
             result = nir_fge(&b, srcs[0], srcs[1]);
-         else if (types[0] == GLSL_TYPE_INT)
+         else if (type_is_signed(types[0]))
             result = nir_ige(&b, srcs[0], srcs[1]);
          else
             result = nir_uge(&b, srcs[0], srcs[1]);

src/compiler/glsl/int64.glsl

@@ -0,0 +1,121 @@
+/* Compile with:
+ *
+ * glsl_compiler --version 140 --dump-builder int64.glsl > builtin_int64.h
+ *
+ * Using version 1.40+ prevents built-in variables from being included.
+ */
+#version 400
+#extension GL_ARB_gpu_shader_int64: require
+#extension GL_ARB_shading_language_420pack: require
+
+uvec2
+umul64(uvec2 a, uvec2 b)
+{
+   uvec2 result;
+
+   umulExtended(a.x, b.x, result.y, result.x);
+   result.y += a.x * b.y + a.y * b.x;
+
+   return result;
+}
+
+ivec2
+sign64(ivec2 a)
+{
+   ivec2 result;
+
+   result.y = a.y >> 31;
+   result.x = result.y | int((a.x | a.y) != 0);
+
+   return result;
+}
+
+uvec4
+udivmod64(uvec2 n, uvec2 d)
+{
+   uvec2 quot = uvec2(0U, 0U);
+   int log2_denom = findMSB(d.y) + 32;
+
+   /* If the upper 32 bits of denom are non-zero, it is impossible for shifts
+    * greater than 32 bits to occur.  If the upper 32 bits of the numerator
+    * are zero, it is impossible for (denom << [63, 32]) <= numer unless
+    * denom == 0.
+    */
+   if (d.y == 0 && n.y >= d.x) {
+      log2_denom = findMSB(d.x);
+
+      /* Since the upper 32 bits of denom are zero, log2_denom <= 31 and we
+       * don't have to compare log2_denom inside the loop as is done in the
+       * general case (below).
+       */
+      for (int i = 31; i >= 1; i--) {
+	 if (log2_denom <= 31 - i && (d.x << i) <= n.y) {
+	    n.y -= d.x << i;
+	    quot.y |= 1U << i;
+	 }
+      }
+
+      /* log2_denom is always <= 31, so manually peel the last loop
+       * iteration.
+       */
+      if (d.x <= n.y) {
+	 n.y -= d.x;
+	 quot.y |= 1U;
+      }
+   }
+
+   uint64_t d64 = packUint2x32(d);
+   uint64_t n64 = packUint2x32(n);
+   for (int i = 31; i >= 1; i--) {
+      if (log2_denom <= 63 - i && (d64 << i) <= n64) {
+	 n64 -= d64 << i;
+	 quot.x |= 1U << i;
+      }
+   }
+
+   /* log2_denom is always <= 63, so manually peel the last loop
+    * iteration.
+    */
+   if (d64 <= n64) {
+      n64 -= d64;
+      quot.x |= 1U;
+   }
+
+   return uvec4(quot, unpackUint2x32(n64));
+}
+
+uvec2
+udiv64(uvec2 n, uvec2 d)
+{
+   return udivmod64(n, d).xy;
+}
+
+ivec2
+idiv64(ivec2 _n, ivec2 _d)
+{
+   const bool negate = (_n.y < 0) != (_d.y < 0);
+   uvec2 n = unpackUint2x32(uint64_t(abs(packInt2x32(_n))));
+   uvec2 d = unpackUint2x32(uint64_t(abs(packInt2x32(_d))));
+
+   uvec2 quot = udivmod64(n, d).xy;
+
+   return negate ? unpackInt2x32(-int64_t(packUint2x32(quot))) : ivec2(quot);
+}
+
+uvec2
+umod64(uvec2 n, uvec2 d)
+{
+   return udivmod64(n, d).zw;
+}
+
+ivec2
+imod64(ivec2 _n, ivec2 _d)
+{
+   const bool negate = (_n.y < 0) != (_d.y < 0);
+   uvec2 n = unpackUint2x32(uint64_t(abs(packInt2x32(_n))));
+   uvec2 d = unpackUint2x32(uint64_t(abs(packInt2x32(_d))));
+
+   uvec2 rem = udivmod64(n, d).zy;
+
+   return negate ? unpackInt2x32(-int64_t(packUint2x32(rem))) : ivec2(rem);
+}

src/compiler/glsl/ir.cpp

@@ -261,6 +261,8 @@ ir_expression::ir_expression(int op, ir_rvalue *op0)
    case ir_unop_find_msb:
    case ir_unop_find_lsb:
    case ir_unop_subroutine_to_int:
+   case ir_unop_i642i:
+   case ir_unop_u642i:
       this->type = glsl_type::get_instance(GLSL_TYPE_INT,
 					   op0->type->vector_elements, 1);
       break;
@@ -271,6 +273,8 @@ ir_expression::ir_expression(int op, ir_rvalue *op0)
    case ir_unop_d2f:
    case ir_unop_bitcast_i2f:
    case ir_unop_bitcast_u2f:
+   case ir_unop_i642f:
+   case ir_unop_u642f:
       this->type = glsl_type::get_instance(GLSL_TYPE_FLOAT,
 					   op0->type->vector_elements, 1);
       break;
@@ -278,6 +282,7 @@ ir_expression::ir_expression(int op, ir_rvalue *op0)
    case ir_unop_f2b:
    case ir_unop_i2b:
    case ir_unop_d2b:
+   case ir_unop_i642b:
       this->type = glsl_type::get_instance(GLSL_TYPE_BOOL,
 					   op0->type->vector_elements, 1);
       break;
@@ -285,6 +290,8 @@ ir_expression::ir_expression(int op, ir_rvalue *op0)
    case ir_unop_f2d:
    case ir_unop_i2d:
    case ir_unop_u2d:
+   case ir_unop_i642d:
+   case ir_unop_u642d:
       this->type = glsl_type::get_instance(GLSL_TYPE_DOUBLE,
 					   op0->type->vector_elements, 1);
       break;
@@ -293,18 +300,43 @@ ir_expression::ir_expression(int op, ir_rvalue *op0)
    case ir_unop_f2u:
    case ir_unop_d2u:
    case ir_unop_bitcast_f2u:
+   case ir_unop_i642u:
+   case ir_unop_u642u:
       this->type = glsl_type::get_instance(GLSL_TYPE_UINT,
 					   op0->type->vector_elements, 1);
       break;
 
+   case ir_unop_i2i64:
+   case ir_unop_u2i64:
+   case ir_unop_b2i64:
+   case ir_unop_f2i64:
+   case ir_unop_d2i64:
+   case ir_unop_u642i64:
+      this->type = glsl_type::get_instance(GLSL_TYPE_INT64,
+					   op0->type->vector_elements, 1);
+      break;
+
+   case ir_unop_i2u64:
+   case ir_unop_u2u64:
+   case ir_unop_f2u64:
+   case ir_unop_d2u64:
+   case ir_unop_i642u64:
+      this->type = glsl_type::get_instance(GLSL_TYPE_UINT64,
+					   op0->type->vector_elements, 1);
+      break;
    case ir_unop_noise:
       this->type = glsl_type::float_type;
       break;
 
    case ir_unop_unpack_double_2x32:
+   case ir_unop_unpack_uint_2x32:
       this->type = glsl_type::uvec2_type;
       break;
 
+   case ir_unop_unpack_int_2x32:
+      this->type = glsl_type::ivec2_type;
+      break;
+
    case ir_unop_pack_snorm_2x16:
    case ir_unop_pack_snorm_4x8:
    case ir_unop_pack_unorm_2x16:
@@ -317,6 +349,14 @@ ir_expression::ir_expression(int op, ir_rvalue *op0)
       this->type = glsl_type::double_type;
       break;
 
+   case ir_unop_pack_int_2x32:
+      this->type = glsl_type::int64_t_type;
+      break;
+
+   case ir_unop_pack_uint_2x32:
+      this->type = glsl_type::uint64_t_type;
+      break;
+
    case ir_unop_unpack_snorm_2x16:
    case ir_unop_unpack_unorm_2x16:
    case ir_unop_unpack_half_2x16:
@@ -347,6 +387,21 @@ ir_expression::ir_expression(int op, ir_rvalue *op0)
       this->type = glsl_type::bool_type;
       break;
 
+   case ir_unop_bitcast_i642d:
+   case ir_unop_bitcast_u642d:
+      this->type = glsl_type::get_instance(GLSL_TYPE_DOUBLE,
+                                           op0->type->vector_elements, 1);
+      break;
+
+   case ir_unop_bitcast_d2i64:
+      this->type = glsl_type::get_instance(GLSL_TYPE_INT64,
+                                           op0->type->vector_elements, 1);
+      break;
+   case ir_unop_bitcast_d2u64:
+      this->type = glsl_type::get_instance(GLSL_TYPE_UINT64,
+                                           op0->type->vector_elements, 1);
+      break;
+
    default:
       assert(!"not reached: missing automatic type setup for ir_expression");
       this->type = op0->type;
@@ -613,6 +668,32 @@ ir_constant::ir_constant(int integer, unsigned vector_elements)
    }
 }
 
+ir_constant::ir_constant(uint64_t u64, unsigned vector_elements)
+   : ir_rvalue(ir_type_constant)
+{
+   assert(vector_elements <= 4);
+   this->type = glsl_type::get_instance(GLSL_TYPE_UINT64, vector_elements, 1);
+   for (unsigned i = 0; i < vector_elements; i++) {
+      this->value.u64[i] = u64;
+   }
+   for (unsigned i = vector_elements; i < 16; i++) {
+      this->value.u64[i] = 0;
+   }
+}
+
+ir_constant::ir_constant(int64_t int64, unsigned vector_elements)
+   : ir_rvalue(ir_type_constant)
+{
+   assert(vector_elements <= 4);
+   this->type = glsl_type::get_instance(GLSL_TYPE_INT64, vector_elements, 1);
+   for (unsigned i = 0; i < vector_elements; i++) {
+      this->value.i64[i] = int64;
+   }
+   for (unsigned i = vector_elements; i < 16; i++) {
+      this->value.i64[i] = 0;
+   }
+}
+
 ir_constant::ir_constant(bool b, unsigned vector_elements)
    : ir_rvalue(ir_type_constant)
 {
@@ -716,6 +797,11 @@ ir_constant::ir_constant(const struct glsl_type *type, exec_list *value_list)
 	    for (unsigned i = 0; i < type->components(); i++)
 	       this->value.d[i] = value->value.d[0];
 	    break;
+	 case GLSL_TYPE_UINT64:
+	 case GLSL_TYPE_INT64:
+	    for (unsigned i = 0; i < type->components(); i++)
+	       this->value.u64[i] = value->value.u64[0];
+	    break;
 	 case GLSL_TYPE_BOOL:
 	    for (unsigned i = 0; i < type->components(); i++)
 	       this->value.b[i] = value->value.b[0];
@@ -778,6 +864,12 @@ ir_constant::ir_constant(const struct glsl_type *type, exec_list *value_list)
 	 case GLSL_TYPE_DOUBLE:
 	    this->value.d[i] = value->get_double_component(j);
 	    break;
+         case GLSL_TYPE_UINT64:
+	    this->value.u64[i] = value->get_uint64_component(j);
+	    break;
+	 case GLSL_TYPE_INT64:
+	    this->value.i64[i] = value->get_int64_component(j);
+	    break;
 	 default:
 	    /* FINISHME: What to do?  Exceptions are not the answer.
 	     */
@@ -831,6 +923,8 @@ ir_constant::get_bool_component(unsigned i) const
    case GLSL_TYPE_FLOAT: return ((int)this->value.f[i]) != 0;
    case GLSL_TYPE_BOOL:  return this->value.b[i];
    case GLSL_TYPE_DOUBLE: return this->value.d[i] != 0.0;
+   case GLSL_TYPE_UINT64: return this->value.u64[i] != 0;
+   case GLSL_TYPE_INT64:  return this->value.i64[i] != 0;
    default:              assert(!"Should not get here."); break;
    }
 
@@ -849,6 +943,8 @@ ir_constant::get_float_component(unsigned i) const
    case GLSL_TYPE_FLOAT: return this->value.f[i];
    case GLSL_TYPE_BOOL:  return this->value.b[i] ? 1.0f : 0.0f;
    case GLSL_TYPE_DOUBLE: return (float) this->value.d[i];
+   case GLSL_TYPE_UINT64: return (float) this->value.u64[i];
+   case GLSL_TYPE_INT64:  return (float) this->value.i64[i];
    default:              assert(!"Should not get here."); break;
    }
 
@@ -867,6 +963,8 @@ ir_constant::get_double_component(unsigned i) const
    case GLSL_TYPE_FLOAT: return (double) this->value.f[i];
    case GLSL_TYPE_BOOL:  return this->value.b[i] ? 1.0 : 0.0;
    case GLSL_TYPE_DOUBLE: return this->value.d[i];
+   case GLSL_TYPE_UINT64: return (double) this->value.u64[i];
+   case GLSL_TYPE_INT64:  return (double) this->value.i64[i];
    default:              assert(!"Should not get here."); break;
    }
 
@@ -885,6 +983,8 @@ ir_constant::get_int_component(unsigned i) const
    case GLSL_TYPE_FLOAT: return (int) this->value.f[i];
    case GLSL_TYPE_BOOL:  return this->value.b[i] ? 1 : 0;
    case GLSL_TYPE_DOUBLE: return (int) this->value.d[i];
+   case GLSL_TYPE_UINT64: return (int) this->value.u64[i];
+   case GLSL_TYPE_INT64:  return (int) this->value.i64[i];
    default:              assert(!"Should not get here."); break;
    }
 
@@ -903,6 +1003,48 @@ ir_constant::get_uint_component(unsigned i) const
    case GLSL_TYPE_FLOAT: return (unsigned) this->value.f[i];
    case GLSL_TYPE_BOOL:  return this->value.b[i] ? 1 : 0;
    case GLSL_TYPE_DOUBLE: return (unsigned) this->value.d[i];
+   case GLSL_TYPE_UINT64: return (unsigned) this->value.u64[i];
+   case GLSL_TYPE_INT64:  return (unsigned) this->value.i64[i];
+   default:              assert(!"Should not get here."); break;
+   }
+
+   /* Must return something to make the compiler happy.  This is clearly an
+    * error case.
+    */
+   return 0;
+}
+
+int64_t
+ir_constant::get_int64_component(unsigned i) const
+{
+   switch (this->type->base_type) {
+   case GLSL_TYPE_UINT:  return this->value.u[i];
+   case GLSL_TYPE_INT:   return this->value.i[i];
+   case GLSL_TYPE_FLOAT: return (int64_t) this->value.f[i];
+   case GLSL_TYPE_BOOL:  return this->value.b[i] ? 1 : 0;
+   case GLSL_TYPE_DOUBLE: return (int64_t) this->value.d[i];
+   case GLSL_TYPE_UINT64: return (int64_t) this->value.u64[i];
+   case GLSL_TYPE_INT64:  return this->value.i64[i];
+   default:              assert(!"Should not get here."); break;
+   }
+
+   /* Must return something to make the compiler happy.  This is clearly an
+    * error case.
+    */
+   return 0;
+}
+
+uint64_t
+ir_constant::get_uint64_component(unsigned i) const
+{
+   switch (this->type->base_type) {
+   case GLSL_TYPE_UINT:  return this->value.u[i];
+   case GLSL_TYPE_INT:   return this->value.i[i];
+   case GLSL_TYPE_FLOAT: return (uint64_t) this->value.f[i];
+   case GLSL_TYPE_BOOL:  return this->value.b[i] ? 1 : 0;
+   case GLSL_TYPE_DOUBLE: return (uint64_t) this->value.d[i];
+   case GLSL_TYPE_UINT64: return this->value.u64[i];
+   case GLSL_TYPE_INT64:  return (uint64_t) this->value.i64[i];
    default:              assert(!"Should not get here."); break;
    }
 
@@ -968,6 +1110,8 @@ ir_constant::copy_offset(ir_constant *src, int offset)
    case GLSL_TYPE_INT:
    case GLSL_TYPE_FLOAT:
    case GLSL_TYPE_DOUBLE:
+   case GLSL_TYPE_UINT64:
+   case GLSL_TYPE_INT64:
    case GLSL_TYPE_BOOL: {
       unsigned int size = src->type->components();
       assert (size <= this->type->components() - offset);
@@ -988,6 +1132,12 @@ ir_constant::copy_offset(ir_constant *src, int offset)
 	 case GLSL_TYPE_DOUBLE:
 	    value.d[i+offset] = src->get_double_component(i);
 	    break;
+         case GLSL_TYPE_UINT64:
+	    value.u64[i+offset] = src->get_uint64_component(i);
+	    break;
+	 case GLSL_TYPE_INT64:
+	    value.i64[i+offset] = src->get_int64_component(i);
+	    break;
 	 default: // Shut up the compiler
 	    break;
 	 }
@@ -1047,6 +1197,12 @@ ir_constant::copy_masked_offset(ir_constant *src, int offset, unsigned int mask)
 	 case GLSL_TYPE_DOUBLE:
 	    value.d[i+offset] = src->get_double_component(id++);
 	    break;
+         case GLSL_TYPE_UINT64:
+	    value.u64[i+offset] = src->get_uint64_component(id++);
+	    break;
+	 case GLSL_TYPE_INT64:
+	    value.i64[i+offset] = src->get_int64_component(id++);
+	    break;
 	 default:
 	    assert(!"Should not get here.");
 	    return;
@@ -1111,6 +1267,14 @@ ir_constant::has_value(const ir_constant *c) const
 	 if (this->value.d[i] != c->value.d[i])
 	    return false;
 	 break;
+      case GLSL_TYPE_UINT64:
+	 if (this->value.u64[i] != c->value.u64[i])
+	    return false;
+	 break;
+      case GLSL_TYPE_INT64:
+	 if (this->value.i64[i] != c->value.i64[i])
+	    return false;
+	 break;
       default:
 	 assert(!"Should not get here.");
 	 return false;
@@ -1152,6 +1316,14 @@ ir_constant::is_value(float f, int i) const
 	 if (this->value.d[c] != double(f))
 	    return false;
 	 break;
+      case GLSL_TYPE_UINT64:
+	 if (this->value.u64[c] != uint64_t(i))
+	    return false;
+	 break;
+      case GLSL_TYPE_INT64:
+	 if (this->value.i64[c] != i)
+	    return false;
+	 break;
       default:
 	 /* The only other base types are structures, arrays, and samplers.
 	  * Samplers cannot be constants, and the others should have been

src/compiler/glsl/ir.h

@@ -2093,6 +2093,8 @@ union ir_constant_data {
       float f[16];
       bool b[16];
       double d[16];
+      uint64_t u64[16];
+      int64_t i64[16];
 };
 
 
@@ -2104,6 +2106,8 @@ public:
    ir_constant(int i, unsigned vector_elements=1);
    ir_constant(float f, unsigned vector_elements=1);
    ir_constant(double d, unsigned vector_elements=1);
+   ir_constant(uint64_t u64, unsigned vector_elements=1);
+   ir_constant(int64_t i64, unsigned vector_elements=1);
 
    /**
     * Construct an ir_constant from a list of ir_constant values
@@ -2154,6 +2158,8 @@ public:
    double get_double_component(unsigned i) const;
    int get_int_component(unsigned i) const;
    unsigned get_uint_component(unsigned i) const;
+   int64_t get_int64_component(unsigned i) const;
+   uint64_t get_uint64_component(unsigned i) const;
    /*@}*/
 
    ir_constant *get_array_element(unsigned i) const;
@@ -2377,25 +2383,6 @@ extern void
 _mesa_glsl_initialize_derived_variables(struct gl_context *ctx,
                                         gl_shader *shader);
 
-extern void
-_mesa_glsl_initialize_builtin_functions();
-
-extern ir_function_signature *
-_mesa_glsl_find_builtin_function(_mesa_glsl_parse_state *state,
-                                 const char *name, exec_list *actual_parameters);
-
-extern ir_function *
-_mesa_glsl_find_builtin_function_by_name(const char *name);
-
-extern gl_shader *
-_mesa_glsl_get_builtin_function_shader(void);
-
-extern ir_function_signature *
-_mesa_get_main_function_signature(glsl_symbol_table *symbols);
-
-extern void
-_mesa_glsl_release_builtin_functions(void);
-
 extern void
 reparent_ir(exec_list *list, void *mem_ctx);
 

src/compiler/glsl/ir_builder.cpp

@@ -518,6 +518,30 @@ b2f(operand a)
    return expr(ir_unop_b2f, a);
 }
 
+ir_expression*
+bitcast_d2i64(operand a)
+{
+   return expr(ir_unop_bitcast_d2i64, a);
+}
+
+ir_expression*
+bitcast_d2u64(operand a)
+{
+   return expr(ir_unop_bitcast_d2u64, a);
+}
+
+ir_expression*
+bitcast_i642d(operand a)
+{
+   return expr(ir_unop_bitcast_i642d, a);
+}
+
+ir_expression*
+bitcast_u642d(operand a)
+{
+   return expr(ir_unop_bitcast_u642d, a);
+}
+
 ir_expression *
 interpolate_at_centroid(operand a)
 {

src/compiler/glsl/ir_builder.h

@@ -191,6 +191,12 @@ ir_expression *f2d(operand a);
 ir_expression *i2d(operand a);
 ir_expression *u2d(operand a);
 
+ir_expression *bitcast_d2i64(operand a);
+ir_expression *bitcast_d2u64(operand a);
+
+ir_expression *bitcast_i642d(operand a);
+ir_expression *bitcast_u642d(operand a);
+
 ir_expression *min2(operand a, operand b);
 ir_expression *max2(operand a, operand b);
 

src/compiler/glsl/ir_builder_print_visitor.cpp

@@ -396,13 +396,24 @@ ir_builder_print_visitor::visit(ir_constant *ir)
 
             memcpy(&v, &ir->value.d[i], sizeof(v));
             if (v != 0)
-               /* FIXME: This won't actually work until ARB_gpu_shader_int64
-                * support lands.
-                */
                print_without_indent("r%04X_data.u64[%u] = 0x%016" PRIx64 "; /* %g */\n",
                                     my_index, i, v, ir->value.d[i]);
             break;
          }
+         case GLSL_TYPE_UINT64:
+            if (ir->value.u64[i] != 0)
+               print_without_indent("r%04X_data.u64[%u] = %" PRIu64 ";\n",
+                                    my_index,
+                                    i,
+                                    ir->value.u64[i]);
+            break;
+         case GLSL_TYPE_INT64:
+            if (ir->value.i64[i] != 0)
+               print_without_indent("r%04X_data.i64[%u] = %" PRId64 ";\n",
+                                    my_index,
+                                    i,
+                                    ir->value.i64[i]);
+            break;
          case GLSL_TYPE_BOOL:
             if (ir->value.u[i] != 0)
                print_without_indent("r%04X_data.u[%u] = 1;\n", my_index, i);

src/compiler/glsl/ir_clone.cpp

@@ -337,6 +337,8 @@ ir_constant::clone(void *mem_ctx, struct hash_table *ht) const
    case GLSL_TYPE_FLOAT:
    case GLSL_TYPE_DOUBLE:
    case GLSL_TYPE_BOOL:
+   case GLSL_TYPE_UINT64:
+   case GLSL_TYPE_INT64:
       return new(mem_ctx) ir_constant(this->type, &this->value);
 
    case GLSL_TYPE_STRUCT: {

src/compiler/glsl/ir_constant_expression.cpp

@@ -88,6 +88,42 @@ bitcast_f2u(float f)
    return u;
 }
 
+static double
+bitcast_u642d(uint64_t u)
+{
+   assert(sizeof(double) == sizeof(uint64_t));
+   double d;
+   memcpy(&d, &u, sizeof(d));
+   return d;
+}
+
+static double
+bitcast_i642d(int64_t i)
+{
+   assert(sizeof(double) == sizeof(int64_t));
+   double d;
+   memcpy(&d, &i, sizeof(d));
+   return d;
+}
+
+static double
+bitcast_d2u64(double d)
+{
+   assert(sizeof(double) == sizeof(uint64_t));
+   uint64_t u;
+   memcpy(&u, &d, sizeof(d));
+   return u;
+}
+
+static double
+bitcast_d2i64(double d)
+{
+   assert(sizeof(double) == sizeof(int64_t));
+   int64_t i;
+   memcpy(&i, &d, sizeof(d));
+   return i;
+}
+
 /**
  * Evaluate one component of a floating-point 4x8 unpacking function.
  */

src/compiler/glsl/ir_expression_operation.py

@@ -80,14 +80,16 @@ class type_signature_iter(object):
 
 uint_type = type("unsigned", "u", "GLSL_TYPE_UINT")
 int_type = type("int", "i", "GLSL_TYPE_INT")
+uint64_type = type("uint64_t", "u64", "GLSL_TYPE_UINT64")
+int64_type = type("int64_t", "i64", "GLSL_TYPE_INT64")
 float_type = type("float", "f", "GLSL_TYPE_FLOAT")
 double_type = type("double", "d", "GLSL_TYPE_DOUBLE")
 bool_type = type("bool", "b", "GLSL_TYPE_BOOL")
 
-all_types = (uint_type, int_type, float_type, double_type, bool_type)
-numeric_types = (uint_type, int_type, float_type, double_type)
-signed_numeric_types = (int_type, float_type, double_type)
-integer_types = (uint_type, int_type)
+all_types = (uint_type, int_type, float_type, double_type, uint64_type, int64_type, bool_type)
+numeric_types = (uint_type, int_type, float_type, double_type, uint64_type, int64_type)
+signed_numeric_types = (int_type, float_type, double_type, int64_type)
+integer_types = (uint_type, int_type, uint64_type, int64_type)
 real_types = (float_type, double_type)
 
 # This template is for operations that can have operands of a several
@@ -418,8 +420,8 @@ ir_expression_operation = [
    operation("bit_not", 1, printable_name="~", source_types=integer_types, c_expression="~ {src0}"),
    operation("logic_not", 1, printable_name="!", source_types=(bool_type,), c_expression="!{src0}"),
    operation("neg", 1, source_types=numeric_types, c_expression={'u': "-((int) {src0})", 'default': "-{src0}"}),
-   operation("abs", 1, source_types=signed_numeric_types, c_expression={'i': "{src0} < 0 ? -{src0} : {src0}", 'f': "fabsf({src0})", 'd': "fabs({src0})"}),
-   operation("sign", 1, source_types=signed_numeric_types, c_expression={'i': "({src0} > 0) - ({src0} < 0)", 'f': "float(({src0} > 0.0F) - ({src0} < 0.0F))", 'd': "double(({src0} > 0.0) - ({src0} < 0.0))"}),
+   operation("abs", 1, source_types=signed_numeric_types, c_expression={'i': "{src0} < 0 ? -{src0} : {src0}", 'f': "fabsf({src0})", 'd': "fabs({src0})", 'i64': "{src0} < 0 ? -{src0} : {src0}"}),
+   operation("sign", 1, source_types=signed_numeric_types, c_expression={'i': "({src0} > 0) - ({src0} < 0)", 'f': "float(({src0} > 0.0F) - ({src0} < 0.0F))", 'd': "double(({src0} > 0.0) - ({src0} < 0.0))", 'i64': "({src0} > 0) - ({src0} < 0)"}),
    operation("rcp", 1, source_types=real_types, c_expression={'f': "{src0} != 0.0F ? 1.0F / {src0} : 0.0F", 'd': "{src0} != 0.0 ? 1.0 / {src0} : 0.0"}),
    operation("rsq", 1, source_types=real_types, c_expression={'f': "1.0F / sqrtf({src0})", 'd': "1.0 / sqrt({src0})"}),
    operation("sqrt", 1, source_types=real_types, c_expression={'f': "sqrtf({src0})", 'd': "sqrt({src0})"}),
@@ -439,7 +441,7 @@ ir_expression_operation = [
    # Boolean-to-float conversion
    operation("b2f", 1, source_types=(bool_type,), dest_type=float_type, c_expression="{src0} ? 1.0F : 0.0F"),
    # int-to-boolean conversion
-   operation("i2b", 1, source_types=integer_types, dest_type=bool_type, c_expression="{src0} ? true : false"),
+   operation("i2b", 1, source_types=(uint_type, int_type), dest_type=bool_type, c_expression="{src0} ? true : false"),
    # Boolean-to-int conversion
    operation("b2i", 1, source_types=(bool_type,), dest_type=int_type, c_expression="{src0} ? 1 : 0"),
    # Unsigned-to-float conversion.
@@ -470,6 +472,37 @@ ir_expression_operation = [
    operation("bitcast_u2f", 1, source_types=(uint_type,), dest_type=float_type, c_expression="bitcast_u2f({src0})"),
    # 'Bit-identical float-to-uint "conversion"
    operation("bitcast_f2u", 1, source_types=(float_type,), dest_type=uint_type, c_expression="bitcast_f2u({src0})"),
+   # Bit-identical u64-to-double "conversion"
+   operation("bitcast_u642d", 1, source_types=(uint64_type,), dest_type=double_type, c_expression="bitcast_u642d({src0})"),
+   # Bit-identical i64-to-double "conversion"
+   operation("bitcast_i642d", 1, source_types=(int64_type,), dest_type=double_type, c_expression="bitcast_i642d({src0})"),
+   # Bit-identical double-to_u64 "conversion"
+   operation("bitcast_d2u64", 1, source_types=(double_type,), dest_type=uint64_type, c_expression="bitcast_d2u64({src0})"),
+   # Bit-identical double-to-i64 "conversion"
+   operation("bitcast_d2i64", 1, source_types=(double_type,), dest_type=int64_type, c_expression="bitcast_d2i64({src0})"),
+   # i64-to-i32 conversion
+   operation("i642i", 1, source_types=(int64_type,), dest_type=int_type, c_expression="{src0}"),
+   # ui64-to-i32 conversion
+   operation("u642i", 1, source_types=(uint64_type,), dest_type=int_type, c_expression="{src0}"),
+   operation("i642u", 1, source_types=(int64_type,), dest_type=uint_type, c_expression="{src0}"),
+   operation("u642u", 1, source_types=(uint64_type,), dest_type=uint_type, c_expression="{src0}"),
+   operation("i642b", 1, source_types=(int64_type,), dest_type=bool_type, c_expression="{src0} != 0"),
+   operation("i642f", 1, source_types=(int64_type,), dest_type=float_type, c_expression="{src0}"),
+   operation("u642f", 1, source_types=(uint64_type,), dest_type=float_type, c_expression="{src0}"),
+   operation("i642d", 1, source_types=(int64_type,), dest_type=double_type, c_expression="{src0}"),
+   operation("u642d", 1, source_types=(uint64_type,), dest_type=double_type, c_expression="{src0}"),
+   operation("i2i64", 1, source_types=(int_type,), dest_type=int64_type, c_expression="{src0}"),
+   operation("u2i64", 1, source_types=(uint_type,), dest_type=int64_type, c_expression="{src0}"),
+   operation("b2i64", 1, source_types=(bool_type,), dest_type=int64_type, c_expression="{src0}"),
+   operation("f2i64", 1, source_types=(float_type,), dest_type=int64_type, c_expression="{src0}"),
+   operation("d2i64", 1, source_types=(double_type,), dest_type=int64_type, c_expression="{src0}"),
+   operation("i2u64", 1, source_types=(int_type,), dest_type=uint64_type, c_expression="{src0}"),
+   operation("u2u64", 1, source_types=(uint_type,), dest_type=uint64_type, c_expression="{src0}"),
+   operation("f2u64", 1, source_types=(float_type,), dest_type=uint64_type, c_expression="{src0}"),
+   operation("d2u64", 1, source_types=(double_type,), dest_type=uint64_type, c_expression="{src0}"),
+   operation("u642i64", 1, source_types=(uint64_type,), dest_type=int64_type, c_expression="{src0}"),
+   operation("i642u64", 1, source_types=(int64_type,), dest_type=uint64_type, c_expression="{src0}"),
+
 
    # Unary floating-point rounding operations.
    operation("trunc", 1, source_types=real_types, c_expression={'f': "truncf({src0})", 'd': "trunc({src0})"}),
@@ -503,10 +536,10 @@ ir_expression_operation = [
    operation("unpack_half_2x16", 1, printable_name="unpackHalf2x16", source_types=(uint_type,), dest_type=float_type, c_expression="unpack_2x16(unpack_half_1x16, op[0]->value.u[0], &data.f[0], &data.f[1])", flags=frozenset((horizontal_operation, non_assign_operation))),
 
    # Bit operations, part of ARB_gpu_shader5.
-   operation("bitfield_reverse", 1, source_types=integer_types, c_expression="bitfield_reverse({src0})"),
-   operation("bit_count", 1, source_types=integer_types, dest_type=int_type, c_expression="_mesa_bitcount({src0})"),
-   operation("find_msb", 1, source_types=integer_types, dest_type=int_type, c_expression={'u': "find_msb_uint({src0})", 'i': "find_msb_int({src0})"}),
-   operation("find_lsb", 1, source_types=integer_types, dest_type=int_type, c_expression="find_msb_uint({src0} & -{src0})"),
+   operation("bitfield_reverse", 1, source_types=(uint_type, int_type), c_expression="bitfield_reverse({src0})"),
+   operation("bit_count", 1, source_types=(uint_type, int_type), dest_type=int_type, c_expression="_mesa_bitcount({src0})"),
+   operation("find_msb", 1, source_types=(uint_type, int_type), dest_type=int_type, c_expression={'u': "find_msb_uint({src0})", 'i': "find_msb_int({src0})"}),
+   operation("find_lsb", 1, source_types=(uint_type, int_type), dest_type=int_type, c_expression="find_msb_uint({src0} & -{src0})"),
 
    operation("saturate", 1, printable_name="sat", source_types=(float_type,), c_expression="CLAMP({src0}, 0.0f, 1.0f)"),
 
@@ -543,12 +576,18 @@ ir_expression_operation = [
    operation("vote_all", 1),
    operation("vote_eq", 1),
 
+   # 64-bit integer packing ops.
+   operation("pack_int_2x32", 1, printable_name="packInt2x32", source_types=(int_type,), dest_type=int64_type, c_expression="memcpy(&data.i64[0], &op[0]->value.i[0], sizeof(int64_t))", flags=frozenset((horizontal_operation, non_assign_operation))),
+   operation("pack_uint_2x32", 1, printable_name="packUint2x32", source_types=(uint_type,), dest_type=uint64_type, c_expression="memcpy(&data.u64[0], &op[0]->value.u[0], sizeof(uint64_t))", flags=frozenset((horizontal_operation, non_assign_operation))),
+   operation("unpack_int_2x32", 1, printable_name="unpackInt2x32", source_types=(int64_type,), dest_type=int_type, c_expression="memcpy(&data.i[0], &op[0]->value.i64[0], sizeof(int64_t))", flags=frozenset((horizontal_operation, non_assign_operation))),
+   operation("unpack_uint_2x32", 1, printable_name="unpackUint2x32", source_types=(uint64_type,), dest_type=uint_type, c_expression="memcpy(&data.u[0], &op[0]->value.u64[0], sizeof(uint64_t))", flags=frozenset((horizontal_operation, non_assign_operation))),
+
    operation("add", 2, printable_name="+", source_types=numeric_types, c_expression="{src0} + {src1}", flags=vector_scalar_operation),
    operation("sub", 2, printable_name="-", source_types=numeric_types, c_expression="{src0} - {src1}", flags=vector_scalar_operation),
    # "Floating-point or low 32-bit integer multiply."
    operation("mul", 2, printable_name="*", source_types=numeric_types, c_expression="{src0} * {src1}"),
    operation("imul_high", 2),       # Calculates the high 32-bits of a 64-bit multiply.
-   operation("div", 2, printable_name="/", source_types=numeric_types, c_expression={'u': "{src1} == 0 ? 0 : {src0} / {src1}", 'i': "{src1} == 0 ? 0 : {src0} / {src1}", 'default': "{src0} / {src1}"}, flags=vector_scalar_operation),
+   operation("div", 2, printable_name="/", source_types=numeric_types, c_expression={'u': "{src1} == 0 ? 0 : {src0} / {src1}", 'i': "{src1} == 0 ? 0 : {src0} / {src1}", 'u64': "{src1} == 0 ? 0 : {src0} / {src1}", 'i64': "{src1} == 0 ? 0 : {src0} / {src1}", 'default': "{src0} / {src1}"}, flags=vector_scalar_operation),
 
    # Returns the carry resulting from the addition of the two arguments.
    operation("carry", 2),
@@ -561,7 +600,7 @@ ir_expression_operation = [
    #
    # We don't use fmod because it rounds toward zero; GLSL specifies the use
    # of floor.
-   operation("mod", 2, printable_name="%", source_types=numeric_types, c_expression={'u': "{src1} == 0 ? 0 : {src0} % {src1}", 'i': "{src1} == 0 ? 0 : {src0} % {src1}", 'f': "{src0} - {src1} * floorf({src0} / {src1})", 'd': "{src0} - {src1} * floor({src0} / {src1})"}, flags=vector_scalar_operation),
+   operation("mod", 2, printable_name="%", source_types=numeric_types, c_expression={'u': "{src1} == 0 ? 0 : {src0} % {src1}", 'i': "{src1} == 0 ? 0 : {src0} % {src1}", 'f': "{src0} - {src1} * floorf({src0} / {src1})", 'd': "{src0} - {src1} * floor({src0} / {src1})", 'u64': "{src1} == 0 ? 0 : {src0} % {src1}", 'i64': "{src1} == 0 ? 0 : {src0} % {src1}"}, flags=vector_scalar_operation),
 
    # Binary comparison operators which return a boolean vector.
    # The type of both operands must be equal.

src/compiler/glsl/ir_optimization.h

@@ -52,6 +52,12 @@
 #define DDIV_TO_MUL_RCP           0x100000
 #define DIV_TO_MUL_RCP            (FDIV_TO_MUL_RCP | DDIV_TO_MUL_RCP)
 
+/* Opertaions for lower_64bit_integer_instructions() */
+#define MUL64                     (1U << 0)
+#define SIGN64                    (1U << 1)
+#define DIV64                     (1U << 2)
+#define MOD64                     (1U << 3)
+
 /**
  * \see class lower_packing_builtins_visitor
  */
@@ -164,3 +170,6 @@ void propagate_invariance(exec_list *instructions);
 ir_rvalue *
 compare_index_block(exec_list *instructions, ir_variable *index,
 		    unsigned base, unsigned components, void *mem_ctx);
+
+bool lower_64bit_integer_instructions(exec_list *instructions,
+                                      unsigned what_to_lower);

src/compiler/glsl/ir_print_visitor.cpp

@@ -21,6 +21,8 @@
  * DEALINGS IN THE SOFTWARE.
  */
 
+#define __STDC_FORMAT_MACROS 1
+#include <inttypes.h> /* for PRIx64 macro */
 #include "ir_print_visitor.h"
 #include "compiler/glsl_types.h"
 #include "glsl_parser_extras.h"
@@ -476,6 +478,8 @@ void ir_print_visitor::visit(ir_constant *ir)
             else
                fprintf(f, "%f", ir->value.f[i]);
             break;
+	 case GLSL_TYPE_UINT64:fprintf(f, "%" PRIu64, ir->value.u64[i]); break;
+	 case GLSL_TYPE_INT64: fprintf(f, "%" PRIi64, ir->value.i64[i]); break;
 	 case GLSL_TYPE_BOOL:  fprintf(f, "%d", ir->value.b[i]); break;
 	 case GLSL_TYPE_DOUBLE:
             if (ir->value.d[i] == 0.0)

src/compiler/glsl/ir_validate.cpp

@@ -253,7 +253,8 @@ ir_validate::visit_leave(ir_expression *ir)
    case ir_unop_sign:
       assert(ir->operands[0]->type->base_type == GLSL_TYPE_INT ||
              ir->operands[0]->type->base_type == GLSL_TYPE_FLOAT ||
-             ir->operands[0]->type->base_type == GLSL_TYPE_DOUBLE);
+             ir->operands[0]->type->base_type == GLSL_TYPE_DOUBLE ||
+             ir->operands[0]->type->base_type == GLSL_TYPE_INT64);
       assert(ir->type == ir->operands[0]->type);
       break;
 
@@ -331,6 +332,102 @@ ir_validate::visit_leave(ir_expression *ir)
       assert(ir->type->base_type == GLSL_TYPE_UINT);
       break;
 
+   case ir_unop_bitcast_u642d:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_UINT64);
+      assert(ir->type->base_type == GLSL_TYPE_DOUBLE);
+      break;
+   case ir_unop_bitcast_i642d:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_INT64);
+      assert(ir->type->base_type == GLSL_TYPE_DOUBLE);
+      break;
+   case ir_unop_bitcast_d2u64:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_DOUBLE);
+      assert(ir->type->base_type == GLSL_TYPE_UINT64);
+      break;
+   case ir_unop_bitcast_d2i64:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_DOUBLE);
+      assert(ir->type->base_type == GLSL_TYPE_INT64);
+      break;
+   case ir_unop_i642i:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_INT64);
+      assert(ir->type->base_type == GLSL_TYPE_INT);
+      break;
+   case ir_unop_u642i:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_UINT64);
+      assert(ir->type->base_type == GLSL_TYPE_INT);
+      break;
+   case ir_unop_i642u:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_INT64);
+      assert(ir->type->base_type == GLSL_TYPE_UINT);
+      break;
+   case ir_unop_u642u:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_UINT64);
+      assert(ir->type->base_type == GLSL_TYPE_UINT);
+      break;
+   case ir_unop_i642b:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_INT64);
+      assert(ir->type->base_type == GLSL_TYPE_BOOL);
+      break;
+   case ir_unop_i642f:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_INT64);
+      assert(ir->type->base_type == GLSL_TYPE_FLOAT);
+      break;
+   case ir_unop_u642f:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_UINT64);
+      assert(ir->type->base_type == GLSL_TYPE_FLOAT);
+      break;
+   case ir_unop_i642d:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_INT64);
+      assert(ir->type->base_type == GLSL_TYPE_DOUBLE);
+      break;
+   case ir_unop_u642d:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_UINT64);
+      assert(ir->type->base_type == GLSL_TYPE_DOUBLE);
+      break;
+   case ir_unop_i2i64:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_INT);
+      assert(ir->type->base_type == GLSL_TYPE_INT64);
+      break;
+   case ir_unop_u2i64:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_UINT);
+      assert(ir->type->base_type == GLSL_TYPE_INT64);
+      break;
+   case ir_unop_b2i64:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_BOOL);
+      assert(ir->type->base_type == GLSL_TYPE_INT64);
+      break;
+   case ir_unop_f2i64:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_FLOAT);
+      assert(ir->type->base_type == GLSL_TYPE_INT64);
+      break;
+   case ir_unop_d2i64:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_DOUBLE);
+      assert(ir->type->base_type == GLSL_TYPE_INT64);
+      break;
+   case ir_unop_i2u64:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_INT);
+      assert(ir->type->base_type == GLSL_TYPE_UINT64);
+      break;
+   case ir_unop_u2u64:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_UINT);
+      assert(ir->type->base_type == GLSL_TYPE_UINT64);
+      break;
+   case ir_unop_f2u64:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_FLOAT);
+      assert(ir->type->base_type == GLSL_TYPE_UINT64);
+      break;
+   case ir_unop_d2u64:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_DOUBLE);
+      assert(ir->type->base_type == GLSL_TYPE_UINT64);
+      break;
+   case ir_unop_u642i64:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_UINT64);
+      assert(ir->type->base_type == GLSL_TYPE_INT64);
+      break;
+   case ir_unop_i642u64:
+      assert(ir->operands[0]->type->base_type == GLSL_TYPE_INT64);
+      assert(ir->type->base_type == GLSL_TYPE_UINT64);
+      break;
    case ir_unop_trunc:
    case ir_unop_round_even:
    case ir_unop_ceil:
@@ -370,6 +467,16 @@ ir_validate::visit_leave(ir_expression *ir)
       assert(ir->operands[0]->type == glsl_type::uvec2_type);
       break;
 
+   case ir_unop_pack_int_2x32:
+      assert(ir->type == glsl_type::int64_t_type);
+      assert(ir->operands[0]->type == glsl_type::ivec2_type);
+      break;
+
+   case ir_unop_pack_uint_2x32:
+      assert(ir->type == glsl_type::uint64_t_type);
+      assert(ir->operands[0]->type == glsl_type::uvec2_type);
+      break;
+
    case ir_unop_unpack_snorm_2x16:
    case ir_unop_unpack_unorm_2x16:
    case ir_unop_unpack_half_2x16:
@@ -388,6 +495,16 @@ ir_validate::visit_leave(ir_expression *ir)
       assert(ir->operands[0]->type == glsl_type::double_type);
       break;
 
+   case ir_unop_unpack_int_2x32:
+      assert(ir->type == glsl_type::ivec2_type);
+      assert(ir->operands[0]->type == glsl_type::int64_t_type);
+      break;
+
+   case ir_unop_unpack_uint_2x32:
+      assert(ir->type == glsl_type::uvec2_type);
+      assert(ir->operands[0]->type == glsl_type::uint64_t_type);
+      break;
+
    case ir_unop_bitfield_reverse:
       assert(ir->operands[0]->type == ir->type);
       assert(ir->type->is_integer());
@@ -537,7 +654,7 @@ ir_validate::visit_leave(ir_expression *ir)
 
    case ir_binop_lshift:
    case ir_binop_rshift:
-      assert(ir->operands[0]->type->is_integer() &&
+      assert(ir->operands[0]->type->is_integer_32_64() &&
              ir->operands[1]->type->is_integer());
       if (ir->operands[0]->type->is_scalar()) {
           assert(ir->operands[1]->type->is_scalar());
@@ -555,7 +672,7 @@ ir_validate::visit_leave(ir_expression *ir)
    case ir_binop_bit_or:
        assert(ir->operands[0]->type->base_type ==
               ir->operands[1]->type->base_type);
-       assert(ir->type->is_integer());
+       assert(ir->type->is_integer_32_64());
        if (ir->operands[0]->type->is_vector() &&
            ir->operands[1]->type->is_vector()) {
            assert(ir->operands[0]->type->vector_elements ==

src/compiler/glsl/link_uniform_initializers.cpp

@@ -64,6 +64,8 @@ copy_constant_to_storage(union gl_constant_value *storage,
          storage[i].f = val->value.f[i];
          break;
       case GLSL_TYPE_DOUBLE:
+      case GLSL_TYPE_UINT64:
+      case GLSL_TYPE_INT64:
          /* XXX need to check on big-endian */
          memcpy(&storage[i * 2].u, &val->value.d[i], sizeof(double));
          break;

src/compiler/glsl/link_uniforms.cpp

@@ -535,7 +535,7 @@ private:
             const char *str_end;
             while((str_start = strchr(name_copy, '[')) &&
                   (str_end = strchr(name_copy, ']'))) {
-               memmove(str_start, str_end + 1, 1 + strlen(str_end + 1));
+               memmove(str_start, str_end + 1, 1 + strlen(str_end));
             }
 
             unsigned index = 0;
@@ -1042,12 +1042,10 @@ find_empty_block(struct gl_shader_program *prog,
 
 static void
 link_setup_uniform_remap_tables(struct gl_context *ctx,
-                                struct gl_shader_program *prog,
-                                unsigned num_explicit_uniform_locs)
+                                struct gl_shader_program *prog)
 {
-   unsigned total_entries = num_explicit_uniform_locs;
-   unsigned empty_locs =
-      prog->NumUniformRemapTable - num_explicit_uniform_locs;
+   unsigned total_entries = prog->NumExplicitUniformLocations;
+   unsigned empty_locs = prog->NumUniformRemapTable - total_entries;
 
    /* Reserve all the explicit locations of the active uniforms. */
    for (unsigned i = 0; i < prog->data->NumUniformStorage; i++) {
@@ -1206,8 +1204,7 @@ link_setup_uniform_remap_tables(struct gl_context *ctx,
 static void
 link_assign_uniform_storage(struct gl_context *ctx,
                             struct gl_shader_program *prog,
-                            const unsigned num_data_slots,
-                            unsigned num_explicit_uniform_locs)
+                            const unsigned num_data_slots)
 {
    /* On the outside chance that there were no uniforms, bail out.
     */
@@ -1266,15 +1263,14 @@ link_assign_uniform_storage(struct gl_context *ctx,
    assert(parcel.values == data_end);
 #endif
 
-   link_setup_uniform_remap_tables(ctx, prog, num_explicit_uniform_locs);
+   link_setup_uniform_remap_tables(ctx, prog);
 
    link_set_uniform_initializers(prog, boolean_true);
 }
 
 void
 link_assign_uniform_locations(struct gl_shader_program *prog,
-                              struct gl_context *ctx,
-                              unsigned int num_explicit_uniform_locs)
+                              struct gl_context *ctx)
 {
    ralloc_free(prog->data->UniformStorage);
    prog->data->UniformStorage = NULL;
@@ -1335,6 +1331,5 @@ link_assign_uniform_locations(struct gl_shader_program *prog,
    hiddenUniforms->iterate(assign_hidden_uniform_slot_id, &uniform_size);
    delete hiddenUniforms;
 
-   link_assign_uniform_storage(ctx, prog, uniform_size.num_values,
-                               num_explicit_uniform_locs);
+   link_assign_uniform_storage(ctx, prog, uniform_size.num_values);
 }

src/compiler/glsl/link_varyings.cpp

@@ -106,8 +106,9 @@ create_xfb_varying_names(void *mem_ctx, const glsl_type *t, char **name,
    }
 }
 
-bool
+static bool
 process_xfb_layout_qualifiers(void *mem_ctx, const gl_linked_shader *sh,
+                              struct gl_shader_program *prog,
                               unsigned *num_tfeedback_decls,
                               char ***varying_names)
 {
@@ -118,8 +119,9 @@ process_xfb_layout_qualifiers(void *mem_ctx, const gl_linked_shader *sh,
     * xfb_stride to interface block members so this will catch that case also.
     */
    for (unsigned j = 0; j < MAX_FEEDBACK_BUFFERS; j++) {
-      if (sh->info.TransformFeedback.BufferStride[j]) {
+      if (prog->TransformFeedback.BufferStride[j]) {
          has_xfb_qualifiers = true;
+         break;
       }
    }
 
@@ -573,7 +575,7 @@ cross_validate_outputs_to_inputs(struct gl_shader_program *prog,
  * Demote shader inputs and outputs that are not used in other stages, and
  * remove them via dead code elimination.
  */
-void
+static void
 remove_unused_shader_inputs_and_outputs(bool is_separate_shader_object,
                                         gl_linked_shader *sh,
                                         enum ir_variable_mode mode)
@@ -743,10 +745,12 @@ tfeedback_decl::assign_location(struct gl_context *ctx,
       unsigned actual_array_size;
       switch (this->lowered_builtin_array_variable) {
       case clip_distance:
-         actual_array_size = prog->LastClipDistanceArraySize;
+         actual_array_size = prog->last_vert_prog ?
+            prog->last_vert_prog->info.clip_distance_array_size : 0;
          break;
       case cull_distance:
-         actual_array_size = prog->LastCullDistanceArraySize;
+         actual_array_size = prog->last_vert_prog ?
+            prog->last_vert_prog->info.cull_distance_array_size : 0;
          break;
       case tess_level_outer:
          actual_array_size = 4;
@@ -1014,7 +1018,7 @@ tfeedback_decl::find_candidate(gl_shader_program *prog,
  * If an error occurs, the error is reported through linker_error() and false
  * is returned.
  */
-bool
+static bool
 parse_tfeedback_decls(struct gl_context *ctx, struct gl_shader_program *prog,
                       const void *mem_ctx, unsigned num_names,
                       char **varying_names, tfeedback_decl *decls)
@@ -1069,11 +1073,14 @@ cmp_xfb_offset(const void * x_generic, const void * y_generic)
  * If an error occurs, the error is reported through linker_error() and false
  * is returned.
  */
-bool
+static bool
 store_tfeedback_info(struct gl_context *ctx, struct gl_shader_program *prog,
                      unsigned num_tfeedback_decls,
                      tfeedback_decl *tfeedback_decls, bool has_xfb_qualifiers)
 {
+   if (!prog->last_vert_prog)
+      return true;
+
    /* Make sure MaxTransformFeedbackBuffers is less than 32 so the bitmask for
     * tracking the number of buffers doesn't overflow.
     */
@@ -1082,7 +1089,7 @@ store_tfeedback_info(struct gl_context *ctx, struct gl_shader_program *prog,
    bool separate_attribs_mode =
       prog->TransformFeedback.BufferMode == GL_SEPARATE_ATTRIBS;
 
-   struct gl_program *xfb_prog = prog->xfb_program;
+   struct gl_program *xfb_prog = prog->last_vert_prog;
    xfb_prog->sh.LinkedTransformFeedback =
       rzalloc(xfb_prog, struct gl_transform_feedback_info);
 
@@ -1985,7 +1992,7 @@ canonicalize_shader_io(exec_list *ir, enum ir_variable_mode io_mode)
  * 64 bit map. Per-vertex and per-patch both have separate location domains
  * with a max of MAX_VARYING.
  */
-uint64_t
+static uint64_t
 reserved_varying_slot(struct gl_linked_shader *stage,
                       ir_variable_mode io_mode)
 {
@@ -2042,7 +2049,7 @@ reserved_varying_slot(struct gl_linked_shader *stage,
  * be NULL.  In this case, varying locations are assigned solely based on the
  * requirements of transform feedback.
  */
-bool
+static bool
 assign_varying_locations(struct gl_context *ctx,
                          void *mem_ctx,
                          struct gl_shader_program *prog,
@@ -2373,3 +2380,160 @@ check_against_input_limit(struct gl_context *ctx,
 
    return true;
 }
+
+bool
+link_varyings(struct gl_shader_program *prog, unsigned first, unsigned last,
+              struct gl_context *ctx, void *mem_ctx)
+{
+   bool has_xfb_qualifiers = false;
+   unsigned num_tfeedback_decls = 0;
+   char **varying_names = NULL;
+   tfeedback_decl *tfeedback_decls = NULL;
+
+   /* From the ARB_enhanced_layouts spec:
+    *
+    *    "If the shader used to record output variables for transform feedback
+    *    varyings uses the "xfb_buffer", "xfb_offset", or "xfb_stride" layout
+    *    qualifiers, the values specified by TransformFeedbackVaryings are
+    *    ignored, and the set of variables captured for transform feedback is
+    *    instead derived from the specified layout qualifiers."
+    */
+   for (int i = MESA_SHADER_FRAGMENT - 1; i >= 0; i--) {
+      /* Find last stage before fragment shader */
+      if (prog->_LinkedShaders[i]) {
+         has_xfb_qualifiers =
+            process_xfb_layout_qualifiers(mem_ctx, prog->_LinkedShaders[i],
+                                          prog, &num_tfeedback_decls,
+                                          &varying_names);
+         break;
+      }
+   }
+
+   if (!has_xfb_qualifiers) {
+      num_tfeedback_decls = prog->TransformFeedback.NumVarying;
+      varying_names = prog->TransformFeedback.VaryingNames;
+   }
+
+   if (num_tfeedback_decls != 0) {
+      /* From GL_EXT_transform_feedback:
+       *   A program will fail to link if:
+       *
+       *   * the <count> specified by TransformFeedbackVaryingsEXT is
+       *     non-zero, but the program object has no vertex or geometry
+       *     shader;
+       */
+      if (first >= MESA_SHADER_FRAGMENT) {
+         linker_error(prog, "Transform feedback varyings specified, but "
+                      "no vertex, tessellation, or geometry shader is "
+                      "present.\n");
+         return false;
+      }
+
+      tfeedback_decls = rzalloc_array(mem_ctx, tfeedback_decl,
+                                      num_tfeedback_decls);
+      if (!parse_tfeedback_decls(ctx, prog, mem_ctx, num_tfeedback_decls,
+                                 varying_names, tfeedback_decls))
+         return false;
+   }
+
+   /* If there is no fragment shader we need to set transform feedback.
+    *
+    * For SSO we also need to assign output locations.  We assign them here
+    * because we need to do it for both single stage programs and multi stage
+    * programs.
+    */
+   if (last < MESA_SHADER_FRAGMENT &&
+       (num_tfeedback_decls != 0 || prog->SeparateShader)) {
+      const uint64_t reserved_out_slots =
+         reserved_varying_slot(prog->_LinkedShaders[last], ir_var_shader_out);
+      if (!assign_varying_locations(ctx, mem_ctx, prog,
+                                    prog->_LinkedShaders[last], NULL,
+                                    num_tfeedback_decls, tfeedback_decls,
+                                    reserved_out_slots))
+         return false;
+   }
+
+   if (last <= MESA_SHADER_FRAGMENT) {
+      /* Remove unused varyings from the first/last stage unless SSO */
+      remove_unused_shader_inputs_and_outputs(prog->SeparateShader,
+                                              prog->_LinkedShaders[first],
+                                              ir_var_shader_in);
+      remove_unused_shader_inputs_and_outputs(prog->SeparateShader,
+                                              prog->_LinkedShaders[last],
+                                              ir_var_shader_out);
+
+      /* If the program is made up of only a single stage */
+      if (first == last) {
+         gl_linked_shader *const sh = prog->_LinkedShaders[last];
+
+         do_dead_builtin_varyings(ctx, NULL, sh, 0, NULL);
+         do_dead_builtin_varyings(ctx, sh, NULL, num_tfeedback_decls,
+                                  tfeedback_decls);
+
+         if (prog->SeparateShader) {
+            const uint64_t reserved_slots =
+               reserved_varying_slot(sh, ir_var_shader_in);
+
+            /* Assign input locations for SSO, output locations are already
+             * assigned.
+             */
+            if (!assign_varying_locations(ctx, mem_ctx, prog,
+                                          NULL /* producer */,
+                                          sh /* consumer */,
+                                          0 /* num_tfeedback_decls */,
+                                          NULL /* tfeedback_decls */,
+                                          reserved_slots))
+               return false;
+         }
+      } else {
+         /* Linking the stages in the opposite order (from fragment to vertex)
+          * ensures that inter-shader outputs written to in an earlier stage
+          * are eliminated if they are (transitively) not used in a later
+          * stage.
+          */
+         int next = last;
+         for (int i = next - 1; i >= 0; i--) {
+            if (prog->_LinkedShaders[i] == NULL && i != 0)
+               continue;
+
+            gl_linked_shader *const sh_i = prog->_LinkedShaders[i];
+            gl_linked_shader *const sh_next = prog->_LinkedShaders[next];
+
+            const uint64_t reserved_out_slots =
+               reserved_varying_slot(sh_i, ir_var_shader_out);
+            const uint64_t reserved_in_slots =
+               reserved_varying_slot(sh_next, ir_var_shader_in);
+
+            do_dead_builtin_varyings(ctx, sh_i, sh_next,
+                      next == MESA_SHADER_FRAGMENT ? num_tfeedback_decls : 0,
+                      tfeedback_decls);
+
+            if (!assign_varying_locations(ctx, mem_ctx, prog, sh_i, sh_next,
+                      next == MESA_SHADER_FRAGMENT ? num_tfeedback_decls : 0,
+                      tfeedback_decls,
+                      reserved_out_slots | reserved_in_slots))
+               return false;
+
+            /* This must be done after all dead varyings are eliminated. */
+            if (sh_i != NULL) {
+               unsigned slots_used = _mesa_bitcount_64(reserved_out_slots);
+               if (!check_against_output_limit(ctx, prog, sh_i, slots_used)) {
+                  return false;
+               }
+            }
+
+            unsigned slots_used = _mesa_bitcount_64(reserved_in_slots);
+            if (!check_against_input_limit(ctx, prog, sh_next, slots_used))
+               return false;
+
+            next = i;
+         }
+      }
+   }
+
+   if (!store_tfeedback_info(ctx, prog, num_tfeedback_decls, tfeedback_decls,
+                             has_xfb_qualifiers))
+      return false;
+
+   return true;
+}

src/compiler/glsl/link_varyings.h

@@ -176,6 +176,14 @@ private:
       case GL_DOUBLE_MAT4:
       case GL_DOUBLE_MAT4x2:
       case GL_DOUBLE_MAT4x3:
+      case GL_INT64_ARB:
+      case GL_INT64_VEC2_ARB:
+      case GL_INT64_VEC3_ARB:
+      case GL_INT64_VEC4_ARB:
+      case GL_UNSIGNED_INT64_ARB:
+      case GL_UNSIGNED_INT64_VEC2_ARB:
+      case GL_UNSIGNED_INT64_VEC3_ARB:
+      case GL_UNSIGNED_INT64_VEC4_ARB:
          return true;
       default:
          return false;
@@ -288,57 +296,13 @@ private:
    unsigned stream_id;
 };
 
+bool
+link_varyings(struct gl_shader_program *prog, unsigned first, unsigned last,
+              struct gl_context *ctx, void *mem_ctx);
 
 void
 cross_validate_outputs_to_inputs(struct gl_shader_program *prog,
                                  gl_linked_shader *producer,
                                  gl_linked_shader *consumer);
 
-bool
-parse_tfeedback_decls(struct gl_context *ctx, struct gl_shader_program *prog,
-                      const void *mem_ctx, unsigned num_names,
-                      char **varying_names, tfeedback_decl *decls);
-
-bool
-process_xfb_layout_qualifiers(void *mem_ctx, const gl_linked_shader *sh,
-                              unsigned *num_tfeedback_decls,
-                              char ***varying_names);
-
-void
-remove_unused_shader_inputs_and_outputs(bool is_separate_shader_object,
-                                        gl_linked_shader *sh,
-                                        enum ir_variable_mode mode);
-
-bool
-store_tfeedback_info(struct gl_context *ctx, struct gl_shader_program *prog,
-                     unsigned num_tfeedback_decls,
-                     tfeedback_decl *tfeedback_decls,
-                     bool has_xfb_qualifiers);
-
-bool
-assign_varying_locations(struct gl_context *ctx,
-			 void *mem_ctx,
-			 struct gl_shader_program *prog,
-                         gl_linked_shader *producer,
-                         gl_linked_shader *consumer,
-                         unsigned num_tfeedback_decls,
-                         tfeedback_decl *tfeedback_decls,
-                         const uint64_t reserved_slots);
-
-uint64_t
-reserved_varying_slot(struct gl_linked_shader *stage,
-                      ir_variable_mode io_mode);
-
-bool
-check_against_output_limit(struct gl_context *ctx,
-                           struct gl_shader_program *prog,
-                           gl_linked_shader *producer,
-                           unsigned num_explicit_locations);
-
-bool
-check_against_input_limit(struct gl_context *ctx,
-                          struct gl_shader_program *prog,
-                          gl_linked_shader *consumer,
-                          unsigned num_explicit_locations);
-
 #endif /* GLSL_LINK_VARYINGS_H */

src/compiler/glsl/linker.h

@@ -35,8 +35,7 @@ link_invalidate_variable_locations(exec_list *ir);
 
 extern void
 link_assign_uniform_locations(struct gl_shader_program *prog,
-                              struct gl_context *ctx,
-                              unsigned int num_explicit_uniform_locs);
+                              struct gl_context *ctx);
 
 extern void
 link_set_uniform_initializers(struct gl_shader_program *prog,

src/compiler/glsl/lower_blend_equation_advanced.cpp

@@ -28,6 +28,7 @@
 #include "program/prog_instruction.h"
 #include "program/prog_statevars.h"
 #include "util/bitscan.h"
+#include "builtin_functions.h"
 
 using namespace ir_builder;
 

src/compiler/glsl/lower_int64.cpp

@@ -0,0 +1,403 @@
+/*
+ * Copyright © 2016 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+
+/**
+ * \file lower_int64.cpp
+ *
+ * Lower 64-bit operations to 32-bit operations.  Each 64-bit value is lowered
+ * to a uvec2.  For each operation that can be lowered, there is a function
+ * called __builtin_foo with the same number of parameters that takes uvec2
+ * sources and produces uvec2 results.  An operation like
+ *
+ *     uint64_t(x) * uint64_t(y)
+ *
+ * becomes
+ *
+ *     packUint2x32(__builtin_umul64(unpackUint2x32(x), unpackUint2x32(y)));
+ */
+
+#include "main/macros.h"
+#include "compiler/glsl_types.h"
+#include "ir.h"
+#include "ir_rvalue_visitor.h"
+#include "ir_builder.h"
+#include "ir_optimization.h"
+#include "util/hash_table.h"
+#include "builtin_functions.h"
+
+typedef ir_function_signature *(*function_generator)(void *mem_ctx,
+                                                     builtin_available_predicate avail);
+
+using namespace ir_builder;
+
+namespace lower_64bit {
+void expand_source(ir_factory &, ir_rvalue *val, ir_variable **expanded_src);
+
+ir_dereference_variable *compact_destination(ir_factory &,
+                                             const glsl_type *type,
+                                             ir_variable *result[4]);
+
+ir_rvalue *lower_op_to_function_call(ir_instruction *base_ir,
+                                     ir_expression *ir,
+                                     ir_function_signature *callee);
+};
+
+using namespace lower_64bit;
+
+namespace {
+
+class lower_64bit_visitor : public ir_rvalue_visitor {
+public:
+   lower_64bit_visitor(void *mem_ctx, exec_list *instructions, unsigned lower)
+      : progress(false), lower(lower), instructions(instructions),
+        function_list(), added_functions(&function_list, mem_ctx)
+   {
+      functions = _mesa_hash_table_create(mem_ctx,
+                                          _mesa_key_hash_string,
+                                          _mesa_key_string_equal);
+
+      foreach_in_list(ir_instruction, node, instructions) {
+         ir_function *const f = node->as_function();
+
+         if (f == NULL || strncmp(f->name, "__builtin_", 10) != 0)
+            continue;
+
+         add_function(f);
+      }
+   }
+
+   ~lower_64bit_visitor()
+   {
+      _mesa_hash_table_destroy(functions, NULL);
+   }
+
+   void handle_rvalue(ir_rvalue **rvalue);
+
+   void add_function(ir_function *f)
+   {
+      _mesa_hash_table_insert(functions, f->name, f);
+   }
+
+   ir_function *find_function(const char *name)
+   {
+      struct hash_entry *const entry =
+         _mesa_hash_table_search(functions, name);
+
+      return entry != NULL ? (ir_function *) entry->data : NULL;
+   }
+
+   bool progress;
+
+private:
+   unsigned lower; /** Bitfield of which operations to lower */
+
+   exec_list *instructions;
+
+   /** Hashtable containing all of the known functions in the IR */
+   struct hash_table *functions;
+
+public:
+   exec_list function_list;
+
+private:
+   ir_factory added_functions;
+
+   ir_rvalue *handle_op(ir_expression *ir, const char *function_name,
+                        function_generator generator);
+};
+
+} /* anonymous namespace */
+
+static bool
+is_integer_64(const glsl_type *t)
+{
+   return t->base_type == GLSL_TYPE_UINT64 || t->base_type == GLSL_TYPE_INT64;
+}
+
+/**
+ * Determine if a particular type of lowering should occur
+ */
+#define lowering(x) (this->lower & x)
+
+bool
+lower_64bit_integer_instructions(exec_list *instructions,
+                                 unsigned what_to_lower)
+{
+   if (instructions->is_empty())
+      return false;
+
+   ir_instruction *first_inst = (ir_instruction *) instructions->get_head_raw();
+   void *const mem_ctx = ralloc_parent(first_inst);
+   lower_64bit_visitor v(mem_ctx, instructions, what_to_lower);
+
+   visit_list_elements(&v, instructions);
+
+   if (v.progress && !v.function_list.is_empty()) {
+      /* Move all of the nodes from function_list to the head if the incoming
+       * instruction list.
+       */
+      exec_node *const after = &instructions->head_sentinel;
+      exec_node *const before = instructions->head_sentinel.next;
+      exec_node *const head = v.function_list.head_sentinel.next;
+      exec_node *const tail = v.function_list.tail_sentinel.prev;
+
+      before->next = head;
+      head->prev = before;
+
+      after->prev = tail;
+      tail->next = after;
+   }
+
+   return v.progress;
+}
+
+
+/**
+ * Expand individual 64-bit values to uvec2 values
+ *
+ * Each operation is in one of a few forms.
+ *
+ *     vector op vector
+ *     vector op scalar
+ *     scalar op vector
+ *     scalar op scalar
+ *
+ * In the 'vector op vector' case, the two vectors must have the same size.
+ * In a way, the 'scalar op scalar' form is special case of the 'vector op
+ * vector' form.
+ *
+ * This method generates a new set of uvec2 values for each element of a
+ * single operand.  If the operand is a scalar, the uvec2 is replicated
+ * multiple times.  A value like
+ *
+ *     u64vec3(a) + u64vec3(b)
+ *
+ * becomes
+ *
+ *     u64vec3 tmp0 = u64vec3(a) + u64vec3(b);
+ *     uvec2 tmp1 = unpackUint2x32(tmp0.x);
+ *     uvec2 tmp2 = unpackUint2x32(tmp0.y);
+ *     uvec2 tmp3 = unpackUint2x32(tmp0.z);
+ *
+ * and the returned operands array contains ir_variable pointers to
+ *
+ *     { tmp1, tmp2, tmp3, tmp1 }
+ */
+void
+lower_64bit::expand_source(ir_factory &body,
+                           ir_rvalue *val,
+                           ir_variable **expanded_src)
+{
+   assert(val->type->base_type == GLSL_TYPE_UINT64 ||
+          val->type->base_type == GLSL_TYPE_INT64);
+
+   ir_variable *const temp = body.make_temp(val->type, "tmp");
+
+   body.emit(assign(temp, val));
+
+   const ir_expression_operation unpack_opcode =
+      val->type->base_type == GLSL_TYPE_UINT64
+      ? ir_unop_unpack_uint_2x32 : ir_unop_unpack_int_2x32;
+
+   const glsl_type *const type =
+      val->type->base_type == GLSL_TYPE_UINT64
+      ? glsl_type::uvec2_type : glsl_type::ivec2_type;
+
+   unsigned i;
+   for (i = 0; i < val->type->vector_elements; i++) {
+      expanded_src[i] = body.make_temp(type, "expanded_64bit_source");
+
+      body.emit(assign(expanded_src[i],
+                       expr(unpack_opcode, swizzle(temp, i, 1))));
+   }
+
+   for (/* empty */; i < 4; i++)
+      expanded_src[i] = expanded_src[0];
+}
+
+/**
+ * Convert a series of uvec2 results into a single 64-bit integer vector
+ */
+ir_dereference_variable *
+lower_64bit::compact_destination(ir_factory &body,
+                                 const glsl_type *type,
+                                 ir_variable *result[4])
+{
+   const ir_expression_operation pack_opcode =
+      type->base_type == GLSL_TYPE_UINT64
+      ? ir_unop_pack_uint_2x32 : ir_unop_pack_int_2x32;
+
+   ir_variable *const compacted_result =
+      body.make_temp(type, "compacted_64bit_result");
+
+   for (unsigned i = 0; i < type->vector_elements; i++) {
+      body.emit(assign(compacted_result,
+                       expr(pack_opcode, result[i]),
+                       1U << i));
+   }
+
+   void *const mem_ctx = ralloc_parent(compacted_result);
+   return new(mem_ctx) ir_dereference_variable(compacted_result);
+}
+
+ir_rvalue *
+lower_64bit::lower_op_to_function_call(ir_instruction *base_ir,
+                                       ir_expression *ir,
+                                       ir_function_signature *callee)
+{
+   const unsigned num_operands = ir->get_num_operands();
+   ir_variable *src[4][4];
+   ir_variable *dst[4];
+   void *const mem_ctx = ralloc_parent(ir);
+   exec_list instructions;
+   unsigned source_components = 0;
+   const glsl_type *const result_type =
+      ir->type->base_type == GLSL_TYPE_UINT64
+      ? glsl_type::uvec2_type : glsl_type::ivec2_type;
+
+   ir_factory body(&instructions, mem_ctx);
+
+   for (unsigned i = 0; i < num_operands; i++) {
+      expand_source(body, ir->operands[i], src[i]);
+
+      if (ir->operands[i]->type->vector_elements > source_components)
+         source_components = ir->operands[i]->type->vector_elements;
+   }
+
+   for (unsigned i = 0; i < source_components; i++) {
+      dst[i] = body.make_temp(result_type, "expanded_64bit_result");
+
+      exec_list parameters;
+
+      for (unsigned j = 0; j < num_operands; j++)
+         parameters.push_tail(new(mem_ctx) ir_dereference_variable(src[j][i]));
+
+      ir_dereference_variable *const return_deref =
+         new(mem_ctx) ir_dereference_variable(dst[i]);
+
+      ir_call *const c = new(mem_ctx) ir_call(callee,
+                                              return_deref,
+                                              &parameters);
+
+      body.emit(c);
+   }
+
+   ir_rvalue *const rv = compact_destination(body, ir->type, dst);
+
+   /* Move all of the nodes from instructions between base_ir and the
+    * instruction before it.
+    */
+   exec_node *const after = base_ir;
+   exec_node *const before = after->prev;
+   exec_node *const head = instructions.head_sentinel.next;
+   exec_node *const tail = instructions.tail_sentinel.prev;
+
+   before->next = head;
+   head->prev = before;
+
+   after->prev = tail;
+   tail->next = after;
+
+   return rv;
+}
+
+ir_rvalue *
+lower_64bit_visitor::handle_op(ir_expression *ir,
+                               const char *function_name,
+                               function_generator generator)
+{
+   for (unsigned i = 0; i < ir->get_num_operands(); i++)
+      if (!is_integer_64(ir->operands[i]->type))
+         return ir;
+
+   /* Get a handle to the correct ir_function_signature for the core
+    * operation.
+    */
+   ir_function_signature *callee = NULL;
+   ir_function *f = find_function(function_name);
+
+   if (f != NULL) {
+      callee = (ir_function_signature *) f->signatures.get_head();
+      assert(callee != NULL && callee->ir_type == ir_type_function_signature);
+   } else {
+      f = new(base_ir) ir_function(function_name);
+      callee = generator(base_ir, NULL);
+
+      f->add_signature(callee);
+
+      add_function(f);
+   }
+
+   return lower_op_to_function_call(this->base_ir, ir, callee);
+}
+
+void
+lower_64bit_visitor::handle_rvalue(ir_rvalue **rvalue)
+{
+   if (*rvalue == NULL || (*rvalue)->ir_type != ir_type_expression)
+      return;
+
+   ir_expression *const ir = (*rvalue)->as_expression();
+   assert(ir != NULL);
+
+   switch (ir->operation) {
+   case ir_unop_sign:
+      if (lowering(SIGN64)) {
+         *rvalue = handle_op(ir, "__builtin_sign64", generate_ir::sign64);
+         this->progress = true;
+      }
+      break;
+
+   case ir_binop_div:
+      if (lowering(DIV64)) {
+         if (ir->type->base_type == GLSL_TYPE_UINT64) {
+            *rvalue = handle_op(ir, "__builtin_udiv64", generate_ir::udiv64);
+         } else {
+            *rvalue = handle_op(ir, "__builtin_idiv64", generate_ir::idiv64);
+         }
+         this->progress = true;
+      }
+      break;
+
+   case ir_binop_mod:
+      if (lowering(MOD64)) {
+         if (ir->type->base_type == GLSL_TYPE_UINT64) {
+            *rvalue = handle_op(ir, "__builtin_umod64", generate_ir::umod64);
+         } else {
+            *rvalue = handle_op(ir, "__builtin_imod64", generate_ir::imod64);
+         }
+         this->progress = true;
+      }
+      break;
+
+   case ir_binop_mul:
+      if (lowering(MUL64)) {
+         *rvalue = handle_op(ir, "__builtin_umul64", generate_ir::umul64);
+         this->progress = true;
+      }
+      break;
+
+   default:
+      break;
+   }
+}

src/compiler/glsl/lower_jumps.cpp

@@ -932,21 +932,14 @@ lower_continue:
              * break statement if necessary.
              */
             return_if->then_instructions.push_tail(new(ir) ir_loop_jump(ir_loop_jump::jump_break));
-         else {
-            /* Otherwise, ensure that the instructions that follow are only
-             * executed if the return flag is clear.  We can do that by moving
-             * those instructions into the else clause of the generated if
+         else
+            /* Otherwise, all we need to do is ensure that the
+             * instructions that follow are only executed if the
+             * return flag is clear.  We can do that by moving those
+             * instructions into the else clause of the generated if
              * statement.
              */
             move_outer_block_inside(ir, &return_if->else_instructions);
-
-            /* In case the loop is embeded inside an if add a new return to
-             * the return flag then branch and let a future pass tidy it up.
-             */
-            if (this->function.signature->return_type->is_void())
-               return_if->then_instructions.push_tail(new(ir) ir_return(NULL));
-         }
-
          ir->insert_after(return_if);
       }
 

src/compiler/glsl/lower_packed_varyings.cpp

@@ -351,6 +351,38 @@ lower_packed_varyings_visitor::bitwise_assign_pack(ir_rvalue *lhs,
             rhs = u2i(expr(ir_unop_unpack_double_2x32, rhs));
          }
          break;
+      case GLSL_TYPE_INT64:
+         assert(rhs->type->vector_elements <= 2);
+         if (rhs->type->vector_elements == 2) {
+            ir_variable *t = new(mem_ctx) ir_variable(lhs->type, "pack", ir_var_temporary);
+
+            assert(lhs->type->vector_elements == 4);
+            this->out_variables->push_tail(t);
+            this->out_instructions->push_tail(
+               assign(t, expr(ir_unop_unpack_int_2x32, swizzle_x(rhs->clone(mem_ctx, NULL))), 0x3));
+            this->out_instructions->push_tail(
+               assign(t,  expr(ir_unop_unpack_int_2x32, swizzle_y(rhs)), 0xc));
+            rhs = deref(t).val;
+         } else {
+            rhs = expr(ir_unop_unpack_int_2x32, rhs);
+         }
+         break;
+      case GLSL_TYPE_UINT64:
+         assert(rhs->type->vector_elements <= 2);
+         if (rhs->type->vector_elements == 2) {
+            ir_variable *t = new(mem_ctx) ir_variable(lhs->type, "pack", ir_var_temporary);
+
+            assert(lhs->type->vector_elements == 4);
+            this->out_variables->push_tail(t);
+            this->out_instructions->push_tail(
+                  assign(t, u2i(expr(ir_unop_unpack_uint_2x32, swizzle_x(rhs->clone(mem_ctx, NULL)))), 0x3));
+            this->out_instructions->push_tail(
+                  assign(t,  u2i(expr(ir_unop_unpack_uint_2x32, swizzle_y(rhs))), 0xc));
+            rhs = deref(t).val;
+         } else {
+            rhs = u2i(expr(ir_unop_unpack_uint_2x32, rhs));
+         }
+         break;
       default:
          assert(!"Unexpected type conversion while lowering varyings");
          break;
@@ -400,6 +432,36 @@ lower_packed_varyings_visitor::bitwise_assign_unpack(ir_rvalue *lhs,
             rhs = expr(ir_unop_pack_double_2x32, i2u(rhs));
          }
          break;
+      case GLSL_TYPE_INT64:
+         assert(lhs->type->vector_elements <= 2);
+         if (lhs->type->vector_elements == 2) {
+            ir_variable *t = new(mem_ctx) ir_variable(lhs->type, "unpack", ir_var_temporary);
+            assert(rhs->type->vector_elements == 4);
+            this->out_variables->push_tail(t);
+            this->out_instructions->push_tail(
+                  assign(t, expr(ir_unop_pack_int_2x32, swizzle_xy(rhs->clone(mem_ctx, NULL))), 0x1));
+            this->out_instructions->push_tail(
+                  assign(t, expr(ir_unop_pack_int_2x32, swizzle(rhs->clone(mem_ctx, NULL), SWIZZLE_ZWZW, 2)), 0x2));
+            rhs = deref(t).val;
+         } else {
+            rhs = expr(ir_unop_pack_int_2x32, rhs);
+         }
+         break;
+      case GLSL_TYPE_UINT64:
+         assert(lhs->type->vector_elements <= 2);
+         if (lhs->type->vector_elements == 2) {
+            ir_variable *t = new(mem_ctx) ir_variable(lhs->type, "unpack", ir_var_temporary);
+            assert(rhs->type->vector_elements == 4);
+            this->out_variables->push_tail(t);
+            this->out_instructions->push_tail(
+                  assign(t, expr(ir_unop_pack_uint_2x32, i2u(swizzle_xy(rhs->clone(mem_ctx, NULL)))), 0x1));
+            this->out_instructions->push_tail(
+                  assign(t, expr(ir_unop_pack_uint_2x32, i2u(swizzle(rhs->clone(mem_ctx, NULL), SWIZZLE_ZWZW, 2))), 0x2));
+            rhs = deref(t).val;
+         } else {
+            rhs = expr(ir_unop_pack_uint_2x32, i2u(rhs));
+         }
+         break;
       default:
          assert(!"Unexpected type conversion while lowering varyings");
          break;

src/compiler/glsl/lower_vertex_id.cpp

@@ -37,6 +37,7 @@
 #include "ir_builder.h"
 #include "linker.h"
 #include "program/prog_statevars.h"
+#include "builtin_functions.h"
 
 namespace {
 

src/compiler/glsl/opt_algebraic.cpp

@@ -472,6 +472,34 @@ ir_algebraic_visitor::handle_expression(ir_expression *ir)
       }
       break;
 
+      /* This macro CANNOT use the do { } while(true) mechanism because
+       * then the breaks apply to the loop instead of the switch!
+       */
+#define HANDLE_PACK_UNPACK_INVERSE(inverse_operation)                   \
+      {                                                                 \
+         ir_expression *const op = ir->operands[0]->as_expression();    \
+         if (op == NULL)                                                \
+            break;                                                      \
+         if (op->operation == (inverse_operation))                      \
+            return op->operands[0];                                     \
+         break;                                                         \
+      }
+
+   case ir_unop_unpack_uint_2x32:
+      HANDLE_PACK_UNPACK_INVERSE(ir_unop_pack_uint_2x32);
+   case ir_unop_pack_uint_2x32:
+      HANDLE_PACK_UNPACK_INVERSE(ir_unop_unpack_uint_2x32);
+   case ir_unop_unpack_int_2x32:
+      HANDLE_PACK_UNPACK_INVERSE(ir_unop_pack_int_2x32);
+   case ir_unop_pack_int_2x32:
+      HANDLE_PACK_UNPACK_INVERSE(ir_unop_unpack_int_2x32);
+   case ir_unop_unpack_double_2x32:
+      HANDLE_PACK_UNPACK_INVERSE(ir_unop_pack_double_2x32);
+   case ir_unop_pack_double_2x32:
+      HANDLE_PACK_UNPACK_INVERSE(ir_unop_unpack_double_2x32);
+
+#undef HANDLE_PACK_UNPACK_INVERSE
+
    case ir_binop_add:
       if (is_vec_zero(op_const[0]))
 	 return ir->operands[1];

src/compiler/glsl/standalone.cpp

@@ -41,6 +41,7 @@
 #include "linker.h"
 #include "glsl_parser_extras.h"
 #include "ir_builder_print_visitor.h"
+#include "builtin_functions.h"
 #include "opt_add_neg_to_sub.h"
 
 class dead_variable_visitor : public ir_hierarchical_visitor {

src/compiler/glsl/standalone_scaffolding.cpp

@@ -178,6 +178,7 @@ void initialize_context_to_defaults(struct gl_context *ctx, gl_api api)
    ctx->Extensions.ARB_fragment_layer_viewport = true;
    ctx->Extensions.ARB_gpu_shader5 = true;
    ctx->Extensions.ARB_gpu_shader_fp64 = true;
+   ctx->Extensions.ARB_gpu_shader_int64 = true;
    ctx->Extensions.ARB_sample_shading = true;
    ctx->Extensions.ARB_shader_bit_encoding = true;
    ctx->Extensions.ARB_shader_draw_parameters = true;

src/compiler/glsl/tests/cache_test.c

@@ -111,7 +111,7 @@ rmrf_local(const char *path)
    if (path == NULL || *path == '\0' || *path != '.')
       return -1;
 
-   return nftw(path, remove_entry, 64, FTW_DEPTH | FTW_PHYS);
+   return nftw(path, remove_entry, 64, FTW_DEPTH | FTW_PHYS | FTW_MOUNT);
 }
 
 #define CACHE_TEST_TMP "./cache-test-tmp"

src/compiler/glsl/tests/lower_int64_test.cpp

@@ -0,0 +1,440 @@
+/*
+ * Copyright © 2013 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ */
+#include <gtest/gtest.h>
+#include "main/compiler.h"
+#include "main/mtypes.h"
+#include "main/macros.h"
+#include "ir.h"
+#include "ir_builder.h"
+
+using namespace ir_builder;
+
+namespace lower_64bit {
+void expand_source(ir_factory &body,
+                   ir_rvalue *val,
+                   ir_variable **expanded_src);
+
+ir_dereference_variable *compact_destination(ir_factory &body,
+                                             const glsl_type *type,
+                                             ir_variable *result[4]);
+
+ir_rvalue *lower_op_to_function_call(ir_instruction *base_ir,
+                                     ir_expression *ir,
+                                     ir_function_signature *callee);
+};
+
+class expand_source : public ::testing::Test {
+public:
+   virtual void SetUp();
+   virtual void TearDown();
+
+   exec_list instructions;
+   ir_factory *body;
+   ir_variable *expanded_src[4];
+   void *mem_ctx;
+};
+
+void
+expand_source::SetUp()
+{
+   mem_ctx = ralloc_context(NULL);
+
+   memset(expanded_src, 0, sizeof(expanded_src));
+   instructions.make_empty();
+   body = new ir_factory(&instructions, mem_ctx);
+}
+
+void
+expand_source::TearDown()
+{
+   delete body;
+   body = NULL;
+
+   ralloc_free(mem_ctx);
+   mem_ctx = NULL;
+}
+
+static ir_dereference_variable *
+create_variable(void *mem_ctx, const glsl_type *type)
+{
+   ir_variable *var = new(mem_ctx) ir_variable(type,
+                                               "variable",
+                                               ir_var_temporary);
+
+   return new(mem_ctx) ir_dereference_variable(var);
+}
+
+static ir_expression *
+create_expression(void *mem_ctx, const glsl_type *type)
+{
+   return new(mem_ctx) ir_expression(ir_unop_neg,
+                                     create_variable(mem_ctx, type));
+}
+
+static void
+check_expanded_source(const glsl_type *type,
+                      ir_variable *expanded_src[4])
+{
+   const glsl_type *const expanded_type =
+      type->base_type == GLSL_TYPE_UINT64
+      ? glsl_type::uvec2_type :glsl_type::ivec2_type;
+
+   for (int i = 0; i < type->vector_elements; i++) {
+      EXPECT_EQ(expanded_type, expanded_src[i]->type);
+
+      /* All elements that are part of the vector must be unique. */
+      for (int j = i - 1; j >= 0; j--) {
+         EXPECT_NE(expanded_src[i], expanded_src[j])
+            << "    Element " << i << " is the same as element " << j;
+      }
+   }
+
+   /* All elements that are not part of the vector must be the same as element
+    * 0.  This is primarily for scalars (where every element is the same).
+    */
+   for (int i = type->vector_elements; i < 4; i++) {
+      EXPECT_EQ(expanded_src[0], expanded_src[i])
+         << "    Element " << i << " should be the same as element 0";
+   }
+}
+
+static void
+check_instructions(exec_list *instructions,
+                   const glsl_type *type,
+                   const ir_instruction *source)
+{
+   const glsl_type *const expanded_type =
+      type->base_type == GLSL_TYPE_UINT64
+      ? glsl_type::uvec2_type : glsl_type::ivec2_type;
+
+   const ir_expression_operation unpack_opcode =
+      type->base_type == GLSL_TYPE_UINT64
+      ? ir_unop_unpack_uint_2x32 : ir_unop_unpack_int_2x32;
+
+   ir_instruction *ir;
+
+   /* The instruction list should contain IR to represent:
+    *
+    *    type tmp1;
+    *    tmp1 = source;
+    *    uvec2 tmp2;
+    *    tmp2 = unpackUint2x32(tmp1.x);
+    *    uvec2 tmp3;
+    *    tmp3 = unpackUint2x32(tmp1.y);
+    *    uvec2 tmp4;
+    *    tmp4 = unpackUint2x32(tmp1.z);
+    *    uvec2 tmp5;
+    *    tmp5 = unpackUint2x32(tmp1.w);
+    */
+   ASSERT_FALSE(instructions->is_empty());
+   ir = (ir_instruction *) instructions->pop_head();
+   ir_variable *const tmp1 = ir->as_variable();
+   EXPECT_EQ(ir_type_variable, ir->ir_type);
+   EXPECT_EQ(type, tmp1->type) <<
+      "    Got " <<
+      tmp1->type->name <<
+      ", expected " <<
+      type->name;
+
+   ASSERT_FALSE(instructions->is_empty());
+   ir = (ir_instruction *) instructions->pop_head();
+   ir_assignment *const assign1 = ir->as_assignment();
+   EXPECT_EQ(ir_type_assignment, ir->ir_type);
+   ASSERT_NE((void *)0, assign1);
+   EXPECT_EQ(tmp1, assign1->lhs->variable_referenced());
+   EXPECT_EQ(source, assign1->rhs);
+
+   for (unsigned i = 0; i < type->vector_elements; i++) {
+      ASSERT_FALSE(instructions->is_empty());
+      ir = (ir_instruction *) instructions->pop_head();
+      ir_variable *const tmp2 = ir->as_variable();
+      EXPECT_EQ(ir_type_variable, ir->ir_type);
+      EXPECT_EQ(expanded_type, tmp2->type);
+
+      ASSERT_FALSE(instructions->is_empty());
+      ir = (ir_instruction *) instructions->pop_head();
+      ir_assignment *const assign2 = ir->as_assignment();
+      EXPECT_EQ(ir_type_assignment, ir->ir_type);
+      ASSERT_NE((void *)0, assign2);
+      EXPECT_EQ(tmp2, assign2->lhs->variable_referenced());
+      ir_expression *unpack = assign2->rhs->as_expression();
+      ASSERT_NE((void *)0, unpack);
+      EXPECT_EQ(unpack_opcode, unpack->operation);
+      EXPECT_EQ(tmp1, unpack->operands[0]->variable_referenced());
+   }
+
+   EXPECT_TRUE(instructions->is_empty());
+}
+
+TEST_F(expand_source, uint64_variable)
+{
+   const glsl_type *const type = glsl_type::uint64_t_type;
+   ir_dereference_variable *const deref = create_variable(mem_ctx, type);
+
+   lower_64bit::expand_source(*body, deref, expanded_src);
+
+   check_expanded_source(type, expanded_src);
+   check_instructions(&instructions, type, deref);
+}
+
+TEST_F(expand_source, u64vec2_variable)
+{
+   const glsl_type *const type = glsl_type::u64vec2_type;
+   ir_dereference_variable *const deref = create_variable(mem_ctx, type);
+
+   lower_64bit::expand_source(*body, deref, expanded_src);
+
+   check_expanded_source(type, expanded_src);
+   check_instructions(&instructions, type, deref);
+}
+
+TEST_F(expand_source, u64vec3_variable)
+{
+   const glsl_type *const type = glsl_type::u64vec3_type;
+
+   /* Generate an operand that is a scalar variable dereference. */
+   ir_variable *const var = new(mem_ctx) ir_variable(type,
+                                                     "variable",
+                                                     ir_var_temporary);
+
+   ir_dereference_variable *const deref =
+      new(mem_ctx) ir_dereference_variable(var);
+
+   lower_64bit::expand_source(*body, deref, expanded_src);
+
+   check_expanded_source(type, expanded_src);
+   check_instructions(&instructions, type, deref);
+}
+
+TEST_F(expand_source, u64vec4_variable)
+{
+   const glsl_type *const type = glsl_type::u64vec4_type;
+   ir_dereference_variable *const deref = create_variable(mem_ctx, type);
+
+   lower_64bit::expand_source(*body, deref, expanded_src);
+
+   check_expanded_source(type, expanded_src);
+   check_instructions(&instructions, type, deref);
+}
+
+TEST_F(expand_source, int64_variable)
+{
+   const glsl_type *const type = glsl_type::int64_t_type;
+   ir_dereference_variable *const deref = create_variable(mem_ctx, type);
+
+   lower_64bit::expand_source(*body, deref, expanded_src);
+
+   check_expanded_source(type, expanded_src);
+   check_instructions(&instructions, type, deref);
+}
+
+TEST_F(expand_source, i64vec2_variable)
+{
+   const glsl_type *const type = glsl_type::i64vec2_type;
+   ir_dereference_variable *const deref = create_variable(mem_ctx, type);
+
+   lower_64bit::expand_source(*body, deref, expanded_src);
+
+   check_expanded_source(type, expanded_src);
+   check_instructions(&instructions, type, deref);
+}
+
+TEST_F(expand_source, i64vec3_variable)
+{
+   const glsl_type *const type = glsl_type::i64vec3_type;
+   ir_dereference_variable *const deref = create_variable(mem_ctx, type);
+
+   lower_64bit::expand_source(*body, deref, expanded_src);
+
+   check_expanded_source(type, expanded_src);
+   check_instructions(&instructions, type, deref);
+}
+
+TEST_F(expand_source, i64vec4_variable)
+{
+   const glsl_type *const type = glsl_type::i64vec4_type;
+   ir_dereference_variable *const deref = create_variable(mem_ctx, type);
+
+   lower_64bit::expand_source(*body, deref, expanded_src);
+
+   check_expanded_source(type, expanded_src);
+   check_instructions(&instructions, type, deref);
+}
+
+TEST_F(expand_source, uint64_expression)
+{
+   const glsl_type *const type = glsl_type::uint64_t_type;
+   ir_expression *const expr = create_expression(mem_ctx, type);
+
+   lower_64bit::expand_source(*body, expr, expanded_src);
+
+   check_expanded_source(type, expanded_src);
+   check_instructions(&instructions, type, expr);
+}
+
+TEST_F(expand_source, u64vec2_expression)
+{
+   const glsl_type *const type = glsl_type::u64vec2_type;
+   ir_expression *const expr = create_expression(mem_ctx, type);
+
+   lower_64bit::expand_source(*body, expr, expanded_src);
+
+   check_expanded_source(type, expanded_src);
+   check_instructions(&instructions, type, expr);
+}
+
+TEST_F(expand_source, u64vec3_expression)
+{
+   const glsl_type *const type = glsl_type::u64vec3_type;
+   ir_expression *const expr = create_expression(mem_ctx, type);
+
+   lower_64bit::expand_source(*body, expr, expanded_src);
+
+   check_expanded_source(type, expanded_src);
+   check_instructions(&instructions, type, expr);
+}
+
+TEST_F(expand_source, u64vec4_expression)
+{
+   const glsl_type *const type = glsl_type::u64vec4_type;
+   ir_expression *const expr = create_expression(mem_ctx, type);
+
+   lower_64bit::expand_source(*body, expr, expanded_src);
+
+   check_expanded_source(type, expanded_src);
+   check_instructions(&instructions, type, expr);
+}
+
+TEST_F(expand_source, int64_expression)
+{
+   const glsl_type *const type = glsl_type::int64_t_type;
+   ir_expression *const expr = create_expression(mem_ctx, type);
+
+   lower_64bit::expand_source(*body, expr, expanded_src);
+
+   check_expanded_source(type, expanded_src);
+   check_instructions(&instructions, type, expr);
+}
+
+TEST_F(expand_source, i64vec2_expression)
+{
+   const glsl_type *const type = glsl_type::i64vec2_type;
+   ir_expression *const expr = create_expression(mem_ctx, type);
+
+   lower_64bit::expand_source(*body, expr, expanded_src);
+
+   check_expanded_source(type, expanded_src);
+   check_instructions(&instructions, type, expr);
+}
+
+TEST_F(expand_source, i64vec3_expression)
+{
+   const glsl_type *const type = glsl_type::i64vec3_type;
+   ir_expression *const expr = create_expression(mem_ctx, type);
+
+   lower_64bit::expand_source(*body, expr, expanded_src);
+
+   check_expanded_source(type, expanded_src);
+   check_instructions(&instructions, type, expr);
+}
+
+TEST_F(expand_source, i64vec4_expression)
+{
+   const glsl_type *const type = glsl_type::i64vec4_type;
+   ir_expression *const expr = create_expression(mem_ctx, type);
+
+   lower_64bit::expand_source(*body, expr, expanded_src);
+
+   check_expanded_source(type, expanded_src);
+   check_instructions(&instructions, type, expr);
+}
+
+class compact_destination : public ::testing::Test {
+public:
+   virtual void SetUp();
+   virtual void TearDown();
+
+   exec_list instructions;
+   ir_factory *body;
+   ir_variable *expanded_src[4];
+   void *mem_ctx;
+};
+
+void
+compact_destination::SetUp()
+{
+   mem_ctx = ralloc_context(NULL);
+
+   memset(expanded_src, 0, sizeof(expanded_src));
+   instructions.make_empty();
+   body = new ir_factory(&instructions, mem_ctx);
+}
+
+void
+compact_destination::TearDown()
+{
+   delete body;
+   body = NULL;
+
+   ralloc_free(mem_ctx);
+   mem_ctx = NULL;
+}
+
+TEST_F(compact_destination, uint64)
+{
+   const glsl_type *const type = glsl_type::uint64_t_type;
+
+   for (unsigned i = 0; i < type->vector_elements; i++) {
+      expanded_src[i] = new(mem_ctx) ir_variable(glsl_type::uvec2_type,
+                                                 "result",
+                                                 ir_var_temporary);
+   }
+
+   ir_dereference_variable *deref =
+      lower_64bit::compact_destination(*body,
+                                       type,
+                                       expanded_src);
+
+   ASSERT_EQ(ir_type_dereference_variable, deref->ir_type);
+   EXPECT_EQ(type, deref->var->type) <<
+      "    Got " <<
+      deref->var->type->name <<
+      ", expected " <<
+      type->name;
+
+   ir_instruction *ir;
+
+   ASSERT_FALSE(instructions.is_empty());
+   ir = (ir_instruction *) instructions.pop_head();
+   ir_variable *const var = ir->as_variable();
+   ASSERT_NE((void *)0, var);
+   EXPECT_EQ(deref->var, var);
+
+   for (unsigned i = 0; i < type->vector_elements; i++) {
+      ASSERT_FALSE(instructions.is_empty());
+      ir = (ir_instruction *) instructions.pop_head();
+      ir_assignment *const assign = ir->as_assignment();
+      ASSERT_NE((void *)0, assign);
+      EXPECT_EQ(deref->var, assign->lhs->variable_referenced());
+   }
+}

src/compiler/glsl/tests/lower_jumps/create_test_cases.py

@@ -236,13 +236,6 @@ def if_execute_flag(statements):
     check_sexp(statements)
     return [['if', ['var_ref', 'execute_flag'], statements, []]]
 
-def if_return_flag(then_statements, else_statements):
-    """Wrap statements in an if test with return_flag as the condition.
-    """
-    check_sexp(then_statements)
-    check_sexp(else_statements)
-    return [['if', ['var_ref', 'return_flag'], then_statements, else_statements]]
-
 def if_not_return_flag(statements):
     """Wrap statements in an if test so that they will only execute if
     return_flag is False.
@@ -459,10 +452,7 @@ def test_lower_pulled_out_jump():
                 loop(simple_if('b', simple_if('c', [], continue_()),
                                lowered_return_simple()) +
                      break_()) +
-
-                if_return_flag(assign_x('return_flag', const_bool(1)) +
-                               assign_x('execute_flag', const_bool(0)),
-                               assign_x('d', const_float(1))))
+                if_not_return_flag(assign_x('d', const_float(1))))
             ))
     create_test_case(doc_string, input_sexp, expected_sexp, 'lower_pulled_out_jump',
                      lower_main_return=True, pull_out_jumps=True)
@@ -593,14 +583,11 @@ def test_lower_return_void_at_end_of_loop():
             assign_x('b', const_float(2))
             ))
     expected_sexp = make_test_case('main', 'void', (
-            declare_execute_flag() +
             declare_return_flag() +
             loop(assign_x('a', const_float(1)) +
                  lowered_return_simple() +
                  break_()) +
-            if_return_flag(assign_x('return_flag', const_bool(1)) +
-                           assign_x('execute_flag', const_bool(0)),
-                           assign_x('b', const_float(2)))
+            if_not_return_flag(assign_x('b', const_float(2)))
             ))
     create_test_case(doc_string, input_sexp, input_sexp, 'return_void_at_end_of_loop_lower_nothing')
     create_test_case(doc_string, input_sexp, expected_sexp, 'return_void_at_end_of_loop_lower_return',

src/compiler/glsl/udivmod64.h

@@ -0,0 +1,206 @@
+ir_function_signature *
+udivmod64(void *mem_ctx, ir_factory &body)
+{
+   ir_function_signature *const sig =
+      new(mem_ctx) ir_function_signature(glsl_type::uvec4_type);
+   exec_list sig_parameters;
+
+   ir_variable *const r0001 = new(mem_ctx) ir_variable(glsl_type::uvec2_type, "numer", ir_var_function_in);
+   sig_parameters.push_tail(r0001);
+   ir_variable *const r0002 = new(mem_ctx) ir_variable(glsl_type::uvec2_type, "denom", ir_var_function_in);
+   sig_parameters.push_tail(r0002);
+   ir_variable *const r0003 = new(mem_ctx) ir_variable(glsl_type::int_type, "i", ir_var_auto);
+   body.emit(r0003);
+   ir_variable *const r0004 = new(mem_ctx) ir_variable(glsl_type::uint64_t_type, "n64", ir_var_auto);
+   body.emit(r0004);
+   ir_variable *const r0005 = new(mem_ctx) ir_variable(glsl_type::int_type, "log2_denom", ir_var_auto);
+   body.emit(r0005);
+   ir_variable *const r0006 = new(mem_ctx) ir_variable(glsl_type::uvec2_type, "quot", ir_var_auto);
+   body.emit(r0006);
+   body.emit(assign(r0006, ir_constant::zero(mem_ctx, glsl_type::uvec2_type), 0x03));
+
+   ir_expression *const r0007 = expr(ir_unop_find_msb, swizzle_y(r0002));
+   body.emit(assign(r0005, add(r0007, body.constant(int(32))), 0x01));
+
+   /* IF CONDITION */
+   ir_expression *const r0009 = equal(swizzle_y(r0002), body.constant(0u));
+   ir_expression *const r000A = nequal(swizzle_y(r0001), body.constant(0u));
+   ir_expression *const r000B = logic_and(r0009, r000A);
+   ir_if *f0008 = new(mem_ctx) ir_if(operand(r000B).val);
+   exec_list *const f0008_parent_instructions = body.instructions;
+
+      /* THEN INSTRUCTIONS */
+      body.instructions = &f0008->then_instructions;
+
+      ir_variable *const r000C = new(mem_ctx) ir_variable(glsl_type::int_type, "i", ir_var_auto);
+      body.emit(r000C);
+      ir_variable *const r000D = body.make_temp(glsl_type::int_type, "findMSB_retval");
+      body.emit(assign(r000D, expr(ir_unop_find_msb, swizzle_x(r0002)), 0x01));
+
+      body.emit(assign(r0005, r000D, 0x01));
+
+      body.emit(assign(r000C, sub(body.constant(int(31)), r000D), 0x01));
+
+      /* LOOP BEGIN */
+      ir_loop *f000E = new(mem_ctx) ir_loop();
+      exec_list *const f000E_parent_instructions = body.instructions;
+
+         body.instructions = &f000E->body_instructions;
+
+         /* IF CONDITION */
+         ir_expression *const r0010 = less(r000C, body.constant(int(1)));
+         ir_if *f000F = new(mem_ctx) ir_if(operand(r0010).val);
+         exec_list *const f000F_parent_instructions = body.instructions;
+
+            /* THEN INSTRUCTIONS */
+            body.instructions = &f000F->then_instructions;
+
+            body.emit(new(mem_ctx) ir_loop_jump(ir_loop_jump::jump_break));
+
+
+         body.instructions = f000F_parent_instructions;
+         body.emit(f000F);
+
+         /* END IF */
+
+         /* IF CONDITION */
+         ir_expression *const r0012 = lshift(swizzle_x(r0002), r000C);
+         ir_expression *const r0013 = lequal(r0012, swizzle_y(r0001));
+         ir_if *f0011 = new(mem_ctx) ir_if(operand(r0013).val);
+         exec_list *const f0011_parent_instructions = body.instructions;
+
+            /* THEN INSTRUCTIONS */
+            body.instructions = &f0011->then_instructions;
+
+            ir_expression *const r0014 = lshift(swizzle_x(r0002), r000C);
+            body.emit(assign(r0001, sub(swizzle_y(r0001), r0014), 0x02));
+
+            ir_expression *const r0015 = lshift(body.constant(1u), r000C);
+            body.emit(assign(r0006, bit_or(swizzle_y(r0006), r0015), 0x02));
+
+
+         body.instructions = f0011_parent_instructions;
+         body.emit(f0011);
+
+         /* END IF */
+
+         body.emit(assign(r000C, add(r000C, body.constant(int(-1))), 0x01));
+
+      /* LOOP END */
+
+      body.instructions = f000E_parent_instructions;
+      body.emit(f000E);
+
+      /* IF CONDITION */
+      ir_expression *const r0017 = lequal(swizzle_x(r0002), swizzle_y(r0001));
+      ir_if *f0016 = new(mem_ctx) ir_if(operand(r0017).val);
+      exec_list *const f0016_parent_instructions = body.instructions;
+
+         /* THEN INSTRUCTIONS */
+         body.instructions = &f0016->then_instructions;
+
+         body.emit(assign(r0001, sub(swizzle_y(r0001), swizzle_x(r0002)), 0x02));
+
+         body.emit(assign(r0006, bit_or(swizzle_y(r0006), body.constant(1u)), 0x02));
+
+
+      body.instructions = f0016_parent_instructions;
+      body.emit(f0016);
+
+      /* END IF */
+
+
+   body.instructions = f0008_parent_instructions;
+   body.emit(f0008);
+
+   /* END IF */
+
+   body.emit(assign(r0004, expr(ir_unop_pack_uint_2x32, r0001), 0x01));
+
+   ir_expression *const r0018 = sub(body.constant(int(63)), r0005);
+   body.emit(assign(r0003, expr(ir_binop_min, body.constant(int(31)), r0018), 0x01));
+
+   /* LOOP BEGIN */
+   ir_loop *f0019 = new(mem_ctx) ir_loop();
+   exec_list *const f0019_parent_instructions = body.instructions;
+
+      body.instructions = &f0019->body_instructions;
+
+      /* IF CONDITION */
+      ir_expression *const r001B = less(r0003, body.constant(int(1)));
+      ir_if *f001A = new(mem_ctx) ir_if(operand(r001B).val);
+      exec_list *const f001A_parent_instructions = body.instructions;
+
+         /* THEN INSTRUCTIONS */
+         body.instructions = &f001A->then_instructions;
+
+         body.emit(new(mem_ctx) ir_loop_jump(ir_loop_jump::jump_break));
+
+
+      body.instructions = f001A_parent_instructions;
+      body.emit(f001A);
+
+      /* END IF */
+
+      ir_variable *const r001C = body.make_temp(glsl_type::uint64_t_type, "assignment_tmp");
+      ir_expression *const r001D = expr(ir_unop_pack_uint_2x32, r0002);
+      body.emit(assign(r001C, lshift(r001D, r0003), 0x01));
+
+      /* IF CONDITION */
+      ir_expression *const r001F = lequal(r001C, r0004);
+      ir_if *f001E = new(mem_ctx) ir_if(operand(r001F).val);
+      exec_list *const f001E_parent_instructions = body.instructions;
+
+         /* THEN INSTRUCTIONS */
+         body.instructions = &f001E->then_instructions;
+
+         body.emit(assign(r0004, sub(r0004, r001C), 0x01));
+
+         ir_expression *const r0020 = lshift(body.constant(1u), r0003);
+         body.emit(assign(r0006, bit_or(swizzle_x(r0006), r0020), 0x01));
+
+
+      body.instructions = f001E_parent_instructions;
+      body.emit(f001E);
+
+      /* END IF */
+
+      body.emit(assign(r0003, add(r0003, body.constant(int(-1))), 0x01));
+
+   /* LOOP END */
+
+   body.instructions = f0019_parent_instructions;
+   body.emit(f0019);
+
+   ir_variable *const r0021 = body.make_temp(glsl_type::uint64_t_type, "packUint2x32_retval");
+   body.emit(assign(r0021, expr(ir_unop_pack_uint_2x32, r0002), 0x01));
+
+   /* IF CONDITION */
+   ir_expression *const r0023 = lequal(r0021, r0004);
+   ir_if *f0022 = new(mem_ctx) ir_if(operand(r0023).val);
+   exec_list *const f0022_parent_instructions = body.instructions;
+
+      /* THEN INSTRUCTIONS */
+      body.instructions = &f0022->then_instructions;
+
+      ir_expression *const r0024 = expr(ir_unop_pack_uint_2x32, r0002);
+      body.emit(assign(r0004, sub(r0004, r0024), 0x01));
+
+      body.emit(assign(r0006, bit_or(swizzle_x(r0006), body.constant(1u)), 0x01));
+
+
+   body.instructions = f0022_parent_instructions;
+   body.emit(f0022);
+
+   /* END IF */
+
+   ir_variable *const r0025 = body.make_temp(glsl_type::uvec4_type, "vec_ctor");
+   body.emit(assign(r0025, r0006, 0x03));
+
+   body.emit(assign(r0025, expr(ir_unop_unpack_uint_2x32, r0004), 0x0c));
+
+   body.emit(ret(r0025));
+
+   sig->replace_parameters(&sig_parameters);
+   return sig;
+}

src/compiler/glsl_types.cpp

@@ -354,6 +354,10 @@ const glsl_type *glsl_type::get_base_type() const
       return double_type;
    case GLSL_TYPE_BOOL:
       return bool_type;
+   case GLSL_TYPE_UINT64:
+      return uint64_t_type;
+   case GLSL_TYPE_INT64:
+      return int64_t_type;
    default:
       return error_type;
    }
@@ -380,6 +384,10 @@ const glsl_type *glsl_type::get_scalar_type() const
       return double_type;
    case GLSL_TYPE_BOOL:
       return bool_type;
+   case GLSL_TYPE_UINT64:
+      return uint64_t_type;
+   case GLSL_TYPE_INT64:
+      return int64_t_type;
    default:
       /* Handle everything else */
       return type;
@@ -519,6 +527,31 @@ glsl_type::bvec(unsigned components)
 }
 
 
+const glsl_type *
+glsl_type::i64vec(unsigned components)
+{
+   if (components == 0 || components > 4)
+      return error_type;
+
+   static const glsl_type *const ts[] = {
+      int64_t_type, i64vec2_type, i64vec3_type, i64vec4_type
+   };
+   return ts[components - 1];
+}
+
+
+const glsl_type *
+glsl_type::u64vec(unsigned components)
+{
+   if (components == 0 || components > 4)
+      return error_type;
+
+   static const glsl_type *const ts[] = {
+      uint64_t_type, u64vec2_type, u64vec3_type, u64vec4_type
+   };
+   return ts[components - 1];
+}
+
 const glsl_type *
 glsl_type::get_instance(unsigned base_type, unsigned rows, unsigned columns)
 {
@@ -542,6 +575,10 @@ glsl_type::get_instance(unsigned base_type, unsigned rows, unsigned columns)
          return dvec(rows);
       case GLSL_TYPE_BOOL:
          return bvec(rows);
+      case GLSL_TYPE_UINT64:
+         return u64vec(rows);
+      case GLSL_TYPE_INT64:
+         return i64vec(rows);
       default:
          return error_type;
       }
@@ -1239,6 +1276,8 @@ glsl_type::component_slots() const
       return this->components();
 
    case GLSL_TYPE_DOUBLE:
+   case GLSL_TYPE_UINT64:
+   case GLSL_TYPE_INT64:
       return 2 * this->components();
 
    case GLSL_TYPE_STRUCT:
@@ -1321,6 +1360,8 @@ glsl_type::uniform_locations() const
    case GLSL_TYPE_INT:
    case GLSL_TYPE_FLOAT:
    case GLSL_TYPE_DOUBLE:
+   case GLSL_TYPE_UINT64:
+   case GLSL_TYPE_INT64:
    case GLSL_TYPE_BOOL:
    case GLSL_TYPE_SAMPLER:
    case GLSL_TYPE_IMAGE:
@@ -1350,6 +1391,8 @@ glsl_type::varying_count() const
    case GLSL_TYPE_FLOAT:
    case GLSL_TYPE_DOUBLE:
    case GLSL_TYPE_BOOL:
+   case GLSL_TYPE_UINT64:
+   case GLSL_TYPE_INT64:
       return 1;
 
    case GLSL_TYPE_STRUCT:
@@ -1923,6 +1966,8 @@ glsl_type::count_attribute_slots(bool is_vertex_input) const
    case GLSL_TYPE_BOOL:
       return this->matrix_columns;
    case GLSL_TYPE_DOUBLE:
+   case GLSL_TYPE_UINT64:
+   case GLSL_TYPE_INT64:
       if (this->vector_elements > 2 && !is_vertex_input)
          return this->matrix_columns * 2;
       else

src/compiler/glsl_types.h

@@ -54,6 +54,8 @@ enum glsl_base_type {
    GLSL_TYPE_INT,
    GLSL_TYPE_FLOAT,
    GLSL_TYPE_DOUBLE,
+   GLSL_TYPE_UINT64,
+   GLSL_TYPE_INT64,
    GLSL_TYPE_BOOL,
    GLSL_TYPE_SAMPLER,
    GLSL_TYPE_IMAGE,
@@ -69,7 +71,9 @@ enum glsl_base_type {
 
 static inline bool glsl_base_type_is_64bit(enum glsl_base_type type)
 {
-   return type == GLSL_TYPE_DOUBLE;
+   return type == GLSL_TYPE_DOUBLE ||
+          type == GLSL_TYPE_UINT64 ||
+          type == GLSL_TYPE_INT64;
 }
 
 enum glsl_sampler_dim {
@@ -225,6 +229,8 @@ struct glsl_type {
    static const glsl_type *ivec(unsigned components);
    static const glsl_type *uvec(unsigned components);
    static const glsl_type *bvec(unsigned components);
+   static const glsl_type *i64vec(unsigned components);
+   static const glsl_type *u64vec(unsigned components);
    /**@}*/
 
    /**
@@ -462,7 +468,7 @@ struct glsl_type {
     */
    bool is_numeric() const
    {
-      return (base_type >= GLSL_TYPE_UINT) && (base_type <= GLSL_TYPE_DOUBLE);
+      return (base_type >= GLSL_TYPE_UINT) && (base_type <= GLSL_TYPE_INT64);
    }
 
    /**
@@ -474,8 +480,17 @@ struct glsl_type {
    }
 
    /**
-    * Query whether or not type is an integral type, or for struct, interface
-    * and array types, contains an integral type.
+    * Query whether or not a type is a 32-bit or 64-bit integer
+    */
+   bool is_integer_32_64() const
+   {
+      return (base_type == GLSL_TYPE_UINT) || (base_type == GLSL_TYPE_INT) ||
+             (base_type == GLSL_TYPE_UINT64) || (base_type == GLSL_TYPE_INT64);
+   }
+
+   /**
+    * Query whether or not type is an integral type, or for struct and array
+    * types, contains an integral type.
     */
    bool contains_integer() const;
 

src/compiler/nir/nir.c

@@ -846,6 +846,8 @@ nir_deref_get_const_initializer_load(nir_shader *shader, nir_deref_var *deref)
    case GLSL_TYPE_INT:
    case GLSL_TYPE_UINT:
    case GLSL_TYPE_DOUBLE:
+   case GLSL_TYPE_UINT64:
+   case GLSL_TYPE_INT64:
    case GLSL_TYPE_BOOL:
       load->value = constant->values[matrix_col];
       break;