summaryrefslogtreecommitdiffstats
path: root/src/gallium/drivers/radeonsi
Commit message (Collapse)AuthorAgeFilesLines
* radeonsi: don't crash on compute shader compile failureMarek Olšák2017-03-241-1/+5
| | | | | Reviewed-by: Samuel Pitoiset <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: don't hang on shader compile failureMarek Olšák2017-03-241-1/+1
| | | | | | Cc: 17.0 <[email protected]> Reviewed-by: Samuel Pitoiset <[email protected]> Reviewed-by: Nicolai Hähnle <[email protected]>
* radeonsi: fix dvec[34] attributes sourced from current attribute stateNicolai Hähnle2017-03-241-3/+4
| | | | | | | | | | | | The state tracker no longer uploads those attributes for us, so we must conservatively upload the size of the largest attribute, which is a dvec4. Fixes a regression of GL45-CTS.gpu_shader_fp64.varyings and GL45-CTS.vertex_attrib_64bit.limits_test. Fixes: 9b91e0b54cc2 ("radeonsi: allow unaligned vertex buffer offsets and strides on CIK-VI") Reviewed-by: Marek Olšák <[email protected]>
* gallivm: remove lp_add_attr_dereferenceable in favor of amd/commonMarek Olšák2017-03-221-1/+1
| | | | | Reviewed-by: Nicolai Hähnle <[email protected]> Reviewed-by: Dave Airlie <[email protected]>
* util/disk_cache: use a helper to compute cache keysGrazvydas Ignotas2017-03-211-2/+4
| | | | | | | | | | This will allow to hash additional data into the cache keys or even change the hashing algorithm easily, should we decide to do so. v2: don't try to compute key (and crash) if cache is disabled Signed-off-by: Grazvydas Ignotas <[email protected]> Reviewed-by: Timothy Arceri <[email protected]>
* radeonsi: check the IR type before waiting for a compute compilation fenceMarek Olšák2017-03-201-1/+3
| | | | | | | This should fix OpenCL getting stuck. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100288 Reviewed-by: Samuel Pitoiset <[email protected]>
* si_descriptor: move velems nullity check before dereferenceJulien Isorce2017-03-201-4/+11
| | | | | | | | | CID 1399479: Dereference before null check (REVERSE_INULL) check_after_deref: Null-checking velems suggests that it may be null, but it has already been dereferenced on all paths leading to the check. Signed-off-by: Julien Isorce <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* si_pipe: remove nullity check after dereferenceJulien Isorce2017-03-201-3/+0
| | | | | | | | | sscreen cannot be NULL CID 1354483 Signed-off-by: Julien Isorce <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* radeonsi: compile all TGSI compute shaders asynchronouslyMarek Olšák2017-03-171-44/+81
| | | | | | required by threaded gallium Reviewed-by: Timothy Arceri <[email protected]>
* radeonsi: require that compiler threads are enabledMarek Olšák2017-03-172-11/+13
| | | | | | | threaded gallium can't use pipe_context's LLVM target machine, because create_shader_selector can be called from a non-driver thread. Reviewed-by: Timothy Arceri <[email protected]>
* radeonsi: implement TGSI opcodes TEX_LZ and TXF_LZMarek Olšák2017-03-152-6/+16
| | | | | | | | | | This massively decreases VGPR spilling for DiRT Showdown, because we no longer have to use v4i32 for 2D fetches when level == 0. We now use v2i32 for those cases. DiRT Showdown - Spilled VGPRs: -26 (-81%) This surprisingly doesn't have any useful effect on performance (+ 0.05%).
* gallium: add PIPE_CAP_TGSI_TEX_TXF_LZMarek Olšák2017-03-151-0/+1
|
* radeonsi: disable sinking common instructions down to the end blockSamuel Pitoiset2017-03-151-0/+11
| | | | | | | | | | | | | | | | | | | | | | | | | | Initially this was a workaround for a bug introduced in LLVM 4.0 in the SimplifyCFG pass that caused image instrinsics to disappear (because they were badly sunk). Finally, this is a win because it decreases SGPR spilling and increases the number of waves a bit. Although, shader-db results are good I think we might want to remove it in the future once the issue is fixed. For now, enable it for LLVM >= 4.0. This also fixes a rendering issue with the speedometer in Dirt Rally. More information can be found here https://reviews.llvm.org/D26348. Thanks to Dave Airlie for the patch. v2: - add a FIXME comment - use if (HAVE_LLVM >= 0x0400) instead Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99484 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97988 Signed-off-by: Samuel Pitoiset <[email protected]> Cc: 17.0 <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* radeonsi: clean up tex_fetch_ptrs()Samuel Pitoiset2017-03-151-6/+4
| | | | | | | | Will also help when the src sampler register will be TGSI_FILE_CONSTANT for bindless. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
* radeonsi: mark all bound shader buffer ranges as initializedMarek Olšák2017-03-131-0/+3
| | | | | | | | This should prevent cases when a buffer was incorrectly mapped without synchronization just because this wasn't done. Cc: 13.0 17.0 <[email protected]> Reviewed-by: Samuel Pitoiset <[email protected]>
* radeonsi: fix elimination of literal VS outputsMarek Olšák2017-03-081-4/+7
| | | | | | broken when switched to the new intrinsics. Reviewed-by: Samuel Pitoiset <[email protected]>
* radeonsi: s/uint/enum pipe_shader_type/Brian Paul2017-03-082-2/+4
| | | | | | This can probably be done in more places in the driver. Reviewed-by: Edward O'Callaghan <[email protected]>
* gallium: s/uint/enum pipe_shader_type/ for set_constant_buffer()Brian Paul2017-03-081-1/+1
| | | | Reviewed-by: Edward O'Callaghan <[email protected]>
* gallium: s/unsigned/enum pipe_shader_type/ for pipe_screen::get_shader_param()Brian Paul2017-03-081-1/+3
| | | | Reviewed-by: Edward O'Callaghan <[email protected]>
* radeonsi/ac: move get_image_intr_name to commonDave Airlie2017-03-071-36/+10
| | | | | | | | This code is used in radv, so move to common build code. Reviewed-by: Bas Nieuwenhuizen <[email protected]> Reviewed-by: Marek Olšák <[email protected]> Signed-off-by: Dave Airlie <[email protected]>
* gallium/util: replace pipe_mutex_unlock() with mtx_unlock()Timothy Arceri2017-03-072-13/+13
| | | | | | | | | | pipe_mutex_unlock() was made unnecessary with fd33a6bcd7f12. Replaced using: find ./src -type f -exec sed -i -- \ 's:pipe_mutex_unlock(\([^)]*\)):mtx_unlock(\&\1):g' {} \; Reviewed-by: Marek Olšák <[email protected]>
* gallium/util: replace pipe_mutex_lock() with mtx_lock()Timothy Arceri2017-03-072-4/+4
| | | | | | | | | | replace pipe_mutex_lock() was made unnecessary with fd33a6bcd7f12. Replaced using: find ./src -type f -exec sed -i -- \ 's:pipe_mutex_lock(\([^)]*\)):mtx_lock(\&\1):g' {} \; Reviewed-by: Marek Olšák <[email protected]>
* gallium/util: replace pipe_mutex_destroy() with mtx_destroy()Timothy Arceri2017-03-072-3/+3
| | | | | | | | | | pipe_mutex_destroy() was made unnecessary with fd33a6bcd7f12. Replace was done with: find ./src -type f -exec sed -i -- \ 's:pipe_mutex_destroy(\([^)]*\)):mtx_destroy(\&\1):g' {} \; Reviewed-by: Marek Olšák <[email protected]>
* gallium/util: replace pipe_mutex_init() with mtx_init()Timothy Arceri2017-03-072-3/+3
| | | | | | | | | | pipe_mutex_init() was made unnecessary with fd33a6bcd7f12. Replace was done using: find ./src -type f -exec sed -i -- \ 's:pipe_mutex_init(\([^)]*\)):(void) mtx_init(\&\1, mtx_plain):g' {} \; Reviewed-by: Marek Olšák <[email protected]>
* gallium/util: replace pipe_mutex with mtx_tTimothy Arceri2017-03-072-3/+3
| | | | | | pipe_mutex was made unnecessary with fd33a6bcd7f12. Reviewed-by: Marek Olšák <[email protected]>
* radeonsi: drop support for LLVM 3.6 & 3.7Marek Olšák2017-03-064-24/+5
| | | | | | They are too old. Reviewed-by: Dave Airlie <[email protected]>
* radeonsi: set the convergent attribute where neededMarek Olšák2017-03-061-1/+1
| | | | Reviewed-by: Dave Airlie <[email protected]>
* radeonsi: fix LLVM 3.9 - don't use non-matching attributes on declarationsMarek Olšák2017-03-061-2/+4
| | | | | | | | | Call site attributes are used since LLVM 4.0. This also reverts commit b19caecbd6f310c1663b0cfe483d113ae3bd5fe2 "radeon/ac: fix intrinsic version check", because this is the correct fix. Reviewed-by: Dave Airlie <[email protected]>
* ac: normalize build helper namesMarek Olšák2017-03-033-35/+35
| | | | | | s/emit/build/ Reviewed-by: Dave Airlie <[email protected]>
* radeonsi: move SI.vs.load.input building into amd/commonMarek Olšák2017-03-031-15/+10
| | | | Reviewed-by: Dave Airlie <[email protected]>
* radeonsi: detect and mark loads/stores from read-only/write-only memoryMarek Olšák2017-03-031-10/+105
|
* ac: replace llvm.SI.tbuffer.store with llvm.amdgcn.buffer.store if ADD_TID=0Marek Olšák2017-03-031-11/+11
| | | | | | | | ADD_TID doesn't work. Needs more investigation. v2: remove leftover dead code Reviewed-by: Dave Airlie <[email protected]> (v1)
* radeonsi: use the writeonly LLVM attributeMarek Olšák2017-03-031-3/+6
|
* ac: remove offen parameter from ac_build_buffer_store_dwordMarek Olšák2017-03-031-13/+12
| | | | Reviewed-by: Dave Airlie <[email protected]>
* radeonsi: enable TC L2 for tessellation offchip storesMarek Olšák2017-03-031-8/+8
| | | | Vulkan does the same thing.
* radeonsi: merge and simplify tbuffer_store functionsMarek Olšák2017-03-031-40/+39
| | | | Reviewed-by: Dave Airlie <[email protected]>
* radeonsi: set noalias on input shader pointersMarek Olšák2017-03-031-0/+1
|
* radeonsi: replace AMDGPU.bfe.* with amdgcn.*bfeMarek Olšák2017-03-031-7/+4
| | | | Reviewed-by: Dave Airlie <[email protected]>
* radeonsi: move kill intrinsic building into amd/commonMarek Olšák2017-03-032-14/+12
| | | | | | just a cleanup Reviewed-by: Dave Airlie <[email protected]>
* radeonsi: set readnone on reads from read-only memoryMarek Olšák2017-03-031-10/+10
|
* radeonsi: replace SI.buffer.load.dword with amdgcn.buffer.loadMarek Olšák2017-03-031-45/+19
|
* radeonsi: replace SI.packf16 with amdgcn.cvt.pkrtzMarek Olšák2017-03-031-5/+1
|
* radeonsi: remove last use of llvm.SI.resinfoMarek Olšák2017-03-031-48/+49
| | | | and move one function up to reuse the code.
* radeonsi: move image intrinsic building to amd/commonMarek Olšák2017-03-031-92/+62
| | | | Reviewed-by: Dave Airlie <[email protected]>
* ac: replace SI.export with amdgcn.exp.*Marek Olšák2017-03-031-3/+5
| | | | Reviewed-by: Dave Airlie <[email protected]>
* radeonsi: move llvm.SI.export building to amd/commonMarek Olšák2017-03-031-162/+144
| | | | Reviewed-by: Dave Airlie <[email protected]>
* ac: unify build_type_name_for_intr functionsMarek Olšák2017-03-031-44/+5
| | | | Reviewed-by: Dave Airlie <[email protected]>
* radeonsi: set unorm=1 for TGSI_TEXTURE_SHADOWRECT as wellMarek Olšák2017-03-031-1/+2
| | | | | | It was harmless, because we also set unorm in the sampler state. Reviewed-by: Dave Airlie <[email protected]>
* tgsi/scan: record load/store/atomic image usageMarek Olšák2017-03-031-1/+3
| | | | Reviewed-by: Dave Airlie <[email protected]>
* radeonsi: add support for an on-disk shader cacheTimothy Arceri2017-03-031-7/+60
| | | | | | | | | | | | | V2: - when loading from disk cache also binary insert into memory cache. - check that the binary loaded from disk is the correct size. If not delete the cache item and skip loading from cache. V3: - remove unrequired variable Reviewed-by: Grigori Goronzy <[email protected]> Reviewed-by: Marek Olšák <[email protected]>
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-19 05:42:42 +0000