summaryrefslogtreecommitdiffstats
path: root/src/amd
Commit message (Collapse)AuthorAgeFilesLines
* radv: Store color/depth surface info in attachment info instead of framebuffer.Bas Nieuwenhuizen2019-08-027-104/+102
| | | | | | That way we can use it for imageless framebuffers. Reviewed-by: Samuel Pitoiset <[email protected]>
* ac/nir,radv: Optimize bounds check for 64 bit CAS.Bas Nieuwenhuizen2019-08-027-17/+36
| | | | | | | | When the application does not ask for robust buffer access. Only implemented the check in radv. Reviewed-by: Samuel Pitoiset <[email protected]>
* radv: fix image_has_{cmask,fmask}() helpersSamuel Pitoiset2019-08-021-2/+2
| | | | | | | | | | | The driver should now rely on cmask_offset because CMASK can be disabled by the driver for some reasons (eg. mipmaps). Apply the same change for FMASK, although it should be useless. Fixes: ad1bc8621df ("radv: remove radv_get_image_fmask_info()") Fixes: 10d08da52c6 ("radv/gfx10: add missing dcc_tile_swizzle tweak") Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv: remove radv_get_image_fmask_info()Samuel Pitoiset2019-08-024-59/+25
| | | | | | | It's unnecessary to duplicate fields in another struct. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv/gfx10: add missing dcc_tile_swizzle tweakSamuel Pitoiset2019-08-021-1/+3
| | | | | | Fixes: c90f46700dd ("radv/gfx10: mask DCC tile swizzle by alignment") Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv: remove radv_get_image_cmask_info()Samuel Pitoiset2019-08-024-45/+21
| | | | | | | It's unnecessary to duplicate fields in another struct. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv: only account for tile_swizzle for color surfaces with DCCSamuel Pitoiset2019-08-021-3/+3
| | | | | | | It's 0 for depth surfaces with TC compat HTILE enabled. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv: Enable VK_KHR_shader_atomic_int64Bas Nieuwenhuizen2019-08-022-6/+3
| | | | Reviewed-by: Samuel Pitoiset <[email protected]>
* ac/nir: Implement LLVM9 64-bit buffer compare & exchange.Bas Nieuwenhuizen2019-08-021-4/+64
| | | | | | | | | | | LLVM 9 does not have a 64-bit buffer compswap intrinsic, so this extracts the ptr, does a bound check and then uses a cmpxchg LLVM instruction. Not ideal, but the earliest release we're going to get a proper intrinsic is LLVM 10. Reviewed-by: Samuel Pitoiset <[email protected]>
* Revert "ac/nir: handle negate modifier"Connor Abbott2019-08-021-12/+1
| | | | This reverts commit bfea7e4d2965269bff8f1f6449cb99c312fd7384.
* Revert "ac/nir: handle abs modifier"Connor Abbott2019-08-021-30/+11
| | | | | | This reverts commit d3c80733cdfe8552b2f447ec8ed62465d0f2af1a. These were only appearing due to memory corruption.
* radv: re-apply "Optimize rebinding the same descriptor set."Samuel Pitoiset2019-08-021-1/+7
| | | | | | | | | | | | | This makes it cheaper to just change the dynamic offsets with the same descriptor sets. This optimization has been reverted a while back because of random GPU hangs on GFX9, no it looks fine, at least CTS no longer hangs on GFX9 and it doesn't hang on GFX10 as well. It fixes a performance problem with Wolfenstein Youngblood. Suggested-by: Philip Rebohle <[email protected]>
* radv/gfx10: use the correct target machine for Wave32Samuel Pitoiset2019-08-023-10/+26
| | | | | Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv/gfx10: add Wave32 support for vertex, tessellation and geometry shadersSamuel Pitoiset2019-08-027-8/+26
| | | | | | | It can be enabled with RADV_PERFTEST=gewave32. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv/gfx10: add Wave32 support for fragment shadersSamuel Pitoiset2019-08-027-2/+16
| | | | | | | It can be enabled with RADV_PERFTEST=pswave32. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv/gfx10: implement a GE bug workaroundSamuel Pitoiset2019-07-311-4/+23
| | | | | Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv/gfx10: remove an obsolete VGT_REUSE_OFF workaroundSamuel Pitoiset2019-07-311-6/+0
| | | | | Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv/gfx10: disable LATE_ALLOC_GS on Navi14Samuel Pitoiset2019-07-311-1/+8
| | | | | Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv/gfx10: implement a bug workaround for GE_PC_ALLOCSamuel Pitoiset2019-07-312-17/+13
| | | | | Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv/gfx10: implement a bug workaround for NGG -> legacy transitionsSamuel Pitoiset2019-07-312-2/+21
| | | | | Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv: skip draw calls with 0-sized index buffersSamuel Pitoiset2019-07-311-0/+6
| | | | | Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* tree-wide: replace MAYBE_UNUSED with ASSERTEDEric Engestrom2019-07-317-27/+27
| | | | | | Suggested-by: Jason Ekstrand <[email protected]> Signed-off-by: Eric Engestrom <[email protected]> Reviewed-by: Matt Turner <[email protected]>
* radv: drop incorrect MAYBE_UNUSEDEric Engestrom2019-07-311-2/+2
| | | | | | | `compressed` is clearly always used on the line right after. Signed-off-by: Eric Engestrom <[email protected]> Reviewed-by: Matt Turner <[email protected]>
* radv/gfx10: add Wave32 support for compute shadersSamuel Pitoiset2019-07-317-6/+53
| | | | | | | It can be enabled with RADV_PERFTEST=cswave32. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* ac/nir: fix incorrect Phis if callbacks use control flow inside control flowMarek Olšák2019-07-301-2/+2
|
* ac/nir: handle abs modifierMarek Olšák2019-07-301-11/+30
|
* ac: fix a memory leak in the error path of ac_build_type_name_for_intrMarek Olšák2019-07-301-0/+1
|
* ac: allow control flow statements in NIR callbacksMarek Olšák2019-07-302-20/+29
| | | | This fixes a crash when compiling geometry shaders on radeonsi.
* ac/nir: handle negate modifierMarek Olšák2019-07-301-1/+12
|
* radeonsi/nir: implement FBFETCH for KHR_blend_equation_advancedMarek Olšák2019-07-302-0/+7
|
* radeonsi: adjust RB+ blend optimization settingsMarek Olšák2019-07-301-1/+1
| | | | based on PAL
* ac/surface: allow linear swizzle mode automatic selection on gfx9 & 10Marek Olšák2019-07-301-1/+0
| | | | let addrlib make the decision to get the same result as PAL.
* radv: Fix descriptor set allocation failure.Bas Nieuwenhuizen2019-07-301-1/+5
| | | | | | | | | | | | | | Set all the handles to VK_NULL_HANDLE: "If the creation of any of those descriptor sets fails, then the implementation must destroy all successfully created descriptor set objects from this command, set all entries of the pDescriptorSets array to VK_NULL_HANDLE and return the error." (Vulkan 1.1.117 Spec, section 13.2) CC: <[email protected]> Reviewed-by: Dave Airlie <[email protected]>
* radv: fix queries with WAIT_BIT returning VK_NOT_READYAndres Rodriguez2019-07-271-1/+1
| | | | | | | | | | | | | | | | | When vkGetQueryPoolResults() is called with VK_QUERY_RESULT_WAIT_BIT set, the driver is supposed to wait for the query to become available before returning. Currently, radv returns once the query is indeed ready, but it returns VK_NOT_READY. It also fails to populate the results. The problem is a missing volatile in the secondary check for query availability. This patch removes the secondary check altogether since it is redundant with the preceding loop. This bug was found with an unreleased version of SteamVR. Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv/gfx10: only compile the GS copy shader on-demandSamuel Pitoiset2019-07-301-1/+2
| | | | | Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv: Don't include radv_private.h from radv_shader.hDaniel Schürmann2019-07-306-94/+143
| | | | | | | This patch decouples radv_shader.h from any LLVM dependency. Reviewed-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* amd: add support for ArcturusMarek Olšák2019-07-293-0/+7
| | | | Reviewed-by: Pierre-Eric Pelloux-Prayer <[email protected]>
* radeonsi: add support for compute-only chipsMarek Olšák2019-07-293-1/+6
| | | | Reviewed-by: Pierre-Eric Pelloux-Prayer <[email protected]>
* radv: implement VK_EXT_index_type_uint8Samuel Pitoiset2019-07-293-6/+61
| | | | | | | Natively supported on VI+. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv: Take variable descriptor counts into account for buffer entries.Bas Nieuwenhuizen2019-07-291-1/+10
| | | | | | Fixes: b5e04e9217b "radv: Support allocating variable size descriptor sets." Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=111019 Reviewed-by: Samuel Pitoiset <[email protected]>
* radv/gfx10: do not use the fast depth or stencil clear bytes pathSamuel Pitoiset2019-07-291-2/+3
| | | | | | | | | It causes issues on GFX10. This fixes rendering issues with vkmark and Wreckfest at least. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]
* ac: do not crash when the buffer data format is invalidSamuel Pitoiset2019-07-291-0/+1
| | | | | | | | | | This might happen when a pipeline doesn't define the vertex input state, so the buffer data format is 0 (aka INVALID). This fixes crashes when compiling some shaders on GFX10. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* ac/nir: fix txf_ms with an offsetRhys Perry2019-07-291-2/+2
| | | | | | | | | | Seems to fix some hair artifacts in Max Payne 3: https://github.com/daniel-schuermann/mesa/issues/76 Signed-off-by: Rhys Perry <[email protected]> Fixes: f4e499ec791 ('radv: add initial non-conformant radv vulkan driver') Reviewed-by: Bas Nieuwenhuizen <[email protected]> Reviewed-by: Samuel Pitoiset <[email protected]>
* radv: Delete unused local variables in optimization loopConnor Abbott2019-07-291-0/+2
| | | | | | | | | | | | | | | | Totals from affected shaders: SGPRS: 376 -> 376 (0.00 %) VGPRS: 620 -> 560 (-9.68 %) Spilled SGPRs: 0 -> 0 (0.00 %) Spilled VGPRs: 0 -> 0 (0.00 %) Private memory VGPRs: 0 -> 0 (0.00 %) Scratch size: 292 -> 292 (0.00 %) dwords per thread Code Size: 20024 -> 20144 (0.60 %) bytes LDS: 0 -> 0 (0.00 %) blocks Max Waves: 25 -> 25 (0.00 %) Wait states: 0 -> 0 (0.00 %) Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* anv+tu+radv: delete unusable dev_icd.jsonEric Engestrom2019-07-261-13/+0
| | | | | | | | | | | As per previous commit, Meson doesn't support using uninstalled libs, they're simply not ready until `ninja install` is ran, so delete them. Suggested-by: Jason Ekstrand <[email protected]> Signed-off-by: Eric Engestrom <[email protected]> Reviewed-by: Jason Ekstrand <[email protected]> # for anv Reviewed-by: Eric Anholt <[email protected]> # for tu Reviewed-by: Bas Nieuwenhuizen <[email protected]> # for radv
* radv: Set correct metadata size for GFX9+.Bas Nieuwenhuizen2019-07-251-1/+2
| | | | | | | | | | | Without correct size, radeonsi assumes the metadata is incorrect, which can and will cause issues. Since the metadata is really incorrect without the size, let us fix that. Fixes: e43cc3e3afc "radv/gfx9: handle GFX9 opaque metadata" Reviewed-by: Samuel Pitoiset <[email protected]>
* radv/gfx10: fix intensity formats by setting ALPHA_IS_ON_MSBSamuel Pitoiset2019-07-251-6/+11
| | | | | | | | | | This fixes dEQP-VK.rasterization.primitive_size.points.point_size_* This also fixes some black squares with the Sascha SSAO demo. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv/gfx10: use L2 for DMA copy/fill operationsSamuel Pitoiset2019-07-251-0/+16
| | | | | | | | | | | | | It's coherent and faster. GFX7-GFX9 should also support this but for now only uses L2 for GFX10 because it's untested on previous gens. This fixes dEQP-VK.memory.pipeline_barrier.transfer_* This also fixes some missing geometry in Dawn Of War III because VBOs weren't updated correctly. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv/gfx10: enable VK_EXT_transform_feedbackSamuel Pitoiset2019-07-241-1/+1
| | | | | | | | | | | | When a pipeline uses transform feedback, the driver fallbacks to the legacy path because NGG support for streamout is a non-trivial amount of work. AMDVLK also uses the legacy path for streamout, while RadeonSI uses the new NGG path. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
* radv/gfx10: do not enable NGG if a pipeline uses XFBSamuel Pitoiset2019-07-241-0/+27
| | | | | | | | NGG GS for streamout requires a bunch of work, so enable it with the legacy path only for now. Signed-off-by: Samuel Pitoiset <[email protected]> Reviewed-by: Bas Nieuwenhuizen <[email protected]>
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-18 13:42:00 +0000