mesa: Changes to 'refs/tags/mesa-17.2.0-rc3-1'
Tag 'mesa-17.2.0-rc3-1' created by Timo Aaltonen <tjaalton@debian.org> at 2017-08-08 06:38 +0000
Tagging upload of mesa 17.2.0~rc3-1 to experimental.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1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=/9/0
-----END PGP SIGNATURE-----
Changes since mesa-17.1.5-1:
Aaron Watry (2):
clover/device: Get device/host unified memory from pipe driver
radeon/winsys: Limit max allocation size to 70% of VRAM
Adam Jackson (2):
egl/x11: Honor the EGL_PLATFORM_X11_SCREEN_EXT attribute
egl/platform/drm: Don't take display ownership until gbm is initialized
Alejandro Piñeiro (1):
egl: remove unused err variable
Aleksander Morgado (2):
etnaviv: don't dereference etna_resource pointer if allocation fails
etnaviv: fix refcnt initialization in etna_screen
Alex Deucher (2):
radeonsi: add new vega10 pci ids
radeonsi: add new polaris12 pci id
Alex Smith (10):
util/vulkan: Move Vulkan utilities to src/vulkan/util
radv/vulkan: Move radv_get_driver_version to src/vulkan/util
anv: Set driver version to Mesa version
anv: Set better descriptor set limits
ac/nir: Use correct LLVM intrinsics for atomic ops on imageBuffers
ac/nir: Fix ordering of parameters for image atomic cmpswap intrinsics
radv: Fix possible invalid free of dynamic descriptors
radv: Fix descriptors for cube images with VK_IMAGE_USAGE_STORAGE_BIT
radv: Set the RADEON_SURF_OPTIMIZE_FOR_SPACE flag for images
radv: Generate storage image descriptors unconditionally
Alexandre Courbot (1):
nvc0: support for GP10B
Alexandre Demers (1):
osmesa: link with libunwind if enabled (v2)
Andres Gomez (38):
docs: add release notes for 17.0.5
docs: add sha256 checksums for 17.0.5
docs: add news item and link release notes for 17.0.5
travis: replace Trusty-based LLVM toolchain apt-get with apt addon
travis: add the possibility of using the txc-dxtn library
glsl: Corrected some typos and error messages
bin/get-fixes-pick-list.sh: don't warn if more than one, go over them
docs/releasing: format/style homogenization
docs/releasing: check in master for forgotten nomination candidates
docs/releasing: further explain the build/check testing process
docs/releasing: if possible, do some every day use on the RC
docs/releasing: added relevant people for build/check with Windows
docs/releasing: added relevant people for build/check with Android
docs/releasing: added relevant people for build/check with MacOSX
bin: use tabs for coding style on *.sh files
bin/*py: honor editorconfig formatting
docs: update "Release manager" column
docs: extend until the end of August
bin/get-fixes-pick-list.sh: bring back the warning
docs: add release notes for 17.0.6
docs: add sha256 checksums for 17.0.6
docs: update calendar, add news item and link release notes for 17.0.6
docs: small release calendar fixes
bin/get-fixes-pick-list.sh: parse just the commit message
bin/get-fixes-pick-list.sh: better identify multiple "fixes:" tags
anv: FORMAT_FEATURE_TRANSFER_SRC/DST_BIT_KHR not used with VkFormatProperties.bufferFeatures
docs: add release notes for 17.1.4
docs: add sha256 checksums for 17.1.4
docs: update calendar, add news item and link release notes for 17.1.4
nir/spirv: Remove unnecessary comment.
docs: add release notes for 17.1.5
docs: add sha256 checksums for 17.1.5
docs: update calendar, add news item and link release notes for 17.1.5
docs: add instructions to specify LLVM version for basic testing
docs: avoid overwrite of LD_LIBRARY_PATH during basic testing
docs: update master's release notes, news and calendar commit
travis: add missing wayland-protocols
broadcom: correct header file in BROADCOM_FILES
Anuj Phogat (38):
i965: Rename brw_format variable names to isl_format
i965: Remove an extra semicolon
i965: Add and initialize l3_banks field for gen7+
i965: Simplify l3 way size computations
intel: Fix broxton 2x6 way size computation
i965/cnl: Define genX(x) and GENX(x) for gen10
i965/cnl: Include gen10_pack.h
i965/cnl: Add gen10 specific function declarations
intel/genxml: Update genx_bits for gen10+
i965/cnl: Wire up Mesa build files for gen10
i965/cnl: Wire up android Mesa build files for gen10
i965/cnl: Add pci id for INTEL_DEVID_OVERRIDE
i965/cnl: Add cnl bits in aubinator
i965/cnl: Update few assertions
i965/cnl: Handle gen10 in switch cases across the driver
i965/cnl: Start using CNL MOCS defines
i965/cnl: Don't resolve single sampled color rb in case of sRGB formats
i965/cnl: Make URB {VS, GS, HS, DS} sizes non multiple of 3
i965/cnl: Reformat surface_format_info table to accomodate gen10+
i965/cnl: Enable CCS_E and RT support for few formats
intel/isl: Use uint64_t to store total surface size
intel/isl: Add the maximum surface size limit
i965: Fix broxton 2x6 l3 config
i965: Add a variable for way size per bank in get_l3_way_size()
intel/genxml: Add better support for MI_MATH in gen10
intel/genxml: Add INSTDONE registers in gen10
intel/genxml: Combine DataDWord{0, 1} fields in to ImmediateData field
intel/genxml: Rename IndirectStatePointer to BorderColorPointer
intel/genxml: Rename StartInstanceLocation to StartingInstanceLocation
intel/genxml: Add Gen10 CACHE_MODE_1 definitions
anv/cnl: Add #defines for MOCS and genX(x)
anv/cnl: Use GENX(xx) in place of GEN9_xx
anv/cnl: Don't set FloatBlendOptimizationEnable{Mask}
anv/cnl: Generate and use gen10 functions
intel: Enable vulkan build for gen10
i965/cnl: Don't write to Cache Mode Register 1 on gen10+
anv/cnl: Don't write to Cache Mode Register 1 on gen10+
intel/compiler: Don't use opt_sampler_eot() optimization on gen10+
Anusha Srivatsa (1):
i965/CFL: Add PCI Ids for Coffee Lake.
Axel Davy (1):
st/nine: Fix a regression and syntax cleanup
Bartosz Tomczyk (3):
mesa/glthread: add async support to ARB_viewport_array functions
mesa/vbo: reduce prim array size
mesa: Avoid leaking surface in st_renderbuffer_delete
Bas Nieuwenhuizen (52):
radv: Don't allocate dynamic descriptors separately.
radv: Use host memory pool for non-freeable descriptors.
radv: Don't try to find gaps for non-freeable descriptors.
radv: Remove binding buffer count.
radv: Add shader prefetch.
radv: Set variant code_size when created from the cache.
radv: Prefetch compute shader too.
radv: Use the correct pipeline for dispatches.
radv: Enable lowering fdiv in nir.
radv: Enable userspace fence checking.
radv: Add NIR loop unrolling.
radv: Add top of pipe timestamp queries.
radv: Use correct stage for ready bit.
radv: Don't set dynamic state for pipelines with rasterizer dicard.
radv: Add userspace fence buffer per context.
radv: Don't use FLAT_SHADE for constants.
radv: Save descriptor set even if vertex buffers are not saved.
radv: Don't use a separate can_expclear.
radv: Specify semantics of HTILE layout helpers.
radv: Add queue masks for htile usage determination.
radv: Use correct clear words for HTILE.
radv: Add compute HTILE fast clear.
radv: Also signal fence if vkAcquireNextImageKHR returns VK_SUBOPTIMAL_KHR.
radv: Reserve space for descriptor and push constant user SGPR setting.
radv: Revert HTILE reset word to 0xFFFFFFFF.
radv: Set both compute and graphics SGPRS on descriptor set flush.
radv: Dirty all descriptors sets when changing the pipeline.
radv: Add support for external queue family.
radv: Add VkPhysicalDeviceIDProperties support.
ac/surface: Fix HTILE for radv.
radv: Don't use a divide by index_size.
radv: Remove vertex_descriptors_dirty.
radv: Add early exit for cache flushes.
radv: Move pipeline stuff from flush_state to emit_graphics_pipeline.
radv: Split out updating the vertex descriptors.
radv: Remove SI num RB override for occlusion queries.
radv: Use correct image layout for blit based copies.
radv: Remove unused args of radv_image_view_init.
radv: Use v4i32 variant of llvm.SI.load.const.
radv: Disable depth & stencil tests when the depthbuffer doesn't support it.
radv: Always set depthbuffer using image format instead of iview format.
ac/nir: Move VS position exports before param exports.
radv: Add compute htile clear for combined depth+stencil surfaces.
radv: Add support for VK_KHR_get_memory_requirements2.
radv: Add support for VK_KHR_dedicated_allocation.
radv: Use the KHR dedicated alloc for the WSI.
radv: Remove NV dedicated alloc extension.
radv: Add VK_KHR_storage_buffer_storage_class support.
radv: Add support for VK_KHR_variable_pointers.
radv: Handle VK_ATTACHMENT_UNUSED in color attachments.
radv: Don't segfault when exporting an image which hasn't been bound yet.
radv: Don't underflow non-visible VRAM size.
Ben Boeckel (1):
scons: update for LLVM 4.0
Ben Crocker (2):
egl_dri2: swrastGetDrawableInfo: set *x, *y [v2]
mapi: Enable assembly language API acceleration for PPC64LE (V2)
Ben Skeggs (2):
nvc0: copy engine surface params are only relevant for tiled surfaces
nvc0: fix transfer of larger rectangles with DmaCopy on gk104 and up
Ben Widawsky (8):
i965: Allocate tile aligned height
i965: Make feature macros gen8 based
i965/cnl: Implement depth count workaround
i965/cnl: Add a preliminary device for Cannonlake
i965/cnl: Add l3 configuration for Cannonlake
i965: Use already existing eu_total
i965/miptree: Add a return for updating of winsys
i965/miptree: Allocate mt earlier in update winsys
Benedikt Schemmer (1):
drirc: Add allow_glsl_builtin_variable_redeclaration for Dead Island Riptide Definitive Edition
Boyan Ding (4):
nouveau: Fix indentation of maxwell compute class definitions
nvc0: Fix index of unk fields in nve4_cp_launch_desc
nvc0: Add new launch descriptor format for GP100
nvc0: Enable compute support for Pascal
Brian Paul (101):
mesa: fix Windows build error related to getuid()
mesa: print target string in glBindTexture() error message
st/mesa: minor clean-ups in st_update_renderbuffer_surface()
draw: whitespace fixes in draw_pipe_vbuf.c
svga: fail screen creation if HW version is too old
svga: add DEBUG_SAMPLERS
svga: improve surface view debug messages
svga: add const qualifiers on svga_check_sampler_view_resource_collision()
gallium/util: add some comments in u_debug_flush.c
gallium/util: reduce util_snprintf() calls in debug_flush_might_flush_cb()
svga: fix vertex buffer binding issue
st/mesa: whitespace clean-ups in st_manager.c
mesa: optimize color_buffer_writes_enabled()
st/mesa: add more fallback gallium formats for GL integer formats
st/wgl: whitespace/formatting fixes in stw_ext_pixelformat.c
st/wgl: use ARRAY_SIZE() macro in wglChoosePixelFormatARB()
st/wgl: whitespace, formatting fixes in stw_pixelformat.c
mesa: add const qualifier on _mesa_valid_to_render()
Revert "gallium: remove unused PIPE_CC_GCC_VERSION"
svga: Fix MSVC build.
svga: log the process command line to the vmware.log file
svga: init local vars to silence uninitialized use warnings
svga: move logging initialization code into new function
svga: init svga_screen::swc_mutex with mtx_recursive
glsl: Fix g++ initializer order warning
st/wgl: whitespace, formatting fixes in stw_device.c
svga: document some incorrect VGPU10 shader translation issues
xlib: fix glXGetCurrentDisplay() failure
mesa: document range of SampleCoverageValue, MinSampleShadingValue
gallium/u_threaded: fixes for MSVC
tgsi: clarify TGSI_SEMANTIC_SAMPLEMASK documentation
gallium/util: whitespace, formatting fixes in u_upload_mgr.c
st/mesa: unmap the stream_uploader buffer before drawing
gallium/docs: clarify TGSI_SEMANTIC_SAMPLEMASK, again
automake: increase the MESA_GIT_SHA1 hash id length from 7 to 10 digits
draw: whitespace and formatting fixes
svga: clamp device line width to at least 1 to fix HWv8 line stippling
draw: check for line_width != 1.0f in validate_pipeline()
svga: add some missing SVGA_STATS_* enum values, prefix strings
gallium/docs: improve docs for SAMPLE_POS, SAMPLE_INFO, TXQS, MSAA semantics
mesa: whitespace fixes in enable.c
svga: add a few minor comments
gallium/hud: support GALLIUM_HUD_DUMP_DIR feature on Windows
svga: add new num-failed-allocations HUD query
softpipe: remove unused softpipe_context::line_stipple_counter
translate: whitespace fixes in translate_generic.c
gallium/vbuf: add some const qualifiers
gallium/vbuf: avoid segfault when we get invalid glDrawRangeElements()
svga: check return value from svga_set_shader( SVGA3D_SHADERTYPE_GS, NULL)
svga: minor whitespace fixes in svga_pipe_vertex.c
pipe_loader_sw: fix compilation warning
scons: add code to generate format_fallback.c file
svga: change error handling convention for svga_set_stream_output()
svga: whitespace fixes in svga_resource_buffer_upload.c
svga: update a few surface format names
gallium/docs: clarify that samplers are not used with PIPE_BUFFER resources
gallium/docs: document that TXF is used with PIPE_BUFFER resources
st/mesa: check for incomplete texture in st_finalize_texture()
st/mesa: fix texture image resource selection in st_render_texture()
svga: add texture size/levels sanity check code in svga_texture_create()
docs: update bug reporting guidelines
svga: fix buffer binding flags initialization
svga: move assertion in draw_vgpu10()
svga: fix texture buffer object regression
svga: don't call svga_texture_device_format_has_alpha() for PIPE_BUFFER
st/mesa: whitespace fixes in st_texture.c
st/mesa: whitespace fixes in st_cb_fbo.c
st/mesa: find proper mipmap level in st_ClearTexSubImage()
mesa: new comments, assertion related to glClearTexSubImage
mesa: simplify get_tex_images_for_clear()
svga: adjust line subpixel position for HWv8
vbo: simplify vbo_save_NotifyBegin()
vbo: whitespace/formatting fixes in vbo_save_loopback.c
vbo: rename target->index in loopback code
mesa: add some braces in _mesa_make_current()
st/mesa: remove unused st_framebuffer::Private field
mesa: finish implementing glPrimitiveRestartNV() for display lists
svga: fix breakage in create_backed_surface_view()
svga: fix PIPE_CAP_MAX_TEXTURE_BUFFER_SIZE value
svga: s/unsigned/enum tgsi_semantic/
svga: s/unsigned/enum tgsi_file_type/
svga: s/unsigned/enum tgsi_interpolate_mode/
svga: s/unsigned/enum tgsi_swizzle
svga: s/unsigned/enum tgsi_texture_type/
svga: fix texture swizzle writemasking
draw: s/unsigned/enum tgsi_semantic/
draw: whitespace, formatting fixes in draw_vs_exec.c
svga: move comment, declaration in svga_init_shader_key_common()
svga: whitespace, formatting fixes in svga_swtnl_state.c
svga: whitespace, formatting fixes in svga_swtnl_draw.c
svga: whitespace, formatting fixes in svga_swtnl_backend.c
svga: add comment about 'extra' constant locations
svga: add some const qualifiers
svga: whitespace clean-up in svga_winsys.h
mesa: fix GL_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION_EXT query
mesa: whitespace fixes in get.c
mesa: check API profile for GL_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION
mesa: include texture size in error messages
svga: only support 4x, 8x, 16x msaa
mesa: remove pointless assignments in init_teximage_fields_ms()
st/mesa: use proper resource target type in st_AllocTextureStorage()
Bruce Cherniak (11):
swr: move msaa resolve to generalized StoreTile
swr: Don't crash when encountering a VBO with stride = 0.
swr: set an explicit clear_rect if scissor is not enabled.
swr: conditionally validate vertex buffer state
swr: Remove need to allocate vertex buffer scratch space all in one go
swr: Minor cleanup of variable usage, no functional change.
swr: Limit memory held by defer deleted resources.
swr: Remove hard-coded constant and "todo" comment.
swr: Move environment config options into separate function.
swr: Add path to draw directly from client memory without copy.
st/osmesa: add osmesa framebuffer iface hash table per st manager
Chad Versace (47):
anv: Implement VK_KHX_external_memory_capabilities
anv: Drop 'x11' prefix from non-X11 WSI funcs
egl: Emit error when EGLSurface is lost
egl/android: Cancel any outstanding ANativeBuffer in surface destructor
egl/android: Mark surface as lost when dequeueBuffer fails
egl: Fix -Wint-to-pointer-cast
egl: Partially revert 23c86c74, fix eglMakeCurrent
i965/dri: Rewrite comment for intelCreateBuffer
i965/dri: Combine declaration and assignment in intelCreateBuffer
i965: Rename intel_create_renderbuffer
i965: Fix type of intel_update_image_buffers::format
i965: Move an 'i' declaration into its 'for' loop
i965: Add whitespace in intel_update_image_buffers()
i965: Document type of GLuint __DRIimage::format
i965: Move func to right comment block in brw_context.h
i965: Fix type of brw_context::render_target_format[]
egl/android: Align channel masks in HAL_PIXEL_FORMAT table
egl/android: Drop unused 'format' param in get_back_bo()
i965: Reject unsupported formats in glEGLImageTargetTexture2D()
i965: Fix return type of brw_isl_format_for_mesa_format() [v2]
i965: Fix return type of translate_tex_format()
i965: Remove bad assert on isl_format
i965: Cleanup in intel_create_winsys_renderbuffer()
i965/dri: Fix bad GL error in intel_create_winsys_renderbuffer()
i965: Replace 0 with ISL_FORMAT_UNSUPPORTED in format table (v2)
egl/android: Change order of EGLConfig generation (v2)
egl/android: Declare loop vars inside their loops (v2)
egl/drm: Move loop vars inside the loop
egl/x11: Declare loop vars inside the loop
egl/surfaceless: Move loop vars inside the loop
egl/wayland: Declare loop vars inside the loop
egl/dri2: Declare loop vars inside the loop
egl/android: Declare EGLConfig attrib array inside loop (v2)
egl/drm: Declare EGLConfig attrib array inside loop
egl/x11: Declare EGLConfig attrib array inside loop
egl: Rename 'count' in ${platform}_add_configs_for_visuals (v2)
i965: Rename some vague format members of brw_context
i965: Move brw_context format arrays to intel_screen
i965/dri: Add intel_screen param to intel_create_winsys_renderbuffer
i965: Fix -Wunused-variable in gen8_write_pma_stall_bits()
anv: Fix -Wswitch in anv_layout_to_aux_usage()
glapi: Fix -Wduplicate-decl-specifier due to double-const
mesa: Add _mesa_format_fallback_rgbx_to_rgba() [v2]
i965: Add a RGBX->RGBA fallback for glEGLImageTextureTarget2D()
i965/dri: Support R8G8B8A8 and R8G8B8X8 configs
mesa: Fix Android build
anv/image: Fix VK_IMAGE_CREATE_CUBE_COMPATIBLE_BIT
Chandu Babu N (2):
st/va: Fix leak in VAAPI subpictures
change va max_entrypoints
Charmaine Lee (25):
svga: Move setting the rendered_to flags to framebuffer emit time
svga: Move rendertarget view related fields to hw_clear state
svga: Set the surface dirty bit for the right surface view
svga: Update the backing resource only if needed
svga: cache the backing surface handle in the texture object
svga: fix format for screen target
svga: use the winsys interface to invalidate surface
svga: Removed the unused label 'done' in svga_validate_surface_view()
st/wgl: allow WGL_BIND_TO_TEXTURE_RGB_ARB for RGBA visuals
svga: fix pre-mature flushing of the command buffer
svga: pass bind_flags to surface create functions
svga: pass tobind_flags to svga_buffer_handle
svga: do not unconditionally enable streamout bind flag
svga: create buffer surfaces for incompatible bind flags
svga: fix the default devcap for SVGA3D_Z_D24S8_INT
svga: clean up format_cap_table
svga: re-validate sampler view at draw time if needed
svga: add mksstats for context creation
svga: fixed surface size to include array size
svga: add mksstats for surface view emulation
st/mesa: add a winsys buffers list in st_context
st/mesa: init winsys buffers list only if context creation succeeds
st/mesa: add destroy_drawable interface
svga: fix constant indices for texcoord scale factors and texture buffer size
st/mesa: create framebuffer iface hash table per st manager
Chih-Wei Huang (3):
Android: correct libz dependency
Android: use bionic pthread_barrier_* if possible
Android: fix spirv_info.c generation
Chris Wilson (15):
i965: Order write of query availablity with earlier writes
i965: Pass the map-mode along to intel_mipmap_tree_map_raw()
i965: Explicitly disallow tiled memcpy path on Gen4 with swizzling.
i965: Mark freshly allocate bo as idle
i965: Discard bo->map_count
i965: Drop global bufmgr lock from brw_bo_map_* functions.
i965: Remove clearing of bo->map_gtt after failure
i965: Use brw_bo_wait() for brw_bo_wait_rendering()
i965: Track when a bo is shared with an external client
i965: Disable access to CPU mmap for async access on non-LLC machines
i965: Resolve framebuffers before signaling the fence
i965: Use VALGRIND_MAKE_MEM_x in place of MALLOCLIKE/FREELIKE
i965: Rename batch->exec_objects to validation_list
i965/bufmgr: Add comments about GTT coherency issues.
i965/blit: Remember to include miptree buffer offset in relocs
Christian Gmeiner (6):
etnaviv: add L8A8_UNORM texture format
etnaviv: add rs-operations sw query
etnaviv: add support for extended texture formats
etnaviv: add support for swizzled texture formats
etnaviv: add R8G8 texture support
etnaviv: add support for snorm textures
Christian König (1):
winsys/amdgpu: align VA allocations to fragment size v2
Christoph Haag (2):
ac: fix build after LLVM 5.0 SVN r300718
gallium/hud: use double values for all graphs
Chuck Atkins (1):
configure.ac: Reduce zlib requirement from 1.2.8 to 1.2.3.
Connor Abbott (8):
spirv: fix OpBitcast when the src and dst bitsize are different (v3)
ac/nir: implement 64-bit packing and unpacking
ac/nir: fix 64-bit shifts
nir: don't segfault when printing variables with no name
nir/lower_io_to_temporaries: don't set compact on shadow vars
ac/nir: rewrite shared variable handling (v2)
nir: fix algebraic optimizations
ac/nir: fix lsb emission
Constantine Charlamov (3):
r600g: get rid of trailing whitespace
r600g: take into account offset to system inputs at tgsi_interp_egcm()
r600g: fix crash when file in R600_TRACE doesn't exist
Constantine Kharlamov (1):
r600g: update dirty_level_mask after the 1-st draw after FB change
Daniel Stone (30):
gbm/dri: Fix sign-extension in modifier query
i965: Use helper function for modifier -> tiling
i965: Set modifier for imported and duplicated images
i915: Fix build break with empty unreachable()
vulkan: Fix Wayland uninitialised registry
vulkan/wsi/wayland: Remove roundtrip when creating image
vulkan/wsi/wayland: Use per-display event queue
vulkan/wsi/wayland: Use proxy wrappers for swapchain
egl/wayland: Don't open-code roundtrip
egl/wayland: Use per-surface event queues
egl/wayland: Ensure we get a back buffer
egl/dri2: Avoid sign extension when building modifier
i965: Invert image modifier/tiling inference
i965: Move fallback size assignment out of bufmgr
i965: Improve same-buffer restriction for imports
util: Remove u_math from u_vector
egl/wayland: Fix whitespace damage
egl/wayland: Use MIN2 for wl_drm version
gbm: Axe buffer import format conversion table
gbm: Pull out FourCC <-> DRIimage format table
gbm: Check harder for supported formats
gbm: Remove is_planar_format dead code
egl/wayland: Make create_wl_buffer more generic
egl/wayland: Remove more surface specifics from create_wl_buffer
egl/wayland: Remove duplicate wl_buffer creation code
egl/wayland: Use linux-dmabuf interface for buffers
egl: Add MKDIR_GEN definition
egl/wayland: Ignore invalid modifiers
st/dri: Check get-handle return value in queryImage
st/dri2: Return invalid modifier when no driver support
Dave Airlie (175):
radv/ac: add initial pre-pass for shader info gathering
radv: start conditionalising vertex inputs. (v2)
radv: optimise compute shader grid size emission.
radv/ac: move needs_push_constants to shader info.
radv/ac: frag shader only needs ring offsets if sample positions enabled
radv/ac: mark used descriptor sets in shader info.
radv: start allocating user sgprs
radv/ac: add support for indirect access of descriptor sets.
radv: add support for 32 descriptor sets.
radv: drop debugging leftovers code in descriptor set patches.
radv/meta: add support for save/restore meta without vertex data.
radv/meta: add common shader vertex generation function
radv/meta: move fast clear to generate vertices in shader.
radv/meta: move depth decompress to using inline vertex data
radv/meta: avoid using vertex buffer for resolve shader.
radv/meta: reduce vertex buffer usage in clear shaders
radv/meta: reduce vertex buffer in blit2d.
radv/meta: generate position for blit shaders.
radv/meta: refactor out some common shaders.
radv/ac: use tex_lz if we can.
radv/ac: copy llvm machine feature flags from radeonsi.
radv/ac: move point coord after layer/viewport.
radv/ac: overhaul vs output/ps input routing
radv/ac: setup mrt exports then export them in one go. (v2)
radv: fix regression in descriptor set freeing.
radeonsi/ac: move vertex export remove to common code.
radv/ac: eliminate unused vertex shader outputs. (v2)
radeon/ac: fix build on llvm 3.8.1
radeon/ac: remove assert causing regression
radv: provide a helper for comparing an image extents.
radv: move some image info into a separate struct.
radv: remove some members of radeon surface.
radv/image: drop blk_d
radv: drop unused surface level members.
radv/image: resize some surface members.
radv: minor pahole related improvements.
radv/wsi: report presentation error per image request
radv: fix stencil only clears.
radv: flush more stages when semaphore are waiting.
radv: enable lower_sub to fix loop unrolling.
radv: flush f32->f16 conversion denormals to zero. (v2)
radv/ac: canonicalize the output for 32-bit float min/max.
radeonsi: constify a bunch of the perfcounter structs.
radv: don't advertise transfer props unless we can do anything else
radv: enable POLARIS12 support.
radv: apply the tess+GS hang workaround to Polaris12 as well
radv: set PERF_MOD in sample state like radeonsi.
radv/meta: add srgb conversion to end of resolve shader.
radv/meta: split out core part of resolve shader
radv/resolve: split resolve emission out for compute
radv: add subpass resolve compute path
radv/meta: add resolve pass using fragment/vertex shaders
radv/meta: select resolve paths
radv: drop resolve hack workarounds
radv: set base/ranges for push constant loads.
radv/meta: use novertex save path for resolve pass.
radv/meta: move clear color to using push constants
radv/meta: move blit2d to using push constants
radv/meta: port blit to using push constants
radv/meta: cleanup some unused code path
radv: fix regression in blit2d push constant change.
radv: handle fragment shader srgb resolve pass better
Revert "amd/common: set vcn dec as hw decode as well"
Revert "amd/common: add vcn dec ip info query"
glsl/lower_int64: only set progress when something is lowered.
r600/eg: add support for tracing IBs after a hang.
radv: factor out si_emit_wait_fence code.
radv: factor out eop event writing code. (v2)
radv: bump some base addresses to 64-bits.
radv: realign cp dma code with radeonsi
radv/formats: reverse how the image format properties KHR2 is handled
radv: add external memory support.
ac: add new helper function to add a integer target dependent function attr.
ac/nir: set workgroup size attribute to correct value.
radv: prepare fmask surface creation
r600: fix incorrect and missing bit field in register headers.
r600/sb: fix typo in field definitions
r600: add missing RAT registers and operations.
r600: add register info for atomic counters.
r600: document some of the missing shader constants.
r600: refactor out some compressed resource state code.
radv: use radv_foreach_stage in a couple of places.
radv: refactor depth/stencil state setup
radv: refactor color surface init before gfx9.
radv: refactor some texture descriptor state.
radv: split metadata struct into legacy/gfx9 parts.
radv: remove doubled up prototype.
radv: add a comment from radeonsi before cp dma function.
radv: fix typo in comment.
radv: set offchip hs param like radeonsi.
radv: add some misc gfx9 pieces.
radv: add GFX9 support for color surfaces.
radv: add gfx9 depth/stencil surface support.
radv: add gfx9 cp dma support.
radv: don't setup raster_config on gfx9.
radv: add GFX9 to initialisation cmd buffer.
radv: add texture descriptor/fmask/cmask support for GFX9
radv: add GFX9 cache flushing support.
radv: add rb+ support for GFX9
radv: add IA_MULTI_VGT_PARAM support for GFX9.
radv: add some GFX9 specific events.
radv: misc GFX9 changes.
radv: turn off geom/tess for gfx9.
radv: enable GFX9 on radv
radv: fix write event eop on vega.
radv: expose integrated device type for APUs.
radv/gfx9: use correct register setting for uconfig regs
radv: rename and make global some functions.
radv: move calculating the vertex sgpr to the pipeline.
radv: move lots of index related things into the bind.
radv: move chip_class extraction down further.
ac/nir: consolidate setting userdata location
ac/nir: assign argument param pointers in one place.
ac/nir: move gpr counting inside argument handling.
st_glsl_to_tgsi: bump index back up to 32-bit
st_glsl_to_tgsi: rewrite rename registers to use array fully.
st_glsl_to_tgsi: replace variables tracking list with a hash table
st_glsl_to_tgsi: cleanup variable storage search.
radv: introduce perf test env var and allow to enable chaining
radv: set fmask state to all 0s when no fmask. (v2)
glsl/lower_distance: only set max_array_access for 1D clip dist arrays
u_dynarray: fix coverity warning about ignoring return value from reralloc
st_glsl_to_tgsi: don't try and pass 32-bit values to get_deref_offsets
radv: return correct error on invalid handle from vkAllocateMemory
radv: move assert down in radv_bind_descriptor_set
ac/gpu: drop duplicated code line.
st/mesa: fix assert to be simpler
radv: set use_prim_id for tess shaders correctly.
radv: set prim_id for geometry shaders
radv: compile fragment shader first.
radv: handle primitive id input into fragment shader with no geom shader
radv/meta: don't need vertex info for resolve shader.
r600: route indirect address register correctly for vtx fetches.
r600: add support for vertex fetches via texture cache
r600: for memory instructions dump index gpr for read indirects also.
r600: don't merge GDS into VTX
r600/asm: add support for other GDS operations.
ac/nir: remove last remnants of v16i8
radv: enable Int64 capability (v2)
radv: add support for cmd predication.
ac/radv: change api to create target machine
ac/llvm: create features list using snprintf.
ac/llvm: set xnack like radeonsi does.
radv: enable sisched toggle in perftest flags.
radv: don't overallocate depth/stencil formats
radv: add support for using addrlib max alignment.
radv/ac: drop setting xnack
radv: allow clear merging for depth/stencil with no care stencil
radv: set cb base tile swizzles for MRT speedups (v4)
amd/addrlib: fix typo in api name.
radv/clear: add r32g32b32a32 fast clear support (v2)
radv: predicate cmask eliminate when using DCC.
radv: don't shadow meta_va.
radv: introduce some wrapper in cs code to make porting off libdrm_amdgpu easier.
radv: port to new libdrm API.
ac/gpu: add code to detect if kernel supports sync objects.
radv/winsys: add syncobj hooks
radv: initial support for shared semaphores (v2)
radv: reset non-syncobj semaphore context after wait.
radv: for external memory imports close the fd on import success
radv: check enabled device features.
radv: fix non-0 based layer clears.
radv: don't set dedicated bit for buffer external memory.
radv: enable sample shading
virgl: encode index buffer offset.
radv: fix buffer views on SI/CIK.
radv: only report external semaphore info for opaque fd.
radv/ac: realign SI workaround with radeonsi.
radv/ac: port SI TC L1 write corruption fix.
virgl: drop precise modifier.
radv: for stencil only set Z tile mode index to same value
radeon/ac: use ds_swizzle for derivs on si/cik.
intel/vec4/gs: reset nr_pull_param if DUAL_INSTANCED compile failed.
radv: handle 10-bit format clamping workaround.
Revert "st_glsl_to_tgsi: rewrite rename registers to use array fully."
Deepak Rawat (1):
winsys/svga: fix error path when kernel is not able to create surface
Dylan Baker (1):
radv: rebase radv_entrypoints_gen.py on anv_entrypoints_gen.py
Edmondo Tommasina (3):
drirc: Add glsl_correct_derivatives_after_discard for The Witcher 2
drirc: whitelist glthread for Civilization 5
drirc: whitelist glthread for The Witcher 2
Eero Tamminen (1):
i965: Fix anisotropic filtering for mag filter
Elie TOURNIER (1):
egl: add gitignore
Elie Tournier (1):
glsl: Fix indent in dump code
Emil Velikov (180):
docs: add 17.2.0-devel release notes template, bump version
docs: add release notes for 17.0.4
docs: add sha256 checksums for 17.0.4
docs: add news item and link release notes for 17.0.4
configure.ac: print deprecation warning as needed
configure.ac: manually expand PKG_CHECK_VAR
configure.ac: check require_basic_egl only if egl enabled
winsys/sw/dri: don't use GNU void pointer arithmetic
st/mesa: automake: honour the vdpau header install location
st/clover: add space between < and ::
mesa/glthread: correctly compare thread handles
ac: include all sources in the tarball
ac: automake: sort sources list alphabetically
travis: explicitly LD_LIBRARY_PATH the local libraries
travis: enable apt cache
travis: automatically manage ccache caching
travis: remove unused -dev packages
travis: rework "if test" blocks in the script section
travis: split out matrix from env
travis: add separate "scons" and "scons llvm" targets
travis: add "scons swr" to the build matrix
travis: add "make swr" to the build matrix
travis: split the make target to three separate ones
travis: model scons check target like the make one
travis: add Gallium state-tracker targets
travis: enable wayland support
travis: bump MAKEFLAGS to -j4
docs: add release calendar page and references to it
egl: polish dri2_to_egl_attribute_map[]
egl: initialise dummy_thread via _eglInitThreadInfo
c11/threads: rework Windows thrd_current() comment
dri_interface.h: define __DRI_ATTRIB_MAX
eg: explicitly size dri2_to_egl_attribute_map[]
anv: fix anv_gem_mmap comment to not mention NULL
gallium/dri: always link against shared glapi
mesa/dri: always link against shared glapi
gallium/dri: remove unneeded HAVE_SHARED_GLAPI guard
mesa/dri: remove unneeded HAVE_SHARED_GLAPI guard
mapi/es*api: remove unneeded HAVE_SHARED_GLAPI guard
glx: glX_proto_send.py: use correct compile guard GLX_INDIRECT_RENDERING
drivers/x11: remove unneeded GLX_SHARED_GLAPI define
targets/libgl-xlib: remove unneeded GLX_SHARED_GLAPI define
glx: automake: scons: remove unneeded GLX_SHARED_GLAPI define
glx/windows: remove empty variable SHARED_GLAPI_CFLAGS
glx/apple: remove empty variable SHARED_GLAPI_CFLAGS
gl_table.py: always regenerate the complete struct _glapi_table
mesa/tests: remove no longer needed HAVE_SHARED_GLAPI define
mapi: replace mapi_table abstraction
mapy_abi.py: remove dead output_for_app generator
mapi_abi.py: remove no longer used --mode option
egl: remove unused _eglSetLogLevel()
egl: fold Android logger into main/
egl: remove no longer needed logger infra
egl: remove unneeded else statement in _eglInitLogger
egl: remove suprous header eglcompiler.h
egl: drop unneeded sentinel from level_strings[]
egl: use designated initializers
docs: Update 17.1.0 release notes
docs: add sha256 checksums for 17.1.0
docs: add news item and link release notes for 17.1.0
egl: simplify the Android logger
anv: document that anv_gem_mmap returns MAP_FAILED on error
radeon: automake: remove unneeded elf Cflags/Libs
virgl: remove unused draw include
docs: remove released versions from the calendar
docs/releasing: don't forget to update the calendar
configure: remove unneeded bits around libunwind handling
egl: add g_egldispatchstubs.h to the release tarball
configure: move platform handling further up
configure: rename remaining HAVE_EGL_PLATFORM_* guards
configure: update remaining --with-egl-platforms references
configure: loosen --with-platforms heuristics
configure: enable the surfaceless platform by default
configure: set HAVE_foo_PLATFORM as applicable
configure: error out when building GLX w/o the X11 platform
configure: check once for DRI3 dependencies
vulkan: automake: remove unused VULKAN_LIB_DEPS variable
loader: build libloader_dri3_helper.la only with HAVE_PLATFORM_X11
configure: error out when building X11 Vulkan without DRI3
auxiliary/vl: use vl_*_screen_create stubs when building w/o platform
st/va: move variable declaration to where its used
st/va: fix misplaced closing bracket
st/va: check if vl_*_screen_create has failed only once
st/omx: remove unused drm_driver.h includes
st/omx: remove unneeded X11 include
st/omx: fix building against X11-less setups
gallium/targets: link against XCB only as needed
st/omx: add DRI3 support
st/xvmc: add DRI3 support
configure: error out if building VA w/o supported platform
configure: error out if building OMX w/o supported platform
configure: error out if building VDPAU w/o supported platform
configure: error out if building XVMC w/o supported platform
configure: trivial whitespace cleanup
travis: remove workarounds for the Vulkan target
automake: add SWR LLVM gen_builder.hpp workaround
configure.ac: add xcb-fixes to the XCB DRI3 list
configure.ac: s/xcb-fixes/xcb-xfixes/
docs: add release notes for 17.1.1
docs: add sha256 checksums for 17.1.1
docs: update calendar, add news item and link release notes for 17.1.1
anv: automake: list shared libraries after the static ones
radv: automake: list shared libraries after the static ones
gbm: remove unused gbm_dri_device::loader
gbm: remove "struct gbm_drm_bo" abstraction
gbm: move gbm_drm_device::driver_name to gbm_dri_device
gbm: remove unneeded gbm_drm_device abstraction
egl: split out a dri2_display_destroy() helper
egl/android: make use of dri2_display_destroy() helper
egl/drm: make use of the dri2_display_destroy() helper
egl/surfaceless: make use of the dri2_display_destroy() helper
egl/wayland: make use of the dri2_display_destroy() helper
egl/x11: make use of the dri2_display_destroy() helper
egl: refactor dri2_create_screen() into three separate functions
egl/drm: use dri2_setup_extensions() over the extensions provided by GBM
gbm: manage only the required set of DRI extensions
egl: error out on eglCreateWaylandBufferFromImageWL
egl/wayland: don't modify the vtbl if an extension is not available
egl: annotate dri2_egl_display_vtbl as const data
egl/drm: check for dri2_dpy->flush before using the flush extension
egl/drm: flatten codeflow
egl/x11: check for dri2_dpy->flush before using the flush extension
egl/x11: flatten codeflow
egl/android: use the image_driver alongside the image_loader
egl/surfaceless: use the image_driver for image_loader
egl/wayland: select the format based on the interface used
egl/wayland: set the resize_callback if the flush extension is available
egl/wayland: use the image_driver alongside the image_loader
docs: add release notes for 17.0.7
docs: add sha256 checksums for 17.0.7
docs: update calendar, add news item and link release notes for 17.0.7
radeon: remove out of date LLVM_REVISION.txt
i965: gen4_blorp_exec.h to the sources list
configure.ac: add -pthread to PTHREAD_LIBS
configure.ac: remove manual AC_SUBST for pthread-stubs
ac: remove amdgpu.h dependency
Revert "amd/common: add missing libdrm include path"
ac: resolve conflicts introduced with "ac: remove amdgpu.h dependency"
r600, radeon: move radeon_shader_binary_{init,clean} back to radeon
radeonsi: include ac_binary.h for struct ac_shader_binary
r600: include libelf headers only as needed
docs: add release notes for 17.1.3
docs: add sha256 checksums for 17.1.3
docs: update calendar, add news item and link release notes for 17.1.3
egl: drop _eglInitImage() return type
egl: fold _eglError() + return EGL_FALSE
egl: move eglCreateDRMImageMESA's malloc later
egl: call _eglError within _eglParseImageAttribList
egl: split _eglParseImageAttribList into per extension functions
egl: simplify EXT_image_dma_buf_import_modifiers attr parsing
egl: polish EXT_image_dma_buf_import attr parsing
egl: add width/height as EXT_image_dma_buf_import attrs
egl: check for extensions' presence during attr parsing
egl: don't set modifier if no modifiers are available
egl: enhance KHR_gl_image extensions checks
egl: set KHR_gl_texture_3D_image only when the requirements are met.
docs: update HTTP -> HTTPS reference to reflect reality
travis: lower SWR requirement to GCC 4.8, aka std=c++11
i965: use strtol to convert the integer deviceID override
anv: advertise v6 of the wayland surface extension
radv: advertise v6 of the wayland surface extension
i965: remove local vblank_mode option
i915: remove local vblank_mode option
radeon: remove local vblank_mode option
swrast: add dri2ConfigQueryExtension to the correct extension list
st/dri: list __DRI2_FENCE extension only where needed
swr: remove unneeded fallback strcasecmp define
egl: propagate EGL_BAD_ATTRIBUTE during EGLImage attr parsing
dri/common: use designated initializers for OptConfElems
configure.ac: rework wayland-protocols handling
swr: don't forget to link AVX/AVX2 against pthreads
swr: don't forget to link KNL/SKX against pthreads
swr: use the correct variable for no undefined symbols
wayland-egl: update the SHA1 of the commit introducing v3
swrast: add dri2ConfigQueryExtension to the correct extension list
egl: guard wayland header dep. tracking behind HAVE_PLATFORM_WAYLAND
intel/blorp: ship blorp_genX_exec.h within the tarball
Update version to 17.2.0-rc1
Update version to 17.2.0-rc2
Update version to 17.2.0-rc3
Eric Anholt (36):
vc4: Enable V3D 2.6.
vc4: Only build the NEON code on arm32.
nir: Pick just the channels we want for bitmap and drawpixels lowering.
gallium: Enable ARM NEON CPU detection.
vc4: Use a wrapper file to set VC4_BUILD_NEON instead of CFLAGS.
vc4: Use runtime CPU detection for whether NEON is available.
glsl: Restrict functions to not return arrays or SOAs in GLSL 1.00.
glsl: Ban #undefining __LINE__ and friends on GLES2.
glsl: Restrict func redeclarations (not just redefinitions) on GLSL 1.00.
glsl: Don't allow redefining builtin functions on GLSL 1.00.
renderonly: Initialize fields of struct winsys_handle.
vc4: Drop the u_resource_vtbl no-op layer.
vc4: Drop pointless indirections around BO import/export.
vc4: Don't allocate new BOs to avoid synchronization when they're shared.
etnaviv: Only use renderonly_get_handle for GEM handles.
gallium: Add renderonly-based support for pl111+vc4.
vc4: Allow VBOs to be mapped during execution.
vc4: Clean up release build warnings using MAYBE_UNUSED.
vc4: Introduce XML-based packet header generation like Intel's.
vc4: Start using the pack header.
vc4: Simplify pack header usage
vc4: Convert the driver to emitting the shader record using pack macros.
vc4: Move rasterizer state packing to CSO creation time.
vc4: Replace a couple of magic numbers with #define usage.
vc4: Start using XML unpack functions in CL dump.
vc4: Remove a stale comment.
intel: Move the DRM uapi headers to a non-Intel location.
vc4: Switch back to using a local copy of vc4_drm.h.
vc4: Make the miptree debug code available under VC4_DEBUG=surf
vc4: Use vc4_setup_slices for resource import
vc4: Set shareable BOs as T tiled if possible
broadcom/genxml: Support unpacking fixed-point fractional values.
broadcom/genxml: Use the same "gen" attr for HW version as Intel does.
broadcom: Introduce a header for talking about chip revisions.
broadcom/vc4: Add more packets to the v2.1 XML.
broadcom/vc4: Prefer blit via rendering to the software fallback.
Eric Engestrom (28):
configure: print LDFLAGS alongside CFLAGS & co.
egl: avoid dereferencing a null display
configure.ac: rephrase 'GLX w/o X11' error message
amd/common: add missing libdrm include path
egl/drm: remove temporary fd variable
configure.ac: simplify --enable-libunwind=auto check
tree-wide: remove trailing backslash
egl: improve dri2_fallback_swap_buffers_with_damage()
egl/display: only detect the platform once
egl/display: make platform detection thread-safe
mesa/format_info: use designated initialiser list
egl: turn boolean `int`s into `bool`s
egl: turn one more boolean `int` into a `bool`
egl: simplify dri_config conditionals
egl: properly count configs
travis: add missing libs: xdamage + xfixes
radv: use Mesa's u_atomic.h header
anv: use Mesa's u_atomic.h header
mesa: add missing include
build systems: uniformize git_sha1.h generation
scons: wait on subprocess' completion
build systems: move git_sha1_gen.sh to bin/
scons: split out check_header() helper
configure: only install khrplatform.h if needed
vulkan/util: fix typo in comment
configure.ac: fix whitespace
gbm: fix typo
egl: fix line continuation
Eric Le Bihan (1):
Fix khrplatform.h not installed if EGL is disabled.
Fabio Estevam (1):
etnaviv: etnaviv_fence: Simplify the return code logic
Francisco Jerez (4):
intel/fs: Use regs_written() in spilling cost heuristic for improved accuracy.
intel/fs: Take into account amount of data read in spilling cost heuristic.
i965/fs: Handle explicit flag destinations in flags_written()
i965/fs: Handle explicit flag sources in flags_read()
Frank Richter (1):
gallium/wgl: Allow context creation even if SetPixelFormat() wasn't called
Fredrik Höglund (1):
radv/meta: fix restoring a push descriptor set
George Kyriazis (6):
swr/rast: support polygonmode point
swr: fix polygonmode for front==back
swr: Include definition of missing function
swr: invalidate attachment on transition change
swr/rast: make SWR_VISIBLE attribute work for windows
swr: fix transform feedback logic
Glenn Kennard (3):
r600g: Add scratch ring register defines
r600g: Add instruction encoding defines for MEM_RD
r600g: Add defines for per-shader engine settings
Grazvydas Ignotas (15):
radv: enable timestampComputeAndGraphics
Reply to: