summaryrefslogtreecommitdiffstats
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
* glsl: make sampler/image types as 64-bitSamuel Pitoiset2017-05-062-2/+8
| | | | | | | | | | | | | | | | | | | | | | | | The ARB_bindless_texture spec says: "Samplers are represented using 64-bit integer handles." and, "Images are represented using 64-bit integer handles." It seems simpler to always consider sampler and image types as 64-bit unsigned integer. This introduces a temporary workaround in _mesa_get_uniform() because at this point no flag are used to distinguish between bound and bindless samplers. This is going to be removed in a separate series. This avoids breaking arb_shader_image_load_store-state. v3: - update the comment slightly Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]> Reviewed-by: Timothy Arceri <[email protected]>
* glsl: add ARB_bindless_texture enableSamuel Pitoiset2017-05-063-0/+9
| | | | | | | | | This also adds the extension to the standalone GLSL compiler. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Ilia Mirkin <[email protected]> Reviewed-by: Timothy Arceri <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]>
* mesa: add ARB_bindless_texture to the extensions listSamuel Pitoiset2017-05-062-0/+2
| | | | | | | | This is required for the following GLSL bits. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]> Reviewed-by: Timothy Arceri <[email protected]>
* radv/meta: fix restoring a push descriptor setFredrik Höglund2017-05-062-2/+7
| | | | | | | | | | | radv_bind_descriptor_set cannot be used to bind a push descriptor set since a push descriptor set does not have a buffer list. However, there is no need to add the buffers again when restoring a set, so this fix is also an optimization. Cc: "17.1" <[email protected]> Signed-off-by: Fredrik Höglund <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* anv/allocator: Only write to _vg_ptr if we have valgrindJason Ekstrand2017-05-051-1/+1
| | | | | | | This fixes the build when not building against valgrind headers. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100945 Reviewed-by: Chad Versace <[email protected]>
* i915: Fix build break with empty unreachable()Daniel Stone2017-05-051-4/+2
| | | | | | | | Actually put something in unreachable(), so as not to break the build on a Friday evening. Signed-off-by: Daniel Stone <[email protected]> Reported-by: Mark Janes <[email protected]>
* radeonsi: apply the tess+GS hang workaround to Polaris12 as wellMarek Olšák2017-05-051-1/+2
| | | | | | Cc: 17.1 <[email protected]> Reviewed-by: Alex Deucher <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]>
* i965: Set modifier for imported and duplicated imagesDaniel Stone2017-05-051-0/+20
| | | | | | | | | | | | | | | | | When a buffer is being created from FD or GEM flink import, the current API makes no provision for passing modifier information along with this. Set the modifier for such images to DRM_FORMAT_MOD_INVALID. Also preserve the modifier when duplicating an image, as will be done by GBM when importing from a wl_buffer. This doubly tripped up Wayland, as the images would first have been created (as wl_buffers) with a 0 modifier, and then lost what modifier they would've had when being duplicated into gbm_bos. Fixes: d78a36ea624 ("i965/dri: Handle the linear fb modifier") Signed-off-by: Daniel Stone <[email protected]> Reviewed-by: Emil Velikov <[email protected]>
* i965: Use helper function for modifier -> tilingDaniel Stone2017-05-051-17/+34
| | | | | | | | Use a helper function and struct to convert between a modifier and tiling mode, so we can use it later for a tiling -> modifier lookup. Signed-off-by: Daniel Stone <[email protected]> Reviewed-by: Emil Velikov <[email protected]>
* radeonsi: fix build with GCC 4.8Samuel Pitoiset2017-05-051-1/+1
| | | | | | | Fixes: 7088b655e8 ("radeonsi: constify a bunch of the perfcounter structs.") Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100937 Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* st/glsl_to_tgsi: fix renumber_registers() in presence of dead codeSamuel Pitoiset2017-05-051-5/+34
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The TGSI DCE pass doesn't eliminate dead assignments like MOV TEMP[0], TEMP[1] in presence of loops because it assumes that the visitor doesn't emit dead code. This assumption is actually wrong and this situation happens. However, it appears that the merge_registers() pass accidentally takes care of this for some weird reasons. But since this pass has been disabled for RadeonSI and Nouveau, the renumber_registers() pass which is called *after*, can't do its job correctly. This is because it assumes that no dead code is present. But if there is still a dead assignment, it might re-use the TEMP register id incorrectly and emits wrong code. This patches fixes the issue by recording writes instead of reads, and this has the advantage to be faster. This should fix Unigine Heaven on RadeonSI and Nouveau. shader-db results with RadeonSI: 47109 shaders in 29632 tests Totals: SGPRS: 1923308 -> 1923316 (0.00 %) VGPRS: 1133843 -> 1133847 (0.00 %) Spilled SGPRs: 2516 -> 2518 (0.08 %) Spilled VGPRs: 65 -> 65 (0.00 %) Private memory VGPRs: 1184 -> 1184 (0.00 %) Scratch size: 1308 -> 1308 (0.00 %) dwords per thread Code Size: 60095968 -> 60096256 (0.00 %) bytes LDS: 1077 -> 1077 (0.00 %) blocks Max Waves: 431889 -> 431889 (0.00 %) Wait states: 0 -> 0 (0.00 %) It's still interesting to disable the merge_registers() pass. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]>
* anv/query: handle more cases of 'out of host memory'Iago Toral Quiroga2017-05-051-0/+10
| | | | Reviewed-by: Samuel Iglesias Gonsálvez <[email protected]>
* egl/android: Set EGLSurface.Lost to EGL_TRUE/EGL_FALSENicolas Boichat2017-05-041-2/+2
| | | | | | | | | Lost is an EGLBoolean, so we should assign it to EGL_TRUE/EGL_FALSE, not true/false. Fixes: e5eace58684 ("egl/android: Mark surface as lost when dequeueBuffer fails") Fixes: 0212db35040 ("egl/android: Cancel any outstanding ANativeBuffer in surface destructor") Reviewed-by: Chad Versace <[email protected]>
* anv/allocator: Improve block pool growing assertsJason Ekstrand2017-05-041-6/+5
| | | | Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv: Drop the instruction pool block sizeJason Ekstrand2017-05-041-2/+1
| | | | | | | Now that we can allocate states larger than the block size, we no longer need a block size of 1MB which can be rather wasteful. Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv/allocator: Add support for large stream allocationsJason Ekstrand2017-05-041-4/+7
| | | | Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv/allocator: Allow state pools to allocate large statesJason Ekstrand2017-05-041-0/+69
| | | | | | | | | | | | | | | | | | | | | | Previously, the maximum size of a state that could be allocated from a state pool was a block. However, this has caused us various issues particularly with shaders which are potentially very large. We've also hit issues with render passes with a large number of attachments when we go to allocate the block of surface state. This effectively removes the restriction on the maximum size of a single state. (There's still a limit of 1MB imposed by a fixed-length bucket array.) For states larger than the block size, we just grab a large block off of the block pool rather than sub-allocating. When we go to allocate some chunk of state and the current bucket does not have state, we try to pull a chunk from some larger bucket and split it up. This should improve memory usage if a client occasionally allocates a large block of state. This commit is inspired by some similar work done by Juan A. Suarez Romero <[email protected]>. Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv/allocator: Support pushing multiple blocks onto a free list at onceJason Ekstrand2017-05-041-3/+16
| | | | Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv/allocator: Add helpers for dealing with bucket sizesJason Ekstrand2017-05-041-10/+20
| | | | Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv/allocator: Add the capability to allocate blocks of different sizesJason Ekstrand2017-05-042-15/+13
| | | | Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv/allocator: Rework a commentJason Ekstrand2017-05-041-4/+5
| | | | | | This commit just fixes up the English a bit and re-flows the comment. Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv/allocator: Tweak the block pool growing algorithmJason Ekstrand2017-05-041-10/+18
| | | | | | | | | The old algorithm worked fine assuming a constant block size. We're about to break that assumption so we need an algorithm that's a bit more robust against suddenly growing by a huge amount compared to the currently allocated quantity of memory. Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv/allocator: Embed the block_pool in the state_poolJason Ekstrand2017-05-0411-86/+66
| | | | | | | Now that the state stream is allocating off of the state pool, there's no reason why we need the block pool to be separate. Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv/allocator: Get rid of the ability to free blocksJason Ekstrand2017-05-042-34/+2
| | | | | | | Now that everything is going through the state pools, the block pool no longer needs to be able to handle re-use. Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv: Allocate binding table blocks through the state poolJason Ekstrand2017-05-042-27/+22
| | | | Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv/allocator: Add support for "back" allocations to state_poolJason Ekstrand2017-05-042-2/+33
| | | | Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv/allocator: Drop the block_size field from block_poolJason Ekstrand2017-05-049-50/+74
| | | | | | | | | | | | | Since the state_stream is now pulling from a state_pool, the only thing pulling directly off the block pool is the state pool so we can just move the block_size there. The one exception is when we allocate binding tables but we can just reference the state pool there as well. The only functional change here is that we no longer grow the block pool immediately upon creation so no BO gets allocated until our first state allocation. Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv/allocator: Pull the userptr part of block_pool_grow into a helperJason Ekstrand2017-05-041-91/+104
| | | | Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv/allocator: Roll fixed_size_state_pool into state_poolJason Ekstrand2017-05-041-50/+35
| | | | | | | The helper functions aren't really gaining us as much as they claim and are actually about to be in the way. Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv/allocator: Remove the state_size field from fixed_size_state_poolJason Ekstrand2017-05-042-12/+10
| | | | Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv: Get rid of a bunch of uses of size_tJason Ekstrand2017-05-044-14/+14
| | | | | | | | | | We should only use size_t when referring to sizes of bits of CPU memory. Anything on the GPU or just a regular array length should be a type that has the same size on both 32 and 64-bit architectures. For state objects, we use a uint32_t because we'll never allocate a piece of driver-internal GPU state larger than 2GB (more like 16KB). Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv/allocator: Convert the state stream to pull from a state poolJason Ekstrand2017-05-044-51/+55
| | | | Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv/allocator: Return a null state for zero-size allocationsJason Ekstrand2017-05-042-0/+11
| | | | Reviewed-by: Juan A. Suarez Romero <[email protected]>
* anv/allocator: Add no-valgrind versions of state_pool_alloc/freeJason Ekstrand2017-05-041-5/+19
| | | | Reviewed-by: Juan A. Suarez Romero <[email protected]>
* radv: enable POLARIS12 support.Dave Airlie2017-05-054-0/+8
| | | | | | | | | | | This just adds the chip in the right places. We don't set the partial_vs_wave workaround, as radeonsi doesn't, but have to confirm it's not required. Reviewed-by: Bas Nieuwenhuizen <[email protected]> Cc: "17.1" <[email protected]> Signed-off-by: Dave Airlie <[email protected]>
* egl/android: Mark surface as lost when dequeueBuffer failsChad Versace2017-05-041-0/+10
| | | | | | | | | | | | | | | This ensures that future calls to eglSwapBuffers and eglMakeCurrent emit an error. This patch is part of a series for fixing android.hardware.camera2.cts.RobustnessTest#testAbandonRepeatingRequestSurface on Chrome OS x86 devices. Cc: [email protected] Cc: Tomasz Figa <[email protected]> Cc: Tapani Pälli <[email protected]> Reviewed-by: Nicolas Boichat <[email protected]> Reviewed-by: Emil Velikov <[email protected]>
* egl/android: Cancel any outstanding ANativeBuffer in surface destructorChad Versace2017-05-041-4/+9
| | | | | | | | | | | | | | | | | | | That is, call ANativeWindow::cancelBuffer in droid_destroy_surface(). This should prevent application deadlock when the app destroys the EGLSurface after EGL has acquired a buffer from SurfaceFlinger (ANativeWindow::dequeueBuffer) but before EGL has released it (ANativeWindow::enqueueBuffer). This patch is part of a series for fixing android.hardware.camera2.cts.RobustnessTest#testAbandonRepeatingRequestSurface on Chrome OS x86 devices. Cc: [email protected] Cc: Tomasz Figa <[email protected]> Cc: Tapani Pälli <[email protected]> Reviewed-by: Nicolas Boichat <[email protected]> Reviewed-by: Emil Velikov <[email protected]>
* egl: Emit error when EGLSurface is lostChad Versace2017-05-043-0/+42
| | | | | | | | | | | | | | | | | | | | | | Add a new bool, _EGLSurface::Lost, and check it in eglMakeCurrent and eglSwapBuffers. The EGL 1.5 spec says that those functions emit errors when the native surface is no longer valid. This patch just updates core EGL. No driver sets _EGLSurface::Lost yet. I discovered that Mesa failed to detect lost surfaces while debugging an Android CTS camera test, android.hardware.camera2.cts.RobustnessTest#testAbandonRepeatingRequestSurface. This patch doesn't fix the test though, though, because the test expects EGL_BAD_SURFACE when the surface becomes lost, and this patch actually complies with the EGL spec. If I interpreted the EGL spec correctly, EGL_BAD_NATIVE_WINDOW or EGL_BAD_CURRENT_SURFACE is the correct error. Cc: [email protected] Cc: Tomasz Figa <[email protected]> Cc: Tapani Pälli <[email protected]> Reviewed-by: Nicolas Boichat <[email protected]> Reviewed-by: Emil Velikov <[email protected]>
* winsys/amdgpu: fix Polaris12 (RX 550) breakageMarek Olšák2017-05-051-0/+1
| | | | | | | reported by Greg White. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100892 Cc: 17.1 <[email protected]>
* anv: Simplify Cherryview line handling.Kenneth Graunke2017-05-041-49/+16
| | | | | | | | | We can just use the new CHVLineWidth field rather than an entirely different generation's packing function. v2: Inline the function (requested by Jason) Reviewed-by: Jason Ekstrand <[email protected]>
* i965: Fix line width on Cherryview.Kenneth Graunke2017-05-042-0/+8
| | | | | | | | We just add another field to gen8.xml for the Cherryview line width, rather than trying to replicate the gymnastics done in the Vulkan driver to use gen9 SF pack functions. Reviewed-by: Jason Ekstrand <[email protected]>
* radeonsi/gfx9: allow the scratch buffer in HS and GSMarek Olšák2017-05-051-10/+0
| | | | | | It works now. Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: prevent race conditions when doing scratch patchingMarek Olšák2017-05-051-2/+30
| | | | Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: separate scratch state patching code into its own functionMarek Olšák2017-05-051-46/+55
| | | | | | | Picked from a different branch. When we stop using the scratch patching, this function will not be called. Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi/gfx9: also apply scratch relocations to the 1st shader of merged ↵Marek Olšák2017-05-051-0/+3
| | | | | | shaders Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi/gfx9: set correct LLVM calling conventions for merged shadersMarek Olšák2017-05-052-2/+18
| | | | | | for scratch support Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: remove unused parameters from si_shader_apply_scratch_relocsMarek Olšák2017-05-054-10/+6
| | | | Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: inline si_llvm_shader_type into si_llvm_create_funcMarek Olšák2017-05-053-33/+22
| | | | Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: don't use util_memcpy_cpu_to_le32 for shader uploadsMarek Olšák2017-05-051-7/+8
| | | | | | at least I think this is correct. Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: make si_compile_llvm staticMarek Olšák2017-05-052-16/+8
| | | | Reviewed-by: Nicolai Hähnle <[email protected]>