Increment version to 10.3.0-rc3

Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>

Makefile.am

@@ -64,14 +64,13 @@ IGNORE_FILES = \
 
 parsers: configure
 	$(MAKE) -C src/glsl glsl_parser.cpp glsl_parser.h glsl_lexer.cpp glcpp/glcpp-lex.c glcpp/glcpp-parse.c glcpp/glcpp-parse.h
-	$(MAKE) -C src/mesa program/lex.yy.c program/program_parse.tab.c program/program_parse.tab.h
 
 # Everything for new a Mesa release:
 ARCHIVES = $(PACKAGE_NAME).tar.gz \
 	$(PACKAGE_NAME).tar.bz2 \
 	$(PACKAGE_NAME).zip
 
-tarballs: md5
+tarballs: checksums
 	rm -f ../$(PACKAGE_DIR) $(PACKAGE_NAME).tar
 
 manifest.txt: .git
@@ -98,9 +97,9 @@ $(PACKAGE_NAME).zip: parsers ../$(PACKAGE_DIR) manifest.txt
 	zip -q -@ $(PACKAGE_NAME).zip < $(PACKAGE_DIR)/manifest.txt ; \
 	mv $(PACKAGE_NAME).zip $(PACKAGE_DIR)
 
-md5: $(ARCHIVES)
-	@-md5sum $(PACKAGE_NAME).tar.gz
-	@-md5sum $(PACKAGE_NAME).tar.bz2
-	@-md5sum $(PACKAGE_NAME).zip
+checksums: $(ARCHIVES)
+	@-sha256sum $(PACKAGE_NAME).tar.gz
+	@-sha256sum $(PACKAGE_NAME).tar.bz2
+	@-sha256sum $(PACKAGE_NAME).zip
 
 .PHONY: tarballs md5

VERSION

@@ -1 +1 @@
-10.3.0-devel
+10.3.0-rc3

configure.ac

@@ -355,6 +355,24 @@ AC_LINK_IFELSE(
 LDFLAGS=$save_LDFLAGS
 AM_CONDITIONAL(HAVE_LD_VERSION_SCRIPT, test "$have_ld_version_script" = "yes")
 
+dnl
+dnl Check if linker supports dynamic list files
+dnl
+AC_MSG_CHECKING([if the linker supports --dynamic-list])
+save_LDFLAGS=$LDFLAGS
+LDFLAGS="$LDFLAGS -Wl,--dynamic-list=conftest.dyn"
+cat > conftest.dyn <<EOF
+{
+	radeon_drm_winsys_create;
+};
+EOF
+AC_LINK_IFELSE(
+    [AC_LANG_SOURCE([int main() { return 0;}])],
+    [have_ld_dynamic_list=yes;AC_MSG_RESULT(yes)],
+    [have_ld_dynamic_list=no; AC_MSG_RESULT(no)])
+LDFLAGS=$save_LDFLAGS
+AM_CONDITIONAL(HAVE_LD_DYNAMIC_LIST, test "$have_ld_dynamic_list" = "yes")
+
 dnl
 dnl compatibility symlinks
 dnl

include/pci_ids/radeonsi_pci_ids.h

@@ -38,6 +38,7 @@ CHIPSET(0x6828, VERDE_6828, VERDE)
 CHIPSET(0x6829, VERDE_6829, VERDE)
 CHIPSET(0x682A, VERDE_682A, VERDE)
 CHIPSET(0x682B, VERDE_682B, VERDE)
+CHIPSET(0x682C, VERDE_682C, VERDE)
 CHIPSET(0x682D, VERDE_682D, VERDE)
 CHIPSET(0x682F, VERDE_682F, VERDE)
 CHIPSET(0x6830, VERDE_6830, VERDE)
@@ -54,8 +55,11 @@ CHIPSET(0x6600, OLAND_6600, OLAND)
 CHIPSET(0x6601, OLAND_6601, OLAND)
 CHIPSET(0x6602, OLAND_6602, OLAND)
 CHIPSET(0x6603, OLAND_6603, OLAND)
+CHIPSET(0x6604, OLAND_6604, OLAND)
+CHIPSET(0x6605, OLAND_6605, OLAND)
 CHIPSET(0x6606, OLAND_6606, OLAND)
 CHIPSET(0x6607, OLAND_6607, OLAND)
+CHIPSET(0x6608, OLAND_6608, OLAND)
 CHIPSET(0x6610, OLAND_6610, OLAND)
 CHIPSET(0x6611, OLAND_6611, OLAND)
 CHIPSET(0x6613, OLAND_6613, OLAND)
@@ -73,6 +77,8 @@ CHIPSET(0x666F, HAINAN_666F, HAINAN)
 
 CHIPSET(0x6640, BONAIRE_6640, BONAIRE)
 CHIPSET(0x6641, BONAIRE_6641, BONAIRE)
+CHIPSET(0x6646, BONAIRE_6646, BONAIRE)
+CHIPSET(0x6647, BONAIRE_6647, BONAIRE)
 CHIPSET(0x6649, BONAIRE_6649, BONAIRE)
 CHIPSET(0x6650, BONAIRE_6650, BONAIRE)
 CHIPSET(0x6651, BONAIRE_6651, BONAIRE)
@@ -132,6 +138,7 @@ CHIPSET(0x1313, KAVERI_1313, KAVERI)
 CHIPSET(0x1315, KAVERI_1315, KAVERI)
 CHIPSET(0x1316, KAVERI_1316, KAVERI)
 CHIPSET(0x1317, KAVERI_1317, KAVERI)
+CHIPSET(0x1318, KAVERI_1318, KAVERI)
 CHIPSET(0x131B, KAVERI_131B, KAVERI)
 CHIPSET(0x131C, KAVERI_131C, KAVERI)
 CHIPSET(0x131D, KAVERI_131D, KAVERI)

src/gallium/auxiliary/vl/vl_compositor.c

@@ -1060,6 +1060,7 @@ vl_compositor_render(struct vl_compositor_state *s,
       s->scissor.maxx = dst_surface->width;
       s->scissor.maxy = dst_surface->height;
    }
+   c->pipe->set_scissor_states(c->pipe, 0, 1, &s->scissor);
 
    gen_vertex_data(c, s, dirty_area);
 
@@ -1072,7 +1073,6 @@ vl_compositor_render(struct vl_compositor_state *s,
       dirty_area->x1 = dirty_area->y1 = MIN_DIRTY;
    }
 
-   c->pipe->set_scissor_states(c->pipe, 0, 1, &s->scissor);
    c->pipe->set_framebuffer_state(c->pipe, &c->fb_state);
    c->pipe->bind_vs_state(c->pipe, c->vs);
    c->pipe->set_vertex_buffers(c->pipe, 0, 1, &c->vertex_buf);

src/gallium/drivers/freedreno/a2xx/fd2_context.c

@@ -98,6 +98,7 @@ fd2_context_create(struct pipe_screen *pscreen, void *priv)
 	pctx = &fd2_ctx->base.base;
 
 	fd2_ctx->base.dev = fd_device_ref(screen->dev);
+	fd2_ctx->base.screen = fd_screen(pscreen);
 
 	pctx->destroy = fd2_context_destroy;
 	pctx->create_blend_state = fd2_blend_state_create;

src/gallium/drivers/freedreno/a3xx/fd3_emit.c

@@ -215,14 +215,19 @@ emit_textures(struct fd_ringbuffer *ring,
 		OUT_RING(ring, CP_LOAD_STATE_1_STATE_TYPE(ST_CONSTANTS) |
 				CP_LOAD_STATE_1_EXT_SRC_ADDR(0));
 		for (i = 0; i < tex->num_textures; i++) {
-			static const struct fd3_pipe_sampler_view dummy_view = {};
+			static const struct fd3_pipe_sampler_view dummy_view = {
+					.base.u.tex.first_level = 1,
+			};
 			const struct fd3_pipe_sampler_view *view = tex->textures[i] ?
 					fd3_pipe_sampler_view(tex->textures[i]) :
 					&dummy_view;
 			struct fd_resource *rsc = view->tex_resource;
+			unsigned start = view->base.u.tex.first_level;
+			unsigned end   = view->base.u.tex.last_level;
 
-			for (j = 0; j < view->mipaddrs; j++) {
-				struct fd_resource_slice *slice = fd_resource_slice(rsc, j);
+			for (j = 0; j < (end - start + 1); j++) {
+				struct fd_resource_slice *slice =
+						fd_resource_slice(rsc, j + start);
 				OUT_RELOC(ring, rsc->bo, slice->offset, 0, 0);
 			}
 

src/gallium/drivers/freedreno/a3xx/fd3_texture.c

@@ -144,7 +144,8 @@ fd3_sampler_view_create(struct pipe_context *pctx, struct pipe_resource *prsc,
 {
 	struct fd3_pipe_sampler_view *so = CALLOC_STRUCT(fd3_pipe_sampler_view);
 	struct fd_resource *rsc = fd_resource(prsc);
-	unsigned miplevels = cso->u.tex.last_level - cso->u.tex.first_level;
+	unsigned lvl = cso->u.tex.first_level;
+	unsigned miplevels = cso->u.tex.last_level - lvl;
 
 	if (!so)
 		return NULL;
@@ -156,7 +157,6 @@ fd3_sampler_view_create(struct pipe_context *pctx, struct pipe_resource *prsc,
 	so->base.context = pctx;
 
 	so->tex_resource =  rsc;
-	so->mipaddrs = 1 + miplevels;
 
 	so->texconst0 =
 			A3XX_TEX_CONST_0_TYPE(tex_type(prsc->target)) |
@@ -170,11 +170,11 @@ fd3_sampler_view_create(struct pipe_context *pctx, struct pipe_resource *prsc,
 
 	so->texconst1 =
 			A3XX_TEX_CONST_1_FETCHSIZE(fd3_pipe2fetchsize(cso->format)) |
-			A3XX_TEX_CONST_1_WIDTH(prsc->width0) |
-			A3XX_TEX_CONST_1_HEIGHT(prsc->height0);
+			A3XX_TEX_CONST_1_WIDTH(u_minify(prsc->width0, lvl)) |
+			A3XX_TEX_CONST_1_HEIGHT(u_minify(prsc->height0, lvl));
 	/* when emitted, A3XX_TEX_CONST_2_INDX() must be OR'd in: */
 	so->texconst2 =
-			A3XX_TEX_CONST_2_PITCH(rsc->slices[0].pitch * rsc->cpp);
+			A3XX_TEX_CONST_2_PITCH(rsc->slices[lvl].pitch * rsc->cpp);
 	so->texconst3 = 0x00000000;  /* ??? */
 
 	return &so->base;

src/gallium/drivers/freedreno/a3xx/fd3_texture.h

@@ -51,7 +51,6 @@ fd3_sampler_stateobj(struct pipe_sampler_state *samp)
 struct fd3_pipe_sampler_view {
 	struct pipe_sampler_view base;
 	struct fd_resource *tex_resource;
-	uint32_t mipaddrs;
 	uint32_t texconst0, texconst1, texconst2, texconst3;
 };
 

src/gallium/drivers/freedreno/freedreno_resource.c

@@ -304,7 +304,36 @@ fail:
 	return NULL;
 }
 
-static bool render_blit(struct pipe_context *pctx, struct pipe_blit_info *info);
+static void fd_blitter_pipe_begin(struct fd_context *ctx);
+static void fd_blitter_pipe_end(struct fd_context *ctx);
+
+/**
+ * _copy_region using pipe (3d engine)
+ */
+static bool
+fd_blitter_pipe_copy_region(struct fd_context *ctx,
+		struct pipe_resource *dst,
+		unsigned dst_level,
+		unsigned dstx, unsigned dsty, unsigned dstz,
+		struct pipe_resource *src,
+		unsigned src_level,
+		const struct pipe_box *src_box)
+{
+	/* not until we allow rendertargets to be buffers */
+	if (dst->target == PIPE_BUFFER || src->target == PIPE_BUFFER)
+		return false;
+
+	if (!util_blitter_is_copy_supported(ctx->blitter, dst, src))
+		return false;
+
+	fd_blitter_pipe_begin(ctx);
+	util_blitter_copy_texture(ctx->blitter,
+			dst, dst_level, dstx, dsty, dstz,
+			src, src_level, src_box);
+	fd_blitter_pipe_end(ctx);
+
+	return true;
+}
 
 /**
  * Copy a block of pixels from one resource to another.
@@ -320,40 +349,33 @@ fd_resource_copy_region(struct pipe_context *pctx,
 		unsigned src_level,
 		const struct pipe_box *src_box)
 {
+	struct fd_context *ctx = fd_context(pctx);
+
 	/* TODO if we have 2d core, or other DMA engine that could be used
 	 * for simple copies and reasonably easily synchronized with the 3d
 	 * core, this is where we'd plug it in..
 	 */
-	struct pipe_blit_info info = {
-		.dst = {
-			.resource = dst,
-			.box = {
-				.x      = dstx,
-				.y      = dsty,
-				.z      = dstz,
-				.width  = src_box->width,
-				.height = src_box->height,
-				.depth  = src_box->depth,
-			},
-			.format = util_format_linear(dst->format),
-		},
-		.src = {
-			.resource = src,
-			.box      = *src_box,
-			.format   = util_format_linear(src->format),
-		},
-		.mask = PIPE_MASK_RGBA,
-		.filter = PIPE_TEX_FILTER_NEAREST,
-	};
-	render_blit(pctx, &info);
+
+	/* try blit on 3d pipe: */
+	if (fd_blitter_pipe_copy_region(ctx,
+			dst, dst_level, dstx, dsty, dstz,
+			src, src_level, src_box))
+		return;
+
+	/* else fallback to pure sw: */
+	util_resource_copy_region(pctx,
+			dst, dst_level, dstx, dsty, dstz,
+			src, src_level, src_box);
 }
 
-/* Optimal hardware path for blitting pixels.
+/**
+ * Optimal hardware path for blitting pixels.
  * Scaling, format conversion, up- and downsampling (resolve) are allowed.
  */
 static void
 fd_blit(struct pipe_context *pctx, const struct pipe_blit_info *blit_info)
 {
+	struct fd_context *ctx = fd_context(pctx);
 	struct pipe_blit_info info = *blit_info;
 
 	if (info.src.resource->nr_samples > 1 &&
@@ -373,21 +395,21 @@ fd_blit(struct pipe_context *pctx, const struct pipe_blit_info *blit_info)
 		info.mask &= ~PIPE_MASK_S;
 	}
 
-	render_blit(pctx, &info);
-}
-
-static bool
-render_blit(struct pipe_context *pctx, struct pipe_blit_info *info)
-{
-	struct fd_context *ctx = fd_context(pctx);
-
-	if (!util_blitter_is_blit_supported(ctx->blitter, info)) {
+	if (!util_blitter_is_blit_supported(ctx->blitter, &info)) {
 		DBG("blit unsupported %s -> %s",
-				util_format_short_name(info->src.resource->format),
-				util_format_short_name(info->dst.resource->format));
-		return false;
+				util_format_short_name(info.src.resource->format),
+				util_format_short_name(info.dst.resource->format));
+		return;
 	}
 
+	fd_blitter_pipe_begin(ctx);
+	util_blitter_blit(ctx->blitter, &info);
+	fd_blitter_pipe_end(ctx);
+}
+
+static void
+fd_blitter_pipe_begin(struct fd_context *ctx)
+{
 	util_blitter_save_vertex_buffer_slot(ctx->blitter, ctx->vertexbuf.vb);
 	util_blitter_save_vertex_elements(ctx->blitter, ctx->vtx);
 	util_blitter_save_vertex_shader(ctx->blitter, ctx->prog.vp);
@@ -407,15 +429,21 @@ render_blit(struct pipe_context *pctx, struct pipe_blit_info *info)
 			ctx->fragtex.num_textures, ctx->fragtex.textures);
 
 	fd_hw_query_set_stage(ctx, ctx->ring, FD_STAGE_BLIT);
-	util_blitter_blit(ctx->blitter, info);
-	fd_hw_query_set_stage(ctx, ctx->ring, FD_STAGE_NULL);
-
-	return true;
 }
 
 static void
-fd_flush_resource(struct pipe_context *ctx, struct pipe_resource *resource)
+fd_blitter_pipe_end(struct fd_context *ctx)
 {
+	fd_hw_query_set_stage(ctx, ctx->ring, FD_STAGE_NULL);
+}
+
+static void
+fd_flush_resource(struct pipe_context *pctx, struct pipe_resource *prsc)
+{
+	struct fd_resource *rsc = fd_resource(prsc);
+
+	if (rsc->dirty)
+		fd_context_render(pctx);
 }
 
 void

src/gallium/drivers/freedreno/ir3/ir3_dump.c

@@ -322,7 +322,8 @@ static void ir3_block_dump(struct ir3_dump_ctx *ctx,
 
 	/* draw instruction graph: */
 	for (i = 0; i < block->noutputs; i++)
-		dump_instr(ctx, block->outputs[i]);
+		if (block->outputs[i])
+			dump_instr(ctx, block->outputs[i]);
 
 	/* draw outputs: */
 	fprintf(ctx->f, "output%lx [shape=record,label=\"outputs", PTRID(block));

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

@@ -174,15 +174,31 @@ NVC0LegalizePostRA::findOverwritingDefs(const Instruction *texi,
 }
 
 void
-NVC0LegalizePostRA::findFirstUses(const Instruction *texi,
-                                  const Instruction *insn,
-                                  std::list<TexUse> &uses)
+NVC0LegalizePostRA::findFirstUses(
+      const Instruction *texi,
+      const Instruction *insn,
+      std::list<TexUse> &uses,
+      std::tr1::unordered_set<const Instruction *>& visited)
 {
    for (int d = 0; insn->defExists(d); ++d) {
       Value *v = insn->getDef(d);
       for (Value::UseIterator u = v->uses.begin(); u != v->uses.end(); ++u) {
          Instruction *usei = (*u)->getInsn();
 
+         /* XXX HACK ALERT XXX
+          *
+          * This shouldn't have to be here, we should always be making forward
+          * progress by looking at the uses. However this somehow does not
+          * appear to be the case. Probably because this is being done right
+          * after RA, when the defs/uses lists have been messed with by node
+          * merging. This should probably be moved to being done right before
+          * RA. But this will do for now.
+          */
+         if (visited.find(usei) != visited.end())
+            continue;
+
+         visited.insert(usei);
+
          if (usei->op == OP_PHI || usei->op == OP_UNION) {
             // need a barrier before WAW cases
             for (int s = 0; usei->srcExists(s); ++s) {
@@ -197,11 +213,11 @@ NVC0LegalizePostRA::findFirstUses(const Instruction *texi,
              usei->op == OP_PHI ||
              usei->op == OP_UNION) {
             // these uses don't manifest in the machine code
-            findFirstUses(texi, usei, uses);
+            findFirstUses(texi, usei, uses, visited);
          } else
          if (usei->op == OP_MOV && usei->getDef(0)->equals(usei->getSrc(0)) &&
              usei->subOp != NV50_IR_SUBOP_MOV_FINAL) {
-            findFirstUses(texi, usei, uses);
+            findFirstUses(texi, usei, uses, visited);
          } else {
             addTexUse(uses, usei, insn);
          }
@@ -257,8 +273,10 @@ NVC0LegalizePostRA::insertTextureBarriers(Function *fn)
    uses = new std::list<TexUse>[texes.size()];
    if (!uses)
       return false;
-   for (size_t i = 0; i < texes.size(); ++i)
-      findFirstUses(texes[i], texes[i], uses[i]);
+   for (size_t i = 0; i < texes.size(); ++i) {
+      std::tr1::unordered_set<const Instruction *> visited;
+      findFirstUses(texes[i], texes[i], uses[i], visited);
+   }
 
    // determine the barrier level at each use
    for (size_t i = 0; i < texes.size(); ++i) {

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

@@ -20,6 +20,8 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  */
 
+#include <tr1/unordered_set>
+
 #include "codegen/nv50_ir.h"
 #include "codegen/nv50_ir_build_util.h"
 
@@ -69,7 +71,8 @@ private:
    bool insertTextureBarriers(Function *);
    inline bool insnDominatedBy(const Instruction *, const Instruction *) const;
    void findFirstUses(const Instruction *tex, const Instruction *def,
-                      std::list<TexUse>&);
+                      std::list<TexUse>&,
+                      std::tr1::unordered_set<const Instruction *>&);
    void findOverwritingDefs(const Instruction *tex, Instruction *insn,
                             const BasicBlock *term,
                             std::list<TexUse>&);

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

@@ -567,6 +567,10 @@ ConstantFolding::expr(Instruction *i,
       ImmediateValue src0;
       if (i->src(0).getImmediate(src0))
          expr(i, src0, *i->getSrc(1)->asImm());
+      if (i->saturate && !prog->getTarget()->isSatSupported(i)) {
+         bld.setPosition(i, false);
+         i->setSrc(1, bld.loadImm(NULL, res.data.u32));
+      }
    } else {
       i->op = i->saturate ? OP_SAT : OP_MOV; /* SAT handled by unary() */
    }

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

@@ -585,9 +585,12 @@ nv50_stage_sampler_states_bind(struct nv50_context *nv50, int s,
          nv50_screen_tsc_unlock(nv50->screen, old);
    }
    assert(nv50->num_samplers[s] <= PIPE_MAX_SAMPLERS);
-   for (; i < nv50->num_samplers[s]; ++i)
-      if (nv50->samplers[s][i])
+   for (; i < nv50->num_samplers[s]; ++i) {
+      if (nv50->samplers[s][i]) {
          nv50_screen_tsc_unlock(nv50->screen, nv50->samplers[s][i]);
+         nv50->samplers[s][i] = NULL;
+      }
+   }
 
    nv50->num_samplers[s] = nr;
 

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

@@ -54,8 +54,8 @@ nv50_validate_fb(struct nv50_context *nv50)
       assert(mt->layout_3d || !array_mode || array_size == 1);
 
       BEGIN_NV04(push, NV50_3D(RT_ADDRESS_HIGH(i)), 5);
-      PUSH_DATAh(push, bo->offset + sf->offset);
-      PUSH_DATA (push, bo->offset + sf->offset);
+      PUSH_DATAh(push, mt->base.address + sf->offset);
+      PUSH_DATA (push, mt->base.address + sf->offset);
       PUSH_DATA (push, nv50_format_table[sf->base.format].rt);
       if (likely(nouveau_bo_memtype(bo))) {
          PUSH_DATA (push, mt->level[sf->base.u.tex.level].tile_mode);
@@ -97,8 +97,8 @@ nv50_validate_fb(struct nv50_context *nv50)
       int unk = mt->base.base.target == PIPE_TEXTURE_3D || sf->depth == 1;
 
       BEGIN_NV04(push, NV50_3D(ZETA_ADDRESS_HIGH), 5);
-      PUSH_DATAh(push, bo->offset + sf->offset);
-      PUSH_DATA (push, bo->offset + sf->offset);
+      PUSH_DATAh(push, mt->base.address + sf->offset);
+      PUSH_DATA (push, mt->base.address + sf->offset);
       PUSH_DATA (push, nv50_format_table[fb->zsbuf->format].rt);
       PUSH_DATA (push, mt->level[sf->base.u.tex.level].tile_mode);
       PUSH_DATA (push, mt->layer_stride >> 2);

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

@@ -114,8 +114,8 @@ nv50_2d_texture_set(struct nouveau_pushbuf *push, int dst,
       PUSH_DATA (push, mt->level[level].pitch);
       PUSH_DATA (push, width);
       PUSH_DATA (push, height);
-      PUSH_DATAh(push, bo->offset + offset);
-      PUSH_DATA (push, bo->offset + offset);
+      PUSH_DATAh(push, mt->base.address + offset);
+      PUSH_DATA (push, mt->base.address + offset);
    } else {
       BEGIN_NV04(push, SUBC_2D(mthd), 5);
       PUSH_DATA (push, format);
@@ -126,8 +126,8 @@ nv50_2d_texture_set(struct nouveau_pushbuf *push, int dst,
       BEGIN_NV04(push, SUBC_2D(mthd + 0x18), 4);
       PUSH_DATA (push, width);
       PUSH_DATA (push, height);
-      PUSH_DATAh(push, bo->offset + offset);
-      PUSH_DATA (push, bo->offset + offset);
+      PUSH_DATAh(push, mt->base.address + offset);
+      PUSH_DATA (push, mt->base.address + offset);
    }
 
 #if 0
@@ -299,8 +299,8 @@ nv50_clear_render_target(struct pipe_context *pipe,
    BEGIN_NV04(push, NV50_3D(RT_CONTROL), 1);
    PUSH_DATA (push, 1);
    BEGIN_NV04(push, NV50_3D(RT_ADDRESS_HIGH(0)), 5);
-   PUSH_DATAh(push, bo->offset + sf->offset);
-   PUSH_DATA (push, bo->offset + sf->offset);
+   PUSH_DATAh(push, mt->base.address + sf->offset);
+   PUSH_DATA (push, mt->base.address + sf->offset);
    PUSH_DATA (push, nv50_format_table[dst->format].rt);
    PUSH_DATA (push, mt->level[sf->base.u.tex.level].tile_mode);
    PUSH_DATA (push, mt->layer_stride >> 2);
@@ -381,8 +381,8 @@ nv50_clear_depth_stencil(struct pipe_context *pipe,
    nv50->scissors_dirty |= 1;
 
    BEGIN_NV04(push, NV50_3D(ZETA_ADDRESS_HIGH), 5);
-   PUSH_DATAh(push, bo->offset + sf->offset);
-   PUSH_DATA (push, bo->offset + sf->offset);
+   PUSH_DATAh(push, mt->base.address + sf->offset);
+   PUSH_DATA (push, mt->base.address + sf->offset);
    PUSH_DATA (push, nv50_format_table[dst->format].rt);
    PUSH_DATA (push, mt->level[sf->base.u.tex.level].tile_mode);
    PUSH_DATA (push, mt->layer_stride >> 2);

src/gallium/drivers/nouveau/nv50/nv50_transfer.c

@@ -24,6 +24,8 @@ nv50_m2mf_rect_setup(struct nv50_m2mf_rect *rect,
    rect->bo = mt->base.bo;
    rect->domain = mt->base.domain;
    rect->base = mt->level[l].offset;
+   if (mt->base.bo->offset != mt->base.address)
+      rect->base += mt->base.address - mt->base.bo->offset;
    rect->pitch = mt->level[l].pitch;
    if (util_format_is_plain(res->format)) {
       rect->width = w << mt->ms_x;

src/gallium/drivers/nouveau/nv50/nv84_video.c

@@ -482,12 +482,14 @@ nv84_create_decoder(struct pipe_context *context,
       mip.level[0].pitch = surf.width * 4;
       mip.base.domain = NOUVEAU_BO_VRAM;
       mip.base.bo = dec->mbring;
+      mip.base.address = dec->mbring->offset;
       context->clear_render_target(context, &surf.base, &color, 0, 0, 64, 4760);
       surf.offset = dec->vpring->size / 2 - 0x1000;
       surf.width = 1024;
       surf.height = 1;
       mip.level[0].pitch = surf.width * 4;
       mip.base.bo = dec->vpring;
+      mip.base.address = dec->vpring->offset;
       context->clear_render_target(context, &surf.base, &color, 0, 0, 1024, 1);
       surf.offset = dec->vpring->size - 0x1000;
       context->clear_render_target(context, &surf.base, &color, 0, 0, 1024, 1);
@@ -683,17 +685,14 @@ nv84_video_buffer_create(struct pipe_context *pipe,
                       bo_size, &cfg, &buffer->full))
       goto error;
 
-   mt0->base.bo = buffer->interlaced;
+   nouveau_bo_ref(buffer->interlaced, &mt0->base.bo);
    mt0->base.domain = NOUVEAU_BO_VRAM;
-   mt0->base.offset = 0;
-   mt0->base.address = buffer->interlaced->offset + mt0->base.offset;
-   nouveau_bo_ref(buffer->interlaced, &empty);
+   mt0->base.address = buffer->interlaced->offset;
 
-   mt1->base.bo = buffer->interlaced;
+   nouveau_bo_ref(buffer->interlaced, &mt1->base.bo);
    mt1->base.domain = NOUVEAU_BO_VRAM;
-   mt1->base.offset = mt0->layer_stride * 2;
-   mt1->base.address = buffer->interlaced->offset + mt1->base.offset;
-   nouveau_bo_ref(buffer->interlaced, &empty);
+   mt1->base.offset = mt0->total_size;
+   mt1->base.address = buffer->interlaced->offset + mt0->total_size;
 
    memset(&sv_templ, 0, sizeof(sv_templ));
    for (component = 0, i = 0; i < 2; ++i ) {

src/gallium/drivers/nouveau/nvc0/nvc0_miptree.c

@@ -261,7 +261,6 @@ nvc0_miptree_create(struct pipe_screen *pscreen,
 
    if (pt->usage == PIPE_USAGE_STAGING) {
       switch (pt->target) {
-      case PIPE_TEXTURE_1D:
       case PIPE_TEXTURE_2D:
       case PIPE_TEXTURE_RECT:
          if (pt->last_level == 0 &&

src/gallium/drivers/r600/r600_blit.c

@@ -440,7 +440,8 @@ static void r600_clear(struct pipe_context *ctx, unsigned buffers,
 	}
 
 	r600_blitter_begin(ctx, R600_CLEAR);
-	util_blitter_clear(rctx->blitter, fb->width, fb->height, 1,
+	util_blitter_clear(rctx->blitter, fb->width, fb->height,
+			   util_framebuffer_get_num_layers(fb),
 			   buffers, color, depth, stencil);
 	r600_blitter_end(ctx);
 

src/gallium/drivers/r600/r600_state_common.c

@@ -1245,12 +1245,6 @@ static bool r600_update_derived_state(struct r600_context *rctx)
 			}
 		}
 
-		if (rctx->b.chip_class >= EVERGREEN) {
-			evergreen_update_db_shader_control(rctx);
-		} else {
-			r600_update_db_shader_control(rctx);
-		}
-
 		if (unlikely(!ps_dirty && rctx->ps_shader && rctx->rasterizer &&
 				((rctx->rasterizer->sprite_coord_enable != rctx->ps_shader->current->sprite_coord_enable) ||
 						(rctx->rasterizer->flatshade != rctx->ps_shader->current->flatshade)))) {
@@ -1264,6 +1258,12 @@ static bool r600_update_derived_state(struct r600_context *rctx)
 		update_shader_atom(ctx, &rctx->pixel_shader, rctx->ps_shader->current);
 	}
 
+	if (rctx->b.chip_class >= EVERGREEN) {
+		evergreen_update_db_shader_control(rctx);
+	} else {
+		r600_update_db_shader_control(rctx);
+	}
+
 	/* on R600 we stuff masks + txq info into one constant buffer */
 	/* on evergreen we only need a txq info one */
 	if (rctx->b.chip_class < EVERGREEN) {

src/gallium/drivers/radeon/r600_query.c

@@ -807,12 +807,40 @@ void r600_suspend_nontimer_queries(struct r600_common_context *ctx)
 	assert(ctx->num_cs_dw_nontimer_queries_suspend == 0);
 }
 
+static unsigned r600_queries_num_cs_dw_for_resuming(struct r600_common_context *ctx)
+{
+	struct r600_query *query;
+	unsigned num_dw = 0;
+
+	LIST_FOR_EACH_ENTRY(query, &ctx->active_nontimer_queries, list) {
+		/* begin + end */
+		num_dw += query->num_cs_dw * 2;
+
+		/* Workaround for the fact that
+		 * num_cs_dw_nontimer_queries_suspend is incremented for every
+		 * resumed query, which raises the bar in need_cs_space for
+		 * queries about to be resumed.
+		 */
+		num_dw += query->num_cs_dw;
+	}
+	/* primitives generated query */
+	num_dw += ctx->streamout.enable_atom.num_dw;
+	/* guess for ZPASS enable or PERFECT_ZPASS_COUNT enable updates */
+	num_dw += 13;
+
+	return num_dw;
+}
+
 void r600_resume_nontimer_queries(struct r600_common_context *ctx)
 {
 	struct r600_query *query;
 
 	assert(ctx->num_cs_dw_nontimer_queries_suspend == 0);
 
+	/* Check CS space here. Resuming must not be interrupted by flushes. */
+	ctx->need_gfx_cs_space(&ctx->b,
+			       r600_queries_num_cs_dw_for_resuming(ctx), TRUE);
+
 	LIST_FOR_EACH_ENTRY(query, &ctx->active_nontimer_queries, list) {
 		r600_emit_query_begin(ctx, query);
 	}

src/gallium/drivers/radeon/radeon_video.c

@@ -251,8 +251,11 @@ int rvid_get_video_param(struct pipe_screen *screen,
 			       profile != PIPE_VIDEO_PROFILE_VC1_MAIN;
 		case PIPE_VIDEO_CAP_PREFERS_INTERLACED:
 		case PIPE_VIDEO_CAP_SUPPORTS_INTERLACED:
-			/* and MPEG2 only with shaders */
-			return codec != PIPE_VIDEO_FORMAT_MPEG12;
+			/* MPEG2 only with shaders and no support for
+			   interlacing on R6xx style UVD */
+			return codec != PIPE_VIDEO_FORMAT_MPEG12 &&
+			       /* TODO: RV770 might actually work */
+			       rscreen->family > CHIP_RV770;
 		default:
 			break;
 		}

src/gallium/state_trackers/dri/dri2.c

@@ -1328,6 +1328,7 @@ dri_kms_init_screen(__DRIscreen * sPriv)
    const __DRIconfig **configs;
    struct dri_screen *screen;
    struct pipe_screen *pscreen = NULL;
+   uint64_t cap;
 
    screen = CALLOC_STRUCT(dri_screen);
    if (!screen)
@@ -1339,6 +1340,13 @@ dri_kms_init_screen(__DRIscreen * sPriv)
    sPriv->driverPrivate = (void *)screen;
 
    pscreen = kms_swrast_create_screen(screen->fd);
+
+   if (drmGetCap(sPriv->fd, DRM_CAP_PRIME, &cap) == 0 &&
+          (cap & DRM_PRIME_CAP_IMPORT)) {
+      dri2ImageExtension.createImageFromFds = dri2_from_fds;
+      dri2ImageExtension.createImageFromDmaBufs = dri2_from_dma_bufs;
+   }
+
    sPriv->extensions = dri_screen_extensions;
 
    /* dri_init_screen_helper checks pscreen for us */

src/gallium/state_trackers/vdpau/device.c

@@ -42,6 +42,8 @@ vdp_imp_device_create_x11(Display *display, int screen, VdpDevice *device,
                           VdpGetProcAddress **get_proc_address)
 {
    struct pipe_screen *pscreen;
+   struct pipe_resource *res, res_tmpl;
+   struct pipe_sampler_view sv_tmpl;
    vlVdpDevice *dev = NULL;
    VdpStatus ret;
 
@@ -79,6 +81,43 @@ vdp_imp_device_create_x11(Display *display, int screen, VdpDevice *device,
       goto no_context;
    }
 
+   memset(&res_tmpl, 0, sizeof(res_tmpl));
+
+   res_tmpl.target = PIPE_TEXTURE_2D;
+   res_tmpl.format = PIPE_FORMAT_R8G8B8A8_UNORM;
+   res_tmpl.width0 = 1;
+   res_tmpl.height0 = 1;
+   res_tmpl.depth0 = 1;
+   res_tmpl.array_size = 1;
+   res_tmpl.bind = PIPE_BIND_SAMPLER_VIEW;
+   res_tmpl.usage = PIPE_USAGE_DEFAULT;
+
+   if (!CheckSurfaceParams(pscreen, &res_tmpl)) {
+      ret = VDP_STATUS_NO_IMPLEMENTATION;
+      goto no_resource;
+   }
+
+   res = pscreen->resource_create(pscreen, &res_tmpl);
+   if (!res) {
+      ret = VDP_STATUS_RESOURCES;
+      goto no_resource;
+   }
+
+   memset(&sv_tmpl, 0, sizeof(sv_tmpl));
+   u_sampler_view_default_template(&sv_tmpl, res, res->format);
+
+   sv_tmpl.swizzle_r = PIPE_SWIZZLE_ONE;
+   sv_tmpl.swizzle_g = PIPE_SWIZZLE_ONE;
+   sv_tmpl.swizzle_b = PIPE_SWIZZLE_ONE;
+   sv_tmpl.swizzle_a = PIPE_SWIZZLE_ONE;
+
+   dev->dummy_sv = dev->context->create_sampler_view(dev->context, res, &sv_tmpl);
+   pipe_resource_reference(&res, NULL);
+   if (!dev->dummy_sv) {
+      ret = VDP_STATUS_RESOURCES;
+      goto no_resource;
+   }
+
    *device = vlAddDataHTAB(dev);
    if (*device == 0) {
       ret = VDP_STATUS_ERROR;
@@ -93,8 +132,9 @@ vdp_imp_device_create_x11(Display *display, int screen, VdpDevice *device,
    return VDP_STATUS_OK;
 
 no_handle:
+   pipe_sampler_view_reference(&dev->dummy_sv, NULL);
+no_resource:
    dev->context->destroy(dev->context);
-   /* Destroy vscreen */
 no_context:
    vl_screen_destroy(dev->vscreen);
 no_vscreen:
@@ -185,6 +225,7 @@ vlVdpDeviceFree(vlVdpDevice *dev)
 {
    pipe_mutex_destroy(dev->mutex);
    vl_compositor_cleanup(&dev->compositor);
+   pipe_sampler_view_reference(&dev->dummy_sv, NULL);
    dev->context->destroy(dev->context);
    vl_screen_destroy(dev->vscreen);
    FREE(dev);

src/gallium/state_trackers/vdpau/output.c

@@ -624,9 +624,9 @@ vlVdpOutputSurfaceRenderOutputSurface(VdpOutputSurface destination_surface,
                                       uint32_t flags)
 {
    vlVdpOutputSurface *dst_vlsurface;
-   vlVdpOutputSurface *src_vlsurface;
 
    struct pipe_context *context;
+   struct pipe_sampler_view *src_sv;
    struct vl_compositor *compositor;
    struct vl_compositor_state *cstate;
 
@@ -639,12 +639,19 @@ vlVdpOutputSurfaceRenderOutputSurface(VdpOutputSurface destination_surface,
    if (!dst_vlsurface)
       return VDP_STATUS_INVALID_HANDLE;
 
-   src_vlsurface = vlGetDataHTAB(source_surface);
-   if (!src_vlsurface)
-      return VDP_STATUS_INVALID_HANDLE;
+   if (source_surface == VDP_INVALID_HANDLE) {
+      src_sv = dst_vlsurface->device->dummy_sv;
 
-   if (dst_vlsurface->device != src_vlsurface->device)
-      return VDP_STATUS_HANDLE_DEVICE_MISMATCH;
+   } else {
+      vlVdpOutputSurface *src_vlsurface = vlGetDataHTAB(source_surface);
+      if (!src_vlsurface)
+         return VDP_STATUS_INVALID_HANDLE;
+
+      if (dst_vlsurface->device != src_vlsurface->device)
+         return VDP_STATUS_HANDLE_DEVICE_MISMATCH;
+
+      src_sv = src_vlsurface->sampler_view;
+   }
 
    pipe_mutex_lock(dst_vlsurface->device->mutex);
    vlVdpResolveDelayedRendering(dst_vlsurface->device, NULL, NULL);
@@ -657,7 +664,7 @@ vlVdpOutputSurfaceRenderOutputSurface(VdpOutputSurface destination_surface,
 
    vl_compositor_clear_layers(cstate);
    vl_compositor_set_layer_blend(cstate, 0, blend, false);
-   vl_compositor_set_rgba_layer(cstate, compositor, 0, src_vlsurface->sampler_view,
+   vl_compositor_set_rgba_layer(cstate, compositor, 0, src_sv,
                                 RectToPipe(source_rect, &src_rect), NULL,
                                 ColorsToPipe(colors, flags, vlcolors));
    STATIC_ASSERT(VL_COMPOSITOR_ROTATE_0 == VDP_OUTPUT_SURFACE_RENDER_ROTATE_0);
@@ -688,9 +695,9 @@ vlVdpOutputSurfaceRenderBitmapSurface(VdpOutputSurface destination_surface,
                                       uint32_t flags)
 {
    vlVdpOutputSurface *dst_vlsurface;
-   vlVdpBitmapSurface *src_vlsurface;
 
    struct pipe_context *context;
+   struct pipe_sampler_view *src_sv;
    struct vl_compositor *compositor;
    struct vl_compositor_state *cstate;
 
@@ -703,12 +710,19 @@ vlVdpOutputSurfaceRenderBitmapSurface(VdpOutputSurface destination_surface,
    if (!dst_vlsurface)
       return VDP_STATUS_INVALID_HANDLE;
 
-   src_vlsurface = vlGetDataHTAB(source_surface);
-   if (!src_vlsurface)
-      return VDP_STATUS_INVALID_HANDLE;
+   if (source_surface == VDP_INVALID_HANDLE) {
+      src_sv = dst_vlsurface->device->dummy_sv;
 
-   if (dst_vlsurface->device != src_vlsurface->device)
-      return VDP_STATUS_HANDLE_DEVICE_MISMATCH;
+   } else {
+      vlVdpBitmapSurface *src_vlsurface = vlGetDataHTAB(source_surface);
+      if (!src_vlsurface)
+         return VDP_STATUS_INVALID_HANDLE;
+
+      if (dst_vlsurface->device != src_vlsurface->device)
+         return VDP_STATUS_HANDLE_DEVICE_MISMATCH;
+
+      src_sv = src_vlsurface->sampler_view;
+   }
 
    context = dst_vlsurface->device->context;
    compositor = &dst_vlsurface->device->compositor;
@@ -721,7 +735,7 @@ vlVdpOutputSurfaceRenderBitmapSurface(VdpOutputSurface destination_surface,
 
    vl_compositor_clear_layers(cstate);
    vl_compositor_set_layer_blend(cstate, 0, blend, false);
-   vl_compositor_set_rgba_layer(cstate, compositor, 0, src_vlsurface->sampler_view,
+   vl_compositor_set_rgba_layer(cstate, compositor, 0, src_sv,
                                 RectToPipe(source_rect, &src_rect), NULL,
                                 ColorsToPipe(colors, flags, vlcolors));
    vl_compositor_set_layer_rotation(cstate, 0, flags & 3);

src/gallium/state_trackers/vdpau/vdpau_private.h

@@ -348,6 +348,7 @@ typedef struct
    struct vl_screen *vscreen;
    struct pipe_context *context;
    struct vl_compositor compositor;
+   struct pipe_sampler_view *dummy_sv;
    pipe_mutex mutex;
 
    struct {

src/gallium/targets/dri/Makefile.am

@@ -26,7 +26,6 @@ gallium_dri_la_LDFLAGS = \
 	-shrext .so \
 	-module \
 	-avoid-version \
-	-Wl,--dynamic-list=$(top_srcdir)/src/gallium/targets/dri-vdpau.dyn \
 	$(GC_SECTIONS)
 
 if HAVE_LD_VERSION_SCRIPT
@@ -34,6 +33,11 @@ gallium_dri_la_LDFLAGS += \
 	-Wl,--version-script=$(top_srcdir)/src/gallium/targets/dri/dri.sym
 endif # HAVE_LD_VERSION_SCRIPT
 
+if HAVE_LD_DYNAMIC_LIST
+gallium_dri_la_LDFLAGS += \
+	-Wl,--dynamic-list=$(top_srcdir)/src/gallium/targets/dri-vdpau.dyn
+endif # HAVE_LD_DYNAMIC_LIST
+
 gallium_dri_la_LIBADD = \
 	$(top_builddir)/src/mesa/libmesagallium.la \
 	$(top_builddir)/src/mesa/drivers/dri/common/libdricommon.la \

src/gallium/targets/vdpau/Makefile.am

@@ -15,7 +15,6 @@ libvdpau_gallium_la_LDFLAGS = \
 	-module \
 	-no-undefined \
 	-version-number $(VDPAU_MAJOR):$(VDPAU_MINOR) \
-	-Wl,--dynamic-list=$(top_srcdir)/src/gallium/targets/dri-vdpau.dyn \
 	$(GC_SECTIONS) \
 	$(LD_NO_UNDEFINED)
 
@@ -24,6 +23,11 @@ libvdpau_gallium_la_LDFLAGS += \
 	-Wl,--version-script=$(top_srcdir)/src/gallium/targets/vdpau/vdpau.sym
 endif # HAVE_LD_VERSION_SCRIPT
 
+if HAVE_LD_DYNAMIC_LIST
+libvdpau_gallium_la_LDFLAGS += \
+	-Wl,--dynamic-list=$(top_srcdir)/src/gallium/targets/dri-vdpau.dyn
+endif # HAVE_LD_DYNAMIC_LIST
+
 libvdpau_gallium_la_LIBADD = \
 	$(top_builddir)/src/gallium/state_trackers/vdpau/libvdpautracker.la \
 	$(top_builddir)/src/gallium/auxiliary/libgallium.la \

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

@@ -238,7 +238,7 @@ out_mip:
 
 static struct svga_winsys_surface *
 vmw_drm_surface_from_handle(struct svga_winsys_screen *sws,
-			    struct winsys_handle *whandle,
+                            struct winsys_handle *whandle,
 			    SVGA3dSurfaceFormat *format)
 {
     struct vmw_svga_winsys_surface *vsrf;
@@ -248,7 +248,8 @@ vmw_drm_surface_from_handle(struct svga_winsys_screen *sws,
     struct drm_vmw_surface_arg *req = &arg.req;
     struct drm_vmw_surface_create_req *rep = &arg.rep;
     uint32_t handle = 0;
-    SVGA3dSize size;
+    struct drm_vmw_size size;
+    SVGA3dSize base_size;
     int ret;
     int i;
 
@@ -274,7 +275,7 @@ vmw_drm_surface_from_handle(struct svga_winsys_screen *sws,
 
     memset(&arg, 0, sizeof(arg));
     req->sid = handle;
-    rep->size_addr = (size_t)&size;
+    rep->size_addr = (unsigned long)&size;
 
     ret = drmCommandWriteRead(vws->ioctl.drm_fd, DRM_VMW_REF_SURFACE,
 			      &arg, sizeof(arg));
@@ -324,7 +325,11 @@ vmw_drm_surface_from_handle(struct svga_winsys_screen *sws,
     *format = rep->format;
 
     /* Estimate usage, for early flushing. */
-    vsrf->size = svga3dsurface_get_serialized_size(rep->format, size,
+
+    base_size.width = size.width;
+    base_size.height = size.height;
+    base_size.depth = size.depth;
+    vsrf->size = svga3dsurface_get_serialized_size(rep->format, base_size,
                                                    rep->mip_levels[0],
                                                    FALSE);
 

src/gallium/winsys/sw/kms-dri/kms_dri_sw_winsys.c

@@ -38,6 +38,7 @@
 #include <sys/mman.h>
 #include <unistd.h>
 #include <dlfcn.h>
+#include <fcntl.h>
 #include <xf86drm.h>
 
 #include "pipe/p_compiler.h"
@@ -121,7 +122,7 @@ kms_sw_displaytarget_create(struct sw_winsys *ws,
    int ret;
 
    kms_sw_dt = CALLOC_STRUCT(kms_sw_displaytarget);
-   if(!kms_sw_dt)
+   if (!kms_sw_dt)
       goto no_dt;
 
    kms_sw_dt->ref_count = 1;
@@ -210,6 +211,38 @@ kms_sw_displaytarget_map(struct sw_winsys *ws,
    return kms_sw_dt->mapped;
 }
 
+static struct kms_sw_displaytarget *
+kms_sw_displaytarget_add_from_prime(struct kms_sw_winsys *kms_sw, int fd)
+{
+   uint32_t handle = -1;
+   struct kms_sw_displaytarget * kms_sw_dt;
+   int ret;
+
+   ret = drmPrimeFDToHandle(kms_sw->fd, fd, &handle);
+
+   if (ret)
+      return NULL;
+
+   kms_sw_dt = CALLOC_STRUCT(kms_sw_displaytarget);
+   if (!kms_sw_dt)
+      return NULL;
+
+   kms_sw_dt->ref_count = 1;
+   kms_sw_dt->handle = handle;
+   kms_sw_dt->size = lseek(fd, 0, SEEK_END);
+
+   if (kms_sw_dt->size == (off_t)-1) {
+      FREE(kms_sw_dt);
+      return NULL;
+   }
+
+   lseek(fd, 0, SEEK_SET);
+
+   list_add(&kms_sw_dt->link, &kms_sw->bo_list);
+
+   return kms_sw_dt;
+}
+
 static void
 kms_sw_displaytarget_unmap(struct sw_winsys *ws,
                            struct sw_displaytarget *dt)
@@ -231,17 +264,34 @@ kms_sw_displaytarget_from_handle(struct sw_winsys *ws,
    struct kms_sw_winsys *kms_sw = kms_sw_winsys(ws);
    struct kms_sw_displaytarget *kms_sw_dt;
 
-   assert(whandle->type == DRM_API_HANDLE_TYPE_KMS);
+   assert(whandle->type == DRM_API_HANDLE_TYPE_KMS ||
+          whandle->type == DRM_API_HANDLE_TYPE_FD);
 
-   LIST_FOR_EACH_ENTRY(kms_sw_dt, &kms_sw->bo_list, link) {
-      if (kms_sw_dt->handle == whandle->handle) {
+   switch(whandle->type) {
+   case DRM_API_HANDLE_TYPE_FD:
+      kms_sw_dt = kms_sw_displaytarget_add_from_prime(kms_sw, whandle->handle);
+      if (kms_sw_dt) {
          kms_sw_dt->ref_count++;
-
-         DEBUG("KMS-DEBUG: imported buffer %u (size %u)\n", kms_sw_dt->handle, kms_sw_dt->size);
-
+         kms_sw_dt->width = templ->width0;
+         kms_sw_dt->height = templ->height0;
+         kms_sw_dt->stride = whandle->stride;
          *stride = kms_sw_dt->stride;
-         return (struct sw_displaytarget *)kms_sw_dt;
       }
+      return (struct sw_displaytarget *)kms_sw_dt;
+   case DRM_API_HANDLE_TYPE_KMS:
+      LIST_FOR_EACH_ENTRY(kms_sw_dt, &kms_sw->bo_list, link) {
+         if (kms_sw_dt->handle == whandle->handle) {
+            kms_sw_dt->ref_count++;
+
+            DEBUG("KMS-DEBUG: imported buffer %u (size %u)\n", kms_sw_dt->handle, kms_sw_dt->size);
+
+            *stride = kms_sw_dt->stride;
+            return (struct sw_displaytarget *)kms_sw_dt;
+         }
+      }
+      /* fallthrough */
+   default:
+      break;
    }
 
    assert(0);
@@ -253,16 +303,26 @@ kms_sw_displaytarget_get_handle(struct sw_winsys *winsys,
                                 struct sw_displaytarget *dt,
                                 struct winsys_handle *whandle)
 {
+   struct kms_sw_winsys *kms_sw = kms_sw_winsys(winsys);
    struct kms_sw_displaytarget *kms_sw_dt = kms_sw_displaytarget(dt);
 
-   if (whandle->type == DRM_API_HANDLE_TYPE_KMS) {
+   switch(whandle->type) {
+   case DRM_API_HANDLE_TYPE_KMS:
       whandle->handle = kms_sw_dt->handle;
       whandle->stride = kms_sw_dt->stride;
-   } else {
+      return TRUE;
+   case DRM_API_HANDLE_TYPE_FD:
+      if (!drmPrimeHandleToFD(kms_sw->fd, kms_sw_dt->handle,
+                             DRM_CLOEXEC, &whandle->handle)) {
+         whandle->stride = kms_sw_dt->stride;
+         return TRUE;
+      }
+      /* fallthrough */
+   default:
       whandle->handle = 0;
       whandle->stride = 0;
+      return FALSE;
    }
-   return TRUE;
 }
 
 static void
@@ -315,4 +375,4 @@ kms_dri_create_winsys(int fd)
    return &ws->base;
 }
 
-/* vim: set sw=3 ts=8 sts=3 expandtab: */
+/* vim: set sw=3 ts=8 sts=3 expandtab: */

src/glsl/glcpp/glcpp-lex.l

@@ -289,8 +289,14 @@ HEXADECIMAL_INTEGER	0[xX][0-9a-fA-F]+[uU]?
 }
 
 	/* Swallow empty #pragma directives, (to avoid confusing the
-	 * downstream compiler). */
-<HASH>pragma{HSPACE}*/{NEWLINE} {
+	 * downstream compiler).
+	 *
+	 * Note: We use a simple regular expression for the lookahead
+	 * here. Specifically, we cannot use the complete {NEWLINE} expression
+	 * since it uses alternation and we've found that there's a flex bug
+	 * where using alternation in the lookahead portion of a pattern
+	 * triggers a buffer overrun. */
+<HASH>pragma{HSPACE}*/[\r\n] {
 	BEGIN INITIAL;
 }
 

src/glsl/lower_variable_index_to_cond_assign.cpp

@@ -76,7 +76,7 @@ compare_index_block(exec_list *instructions, ir_variable *index,
    ir_rvalue *broadcast_index = new(mem_ctx) ir_dereference_variable(index);
 
    assert(index->type->is_scalar());
-   assert(index->type->base_type == GLSL_TYPE_INT);
+   assert(index->type->base_type == GLSL_TYPE_INT || index->type->base_type == GLSL_TYPE_UINT);
    assert(components >= 1 && components <= 4);
 
    if (components > 1) {

src/mesa/drivers/common/meta.c

@@ -396,25 +396,6 @@ _mesa_meta_init(struct gl_context *ctx)
    ctx->Meta = CALLOC_STRUCT(gl_meta_state);
 }
 
-static GLenum
-gl_buffer_index_to_drawbuffers_enum(gl_buffer_index bufindex)
-{
-   assert(bufindex < BUFFER_COUNT);
-
-   if (bufindex >= BUFFER_COLOR0)
-      return GL_COLOR_ATTACHMENT0 + bufindex - BUFFER_COLOR0;
-   else if (bufindex == BUFFER_FRONT_LEFT)
-      return GL_FRONT_LEFT;
-   else if (bufindex == BUFFER_FRONT_RIGHT)
-      return GL_FRONT_RIGHT;
-   else if (bufindex == BUFFER_BACK_LEFT)
-      return GL_BACK_LEFT;
-   else if (bufindex == BUFFER_BACK_RIGHT)
-      return GL_BACK_RIGHT;
-
-   return GL_NONE;
-}
-
 /**
  * Free context meta-op state.
  * To be called once during context destruction.
@@ -806,20 +787,9 @@ _mesa_meta_begin(struct gl_context *ctx, GLbitfield state)
    }
 
    if (state & MESA_META_DRAW_BUFFERS) {
-      int buf, real_color_buffers = 0;
-      memset(save->ColorDrawBuffers, 0, sizeof(save->ColorDrawBuffers));
-
-      for (buf = 0; buf < ctx->Const.MaxDrawBuffers; buf++) {
-         int buf_index = ctx->DrawBuffer->_ColorDrawBufferIndexes[buf];
-         if (buf_index == -1)
-            continue;
-
-         save->ColorDrawBuffers[buf] =
-            gl_buffer_index_to_drawbuffers_enum(buf_index);
-
-         if (++real_color_buffers >= ctx->DrawBuffer->_NumColorDrawBuffers)
-            break;
-      }
+      struct gl_framebuffer *fb = ctx->DrawBuffer;
+      memcpy(save->ColorDrawBuffers, fb->ColorDrawBuffer,
+             sizeof(save->ColorDrawBuffers));
    }
 
    /* misc */
@@ -1224,7 +1194,7 @@ _mesa_meta_end(struct gl_context *ctx)
       _mesa_BindRenderbuffer(GL_RENDERBUFFER, save->RenderbufferName);
 
    if (state & MESA_META_DRAW_BUFFERS) {
-      _mesa_DrawBuffers(ctx->Const.MaxDrawBuffers, save->ColorDrawBuffers);
+      _mesa_drawbuffers(ctx, ctx->Const.MaxDrawBuffers, save->ColorDrawBuffers, NULL);
    }
 
    ctx->Meta->SaveStackDepth--;

src/mesa/drivers/common/meta_copy_image.c

@@ -74,7 +74,7 @@ make_view(struct gl_context *ctx, struct gl_texture_image *tex_image,
                               tex_image->Depth,
                               0, internal_format, tex_format);
 
-   view_tex_obj->MinLevel = 0;
+   view_tex_obj->MinLevel = tex_image->Level;
    view_tex_obj->NumLevels = 1;
    view_tex_obj->MinLayer = tex_obj->MinLayer;
    view_tex_obj->NumLayers = tex_obj->NumLayers;

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

@@ -1029,19 +1029,17 @@ update_uip_jip(struct brw_context *brw, brw_inst *insn,
 {
    int scale = brw->gen >= 8 ? sizeof(brw_compact_inst) : 1;
 
-   int32_t jip = brw_inst_jip(brw, insn);
-   jip -= scale *
-      compacted_between(this_old_ip, this_old_ip + jip, compacted_counts);
-   brw_inst_set_jip(brw, insn, jip);
+   int32_t jip = brw_inst_jip(brw, insn) / scale;
+   jip -= compacted_between(this_old_ip, this_old_ip + jip, compacted_counts);
+   brw_inst_set_jip(brw, insn, jip * scale);
 
    if (brw_inst_opcode(brw, insn) == BRW_OPCODE_ENDIF ||
        brw_inst_opcode(brw, insn) == BRW_OPCODE_WHILE)
       return;
 
-   int32_t uip = brw_inst_uip(brw, insn);
-   uip -= scale *
-      compacted_between(this_old_ip, this_old_ip + uip, compacted_counts);
-   brw_inst_set_uip(brw, insn, uip);
+   int32_t uip = brw_inst_uip(brw, insn) / scale;
+   uip -= compacted_between(this_old_ip, this_old_ip + uip, compacted_counts);
+   brw_inst_set_uip(brw, insn, uip * scale);
 }
 
 void

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

@@ -2246,10 +2246,10 @@ fs_visitor::emit_bool_to_cond_code(ir_rvalue *ir)
       return;
    }
 
-   fs_reg op[2];
+   fs_reg op[3];
    fs_inst *inst;
 
-   assert(expr->get_num_operands() <= 2);
+   assert(expr->get_num_operands() <= 3);
    for (unsigned int i = 0; i < expr->get_num_operands(); i++) {
       assert(expr->operands[i]->type->is_scalar());
 
@@ -2336,6 +2336,22 @@ fs_visitor::emit_bool_to_cond_code(ir_rvalue *ir)
                brw_conditional_for_comparison(expr->operation)));
       break;
 
+   case ir_triop_csel: {
+      /* Expand the boolean condition into the flag register. */
+      inst = emit(MOV(reg_null_d, op[0]));
+      inst->conditional_mod = BRW_CONDITIONAL_NZ;
+
+      /* Select which boolean to return. */
+      fs_reg temp(this, expr->operands[1]->type);
+      inst = emit(SEL(temp, op[1], op[2]));
+      inst->predicate = BRW_PREDICATE_NORMAL;
+
+      /* Expand the result to a condition code. */
+      inst = emit(MOV(reg_null_d, temp));
+      inst->conditional_mod = BRW_CONDITIONAL_NZ;
+      break;
+   }
+
    default:
       unreachable("not reached");
    }

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

@@ -282,6 +282,7 @@ get_fast_clear_rect(struct brw_context *brw, struct gl_framebuffer *fb,
        * factor is 2 vertically and either 2 or 8 horizontally.
        */
       switch (irb->mt->num_samples) {
+      case 2:
       case 4:
          x_scaledown = 8;
          break;

src/mesa/drivers/dri/i965/brw_vec4_visitor.cpp

@@ -777,10 +777,10 @@ vec4_visitor::emit_bool_to_cond_code(ir_rvalue *ir,
    *predicate = BRW_PREDICATE_NORMAL;
 
    if (expr) {
-      src_reg op[2];
+      src_reg op[3];
       vec4_instruction *inst;
 
-      assert(expr->get_num_operands() <= 2);
+      assert(expr->get_num_operands() <= 3);
       for (unsigned int i = 0; i < expr->get_num_operands(); i++) {
 	 expr->operands[i]->accept(this);
 	 op[i] = this->result;
@@ -852,6 +852,22 @@ vec4_visitor::emit_bool_to_cond_code(ir_rvalue *ir,
 		  brw_conditional_for_comparison(expr->operation)));
 	 break;
 
+      case ir_triop_csel: {
+         /* Expand the boolean condition into the flag register. */
+         inst = emit(MOV(dst_null_d(), op[0]));
+         inst->conditional_mod = BRW_CONDITIONAL_NZ;
+
+         /* Select which boolean to return. */
+         dst_reg temp(this, expr->operands[1]->type);
+         inst = emit(BRW_OPCODE_SEL, temp, op[1], op[2]);
+         inst->predicate = BRW_PREDICATE_NORMAL;
+
+         /* Expand the result to a condition code. */
+         inst = emit(MOV(dst_null_d(), src_reg(temp)));
+         inst->conditional_mod = BRW_CONDITIONAL_NZ;
+         break;
+      }
+
       default:
 	 unreachable("not reached");
       }
@@ -1135,6 +1151,13 @@ vec4_visitor::try_emit_mad(ir_expression *ir)
 bool
 vec4_visitor::try_emit_b2f_of_compare(ir_expression *ir)
 {
+   /* This optimization relies on CMP setting the destination to 0 when
+    * false.  Early hardware only sets the least significant bit, and
+    * leaves the other bits undefined.  So we can't use it.
+    */
+   if (brw->gen < 6)
+      return false;
+
    ir_expression *const cmp = ir->operands[0]->as_expression();
 
    if (cmp == NULL)
@@ -2589,7 +2612,7 @@ vec4_visitor::visit(ir_texture *ir)
       } else if (ir->op == ir_txf_ms) {
          emit(MOV(dst_reg(MRF, param_base + 1, sample_index_type, WRITEMASK_X),
                   sample_index));
-         if (brw->gen >= 7)
+         if (brw->gen >= 7) {
             /* MCS data is in the first channel of `mcs`, but we need to get it into
              * the .y channel of the second vec4 of params, so replicate .x across
              * the whole vec4 and then mask off everything except .y
@@ -2597,6 +2620,7 @@ vec4_visitor::visit(ir_texture *ir)
             mcs.swizzle = BRW_SWIZZLE_XXXX;
             emit(MOV(dst_reg(MRF, param_base + 1, glsl_type::uint_type, WRITEMASK_Y),
                      mcs));
+         }
          inst->mlen++;
       } else if (ir->op == ir_txd) {
 	 const glsl_type *type = lod_type;

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

@@ -40,6 +40,7 @@ copy_image_with_blitter(struct brw_context *brw,
                         int src_width, int src_height)
 {
    GLuint bw, bh;
+   uint32_t src_image_x, src_image_y, dst_image_x, dst_image_y;
    int cpp;
 
    /* The blitter doesn't understand multisampling at all. */
@@ -70,43 +71,53 @@ copy_image_with_blitter(struct brw_context *brw,
       return false;
    }
 
+   intel_miptree_get_image_offset(src_mt, src_level, src_z,
+                                  &src_image_x, &src_image_y);
+
    if (_mesa_is_format_compressed(src_mt->format)) {
       _mesa_get_format_block_size(src_mt->format, &bw, &bh);
 
       assert(src_x % bw == 0);
-      assert(src_y % bw == 0);
+      assert(src_y % bh == 0);
       assert(src_width % bw == 0);
-      assert(src_height % bw == 0);
+      assert(src_height % bh == 0);
 
       src_x /= (int)bw;
-      src_y /= (int)bw;
+      src_y /= (int)bh;
       src_width /= (int)bw;
-      src_height /= (int)bw;
+      src_height /= (int)bh;
+
+      /* Inside of the miptree, the x offsets are stored in pixels while
+       * the y offsets are stored in blocks.  We need to scale just the x
+       * offset.
+       */
+      src_image_x /= bw;
 
       cpp = _mesa_get_format_bytes(src_mt->format);
    } else {
       cpp = src_mt->cpp;
    }
+   src_x += src_image_x;
+   src_y += src_image_y;
+
+   intel_miptree_get_image_offset(dst_mt, dst_level, dst_z,
+                                  &dst_image_x, &dst_image_y);
 
    if (_mesa_is_format_compressed(dst_mt->format)) {
       _mesa_get_format_block_size(dst_mt->format, &bw, &bh);
 
       assert(dst_x % bw == 0);
-      assert(dst_y % bw == 0);
+      assert(dst_y % bh == 0);
 
       dst_x /= (int)bw;
-      dst_y /= (int)bw;
+      dst_y /= (int)bh;
+
+      /* Inside of the miptree, the x offsets are stored in pixels while
+       * the y offsets are stored in blocks.  We need to scale just the x
+       * offset.
+       */
+      dst_image_x /= bw;
    }
-
-   uint32_t src_image_x, src_image_y;
-   intel_miptree_get_image_offset(src_mt, src_level, src_z,
-                                  &src_image_x, &src_image_y);
-   src_x += src_image_x;
-   src_y += src_image_y;
-
-   uint32_t dst_image_x, dst_image_y;
-   intel_miptree_get_image_offset(dst_mt, dst_level, dst_z,
-                                  &dst_image_x, &dst_image_y);
    dst_x += dst_image_x;
    dst_y += dst_image_y;
 
@@ -243,9 +254,11 @@ intel_copy_image_sub_data(struct gl_context *ctx,
    intel_miptree_all_slices_resolve_depth(brw, intel_dst_image->mt);
    intel_miptree_resolve_color(brw, intel_dst_image->mt);
 
-   if (copy_image_with_blitter(brw, intel_src_image->mt, src_image->Level,
+   unsigned src_level = src_image->Level + src_image->TexObject->MinLevel;
+   unsigned dst_level = dst_image->Level + dst_image->TexObject->MinLevel;
+   if (copy_image_with_blitter(brw, intel_src_image->mt, src_level,
                                src_x, src_y, src_z,
-                               intel_dst_image->mt, src_image->Level,
+                               intel_dst_image->mt, dst_level,
                                dst_x, dst_y, dst_z,
                                src_width, src_height))
       return;
@@ -253,9 +266,9 @@ intel_copy_image_sub_data(struct gl_context *ctx,
    /* This is a worst-case scenario software fallback that maps the two
     * textures and does a memcpy between them.
     */
-   copy_image_with_memcpy(brw, intel_src_image->mt, src_image->Level,
+   copy_image_with_memcpy(brw, intel_src_image->mt, src_level,
                           src_x, src_y, src_z,
-                          intel_dst_image->mt, src_image->Level,
+                          intel_dst_image->mt, dst_level,
                           dst_x, dst_y, dst_z,
                           src_width, src_height);
 }

src/mesa/drivers/dri/i965/test_vec4_register_coalesce.cpp

@@ -241,8 +241,5 @@ TEST_F(register_coalesce_test, test_channel_mul_grf)
 
    register_coalesce(v);
 
-   /* This path isn't supported yet in the reswizzling code, so we're checking
-    * that we haven't done anything bad to scalar non-DP[234]s.
-    */
-   EXPECT_NE(mul->dst.reg, to.reg);
+   EXPECT_EQ(mul->dst.reg, to.reg);
 }

src/mesa/main/attrib.c

@@ -1488,6 +1488,10 @@ copy_array_attrib(struct gl_context *ctx,
 
    /* skip ArrayBufferObj */
    /* skip IndexBufferObj */
+
+   /* Invalidate draw state. It will be updated during the next draw. */
+   dest->DrawMethod = DRAW_NONE;
+   dest->_DrawArrays = NULL;
 }
 
 /**

src/mesa/main/context.c

@@ -653,6 +653,9 @@ _mesa_init_constants(struct gl_constants *consts, gl_api api)
    /* GL_ARB_framebuffer_object */
    consts->MaxSamples = 0;
 
+   /* GLSL default if NativeIntegers == FALSE */
+   consts->UniformBooleanTrue = FLT_AS_UINT(1.0f);
+
    /* GL_ARB_sync */
    consts->MaxServerWaitTimeout = 0x1fff7fffffffULL;
 

src/mesa/main/macros.h

@@ -184,6 +184,13 @@ static inline GLfloat UINT_AS_FLT(GLuint u)
    return tmp.f;
 }
 
+static inline unsigned FLT_AS_UINT(float f)
+{
+   fi_type tmp;
+   tmp.f = f;
+   return tmp.u;
+}
+
 /**
  * Convert a floating point value to an unsigned fixed point value.
  *

src/mesa/state_tracker/st_extensions.c

@@ -34,6 +34,7 @@
 #include "pipe/p_context.h"
 #include "pipe/p_defines.h"
 #include "pipe/p_screen.h"
+#include "util/u_math.h"
 
 #include "st_context.h"
 #include "st_extensions.h"
@@ -274,8 +275,6 @@ void st_init_limits(struct pipe_screen *screen,
    c->MinProgramTextureGatherOffset = screen->get_param(screen, PIPE_CAP_MIN_TEXTURE_GATHER_OFFSET);
    c->MaxProgramTextureGatherOffset = screen->get_param(screen, PIPE_CAP_MAX_TEXTURE_GATHER_OFFSET);
 
-   c->UniformBooleanTrue = ~0;
-
    c->MaxTransformFeedbackBuffers =
       screen->get_param(screen, PIPE_CAP_MAX_STREAM_OUTPUT_BUFFERS);
    c->MaxTransformFeedbackBuffers = MIN2(c->MaxTransformFeedbackBuffers, MAX_FEEDBACK_BUFFERS);
@@ -697,6 +696,8 @@ void st_init_extensions(struct pipe_screen *screen,
       }
    }
 
+   consts->UniformBooleanTrue = consts->NativeIntegers ? ~0 : fui(1.0f);
+
    /* Below are the cases which cannot be moved into tables easily. */
 
    if (!has_lib_dxtc && !options->force_s3tc_enable) {

src/mesa/state_tracker/st_glsl_to_tgsi.cpp

@@ -74,14 +74,6 @@ extern "C" {
                            (1 << PROGRAM_CONSTANT) |     \
                            (1 << PROGRAM_UNIFORM))
 
-/**
- * Maximum number of temporary registers.
- *
- * It is too big for stack allocated arrays -- it will cause stack overflow on
- * Windows and likely Mac OS X.
- */
-#define MAX_TEMPS         4096
-
 /**
  * Maximum number of arrays
  */
@@ -3301,14 +3293,10 @@ get_src_arg_mask(st_dst_reg dst, st_src_reg src)
 void
 glsl_to_tgsi_visitor::simplify_cmp(void)
 {
-   unsigned *tempWrites;
+   int tempWritesSize = 0;
+   unsigned *tempWrites = NULL;
    unsigned outputWrites[MAX_PROGRAM_OUTPUTS];
 
-   tempWrites = new unsigned[MAX_TEMPS];
-   if (!tempWrites) {
-      return;
-   }
-   memset(tempWrites, 0, sizeof(unsigned) * MAX_TEMPS);
    memset(outputWrites, 0, sizeof(outputWrites));
 
    foreach_in_list(glsl_to_tgsi_instruction, inst, &this->instructions) {
@@ -3330,7 +3318,19 @@ glsl_to_tgsi_visitor::simplify_cmp(void)
          prevWriteMask = outputWrites[inst->dst.index];
          outputWrites[inst->dst.index] |= inst->dst.writemask;
       } else if (inst->dst.file == PROGRAM_TEMPORARY) {
-         assert(inst->dst.index < MAX_TEMPS);
+         if (inst->dst.index >= tempWritesSize) {
+            const int inc = 4096;
+
+            tempWrites = (unsigned*)
+                         realloc(tempWrites,
+                                 (tempWritesSize + inc) * sizeof(unsigned));
+            if (!tempWrites)
+               return;
+
+            memset(tempWrites + tempWritesSize, 0, inc * sizeof(unsigned));
+            tempWritesSize += inc;
+         }
+
          prevWriteMask = tempWrites[inst->dst.index];
          tempWrites[inst->dst.index] |= inst->dst.writemask;
       } else
@@ -3349,7 +3349,7 @@ glsl_to_tgsi_visitor::simplify_cmp(void)
       }
    }
 
-   delete [] tempWrites;
+   free(tempWrites);
 }
 
 /* Replaces all references to a temporary register index with another index. */
@@ -4158,7 +4158,9 @@ struct label {
 struct st_translate {
    struct ureg_program *ureg;
 
-   struct ureg_dst temps[MAX_TEMPS];
+   unsigned temps_size;
+   struct ureg_dst *temps;
+
    struct ureg_dst arrays[MAX_ARRAYS];
    struct ureg_src *constants;
    struct ureg_src *immediates;
@@ -4299,7 +4301,19 @@ dst_register(struct st_translate *t,
       return ureg_dst_undef();
 
    case PROGRAM_TEMPORARY:
-      assert(index < Elements(t->temps));
+      /* Allocate space for temporaries on demand. */
+      if (index >= t->temps_size) {
+         const int inc = 4096;
+
+         t->temps = (struct ureg_dst*)
+                    realloc(t->temps,
+                            (t->temps_size + inc) * sizeof(struct ureg_dst));
+         if (!t->temps)
+            return ureg_dst_undef();
+
+         memset(t->temps + t->temps_size, 0, inc * sizeof(struct ureg_dst));
+         t->temps_size += inc;
+      }
 
       if (ureg_dst_is_undef(t->temps[index]))
          t->temps[index] = ureg_DECL_local_temporary(t->ureg);
@@ -5158,6 +5172,7 @@ st_translate_program(
 
 out:
    if (t) {
+      free(t->temps);
       free(t->insn);
       free(t->labels);
       free(t->constants);