VERSION | 2 src/gallium/docs/source/screen.rst | 6 + src/gallium/drivers/freedreno/freedreno_screen.c | 5 src/gallium/drivers/i915/i915_screen.c | 5 src/gallium/drivers/ilo/ilo_screen.c | 3 src/gallium/drivers/llvmpipe/lp_screen.c | 3 src/gallium/drivers/nouveau/codegen/nv50_ir_ra.cpp | 39 ++++++- src/gallium/drivers/nouveau/nouveau_screen.h | 5 src/gallium/drivers/nouveau/nouveau_vp3_video.c | 78 ++++++++++++++- src/gallium/drivers/nouveau/nv30/nv30_screen.c | 4 src/gallium/drivers/nouveau/nv50/nv50_screen.c | 3 src/gallium/drivers/nouveau/nv50/nv84_video.c | 68 ++++++++++++- src/gallium/drivers/nouveau/nvc0/nvc0_screen.c | 3 src/gallium/drivers/r300/r300_screen.c | 2 src/gallium/drivers/r600/r600_hw_context.c | 2 src/gallium/drivers/r600/r600_pipe.c | 14 +- src/gallium/drivers/r600/r600_pipe.h | 1 src/gallium/drivers/radeon/r600_buffer_common.c | 2 src/gallium/drivers/radeon/r600_pipe_common.c | 2 src/gallium/drivers/radeon/r600_pipe_common.h | 3 src/gallium/drivers/radeon/r600_texture.c | 2 src/gallium/drivers/radeonsi/si_hw_context.c | 4 src/gallium/drivers/radeonsi/si_pipe.c | 6 + src/gallium/drivers/softpipe/sp_screen.c | 3 src/gallium/drivers/svga/svga_screen.c | 2 src/gallium/include/pipe/p_defines.h | 4 src/gallium/targets/r600/vdpau/Makefile.am | 6 - src/glsl/ast_to_hir.cpp | 13 +- src/glsl/builtin_functions.cpp | 11 +- src/glsl/glcpp/glcpp-parse.y | 22 +++- src/glsl/glcpp/tests/086-reserved-macro-names.c.expected | 4 src/glsl/linker.cpp | 9 - src/glsl/opt_vectorize.cpp | 14 ++ src/mesa/drivers/dri/i915/i830_vtbl.c | 8 - src/mesa/drivers/dri/i915/i915_context.c | 2 src/mesa/drivers/dri/i915/i915_texstate.c | 2 src/mesa/drivers/dri/i915/i915_vtbl.c | 8 - src/mesa/drivers/dri/i915/intel_fbo.c | 2 src/mesa/drivers/dri/i915/intel_screen.c | 2 src/mesa/drivers/dri/i965/brw_blorp.cpp | 4 src/mesa/drivers/dri/i965/brw_blorp_blit.cpp | 6 - src/mesa/drivers/dri/i965/brw_clear.c | 2 src/mesa/drivers/dri/i965/brw_context.c | 12 +- src/mesa/drivers/dri/i965/brw_misc_state.c | 6 - src/mesa/drivers/dri/i965/brw_surface_formats.c | 12 +- src/mesa/drivers/dri/i965/intel_fbo.c | 2 src/mesa/drivers/dri/i965/intel_mipmap_tree.c | 8 - src/mesa/drivers/dri/i965/intel_mipmap_tree.h | 4 src/mesa/drivers/dri/i965/intel_screen.c | 4 src/mesa/drivers/dri/nouveau/nouveau_context.c | 7 - src/mesa/drivers/dri/nouveau/nouveau_context.h | 3 src/mesa/drivers/dri/nouveau/nouveau_driver.h | 1 src/mesa/drivers/dri/nouveau/nv04_context.c | 5 src/mesa/drivers/dri/nouveau/nv04_surface.c | 8 - src/mesa/drivers/dri/nouveau/nv10_context.c | 13 +- src/mesa/drivers/dri/nouveau/nv10_state_fb.c | 4 src/mesa/drivers/dri/nouveau/nv10_state_raster.c | 3 src/mesa/drivers/dri/nouveau/nv20_context.c | 5 src/mesa/drivers/dri/radeon/radeon_fbo.c | 10 - src/mesa/drivers/dri/radeon/radeon_screen.c | 4 src/mesa/drivers/dri/radeon/radeon_texture.c | 4 src/mesa/main/attrib.c | 1 src/mesa/main/format_pack.c | 22 ++-- src/mesa/main/format_unpack.c | 12 +- src/mesa/main/formats.c | 16 +-- src/mesa/main/formats.h | 4 src/mesa/main/readpix.c | 2 src/mesa/main/shaderapi.c | 1 src/mesa/main/shaderobj.c | 1 src/mesa/main/texformat.c | 6 - src/mesa/main/texparam.c | 3 src/mesa/main/texstore.c | 8 - src/mesa/state_tracker/st_extensions.c | 2 src/mesa/state_tracker/st_format.c | 8 - src/mesa/swrast/s_depth.c | 14 +- src/mesa/swrast/s_drawpix.c | 4 src/mesa/swrast/s_renderbuffer.c | 2 src/mesa/swrast/s_stencil.c | 2 src/mesa/swrast/s_texfetch.c | 4 src/mesa/swrast/s_texfetch_tmp.h | 4 80 files changed, 450 insertions(+), 167 deletions(-)
New commits: commit fcb4eabb5fc31f8902a6e917a4a008aaa90819c6 Author: Ian Romanick <[email protected]> Date: Fri Feb 21 14:29:44 2014 -0800 mesa: Bump version to 10.1-rc2 Signed-off-by: Ian Romanick <[email protected]> diff --git a/VERSION b/VERSION index ff3cc45..5364038 100644 --- a/VERSION +++ b/VERSION @@ -1 +1 @@ -10.1.0-rc1 +10.1.0-rc2 commit 7cb4dea765b5b1fe8f14197084e2ccc864d525dd Author: Kenneth Graunke <[email protected]> Date: Wed Feb 19 17:20:11 2014 -0800 i965: Create a hardware context before initializing state module. brw_init_state() calls brw_upload_initial_gpu_state(). If hardware contexts are enabled (brw->hw_ctx != NULL), this will upload some initial invariant state for the GPU. Without hardware contexts, we rely on this state being uploaded via atoms that subscribe to the BRW_NEW_CONTEXT bit. Commit 46d3c2bf4ddd227193b98861f1e632498fe547d8 accidentally moved the call to brw_init_state() before creating a hardware context. This meant brw_upload_initial_gpu_state would always early return. Except on Gen6+, we stopped uploading the initial GPU state via state atoms, so it never happened. Fixes a regression since 46d3c2bf4ddd227193b98861f1e632498fe547d8. Cc: "10.0 10.1" <[email protected]> Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Ian Romanick <[email protected]> Reviewed-by: Eric Anholt <[email protected]> (cherry picked from commit 3663bbe773187dee341556ef29e58b1143ef2f5c) diff --git a/src/mesa/drivers/dri/i965/brw_context.c b/src/mesa/drivers/dri/i965/brw_context.c index d9d1ae5..d5dccaf 100644 --- a/src/mesa/drivers/dri/i965/brw_context.c +++ b/src/mesa/drivers/dri/i965/brw_context.c @@ -682,12 +682,6 @@ brwCreateContext(gl_api api, intel_batchbuffer_init(brw); - brw_init_state(brw); - - intelInitExtensions(ctx); - - intel_fbo_init(brw); - if (brw->gen >= 6) { /* Create a new hardware context. Using a hardware context means that * our GPU state will be saved/restored on context switch, allowing us @@ -705,6 +699,12 @@ brwCreateContext(gl_api api, } } + brw_init_state(brw); + + intelInitExtensions(ctx); + + intel_fbo_init(brw); + brw_init_surface_formats(brw); if (brw->is_g4x || brw->gen >= 5) { commit b498fb95867f2b939c8e257a5cda959e42eae7b5 Author: Ian Romanick <[email protected]> Date: Tue Feb 18 09:36:08 2014 -0800 glsl: Only warn for macro names containing __ From page 14 (page 20 of the PDF) of the GLSL 1.10 spec: "In addition, all identifiers containing two consecutive underscores (__) are reserved as possible future keywords." The intention is that names containing __ are reserved for internal use by the implementation, and names prefixed with GL_ are reserved for use by Khronos. Names simply containing __ are dangerous to use, but should be allowed. Per the Khronos bug mentioned below, a future version of the GLSL specification will clarify this. Signed-off-by: Ian Romanick <[email protected]> Cc: "9.2 10.0 10.1" <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]> Tested-by: Kenneth Graunke <[email protected]> Reviewed-by: Anuj Phogat <[email protected]> Tested-by: Darius Spitznagel <[email protected]> Cc: Tapani Pälli <[email protected]> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=71870 Bugzilla: Khronos #11702 (cherry picked from commit 2c85fd5a964a78c9f7a93994fb79f1723c6f45b5) diff --git a/src/glsl/ast_to_hir.cpp b/src/glsl/ast_to_hir.cpp index 8d096ad..624b813 100644 --- a/src/glsl/ast_to_hir.cpp +++ b/src/glsl/ast_to_hir.cpp @@ -2853,10 +2853,17 @@ validate_identifier(const char *identifier, YYLTYPE loc, * "In addition, all identifiers containing two * consecutive underscores (__) are reserved as * possible future keywords." + * + * The intention is that names containing __ are reserved for internal + * use by the implementation, and names prefixed with GL_ are reserved + * for use by Khronos. Names simply containing __ are dangerous to use, + * but should be allowed. + * + * A future version of the GLSL specification will clarify this. */ - _mesa_glsl_error(&loc, state, - "identifier `%s' uses reserved `__' string", - identifier); + _mesa_glsl_warning(&loc, state, + "identifier `%s' uses reserved `__' string", + identifier); } } commit 3731a4fae434ac61bd951009bd776d13ae3833b3 Author: Ian Romanick <[email protected]> Date: Tue Feb 18 09:10:36 2014 -0800 glcpp: Only warn for macro names containing __ Section 3.3 (Preprocessor) of the GLSL 1.30 spec (and later) and the GLSL ES spec (all versions) say: "All macro names containing two consecutive underscores ( __ ) are reserved for future use as predefined macro names. All macro names prefixed with "GL_" ("GL" followed by a single underscore) are also reserved." The intention is that names containing __ are reserved for internal use by the implementation, and names prefixed with GL_ are reserved for use by Khronos. Since every extension adds a name prefixed with GL_ (i.e., the name of the extension), that should be an error. Names simply containing __ are dangerous to use, but should be allowed. In similar cases, the C++ preprocessor specification says, "no diagnostic is required." Per the Khronos bug mentioned below, a future version of the GLSL specification will clarify this. Signed-off-by: Ian Romanick <[email protected]> Cc: "9.2 10.0 10.1" <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]> Tested-by: Kenneth Graunke <[email protected]> Reviewed-by: Anuj Phogat <[email protected]> Tested-by: Darius Spitznagel <[email protected]> Cc: Tapani Pälli <[email protected]> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=71870 Bugzilla: Khronos #11702 (cherry picked from commit 0bd78926304e72ef3566e977d0cb5a959d86b809) diff --git a/src/glsl/glcpp/glcpp-parse.y b/src/glsl/glcpp/glcpp-parse.y index 0393ac7..7f4691a 100644 --- a/src/glsl/glcpp/glcpp-parse.y +++ b/src/glsl/glcpp/glcpp-parse.y @@ -1770,11 +1770,27 @@ static void _check_for_reserved_macro_name (glcpp_parser_t *parser, YYLTYPE *loc, const char *identifier) { - /* According to the GLSL specification, macro names starting with "__" - * or "GL_" are reserved for future use. So, don't allow them. + /* Section 3.3 (Preprocessor) of the GLSL 1.30 spec (and later) and + * the GLSL ES spec (all versions) say: + * + * "All macro names containing two consecutive underscores ( __ ) + * are reserved for future use as predefined macro names. All + * macro names prefixed with "GL_" ("GL" followed by a single + * underscore) are also reserved." + * + * The intention is that names containing __ are reserved for internal + * use by the implementation, and names prefixed with GL_ are reserved + * for use by Khronos. Since every extension adds a name prefixed + * with GL_ (i.e., the name of the extension), that should be an + * error. Names simply containing __ are dangerous to use, but should + * be allowed. + * + * A future version of the GLSL specification will clarify this. */ if (strstr(identifier, "__")) { - glcpp_error (loc, parser, "Macro names containing \"__\" are reserved.\n"); + glcpp_warning(loc, parser, + "Macro names containing \"__\" are reserved " + "for use by the implementation.\n"); } if (strncmp(identifier, "GL_", 3) == 0) { glcpp_error (loc, parser, "Macro names starting with \"GL_\" are reserved.\n"); diff --git a/src/glsl/glcpp/tests/086-reserved-macro-names.c.expected b/src/glsl/glcpp/tests/086-reserved-macro-names.c.expected index d8aa9f0..5ca42a9 100644 --- a/src/glsl/glcpp/tests/086-reserved-macro-names.c.expected +++ b/src/glsl/glcpp/tests/086-reserved-macro-names.c.expected @@ -1,8 +1,8 @@ -0:1(10): preprocessor error: Macro names containing "__" are reserved. +0:1(10): preprocessor warning: Macro names containing "__" are reserved for use by the implementation. 0:2(9): preprocessor error: Macro names starting with "GL_" are reserved. -0:3(9): preprocessor error: Macro names containing "__" are reserved. +0:3(9): preprocessor warning: Macro names containing "__" are reserved for use by the implementation. commit d623eeb37a99ccc419e7475693b2712930aae8bf Author: Anuj Phogat <[email protected]> Date: Thu Dec 19 14:17:19 2013 -0800 glsl: Fix condition to generate shader link error GL_ARB_ES2_compatibility doesn't say anything about shader linking when one of the shaders (vertex or fragment shader) is absent. So, the extension shouldn't change the behavior specified in GLSL specification. Tested the behavior on proprietary linux drivers of NVIDIA and AMD. Both of them allow linking a version 100 shader program in OpenGL context, when one of the shaders is absent. Makes following Khronos CTS tests to pass: successfulcompilevert_linkprogram.test successfulcompilefrag_linkprogram.test Cc: [email protected] Signed-off-by: Anuj Phogat <[email protected]> Reviewed-by: Ian Romanick <[email protected]> (cherry picked from commit 03597cf802a7a89c4853794e6206ab8ab003898d) diff --git a/src/glsl/linker.cpp b/src/glsl/linker.cpp index 93b4754..5635b1d 100644 --- a/src/glsl/linker.cpp +++ b/src/glsl/linker.cpp @@ -2329,11 +2329,12 @@ link_shaders(struct gl_context *ctx, struct gl_shader_program *prog) goto done; /* OpenGL ES requires that a vertex shader and a fragment shader both be - * present in a linked program. By checking prog->IsES, we also - * catch the GL_ARB_ES2_compatibility case. + * present in a linked program. GL_ARB_ES2_compatibility doesn't say + * anything about shader linking when one of the shaders (vertex or + * fragment shader) is absent. So, the extension shouldn't change the + * behavior specified in GLSL specification. */ - if (!prog->InternalSeparateShader && - (ctx->API == API_OPENGLES2 || prog->IsES)) { + if (!prog->InternalSeparateShader && ctx->API == API_OPENGLES2) { if (prog->_LinkedShaders[MESA_SHADER_VERTEX] == NULL) { linker_error(prog, "program lacks a vertex shader\n"); } else if (prog->_LinkedShaders[MESA_SHADER_FRAGMENT] == NULL) { commit 6534d80ccac4a23f7959365bbfcd37626accc807 Author: Anuj Phogat <[email protected]> Date: Fri Feb 14 17:27:29 2014 -0800 mesa: Add GL_TEXTURE_CUBE_MAP_ARRAY to legal_get_tex_level_parameter_target() Fixes failing Khronos CTS test packed_depth_stencil_init.test Cc: <[email protected]> Signed-off-by: Anuj Phogat <[email protected]> Reviewed-by: Ian Romanick <[email protected]> (cherry picked from commit 6bd2472a8b9a969d06dc110bd30ccd8daf713959) diff --git a/src/mesa/main/texparam.c b/src/mesa/main/texparam.c index b7ed50d..bbdbc27 100644 --- a/src/mesa/main/texparam.c +++ b/src/mesa/main/texparam.c @@ -986,6 +986,9 @@ legal_get_tex_level_parameter_target(struct gl_context *ctx, GLenum target) case GL_TEXTURE_CUBE_MAP_NEGATIVE_Z_ARB: case GL_PROXY_TEXTURE_CUBE_MAP_ARB: return ctx->Extensions.ARB_texture_cube_map; + case GL_TEXTURE_CUBE_MAP_ARRAY_ARB: + case GL_PROXY_TEXTURE_CUBE_MAP_ARRAY_ARB: + return ctx->Extensions.ARB_texture_cube_map_array; case GL_TEXTURE_RECTANGLE_NV: case GL_PROXY_TEXTURE_RECTANGLE_NV: return ctx->Extensions.NV_texture_rectangle; commit 20eb46699905e4ed9a4886fdb83caf72a344fb71 Author: Michel Dänzer <[email protected]> Date: Thu Feb 13 11:51:09 2014 +0900 r600g,radeonsi: Consolidate logic for short-circuiting flushes Fixes radeonsi emitting command streams to the kernel even when there have been no draw calls before a flush, potentially powering up the GPU needlessly. Incidentally, this also cuts the runtime of piglit gpu.py in about half on my Kaveri system, probably because an X11 client going away no longer always results in a command stream being submitted to the kernel via glamor. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=65761 Cc: "10.1" [email protected] Reviewed-by: Marek Olšák <[email protected]> (cherry picked from commit cf0172d46ab940a691da6516057c81f28961482f) diff --git a/src/gallium/drivers/r600/r600_hw_context.c b/src/gallium/drivers/r600/r600_hw_context.c index fc81e95..358282a 100644 --- a/src/gallium/drivers/r600/r600_hw_context.c +++ b/src/gallium/drivers/r600/r600_hw_context.c @@ -352,7 +352,7 @@ void r600_begin_new_cs(struct r600_context *ctx) ctx->last_primitive_type = -1; ctx->last_start_instance = -1; - ctx->initial_gfx_cs_size = ctx->b.rings.gfx.cs->cdw; + ctx->b.initial_gfx_cs_size = ctx->b.rings.gfx.cs->cdw; } /* The max number of bytes to copy per packet. */ diff --git a/src/gallium/drivers/r600/r600_pipe.c b/src/gallium/drivers/r600/r600_pipe.c index 66996bd..4ca6a22 100644 --- a/src/gallium/drivers/r600/r600_pipe.c +++ b/src/gallium/drivers/r600/r600_pipe.c @@ -73,7 +73,7 @@ static void r600_flush(struct pipe_context *ctx, unsigned flags) unsigned render_cond_mode = 0; boolean render_cond_cond = FALSE; - if (rctx->b.rings.gfx.cs->cdw == rctx->initial_gfx_cs_size) + if (rctx->b.rings.gfx.cs->cdw == rctx->b.initial_gfx_cs_size) return; rctx->b.rings.gfx.flushing = true; @@ -94,7 +94,7 @@ static void r600_flush(struct pipe_context *ctx, unsigned flags) ctx->render_condition(ctx, render_cond, render_cond_cond, render_cond_mode); } - rctx->initial_gfx_cs_size = rctx->b.rings.gfx.cs->cdw; + rctx->b.initial_gfx_cs_size = rctx->b.rings.gfx.cs->cdw; } static void r600_flush_from_st(struct pipe_context *ctx, diff --git a/src/gallium/drivers/r600/r600_pipe.h b/src/gallium/drivers/r600/r600_pipe.h index 9f27a17..0cc91dd 100644 --- a/src/gallium/drivers/r600/r600_pipe.h +++ b/src/gallium/drivers/r600/r600_pipe.h @@ -375,7 +375,6 @@ struct r600_context { struct r600_screen *screen; struct blitter_context *blitter; struct u_suballocator *allocator_fetch_shader; - unsigned initial_gfx_cs_size; /* Hardware info. */ boolean has_vertex_cache; diff --git a/src/gallium/drivers/radeon/r600_buffer_common.c b/src/gallium/drivers/radeon/r600_buffer_common.c index d29671e..15eec55 100644 --- a/src/gallium/drivers/radeon/r600_buffer_common.c +++ b/src/gallium/drivers/radeon/r600_buffer_common.c @@ -59,7 +59,7 @@ void *r600_buffer_map_sync_with_rings(struct r600_common_context *ctx, rusage = RADEON_USAGE_WRITE; } - if (ctx->rings.gfx.cs->cdw && + if (ctx->rings.gfx.cs->cdw != ctx->initial_gfx_cs_size && ctx->ws->cs_is_buffer_referenced(ctx->rings.gfx.cs, resource->cs_buf, rusage)) { if (usage & PIPE_TRANSFER_DONTBLOCK) { diff --git a/src/gallium/drivers/radeon/r600_pipe_common.h b/src/gallium/drivers/radeon/r600_pipe_common.h index aa1f3ed..3c22c0f 100644 --- a/src/gallium/drivers/radeon/r600_pipe_common.h +++ b/src/gallium/drivers/radeon/r600_pipe_common.h @@ -241,6 +241,7 @@ struct r600_common_context { enum radeon_family family; enum chip_class chip_class; struct r600_rings rings; + unsigned initial_gfx_cs_size; struct u_upload_mgr *uploader; struct u_suballocator *allocator_so_filled_size; diff --git a/src/gallium/drivers/radeonsi/si_hw_context.c b/src/gallium/drivers/radeonsi/si_hw_context.c index 7e40255..3437f59 100644 --- a/src/gallium/drivers/radeonsi/si_hw_context.c +++ b/src/gallium/drivers/radeonsi/si_hw_context.c @@ -81,7 +81,7 @@ void si_context_flush(struct si_context *ctx, unsigned flags) { struct radeon_winsys_cs *cs = ctx->b.rings.gfx.cs; - if (!cs->cdw) + if (cs->cdw == ctx->b.initial_gfx_cs_size) return; /* suspend queries */ @@ -177,6 +177,8 @@ void si_begin_new_cs(struct si_context *ctx) } si_all_descriptors_begin_new_cs(ctx); + + ctx->b.initial_gfx_cs_size = ctx->b.rings.gfx.cs->cdw; } #if SI_TRACE_CS commit 706eef0cfea4973e167bf531f4cf20683c4960bf Author: Kusanagi Kouichi <[email protected]> Date: Wed Feb 12 16:07:55 2014 +0900 targets/vdpau: Always use c++ to link If built without llvm, the following error occurs with mplayer: Failed to open VDPAU backend .../libvdpau_r600.so: undefined symbol: _ZTVN10__cxxabiv117__class_type_infoE [vo/vdpau] Error when calling vdp_device_create_x11: 1 Cc: <[email protected]> Signed-off-by: Kusanagi Kouichi <[email protected]> Reviewed-by: Emil Velikov <[email protected]> (cherry picked from commit 61f6cddef72a49d87e2b31ed49cdbc941d3b1efc) diff --git a/src/gallium/targets/r600/vdpau/Makefile.am b/src/gallium/targets/r600/vdpau/Makefile.am index 7f43fbb..6e9f117 100644 --- a/src/gallium/targets/r600/vdpau/Makefile.am +++ b/src/gallium/targets/r600/vdpau/Makefile.am @@ -44,17 +44,13 @@ libvdpau_r600_la_LIBADD = \ $(GALLIUM_DRI_LIB_DEPS) \ $(RADEON_LIBS) -if HAVE_MESA_LLVM libvdpau_r600_la_LINK = $(CXXLINK) $(libvdpau_r600_la_LDFLAGS) # Mention a dummy pure C++ file to trigger generation of the $(LINK) variable nodist_EXTRA_libvdpau_r600_la_SOURCES = dummy-cpp.cpp +if HAVE_MESA_LLVM libvdpau_r600_la_LDFLAGS += $(LLVM_LDFLAGS) libvdpau_r600_la_LIBADD += $(LLVM_LIBS) -else -libvdpau_r600_la_LINK = $(LINK) $(libvdpau_r600_la_LDFLAGS) -# Mention a dummy pure C file to trigger generation of the $(LINK) variable -nodist_EXTRA_libvdpau_r600_la_SOURCES = dummy-c.c endif # Provide compatibility with scripts for the old Mesa build system for commit a889b8e9aa76dbe960e6117347217529e2c4f714 Author: Carl Worth <[email protected]> Date: Thu Feb 13 09:49:27 2014 -0800 main: Avoid double-free of shader Label As documented, the _mesa_free_shader_program_data function: "Frees all the data that hangs off a shader program object, but not the object itself." This means that this function may be called multiple times on the same object, (and has been observed to). Meanwhile, the shProg->Label field was not being set to NULL after its free(). This led to a second call to free() of the same address on the second call to this function. Fix this by setting this field to NULL after free(), (just as with all other calls to free() in this function). Reviewed-by: Brian Paul <[email protected]> CC: [email protected] (cherry picked from commit a92581acf2aba5e5e9fa199b778e649d5741754d) diff --git a/src/mesa/main/shaderobj.c b/src/mesa/main/shaderobj.c index 4f4bb69..d5c3d80 100644 --- a/src/mesa/main/shaderobj.c +++ b/src/mesa/main/shaderobj.c @@ -355,6 +355,7 @@ _mesa_free_shader_program_data(struct gl_context *ctx, } free(shProg->Label); + shProg->Label = NULL; } commit 02d96b7e9fa70c5af6179144e80ad2e7c4f4198a Author: Alex Deucher <[email protected]> Date: Wed Feb 12 12:00:17 2014 -0500 radeon: reverse DBG_NO_HYPERZ logic Change the flag to DBG_HYPERZ and reverse the logic so setting the flag enabled the feature. This disables hyperz on r600g and radeonsi by default. It can be enabled by setting the env var. There are just too many issues with certain apps so leave it disabled for now until we sort out the issues with the problematic apps. Bugs: https://bugs.freedesktop.org/show_bug.cgi?id=58660 https://bugs.freedesktop.org/show_bug.cgi?id=64471 https://bugs.freedesktop.org/show_bug.cgi?id=66352 https://bugs.freedesktop.org/show_bug.cgi?id=68799 https://bugs.freedesktop.org/show_bug.cgi?id=72685 https://bugs.freedesktop.org/show_bug.cgi?id=73088 https://bugs.freedesktop.org/show_bug.cgi?id=74428 https://bugs.freedesktop.org/show_bug.cgi?id=74803 https://bugs.freedesktop.org/show_bug.cgi?id=74863 https://bugs.freedesktop.org/show_bug.cgi?id=74892 https://bugzilla.kernel.org/show_bug.cgi?id=70411 Signed-off-by: Alex Deucher <[email protected]> Cc: "10.1" "10.0" <[email protected]> Acked-by: Marek Olšák <[email protected]> (cherry picked from commit 01e637114914453451becc0dc8afe60faff48d84) diff --git a/src/gallium/drivers/r600/r600_pipe.c b/src/gallium/drivers/r600/r600_pipe.c index d4e06c5..66996bd 100644 --- a/src/gallium/drivers/r600/r600_pipe.c +++ b/src/gallium/drivers/r600/r600_pipe.c @@ -585,8 +585,8 @@ struct pipe_screen *r600_screen_create(struct radeon_winsys *ws) rscreen->b.debug_flags |= DBG_COMPUTE; if (debug_get_bool_option("R600_DUMP_SHADERS", FALSE)) rscreen->b.debug_flags |= DBG_FS | DBG_VS | DBG_GS | DBG_PS | DBG_CS; - if (!debug_get_bool_option("R600_HYPERZ", TRUE)) - rscreen->b.debug_flags |= DBG_NO_HYPERZ; + if (debug_get_bool_option("R600_HYPERZ", FALSE)) + rscreen->b.debug_flags |= DBG_HYPERZ; if (!debug_get_bool_option("R600_LLVM", TRUE)) rscreen->b.debug_flags |= DBG_NO_LLVM; diff --git a/src/gallium/drivers/radeon/r600_pipe_common.c b/src/gallium/drivers/radeon/r600_pipe_common.c index 98164f0..bfdfaf5 100644 --- a/src/gallium/drivers/radeon/r600_pipe_common.c +++ b/src/gallium/drivers/radeon/r600_pipe_common.c @@ -137,7 +137,7 @@ static const struct debug_named_value common_debug_options[] = { { "ps", DBG_PS, "Print pixel shaders" }, { "cs", DBG_CS, "Print compute shaders" }, - { "nohyperz", DBG_NO_HYPERZ, "Disable Hyper-Z" }, + { "hyperz", DBG_HYPERZ, "Enable Hyper-Z" }, /* GL uses the word INVALIDATE, gallium uses the word DISCARD */ { "noinvalrange", DBG_NO_DISCARD_RANGE, "Disable handling of INVALIDATE_RANGE map flags" }, diff --git a/src/gallium/drivers/radeon/r600_pipe_common.h b/src/gallium/drivers/radeon/r600_pipe_common.h index 9fdfdfd..aa1f3ed 100644 --- a/src/gallium/drivers/radeon/r600_pipe_common.h +++ b/src/gallium/drivers/radeon/r600_pipe_common.h @@ -83,7 +83,7 @@ #define DBG_PS (1 << 11) #define DBG_CS (1 << 12) /* features */ -#define DBG_NO_HYPERZ (1 << 13) +#define DBG_HYPERZ (1 << 13) #define DBG_NO_DISCARD_RANGE (1 << 14) /* The maximum allowed bit is 15. */ diff --git a/src/gallium/drivers/radeon/r600_texture.c b/src/gallium/drivers/radeon/r600_texture.c index 878b26f..8ea1565 100644 --- a/src/gallium/drivers/radeon/r600_texture.c +++ b/src/gallium/drivers/radeon/r600_texture.c @@ -596,7 +596,7 @@ r600_texture_create_object(struct pipe_screen *screen, if (rtex->is_depth) { if (!(base->flags & (R600_RESOURCE_FLAG_TRANSFER | R600_RESOURCE_FLAG_FLUSHED_DEPTH)) && - !(rscreen->debug_flags & DBG_NO_HYPERZ)) { + (rscreen->debug_flags & DBG_HYPERZ)) { r600_texture_allocate_htile(rscreen, rtex); } commit a1b6aa9fe254274dea1346332037bc3db2f924d3 Author: Ilia Mirkin <[email protected]> Date: Wed Feb 12 18:22:56 2014 -0500 nouveau: fix chipset checks for nv1a by using the oclass instead Commit f4ebcd133b9 ("dri/nouveau: NV17_3D class is not available for NV1a chipset") fixed this partially by using the correct 3d class. However there were a lot of checks left over comparing against the chipset. Reported-and-tested-by: John F. Godfrey <[email protected]> Signed-off-by: Ilia Mirkin <[email protected]> Cc: 9.2 10.0 10.1 <[email protected]> Reviewed-by: Francisco Jerez <[email protected]> (cherry picked from commit 0c8b165366d68291e3013c7308b8b1fdd5ade2a2) diff --git a/src/mesa/drivers/dri/nouveau/nv10_context.c b/src/mesa/drivers/dri/nouveau/nv10_context.c index 904a410..00a9358 100644 --- a/src/mesa/drivers/dri/nouveau/nv10_context.c +++ b/src/mesa/drivers/dri/nouveau/nv10_context.c @@ -63,7 +63,7 @@ nv10_use_viewport_zclear(struct gl_context *ctx) struct gl_framebuffer *fb = ctx->DrawBuffer; struct gl_renderbuffer *depthRb = fb->Attachment[BUFFER_DEPTH].Renderbuffer; - return context_chipset(ctx) < 0x17 && + return context_eng3d(ctx)->oclass < NV17_3D_CLASS && !nctx->hierz.clear_blocked && depthRb && (_mesa_get_format_bits(depthRb->Format, GL_DEPTH_BITS) >= 24); @@ -184,7 +184,7 @@ nv10_clear(struct gl_context *ctx, GLbitfield buffers) } if ((buffers & BUFFER_BIT_DEPTH) && ctx->Depth.Mask) { - if (context_chipset(ctx) >= 0x17) + if (context_eng3d(ctx)->oclass >= NV17_3D_CLASS) nv17_zclear(ctx, &buffers); else nv10_zclear(ctx, &buffers); @@ -245,7 +245,7 @@ nv10_hwctx_init(struct gl_context *ctx) BEGIN_NV04(push, NV04_GRAPH(3D, NOP), 1); PUSH_DATA (push, 0); - if (context_chipset(ctx) >= 0x17) { + if (context_eng3d(ctx)->oclass >= NV17_3D_CLASS) { BEGIN_NV04(push, NV17_3D(UNK01AC), 2); PUSH_DATA (push, fifo->vram); PUSH_DATA (push, fifo->vram); @@ -257,7 +257,7 @@ nv10_hwctx_init(struct gl_context *ctx) PUSH_DATA (push, 1); } - if (context_chipset(ctx) >= 0x11) { + if (context_eng3d(ctx)->oclass >= NV15_3D_CLASS) { BEGIN_NV04(push, SUBC_3D(0x120), 3); PUSH_DATA (push, 0); PUSH_DATA (push, 1); diff --git a/src/mesa/drivers/dri/nouveau/nv10_state_fb.c b/src/mesa/drivers/dri/nouveau/nv10_state_fb.c index 19769e5..fb66b2d 100644 --- a/src/mesa/drivers/dri/nouveau/nv10_state_fb.c +++ b/src/mesa/drivers/dri/nouveau/nv10_state_fb.c @@ -106,7 +106,7 @@ nv10_emit_framebuffer(struct gl_context *ctx, int emit) /* At least nv11 seems to get sad if we don't do this before * swapping RTs.*/ - if (context_chipset(ctx) < 0x17) { + if (context_eng3d(ctx)->oclass < NV17_3D_CLASS) { int i; for (i = 0; i < 6; i++) { @@ -140,7 +140,7 @@ nv10_emit_framebuffer(struct gl_context *ctx, int emit) PUSH_MTHDl(push, NV10_3D(ZETA_OFFSET), BUFCTX_FB, s->bo, 0, bo_flags); - if (context_chipset(ctx) >= 0x17) { + if (context_eng3d(ctx)->oclass >= NV17_3D_CLASS) { setup_hierz_buffer(ctx); context_dirty(ctx, ZCLEAR); } diff --git a/src/mesa/drivers/dri/nouveau/nv10_state_raster.c b/src/mesa/drivers/dri/nouveau/nv10_state_raster.c index bd4a78e..da414a0 100644 --- a/src/mesa/drivers/dri/nouveau/nv10_state_raster.c +++ b/src/mesa/drivers/dri/nouveau/nv10_state_raster.c @@ -28,6 +28,7 @@ #include "nouveau_context.h" #include "nouveau_gldefs.h" #include "nouveau_util.h" +#include "nv_object.xml.h" #include "nv10_3d.xml.h" #include "nv10_driver.h" @@ -120,7 +121,7 @@ nv10_emit_logic_opcode(struct gl_context *ctx, int emit) struct nouveau_pushbuf *push = context_push(ctx); assert(!ctx->Color.ColorLogicOpEnabled - || context_chipset(ctx) >= 0x11); + || context_eng3d(ctx)->oclass >= NV15_3D_CLASS); BEGIN_NV04(push, NV11_3D(COLOR_LOGIC_OP_ENABLE), 2); PUSH_DATAb(push, ctx->Color.ColorLogicOpEnabled); commit 150b1f0aac70921ff52c7dd9b9563981976c7982 Author: Fredrik Höglund <[email protected]> Date: Fri Feb 7 20:34:08 2014 +0100 mesa: Preserve the NewArrays state when copying a VAO Cc: "10.1" "10.0" <[email protected]> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=72895 Reviewed-by: Brian Paul <[email protected]> (cherry picked from commit 9afbd04d892f96e7fc6b689ca57ea5da124f7560) diff --git a/src/mesa/main/attrib.c b/src/mesa/main/attrib.c index 7b7cf0e..fa19288 100644 --- a/src/mesa/main/attrib.c +++ b/src/mesa/main/attrib.c @@ -1457,6 +1457,7 @@ copy_array_object(struct gl_context *ctx, /* _Enabled must be the same than on push */ dest->_Enabled = src->_Enabled; + dest->NewArrays = src->NewArrays; dest->_MaxElement = src->_MaxElement; } commit 50066dc544517cff09cdb11203d8d7213126ee78 Author: Matt Turner <[email protected]> Date: Mon Feb 10 12:09:19 2014 -0800 glsl: Do not vectorize vector array dereferences. Array dereferences must have scalar indices, so we cannot vectorize them. Cc: "10.1" <[email protected]> Reported-by: Andrew Guertin <[email protected]> Tested-by: Andrew Guertin <[email protected]> Reviewed-by: Ian Romanick <[email protected]> (cherry picked from commit 025d99ce3cea21c9ecea77b95655c0c838b2fa3c) diff --git a/src/glsl/opt_vectorize.cpp b/src/glsl/opt_vectorize.cpp index 8ee81f1..dba303d 100644 --- a/src/glsl/opt_vectorize.cpp +++ b/src/glsl/opt_vectorize.cpp @@ -82,6 +82,7 @@ public: virtual ir_visitor_status visit_enter(ir_assignment *); virtual ir_visitor_status visit_enter(ir_swizzle *); + virtual ir_visitor_status visit_enter(ir_dereference_array *); virtual ir_visitor_status visit_enter(ir_if *); virtual ir_visitor_status visit_enter(ir_loop *); @@ -289,6 +290,19 @@ ir_vectorize_visitor::visit_enter(ir_swizzle *ir) return visit_continue; } +/* Upon entering an ir_array_dereference, remove the current assignment from + * further consideration. Since the index of an array dereference must scalar, + * we are not able to vectorize it. + * + * FINISHME: If all of scalar indices are identical we could vectorize. + */ +ir_visitor_status +ir_vectorize_visitor::visit_enter(ir_dereference_array *ir) +{ + this->current_assignment = NULL; + return visit_continue_with_parent; +} + /* Since there is no statement to visit between the "then" and "else" * instructions try to vectorize before, in between, and after them to avoid * combining statements from different basic blocks. commit 088d642b8fc1e1cf5ae0ec7b222413218b177b63 Author: Emil Velikov <[email protected]> Date: Sat Jan 25 15:23:41 2014 +0000 dri/nouveau: Pass the API into _mesa_initialize_context Currently we create a OPENGL_COMPAT context regardless of what was requested by the program. Correct that by retaining the program's request and passing it into _mesa_initialize_context. Based on a similar commit for radeon/r200 by Ian Romanick. Cc: "9.1 9.2 10.0" <[email protected]> Signed-off-by: Emil Velikov <[email protected]> (cherry picked from commit 76d9f6d9729db1c999317a6b44818aa90c30a0b3) diff --git a/src/mesa/drivers/dri/nouveau/nouveau_context.c b/src/mesa/drivers/dri/nouveau/nouveau_context.c index b84205b..f8c8dc3 100644 --- a/src/mesa/drivers/dri/nouveau/nouveau_context.c +++ b/src/mesa/drivers/dri/nouveau/nouveau_context.c @@ -72,7 +72,7 @@ nouveau_context_create(gl_api api, return false; } - ctx = screen->driver->context_create(screen, visual, share_ctx); + ctx = screen->driver->context_create(screen, api, visual, share_ctx); if (!ctx) { *error = __DRI_CTX_ERROR_NO_MEMORY; return GL_FALSE; @@ -107,7 +107,8 @@ nouveau_context_create(gl_api api, } GLboolean -nouveau_context_init(struct gl_context *ctx, struct nouveau_screen *screen, +nouveau_context_init(struct gl_context *ctx, gl_api api, + struct nouveau_screen *screen, const struct gl_config *visual, struct gl_context *share_ctx) { struct nouveau_context *nctx = to_nouveau_context(ctx); @@ -125,7 +126,7 @@ nouveau_context_init(struct gl_context *ctx, struct nouveau_screen *screen, nouveau_fbo_functions_init(&functions); /* Initialize the mesa context. */ - _mesa_initialize_context(ctx, API_OPENGL_COMPAT, visual, + _mesa_initialize_context(ctx, api, visual, share_ctx, &functions); nouveau_state_init(ctx); diff --git a/src/mesa/drivers/dri/nouveau/nouveau_context.h b/src/mesa/drivers/dri/nouveau/nouveau_context.h index 07d9605..8ea431b 100644 --- a/src/mesa/drivers/dri/nouveau/nouveau_context.h +++ b/src/mesa/drivers/dri/nouveau/nouveau_context.h @@ -115,7 +115,8 @@ nouveau_context_create(gl_api api, void *share_ctx); GLboolean -nouveau_context_init(struct gl_context *ctx, struct nouveau_screen *screen, +nouveau_context_init(struct gl_context *ctx, gl_api api, + struct nouveau_screen *screen, const struct gl_config *visual, struct gl_context *share_ctx); void diff --git a/src/mesa/drivers/dri/nouveau/nouveau_driver.h b/src/mesa/drivers/dri/nouveau/nouveau_driver.h index e03b2c1..b6a8276 100644 --- a/src/mesa/drivers/dri/nouveau/nouveau_driver.h +++ b/src/mesa/drivers/dri/nouveau/nouveau_driver.h @@ -48,6 +48,7 @@ struct nouveau_driver { struct gl_context *(*context_create)(struct nouveau_screen *screen, + gl_api api, const struct gl_config *visual, struct gl_context *share_ctx); void (*context_destroy)(struct gl_context *ctx); diff --git a/src/mesa/drivers/dri/nouveau/nv04_context.c b/src/mesa/drivers/dri/nouveau/nv04_context.c index 93dcd87..cd6e876 100644 --- a/src/mesa/drivers/dri/nouveau/nv04_context.c +++ b/src/mesa/drivers/dri/nouveau/nv04_context.c @@ -138,7 +138,8 @@ nv04_context_destroy(struct gl_context *ctx) } static struct gl_context * -nv04_context_create(struct nouveau_screen *screen, const struct gl_config *visual, +nv04_context_create(struct nouveau_screen *screen, gl_api api, + const struct gl_config *visual, struct gl_context *share_ctx) { struct nv04_context *nctx; @@ -153,7 +154,7 @@ nv04_context_create(struct nouveau_screen *screen, const struct gl_config *visua ctx = &nctx->base.base; hw = &nctx->base.hw; - if (!nouveau_context_init(ctx, screen, visual, share_ctx)) + if (!nouveau_context_init(ctx, api, screen, visual, share_ctx)) goto fail; /* GL constants. */ diff --git a/src/mesa/drivers/dri/nouveau/nv10_context.c b/src/mesa/drivers/dri/nouveau/nv10_context.c index 8582cb2..904a410 100644 --- a/src/mesa/drivers/dri/nouveau/nv10_context.c +++ b/src/mesa/drivers/dri/nouveau/nv10_context.c @@ -427,7 +427,8 @@ nv10_context_destroy(struct gl_context *ctx) } static struct gl_context * -nv10_context_create(struct nouveau_screen *screen, const struct gl_config *visual, +nv10_context_create(struct nouveau_screen *screen, gl_api api, + const struct gl_config *visual, struct gl_context *share_ctx) { struct nouveau_context *nctx; @@ -441,7 +442,7 @@ nv10_context_create(struct nouveau_screen *screen, const struct gl_config *visua ctx = &nctx->base; - if (!nouveau_context_init(ctx, screen, visual, share_ctx)) + if (!nouveau_context_init(ctx, api, screen, visual, share_ctx)) goto fail; ctx->Extensions.ARB_texture_env_crossbar = true; diff --git a/src/mesa/drivers/dri/nouveau/nv20_context.c b/src/mesa/drivers/dri/nouveau/nv20_context.c index 618280f..14329a2 100644 --- a/src/mesa/drivers/dri/nouveau/nv20_context.c +++ b/src/mesa/drivers/dri/nouveau/nv20_context.c @@ -438,7 +438,8 @@ nv20_context_destroy(struct gl_context *ctx) } static struct gl_context * -nv20_context_create(struct nouveau_screen *screen, const struct gl_config *visual, +nv20_context_create(struct nouveau_screen *screen, gl_api api, + const struct gl_config *visual, struct gl_context *share_ctx) { struct nouveau_context *nctx; @@ -452,7 +453,7 @@ nv20_context_create(struct nouveau_screen *screen, const struct gl_config *visua ctx = &nctx->base; - if (!nouveau_context_init(ctx, screen, visual, share_ctx)) + if (!nouveau_context_init(ctx, api, screen, visual, share_ctx)) goto fail; ctx->Extensions.ARB_texture_env_crossbar = true; commit 69bd4ed017f7cc8bae20fb6da0aba249025b0776 Author: Daniel Kurtz <[email protected]> Date: Fri Feb 7 19:29:06 2014 +0800 glsl: Add locking to builtin_builder singleton Consider a multithreaded program with two contexts A and B, and the following scenario: 1. Context A calls initialize(), which allocates mem_ctx and starts building built-ins. 2. Context B calls initialize(), which sees mem_ctx != NULL and assumes everything is already set up. It returns. 3. Context B calls find(), which fails to find the built-in since it hasn't been created yet. 4. Context A finally finishes initializing the built-ins. This will break at step 3. Adding a lock ensures that subsequent callers of initialize() will wait until initialization is actually complete. Similarly, if any thread calls release while another thread is still initializing, or calling find(), the mem_ctx/shader would get free'd while from under it, leading to corruption or use-after-free crashes. Fixes sporadic failures in Piglit's glx-multithread-shader-compile. Bugzilla: https://bugs.freedesktop.org/69200 Signed-off-by: Daniel Kurtz <[email protected]> Signed-off-by: Kenneth Graunke <[email protected]> -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected] Archive: http://lists.debian.org/[email protected]
