Update version to 17.1.0-rc3

Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
travis: bump MAKEFLAGS to -j4
2017-04-30 09:46:32 +01:00 · 2017-04-30 09:46:32 +01:00 · 2017-04-30 09:46:32 +01:00 · 2017-04-30 09:46:32 +01:00 · 2017-04-30 09:46:32 +01:00 · 2017-04-30 09:46:32 +01:00
46 changed files with 715 additions and 203 deletions
--- a/.travis.yml
+++ b/.travis.yml
@@ -1,24 +1,11 @@
 language: c

-sudo: required
+sudo: false
 dist: trusty

 cache:
-  directories:
-    - $HOME/.ccache
-
-addons:
-  apt:
-    packages:
-      - libdrm-dev
-      - x11proto-xf86vidmode-dev
-      - libexpat1-dev
-      - libxcb-dri2-0-dev
-      - libx11-xcb-dev
-      # LLVM packaging is broken and misses these dependencies
-      - libedit-dev
-      - libelf-dev
-      - scons
+  apt: true
+  ccache: true

 env:
  global:
@@ -32,17 +19,260 @@ env:
    - XCBPROTO_VERSION=xcb-proto-1.11
    - LIBXCB_VERSION=libxcb-1.11
    - LIBXSHMFENCE_VERSION=libxshmfence-1.2
-    - LLVM_VERSION=3.9
-    - LLVM_PACKAGE="llvm-${LLVM_VERSION} llvm-${LLVM_VERSION}-dev"
-    - LLVM_CONFIG="llvm-config-${LLVM_VERSION}"
+    - LIBTXC_DXTN_VERSION=libtxc_dxtn-1.0.1
+    - LIBVDPAU_VERSION=libvdpau-1.1
+    - LIBVA_VERSION=libva-1.6.2
+    - LIBWAYLAND_VERSION=wayland-1.11.1
    - PKG_CONFIG_PATH=$HOME/prefix/lib/pkgconfig
-    - MAKEFLAGS=-j2
-  matrix:
-    - BUILD=make
-    - BUILD=scons
+    - 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-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-platforms=x11,wayland"
+        - DRI_DRIVERS=""
+        # 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

 install:
-  - export PATH="/usr/lib/ccache:$PATH"
  - pip install --user mako

  # Since libdrm gets updated in configure.ac regularly, try to pick up the
@@ -90,25 +320,64 @@ install:
  - tar -jxvf $LIBXSHMFENCE_VERSION.tar.bz2
  - (cd $LIBXSHMFENCE_VERSION && ./configure --prefix=$HOME/prefix && make install)

-  # Install LLVM directly via apt-get (not Travis-CI's apt addon)
-  # See https://github.com/travis-ci/apt-source-whitelist/pull/205#issuecomment-216054237
+  # 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 -nv -O - http://llvm.org/apt/llvm-snapshot.gpg.key | sudo apt-key add -
-  - sudo apt-add-repository -y 'deb http://llvm.org/apt/trusty llvm-toolchain-trusty-3.9 main'
-  - sudo apt-add-repository -y 'deb http://llvm.org/apt/trusty llvm-toolchain-trusty main'
-  - sudo apt-get update -qq
-  - sudo apt-get install -qq -y $LLVM_PACKAGE
+  - wget http://people.freedesktop.org/~aplattner/vdpau/$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
-        --with-platforms=x11,drm
-        --with-dri-drivers=i915,i965,radeon,r200,swrast,nouveau
-        --with-gallium-drivers=i915,nouveau,r300,r600,radeonsi,freedreno,svga,swrast,vc4,virgl,etnaviv,imx
-        --with-vulkan-drivers=radeon
+        $DRI_LOADERS
+        --with-dri-drivers=$DRI_DRIVERS
+        $GALLIUM_ST
+        --with-gallium-drivers=$GALLIUM_DRIVERS
+        --with-vulkan-drivers=$VULKAN_DRIVERS
        --disable-llvm-shared-libs
-        ;
-      make && make check;
-    elif test x$BUILD = xscons; then
-      scons llvm=1 && scons llvm=1 check;
+        &&
+      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 $SCONS_CHECK_COMMAND;
    fi
--- a/2
+++ b/2
@@ -1 +1 @@
-17.1.0-devel
+17.1.0-rc3
--- a/configure.ac
+++ b/configure.ac
@@ -724,7 +724,7 @@ dnl Arch/platform-specific settings
 dnl
 AC_ARG_ENABLE([asm],
    [AS_HELP_STRING([--disable-asm],
-        [disable assembly usage @<:@default=enabled on supported plaforms@:>@])],
+        [disable assembly usage @<:@default=enabled on supported platforms@:>@])],
    [enable_asm="$enableval"],
    [enable_asm=yes]
 )
@@ -1367,7 +1367,7 @@ if test "x$enable_libglvnd" = xyes ; then
    esac

    PKG_CHECK_MODULES([GLVND], libglvnd >= 0.2.0)
-    PKG_CHECK_VAR(LIBGLVND_DATADIR, libglvnd, datadir)
+    LIBGLVND_DATADIR=`$PKG_CONFIG --variable=datadir libglvnd`
    AC_SUBST([LIBGLVND_DATADIR])

    DEFINES="${DEFINES} -DUSE_LIBGLVND=1"
@@ -2146,12 +2146,11 @@ dnl DEPRECATED: EGL Platforms configuration
 dnl
 AC_ARG_WITH([egl-platforms],
    [AS_HELP_STRING([--with-egl-platforms@<:@=DIRS...@:>@],
-        [DEPRECATED: use --with-plaforms instead@<:@default=auto@:>@])],
+        [DEPRECATED: use --with-platforms instead@<:@default=auto@:>@])],
    [with_egl_platforms="$withval"],
    [with_egl_platforms=auto])

 if test "x$with_egl_platforms" = xauto; then
-    AC_MSG_WARN([--with-egl-platforms is deprecated. Use --with-plaforms instead.])
    if test "x$enable_egl" = xyes; then
        if test "x$enable_gbm" = xyes; then
           with_egl_platforms="x11,drm"
@@ -2161,6 +2160,8 @@ if test "x$with_egl_platforms" = xauto; then
    else
        with_egl_platforms=""
    fi
+else
+    AC_MSG_WARN([--with-egl-platforms is deprecated. Use --with-platforms instead.])
 fi

 dnl
@@ -2465,7 +2466,9 @@ if test -n "$with_gallium_drivers"; then
        xvirgl)
            HAVE_GALLIUM_VIRGL=yes
            require_libdrm "virgl"
-            require_basic_egl "virgl"
+            if test "x$enable_egl" = xyes; then
+                require_basic_egl "virgl"
+            fi
            ;;
        *)
            AC_MSG_ERROR([Unknown Gallium driver: $driver])
--- a/src/gallium/auxiliary/pipe-loader/pipe_loader_drm.c
+++ b/src/gallium/auxiliary/pipe-loader/pipe_loader_drm.c
@@ -64,13 +64,13 @@ static const struct pipe_loader_ops pipe_loader_drm_ops;

 #ifdef GALLIUM_STATIC_TARGETS
 static const struct drm_conf_ret throttle_ret = {
-   DRM_CONF_INT,
-   {2},
+   .type = DRM_CONF_INT,
+   .val.val_int = 2,
 };

 static const struct drm_conf_ret share_fd_ret = {
-   DRM_CONF_BOOL,
-   {true},
+   .type = DRM_CONF_BOOL,
+   .val.val_bool = true,
 };

 static inline const struct drm_conf_ret *
--- a/src/gallium/drivers/freedreno/freedreno_context.c
+++ b/src/gallium/drivers/freedreno/freedreno_context.c
@@ -112,7 +112,7 @@ fd_context_destroy(struct pipe_context *pctx)

 	DBG("");

-	if (ctx->screen->reorder)
+	if (ctx->screen->reorder && util_queue_is_initialized(&ctx->flush_queue))
 		util_queue_destroy(&ctx->flush_queue);

 	fd_batch_reference(&ctx->batch, NULL);  /* unref current batch */
--- a/src/gallium/drivers/radeonsi/si_descriptors.c
+++ b/src/gallium/drivers/radeonsi/si_descriptors.c
@@ -765,7 +765,7 @@ static void si_set_shader_image(struct si_context *ctx,
 		static const unsigned char swizzle[4] = { 0, 1, 2, 3 };
 		struct r600_texture *tex = (struct r600_texture *)res;
 		unsigned level = view->u.tex.level;
-		unsigned width, height, depth;
+		unsigned width, height, depth, hw_level;
 		bool uses_dcc = vi_dcc_enabled(tex, level);

 		assert(!tex->is_depth);
@@ -794,20 +794,31 @@ static void si_set_shader_image(struct si_context *ctx,
 		    p_atomic_read(&tex->framebuffers_bound))
 			ctx->need_check_render_feedback = true;

-		/* Always force the base level to the selected level.
-		 *
-		 * This is required for 3D textures, where otherwise
-		 * selecting a single slice for non-layered bindings
-		 * fails. It doesn't hurt the other targets.
-		 */
-		width = u_minify(res->b.b.width0, level);
-		height = u_minify(res->b.b.height0, level);
-		depth = u_minify(res->b.b.depth0, level);
+		if (ctx->b.chip_class >= GFX9) {
+			/* Always set the base address. The swizzle modes don't
+			 * allow setting mipmap level offsets as the base.
+			 */
+			width = res->b.b.width0;
+			height = res->b.b.height0;
+			depth = res->b.b.depth0;
+			hw_level = level;
+		} else {
+			/* Always force the base level to the selected level.
+			 *
+			 * This is required for 3D textures, where otherwise
+			 * selecting a single slice for non-layered bindings
+			 * fails. It doesn't hurt the other targets.
+			 */
+			width = u_minify(res->b.b.width0, level);
+			height = u_minify(res->b.b.height0, level);
+			depth = u_minify(res->b.b.depth0, level);
+			hw_level = 0;
+		}

 		si_make_texture_descriptor(screen, tex,
 					   false, res->b.b.target,
 					   view->format, swizzle,
-					   0, 0,
+					   hw_level, hw_level,
 					   view->u.tex.first_layer,
 					   view->u.tex.last_layer,
 					   width, height, depth,
--- a/src/gallium/drivers/radeonsi/si_shader.c
+++ b/src/gallium/drivers/radeonsi/si_shader.c
@@ -3400,7 +3400,7 @@ image_fetch_rsrc(
 static LLVMValueRef image_fetch_coords(
 		struct lp_build_tgsi_context *bld_base,
 		const struct tgsi_full_instruction *inst,
-		unsigned src)
+		unsigned src, LLVMValueRef desc)
 {
 	struct si_shader_context *ctx = si_shader_context(bld_base);
 	struct gallivm_state *gallivm = &ctx->gallivm;
@@ -3417,14 +3417,30 @@ static LLVMValueRef image_fetch_coords(
 		coords[chan] = tmp;
 	}

-	/* 1D textures are allocated and used as 2D on GFX9. */
 	if (ctx->screen->b.chip_class >= GFX9) {
+		/* 1D textures are allocated and used as 2D on GFX9. */
 		if (target == TGSI_TEXTURE_1D) {
 			coords[1] = ctx->i32_0;
 			num_coords++;
 		} else if (target == TGSI_TEXTURE_1D_ARRAY) {
 			coords[2] = coords[1];
 			coords[1] = ctx->i32_0;
+			num_coords++;
+		} else if (target == TGSI_TEXTURE_2D) {
+			/* The hw can't bind a slice of a 3D image as a 2D
+			 * image, because it ignores BASE_ARRAY if the target
+			 * is 3D. The workaround is to read BASE_ARRAY and set
+			 * it as the 3rd address operand for all 2D images.
+			 */
+			LLVMValueRef first_layer, const5, mask;
+
+			const5 = LLVMConstInt(ctx->i32, 5, 0);
+			mask = LLVMConstInt(ctx->i32, S_008F24_BASE_ARRAY(~0), 0);
+			first_layer = LLVMBuildExtractElement(builder, desc, const5, "");
+			first_layer = LLVMBuildAnd(builder, first_layer, mask, "");
+
+			coords[2] = first_layer;
+			num_coords++;
 		}
 	}

@@ -3539,7 +3555,7 @@ static void load_fetch_args(
 		LLVMValueRef coords;

 		image_fetch_rsrc(bld_base, &inst->Src[0], false, target, &rsrc);
-		coords = image_fetch_coords(bld_base, inst, 1);
+		coords = image_fetch_coords(bld_base, inst, 1, rsrc);

 		if (target == TGSI_TEXTURE_BUFFER) {
 			buffer_append_args(ctx, emit_data, rsrc, coords,
@@ -3814,16 +3830,15 @@ static void store_fetch_args(
 		 */
 		bool force_glc = ctx->screen->b.chip_class == SI;

-		coords = image_fetch_coords(bld_base, inst, 0);
+		image_fetch_rsrc(bld_base, &memory, true, target, &rsrc);
+		coords = image_fetch_coords(bld_base, inst, 0, rsrc);

 		if (target == TGSI_TEXTURE_BUFFER) {
-			image_fetch_rsrc(bld_base, &memory, true, target, &rsrc);
 			buffer_append_args(ctx, emit_data, rsrc, coords,
 					   ctx->i32_0, false, force_glc);
 		} else {
 			emit_data->args[1] = coords;
-			image_fetch_rsrc(bld_base, &memory, true, target,
-					 &emit_data->args[2]);
+			emit_data->args[2] = rsrc;
 			emit_data->args[3] = LLVMConstInt(ctx->i32, 15, 0); /* dmask */
 			emit_data->arg_count = 4;

@@ -4027,7 +4042,7 @@ static void atomic_fetch_args(
 		LLVMValueRef coords;

 		image_fetch_rsrc(bld_base, &inst->Src[0], true, target, &rsrc);
-		coords = image_fetch_coords(bld_base, inst, 1);
+		coords = image_fetch_coords(bld_base, inst, 1, rsrc);

 		if (target == TGSI_TEXTURE_BUFFER) {
 			buffer_append_args(ctx, emit_data, rsrc, coords,
--- a/src/gallium/drivers/radeonsi/si_state.c
+++ b/src/gallium/drivers/radeonsi/si_state.c
@@ -2976,7 +2976,40 @@ si_make_buffer_descriptor(struct si_screen *screen, struct r600_resource *buf,
 	num_records = size / stride;
 	num_records = MIN2(num_records, (buf->b.b.width0 - offset) / stride);

-	if (screen->b.chip_class == VI)
+	/* The NUM_RECORDS field has a different meaning depending on the chip,
+	 * instruction type, STRIDE, and SWIZZLE_ENABLE.
+	 *
+	 * SI-CIK:
+	 * - If STRIDE == 0, it's in byte units.
+	 * - If STRIDE != 0, it's in units of STRIDE, used with inst.IDXEN.
+	 *
+	 * VI:
+	 * - For SMEM and STRIDE == 0, it's in byte units.
+	 * - For SMEM and STRIDE != 0, it's in units of STRIDE.
+	 * - For VMEM and STRIDE == 0 or SWIZZLE_ENABLE == 0, it's in byte units.
+	 * - For VMEM and STRIDE != 0 and SWIZZLE_ENABLE == 1, it's in units of STRIDE.
+	 * NOTE: There is incompatibility between VMEM and SMEM opcodes due to SWIZZLE_-
+	 *       ENABLE. The workaround is to set STRIDE = 0 if SWIZZLE_ENABLE == 0 when
+	 *       using SMEM. This can be done in the shader by clearing STRIDE with s_and.
+	 *       That way the same descriptor can be used by both SMEM and VMEM.
+	 *
+	 * GFX9:
+	 * - For SMEM and STRIDE == 0, it's in byte units.
+	 * - For SMEM and STRIDE != 0, it's in units of STRIDE.
+	 * - For VMEM and inst.IDXEN == 0 or STRIDE == 0, it's in byte units.
+	 * - For VMEM and inst.IDXEN == 1 and STRIDE != 0, it's in units of STRIDE.
+	 */
+	if (screen->b.chip_class >= GFX9)
+		/* When vindex == 0, LLVM sets IDXEN = 0, thus changing units
+		 * from STRIDE to bytes. This works around it by setting
+		 * NUM_RECORDS to at least the size of one element, so that
+		 * the first element is readable when IDXEN == 0.
+		 *
+		 * TODO: Fix this in LLVM, but do we need a new intrinsic where
+		 *       IDXEN is enforced?
+		 */
+		num_records = num_records ? MAX2(num_records, stride) : 0;
+	else if (screen->b.chip_class == VI)
 		num_records *= stride;

 	state[4] = 0;
@@ -3156,7 +3189,8 @@ si_make_texture_descriptor(struct si_screen *screen,
 	if (!sampler &&
 	    (res->target == PIPE_TEXTURE_CUBE ||
 	     res->target == PIPE_TEXTURE_CUBE_ARRAY ||
-	     res->target == PIPE_TEXTURE_3D)) {
+	     (screen->b.chip_class <= VI &&
+	      res->target == PIPE_TEXTURE_3D))) {
 		/* For the purpose of shader images, treat cube maps and 3D
 		 * textures as 2D arrays. For 3D textures, the address
 		 * calculations for mipmaps are different, so we rely on the
--- a/src/gallium/drivers/radeonsi/si_state_draw.c
+++ b/src/gallium/drivers/radeonsi/si_state_draw.c
@@ -291,7 +291,8 @@ si_get_init_multi_vgt_param(struct si_screen *sscreen,
 		/* Needed for 028B6C_DISTRIBUTION_MODE != 0 */
 		if (sscreen->has_distributed_tess) {
 			if (key->u.uses_gs) {
-				partial_es_wave = true;
+				if (sscreen->b.chip_class <= VI)
+					partial_es_wave = true;

 				/* GPU hang workaround. */
 				if (sscreen->b.family == CHIP_TONGA ||
@@ -371,7 +372,7 @@ si_get_init_multi_vgt_param(struct si_screen *sscreen,
 	}

 	/* If SWITCH_ON_EOI is set, PARTIAL_ES_WAVE must be set too. */
-	if (ia_switch_on_eoi)
+	if (sscreen->b.chip_class <= VI && ia_switch_on_eoi)
 		partial_es_wave = true;

 	return S_028AA8_SWITCH_ON_EOP(ia_switch_on_eop) |
@@ -379,7 +380,8 @@ si_get_init_multi_vgt_param(struct si_screen *sscreen,
 		S_028AA8_PARTIAL_VS_WAVE_ON(partial_vs_wave) |
 		S_028AA8_PARTIAL_ES_WAVE_ON(partial_es_wave) |
 		S_028AA8_WD_SWITCH_ON_EOP(sscreen->b.chip_class >= CIK ? wd_switch_on_eop : 0) |
-		S_028AA8_MAX_PRIMGRP_IN_WAVE(sscreen->b.chip_class >= VI ?
+		/* The following field was moved to VGT_SHADER_STAGES_EN in GFX9. */
+		S_028AA8_MAX_PRIMGRP_IN_WAVE(sscreen->b.chip_class == VI ?
 					     max_primgroup_in_wave : 0) |
 		S_030960_EN_INST_OPT_BASIC(sscreen->b.chip_class >= GFX9) |
 		S_030960_EN_INST_OPT_ADV(sscreen->b.chip_class >= GFX9);
--- a/src/gallium/drivers/radeonsi/si_state_shaders.c
+++ b/src/gallium/drivers/radeonsi/si_state_shaders.c
@@ -2115,7 +2115,10 @@ static bool si_update_gs_ring_buffers(struct si_context *sctx)
 	unsigned num_se = sctx->screen->b.info.max_se;
 	unsigned wave_size = 64;
 	unsigned max_gs_waves = 32 * num_se; /* max 32 per SE on GCN */
-	unsigned gs_vertex_reuse = 16 * num_se; /* GS_VERTEX_REUSE register (per SE) */
+	/* On SI-CI, the value comes from VGT_GS_VERTEX_REUSE = 16.
+	 * On VI+, the value comes from VGT_VERTEX_REUSE_BLOCK_CNTL = 30 (+2).
+	 */
+	unsigned gs_vertex_reuse = (sctx->b.chip_class >= VI ? 32 : 16) * num_se;
 	unsigned alignment = 256 * num_se;
 	/* The maximum size is 63.999 MB per SE. */
 	unsigned max_size = ((unsigned)(63.999 * 1024 * 1024) & ~255) * num_se;
@@ -2542,6 +2545,9 @@ static void si_update_vgt_shader_config(struct si_context *sctx)
 			          S_028B54_VS_EN(V_028B54_VS_STAGE_COPY_SHADER);
 		}

+		if (sctx->b.chip_class >= GFX9)
+			stages |= S_028B54_MAX_PRIMGRP_IN_WAVE(2);
+
 		si_pm4_set_reg(*pm4, R_028B54_VGT_SHADER_STAGES_EN, stages);
 	}
 	si_pm4_bind_state(sctx, vgt_shader_config, *pm4);
--- a/src/gallium/state_trackers/clover/llvm/codegen.hpp
+++ b/src/gallium/state_trackers/clover/llvm/codegen.hpp
@@ -49,7 +49,7 @@ namespace clover {
      build_module_library(const ::llvm::Module &mod,
                           enum module::section::type section_type);

-      std::unique_ptr<::llvm::Module>
+      std::unique_ptr< ::llvm::Module>
      parse_module_library(const module &m, ::llvm::LLVMContext &ctx,
                           std::string &r_log);

--- a/src/gallium/state_trackers/clover/llvm/codegen/bitcode.cpp
+++ b/src/gallium/state_trackers/clover/llvm/codegen/bitcode.cpp
@@ -94,7 +94,7 @@ clover::llvm::build_module_library(const ::llvm::Module &mod,
   return m;
 }

-std::unique_ptr<::llvm::Module>
+std::unique_ptr< ::llvm::Module>
 clover::llvm::parse_module_library(const module &m, ::llvm::LLVMContext &ctx,
                                   std::string &r_log) {
   auto mod = ::llvm::parseBitcodeFile(::llvm::MemoryBufferRef(
@@ -104,5 +104,5 @@ clover::llvm::parse_module_library(const module &m, ::llvm::LLVMContext &ctx,
         fail(r_log, error(CL_INVALID_PROGRAM), s);
      });

-   return std::unique_ptr<::llvm::Module>(std::move(*mod));
+   return std::unique_ptr< ::llvm::Module>(std::move(*mod));
 }
--- a/src/gallium/state_trackers/clover/llvm/codegen/common.cpp
+++ b/src/gallium/state_trackers/clover/llvm/codegen/common.cpp
@@ -121,12 +121,12 @@ namespace {
         } else {
            // Other types.
            const auto actual_type =
-               isa<::llvm::PointerType>(arg_type) && arg.hasByValAttr() ?
-               cast<::llvm::PointerType>(arg_type)->getElementType() : arg_type;
+               isa< ::llvm::PointerType>(arg_type) && arg.hasByValAttr() ?
+               cast< ::llvm::PointerType>(arg_type)->getElementType() : arg_type;

            if (actual_type->isPointerTy()) {
               const unsigned address_space =
-                  cast<::llvm::PointerType>(actual_type)->getAddressSpace();
+                  cast< ::llvm::PointerType>(actual_type)->getAddressSpace();

               if (address_space == address_spaces[clang::LangAS::opencl_local
                                                   - compat::lang_as_offset]) {
--- a/src/gallium/state_trackers/clover/llvm/codegen/native.cpp
+++ b/src/gallium/state_trackers/clover/llvm/codegen/native.cpp
@@ -156,7 +156,7 @@ clover::llvm::print_module_native(const ::llvm::Module &mod,
                                  const target &target) {
   std::string log;
   try {
-      std::unique_ptr<::llvm::Module> cmod { CloneModule(&mod) };
+      std::unique_ptr< ::llvm::Module> cmod { CloneModule(&mod) };
      return as_string(emit_code(*cmod, target,
                                 TargetMachine::CGFT_AssemblyFile, log));
   } catch (...) {
--- a/src/gallium/state_trackers/clover/llvm/compat.hpp
+++ b/src/gallium/state_trackers/clover/llvm/compat.hpp
@@ -132,18 +132,18 @@ namespace clover {
 #endif
         }

-         inline std::unique_ptr<::llvm::Linker>
+         inline std::unique_ptr< ::llvm::Linker>
         create_linker(::llvm::Module &mod) {
 #if HAVE_LLVM >= 0x0308
-            return std::unique_ptr<::llvm::Linker>(new ::llvm::Linker(mod));
+            return std::unique_ptr< ::llvm::Linker>(new ::llvm::Linker(mod));
 #else
-            return std::unique_ptr<::llvm::Linker>(new ::llvm::Linker(&mod));
+            return std::unique_ptr< ::llvm::Linker>(new ::llvm::Linker(&mod));
 #endif
         }

         inline bool
         link_in_module(::llvm::Linker &linker,
-                        std::unique_ptr<::llvm::Module> mod) {
+                        std::unique_ptr< ::llvm::Module> mod) {
 #if HAVE_LLVM >= 0x0308
            return linker.linkInModule(std::move(mod));
 #else
--- a/src/gallium/state_trackers/clover/llvm/metadata.hpp
+++ b/src/gallium/state_trackers/clover/llvm/metadata.hpp
@@ -51,7 +51,7 @@ namespace clover {
         is_kernel_node_for(const ::llvm::Function &f) {
            return [&](const ::llvm::MDNode *n) {
               using ::llvm::mdconst::dyn_extract;
-               return &f == dyn_extract<::llvm::Function>(n->getOperand(0));
+               return &f == dyn_extract< ::llvm::Function>(n->getOperand(0));
            };
         }

@@ -65,7 +65,7 @@ namespace clover {
 #endif
         }

-         inline iterator_range<::llvm::MDNode::op_iterator>
+         inline iterator_range< ::llvm::MDNode::op_iterator>
         get_kernel_metadata_operands(const ::llvm::Function &f,
                                      const std::string &name) {
 #if HAVE_LLVM >= 0x0309
@@ -79,11 +79,11 @@ namespace clover {
            const auto kernel_node = find(is_kernel_node_for(f),
                                          get_kernel_nodes(*f.getParent()));

-            const auto data_node = cast<::llvm::MDNode>(
+            const auto data_node = cast< ::llvm::MDNode>(
               find([&](const ::llvm::MDOperand &op) {
-                     if (auto m = dyn_cast<::llvm::MDNode>(op))
+                     if (auto m = dyn_cast< ::llvm::MDNode>(op))
                        if (m->getNumOperands())
-                           if (auto m_name = dyn_cast<::llvm::MDString>(
+                           if (auto m_name = dyn_cast< ::llvm::MDString>(
                                  m->getOperand(0).get()))
                              return m_name->getString() == name;

@@ -106,7 +106,7 @@ namespace clover {
      get_argument_metadata(const ::llvm::Function &f,
                            const ::llvm::Argument &arg,
                            const std::string &name) {
-         return ::llvm::cast<::llvm::MDString>(
+         return ::llvm::cast< ::llvm::MDString>(
               detail::get_kernel_metadata_operands(f, name)[arg.getArgNo()])
            ->getString();
      }
--- a/src/gallium/targets/pipe-loader/pipe_i915.c
+++ b/src/gallium/targets/pipe-loader/pipe_i915.c
@@ -30,7 +30,7 @@ static const struct drm_conf_ret throttle_ret = {

 static const struct drm_conf_ret share_fd_ret = {
   .type = DRM_CONF_BOOL,
-   .val.val_int = true,
+   .val.val_bool = true,
 };

 static const struct drm_conf_ret *drm_configuration(enum drm_conf conf)
--- a/src/gallium/targets/pipe-loader/pipe_msm.c
+++ b/src/gallium/targets/pipe-loader/pipe_msm.c
@@ -24,7 +24,7 @@ static const struct drm_conf_ret throttle_ret = {

 static const struct drm_conf_ret share_fd_ret = {
   .type = DRM_CONF_BOOL,
-   .val.val_int = true,
+   .val.val_bool = true,
 };

 static const struct drm_conf_ret *drm_configuration(enum drm_conf conf)
--- a/src/gallium/targets/pipe-loader/pipe_nouveau.c
+++ b/src/gallium/targets/pipe-loader/pipe_nouveau.c
@@ -24,7 +24,7 @@ static const struct drm_conf_ret throttle_ret = {

 static const struct drm_conf_ret share_fd_ret = {
   .type = DRM_CONF_BOOL,
-   .val.val_int = true,
+   .val.val_bool = true,
 };

 static const struct drm_conf_ret *drm_configuration(enum drm_conf conf)
--- a/src/gallium/targets/pipe-loader/pipe_r300.c
+++ b/src/gallium/targets/pipe-loader/pipe_r300.c
@@ -20,7 +20,7 @@ static const struct drm_conf_ret throttle_ret = {

 static const struct drm_conf_ret share_fd_ret = {
   .type = DRM_CONF_BOOL,
-   .val.val_int = true,
+   .val.val_bool = true,
 };

 static const struct drm_conf_ret *drm_configuration(enum drm_conf conf)
--- a/src/gallium/targets/pipe-loader/pipe_r600.c
+++ b/src/gallium/targets/pipe-loader/pipe_r600.c
@@ -20,7 +20,7 @@ static const struct drm_conf_ret throttle_ret = {

 static const struct drm_conf_ret share_fd_ret = {
   .type = DRM_CONF_BOOL,
-   .val.val_int = true,
+   .val.val_bool = true,
 };

 static const struct drm_conf_ret *drm_configuration(enum drm_conf conf)
--- a/src/gallium/targets/pipe-loader/pipe_radeonsi.c
+++ b/src/gallium/targets/pipe-loader/pipe_radeonsi.c
@@ -26,7 +26,7 @@ static const struct drm_conf_ret throttle_ret = {

 static const struct drm_conf_ret share_fd_ret = {
   .type = DRM_CONF_BOOL,
-   .val.val_int = true,
+   .val.val_bool = true,
 };

 static const struct drm_conf_ret *drm_configuration(enum drm_conf conf)
--- a/src/gallium/targets/pipe-loader/pipe_vmwgfx.c
+++ b/src/gallium/targets/pipe-loader/pipe_vmwgfx.c
@@ -30,7 +30,7 @@ static const struct drm_conf_ret throttle_ret = {

 static const struct drm_conf_ret share_fd_ret = {
   .type = DRM_CONF_BOOL,
-   .val.val_int = true,
+   .val.val_bool = true,
 };

 static const struct drm_conf_ret *drm_configuration(enum drm_conf conf)
--- a/src/gallium/winsys/sw/dri/dri_sw_winsys.c
+++ b/src/gallium/winsys/sw/dri/dri_sw_winsys.c
@@ -200,7 +200,7 @@ dri_sw_displaytarget_display(struct sw_winsys *ws,

   if (box) {
       void *data;
-       data = dri_sw_dt->data + (dri_sw_dt->stride * box->y) + box->x * blsize;
+       data = (char *)dri_sw_dt->data + (dri_sw_dt->stride * box->y) + box->x * blsize;
       dri_sw_ws->lf->put_image2(dri_drawable, data,
                                 box->x, box->y, box->width, box->height, dri_sw_dt->stride);
   } else {
--- a/src/intel/compiler/brw_fs_reg_allocate.cpp
+++ b/src/intel/compiler/brw_fs_reg_allocate.cpp
@@ -822,12 +822,11 @@ fs_visitor::choose_spill_reg(struct ra_graph *g)
   foreach_block_and_inst(block, fs_inst, inst, cfg) {
      for (unsigned int i = 0; i < inst->sources; i++) {
 	 if (inst->src[i].file == VGRF)
-            spill_costs[inst->src[i].nr] += block_scale;
+            spill_costs[inst->src[i].nr] += regs_read(inst, i) * block_scale;
      }

      if (inst->dst.file == VGRF)
-         spill_costs[inst->dst.nr] += DIV_ROUND_UP(inst->size_written, REG_SIZE)
-                                      * block_scale;
+         spill_costs[inst->dst.nr] += regs_written(inst) * block_scale;

      switch (inst->opcode) {

--- a/src/intel/compiler/brw_vec4.cpp
+++ b/src/intel/compiler/brw_vec4.cpp
@@ -1071,6 +1071,13 @@ vec4_instruction::can_reswizzle(const struct gen_device_info *devinfo,
   if (devinfo->gen == 6 && is_math() && swizzle != BRW_SWIZZLE_XYZW)
      return false;

+   /* We can't swizzle implicit accumulator access.  We'd have to
+    * reswizzle the producer of the accumulator value in addition
+    * to the consumer (i.e. both MUL and MACH).  Just skip this.
+    */
+   if (reads_accumulator_implicitly())
+      return false;
+
   if (!can_do_writemask(devinfo) && dst_writemask != WRITEMASK_XYZW)
      return false;

--- a/src/intel/vulkan/anv_allocator.c
+++ b/src/intel/vulkan/anv_allocator.c
@@ -993,6 +993,25 @@ anv_scratch_pool_alloc(struct anv_device *device, struct anv_scratch_pool *pool,

   anv_bo_init_new(&bo->bo, device, size);

+   /* Even though the Scratch base pointers in 3DSTATE_*S are 64 bits, they
+    * are still relative to the general state base address.  When we emit
+    * STATE_BASE_ADDRESS, we set general state base address to 0 and the size
+    * to the maximum (1 page under 4GB).  This allows us to just place the
+    * scratch buffers anywhere we wish in the bottom 32 bits of address space
+    * and just set the scratch base pointer in 3DSTATE_*S using a relocation.
+    * However, in order to do so, we need to ensure that the kernel does not
+    * place the scratch BO above the 32-bit boundary.
+    *
+    * NOTE: Technically, it can't go "anywhere" because the top page is off
+    * limits.  However, when EXEC_OBJECT_SUPPORTS_48B_ADDRESS is set, the
+    * kernel allocates space using
+    *
+    *    end = min_t(u64, end, (1ULL << 32) - I915_GTT_PAGE_SIZE);
+    *
+    * so nothing will ever touch the top page.
+    */
+   bo->bo.flags &= ~EXEC_OBJECT_SUPPORTS_48B_ADDRESS;
+
   /* Set the exists last because it may be read by other threads */
   __sync_synchronize();
   bo->exists = true;
--- a/src/intel/vulkan/anv_batch_chain.c
+++ b/src/intel/vulkan/anv_batch_chain.c
@@ -1265,7 +1265,7 @@ anv_cmd_buffer_execbuf(struct anv_device *device,
                                      cmd_buffer->last_ss_pool_center);
   VkResult result =
      anv_execbuf_add_bo(&execbuf, &ss_pool->bo, &cmd_buffer->surface_relocs,
-                         &cmd_buffer->pool->alloc);
+                         &device->alloc);
   if (result != VK_SUCCESS)
      return result;

@@ -1278,7 +1278,7 @@ anv_cmd_buffer_execbuf(struct anv_device *device,
                                       cmd_buffer->last_ss_pool_center);

      result = anv_execbuf_add_bo(&execbuf, &(*bbo)->bo, &(*bbo)->relocs,
-                                  &cmd_buffer->pool->alloc);
+                                  &device->alloc);
      if (result != VK_SUCCESS)
         return result;
   }
@@ -1387,7 +1387,7 @@ anv_cmd_buffer_execbuf(struct anv_device *device,

   result = anv_device_execbuf(device, &execbuf.execbuf, execbuf.bos);

-   anv_execbuf_finish(&execbuf, &cmd_buffer->pool->alloc);
+   anv_execbuf_finish(&execbuf, &device->alloc);

   return result;
 }
--- a/src/intel/vulkan/anv_blorp.c
+++ b/src/intel/vulkan/anv_blorp.c
@@ -701,10 +701,10 @@ void anv_CmdUpdateBuffer(
      struct anv_state tmp_data =
         anv_cmd_buffer_alloc_dynamic_state(cmd_buffer, copy_size, 64);

-      anv_state_flush(cmd_buffer->device, tmp_data);
-
      memcpy(tmp_data.map, pData, copy_size);

+      anv_state_flush(cmd_buffer->device, tmp_data);
+
      int bs = 16;
      bs = gcd_pow2_u64(bs, dstOffset);
      bs = gcd_pow2_u64(bs, copy_size);
@@ -1398,7 +1398,6 @@ ccs_resolve_attachment(struct anv_cmd_buffer *cmd_buffer,
    * still hot in the cache.
    */
   bool found_draw = false;
-   bool self_dep = false;
   enum anv_subpass_usage usage = 0;
   for (uint32_t s = subpass_idx + 1; s < pass->subpass_count; s++) {
      usage |= pass->attachments[att].subpass_usage[s];
@@ -1408,8 +1407,6 @@ ccs_resolve_attachment(struct anv_cmd_buffer *cmd_buffer,
          * wait to resolve until then.
          */
         found_draw = true;
-         if (pass->attachments[att].subpass_usage[s] & ANV_SUBPASS_USAGE_INPUT)
-            self_dep = true;
         break;
      }
   }
@@ -1468,14 +1465,6 @@ ccs_resolve_attachment(struct anv_cmd_buffer *cmd_buffer,
          *    binding this surface to Sampler."
          */
         resolve_op = BLORP_FAST_CLEAR_OP_RESOLVE_PARTIAL;
-      } else if (cmd_buffer->device->info.gen == 8 && self_dep &&
-                 att_state->input_aux_usage == ISL_AUX_USAGE_CCS_D) {
-         /* On Broadwell we still need to do resolves when there is a
-          * self-dependency because HW could not see fast-clears and works
-          * on the render cache as if there was regular non-fast-clear surface.
-          * To avoid any inconsistency, we force the resolve.
-          */
-         resolve_op = BLORP_FAST_CLEAR_OP_RESOLVE_FULL;
      }
   }

--- a/src/intel/vulkan/genX_cmd_buffer.c
+++ b/src/intel/vulkan/genX_cmd_buffer.c
@@ -291,27 +291,21 @@ color_attachment_compute_aux_usage(struct anv_device *device,
      att_state->input_aux_usage = ISL_AUX_USAGE_CCS_E;
   } else if (att_state->fast_clear) {
      att_state->aux_usage = ISL_AUX_USAGE_CCS_D;
-      if (GEN_GEN >= 9 &&
-          !isl_format_supports_ccs_e(&device->info, iview->isl.format)) {
-         /* From the Sky Lake PRM, RENDER_SURFACE_STATE::AuxiliarySurfaceMode:
-          *
-          *    "If Number of Multisamples is MULTISAMPLECOUNT_1, AUX_CCS_D
-          *    setting is only allowed if Surface Format supported for Fast
-          *    Clear. In addition, if the surface is bound to the sampling
-          *    engine, Surface Format must be supported for Render Target
-          *    Compression for surfaces bound to the sampling engine."
-          *
-          * In other words, we can't sample from a fast-cleared image if it
-          * doesn't also support color compression.
-          */
-         att_state->input_aux_usage = ISL_AUX_USAGE_NONE;
-      } else if (GEN_GEN >= 8) {
-         /* Broadwell/Skylake can sample from fast-cleared images */
+      /* From the Sky Lake PRM, RENDER_SURFACE_STATE::AuxiliarySurfaceMode:
+       *
+       *    "If Number of Multisamples is MULTISAMPLECOUNT_1, AUX_CCS_D
+       *    setting is only allowed if Surface Format supported for Fast
+       *    Clear. In addition, if the surface is bound to the sampling
+       *    engine, Surface Format must be supported for Render Target
+       *    Compression for surfaces bound to the sampling engine."
+       *
+       * In other words, we can only sample from a fast-cleared image if it
+       * also supports color compression.
+       */
+      if (isl_format_supports_ccs_e(&device->info, iview->isl.format))
         att_state->input_aux_usage = ISL_AUX_USAGE_CCS_D;
-      } else {
-         /* Ivy Bridge and Haswell cannot */
+      else
         att_state->input_aux_usage = ISL_AUX_USAGE_NONE;
-      }
   } else {
      att_state->aux_usage = ISL_AUX_USAGE_NONE;
      att_state->input_aux_usage = ISL_AUX_USAGE_NONE;
--- a/src/mapi/glapi/gen/gl_API.xml
+++ b/src/mapi/glapi/gen/gl_API.xml
@@ -10219,7 +10219,7 @@
 </category>

 <category name="GL_EXT_multi_draw_arrays" number="148">
-    <function name="MultiDrawArraysEXT" es1="1.0" es2="2.0" alias="MultiDrawArrays">
+    <function name="MultiDrawArraysEXT" es1="1.0" es2="2.0" exec="dynamic" alias="MultiDrawArrays">
        <param name="mode" type="GLenum"/>
        <param name="first" type="const GLint *"/>
        <param name="count" type="const GLsizei *"/>
--- a/src/mesa/Makefile.am
+++ b/src/mesa/Makefile.am
@@ -96,6 +96,7 @@ endif

 AM_CPPFLAGS = $(DEFINES) $(INCLUDE_DIRS)
 AM_CFLAGS = \
+	$(VDPAU_CFLAGS) \
 	$(LLVM_CFLAGS) \
 	$(VISIBILITY_CFLAGS) \
 	$(MSVC2013_COMPAT_CFLAGS)
--- a/src/mesa/main/api_validate.c
+++ b/src/mesa/main/api_validate.c
@@ -816,25 +816,10 @@ _mesa_validate_DrawRangeElements(struct gl_context *ctx, GLenum mode,
                                       "glDrawRangeElements");
 }

+
 static bool
-validate_draw_arrays(struct gl_context *ctx, const char *func,
-                     GLenum mode, GLsizei count, GLsizei numInstances)
+need_xfb_remaining_prims_check(const struct gl_context *ctx)
 {
-   struct gl_transform_feedback_object *xfb_obj
-      = ctx->TransformFeedback.CurrentObject;
-   FLUSH_CURRENT(ctx, 0);
-
-   if (count < 0) {
-      _mesa_error(ctx, GL_INVALID_VALUE, "%s(count)", func);
-      return false;
-   }
-
-   if (!_mesa_valid_prim_mode(ctx, mode, func))
-      return false;
-
-   if (!check_valid_to_render(ctx, func))
-      return false;
-
   /* From the GLES3 specification, section 2.14.2 (Transform Feedback
    * Primitive Capture):
    *
@@ -862,10 +847,33 @@ validate_draw_arrays(struct gl_context *ctx, const char *func,
    *     is removed and replaced with the GL behavior (primitives are not
    *     written and the corresponding counter is not updated)..."
    */
-   if (_mesa_is_gles3(ctx) && _mesa_is_xfb_active_and_unpaused(ctx) &&
-       !_mesa_has_OES_geometry_shader(ctx) &&
-       !_mesa_has_OES_tessellation_shader(ctx)) {
-      size_t prim_count = vbo_count_tessellated_primitives(mode, count, 1);
+   return _mesa_is_gles3(ctx) && _mesa_is_xfb_active_and_unpaused(ctx) &&
+          !_mesa_has_OES_geometry_shader(ctx) &&
+          !_mesa_has_OES_tessellation_shader(ctx);
+}
+
+
+static bool
+validate_draw_arrays(struct gl_context *ctx, const char *func,
+                     GLenum mode, GLsizei count, GLsizei numInstances)
+{
+   FLUSH_CURRENT(ctx, 0);
+
+   if (count < 0) {
+      _mesa_error(ctx, GL_INVALID_VALUE, "%s(count)", func);
+      return false;
+   }
+
+   if (!_mesa_valid_prim_mode(ctx, mode, func))
+      return false;
+
+   if (!check_valid_to_render(ctx, func))
+      return false;
+
+   if (need_xfb_remaining_prims_check(ctx)) {
+      struct gl_transform_feedback_object *xfb_obj
+         = ctx->TransformFeedback.CurrentObject;
+      size_t prim_count = vbo_count_tessellated_primitives(mode, count, numInstances);
      if (xfb_obj->GlesRemainingPrims < prim_count) {
         _mesa_error(ctx, GL_INVALID_OPERATION,
                     "%s(exceeds transform feedback size)", func);
@@ -913,6 +921,60 @@ _mesa_validate_DrawArraysInstanced(struct gl_context *ctx, GLenum mode, GLint fi
 }


+/**
+ * Called to error check the function parameters.
+ *
+ * Note that glMultiDrawArrays is not part of GLES, so there's limited scope
+ * for sharing code with the validation of glDrawArrays.
+ */
+bool
+_mesa_validate_MultiDrawArrays(struct gl_context *ctx, GLenum mode,
+                               const GLsizei *count, GLsizei primcount)
+{
+   int i;
+
+   FLUSH_CURRENT(ctx, 0);
+
+   if (!_mesa_valid_prim_mode(ctx, mode, "glMultiDrawArrays"))
+      return false;
+
+   if (!check_valid_to_render(ctx, "glMultiDrawArrays"))
+      return false;
+
+   if (primcount < 0) {
+      _mesa_error(ctx, GL_INVALID_VALUE, "glMultiDrawArrays(primcount=%d)",
+                  primcount);
+      return false;
+   }
+
+   for (i = 0; i < primcount; ++i) {
+      if (count[i] < 0) {
+         _mesa_error(ctx, GL_INVALID_VALUE, "glMultiDrawArrays(count[%d]=%d)",
+                     i, count[i]);
+         return false;
+      }
+   }
+
+   if (need_xfb_remaining_prims_check(ctx)) {
+      struct gl_transform_feedback_object *xfb_obj
+         = ctx->TransformFeedback.CurrentObject;
+      size_t xfb_prim_count = 0;
+
+      for (i = 0; i < primcount; ++i)
+         xfb_prim_count += vbo_count_tessellated_primitives(mode, count[i], 1);
+
+      if (xfb_obj->GlesRemainingPrims < xfb_prim_count) {
+         _mesa_error(ctx, GL_INVALID_OPERATION,
+                     "glMultiDrawArrays(exceeds transform feedback size)");
+         return false;
+      }
+      xfb_obj->GlesRemainingPrims -= xfb_prim_count;
+   }
+
+   return true;
+}
+
+
 GLboolean
 _mesa_validate_DrawElementsInstanced(struct gl_context *ctx,
                                     GLenum mode, GLsizei count, GLenum type,
--- a/src/mesa/main/api_validate.h
+++ b/src/mesa/main/api_validate.h
@@ -48,6 +48,10 @@ _mesa_valid_prim_mode(struct gl_context *ctx, GLenum mode, const char *name);
 extern GLboolean
 _mesa_validate_DrawArrays(struct gl_context *ctx, GLenum mode, GLsizei count);

+extern bool
+_mesa_validate_MultiDrawArrays(struct gl_context *ctx, GLenum mode,
+                               const GLsizei *count, GLsizei primcount);
+
 extern GLboolean
 _mesa_validate_DrawElements(struct gl_context *ctx,
 			    GLenum mode, GLsizei count, GLenum type,
--- a/src/mesa/main/glthread.c
+++ b/src/mesa/main/glthread.c
@@ -265,7 +265,7 @@ _mesa_glthread_finish(struct gl_context *ctx)
    * dri interface entrypoints), in which case we don't need to actually
    * synchronize against ourself.
    */
-   if (pthread_self() == glthread->thread)
+   if (pthread_equal(pthread_self(), glthread->thread))
      return;

   pthread_mutex_lock(&glthread->mutex);
--- a/src/mesa/main/uniform_query.cpp
+++ b/src/mesa/main/uniform_query.cpp
@@ -945,6 +945,8 @@ _mesa_uniform(GLint location, GLsizei count, const GLvoid *values,
    */
   if (uni->type->is_sampler()) {
      bool flushed = false;
+      shProg->SamplersValidated = GL_TRUE;
+
      for (int i = 0; i < MESA_SHADER_STAGES; i++) {
 	 struct gl_linked_shader *const sh = shProg->_LinkedShaders[i];

--- a/src/mesa/main/uniforms.c
+++ b/src/mesa/main/uniforms.c
@@ -67,11 +67,15 @@ void
 _mesa_update_shader_textures_used(struct gl_shader_program *shProg,
                                  struct gl_program *prog)
 {
+   GLbitfield mask = prog->SamplersUsed;
+   gl_shader_stage prog_stage =
+      _mesa_program_enum_to_shader_stage(prog->Target);
+   struct gl_linked_shader *shader = shProg->_LinkedShaders[prog_stage];
+
+   assert(shader);
+
   memset(prog->TexturesUsed, 0, sizeof(prog->TexturesUsed));

-   shProg->SamplersValidated = GL_TRUE;
-
-   GLbitfield mask = prog->SamplersUsed;
   while (mask) {
      const int s = u_bit_scan(&mask);
      GLuint unit = prog->SamplerUnits[s];
@@ -87,8 +91,20 @@ _mesa_update_shader_textures_used(struct gl_shader_program *shProg,
       *     types pointing to the same texture image unit within a program
       *     object."
       */
-      if (prog->TexturesUsed[unit] & ~(1 << tgt))
-         shProg->SamplersValidated = GL_FALSE;
+      unsigned stages_mask = shProg->data->linked_stages;
+      while (stages_mask) {
+         const int stage = u_bit_scan(&stages_mask);
+
+         /* Skip validation if we are yet to update textures used in this
+          * stage.
+          */
+         if (prog_stage < stage)
+            break;
+
+         struct gl_program *glprog = shProg->_LinkedShaders[stage]->Program;
+         if (glprog->TexturesUsed[unit] & ~(1 << tgt))
+            shProg->SamplersValidated = GL_FALSE;
+      }

      prog->TexturesUsed[unit] |= (1 << tgt);
   }
--- a/src/mesa/main/varray.c
+++ b/src/mesa/main/varray.c
@@ -1539,24 +1539,6 @@ _mesa_UnlockArraysEXT( void )
 }


-/* GL_EXT_multi_draw_arrays */
-void GLAPIENTRY
-_mesa_MultiDrawArrays( GLenum mode, const GLint *first,
-                          const GLsizei *count, GLsizei primcount )
-{
-   GET_CURRENT_CONTEXT(ctx);
-   GLint i;
-
-   FLUSH_VERTICES(ctx, 0);
-
-   for (i = 0; i < primcount; i++) {
-      if (count[i] > 0) {
-         CALL_DrawArrays(ctx->CurrentClientDispatch, (mode, first[i], count[i]));
-      }
-   }
-}
-
-
 /* GL_IBM_multimode_draw_arrays */
 void GLAPIENTRY
 _mesa_MultiModeDrawArraysIBM( const GLenum * mode, const GLint * first,
--- a/src/mesa/program/ir_to_mesa.cpp
+++ b/src/mesa/program/ir_to_mesa.cpp
@@ -3117,6 +3117,11 @@ _mesa_glsl_link_shader(struct gl_context *ctx, struct gl_shader_program *prog)
   }

   if (prog->data->LinkStatus) {
+      /* Reset sampler validated to true, validation happens via the
+       * LinkShader call below.
+       */
+      prog->SamplersValidated = GL_TRUE;
+
      if (!ctx->Driver.LinkShader(ctx, prog)) {
         prog->data->LinkStatus = linking_failure;
      }
--- a/src/mesa/state_tracker/st_cb_fbo.h
+++ b/src/mesa/state_tracker/st_cb_fbo.h
@@ -85,7 +85,8 @@ static inline struct st_framebuffer *
 st_ws_framebuffer(struct gl_framebuffer *fb)
 {
   /* FBO cannot be casted.  See st_new_framebuffer */
-   if (fb && _mesa_is_winsys_fbo(fb))
+   if (fb && _mesa_is_winsys_fbo(fb) &&
+       fb != _mesa_get_incomplete_framebuffer())
      return (struct st_framebuffer *) fb;
   return NULL;
 }
--- a/src/mesa/state_tracker/st_draw.c
+++ b/src/mesa/state_tracker/st_draw.c
@@ -282,6 +282,8 @@ st_indirect_draw_vbo(struct gl_context *ctx,
   assert(ctx->NewState == 0x0);
   assert(stride);

+   st_invalidate_readpix_cache(st);
+
   /* Validate state. */
   if ((st->dirty | ctx->NewDriverState) & ST_PIPELINE_RENDER_STATE_MASK ||
       st->gfx_shaders_may_be_dirty) {
--- a/src/mesa/state_tracker/st_manager.c
+++ b/src/mesa/state_tracker/st_manager.c
@@ -844,13 +844,14 @@ st_manager_flush_frontbuffer(struct st_context *st)
   struct st_framebuffer *stfb = st_ws_framebuffer(st->ctx->DrawBuffer);
   struct st_renderbuffer *strb = NULL;

+   assert(st->ctx->DrawBuffer != _mesa_get_incomplete_framebuffer());
+
   if (stfb)
      strb = st_renderbuffer(stfb->Base.Attachment[BUFFER_FRONT_LEFT].Renderbuffer);
   if (!strb)
      return;

   /* never a dummy fb */
-   assert(&stfb->Base != _mesa_get_incomplete_framebuffer());
   stfb->iface->flush_front(&st->iface, stfb->iface, ST_ATTACHMENT_FRONT_LEFT);
 }

--- a/src/mesa/vbo/vbo_exec_array.c
+++ b/src/mesa/vbo/vbo_exec_array.c
@@ -404,7 +404,8 @@ vbo_bind_arrays(struct gl_context *ctx)
 */
 static void
 vbo_draw_arrays(struct gl_context *ctx, GLenum mode, GLint start,
-                GLsizei count, GLuint numInstances, GLuint baseInstance)
+                GLsizei count, GLuint numInstances, GLuint baseInstance,
+                GLuint drawID)
 {
   struct vbo_context *vbo = vbo_context(ctx);
   struct _mesa_prim prim[2];
@@ -420,6 +421,7 @@ vbo_draw_arrays(struct gl_context *ctx, GLenum mode, GLint start,
   prim[0].mode = mode;
   prim[0].num_instances = numInstances;
   prim[0].base_instance = baseInstance;
+   prim[0].draw_id = drawID;
   prim[0].is_indirect = 0;
   prim[0].start = start;
   prim[0].count = count;
@@ -572,7 +574,7 @@ vbo_exec_DrawArrays(GLenum mode, GLint start, GLsizei count)
   if (0)
      check_draw_arrays_data(ctx, start, count);

-   vbo_draw_arrays(ctx, mode, start, count, 1, 0);
+   vbo_draw_arrays(ctx, mode, start, count, 1, 0, 0);

   if (0)
      print_draw_arrays(ctx, mode, start, count);
@@ -600,7 +602,7 @@ vbo_exec_DrawArraysInstanced(GLenum mode, GLint start, GLsizei count,
   if (0)
      check_draw_arrays_data(ctx, start, count);

-   vbo_draw_arrays(ctx, mode, start, count, numInstances, 0);
+   vbo_draw_arrays(ctx, mode, start, count, numInstances, 0, 0);

   if (0)
      print_draw_arrays(ctx, mode, start, count);
@@ -630,13 +632,52 @@ vbo_exec_DrawArraysInstancedBaseInstance(GLenum mode, GLint first,
   if (0)
      check_draw_arrays_data(ctx, first, count);

-   vbo_draw_arrays(ctx, mode, first, count, numInstances, baseInstance);
+   vbo_draw_arrays(ctx, mode, first, count, numInstances, baseInstance, 0);

   if (0)
      print_draw_arrays(ctx, mode, first, count);
 }


+/**
+ * Called from glMultiDrawArrays when in immediate mode.
+ */
+static void GLAPIENTRY
+vbo_exec_MultiDrawArrays(GLenum mode, const GLint *first,
+                         const GLsizei *count, GLsizei primcount)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   GLint i;
+
+   if (MESA_VERBOSE & VERBOSE_DRAW)
+      _mesa_debug(ctx,
+                  "glMultiDrawArrays(%s, %p, %p, %d)\n",
+                  _mesa_enum_to_string(mode), first, count, primcount);
+
+   if (!_mesa_validate_MultiDrawArrays(ctx, mode, count, primcount))
+      return;
+
+   for (i = 0; i < primcount; i++) {
+      if (count[i] > 0) {
+         if (0)
+            check_draw_arrays_data(ctx, first[i], count[i]);
+
+         /* The GL_ARB_shader_draw_parameters spec adds the following after the
+          * pseudo-code describing glMultiDrawArrays:
+          *
+          *    "The index of the draw (<i> in the above pseudo-code) may be
+          *     read by a vertex shader as <gl_DrawIDARB>, as described in
+          *     Section 11.1.3.9."
+          */
+         vbo_draw_arrays(ctx, mode, first[i], count[i], 1, 0, i);
+
+         if (0)
+            print_draw_arrays(ctx, mode, first[i], count[i]);
+      }
+   }
+}
+
+

 /**
 * Map GL_ELEMENT_ARRAY_BUFFER and print contents.
@@ -1230,7 +1271,7 @@ vbo_draw_transform_feedback(struct gl_context *ctx, GLenum mode,
        !_mesa_all_varyings_in_vbos(ctx->Array.VAO))) {
      GLsizei n =
         ctx->Driver.GetTransformFeedbackVertexCount(ctx, obj, stream);
-      vbo_draw_arrays(ctx, mode, 0, n, numInstances, 0);
+      vbo_draw_arrays(ctx, mode, 0, n, numInstances, 0, 0);
      return;
   }

@@ -1641,6 +1682,7 @@ vbo_initialize_exec_dispatch(const struct gl_context *ctx,
      SET_DrawRangeElements(exec, vbo_exec_DrawRangeElements);
   }

+   SET_MultiDrawArrays(exec, vbo_exec_MultiDrawArrays);
   SET_MultiDrawElementsEXT(exec, vbo_exec_MultiDrawElements);

   if (ctx->API == API_OPENGL_COMPAT) {
--- a/src/mesa/vbo/vbo_save_api.c
+++ b/src/mesa/vbo/vbo_save_api.c
@@ -1176,6 +1176,40 @@ _save_OBE_DrawArrays(GLenum mode, GLint start, GLsizei count)
 }


+static void GLAPIENTRY
+_save_OBE_MultiDrawArrays(GLenum mode, const GLint *first,
+                          const GLsizei *count, GLsizei primcount)
+{
+   GET_CURRENT_CONTEXT(ctx);
+   GLint i;
+
+   if (!_mesa_is_valid_prim_mode(ctx, mode)) {
+      _mesa_compile_error(ctx, GL_INVALID_ENUM, "glMultiDrawArrays(mode)");
+      return;
+   }
+
+   if (primcount < 0) {
+      _mesa_compile_error(ctx, GL_INVALID_VALUE,
+                          "glMultiDrawArrays(primcount<0)");
+      return;
+   }
+
+   for (i = 0; i < primcount; i++) {
+      if (count[i] < 0) {
+         _mesa_compile_error(ctx, GL_INVALID_VALUE,
+                             "glMultiDrawArrays(count[i]<0)");
+         return;
+      }
+   }
+
+   for (i = 0; i < primcount; i++) {
+      if (count[i] > 0) {
+         _save_OBE_DrawArrays(mode, first[i], count[i]);
+      }
+   }
+}
+
+
 /* Could do better by copying the arrays and element list intact and
 * then emitting an indexed prim at runtime.
 */
@@ -1484,6 +1518,7 @@ vbo_initialize_save_dispatch(const struct gl_context *ctx,
                             struct _glapi_table *exec)
 {
   SET_DrawArrays(exec, _save_OBE_DrawArrays);
+   SET_MultiDrawArrays(exec, _save_OBE_MultiDrawArrays);
   SET_DrawElements(exec, _save_OBE_DrawElements);
   SET_DrawElementsBaseVertex(exec, _save_OBE_DrawElementsBaseVertex);
   SET_DrawRangeElements(exec, _save_OBE_DrawRangeElements);
--- a/src/util/disk_cache.c
+++ b/src/util/disk_cache.c
@@ -31,7 +31,6 @@
 #include <sys/file.h>
 #include <sys/types.h>
 #include <sys/stat.h>
-#include <sys/statvfs.h>
 #include <sys/mman.h>
 #include <unistd.h>
 #include <fcntl.h>
@@ -47,7 +46,6 @@
 #include "util/mesa-sha1.h"
 #include "util/ralloc.h"
 #include "main/errors.h"
-#include "util/macros.h"

 #include "disk_cache.h"

@@ -171,7 +169,6 @@ disk_cache_create(const char *gpu_name, const char *timestamp)
   uint64_t max_size;
   int fd = -1;
   struct stat sb;
-   struct statvfs vfs = { 0 };
   size_t size;

   /* If running as a users other than the real user disable cache */
@@ -331,10 +328,9 @@ disk_cache_create(const char *gpu_name, const char *timestamp)
      }
   }

-   /* Default to 1GB or 10% of filesystem for maximum cache size. */
+   /* Default to 1GB for maximum cache size. */
   if (max_size == 0) {
-      statvfs(path, &vfs);
-      max_size = MAX2(1024*1024*1024, vfs.f_blocks * vfs.f_bsize / 10);
+      max_size = 1024*1024*1024;
   }

   cache->max_size = max_size;
@@ -532,7 +528,7 @@ unlink_lru_file_from_directory(const char *path)
   unlink(filename);
   free (filename);

-   return sb.st_size;
+   return sb.st_blocks * 512;
 }

 /* Is entry a directory with a two-character name, (and not the
@@ -637,8 +633,8 @@ disk_cache_remove(struct disk_cache *cache, const cache_key key)
   unlink(filename);
   free(filename);

-   if (sb.st_size)
-      p_atomic_add(cache->size, - (uint64_t)sb.st_size);
+   if (sb.st_blocks)
+      p_atomic_add(cache->size, - (uint64_t)sb.st_blocks * 512);
 }

 static ssize_t
@@ -880,8 +876,14 @@ cache_put(void *job, int thread_index)
      goto done;
   }

-   file_size += cf_data_size + dc_job->cache->driver_keys_blob_size;
-   p_atomic_add(dc_job->cache->size, file_size);
+   struct stat sb;
+   if (stat(filename, &sb) == -1) {
+      /* Something went wrong remove the file */
+      unlink(filename);
+      goto done;
+   }
+
+   p_atomic_add(dc_job->cache->size, sb.st_blocks * 512);

 done:
   if (fd_final != -1)
--- a/src/util/u_queue.c
+++ b/src/util/u_queue.c
@@ -298,9 +298,18 @@ util_queue_add_job(struct util_queue *queue,
   struct util_queue_job *ptr;

   assert(fence->signalled);
-   fence->signalled = false;

   mtx_lock(&queue->lock);
+   if (queue->kill_threads) {
+      mtx_unlock(&queue->lock);
+      /* well no good option here, but any leaks will be
+       * short-lived as things are shutting down..
+       */
+      return;
+   }
+
+   fence->signalled = false;
+
   assert(queue->num_queued >= 0 && queue->num_queued <= queue->max_jobs);

   /* if the queue is full, wait until there is space */