Mesa 7.10 Release Notes / January 7, 2011
Mesa 7.10 is a new development release.
People who are concerned with stability and reliability should stick
with a previous release or wait for Mesa 7.10.1.
Mesa 7.10 implements the OpenGL 2.1 API, but the version reported by
glGetString(GL_VERSION) depends on the particular driver being used.
Some drivers don't support all the features required in OpenGL 2.1.
See the Compiling/Installing page for prerequisites
for DRI hardware acceleration.
MD5 checksums
tbd
New features
- GL_ARB_explicit_attrib_location extension (Intel and software drivers).
- GL_ARB_texture_rg (Intel, software drivers, gallium drivers).
- GL_EXT_separate_shader_objects extension (Intel and software drivers).
- GL_NV_primitive_restart extension (Gallium softpipe, llvmpipe).
- New fragment shader back-end for i965-class hardware.
- Support for Sandybridge chipset in i965 DRI driver.
Bug fixes
This list is likely incomplete.
- Bug 28800 - [r300c, r300g] Texture corruption with World of Warcraft
- Bug 29420 - Amnesia / HPL2 RendererFeatTest - not rendering correctly
- Bug 29946 - [swrast] piglit valgrind glsl-array-bounds-04 fails
- Bug 30261 - [GLSL 1.20] allowing inconsistent invariant declaration between two vertex shaders
- Bug 30632 - [softpipe] state_tracker/st_manager.c:489: st_context_notify_invalid_framebuffer: Assertion `stfb && stfb->iface == stfbi' failed.
- Bug 30694 - wincopy will crash on Gallium drivers when going to front buffer
- Bug 30771 - [r600g] vert-tex glsl demo
- Bug 30787 - Invalid asm shader does not generate draw-time error when used with GLSL shader
- Bug 30974 - [llvmpipe] SIGABRT src/gallium/drivers/llvmpipe/lp_state_fs.c:779
- Bug 30993 - getFramebufferAttachmentParameteriv wrongly generates error
- Bug 31101 - [glsl2] abort() in ir_validate::visit_enter(ir_assignment *ir)
- Bug 31193 - [regression] aa43176e break water reflections
- Bug 31194 - The mesa meta save/restore code doesn't ref the current GLSL program
- Bug 31371 - glslparsertest: ir.cpp:358: ir_constant::ir_constant(const glsl_type*, const ir_constant_data*): Assertion `(type->base_type >= 0) && (type->base_type <= 3)' failed.
- Bug 31439 - Crash in glBufferSubData() with size == 0
- Bug 31495 - [i965 gles2c bisected] OpenGL ES 2.0 conformance GL2Tests_GetBIFD_input.run regressed
- Bug 31514 - isBuffer returns true for unbound buffers
- Bug 31560 - [tdfx] tdfx_tex.c:702: error: ‘const struct gl_color_table’ has no member named ‘Format’
- Bug 31617 - Radeon/Compiz: 'failed to attach dri2 front buffer', error case not handled
- Bug 31648 - [GLSL] array-struct-array gets assertion: `(size >= 1) && (size <= 4)' failed.
- Bug 31650 - [GLSL] varying gl_TexCoord fails to be re-declared to different size in the second shader
- Bug 31673 - GL_FRAGMENT_PRECISION_HIGH preprocessor macro undefined in GLSL ES
- Bug 31690 - i915 shader compiler fails to flatten if in Aquarium webgl demo.
- Bug 31832 - [i915] Bad renderbuffer format: 21
- Bug 31841 - [drm:radeon_cs_ioctl] *ERROR* Invalid command stream !
- Bug 31894 - Writing to gl_PointSize with GLES2 corrupts other varyings
- Bug 31909 - [i965] brw_fs.cpp:1461: void fs_visitor::emit_bool_to_cond_code(ir_rvalue*): Assertion `expr->operands[i]->type->is_scalar()' failed.
- Bug 31934 - [gallium] Mapping empty buffer object causes SIGSEGV
- Bug 31983 - [i915 gles2] "if (expression with builtin/varying variables) discard" breaks linkage
- Bug 31985 - [GLSL 1.20] initialized uniform array considered as "unsized"
- Bug 31987 - [gles2] if input a wrong pname(GL_NONE) to glGetBoolean, it will not case GL_INVALID_ENUM
- Bug 32035 - [GLSL bisected] comparing unsized array gets segfault
- Bug 32070 - llvmpipe renders stencil demo incorrectly
- Bug 32273 - assertion fails when starting vdrift 2010 release with shaders enabled
- Bug 32287 - [bisected GLSL] float-int failure
- Bug 32311 - [965 bisected] Array look-ups broken on GM45
- Bug 32520 - [gles2] glBlendFunc(GL_ZERO, GL_DST_COLOR) will result in GL_INVALID_ENUM
- Bug 32825 - egl_glx driver completely broken in 7.9 branch [fix in master]
- Bug -
- Bug -
Changes
Adam Jackson (2):
- i965: Update renderer strings for sandybridge
- drivers/x11: unifdef XFree86Server
Alex Deucher (30):
- r600c: fix mipmap stride on evergreen
- r600c: add reloc for CB_COLOR0_ATTRIB
- r600c: pull over 6xx/7xx vertex fixes for evergreen
- r600c: fix segfault in evergreen stencil code
- r100: revalidate after radeon_update_renderbuffers
- r600c: add missing radeon_prepare_render() call on evergreen
- r600c: properly align mipmaps to group size
- egl_dri2: Add radeon chip ids
- r600c/evergreen: texture align is group_bytes just like 6xx/7xx
- r600g: fix buffer alignment
- r600g: All EVENT_WRITE packets need the EVENT_INDEX field
- r600g: translate ARR instruction for evergreen
- r600g: use meaningful defines for chiprev
- r600g: use full range of VS resources for vertex samplers
- r600g: fix additional EVENT_WRITE packet
- r600g: fix some winsys functions to deal properly with evergreen
- r600c: add Ontario Fusion APU support
- r600g: add support for ontario APUs
- r600c: fix VC flush on cedar and palm
- gallium/egl: fix r300 vs r600 loading
- r600c: fix some opcodes on evergreen
- r600c: bump texture limits to hw limits
- r600g: bump texture/cb limits appropriately for evergreen
- radeon: bump mip tree levels to 15
- r600g: fix rendering with a vertex attrib having a zero stride
- r600g: remove useless switch statements
- r600g: add support for NI (northern islands) asics
- r600c: add support for NI asics
- r600g: support up to 64 shader constants
- r600c: fix up SQ setup in blit code for Ontario/NI
Andre Maasikas (3):
- r600c: fix buffer height setting in dri2 case
- r600g: break alu clause earlier
- r600g: fix evergreen interpolation setup
Andrew Randrianasulu (2):
- dri/nv04: Don't expose ARB_texture_env_combine/dot3.
- dri/nv04: Enable eng3dm for A8/L8 textures.
Aras Pranckevicius (2):
- glsl: fix crash in loop analysis when some controls can't be determined
- glsl: fix matrix type check in ir_algebraic
Bas Nieuwenhuizen (3):
- r600g: set ENABLE_KILL in the shader state in the new design
- r600g: set ENABLE_KILL on evergreen too
- r600g: use dirty list to track dirty blocks
Ben Skeggs (3):
- nv50: DST
- nv50: DPH
- nv50: silence some unknown get_param warnings
Benjamin Franzke (2):
- st/egl image: multiply drm buf-stride with blocksize
- r600g: implement texture_get_handle (needed for eglExportDRMImageMESA)
Brian Paul (296):
- glx: add const qualifiers to __indirect_glMultiDrawArraysEXT()
- glsl2: fix signed/unsigned comparison warning
- llvmpipe: cast to silence warning
- llvmpipe: s/boolean/unsigned/ in bitfield to silence warning
- nv50: use unsigned int for bitfields to silence warnings
- tgsi: fix incorrect usage_mask for shadow tex instructions
- gallivm: expand AoS sampling to cover all filtering modes
- gallivm: fix incorrect vector shuffle datatype
- gallivm: move i32_vec_type inside the #ifdef
- mesa: include mfeatures.h in formats.c
- gallivm: fix wrong return value in bitwise functions
- tgsi/sse: fix aos_to_soa() loop to handle num_inputs==0
- gallivm: added missing case for PIPE_TEXTURE_RECT
- gallium: better docs for pipe_rasterizer_state::sprite_coord_enable
- gallium: rework handling of sprite_coord_enable state
- gallium/docs: added new pipeline.txt diagram
- mesa: don't call valid_texture_object() in non-debug builds
- glsl2: silence compiler warnings in printf() calls
- docs: remove old broken link
- docs: mark as obsolete, remove dead links
- llvmpipe: fix query bug when no there's no scene
- gallivm: remove debug code
- llvmpipe: maintain fragment shader state for draw module
- llvmpipe: indentation fix
- llvmpipe: reformatting, remove trailing whitespace, etc
- llvmpipe: clean-up, comments in setup_point_coefficient()
- llvmpipe: rename sprite field, add sprite_coord_origin
- llvmpipe: implement sprite coord origin modes
- draw: fix test for using the wide-point stage
- llvmpipe: check bitshift against PIPE_MAX_SHADER_OUTPUTS
- draw: check bitshift against PIPE_MAX_SHADER_OUTPUS
- Merge branch 'sprite-coord'
- draw: new draw_fs.[ch] files
- glsl2: fix typo in error msg
- gallivm: fix lp_build_sample_compare()
- softpipe: add missing calls to set draw vertex samplers/views
- mesa: don't advertise bogus GL_ARB_shading_language_120 extension
- configs: remove egl-swrast target from linux-dri config
- llvmpipe: fix sprite texcoord setup for non-projective texturing
- mesa: fix assertions to handle srgb formats
- st/mesa: add missing MESA_FORMAT_S8 case in st_mesa_format_to_pipe_format()
- st/mesa: use the wrapped renderbuffer in CopyPixels()
- llvmpipe: make min/max lod and lod bias dynamic state
- llvmpipe: make texture border_color dynamic state
- softpipe: fix repeat() function for NPOT textures
- gallivm: fix repeat() function for NPOT textures
- swrast: update comments for REMAINDER() macro
- softpipe: rename sp_state_fs.c -> sp_state_shader.c
- softpipe: make shader-related functions static
- softpipe: make blend/stencil/depth functions static
- softpipe: make sampler state functions static
- softpipe: make vertex state functions static
- softpipe: make rasterizer state functions static
- softpipe: make stream out state functions static
- softpipe: make clip state functions static
- softpipe: minor asst. clean-ups
- softpipe: allocate tile data on demand
- llvmpipe: fix swizzling of texture border color
- softpipe: fix swizzling of texture border color
- draw: pass sampler state down to llvm jit state
- gallivm: check for level=0 case in lp_build_minify()
- gallivm: added some comments
- draw: check for null sampler pointers
- swrast: fix choose_depth_texture_level() to respect mipmap filtering state
- st/mesa: replace assertion w/ conditional in framebuffer invalidation
- glsl2: fix signed/unsigned comparison warning
- st/xlib: add some comments
- ir_to_mesa: assorted clean-ups, const qualifiers, new comments
- mesa: remove assertion w/ undeclared variable texelBytes
- gallivm: remove newlines
- draw/llvmpipe: replace DRAW_MAX_TEXTURE_LEVELS with PIPE_MAX_TEXTURE_LEVELS
- mesa: reformatting, comments, code movement
- x11: fix breakage from gl_config::visualType removal
- gallivm: work-around trilinear mipmap filtering regression with LLVM 2.8
- mesa: remove post-convolution width/height vars
- gallivm: add compile-time option to emit inst addrs and/or line numbers
- llvmpipe: code to dump bytecode to file (disabled)
- gallivm: added lp_build_print_vec4()
- gallivm: added lp_build_load_volatile()
- glsl: add ir_unop_round_even case to silence unhandled enum warning
- st/mesa: fix regressions in glDrawPixels(GL_STENCIL_INDEX)
- st/mesa: reformatting in st_cb_drawpixels.c
- st/mesa: use GLuint to avoid problem w/ uint not defined on mingw32
- st/mesa: update function name, comments
- gallivm: use util_snprintf()
- llvmpipe: remove lp_setup_coef*.c files from Makefile
- mesa: fix mesa version string construction
- gallivm: fix incorrect type for zero vector in emit_kilp()
- llvmpipe/draw: always enable LLVMAddInstructionCombiningPass()
- draw: use float version of LLVM Mul/Add instructions
- draw: fix typo in comment
- mesa: add GL_RG case to _mesa_source_buffer_exists()
- mesa: add missing cases for packing red/green images
- st/mesa: added cases for GL_COMPRESSED_RED/RG in st_choose_format()
- docs: update texture red/green support in GL3.txt
- docs: add GL_ARB_texture_rg to release notes
- mesa: driver hook for primitive restart
- mesa: set/get primitive restart state
- mesa: API spec for primitive restart
- mesa: regenerated files with primitive restart
- mesa: plug in primitive restart function
- vbo: support for primitive restart
- gallium: new CAP, state for primitive restart
- st/mesa: support for primitive restart
- draw: implement primitive splitting for primitive restart
- softpipe: enable primitive restart
- llvmpipe: enable primitive restart
- docs: added GL_NV_primitive_restart extension
- Merge branch 'primitive-restart-cleanup'
- winsys/xlib: formatting fixes
- winsys/xlib: use Bool type for shm field
- winsys/xlib: fix up allocation/dealloction of XImage
- winsys/xlib: rename xm->xlib
- galahad: silence warnings
- mesa: move declaration before code
- docs: updated GL3 status for primitive restart
- mesa: 80-column wrapping
- mesa: simplify fbo format checking code
- mesa: split up the image.c file
- mesa: add pixel packing for unscaled integer types
- mesa: _mesa_ClearColorIuiEXT() and _mesa_ClearColorIiEXT()
- mesa: _mesa_is_format_integer() function
- mesa: minor reformatting, clean-ups
- mesa: added _mesa_is_fragment_shader_active() helper
- mesa: new glDrawPixels error check for integer formats
- softpipe: added some texture sample debug code (disabled)
- mesa: added new gl_extensions::EXT_gpu_shader4 field
- mesa: added new gl_framebuffer::_IntegerColor field
- mesa: added glGet query for GL_RGBA_INTEGER_MODE_EXT
- mesa: compute _IntegerColor field in _mesa_test_framebuffer_completeness()
- mesa: added cases for GL_EXT_texture_integer formats
- mesa: added cases for GL_EXT_texture_integer
- st/mesa: add format selection for signed/unsigned integer formats
- mesa: simplify target_can_be_compressed() function
- glapi: GL_EXT_texture_integer API
- glapi: include/build EXT_texture_integer.xml
- mesa: regenerated API files for GL_EXT_texture_integer
- mesa: plug in GL_EXT_texture_integer functions
- mesa: display list support for GL_EXT_texture_integer
- st/mesa: be smarter choosing texture format for glDrawPixels()
- softpipe: remove >32bpp color restriction
- mesa: silence enum comparison warning
- mesa: fix uninitialized var warning
- xlib: silence unused var warning
- util: use pointer_to_func() to silence warning
- rtasm: use pointer_to_func() to silence warning
- translate: use function typedefs, casts to silence warnings
- translate: remove unused prototypes
- mesa: additional glReadPixels error checks for GL_EXT_texture_integer
- mesa: additional switch cases for GL_EXT_texture_integer
- mesa: additional teximage error checks for GL_EXT_texture_integer
- mesa: do integer FB / shader validation check in _mesa_valid_to_render()
- mesa: call _mesa_valid_to_render() in glDrawPixels, glCopyPixels, glBitmap
- mesa: remove the unused _mesa_is_fragment_shader_active() function
- mesa: fix bug in _mesa_is_format_integer()
- mesa: rename function to _mesa_is_format_integer_color()
- mesa: remove 'normalized' parameter from _mesa_VertexAttribIPointer()
- vbo: re-indent file
- glapi: xml spec file for GL_EXT_gpu_shader4
- glapi: include EXT_gpu_shader4.xml
- glapi: regenerated API files
- mesa: plug in stubs for glBindFragDataLocation(), glGetFragDataLocation()
- mesa: add glGetUniformuiv(), plug in uint glUniform funcs
- mesa: plug in more GL_EXT_gpu_shader4 functions
- mesa: add new GLvertexformat entries for integer-valued attributes
- mesa: implement integer-valued vertex attribute functions
- mesa: add gl_client_array::Integer field and related vertex array state code
- mesa: consolidate glVertex/Color/etcPointer() code
- mesa: state/queries for GL_MIN/MAX_PROGRAM_TEXEL_OFFSET_EXT
- mesa: glArrayElement support for integer-valued arrays
- mesa: clean-up array element code
- mesa: add extension table entry for GL_EXT_gpu_shader4
- mesa: remove obsolete comment
- mesa: fix incorrect type in _mesa_texstore_rgba_int16()
- mesa: fix integer cases in _mesa_is_legal_format_and_type()
- mesa: add const qualifier to _mesa_is_legal_format_and_type()
- mesa: additional integer formats in _mesa_bytes_per_pixel()
- mesa: pixel transfer ops do not apply to integer-valued textures
- mesa: remove dead code
- osmesa: fix renderbuffer memleak in OSMesaMakeCurrent()
- mesa: use GLubyte for edge flag arrays
- mesa: move the gl_config struct declaration
- dri/util: add a bunch of comments
- mesa: remove always-false conditional in check_compatible()
- mesa: fix aux/accum comment and error message mixups
- llvmpipe: assign context's frag shader pointer before using it
- llvmpipe: add a cast
- mesa: silence new warnings in texobj.c
- egl/gdi: fix typo: xsurf->gsurf
- mesa: code to unpack RGBA as uints
- gallivm: implement scatter stores into temp register file
- gallivm: add some LLVM var labels
- gallivm: added debug code to dump temp registers
- gallivm: add pixel offsets in scatter stores
- gallivm: added lp_elem_type()
- gallivm: implement execution mask for scatter stores
- tgsi: remove unused function
- llvmpipe: added some debug assertions, but disabled
- gallivm: alloca() was called too often for temporary arrays
- gallivm: add const qualifiers, fix comment string
- softpipe: disable vertex texturing with draw/llvm
- mesa: consolidate pixel packing/unpacking code
- mesa: rename vars in pixel pack/unpack code
- mesa: implement uint texstore code
- mesa: remove stray GL_FLOAT case in _mesa_is_legal_format_and_type()
- mesa: make fixed-pt and byte-valued arrays a runtime feature
- softpipe: can't no-op depth test stage when occlusion query is enabled
- mesa: no-op glBufferSubData() on size==0
- mesa: #include mfeatures.h in enums.h
- mesa: improve error message
- mesa: add missing formats in _mesa_format_to_type_and_comps()
- mesa: handle more pixel types in mipmap generation code
- mesa: make glIsBuffer() return false for never bound buffers
- mesa: fix glDeleteBuffers() regression
- tdfx: s/Format/_BaseFormat/
- mesa: consolidate assertions in teximage code
- radeon: set gl_texture_image::TexFormat field in radeonSetTexBuffer2()
- r600: set gl_texture_image::TexFormat field in r600SetTexBuffer2()
- r200: set gl_texture_image::TexFormat field in r200SetTexBuffer2()
- r300: set gl_texture_image::TexFormat field in r300SetTexBuffer2()
- evergreen: set gl_texture_image::TexFormat field in evergreenSetTexBuffer()
- st/mesa: fix glDrawPixels(depth/stencil) bugs
- glsl: fix assorted MSVC warnings
- mesa: add more work-arounds for acoshf(), asinhf(), atahf()
- glsl: remove opt_constant_expression.cpp from SConscript
- mesa: fix error messages and minor reindenting
- mesa: whitespace cleanups
- mesa: 80-column wrapping
- mesa: reorder texture_error_check() params
- mesa: minor clean-ups in context code
- mesa: upgrade to glext.h version 66
- mesa: pass gl_format to _mesa_init_teximage_fields()
- mesa: fix error msg typo
- glapi: rename GL3.xml to GL3x.xml as it covers all GL 3.x versions
- mesa: hook up GL 3.x entrypoints
- docs: update some GL 3.0 status
- mesa: fix get_texture_dimensions() for texture array targets
- swrast: init alpha value to 1.0 in opt_sample_rgb_2d()
- glsl: fix off by one in register index assertion
- glsl: use gl_register_file in a few places
- mesa: rename, make _mesa_register_file_name() non-static
- mesa: _mesa_valid_register_index() to validate register indexes
- mesa: replace #defines with new gl_shader_type enum
- mesa: use gl_shader_type enum
- glsl: better handling of linker failures
- glsl: start restoring some geometry shader code
- mesa: add assertion and update comment in _mesa_format_image_size()
- mesa: added _mesa_format_image_size64()
- x11: remove test_proxy_teximage() function
- st/mesa: fix mapping of zero-sized buffer objects
- gallivm/llvmpipe: squash merge of the llvm-context branch
- mesa: raise max texture sizes to 16K
- softpipe: increase max texture size to 16K
- mesa: replace large/MAX_WIDTH stack allocations with heap allocations
- mesa: replace large/MAX_WIDTH stack allocations with heap allocations
- swrast: avoid large stack allocations in blend code
- swrast: avoid large stack allocations in tex combine code
- st/mesa: avoid large stack allocations in readpixels code
- mesa: replace more MAX_WIDTH stack allocations with heap allocations
- gallivm/llvmpipe: remove lp_build_context::builder
- gallivm: fix null builder pointers
- mesa: fix GL_FRAMEBUFFER_ATTACHMENT_OBJECT_NAME query
- mesa: return GL_FRAMEBUFFER_DEFAULT as FBO attachment type
- llvmpipe: fix broken stencil writemask
- mesa: consolidate some compiler -D flags
- swrast: allow GL_RG format in glDrawPixels()
- swrast: fix indentation
- swrast: accept GL_RG in glReadPixels()
- swrast: restructure some glReadPixels() code
- mesa: make glGet*(GL_NONE) generate GL_INVALID_ENUM
- mesa: remove unneeded cast
- mesa: update comments, remove dead code
- st/mesa: new comment about updating state vars
- mesa: add error margin to clip mask debug/check code
- gallium/util: minor formatting fixes
- mesa/llvm: use llvm-config --cppflags
- st/mesa: fix mipmap generation bug
- mesa: test for cube map completeness in glGenerateMipmap()
- mesa: set gl_texture_object::_Complete=FALSE in incomplete()
- mesa: consolidate glTexImage1/2/3D() code
- mesa: simplify proxy texture code in texture_error_check()
- mesa: consolidate the glTexSubImage1/2/3D() functions
- mesa: consolidate glCopyTexImage1/2D() code
- mesa: consolidate glCopyTexSubImage1/2/3D() functions
- mesa: consolidate glCompressedTexImage1/2/3D() functions
- mesa: make _mesa_test_proxy_teximage() easier to read
- configure: use llvm-config --cppflags instead of --cflags
- mesa: revamp error checking for compressed texture images
- mesa: simplify target checking for TexImage functions
- draw/llvm: don't flush in vs_llvm_delete()
- tnl: Initialize gl_program_machine memory in run_vp.
- tnl: a better way to initialize the gl_program_machine memory
- mesa, st/mesa: disable GL_ARB_geometry_shader4
- mesa/meta: fix broken assertion, rename stack depth var
- glsl: new glsl_strtod() wrapper to fix decimal point interpretation
- st/mesa: fix renderbuffer pointer check in st_Clear()
Brian Rogers (1):
- mesa: Add missing else in do_row_3D
Chad Versace (25):
- intel_extensions: Add ability to set GLSL version via environment
- glsl: Add glsl_type::uvecN_type for N=2,3
- glsl: Add lexer rules for uint and uvecN (N=2..4)
- glsl: Changes in generated file glsl_lexer.cpp
- glsl: Add lexer rules for << and >> in GLSL 1.30
- glsl: Change generated file glsl_lexer.cpp
- glsl: Implement ast-to-hir for binary shifts in GLSL 1.30
- glsl: Implement constant expr evaluation for bitwise-not
- glsl: Implement constant expr evaluation for bit-shift ops
- glsl: Implement constant expr evaluation for bitwise logic ops
- glsl: Fix ir validation for bit logic ops
- glsl: Define shift_result_type() in ast_to_hir.cpp
- glsl: Implement ast-to-hir for bit-shift-assignment
- glsl: Define bit_logic_result_type() in ast_to_hir.cpp
- glsl: Implement ast-to-hir for bit-logic ops
- glsl: Fix lexer rule for ^=
- glsl: Commit generated file glsl_lexer.cpp
- glsl: Fix ast-to-hir for ARB_fragment_coord_conventions
- mesa: Fix C++ includes in sampler.cpp
- glsl: Fix ir_expression::constant_expression_value()
- glsl: Fix erroneous cast in ast_jump_statement::hir()
- glsl: Fix Doxygen tag file in recently renamed files
- glsl: Improve usage message for glsl_compiler
- glsl: Fix linker bug in cross_validate_globals()
- glsl: In ast_to_hir, check sampler array indexing
Chia-I Wu (149):
- glapi: Fix build errors for ES.
- glapi: Fix ES build errors again.
- mesa: Update ES APIspec.xml.
- st/xlib: Notify the context when the front/back buffers are swapped.
- targets/egl: Use C++ compiler to link GL/ES state trackers.
- libgl-xlib: Remove unused st_api_create_OpenGL.
- st/egl: Split modeset code support to modeset.c.
- st/egl: Rename kms backend to drm.
- st/egl: s/kms/drm/ on the drm backend.
- egl: Enable drm platform by default.
- egl: Check extensions.
- st/egl: Skip single-buffered configs in EGL.
- mapi: Fix compiler warnings.
- st/egl: Drop context argument from egl_g3d_get_egl_image.
- targets/egl: Fix linking with libdrm.
- st/vega: Fix version check in context creation.
- egl: Use attribute names as the _EGLConfig member names.
- egl: Access config attributes directly.
- st/egl: Access _EGLConfig directly.
- st/egl: Do not finish a fence that is NULL.
- mesa: Remove unused vtxfmt_tmp.h.
- egl_dri2: Drop the use of _egl[SG]etConfigKey.
- egl_glx: Drop the use of [SG]ET_CONFIG_ATTRIB.
- egl_glx: Fix borken driver.
- egl: Move attributes in _EGLImage to _EGLImageAttribs.
- egl: Parse image attributes with _eglParseImageAttribList.
- egl: Move fallback routines to eglfallbacks.c.
- egl: Drop dpy argument from the link functions.
- egl: Minor changes to the _EGLConfig interface.
- egl: Minor changes to the _EGLScreen interface.
- egl: Fix _eglModeLookup.
- st/egl: Fix native_mode refresh mode.
- egl: Add reference count for resources.
- egl: Use reference counting to replace IsLinked or IsBound.
- egl: Fix a false negative check in _eglCheckMakeCurrent.
- st/egl: Use resource reference count for egl_g3d_sync.
- egl_dri2: Fix a typo that make glFlush be called at wrong time.
- glapi: Do not use glapidispatch.h.
- glapi: Move glapidispatch.h to core mesa.
- glapi: Do not use glapioffsets.h.
- glapi: Merge glapioffsets.h into glapidispath.h.
- vbo: Use CALL_* macros.
- mesa: Remove unnecessary glapitable.h includes.
- autoconf: Better client API selection.
- docs: Update egl and openvg docs.
- autoconf: Update configuration info.
- Merge branch 'glapi-reorg'
- targets: Add missing quotes to Makefile.xorg.
- autoconf: st/vega requires --enable-openvg.
- st/mesa: Unreference the sampler view in st_bind_surface.
- autoconf: Tidy configure output for EGL.
- targets/egl: Fix a warning with --disable-opengl build.
- egl: Rework _eglGetSearchPath.
- mesa: Select FEATURE_remap_table when multiple APIs are enabled.
- mesa: Allow contexts of different APIs to coexist.
- egl: Set up the pthread key even TLS is used.
- st/egl: Add native_surface::present callback.
- st/egl: Use native_surface::present callback.
- d3d1x: Use native_surface::present.
- st/egl: Remove flush_frontbuffer and swap_buffers.
- st/egl: Add support for swap interval and swap behavior.
- st/egl: Add support for EGL_MATCH_NATIVE_PIXMAP.
- st/egl: Add extern "C" wrapper to native.h.
- st/egl: Add native_display_buffer interface.
- st/egl: Use native_display_buffer for EGL_MESA_drm_image.
- autoconf: Add --enable-gallium-egl.
- docs: Update egl docs.
- st/dri: Add support for surfaceless current contexts.
- egl_dri2: Fix __DRI_DRI2 version 1 support.
- st/vega: Do not wait NULL fences.
- gallium: Add st_api::name.
- gallium: Add st_context_iface::share to st_api.
- st/wgl: Use st_context_iface::share for DrvShareLists.
- st/glx: Replace MESA_VERSION_STRING by xmesa_get_name.
- mesa: Clean up core.h.
- scons: Define IN_DRI_DRIVER.
- tgsi: Add STENCIL to text parser.
- st/vega: vegaLookupSingle should validate the state.
- st/vega: Set wrap_r for mask and blend samplers.
- st/vega: Fix vgReadPixels with a subrectangle.
- egl_dri2: Fix one context, multiple surfaces.
- auxiliary: util_blit_pixels_tex should restore the viewport.
- st/vega: Fix a crash with empty paths.
- st/vega: Masks and surfaces should share orientation.
- st/vega: No flipping in vg_prepare_blend_surface.
- st/vega: Fix a typo in EXTENDED_BLENDER_OVER_FUNC.
- llvmpipe: Fix build errors on x86.
- st/vega: Overhaul renderer with renderer states.
- st/vega: Add DRAWTEX renderer state.
- st/vega: Add SCISSOR renderer state.
- st/vega: Add CLEAR renderer state for vgClear.
- st/vega: Add FILTER renderer state for image filtering.
- st/vega: Use the renderer for vgMask.
- st/vega: Add POLYGON_STENCIL and POLYGON_FILL renderer state.
- st/vega: Delay fb state update to vg_validate_state.
- st/vega: Use st_framebuffer for fb width/height.
- st/vega: Move g3d states to renderer.
- st/vega: Make shader_bind call into the renderer.
- st/vega: vg_manager should care about only the color buffer.
- st/vega: Clean up vg_context fields and functions.
- st/vega: Clean up renderer fields and functions.
- st/vega: vg_copy_texture and vg_copy_surface should share code.
- st/vega: Get rid of renderer_copy_texture.
- st/vega: Update to latest headers.
- st/vega: Fix image sampler views for alpha-only formats.
- st/vega: Make path_render and path_stroke take a matrix.
- st/vega: Make image_draw take a matrix.
- st/vega: Add primitive text support.
- st/vega: Revive mask layer support.
- st/vega: More flexible shader selection.
- st/vega: Add color transformation support.
- st/vega: Bump version to 1.1.
- st/vega: Fix paint coordinates transformations.
- st/vega: Fix negated logic in image_draw.
- st/vega: Fix degenerate paints.
- st/vega: Simplify radial gradient.
- st/vega: Remove st_inlines.h.
- st/vega: Delay blend texture creation until needed.
- st/vega: Create drawing surface mask as needed.
- st/vega: Initialize pipe states with renderer.
- st/vega: Avoid unnecessary constant bufer upload.
- st/vega: Destroy the pipe context with vg_context.
- st/vega: polygon_array requires a deep free.
- st/egl: Set pipe_resource::array_size to 1.
- st/vega: Set pipe_resource::array_size to 1.
- st/vega: Move vertex transformation to shader.
- st/vega: Add a missing break.
- st/vega: Add some comments to pipeline shaders.
- st/vega: Refactor blend shaders.
- st/vega: Move masking after blending.
- st/vega: Add support for per-channel alpha.
- st/vega: Blending should use premultiplied alpha.
- st/vega: Fix VG_BLEND_MULTIPLY.
- st/vega: Add blend shaders for all blend modes.
- st/vega: Fix pipe blend state for various blend modes.
- egl: _eglFilterArray should not allocate.
- mapi: Rewrite mapi_abi.py to get rid of preprocessor magic.
- vbo: Drop second ATTR macro.
- vbo: Fix GLES2 glVertexAttrib.
- mesa: Do not advertise GL_OES_texture_3D.
- mesa: Fix GL_FIXED arrays.
- mesa: Fix glTexCoordPointer with type GL_FIXED.
- st/egl: Plug pbuffer leaks.
- st/egl: Fix eglCopyBuffers.
- st/egl: Assorted fixes for dri2_display_get_configs.
- docs/egl: Update egl.html.
- st/egl: Fix eglChooseConfig when configs is NULL.
- docs: Add an example for EGL_DRIVERS_PATH.
- autoconf: Fix --with-driver=xlib --enable-openvg.
Chris Wilson (2):
- i915g: Fix closure of full batch buffers
- intel: Check for unsupported texture when finishing using as a render target
Christoph Bumiller (80):
- nv50: import new compiler
- nouveau: update nouveau_class.h
- nv50: introduce the big formats table
- nv50: don't produce MOV immediate to output reg in store opt
- nv50: change back accidentally swapped UNORM,SNORM vertex type
- nv50: add/fix some license headers
- nv50: simple reload elimination and local CSE
- nv50: fix constant_operand opt mul by 2 case
- nv50: permit usage of undefined TGSI TEMPs
- nv50: add missing 2nd source for POW multiplication
- nv50: add signed RGTC1 to format table, allow 2_10_10_10 for vbufs
- nv50: fix for empty BBs
- nv50: insert MOVs also for PHI sources from dominating block
- nv50: explicitly set src type for SET ops
- nv50: fixes for nested IFs
- nv50: don't eliminate loads to dedicated values
- nv50: fix constbuf validation
- nv50: build proper phi functions in the first place
- nv50: fix reg count
- nv50: begin implementing loops
- nv50: more constant folding
- nv50: loops part 2
- nv50: flatten simple IF/ELSE/ENDIF constructs
- nv50: fix thinko in store to output reg possible check
- nv50: generate JOINs for outermost IF clauses
- nv50: more TGSI opcodes (SIN, SCS, ARL, RET, KILP)
- nv50: fix PSIZ and PRIMID mapping
- nv50: check dst compatibility in CSE
- nv50: initialize edgeflag input index
- nv50: emit predicate for interp
- Merge remote branch 'origin/master' into nv50-compiler
- nv50: DP2, fix ARL
- nv50: yet another case we need a nop.exit
- nv50: fix check for sprite/point coord enable
- nv50: handle TEXTURE_SWIZZLE and GEOMETRY_SHADER4 caps
- nv50: set the FragDepth output index
- nv50: turn off verbose debug output by default
- nv50: attempt at making more complicated loops work
- nv50: SSG
- nv50: make FrontFacing -1 or +1
- nv50: re-add proper TEXBIAS sequence
- nv50: make use of TGSI immediate type
- nv50: must join SELECT inputs before MOV inputs
- nv50: fix XPD, was negated
- nv50: fix find_dom_frontier
- nv50: fix build-predicate function
- Merge remote branch 'origin/master' into nv50-compiler
- nv50: load address register before using it, not after
- nv50: save tgsi instructions
- nv50: prepare for having multiple functions
- nv50: don't parse again in tgsi_2_nc
- nv50: use actual loads/stores if TEMPs are accessed indirectly
- nv50: create value references with the right type
- nv50: duplicate interps in load_proj_tex_coords
- nv50: address regs are 16 bit
- nv50: fix can_load check for 3rd source
- nv50: reduce bb_reachable_by runtime from pot to linear
- nv50: minor compiler fixes and cleanups
- nv50: cannot move from local mem to output reg directly
- nv50: newlines in shader bincode printing
- nv50: match TEMP limit with nv50 ir builder
- nv50: handle TGSI EXP and LOG again
- nv50: check for immediates when turning MUL ADD into MAD
- nv50: interp cannot write flags reg
- nv50: MOV TEMP[0], -CONST[0] must be float32 negation
- nv50: fix indirect CONST access with large or negative offsets
- nv50: fix TXP depth comparison value
- nv50: consider address register in reload elimination
- nv50: improve and fix modifier folding optimization
- nv50: put low limit on REG_ALLOC_TEMP and FP_RESULT_COUNT
- Merge remote branch 'origin/nv50-compiler'
- nv50: don't segfault on shaders with 0 instructions
- nv50: get shader fixups/relocations into working state
- nv50: add relocs for stack and local mem buffers
- nv50: emit constbuf relocs before uploading constants
- nv50: fix typo in fifo packet length limit
- nv50: use formats table in nv50_surface.c
- nv50: use CLEAR_BUFFERS for surface fills
- nv50: fix/handle a few more PIPE_CAPs
- nv50: fix GP state bind and validate
Corbin Simpson (8):
- r600g: Use align() instead of handrolled code.
- r600g: Trivially deobfuscate r600_hw_states.
- r600g: Deobfuscate and comment a few more functions in r600_hw_states.
- r600g: Clean up some indentation and |= vs. | usage.
- r600g: Fix false and true.
- r600g: "tmp" is such a bad name for a texture.
- r600g: Clean up PS setup.
- r600g: Cleanup viewport floats.
Daniel Lichtenberger (1):
- radeon: fix potential segfault in renderbuffer update
Daniel Vetter (21):
- r200: revalidate after radeon_update_renderbuffers
- i915g: rip out ->sw_tiled
- i915g: s/hw_tiled/tiling
- i915g: add pineview pci ids
- i915g: kill RGBA/X formats
- i915g: kill buf->map_gtt
- i915g: kill idws->pool
- i915g: drop alignment parameter from iws->buffer_create
- i915g: add winsys function to create tiled buffers
- i915g: switch to tiled allocations, kill set_fence
- i915g: prepare winsys/batchbuffer for execbuf2
- i915g: return tiling in iws->buffer_from_handle
- i915g: implement unfenced color&depth buffer using tiling bits
- i915g: implement unfenced relocs for textures using tiling bits
- i915g: postpone mipmap/face offset calculation
- i915g: don't pot-align stride for tiled buffers
- i915g: enable X-tiling for textures
- i915g: switch rendering to mipmapped textures to (x,y) offsets
- i915g: enable x-tiling for render targets
- i915g: assert(depth_surface->offset == 0)
- i915g: track TODO items
Dave Airlie (182):
- r300g: fix buffer reuse issue caused by previous commit
- r600g: pull r600_draw struct out into header
- r600g: use index min/max + index buffer offset.
- r600g: add vgt dma src defines
- r600g: fixup texture state on evergreen.
- r600g: fix texture bos and avoid doing depth blit on evergreen
- r600g: hide radeon_ctx inside winsys.
- r600g: attempt to abstract kernel bos from pipe driver.
- r600g: move constant buffer creation behind winsys abstraction.
- r600g: use malloc bufmgr for constant buffers
- r600g: add support for kernel bo
- r600g: add winsys bo caching.
- r600g: add upload manager support.
- r600g: fixup map flushing.
- r600g: use calloc for ctx bo allocations
- r600g: oops got the use_mem_constant the wrong way around.
- r600g; add uses waterfall to asm cf for r6xx.
- r600g: only emit uses waterfall on r6xx hw.
- util/r300g: split the r300 index buffer modifier functions out to util
- r600g: modify index buffers for sizes the hw can't deal with.
- r600g: send correct surface base update for multi-cbufs
- r600g: fix fbo-drawbuffers-maxtargets
- r600g: clean up valgrind issues on maxtargets test.
- r600g: drop debugging that snuck in
- r600g: fix tiling support for ddx supplied buffers
- r600g: add z16 to color setup
- r600g: add color/texture support for more depth formats.
- r600g: fix r700 cube map sizing.
- r600g: fixup r700 CB_SHADER_CONTROL register.
- r600g: add missing BC_INST wrapper for evergreen
- r600g: only flush for the correct colorbuffer, not all of them.
- r600g: deal with overflow of VTX/TEX CF clauses.
- r600g: set back to correct codepaths.
- r600g: fixup evergreen miptree setup.
- r600g: fix eg texture borders.
- r600g: fix typo in struct member name
- r600g: cleanup some of the DB blit code
- r600g: make stencil readback work
- r600g: disable dirty handling on texture from depth code.
- r600g: use floats instead of hex for blit vbo
- r600g: fix depth readback on rv610 and other quirky variants.
- r600g: fix typo in evergreen register list
- u_blitter: add a custom blitter call passing a dsa cso
- r600g: use blitter to do db->cb flushing.
- r600g: fix warnings since last commit.
- egl: fix build since 17eace581d25a626a7d75d9d1205d012cbb14a6e
- r300g: fix point sprite coord.
- r600g: add vert support for 16/16 and 16/16/16 floats.
- r600g: add some more vertex format support.
- r600g: some more vertex formats
- r600g: fix draw-elements and draw-elements-base-vertex
- r600g: drop index_offset parameter to index buffer translate.
- r600g: fixup tex wrapping.
- r600g: fixup VP->FP output->input routing.
- r600g: fix typo in r700 alu emit
- r600g: fixup sprite coord enable.
- r600g: fix polygon mode
- mesa/mipmap: fix warning since 1acadebd6270d3604b026842b8a21360968618a0
- r600g: add eg poly mode code.
- r600g: make index bias fix for evergreen
- r600g: add eg db count control register.
- r300g: fix glsl-fs-pointcoord
- r600g: add evergreen texture resource properly.
- r600g: fix db flush breaking config state
- r600g: on evergreen the centroid isn't set in this register.
- r600g: add back evergreen name.
- r600g: add evergreen texture border support to new path
- r600g: move radeon.h members around to add back map flushing.
- r600g: add initial vertex translate support.
- r600g: remove old assert from new codepath
- Revert "r600g: add initial vertex translate support."
- r600g: port r300g fix for X* formats in texformat code
- r600g: add L8A8 unorm.
- r600g: clean up some code from move to new paths.
- r600g: return string for chip family
- r600g: use Elements macro instead of manual sizeofs
- r600g: fix evergreen depth flushing.
- r600g: add winsys support for CTL constants.
- r600g: drop depth quirk on evergreen
- r600g: add reloc for evergreen color attrib
- r600g: realign evergreen code with r600 code.
- r600g: add assembler support for other vtx fetch fields.
- r600g: fixup vertex format picking.
- r600g: sync vertex/texture cache on resources on evergreen
- r600g: add cb flushing for extra buffers + depth buffer on r600/evergreen
- r600g: fix evergreen draw-buffers
- r600g: flush SH cache on constant change on evergreen
- r600g: only set the Z export if shader exports it.
- r600g: setup basic loop consts on r600 + evergreen.
- mesa/st: initial attempt at RG support for gallium drivers
- r600g: break out of search for reloc bo after finding it.
- r600g: the code to check whether a new vertex shader is needed was wrong
- r600g: fix wwarning in bo_map function
- r600g: TODO domain management
- r600g: add bo fenced list.
- pb: don't keep checking buffers after first busy
- r600g: add bo busy backoff.
- r600g: drop mman allocator
- r600g: drop use_mem_constant.
- r600g: avoid unneeded bo wait
- pb: fix numDelayed accounting
- r600g: add evergreen stencil support.
- r600g: use format from the sampler view not from the texture.
- r600g: fix Z export enable bits.
- r600g: add some RG texture format support.
- r600g: drop width/height per level storage.
- r600g: fix input/output Z export mixup for evergreen.
- r600g: evergreen has no request size bit in texture word4
- r600g: enable vertex samplers.
- r600g: add TXL opcode support.
- r600g: don't run with scissors.
- r600g: fix typo in vertex sampling on r600
- gallium/tgsi: add support for stencil writes.
- gallium/format: add support for X24S8 and S8X24 formats.
- gallium/format: add X32_S8X24_USCALED format.
- gallium/util: add S8 tile sampling support.
- mesa: add support for FRAG_RESULT_STENCIL.
- mesa: improve texstore for 8/24 formats and add texstore for S8.
- softpipe: add support for shader stencil export capability
- st/mesa: add option to choose a texture format that we won't render to.
- st/mesa: use shader stencil export to accelerate shader drawpixels.
- r600g: add support for S8, X24S8 and S8X24 sampler formats.
- r600g: add shader stencil export support.
- glsl: add support for shader stencil export
- st/mesa: enable stencil shader export extension if supported
- r600g: fix depth0 setting
- r600g: fix scissor/cliprect confusion
- r600g: store samplers/views across blit when we need to modify them
- r600g: reduce size of context structure.
- r600g: the vs/ps const arrays weren't actually being used.
- r600g: add copy into tiled texture
- r600g: split out miptree setup like r300g
- r600g: use common texture object create function
- r600g: rename pitch in texture to pitch_in_bytes
- r600g: remove bpt and start using pitch_in_bytes/pixels.
- r600g: fix transfer stride.
- r600g: drop all use of unsigned long
- r600g: use blitter for hw copy region
- r600g: evergreen add stencil export bit
- r600g: add missing eg reg definition
- r600g: fix stencil export for evergreen harder
- r600g: drop unused context members
- r600g: only pick centroid coordinate when asked.
- r600g: fixup pos/face ena/address properly
- r600g: fixup typo in macro name
- r600g: select linear interpolate if tgsi input requests it
- r300g: clean up warning due to unknown cap.
- tgsi: add scanner support for centroid inputs
- r600g: evergreen interpolation support.
- r600g: add evergreen ARL support.
- r600g: switch to a common formats.h file since they are in different regs
- r600g: add defines for tiling
- r600g: get tiling info from kernel
- r600g: set tiling bits in hw state
- r600g: do proper tracking of views/samplers.
- r600g: fix typo in tiling setup cb code.
- r600g: depth needs to bound to ds
- r600g: attempt to cleanup depth blit
- r600g: fix transfer function for tiling.
- r600g: retrieve tiling info from kernel for shared buffers.
- r600g: all non-0 mipmap levels need to be w/h aligned to POT.
- r600g: move to per-miplevel array mode.
- r600g: start adding hooks for aligning width/height for tiles.
- r600g: add r600 surface to store the aligned height.
- r600g: introduce a per-driver resource flag for transfers.
- r600g: add texture tiling alignment support.
- r600g: add texture tiling enable under a debug option.
- r600g: initial translate state support
- r600g: start splitting out common code from eg/r600.
- r600g: not fatal if we can't get tiling info from kernel
- r600g: merge more of the common r600/evergreen state handling
- r600g: drop more common state handling code
- r600g: fix magic 0x1 ->flat shade ena
- r600g: add assembler support for all the kcache fields.
- gallium/noop: report GL 2.1
- r600g: pick correct color swap for A8 fbos.
- r300g/r600g: bump cache manager timeouts to 1s
- r600g: it looks like r600 can handle dword offsets in the indices.
- r300g: try and use all of vertex constant space
- r300g: fixup rs690 tiling stride alignment calculations.
- r600g: fix evergreen segfaults.
- r600g: hack around property unknown issues.
Eric Anholt (300):
- glsl: Add definition of gl_TextureMatrix inverse/transpose builtins.
- i965: Share the KIL_NV implementation between glsl and non-glsl.
- i965: Also enable CC statistics when doing OQs.
- i965: Track the windowizer's dispatch for kill pixel, promoted, and OQ
- glsl: Rework assignments with write_masks to have LHS chan count match RHS.
- glsl: Fix copy'n'wasted ir_noop_swizzle conditions.
- ir_to_mesa: Only compare vector_elements present for any_nequal/all_equal
- i965: Fix the vector/expression splitting for the write_mask change.
- i965: When splitting vector variable assignment, ignore unset channels.
- i965: Update expression splitting for the vector-result change to compares.
- i965: Warning fix for vector result any_nequal/all_equal change.
- mesa: Remove the non-required ARB_imaging extension.
- mesa: Remove EXT_histogram.
- mesa: Remove SGI_color_table.
- mesa: Remove SGI_color_matrix.
- mesa: Remove EXT_convolution.
- intel: Remove disabled stencil drawpixels acceleration.
- intel: Remove unnecessary minimum pitch alignment to 32 bytes.
- intel: Replace my intel_texture_bitmap code with _mesa_meta_Bitmap.
- radeon: Remove copied minimum pitch alignment code.
- unichrome: Mostly revert my convolution removal changes.
- intel: Remove dead intelIsTextureResident().
- i915: Remove a dead if (0) block.
- intel: Dead comment removal.
- intel: Corresponding FinishRenderTexture debug to BeginRenderTexture.
- i965: Add support for rendering to SARGB8 FBOs.
- intel: Fix segfault on INTEL_DEBUG=fbo with unsupported framebuffers.
- intel: Add fallback debug to glGenerateMipmap.
- intel: More reverting of the sw fallback for depth texture border color.
- intel: Improve some of the miptree debugging.
- mesa: Fix type typo in glGenerateMipmap handling of GL_UNSIGNED_INT data.
- glsl: Fix broadcast_index of lower_variable_index_to_cond_assign.
- glsl: Add validation that a swizzle only references valid channels.
- i965: Fix up writemasked assignments in the new FS.
- i965: Remove swizzling of assignment to vector-splitting single-channel LHS.
- i965: Handle all_equal/any_nequal in the new FS.
- i965: Fix vector splitting RHS channel selection with sparse writemasks.
- i965: Add support for dFdx()/dFdy() to the FS backend.
- i965: Add support for attribute interpolation on Sandybridge.
- i965: Set up inputs to the fragment shader according to FP InputsRead.
- i965: Add support for POW in gen6 FS.
- i965: Fix negation in the new FS backend.
- i965: Actually track the "if" depth in loop in the new FS backend.
- i965: Apply the same set of lowering passes to new FS as to Mesa IR.
- i965: Fix valgrind complaint about base_ir for new FS debugging.
- i965: Fix up the FS backend for the variable array indexing pass.
- i965: Set the variable type when dereferencing an array.
- i965: Add support for dereferencing structs to the new FS backend.
- i965: Add support for struct, array, and matrix uniforms to FS backend.
- i965: Fix all non-snb regression in the snb attribute interpolation commit.
- i965: Fix up part of my Sandybridge attributes support patch.
- i965: Add support for gl_FrontFacing to the new FS backend.
- i965: Subtract instead of adding when computing y delta in new FS backend.
- mesa: Pull ir_to_mesa's sampler number fetcher out to shared code.
- i965: Set up sampler numbers in the FS backend.
- i965: Add support for non-color render target write data to new FS backend.
- i965: Add support for MRT to the new FS backend.
- i965: Add support for ir_loop counters to the new FS backend.
- i965: Add support for ARB_fragment_coord_conventions to the new FS backend.
- glsl: Also update implicit sizes of varyings at link time.
- i965: Do interpolation for varying matrices and arrays in the FS backend.
- i965: Don't try to emit interpolation for unused varying slots.
- i965: Fix array indexing of arrays of matrices.
- i965: Clean up obsolete FINISHME comment.
- mesa: Move the list of builtin uniform info from ir_to_mesa to shared code.
- i965: Add support for builtin uniforms to the new FS backend.
- i965: Fix use of undefined mem_ctx in vector splitting.
- i956: Make new FS discard do its work in a temp, not the null reg!
- i965: Clean up the virtual GRF handling.
- ra: First cut at a graph-coloring register allocator for mesa.
- i965: First cut at register allocation using graph coloring.
- i965: Add live interval analysis and hook it up to the register allocator.
- i965: Remove my "safety counter" code from loops.
- i965: Fix whole-structure/array assignment in new FS.
- mesa: Don't reference a W component in setting up a vec3 uniform component.
- i965: Fix new FS handling of builtin uniforms with packed scalars in structs.
- glsl: Add a lowering pass for texture projection.
- i965: Use the lowering pass for texture projection.
- i965: Split the gen4 and gen5 sampler handling apart.
- i965: Add gen6 attribute interpolation to new FS backend.
- i965: Fix the gen6 jump size for BREAK/CONT in new FS.
- i965: Also increment attribute location when skipping unused slots.
- i965: Pre-gen6, map VS outputs (not FS inputs) to URB setup in the new FS.
- i965: Add real support for pre-gen5 texture sampling to the new FS.
- i965: Fix up copy'n'pasteo from moving coordinate setup around for gen4.
- i965: Restore the forcing of aligned pairs for delta_xy on chips with PLN.
- i965: When producing a single channel swizzle, don't make a temporary.
- i965: Add a sanity check for register allocation sizes.
- i965: Fix off-by-ones in handling the last members of register classes.
- i965: Don't try to emit code if we failed register allocation.
- i965: Add support for EXT_texture_swizzle to the new FS backend.
- i965: Set up swizzling of shadow compare results for GL_DEPTH_TEXTURE_MODE.
- i965: Fix glean/texSwizzle regression in previous commit.
- i965: Be more conservative on live interval calculation.
- i965: Add trivial dead code elimination in the new FS backend.
- i965: Add initial folding of constants into operand immediate slots.
- i965: In disasm, gen6 fb writes don't put msg reg # in destreg_conditionalmod.
- i965: Add support for gen6 FB writes to the new FS.
- i965: Enable the constant propagation code.
- i965: Also do constant propagation for the second operand of CMP.
- i965: Add back gen6 headerless FB writes to the new FS backend.
- i965: Gen6 no longer has the IFF instruction; always use IF.
- i965: Fix up IF/ELSE/ENDIF for gen6.
- i965: Fix botch in the header_present case in the new FS.
- i965: Add some clarification of the WECtrl field.
- i965: Don't do 1/w multiplication in new FS for gen6
- i965: Gen6's sampler messages are the same as Ironlake.
- i965: Refactor gl_FrontFacing setup out of general variable setup.
- i965: Add support for gl_FrontFacing on gen6.
- i965: Don't assume that WPOS is always provided on gen6 in the new FS.
- i965: Fix gen6 pointsize handling to match pre-gen6.
- i965: Disable emitting if () statements on gen6 until we really fix them.
- i965: Normalize cubemap coordinates like is done in the Mesa IR path.
- mesa: Simplify a bit of _mesa_add_state_reference using memcmp.
- i965: Drop the check for duplicate _mesa_add_state_reference.
- i965: Drop the check for YUV constants in the param list.
- i965: Handle swizzles in the addition of YUV texture constants.
- i965: Fix gen6 WM push constants updates.
- i965: Fix new FS gen6 interpolation for sparsely-populated arrays.
- i965: Enable attribute swizzling (repositioning) in the gen6 SF.
- i965: Add register coalescing to the new FS backend.
- i965: Split FS_OPCODE_DISCARD into two steps.
- i965: Reduce register interference checks for changed FS_OPCODE_DISCARD.
- i965: Move FS backend structures to a header.
- i965: Give the math opcodes information on base mrf/mrf len.
- i965: Give the FB write and texture opcodes the info on base MRF, like math.
- i965: Compute to MRF in the new FS backend.
- i965: Don't consider gen6 math instructions to write to MRFs.
- i965: Add a couple of checks for gen6 math instruction limits.
- i965: Don't compute-to-MRF in gen6 math instructions.
- i965: Expand uniform args to gen6 math to full registers to get hstride == 1.
- i965: Don't compute-to-MRF in gen6 VS math.
- i965: Fix gen6 pixel_[xy] setup to avoid mixing int and float src operands.
- i965: Always use the new FS backend on gen6.
- i965: Fix missing "break;" in i2b/f2b, and missing AND of CMP result.
- intel: Allow CopyTexSubImage to InternalFormat 3/4 textures, like RGB/RGBA.
- i965: Don't rebase the index buffer to min 0 if any arrays are in VBOs.
- i965: Add support for rescaling GL_TEXTURE_RECTANGLE coords to new FS.
- i965: Set class_sizes[] for the aligned reg pair class.
- i965: Update the live interval when coalescing regs.
- i965: Add a pass to the FS to split virtual GRFs to float channels.
- i965: Add a function for handling the move of boolean values to flag regs.
- i965: Add peepholing of conditional mod generation from expressions.
- i965: Enable the new FS backend on pre-gen6 as well.
- i965: Fix texturing on pre-gen5.
- i965: Set the type of the null register to fix gen6 FS comparisons.
- i965: Disable the debug printf I added for FS disasm.
- i965: Fix a weirdness in NOT handling.
- i965: Fix assertion failure on gen6 BufferSubData to busy BO.
- i965: Assert out on gen6 VS constant buffer reads that hang the GPU for now.
- i965: Fix scissor-offscreen on gen6 like we did pre-gen6.
- i965: Avoid blits in BufferCopySubdata on gen6.
- i965: Tell the shader compiler when we expect depth writes for gen6.
- i965: Remove the gen6 emit_mi_flushes I sprinkled around the driver.
- i965: Disable thread dispatch when the FS doesn't do any work.
- i965: Add EU emit support for gen6's new IF instruction with comparison.
- i965: Set the source operand types for gen6 if/else/endif to integer.
- i965: Use the new style of IF statement with embedded comparison on gen6.
- i965: Split register allocation out of the ever-growing brw_fs.cpp.
- i965: Fix gl_FrontFacing emit on pre-gen6.
- i965: Add support for register spilling.
- i965: Don't emit register spill offsets directly into g0.
- i965: Correct scratch space allocation.
- i965: Be more aggressive in tracking live/dead intervals within loops.
- i965: Move the FS disasm/annotation printout to codegen time.
- i965: Add support for pull constants to the new FS backend.
- i965: Add EU code for dword scattered reads (constant buffer array indexing).
- i965: Clarify an XXX comment in FB writes with real info.
- i965: Use SENDC on the first render target write on gen6.
- i965: Clear some undefined fields of g0 when using them for gen6 FB writes.
- i965: Add disasm for the flag register.
- i965: Add support for discard instructions on gen6.
- i965: Handle new ir_unop_round_even in channel expression splitting.
- i965: Fix typo in comment about state flags.
- i965: Set up the constant buffer on gen6 when it's needed.
- i965: Add support for constant buffer loads on gen6.
- i965: Drop the eot argument to read messages, which can never be set.
- i965: Fix VS URB entry sizing.
- i965: Disable register spilling on gen6 until it's fixed.
- i965: Make FS uniforms be the actual type of the uniform at upload time.
- i965: Add user clip planes support to gen6.
- i965: Update gen6 SF state when point state (sprite or attenuation) changes.
- i965: Upload required gen6 VS push constants even when using pull constants.
- i965: Update the gen6 stencil ref state when stencil state changes.
- mesa: Make metaops use program refcounts instead of names.
- mesa: Don't compute an unused texture completeness debug string.
- intel: For batch, use GTT mapping instead of writing to a malloc and copying.
- intel: Annotate debug printout checks with unlikely().
- intel: Remove the magic unaligned memcpy code.
- i965: Remove dead intel_structs.h file.
- intel: Avoid taking logbase2 of several things that we max.
- intel: Remove duplicated teximage miptree to object miptree promotion.
- intel: Remove leftover dri1 locking fields in the context.
- mesa: Fix delayed state flagging for EXT_sso-related program changes.
- intel: Fix the client-side swapbuffers throttling.
- Revert "intel: Fix the client-side swapbuffers throttling."
- i965: Allow OPCODE_SWZ to put immediates in the first arg.
- i965: Add support for math on constants in gen6 brw_wm_glsl.c path.
- i965: Work around strangeness in swizzling/masking of gen6 math.
- i965: re-enable gen6 IF statements in the fragment shader.
- glsl: Free the loop state context when we free the loop state.
- i965: Fix gl_FragCoord inversion when drawing to an FBO.
- i965: Shut up spurious gcc warning about GLSL_TYPE enums.
- mesa: Don't spam the console in a debug build unless some spam is requested.
- i965: Add state dumping for sampler state.
- i965: Add dumping of the sampler default color.
- i965: Fail on loops on gen6 for now until we write the EU emit code for it.
- i965: Eliminate dead code more aggressively.
- mesa: Include C++ files in the makedepend of DRI drivers.
- i965: Fix compute_to_mrf to not move a MRF write up into another live range.
- i965: Just use memset() to clear most members in FS constructors.
- i965: Remove extra n at the end of every instruction in INTEL_DEBUG=wm.
- i965: Fold constants into the second arg of BRW_SEL as well.
- glsl: Add a helper function for determining if an rvalue could be a saturate.
- i965: Recognize saturates and turn them into a saturated mov.
- ir_to_mesa: Detect and emit MOV_SATs for saturate constructs.
- i965: Improve compute-to-mrf.
- i965: Remove duplicate MRF writes in the FS backend.
- i965: Move gen4 blend constant color to the gen4 blending file.
- i965: Don't upload polygon stipple unless required.
- i965: Don't upload line stipple pattern unless we're stippling.
- i965: Don't upload line smooth params unless we're line smoothing.
- i965: Use the new embedded compare in SEL on gen6 for VS MIN and MAX opcodes.
- i965: Fix type of gl_FragData[] dereference for FB write.
- glsl: Make the symbol table's add_function just use the function's name.
- glsl: Make the symbol table's add_variable just use the variable's name.
- glsl: Add a helper constructor for expressions that works out result type.
- glsl: Fix structure and array comparisions.
- glsl: Quiet unreachable no-return-from-function warning.
- i965: Dump the WHILE jump distance on gen6.
- i965: Add support for gen6 DO/WHILE ISA emit.
- i965: Add support for gen6 BREAK ISA emit.
- i965: Add support for gen6 CONTINUE instruction emit.
- i965: Enable IF statements in the VS.
- i965: Add support for loops in the VS.
- glsl: Mark the array access for whole-array comparisons.
- glsl: Fix flipped return of has_value() for array constants.
- mesa: Add getters for the rest of the supported draw buffers.
- mesa: Add getters for ARB_copy_buffer's attachment points.
- intel: Add an env var override to execute for a different GPU revision.
- i965: Update gen6 WM state on compiled program change, not just FP change.
- i965: Update gen6 SF state on fragment program change too.
- i965: Fix compile warning about missing opcodes.
- i965: Move payload reg setup to compile, not lookup time.
- i965: Provide delta_xy reg to gen6 non-GLSL path PINTERP.
- i965: Fix up 16-wide gen6 FB writes after various refactoring.
- i965: Don't smash a group of coordinates doing gen6 16-wide sampler headers.
- i965: Fix gen6 interpolation setup for 16-wide.
- i965: Fix up gen6 samplers for their usage by brw_wm_emit.c
- i965: Make the sampler's implied move on gen6 be a raw move.
- i965: Align gen6 push constant size to dispatch width.
- i965: Add support for the instruction compression bits on gen6.
- i965: Nuke brw_wm_glsl.c.
- i965: Remove INTEL_DEBUG=glsl_force now that there's no brw_wm_glsl.c
- i965: Fix comment about gen6_wm_constants.
- i965: Handle saturates on gen6 math instructions.
- i965: Always hand the absolute value to RSQ.
- i965: Add disabled debug code for dumping out the WM constant payload.
- i965: Work around gen6 ignoring source modifiers on math instructions.
- i965: Fix flipped value of the not-embedded-in-if on gen6.
- i965: Don't try to store gen6 (float) blend constant color in bytes.
- i965: Set up the color masking for the first drawbuffer on gen6.
- i965: Set up the per-render-target blend state on gen6.
- i965: Set the render target index in gen6 fixed-function/ARB_fp path.
- i965: Use the new pixel mask location for gen6 ARB_fp KIL instructions.
- i965: Drop KIL_NV from the ff/ARB_fp path since it was only used for GLSL.
- i965: Drop push-mode reladdr constant loading and always use constant_map.
- i965: Fix VS constants regression pre-gen6.
- i965: Clean up VS constant buffer location setup.
- i965: Set up the correct texture border color state struct for Ironlake.
- i965: Set render_cache_read_write surface state bit on gen6 constant surfs.
- i965: remove unused variable since brw_wm_glsl.c removal.
- intel: Use plain R8 and RG8 for COMPRESSED_RED and COMPRESSED_RG.
- intel: Set the swizzling for depth textures using the GL_RED depth mode.
- glsl: Correct the marking of InputsRead/OutputsWritten on in/out matrices.
- i965: Correct the dp_read message descriptor setup on g4x.
- intel: Include stdbool so we can stop using GLboolean when we want to.
- i965: Fix ARL to work on gen6.
- i956: Fix the old FP path fragment position setup on gen6.
- i965: Fix gl_FragCoord.z setup on gen6.
- i965: Add support for using the BLT ring on gen6.
- intel: Update renderbuffers before looking up CopyTexImage's read buffer.
- intel: Drop commented intel_flush from copy_teximage.
- intel: Try to sanely check that formats match for CopyTexImage.
- intel: Support glCopyTexImage() from XRGB8888 to ARGB8888.
- i965: Avoid using float type for raw moves, to work around SNB issue.
- i965: Set the alternative floating point mode on gen6 VS and WM.
- i965: Add support for gen6 constant-index constant loading.
- i965: Add support for gen6 reladdr VS constant loading.
- i965: Improve the hacks for ARB_fp scalar^scalar POW on gen6.
- i965: Factor out the ir comparision to BRW_CONDITIONAL_* code.
- i965: Fix regression in FS comparisons on original gen4 due to gen6 changes.
- i965: Do lowering of array indexing of a vector in the FS.
- intel: Only do frame throttling at glFlush time when using frontbuffer.
- intel: Handle forced swrast clears before other clear bits.
- intel: Use tri clears when we don't know how to blit clear the format.
- intel: Add spans code for the ARB_texture_rg support.
- intel: Add a couple of helper functions to reduce rb code duplication.
- intel: Fix segfaults from trying to use _ColorDrawBuffers in FBO validation.
- intel: When validating an FBO's combined depth/stencil, use the given FBO.
Fabian Bieler (2):
- r600g: set address of pop instructions to next instruction
- glsl: fix lowering conditional returns in subroutines
Francisco Jerez (51):
- dri/nv04: Fix PGRAPH_ERRORs when running OA.
- dri/nv04: Mipmapping fixes.
- dri/nv04: Align SIFM transfer dimensions.
- dri/nv04: Fix up color mask.
- dri/nv04: Fix maximum texture size.
- dri/nv04: Fix provoking vertex.
- dri/nouveau: Update nouveau_class.h.
- dri/nouveau: Add some more extensions.
- dri/nouveau: Fix glRenderbufferStorage with DEPTH_COMPONENT as internal format.
- dri/nouveau: Don't request a fake front unnecessarily.
- dri/nouveau: Don't reemit the BO state in nouveau_state_emit().
- dri/nouveau: Cleanup references to the old FBOs on glMakeCurrent().
- meta: Don't bind the created texture object in init_temp_texture().
- dri/nv10: Fix the CLAMP texture wrap mode.
- dri/nv04: Use nvgl_wrap_mode().
- dri/nouveau: Remove unnecessary assertion.
- dri/nouveau: Cleanup more references to old FBOs and VBOs.
- dri/nv10-nv20: Fix texturing in some cases after a base level change.
- dri/nouveau: Fix software mipmap generation on 1x1 textures.
- dri/nouveau: Have a smaller amount of larger scratch buffers.
- dri/nouveau: Remove unnecessary flush.
- dri/nv10: Use fast Z clears.
- dri/nouveau: Minor cleanup.
- dri/nv10: Fake fast Z clears for pre-nv17 cards.
- dri/nouveau: Initialize tile_flags when allocating a render target.
- nouveau: Get larger push buffers.
- dri/nouveau: Force a "slow" Z clear if we're getting a new depth buffer.
- dri/nv20: Clear with the 3D engine.
- dri/nouveau: Don't assert(0) on compressed internal formats.
- dri/nv25: Bind a hierarchical depth buffer.
- dri/nouveau: Call _mesa_update_state() after framebuffer invalidation.
- dri/nouveau: Honor the access flags in nouveau_bufferobj_map_range.
- dri/nouveau: Tell the vbo module we want real hardware BOs.
- dri/nouveau: Split out the scratch helpers to a separate file.
- dri/nouveau: Avoid recursion in nouveau_bo_context_reset().
- dri/nouveau: Use a macro to iterate over the bound vertex attributes.
- dri/nouveau: Split out array handling to its own file.
- dri/nouveau: Optimize VBO binding re-emission.
- dri/nouveau: Keep small DYNAMIC_DRAW vertex buffers in system ram.
- dri/nouveau: Pipeline glTexSubImage texture transfers.
- dri/nouveau: Fix type promotion issue on 32bit platforms.
- dri/nouveau: Validate the framebuffer state on read buffer changes.
- dri/nouveau: Re-emit the BO state when coming back from a software fallback.
- meta: Don't leak alpha function/reference value changes.
- meta: Fix incorrect rendering of the bitmap alpha component.
- vbo: Avoid unnecessary copy to/from current in vertex format upgrade.
- meta: Don't try to disable cube maps if the driver doesn't expose the extension.
- meta: Handle bitmaps with alpha test enabled.
- dri/nouveau: Split hardware/software TNL instantiation more cleanly.
- dri/nouveau: Fix typo.
- dri/nouveau: Kill a bunch of ternary operators.
Fredrik Höglund (2):
- r600g: Fix texture sampling with swizzled coords
- r600g: fix pow(0, 0) evaluating to NaN
Guillermo S. Romero (1):
- r300g: Do not use buf param before checking for NULL.
Henri Verbeet (19):
- r600g: Flush upload buffers before draws instead of before flushes.
- r600g: Check for other references before checking for existing mappings in radeon_bo_pb_map_internal().
- r600g: Remove a redundant flush in r600_texture_transfer_map().
- r600g: Buffer object maps imply a wait.
- r600g: Respect PB_USAGE_UNSYNCHRONIZED in radeon_bo_pb_map_internal().
- Revert "r600g: Flush upload buffers before draws instead of before flushes."
- r600g: fix exports_ps to export a number not a mask.
- r600g: Mention AMD in the renderer string.
- r600g: Cleanup the fenced_bo list in r600_context_fini().
- r600g: Evergreen has two extra frac_bits for the sampler LOD state.
- r600: Evergreen has two extra frac_bits for the sampler LOD state.
- r600g: Add PIPE_FORMAT_L8A8_UNORM for Evergreen as well.
- r600g: Swizzle vertex data only once.
- r600g: Synchronize supported color formats between Evergreen and r600/r700.
- r600g: Fix the PIPE_FORMAT_L8A8_UNORM color swaps.
- r600g: Fix the PIPE_FORMAT_A8_UNORM color swap for Evergreen as well.
- r600g: Cleanup block bo references in r600_context_fini().
- r600g: Cleanup fetch shader resources in r600_pipe_shader_destroy().
- st/mesa: Handle wrapped depth buffers in st_copy_texsubimage().
Hui Qi Tay (10):
- llvmpipe: minor changes in llvm coefficient calcs
- draw: cliptest and viewport done in a single loop in vertex shader
- draw: added viewport and cliptest flags
- draw: sanitize llvm variant key
- draw: corrections for w coordinate
- draw: corrections to allow for different cliptest cases
- llvmpipe: Moved draw pipeline twoside function to llvm setup code
- llvmpipe: added llvm offset setup code
- llvmpipe: clean up polygon offset function in lp setup code
- llvmpipe: fix such that offset/twoside function only does in-place modification
Ian Romanick (102):
- glsl2: Refactor testing for whether a deref is of a matrix or array
- glsl2: Add flags to enable variable index lowering
- glsl: Add doxygen comments
- EGL DRI2: Silence piles of 'unused variable' warnings
- EGL DRI2: Silence 'missing initializer' warnings
- egl_glx: Silence piles of 'unused variable' warnings
- egl: Fix several 'comparison between signed and unsigned integer' warnings
- dri: Ensure that DRI driver cpp files are in tarballs
- mesa: Force GL_ARB_copy_buffer to always be enabled
- mesa: Force GL_SGIS_generate_mipmap to always be enabled
- Remove GL_MESA_packed_depth_stencil
- Remove GL_EXT_cull_vertex
- Regenerate files changed by previous commit
- Remove unnescessary initializations of UpdateTexturePalette
- ARB_texture_rg: Add GLX protocol support
- ARB_texture_rg: Correct some errors in RED / RG internal format handling
- ARB_texture_rg: Add GL_TEXTURE_{RED,GREEN}_SIZE query support
- ARB_texture_rg: Add GL_RED as a valid GL_DEPTH_TEXTURE_MODE
- ARB_texture_rg: Handle RED and RG the same as RGB for tex env
- ARB_texture_rg: Add R8, R16, RG88, and RG1616 internal formats
- ARB_texture_rg: Allow RED and RG textures as FBO color buffer attachments
- mesa: Enable GL_ARB_texture_rg in software paths
- i965: Enable GL_ARB_texture_rg
- mesa: Add ARB_texture_compression_rgtc as an alias for EXT_texture_compression_rgtc
- ARB_texture_rg: Add GL_COMPRESSED_{RED,RG} cases in _mesa_is_color_format
- mesa: Fix misplaced #endif
- mesa: Trivial correction to comment
- rgtc: Detect RGTC formats as color formats and as compressed formats
- docs: Add list of bugs fixed in 7.9
- docs: Import 7.9 release notes from 7.9 branch.
- docs: Import 7.8.x release notes from 7.8 branch.
- docs: download.html does not need to be updated for each release
- docs: Update mailing lines from sf.net to freedesktop.org
- docs: Import news updates from 7.9 branch
- docs: added news item for 7.9 release
- glsl: Fail linking if assign_attribute_locations fails
- glsl: Refactor 'layout' grammar to match GLSL 1.60 spec grammar
- glsl: Slight refactor of error / warning checking for ARB_fcc layout
- glsl: Clear type_qualifier using memset
- glsl: Wrap ast_type_qualifier contents in a struct in a union
- glsl: Regenerate files modified by previous commits
- glcpp: Add the define for ARB_explicit_attrib_location when present
- glcpp: Regenerate files changes by previous commit
- glsl: Add parser support for GL_ARB_explicit_attrib_location layouts
- glsl: Regenerate files changes by previous commit
- glsl: Track explicit location in AST to IR translation
- glsl: Add linker support for explicit attribute locations
- main: Enable GL_ARB_explicit_attrib_location for swrast
- intel: Enable GL_ARB_explicit_attrib_location
- glsl: Remove const decoration from inlined function parameters
- docs: skeleton for 7.10 release notes
- docs: Update status of GL 3.x related extensions
- mesa: Validate assembly shaders when GLSL shaders are used
- glsl: Fix incorrect assertion
- linker: Reject shaders that have unresolved function calls
- mesa: Silence unused variable warning
- mesa: Refactor validation of shader targets
- mesa: Clean up two 'comparison between signed and unsigned' warnings
- mesa: Clean up various 'unused parameter' warnings in shaderapi
- glsl: Slightly change the semantic of _LinkedShaders
- linker: Trivial indention fix
- i965: Fix indentation after commit 3322fbaf
- linker: Improve handling of unread/unwritten shader inputs/outputs
- glapi: Add GL_EXT_separate_shader_objects
- glapi: Commit files changed by previous commit
- mesa: Add infrastructure to track GL_EXT_separate_shader_objects
- mesa: Skeletal support for GL_EXT_separate_shader_objects
- mesa: Add display list support for GL_EXT_separate_shader_objects functions
- mesa: Track an ActiveProgram distinct from CurrentProgram
- Track separate programs for each stage
- swrast: Enable GL_EXT_separate_shader_objects in software paths
- intel: Enable GL_EXT_separate_shader_objects in Intel drivers
- docs: add GL_EXT_separate_shader_objects to release notes
- glsl: Fix incorrect gl_type of sampler2DArray and sampler1DArrayShadow
- ir_to_mesa: Refactor code for emitting DP instructions
- mesa: Allow query of MAX_SAMPLES with EXT_framebuffer_multisample
- glsl: Refactor is_vec_{zero,one} to be methods of ir_constant
- glsl: Simplify generation of swizzle for vector constructors
- glsl: Make is_zero and is_one virtual methods of ir_rvalue
- ir_to_mesa: Generate smarter code for some conditional moves
- glsl: Add ir_unop_sin_reduced and ir_unop_cos_reduced
- glsl: Eliminate assumptions about size of ir_expression::operands
- glsl: Add ir_rvalue::is_negative_one predicate
- glsl: Add unary ir_expression constructor
- glsl: Add ir_quadop_vector expression
- glsl: Fix matrix constructors with vector parameters
- i915: Disallow alpha, red, RG, and sRGB as render targets
- glsl: Use M_LOG2E constant instead of calling log2
- glsl: Lower ir_binop_pow to a sequence of EXP2 and LOG2
- i915: Request that POW instructions be lowered
- i915: Correctly generate unconditional KIL instructions
- glsl: Ensure that equality comparisons don't return a NULL IR tree
- i965: Correctly emit constants for aggregate types (array, matrix, struct)
- glsl: Inherrit type of declared variable from initializer
- linker: Ensure that unsized arrays have a size after linking
- linker: Fix regressions caused by previous commit
- glsl: Inherrit type of declared variable from initializer after processing assignment
- linker: Allow built-in arrays to have different sizes between shader stages
- ir_to_mesa: Don't generate swizzles for record derefs of non-scalar/vectors
- Refresh autogenerated file builtin_function.cpp.
- glsl: Allow less restrictive uses of sampler array indexing in GLSL <= 1.20
- docs: Import 7.9.1 release notes from 7.9 branch
Jakob Bornecrantz (27):
- rbug: Cast opcode to corrent int size
- rbug: Add function to get opcode name string
- scons: Link against talloc in the Gallium DRI drivers
- i915g: Link with wrapper sw winsys with scons
- tgsi: Actually care what check_soa_dependencies says
- tgsi: Fix missing test before check
- llvmpipe: Move makefile include to before targets
- wrapper: Fix spelling
- wrapper: Add a way to dewrap a pipe screen without destroying it
- egl: Remove unnecessary headers
- target-helpers: Remove per target software wrapper check
- graw: Tidy graw xlib scons file a bit
- scons: Remove old pipebuffer SConscript
- scons: Detabify
- scons: Check for pkg-config before trying to use it
- scons: Check for libdrm_[intel|radeon] as well
- scons: Move dependancy checks to the main gallium scons file
- scons: Unify state tracker SConscripts
- galahad: Correct the name of the scons library
- graw: Use inline sw helper instead of roll your own loader
- libgl-xlib: Use sw helper instead of roll your own
- libgl-xlib: Use inline debug helper instead of non-inline version
- graw: Use inline debug helper instead of non-inline version
- gallium: Remove redundant sw and debug target helpers
- i915g: Improve debug printing for textures
- i915g: Make sure that new vbo gets updated
- st/mesa: Unbind all constant buffers
Jerome Glisse (75):
- r600g: alternative command stream building from context
- r600g: move chip class to radeon common structure
- r600g: use pipe context for flushing inside map
- r600g: add back reference check when mapping buffer
- r600g: directly allocate bo for user buffer
- r600g: fix multi buffer rendering
- r600g: occlusion query for new design
- r600g: flush color buffer after draw command
- r600g: disable shader rebuild optimization & account cb flush packet
- r600g: fix multiple occlusion query on same id
- r600g: initial evergreen support in new path
- r600g: fix typo in evergreen define (resource are in x range)
- r600g: move use_mem_constants flags for new designs structure alignment
- r600g: evergreen fix for new design
- r600g: fix compilation after change to evergreend.h
- r600g: fixup some evergreen register definitions
- r600g: fix evergreen new path
- r600g: fix reg definition
- r600g: fix evergreen new path
- r600g: bring over fix from old path to new path
- r600g: fix vertex resource & polygon offset
- r600g: disable early cull optimization when occlusion query running
- r600g: move around variables to share depth uncompression code
- r600g: use depth decompression in new path
- r600g: fix index buffer drawing
- r600g: build packet header once
- r600g: fix pointsprite & resource unbinding
- r600g: fix routing btw vertex & pixel shader
- r600g: fix occlusion query after change to block structure
- r600g: use ptr for blit depth uncompress function
- r600g: fix remaining piglit issue in new design
- r600g: switch to new design
- r600g: suspend/resume occlusion query around clear/copy
- r600g: avoid rebuilding the vertex shader if no change to input format
- r600g: use a hash table instead of group
- r600g: delete old path
- r600g: cleanup
- r600g: more cleanup
- r600g: use constant buffer instead of register for constant
- r600g: fix constant & literal src splitting, also fix mplayer gl2 shader
- evergreeng: avoid overlapping border color btw VS & PS
- r600g: indentation fixes
- r600g: rename radeon_ws_bo to r600_bo
- r600g: allow r600_bo to be a sub allocation of a big bo
- r600g: use r600_bo for relocation argument, simplify code
- r600g: rename radeon_ws_bo to r600_bo
- r600g: remove dead label & fix indentation
- r600g: store reloc information in bo structure
- r600g: improve bo flushing
- r600g: simplify block relocation
- r600g: userspace fence to avoid kernel call for testing bo busy status
- r600g: avoid segfault due to unintialized list pointer
- r600g: fix dirty state handling
- r600g: allow driver to work without submitting cmd to GPU
- gallium/noop: no operation gallium driver
- r600g: code cleanup (indent, trailing space, empty line ...)
- r600g: fix occlusion query on evergreen (avoid lockup)
- r600g: add fetch shader capabilities
- r600g: dump raw shader output for debugging
- r600g: update polygon offset only when rasterizer or zbuffer change
- r600g: indentation fix
- r600g: more indentation fix + warning silencing + dead code removal
- r600g: build fetch shader from vertex elements
- r600g: avoid useless shader rebuild at draw call
- r600g: remove useless flush map
- r600g: remove dead code
- r600g: fix userspace fence against lastest kernel
- r600g: avoid using pb* helper we are loosing previous cpu cycle with it
- r600g: specialized upload manager
- r600g: indentation cleanup
- r600g: fix bo size when creating bo from handle
- r600g: fix segfault when translating vertex buffer
- r600g: need to reference upload buffer as the might still live accross flush
- r600g: properly unset vertex buffer
- r600g: avoid segfault
Joakim Sindholt (3):
- util/u_blitter: fix leak
- radeong: fix leaks
- r300g: silence guard band cap errors
Johann Rudloff (3):
- radeon: Implement EGL_MESA_no_surface_extension
- radeon: Implement __DRI_IMAGE and EGL_MESA_image_drm
- radeon: Implement GL_OES_EGL_image
John Doe (3):
- r600g: misc cleanup
- r600g: don't double count dirty block
- r600g: keep a mapping around for each bo
Jon TURNEY (1):
- Ensure -L$(TOP)/$(LIB_DIR) appears in link line before any -L in $LDFLAGS
José Fonseca (128):
- gallivm: Fix address register swizzle.
- gallivm: Start collecting bitwise arithmetic helpers in a new module.
- gallivm: Clamp indirect register indices to file_max.
- util: linearized sRGB values don't fit into 8bits
- llvmpipe: Default to no threading on single processor systems.
- tgsi: Don't ignore indirect registers in tgsi_check_soa_dependencies
- llvmpipe: Describe how to profile llvmpipe.
- llvmpipe: When failing free fs shader too.
- util: Flush stdout on util_format.
- gallivm: Add unorm support to lp_build_lerp()
- llvmpipe: Special case complementary and identify blend factors in SoA.
- llvmpipe: Make rgb/alpha bland func/factors match, when there is no alpha.
- draw: Prevent clipped vertices overflow.
- draw: Fullfil the new min_lod/max_lod/lod_bias/border_color dynamic state
- gallivm: Fetch the lod from the dynamic state when min_lod == max_lod.
- gallivm: Remove dead experimental code.
- llvmpipe: Decouple sampler view and sampler state updates.
- scons: New build= option, with support for checked builds.
- scons: New build= option, with support for checked builds.
- trace: Fix set_index_buffer and draw_vbo tracing.
- python/retrace: Handle set_index_buffer and draw_vbo.
- gallivm: Use SSE4.1's ROUNDSS/ROUNDSD for scalar rounding.
- gallivm: More comprehensive border usage logic.
- retrace: Handle clear_render_target and clear_depth_stencil.
- llvmpipe: Dump a few missing shader key flags.
- llvmpipe: Fix perspective interpolation for point sprites.
- llvmpipe: Fix sprite coord perspective interpolation of Q.
- gallivm: Take the type signedness in consideration in round/ceil/floor.
- gallivm: Use a faster (and less accurate) log2 in lod computation.
- gallivm: Fast implementation of iround(log2(x))
- gallivm: Combined ifloor & fract helper.
- gallivm: Only apply min/max_lod when necessary.
- gallivm: Compute lod as integer whenever possible.
- util: Cleanup util_pack_z_stencil and friends.
- llvmpipe: Cleanup depth-stencil clears.
- gallivm: Vectorize the rho computation.
- gallivm: Do not do mipfiltering when magnifying.
- gallivm: Simplify lp_build_mipmap_level_sizes' interface.
- gallivm: Don't compute the second mipmap level when frac(lod) == 0
- gallivm: Use lp_build_ifloor_fract for lod computation.
- gallivm: Clamp mipmap level and zero mip weight simultaneously.
- gallivm: Fix copy'n'paste typo in previous commit.
- gallivm: Implement brilinear filtering.
- gallivm: Use the wrappers for SSE pack intrinsics.
- gallivm: Avoid control flow for two-sided stencil test.
- gallivm: Warn when doing inefficient integer comparisons.
- gallivm: Move into the as much of the second level code as possible.
- llvmpipe: First minify the texture size, then broadcast.
- gallivm: Help for combined extraction and broadcasting.
- gallivm: Do size computations simultanously for all dimensions (AoS).
- llvmpipe: Prevent z > 1.0
- llvmpipe: Fix MSVC build. Enable the new SSE2 code on non SSE3 systems.
- gallivm: Handle code have ret correctly.
- util: Defined M_SQRT2 when not available.
- gallivm: Less code duplication in log computation.
- gallivm: Special bri-linear computation path for unmodified rho.
- gallivm: Don't generate Phis for execution mask.
- gallivm: Use varilables instead of Phis for cubemap selection.
- gallivm: Remove support for Phi generation.
- gallivm: Factor out the SI->FP texture size conversion for SoA path too
- gallivm: Simplify if/then/else implementation.
- gallivm: Cleanup the rest of the flow module.
- gallivm: Fix a long standing bug with nested if-then-else emission.
- gallivm: Allow to disable bri-linear filtering with GALLIVM_DEBUG=no_brilinear runtime option
- gallivm: Use variables instead of Phis in loops.
- gallivm: Pass texture coords derivates as scalars.
- llvmpipe: Remove outdated comment about stencil testing.
- gallivm: Eliminate unsigned integer arithmetic from texture coordinates.
- gallium: Define C99 restrict keyword where absent.
- tgsi: Export some names for some tgsi enums.
- gallivm: More detailed analysis of tgsi shaders.
- llvmpipe: Use lp_tgsi_info.
- llvmpipe: Do not dispose the execution engine.
- llvmpipe: Fix MSVC build.
- llmvpipe: improve mm_mullo_epi32
- gallivm: Name anonymous union.
- llvmpipe: Unbreak Z32_FLOAT.
- gallivm: More accurate float -> 24bit & 32bit unorm conversion.
- llvmpipe: Generalize the x8z24 fast path to all depth formats.
- llvmpipe: Fix depth-stencil regression.
- llvmpipe: Ensure z_shift and z_width is initialized.
- gallivm: Fix SoA cubemap derivative computation.
- llvmpipe: Fix bad refactoring.
- llvmpipe: Initialize bld ctx via lp_build_context_init instead of ad-hoc and broken code.
- gallivm: Comment lp_build_insert_new_block().
- gallivm: Add a note about SSE4.1's nearest mode rounding.
- llvmpipe: Don't test rounding of x.5 numbers.
- gallium: Avoid using __doc__ in python scripts.
- gallivm: always enable LLVMAddInstructionCombiningPass()
- gallivm: Remove the EMMS opcodes.
- mesa: Fix windows build (uint -> GLuint).
- scons: Revamp how to specify targets to build.
- scons: Fix MinGW cross-compilation.
- scons: Some pipe drivers are not portable for MSVC
- scons: Restore x11 tool behavior for backwards compatability.
- scons: Disable python state tracker when swig is not present.
- r600g: List recently added files in SConscript.
- scons: Add aliases for several pipe drivers.
- scons: i915 can't build on MSVC either.
- scons: Propagate installation targets.
- xorg/vmwgfx: Add missing source file to SConscript.
- st/xorg: Add missing n to error message.
- st/xorg: Detect libkms with scons too.
- xorg/vmwgfx: Link libkms when available.
- r600g: Swap the util_blitter_destroy call order.
- gallivm: Allocate TEMP/OUT arrays only once.
- libgl-gdi: Allow to pick softpipe/llvmpipe on runtime.
- scons: Use inline wrap helpers more consistently.
- svga: Use consistent hexadecimal representation on debug output.
- scons: Alias for svga
- wgl: Stub WGL_ARB_pbuffer support.
- wgl: More complete WGL_ARB_pbuffer support.
- svga: Silence debug printf.
- scons: Move MSVS_VERSION option to common module.
- vega: Remove extraneous ;
- retrace: Some fixes.
- util: C++ safe.
- wgl: Fix double free. Remove dead code.
- util: Plug leaks in util_destroy_gen_mipmap.
- util: __builtin_frame_address() doesn't work on mingw.
- util: Don't try to use imagehlp on mingw.
- wgl: Unreference the current framebuffer after the make_current call.
- WIN32_THREADS -> WIN32
- mapi: Hack to avoid vgCreateFont being generated as vgCreateFontA.
- wgl: Fix visual's buffer_mask configuration.
- mesa: Temporary hack to prevent stack overflow on windows
- mesa: Bump the number of bits in the register index.
- llvmpipe: Plug fence leaks.
Julien Cristau (1):
- Makefile: don't include the same files twice in the tarball
Keith Whitwell (89):
- llvmpipe: brackets around macro arg
- llvmpipe: remove duplicate code
- llvmpipe: return zero from floor_pot(zero)
- gallivm: make lp_build_sample_nop public
- llvmpipe: add LP_PERF flag to disable various aspects of rasterization
- llvmpipe: add DEBUG_FS to dump variant information
- llvmpipe: use llvm for attribute interpolant calculation
- graw: add frag-face shader
- llvmpipe: fix flatshading in new line code
- draw: don't apply flatshading to clipped tris with <3 verts
- llvmpipe: handle FACING interpolants in line and point setup
- llvmpipe: handle up to 8 planes in triangle binner
- llvmpipe: make debug_fs_variant respect variant->nr_samplers
- gallivm: don't apply zero lod_bias
- llvmpipe: fail gracefully on oom in scene creation
- llvmpipe: avoid overflow in triangle culling
- gallivm: special case conversion 4x4f to 1x16ub
- gallivm: round rather than truncate in new 4x4f->1x16ub conversion path
- llvmpipe: clean up setup_tri a little
- llvmpipe: add rast_tri_4_16 for small lines and points
- llvmpipe: fix off-by-one in tri_16
- llvmpipe: defer attribute interpolation until after mask and ztest
- llvmpipe: use alloca for fs color outputs
- llvmpipe: store zero into all alloca'd values
- llvmpipe: dump fragment shader ir and asm when LP_DEBUG=fs
- gallivm: specialized x8z24 depthtest path
- gallivm: prefer blendvb for integer arguments
- gallivm: simpler uint8->float conversions
- llvmpipe: try to be sensible about whether to branch after mask updates
- llvmpipe: clean up shader pre/postamble, try to catch more early-z
- llvmpipe: simplified SSE2 swz/unswz routines
- llvmpipe: try to do more of rast_tri_3_16 with intrinsics
- llvmpipe: add debug helpers for epi32 etc
- llvmpipe: try to keep plane c values small
- llvmpipe: fix typo in last commit
- gallium: move sse intrinsics debug helpers to u_sse.h
- r600g: add missing file to sconscript
- gallivm: don't branch on KILLs near end of shader
- Revert "llvmpipe: try to keep plane c values small"
- llvmpipe: make sure intrinsics code is guarded with PIPE_ARCH_SSE
- llvmpipe: don't try to emit non-existent color outputs
- r600/drm: fix segfaults in winsys create failure path
- r600g: emit hardware linewidth
- r600g: handle absolute modifier in shader translator
- llvmpipe: reintroduce SET_STATE binner command
- llvmpipe: don't pass frontfacing as a float
- llvmpipe: slightly shrink the size of a binned triangle
- llvmpipe: don't store plane.ei value in binned data
- gallium: move some intrinsics helpers to u_sse.h
- llvmpipe: do plane calculations with intrinsics
- llvmpipe: use aligned loads/stores for plane values
- llvmpipe: fix non-sse build after recent changes
- llvmpipe: check shader outputs are non-null before using
- llvmpipe: validate color outputs against key->nr_cbufs
- llvmpipe: clean up fields in draw_llvm_variant_key
- llvmpipe: remove setup fallback path
- llvmpipe: fail cleanly on malloc failure in lp_setup_alloc_triangle
- Merge remote branch 'origin/master' into lp-setup-llvm
- llvmpipe: remove unused file
- llvmpipe: remove unused arg from jit_setup_tri function
- Merge branch 'llvm-cliptest-viewport'
- draw: make sure viewport gets updated in draw llvm shader
- llvmpipe: turn off draw offset/twoside when we can handle it
- llvmpipe: avoid generating tri_16 for tris which extend past tile bounds
- llvmpipe: guard against NULL task->query pointer
- st/mesa: unbind constant buffer when not in use
- r600g: propagate usage flags in texture transfers
- r600g: propogate resource usage flags to winsys, use to choose bo domains
- r600g: use a buffer in GTT as intermediate on texture up and downloads
- r600g: remove unused flink, domain fields from r600_resource
- r600g: set hardware pixel centers according to gl_rasterization_rules
- evergreeng: protect against null constant buffers
- r600g: don't call debug_get_bool_option for tiling more than once
- evergreeng: respect linewidth state, use integer widths only
- evergreeng: set hardware pixelcenters according to gl_rasterization_rules
- r600g: avoid recursion with staged uploads
- r600g: attempt to turn on DXTn formats
- r600g: translate ARR instruction
- r600: fix my pessimism about PIPE_TRANSFER_x flags
- ws/r600: match bo_busy shared/fence logic in bo_wait
- r600g: guard experimental s3tc code with R600_ENABLE_S3TC
- r600g: do not try to use staging resource for depth textures
- r600g: enforce minimum stride on render target texture images
- llvmpipe: fix up twoside after recent changes
- llvmpipe: twoside for specular color also
- Merge branch 'lp-offset-twoside'
- llvmpipe: raise dirty flag on transfers to bound constbuf
- llvmpipe: remove misleading debug string
- llvmpipe: shortcircuit some calls to set_scene_state
Kenneth Graunke (94):
- glsl: Change from has_builtin_signature to has_user_signature.
- glsl: Don't print blank (function ...) headers for built-ins.
- glsl: Properly handle nested structure types.
- glsl/builtins: Fix equal and notEqual builtins.
- glsl/builtins: Switch comparison functions to just return an expression.
- glsl: Add comments to clarify the types of comparison binops.
- glsl: Fix broken handling of ir_binop_equal and ir_binop_nequal.
- glsl: "Copyright", not "Constantright"
- i965: Fix incorrect batchbuffer size in gen6 clip state command.
- i965: Use logical-not when emitting ir_unop_ceil.
- glsl: Add front-end support for the "trunc" built-in.
- glsl: Refresh autogenerated file builtin_function.cpp.
- i965: Use RNDZ for ir_unop_trunc in the new FS.
- i965: Correctly emit the RNDZ instruction.
- i965: Clean up a warning in the old fragment backend.
- glsl: Add a new ir_unop_round_even opcode for GLSL 1.30's roundEven.
- glsl: Add front-end support for GLSL 1.30's roundEven built-in.
- i965: Add support for ir_unop_round_even via the RNDE instruction.
- glsl: Add support for the 1.30 round() built-in.
- glsl: Refresh autogenerated file builtin_function.cpp.
- glsl: Don't return NULL IR for erroneous bit-shift operators.
- i965: Add missing "break" statement.
- glsl: Fix copy and paste error in ast_bit_and node creation.
- glsl: Regenerate parser files.
- i965: Remove unused variable.
- glsl: Remove useless ir_shader enumeration value.
- mesa: Remove FEATURE_ARB_shading_language_120 macro.
- glcpp: Return NEWLINE token for newlines inside multi-line comments.
- glcpp: Refresh autogenerated lexer file.
- glsl: Add support for GLSL 1.30's modf built-in.
- glsl: Refresh autogenerated file builtin_function.cpp.
- generate_builtins.py: Output large strings as arrays of characters.
- Refresh autogenerated file builtin_function.cpp.
- glsl: Fix constant component count in vector constructor emitting.
- Fix build on systems where "python" is python 3.
- i965: Add bit operation support to the fragment shader backend.
- glsl: Remove unused ARRAY_SIZE macro.
- glsl/builtins: Rename 'x' to 'y_over_x' in atan(float) implementation.
- glsl/builtins: Clean up some ugly autogenerated code in atan.
- Refresh autogenerated file builtin_function.cpp.
- glsl: Don't print a useless space at the end of an S-Expression list.
- ir_reader: Return a specific ir_dereference variant.
- ir_reader: Remove useless error check.
- ir_reader: Fix some potential NULL pointer dereferences.
- ir_dead_functions: Actually free dead functions and signatures.
- glsl: Remove unnecessary "unused variable" warning suppression.
- glsl: Remove GLSL_TYPE_FUNCTION define.
- glsl: Convert glsl_type::base_type from #define'd constants to an enum.
- glsl: Rework reserved word/keyword handling in the lexer.
- glsl: Add new keywords and reserved words for GLSL 1.30.
- glsl: Add support for the 'u' and 'U' unsigned integer suffixes.
- glsl: Refresh autogenerated lexer and parser files.
- generate_builtins.py: Fix inconsistent use of tabs and spaces warning.
- glsl: Implement the asinh, acosh, and atanh built-in functions.
- glsl: Refresh autogenerated file builtin_function.cpp.
- glsl: Add constant expression handling for asinh, acosh, and atanh.
- glsl: Remove unused and out of date Makefile.am.
- glsl: Rename various ir_* files to lower_* and opt_*.
- glcpp: Define GL_FRAGMENT_PRECISION_HIGH if GLSL version >= 1.30.
- Refresh autogenerated glcpp parser.
- glsl: Fix constant expression handling for <, >, <=, >= on vectors.
- glsl: Unconditionally define GL_FRAGMENT_PRECISION_HIGH in ES2 shaders.
- Regenerate glcpp parser.
- glsl: Reimplement the "cross" built-in without ir_binop_cross.
- Refresh autogenerated file builtin_function.cpp.
- glsl: Remove the ir_binop_cross opcode.
- glsl: Refactor get_num_operands.
- glsl: Simplify a type check by using type->is_integer().
- glsl: Combine many instruction lowering passes into one.
- mesa: Fix glGet of ES2's GL_MAX_*_VECTORS properties.
- glsl: Don't inline function prototypes.
- glsl: Use do_common_optimization in the standalone compiler.
- glsl: Add a virtual as_discard() method.
- glsl: Refactor out cloning of function prototypes.
- glsl: Lazily import built-in function prototypes.
- glsl: Remove anti-built-in hacks from the print visitor.
- glsl/linker: Free any IR discarded by optimization passes.
- glsl: Add an optimization pass to simplify discards.
- glsl: Add a lowering pass to move discards out of if-statements.
- glsl: Remove "discard" support from lower_jumps.
- glsl: Add comments to lower_jumps (from the commit message).
- ir_print_visitor: Print out constant structure values.
- glsl: Factor out code which emits a new function into the IR stream.
- symbol_table: Add support for adding a symbol at top-level/global scope.
- glsl: Properly add functions during lazy built-in prototype importing.
- glcpp: Don't emit SPACE tokens in conditional_tokens production.
- Refresh autogenerated glcpp parser.
- glsl: Clean up code by adding a new is_break() function.
- glsl: Consider the "else" branch when looking for loop breaks.
- Remove OES_compressed_paletted_texture from the ES2 extension list.
- glsl/builtins: Compute the correct value for smoothstep(vec, vec, vec).
- glsl: Support if-flattening beyond a given maximum nesting depth.
- i965: Flatten if-statements beyond depth 16 on pre-gen6.
- i965: Internally enable GL_NV_blend_square on ES2.
Kristian Høgsberg (16):
- glx: Hold on to drawables if we're just switching to another context
- intel: Fix GL_ARB_shading_language_120 commit
- dri2: Make createImageFromName() take a __DRIscreen instead of __DRIcontext
- glx: Invalidate buffers after binding a drawable
- dri: Pass the __DRIscreen and the __DRIscreen private back to image lookup
- glx: Only remove drawables from the hash when we actually delete them
- gles2: Add GL_EXT_texture_format_BGRA8888 support
- Get rid of GL/internal/glcore.h
- gl: Remove unused GLcontextModes fields
- Rename GLvisual and __GLcontextModes to struct gl_config
- Drop GLframebuffer typedef and just use struct gl_framebuffer
- Drop GLcontext typedef and use struct gl_context instead
- Drop the "neutral" tnl module
- Only install vtxfmt tables for OpenGL
- i965: Don't write mrf assignment for pointsize output
- docs: Fix MESA_drm_image typo
Krzysztof Smiechowicz (1):
- nvfx: Pair os_malloc_aligned() with os_free_aligned().
Luca Barbieri (84):
- auxiliary: fix unintended fallthrough
- glsl: add pass to lower variable array indexing to conditional assignments
- auxiliary: fix depth-only and stencil-only clears
- gallium: avoid the C++ keyword "template" in sw_winsys.h
- softpipe: make z/s test always pass if no zsbuf, instead of crashing
- tgsi: add switch/case opcodes to tgsi_opcode_tmp.h
- softpipe: fix whitespace
- d3d1x: add new Direct3D 10/11 COM state tracker for Gallium
- d3d1x: add blob and signature extraction APIs
- d3d1x: fix compilation with recent Wine versions installed
- d3d1x: add missing file
- d3d1x: actually enable and fix blob apis
- d3d1x: fix build with compilers other than GCC 4.5
- d3d1x: add template parameters to base class ctor calls for GCC 4.4
- d3d1x: fix GCC 4.1/4.2 build
- d3d1x: ignore errors while building docs
- d3d1x: attempt to fix/workaround bug #30322
- nvfx: remove gl_PointCoord hack
- glx: decouple dri2.c and GLX, fixing Gallium EGL and d3d1x build
- winsys: automatically build sw winsys needed by EGL and d3d1x
- d3d1x: don't build progs automatically
- d3d1x: add missing memory barrier
- d3d1x: link with CXXFLAGS
- d3d1x: fix cf analysis
- d3d1x: fix warning
- d3d1x: fix segfault when hashing
- d3d1x: destroy native_display on adapter destruction
- d3d1x: fix GUID declarations
- d3d1x: redesign the HWND resolver interface
- d3d1x: fix API name
- d3d1x: define GUIDs in the normal way
- d3d1x: add Wine dlls (tri, tex working, but no other testing)
- d3d1x: properly reference count the backend
- d3d1x: fix deadlocks on non-recursive mutex
- d3d1x: bind NULL CSOs before destroying default CSOs on context dtor
- d3d1x: initialize the mutex
- d3d1x: autogenerate shader enums and text from def files
- d3d1x: s/tpf/sm4/g
- d3d1x: normalize whitespace
- d3d1x: remove specstrings
- d3d1x: minifix
- d3d1x: rename context params
- d3d11: rename screen params
- d3d1x: rename params in misc and objects
- d3d1x: rename parameters in dxgi
- d3d11: obliterate IDL parameter names
- d3d1x: remove specstrings.h include
- d3d1x: flush the pipe context when presenting
- d3d1x: remove another include specstrings.h
- d3d1x: flush properly
- d3d1x: add missing guid.cpp
- d3d1x: fix build without system EGL/egl.h
- d3d1x: add autogenerated files as prerequisites, so make builds them
- d3d1x: obliterate IDL parameter names from d3d10.idl from Wine too
- d3d1x: add shader dumping
- d3d1x: add untested support for geometry shader translation
- d3d1x: don't assert on unsupported resource types
- d3d1x: fix CheckMultisampleQualityLevels
- d3d1x: draw to the correct buffer
- d3d1x: fix linking of dxbc2tgsi
- nvfx: allow setting NULL constant buffers
- nvfx: add RGB framebuffer format support in addition to BGR
- d3d1x: don't crash on drivers not supporting vertex or geometry sampling
- d3d1x: assert if X visual is not among enumerated visuals
- d3d1x: stop using GLX in demos, just use the default visual
- d3d1x: CRLF -> LF in progs
- mesa: make makedepend an hard requirement
- gallium: add $(PROGS_DEPS) as dependencies for $(PROGS)
- d3d1x: fix parallel build
- d3d1x: add private gitignore file
- d3d1x: fix progs linking if not all EGL platforms are enabled
- d3d1x: link progs with CXXFLAGS
- d3d11: advertise IDXGIDevice1, not just IDXGIDevice
- d3d11: ignore StructureByteStride
- d3d1x: link to libdrm for X11 platform too
- ureg: support centroid interpolation
- d3d1x: support centroid interpolation
- d3d1x: properly support specifying MipLevels as 0
- d3d1x: put proper calling convention in headers, fixes 64-bit builds
- d3d1x: rework DXGI for occlusion testing and default width/height
- d3d1x: fix Map
- d3d11: fix reference counting so devices get freed
- d3d1x: work around crash in widl
- glsl: Unroll loops with conditional breaks anywhere (not just the end)
Lucas Stach (1):
- nvfx: fill PIPE_CAP_PRIMITIVE_RESTART and PIPE_CAP_SHADER_STENCIL_EXPORT
Marek Olšák (100):
- r300g: prevent creating multiple winsys BOs for the same handle
- r300g/swtcl: fix CS overrun
- st/mesa: fix assertion failure in GetTexImage for cubemaps
- util: make calling remove_from_list multiple times in a row safe
- r300g: fixup long-lived BO maps being incorrectly unmapped when flushing
- r300g: make accessing map_list and buffer_handles thread-safe
- r300g: fix a copy-paste typo for logging
- r300g: fix the border color for every format other than PIPE_FORMAT_B8G8R8A8
- Build r300g by default
- util: fix util_pack_color for B4G4R4A4
- r300g: fix macrotiling on R350
- r300g: code cleanups
- r300/compiler: fix projective mapping of 2D NPOT textures
- r300/compiler: do not use copy propagation if SaturateMode is used
- r300/compiler: fix shadow sampling with swizzled coords
- r300g: add support for 3D NPOT textures without mipmapping
- r300g: fix swizzling of texture border color
- configure.ac: look for libdrm_radeon before building gallium/r300,r600
- configure.ac: do not build xorg-r300g by default
- Makefile: ensure Gallium's Makefile.xorg and SConscript.dri are in the tarball
- r300g: add support for formats beginning with X, like X8R8G8B8
- r300g: fix conditional rendering in non-wait path
- r300g: add support for R8G8 colorbuffers
- r300g: add support for L8A8 colorbuffers
- update release notes for Gallium
- r300g: fix microtiling for 16-bits-per-channel formats
- r300g: do not print get_param errors in non-debug build
- r300g: say no to PIPE_CAP_STREAM_OUTPUT and PIPE_CAP_PRIMITIVE_RESTART
- mesa: allow FBO attachments of formats LUMINANCE, LUMINANCE_ALPHA, and INTENSITY
- r300g: fix texture border for 16-bits-per-channel formats
- st/mesa: support RGBA16 and use it for RGBA12 as well
- r300g: add a default channel ordering of texture border for unhandled formats
- r300g: mention ATI in the renderer string
- r300g: rename has_hyperz -> can_hyperz
- r300g: turn magic numbers into names in the hyperz code
- gallium: add CAPs for indirect addressing and lower it in st/mesa when needed
- tgsi: fill out CAPs for indirect addressing
- i915g: fill out CAPs for indirect addressing
- i965g: fill out CAPs for indirect addressing
- nv50: fill out CAPs for indirect addressing
- nvfx: fill out CAPs for indirect addressing
- r300g: fill out CAPs for indirect addressing
- r600g: fill out CAPs for indirect addressing
- svga: fill out CAPs for indirect addressing
- r300g: fix texture border color for all texture formats
- r300g: clean up redundancy in draw functions
- r300g: return shader caps from Draw for SWTCL vertex shaders
- r300g: remove the hack with OPCODE_RET
- r300g: print FS inputs uninitialized due to hardware limits to stderr
- r300g: fix rendering with no vertex elements
- st/mesa: enable ARB_explicit_attrib_location and EXT_separate_shader_objects
- docs: add GL 4.1 status
- gallium: add PIPE_SHADER_CAP_SUBROUTINES
- st/mesa: set MaxUniformComponents
- u_blitter: use PIPE_TRANSFER_DISCARD to prevent cpu/gpu stall
- r300/compiler: fix rc_rewrite_depth_out for it to work with any instruction
- r300/compiler: remove duplicate function rc_mask_to_swz
- r300/compiler: add a function for swizzling a mask
- r300/compiler: move util functions to radeon_compiler_util
- u_blitter: interpolate clear color using a GENERIC varying instead of COLOR
- st/mesa: fix texture border color for RED and RG base formats
- util: rename u_mempool -> u_slab
- r300g: fix texture border color once again
- r300/compiler: implement and lower OPCODE_CLAMP
- ir_to_mesa: Add support for conditional discards.
- r300g: fix texture swizzling with compressed textures on r400-r500
- r300g: disable ARB_texture_swizzle if S3TC is enabled on r3xx-only
- r300g: fix up cubemap texture offset computation
- r300/compiler: disable the swizzle lowering pass in vertex shaders
- r300g: fix build
- r300g: use internal BO handle for add_buffer and write_reloc
- r300g: implement simple transfer_inline_write for buffers
- mesa, st/mesa: fix gl_FragCoord with FBOs in Gallium
- r300g: fix pointer arithmetic with void* in transfer_inline_write
- r300g: do not remove unused constants if we are not near the limit
- r300g: add capability bit index_bias_supported
- r300g: one more r500_index_bias_supported leftover
- r300g: do not use the index parameter in set_constant_buffer
- r300g: cleanup winsys
- r300g: optimize looping over atoms
- st/mesa: initialize key in st_vp_varient
- u_blitter: use util_is_format_compatible in the assert
- r300g: cache packet dwords of 3D_LOAD_VBPNTR in a command buffer if possible
- r300g: validate buffers only if any of bound buffers is changed
- r300g: also revalidate the SWTCL vertex buffer after its reallocation
- r300/compiler: don't terminate regalloc if we surpass max temps limit
- r300/compiler: add a function to query program stats (alu, tex, temps..)
- r300/compiler: cleanup rc_run_compiler
- r300/compiler: do not print pair/tex/presub program stats for vertex shaders
- r300/compiler: handle DPH and XPD in rc_compute_sources_for_writemask
- r300/compiler: make lowering passes possibly use up to two less temps
- r300/compiler: remove at least unused immediates if externals cannot be removed
- r300/compiler: fix LIT in VS
- r300/compiler: fix swizzle lowering with a presubtract source operand
- r300g: fix rendering with a vertex attrib having a zero stride
- r300g: finally fix the texture corruption on r3xx-r4xx
- r300g/swtcl: re-enable LLVM
- r300g: mark vertex arrays as dirty after a buffer_offset change
- mesa: fix texel store functions for some float formats
- r300/compiler: disable the rename_regs pass for loops
Mario Kleiner (1):
- mesa/r300classic: Fix dri2Invalidate/radeon_prepare_render for page flipping.
Mathias Fröhlich (3):
- r300g: Avoid returning values in a static array, fixing a potential race
- r600g: Only compare active vertex elements
- st/mesa: Set PIPE_TRANSFER_DISCARD for GL_MAP_INVALIDATE_RANGE/BUFFFER_BIT
Michal Krol (10):
- svga: Fix relative addressing translation for pixel shaders.
- svga: Integer constant register file has a separate namespace.
- tgsi/exec: Cleanup the remaining arithmetic instructions.
- tgsi/exec: Get rid of obsolete condition codes.
- tgsi/build: Reduce interface clutter.
- graw/gdi: Initial commit.
- scons: Hook-up graw-gdi target.
- graw/gdi: Fix window dimensions.
- os: Open file streams in binary mode.
- graw: Export graw_save_surface_to_file().
Nicolas Kaiser (26):
- swrast: remove duplicated include
- egl: remove duplicated include
- gallium/rtasm: remove duplicated include
- gallium/util: remove duplicated include
- gallium/i915: remove duplicated include
- gallium/llvmpipe: remove duplicated include
- gallium/softpipe: remove duplicated include
- gallium/st: remove duplicated includes
- gallium/winsys: remove duplicated include
- glx: remove duplicated include
- dri/common: remove duplicated include
- dri/i810: remove duplicated include
- dri/i915: remove duplicated include
- dri/i965: remove duplicated include
- dri/intel: remove duplicated include
- dri/mga: remove duplicated include
- dri/r128: remove duplicated include
- dri/r300: remove duplicated include
- dri/r600: remove duplicated include
- dri/radeon: remove duplicated includes
- dri/savage: remove duplicated include
- main: remove duplicated includes
- math: remove duplicated includes
- st: remove duplicated include
- i965g: use Elements macro instead of manual sizeofs
- nv50: fix always true conditional in shader optimization
Orion Poplawski (1):
- osmesa: link against libtalloc
Owen W. Taylor (1):
- r600g: Fix location for clip plane registers
Peter Clifton (3):
- intel: Fix emit_linear_blit to use DWORD aligned width blits
- intel: Add assert check for blitting alignment.
- meta: Mask Stencil.Clear against stencilMax in _mesa_meta_Clear
Robert Hooker (2):
- intel: Add a new B43 pci id.
- egl_dri2: Add missing intel chip ids.
Roland Scheidegger (16):
- gallivm: fix copy&paste bug
- gallivm: don't use URem/UDiv when calculating offsets for blocks
- gallivm: optimize yuv decoding
- gallivm: fix trunc/itrunc comment
- gallivm: faster iround implementation for sse2
- gallivm: replace sub/floor/ifloor combo with ifloor_fract
- gallivm: optimize some tex wrap mode calculations a bit
- gallivm: more linear tex wrap mode calculation simplification
- gallivm: avoid unnecessary URem in linear wrap repeat case
- gallivm: optimize soa linear clamp to edge wrap mode a bit
- gallivm: make use of new iround code in lp_bld_conv.
- gallivm: fix different handling of [non]normalized coords in linear soa path
- gallivm: only use lp_build_conv 4x4f -> 1x16 ub fastpath with sse2
- r200: fix r200 large points
- mesa: remove unneeded DD_POINT_SIZE and DD_LINE_WIDTH tricaps
- gallium: support for array textures and related changes
Shuang He (1):
- mesa: allow GLfixed arrays for OpenGL ES 2.0
Stephan Schmid (1):
- r600g: fix relative addressing when splitting constant accesses
Thomas Hellstrom (21):
- st/xorg: Don't try to use option values before processing options
- xorg/vmwgfx: Make vmwarectrl work also on 64-bit servers
- st/xorg: Add a customizer option to get rid of annoying cursor update flicker
- xorg/vmwgfx: Don't hide HW cursors when updating them
- st/xorg: Don't try to remove invalid fbs
- st/xorg: Fix typo
- st/xorg, xorg/vmwgfx: Be a bit more frendly towards cross-compiling environments
- st/xorg: Fix compilation errors for Xservers compiled without Composite
- st/xorg: Don't use deprecated x*alloc / xfree functions
- xorg/vmwgfx: Don't use deprecated x*alloc / xfree functions
- st/xorg: Fix compilation for Xservers >= 1.10
- mesa: Make sure we have the talloc cflags when using the talloc headers
- egl: Add an include for size_t
- mesa: Add talloc includes for gles
- st/egl: Fix build for include files in nonstandard places
- svga/drm: Optionally resolve calls to powf during link-time
- gallium/targets: Trivial crosscompiling fix
- st/xorg: Add a function to flush pending rendering and damage
- gallium/targets/xorg-vmwgfx: Xv fixes
- xorg/vmwgfx: Flush even if we don't autopaint the color key
- xorg/vmwgfx: Don't clip video to viewport
Tilman Sauerbeck (35):
- r600g: Fixed a bo leak in r600_blit_state_ps_shader().
- r600g: Use clamped math for RCP and RSQ.
- r600g: Formatting fixes.
- r600g: Added DB_SHADER_CONTROL defines.
- r600g: Only set PA_SC_EDGERULE on rv770 and greater.
- r600g: Enable PIPE_SHADER_CAP_TGSI_CONT_SUPPORTED.
- r600g: Fixed the shift in S_02880C_KILL_ENABLE.
- glsl2: Empty functions can be inlined.
- glsl2: Fixed cloning of ir_call error instructions.
- r600g: Added support for TGSI_SEMANTIC_FACE.
- gallium/docs: Fixed a typo in the SCS opcode description.
- r600g: Honour destination operand's writemask in the SCS implementation.
- r600g: Implemented the Z and W component write for the SCS opcode.
- python/tests: Fixed tri.py for API and TGSI syntax changes.
- r600g: Removed debug code.
- gallium/docs: The RET opcode may appear anywhere in a subroutine.
- r600g: Destroy the blitter.
- r600g: Fixed two texture surface leaks in r600_blit_uncompress_depth().
- r600g: Cleaned up index buffer reference handling in the draw module.
- r600g: Fixed r600_vertex_element leak.
- r600g: Added r600_pipe_shader_destroy().
- r600g: Also clear bc data when we're destroying a shader.
- r600g: In radeon_bo(), call LIST_INITHEAD early.
- r600g: Destroy the blitter.
- r600g: Removed unused 'ptr' argument from radeon_bo().
- r600g: Made radeon_bo_pb_map_internal() actually call radeon_bo_map().
- r600g: Fixed unmap condition in radeon_bo_pb_destroy().
- r600g: Made radeon_bo::map_count signed.
- r600g: We don't support PIPE_CAP_PRIMITIVE_RESTART.
- r600g: Delete custom_dsa_flush on shutdown.
- r600g: Fixed two memory leaks in winsys.
- r600g: Destroy the winsys in r600_destroy_screen().
- st/mesa: Reset the index buffer before destroying the pipe context.
- st/mesa: Reset the constant buffers before destroying the pipe context.
- r600g: Removed duplicated call to tgsi_split_literal_constant().
Timo Wiren (1):
- Fix typos in comments and debug output strings.
Tom Fogal (3):
- Implement x86_64 atomics for compilers w/o intrinsics.
- Prefer intrinsics to handrolled atomic ops.
- Revert "Prefer intrinsics to handrolled atomic ops."
Tom Stellard (32):
- r300/compiler: Refactor the pair instruction data structures
- r300g: Always try to build libr300compiler.a
- r300/compiler: Fix two mistakes in the presubtract optimization pass.
- r300/compiler: Add more helper functions for iterating through sources
- r300/compiler: Print immediate values after "dead constants" pass
- r300/compiler: radeon_remove_constants.c: fix indentation
- r300/compiler: Use rc_for_all_reads_src() in "dead constants" pass
- r300/compiler: Fix segfault in error path
- r300/compiler: Don't use rc_error() unless the error is unrecoverable
- r300/compiler: Don't merge instructions that write output regs and ALU result
- r300/compiler: Create a helper function for merging presubtract sources
- r300/compiler: Fix incorrect assumption
- r300/compiler: Clear empty registers after constant folding
- r300/compiler: Add a new function for more efficient dataflow analysis
- r300g: Add new debug option for logging vertex/fragment program stats
- r300/compiler: Use rc_get_readers_normal() for presubtract optimizations
- r300/compiler: Don't clobber presubtract sources during optimizations
- r300/compiler: Don't track readers into an IF block.
- r300/compiler: Make sure presubtract sources use supported swizzles
- r300/compiler: Fix register allocator's handling of loops
- r300/compiler: Fix instruction scheduling within IF blocks
- r300/compiler: Use zero as the register index for unused sources
- r300/compiler: Ignore alpha dest register when replicating the result
- r300/compiler: Add rc_get_readers()
- r300/compiler: Handle BREAK and CONTINUE in rc_get_readers()
- r300/compiler: Track readers through branches in rc_get_readers()
- r300/compiler: Convert RGB to alpha in the scheduler
- r300/compiler: Use presubtract operations as much as possible
- r300/compiler: Enable rename_reg pass for r500 cards
- r300/compiler: Add a more efficient version of rc_find_free_temporary()
- r300/compiler: Don't allow presubtract sources to be remapped twice
- r300/compiler: Fix black terrain in Civ4
Victor Tseng (1):
- egl/i965: include inline_wrapper_sw_helper.h
Viktor Novotný (6):
- dri/nouveau: Import headers from rules-ng-ng
- dri/nouveau: nv04: Use rules-ng-ng headers
- dri/nouveau: nv10: Use rules-ng-ng headers
- dri/nouveau nv20: Use rules-ng-ng headers
- dri/nouveau: Remove nouveau_class.h, finishing switch to rules-ng-ng headers
- dri/nouveau: Clean up magic numbers in get_rt_format
Vinson Lee (214):
- llvmpipe: Remove unnecessary header.
- r600g: Remove unnecessary headers.
- mesa: Include missing header in program.h.
- glsl: Fix 'format not a string literal and no format arguments' warning.
- r600g: Silence uninitialized variable warning.
- r600g: Silence uninitialized variable warning.
- nvfx: Silence uninitialized variable warnings.
- r600g: Silence uninitialized variable warning.
- r600g: Silence uninitialized variable warning.
- r600g: Silence uninitialized variable warning.
- r600g: Silence unused variable warning.
- nv50: Update files in SConscript to match Makefile.
- nv50: Remove unnecessary headers.
- nv50: Silence uninitialized variable warning.
- nv50: Silence uninitialized variable warning.
- nv50: Silence uninitialized variable warning.
- gallivm: Remove unnecessary headers.
- draw: Remove unnecessary header.
- nv50: Silence uninitialized variable warnings.
- nv50: Fix 'control reaches end of non-void function' warning.
- mesa/st: Silence uninitialized variable warning.
- gallivm: Remove unnecessary header.
- r600g: Remove unnecessary header.
- r600g: Remove unnecessary headers.
- r600g: Fix implicit declaration warning.
- r600g: Fix memory leak on error path.
- r600g: Silence uninitialized variable warning.
- r600g: Silence unused variable warnings.
- mesa: bump version to 7.10
- ir_to_mesa: Remove unused member array_indexed from struct statevar_element.
- mesa: Silence "'valid_texture_object' defined but not used" warning.
- x86: Silence unused variable warning on Mac OS X.
- glsl: Fix 'control reaches end of non-void function' warning.
- nvfx: Remove const qualifer from nvfx_vertprog_translate.
- nvfx: Silence uninitialized variable warnings.
- r600g: Remove unused variable.
- nv50: Silence missing initializer warning.
- nv50: Remove dead initialization.
- nv50: Remove dead initialization.
- tgsi: Remove duplicate case value.
- glut: Define markWindowHidden for non-Windows only.
- glut: Define eventParser for non-Windows only.
- r300g: Silence uninitialized variable warning.
- intel: Fix implicit declaration of function '_mesa_meta_Bitmap' warning.
- mesa: Remove unnecessary headers.
- r600g: Remove unnecessary header.
- unichrome: Remove unnecessary header.
- intel: Remove unnecessary headers.
- r600g: Remove unused variable.
- r600g: Disable unused variables.
- r600g: Remove unused variable.
- r600g: Silence 'control reaches end of non-void function' warning.
- r600g: Remove unused variable.
- r600g: Remove unused variable.
- r600g: Disable unused variables.
- intel: Remove unnecessary header.
- st/dri: Remove unnecessary header.
- r600g: Remove unused variable.
- r300g: Remove unused variable.
- r600g: Don't return a value in function returning void.
- r600g: Remove unused variables.
- r600g: Include p_compiler.h instead of malloc.h.
- r600g: Silence uninitialized variable warnings.
- scons: Add MinGW-w64 prefixes for MinGW build.
- dri: Add GET_PROGRAM_NAME definition for Mac OS X.
- scons: Add program/sampler.cpp to SCons build.
- mesa: Fix printf format warning.
- mesa: Fix printf format warning.
- mesa: Fix printf format warning.
- r300/compiler: Move declaration before code.
- r300/compiler: Move declaration before code.
- r300/compiler: Move declaration before code.
- r300/compiler: Move declaration before code.
- r600g: Update SConscript.
- r300/compiler: Move declaration before code.
- r600g: Update SConscript.
- r300/compiler: Move declaration before code.
- r600g: Update SConscript.
- r300/compiler: Move declaration before code.
- r600g: Fix SCons build.
- r300/compiler: Move declaration before code.
- r300/compiler: Move declaration before code.
- r300/compiler: Move declaration before code.
- r300/compiler: Remove declaration before code.
- r300/compiler: Move declaration before code.
- r300/compiler: Move declaration before code.
- r300/compiler: Move declaration before code.
- r300/compiler: Move declaration before code.
- glsl: Remove unnecessary header.
- savage: Remove unnecessary header.
- r600g: Remove unused variable.
- r600g: Remove unnecessary headers.
- r600g: Fix SCons build.
- r600g: Remove unnecessary header.
- gallivm: Remove unnecessary header.
- r600g: Silence uninitialized variable warning.
- r600g: Silence uninitialized variable warning.
- r600g: Silence uninitialized variable warning.
- i915: Silence unused variable warning in non-debug builds.
- i915: Silence unused variable warning in non-debug builds.
- i965: Silence unused variable warning on non-debug builds.
- i965: Silence unused variable warning on non-debug builds.
- i965: Initialize member variables.
- r300: Silence uninitialized variable warning.
- tdfx: Silence unused variable warning on non-debug builds.
- gallivm: Remove unnecessary header.
- glsl: Initialize variable in ir_derefence_array::constant_expression_value
- mesa: Add missing header to shaderobj.h.
- llvmpipe: Return non-zero exit code for lp_test_round failures.
- r300/compiler: Remove unused variable.
- st/xorg: Fix memory leak on error path.
- llvmpipe: Initialize state variable in debug_bin function.
- llvmpipe: Initialize variable.
- draw: Move loop variable declaration outside for loop.
- r600g: Ensure r600_src is initialized in tgsi_exp function.
- glsl: Add assert for unhandled ir_shader case.
- swrast: Print out format on unexpected failure in _swrast_DrawPixels.
- llvmpipe: Remove unnecessary header.
- draw: Remove unnecessary header.
- gallivm: Silence uninitialized variable warnings.
- gallivm: Silence uninitialized variable warnings.
- gallivm: Silence uninitialized variable warning.
- r300g: Silence uninitialized variable warning.
- mesa: Remove unnecessary headers.
- r600g: Silence uninitialized variable warnings.
- st/mesa: Remove unnecessary header.
- mesa: Remove unnecessary header.
- egl: Remove unnecessary headers.
- swrast: Print out format on unexpected failure in _swrast_ReadPixels.
- st/mesa: Silence uninitialized variable warning.
- savage: Remove unnecessary header.
- st/vega: Remove unnecessary headers.
- dri/nouveau: Silence uninitialized variable warning.
- r300/compiler: Move declaration before code.
- r300/compiler: Move declaration before code.
- i965: Silence uninitialized variable warning.
- i965: Silence uninitialized variable warning.
- mesa: Clean up header file inclusion in accum.h.
- mesa: Clean up header file inclusion in version.h.
- mesa: Clean up header file inclusion in api_loopback.h.
- mesa: Clean up header file inclusion in api_validate.h.
- mesa: Include mfeatures.h in api_loopback for FEATURE_beginend.
- mesa: Include mfeatures.h in api_validate.c for FEATURE_* symbols.
- mesa: Clean up header file inclusion in arrayobj.h.
- mesa: Clean up header file inclusion in atifragshader.h.
- mesa: Clean up header file inclusion in attrib.h.
- mesa: Clean up header file inclusion in blend.h.
- mesa: Clean up header file inclusion in buffers.h.
- mesa: Clean up header file inclusion in colortab.h.
- mesa: Clean up header file inclusion in convolve.h.
- mesa: Clean up header file inclusion in debug.h.
- mesa: Clean up header file inclusion in depth.h.
- mesa: Clean up header file inclusion in depthstencil.h.
- mesa: Clean up header file inclusion in drawpix.h.
- mesa: Clean up header file inclusion in drawtex.h.
- mesa: Clean up header file inclusion in enable.h.
- mesa: Clean up header file inclusion in extensions.h.
- graw: Add struct pipe_surface forward declaration.
- mesa: Clean up header file inclusion in fbobject.h.
- mesa: Clean up header file inclusion in ffvertex_prog.h.
- mesa: Clean up header file inclusion in fog.h.
- mesa: Clean up header file inclusion in framebuffer.h.
- mesa: Clean up header file inclusion in hint.h.
- mesa: Clean up header file inclusion in histogram.h.
- mesa: Clean up header file inclusion in image.h.
- mesa: Add missing header and forward declarations in dd.h.
- mesa: Clean up header file inclusion in light.h.
- mesa: Clean up header file inclusion in lines.h.
- mesa: Clean up header file inclusion in matrix.h.
- mesa: Clean up header file inclusion in multisample.h.
- mesa: Clean up header file inclusion in nvprogram.h.
- winsys/xlib: Add cygwin to SConscript.
- mesa: Clean up header file inclusion in pixel.h.
- mesa: Clean up header file inclusion in pixelstore.h.
- mesa: Fix printf format warnings.
- mesa: Clean up header file inclusion in points.h.
- i965: Silence uninitialized variable warning.
- glsl: Add ir_constant_expression.cpp to SConscript.
- mesa: Add definitions for inverse hyperbolic function on MSVC.
- glsl: Fix 'control reaches end of non-void function' warning.
- glsl: Add lower_vector.cpp to SConscript.
- glsl: Fix type of label 'default' in switch statement.
- st/mesa: Remove unnecessary headers.
- swrast: Remove unnecessary header.
- r600: Remove unnecesary header.
- intel: Remove unnecessary header.
- mesa: Clean up header file inclusion in polygon.h.
- mesa: Clean up header file inclusion in rastpos.h.
- mesa: Clean up header file inclusion in readpix.h.
- mesa: Clean up header file inclusion in renderbuffer.h.
- mesa: Clean up header file inclusion in scissor.h.
- mesa: Clean up header file inclusion in shaderapi.h.
- mesa: Clean up header file inclusion in shared.h.
- mesa: Clean up header file inclusion in stencil.h.
- r600: Remove unnecessary header.
- llvmpipe: Remove unnecessary headers.
- mesa: Clean up header file inclusion in syncobj.h.
- r300/compiler: Move declaration before code.
- r300/compiler: Move declaration before code.
- mesa: Clean up header file inclusion in texcompress.h.
- st/vega: Silence uninitialized variable warning.
- mesa: Clean up header file inclusion in texcompress_s3tc.h.
- mesa: Clean up header file inclusion in texenvprogram.h.
- mesa: Clean up header file inclusion in texformat.h.
- mesa: Clean up header file inclusion in texgetimage.h.
- mesa: Clean up header file inclusion in texobj.h.
- gallium/noop: Add prototype for noop_init_state_functions.
- mesa: Clean up header file inclusion in texrender.h.
- mesa: Clean up header file inclusion in transformfeedback.h.
- mesa: Clean up header file inclusion in varray.h.
- mesa: Clean up header file inclusion in viewport.h.
- r200: Silence uninitialized variable warning.
- r600g: Fix SCons build.
- i965: Silence uninitialized variable warning.
Xavier Chantry (8):
- nv50: fix size of outputs_written array
- nv50: apply layout_mask to tile_flags
- nvfx: only expose one rt on nv30
- nvfx: fb->nr_cbufs <= 1 on nv30
- nvfx: reset nvfx->hw_zeta
- nvfx: fixes after array textures merge
- init ps->context with util_surfaces_get and do_get
- gallium/trace: check bind_vertex_sampler_states and set_vertex_sampler_views
Xiang, Haihao (10):
- mesa: fix regression from b4bb6680200b5a898583392f4c831c02f41e63f7
- i965: add support for polygon mode on Sandybridge.
- i965: fix for flat shading on Sandybridge
- i965: set minimum/maximum Point Width on Sandybridge
- meta: allow nested meta operations
- i965: support for two-sided lighting on Sandybridge
- i965: fix register region description
- i965: use align1 access mode for instructions with execSize=1 in VS
- i965: don't spawn GS thread for LINELOOP on Sandybridge
- i965: use BLT to clear buffer if possible on Sandybridge
Zack Rusin (8):
- rbug: fix rbug when contexts are being destroyed
- llvmpipe: fix rasterization of vertical lines on pixel boundaries
- scons: build the xorg state trackers only when env includes drm
- gallivm: implement indirect addressing of the output registers
- gallivm: implement indirect addressing over inputs
- gallivm: fix storing of the addr register
- scons: add alias for identity
- gallium/util: add states relevant to geometry shaders
Zhenyu Wang (40):
- i965: disasm quarter and write enable instruction control on sandybridge
- i965: new state dump for sandybridge
- i965: enable accumulator update in PS kernel too on sandybridge
- i965: Fix color interpolation on sandybridge
- i965: force zero in clipper to ignore RTAIndex on sandybridge
- i965: fix point size setting in header on sandybridge
- i965: ff sync message change for sandybridge
- i965: ignore quads for GS kernel on sandybridge
- i965: add sandybridge viewport state bo into validation list
- i965: VS use SPF mode on sandybridge for now
- i965: fix jump count on sandybridge
- i965: Fix sampler on sandybridge
- i965: fix const register count for sandybridge
- i965: Add all device ids for sandybridge
- i965: sandybridge pipe control workaround before write cache flush
- i965: only allow SIMD8 kernel on sandybridge now
- i965: don't do calculation for delta_xy on sandybridge
- i965: fix pixel w interpolation on sandybridge
- i965: enable polygon offset on sandybridge
- i965: fix scissor state on sandybridge
- i965: fix point sprite on sandybridge
- i965: fix occlusion query on sandybridge
- i965: fallback bitmap operation on sandybridge
- i965: Always set tiling for depth buffer on sandybridge
- i965: fallback lineloop on sandybridge for now
- Revert "i965: Always set tiling for depth buffer on sandybridge"
- i965: always set tiling for fbo depth buffer on sandybridge
- i965: Fix GS hang on Sandybridge
- Revert "i965: fallback lineloop on sandybridge for now"
- i965: refresh wm push constant also for BRW_NEW_FRAMENT_PROGRAM on gen6
- i965: fix dest type of 'endif' on sandybridge
- Revert "i965: VS use SPF mode on sandybridge for now"
- i965: also using align1 mode for math2 on sandybridge
- i965: Fix GS state uploading on Sandybridge
- i965: upload WM state for _NEW_POLYGON on sandybridge
- i965: Use MI_FLUSH_DW for blt ring flush on sandybridge
- i965: explicit tell header present for fb write on sandybridge
- i965: Fix occlusion query on sandybridge
- i965: Use last vertex convention for quad provoking vertex on sandybridge
- i965: Fix provoking vertex select in clip state for sandybridge
Zou Nan hai (1):
- i965: skip too small size mipmap
delphi (2):
- draw: added userclip planes and updated variant_key
- draw: some changes to allow for runtime changes to userclip planes
nobled (3):
- r300g: Abort if atom allocations fail
- r300g: Abort if draw_create() fails
- r300g: Drop unnecessary cast
pontus lidman (1):
- mesa: check for posix_memalign() errors
richard (2):
- evergreen : fix z format setting, enable stencil.
- r600c : inline vertex format is not updated in an app, switch to use vfetch constants. For the 7.9 and 7.10 branches as well.