- Jun 16, 2022
-
-
Dylan Baker authored
-
Dylan Baker authored
-
- Jun 15, 2022
-
-
this was correct for 64bit loads and manually converted 32bit loads (e.g., bindless), but it was broken for the case where 64bit was not supported, as the offset wasn't being correctly adjusted break out the offset division to hopefully make this a little clearer Fixes: 150d6ee9 ("zink: move all 64-32bit shader load rewriting to nir pass") Reviewed-by: Adam Jackson <ajax@redhat.com> Part-of: <mesa/mesa!16669> (cherry picked from commit bbe51366) Conflicts: src/gallium/drivers/zink/ci/zink-tu-a630-fails.txt CI file removed as it doesn't exist in 22.1
-
this probably generated some pretty gnarly code cc: mesa-stable Reviewed-by: Adam Jackson <ajax@redhat.com> Part-of: <mesa/mesa!16669> (cherry picked from commit 02a3a5ff) Conflicts: src/gallium/drivers/zink/zink_compiler.c
-
This context needs to be locked before usage, and flushed after. If it's forgotten, radeonsi may crash (eg #6666). To avoid this kind of error, introduce 2 helpers. cc: mesa-stable Reviewed-by: Marek Olšák <marek.olsak@amd.com> Part-of: <mesa/mesa!17032> (cherry picked from commit bda1c081)
-
emit_barrier_impl() was still checking the nir_var_uniform flag to detect images, which is no longer correct. Fixes: cfdc7ee0 ("spirv: Use nir_var_mem_image") Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com> Reviewed-by: Jesse Natalie <jenatali@microsoft.com> Part-of: <mesa/mesa!16926> (cherry picked from commit 303175cf)
-
This field copy was missing in vk_get_physical_device_core_1_1_property_ext(). Fixes: 19ff5019 ("vulkan: Add helpers for filling exts for core features and properties.") Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com> Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com> Reviewed-by: Jesse Natalie <jenatali@microsoft.com> Part-of: <mesa/mesa!16926> (cherry picked from commit b78d3ebe)
-
this isn't always 32 Fixes: 2d745904 ("zink: add a gently mangled version of the d3d12 cubemap -> array compiler pass") fixes: dEQP-GL45-ES31.functional.program_uniform.by_pointer.render.struct_in_array.sampler2D_samplerCube_* Reviewed-by: Adam Jackson <ajax@redhat.com> Part-of: <mesa/mesa!17008> (cherry picked from commit 87c7e757)
-
each sampler is 1 driver location, so use the base variable Fixes: 2d745904 ("zink: add a gently mangled version of the d3d12 cubemap -> array compiler pass") fixes: dEQP-GL45-ES31.functional.shaders.opaque_type_indexing.sampler.const_expression.*.samplercubearray Reviewed-by: Adam Jackson <ajax@redhat.com> Part-of: <mesa/mesa!17008> (cherry picked from commit de6af395)
-
this is covered by base spec and doesn't require the extension cc: mesa-stable Reviewed-by: Adam Jackson <ajax@redhat.com> Part-of: <mesa/mesa!17010> (cherry picked from commit 03d7f310)
-
cc: mesa-stable Reviewed-by: Adam Jackson <ajax@redhat.com> Part-of: <mesa/mesa!17010> (cherry picked from commit d8877437)
-
deleting the generated shader on the first loop iteration like this was broken if the shader was used in multiple programs, so delete at the end cc: mesa-stable Reviewed-by: Adam Jackson <ajax@redhat.com> Part-of: <mesa/mesa!17010> (cherry picked from commit 44cf6f75)
-
otherwise this will free the generated tcs multiple times if the tes is used by multiple programs cc: mesa-stable Reviewed-by: Adam Jackson <ajax@redhat.com> Part-of: <mesa/mesa!17010> (cherry picked from commit cf34d3fe)
-
is_generated is in a union with the generated tcs pointer for tes shaders, so check the shader stage too cc: mesa-stable Reviewed-by: Adam Jackson <ajax@redhat.com> Part-of: <mesa/mesa!17010> (cherry picked from commit 338d9864)
-
This hasn't been true for a long time, let's remove the comment. Fixes: 0d4898ae ("include,gallium: Remove pre-MSVC 2013 compatibility.") Reviewed-by: Jesse Natalie <jenatali@microsoft.com> Part-of: <mesa/mesa!16908> (cherry picked from commit d791184b)
-
This is not safe because it may skip regenerating the flags for the loop condition in the loop continue block and these flags may be stomped in the loop body by other conditionals. Fixes: 9909fe6b ('broadcom/compiler: Skip bool_to_cond where possible') Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com> Part-of: <mesa/mesa!17020> (cherry picked from commit a97f78eb)
-
We missed this one, whoops. Fixes: a012b219 ("microsoft: Initial vulkan-on-12 driver") Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com> Reviewed-by: Jesse Natalie <jenatali@microsoft.com> Part-of: <mesa/mesa!16997> (cherry picked from commit 4ab49e25)
-
The order of calls was incorrect. Fixes: e9c41b32 - gallium/u_threaded: add buffer lists - tracking of buffers referenced by tc Closes: mesa/mesa#5327 Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com> Reviewed-By: Jonathan Strobl <jonathan.strobl@gmx.de> Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com> Part-of: <mesa/mesa!16989> (cherry picked from commit 4c5f7b86)
-
cc: mesa-stable Reviewed-by: Adam Jackson <ajax@redhat.com> Part-of: <mesa/mesa!17009> (cherry picked from commit 399104f0)
-
this otherwise causes the driver to make bad assumptions about how the resource should be handled cc: mesa-stable Reviewed-by: Adam Jackson <ajax@redhat.com> Part-of: <mesa/mesa!17009> (cherry picked from commit 23892bff)
-
implicit sync is hard, and many drivers get it wrong, so assume that anyone who isn't mesa might need some hand-holding cc: mesa-stable Reviewed-by: Adam Jackson <ajax@redhat.com> Part-of: <mesa/mesa!17009> (cherry picked from commit ea9e30f9)
-
Dylan Baker authored
-
avoid doing any sort of extra work here and avoid multi-context weirdness cc: mesa-stable Reviewed-by: Dave Airlie <airlied@redhat.com> Part-of: <mesa/mesa!16900>
-
dummy attachments never overwrite existing attachments, so they must be explicitly compared like this to avoid erroneous crashing Fixes: 3892c133 ("zink: add an alternate path for EXT_color_write_enable usage") Reviewed-by: Dave Airlie <airlied@redhat.com> Part-of: <mesa/mesa!16848>
-
This avoids u_blitter recursion: #0 util_blitter_set_running_flag #1 util_blitter_custom_color #2 si_blit_decompress_color #3 si_decompress_dcc #4 si_texture_disable_dcc #5 si_update_ps_colorbuf0_slot #6 si_bind_ps_shader #7 util_blitter_restore_fragment_states #8 util_blitter_custom_color #9 si_blit_decompress_color #10 si_decompress_dcc #11 si_sdma_copy_image #12 si_blit cc: mesa-stable Reviewed-by: Marek Olšák <marek.olsak@amd.com> Part-of: <mesa/mesa!16962> (cherry picked from commit 3d37291e) Conflicts: src/gallium/drivers/radeonsi/si_blit.c
-
Looks like some hardware needs this info in the shader to match the topology. Since there's no spot in the shader info for it, we're currently using the array size of the TCS input vars to store it. Cc: mesa-stable Reviewed-by: Paul Dodzweit <paul.dodzweit@amd.com> Tested-by: Paul Dodzweit <paul.dodzweit@amd.com> Part-of: <mesa/mesa!16920> (cherry picked from commit cc805aef) Conflicts: src/microsoft/compiler/dxil_nir.h
-
ensure that swapchains that are about to be presented asynchronously cannot be destroyed Fixes: 8ade5588 ("zink: add kopper api") Reviewed-by: Adam Jackson <ajax@redhat.com> Part-of: <mesa/mesa!16932> (cherry picked from commit 4bee678d)
-
this provides info about whether a swapchain is in use in another thread Fixes: 8ade5588 ("zink: add kopper api") Reviewed-by: Adam Jackson <ajax@redhat.com> Part-of: <mesa/mesa!16932> (cherry picked from commit 1dfa19c5)
-
this ensures that the same pointer is being accessed from the present thread as was intended when the present was queued Fixes: 8ade5588 ("zink: add kopper api") Reviewed-by: Adam Jackson <ajax@redhat.com> Part-of: <mesa/mesa!16932> (cherry picked from commit 892f4227)
-
The hardware writes one CRC per (effective) tile, the tile size of the CRC buffer is the same as the configured effective tile size. However, all our CRC infrastructure assumes 16x16 tiles. In case CRC is used with smaller tiles, buffer overflows and incorrect rendering are all possible. Don't use CRC at smaller tile sizes. Note disabling CRC correctly invalidates any bound CRC buffers. Fixes: 2e97d7c8 ("panfrost: Transaction elimination support") Closes: #6332 Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com> Part-of: <mesa/mesa!16983> (cherry picked from commit 44223e5f)
-
It has a single user -- in a section of code that only runs for MFBD GPUs and that has already decided whether to use CRCs -- so inlining it simplifies its definition greatly and may avoid redeciding the CRC setting. [Note for mesa-stable maintainers: This is not a bug fix but is marked for backport so the next patch applies cleanly.] Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com> Cc: mesa-stable Part-of: <mesa/mesa!16983> (cherry picked from commit cac0578e)
-
This was accidentally broken while refactoring the script. Fixes: bf3c772e ("ci: Improve vs2019 mesa_build.ps1 for remove the need of cmd.exe") Reviewed-by: Yonggang Luo <luoyonggang@gmail.com> Part-of: <mesa/mesa!16965> (cherry picked from commit 003d0528)
-
If the type is not an array, glsl_get_length() returns 0 and we don't update the new_vars[]/flat_vars[] entries. Fixes: bcd14756 ("nir/lower_io_to_vector: add flat mode") Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com> Reviewed-by: Jesse Natalie <jenatali@microsoft.com> Part-of: <mesa/mesa!16960> (cherry picked from commit d9ec7df2)
-
all of ntv requires scalarized io since the offsets are now array indices instead of byte offsets, so enforce scalarization here to avoid breaking the universe Fixes: 150d6ee9 ("zink: move all 64-32bit shader load rewriting to nir pass") Reviewed-by: Adam Jackson <ajax@redhat.com> Part-of: <mesa/mesa!16669> (cherry picked from commit cdaa601d)
-
the last output isn't always only consuming 1 slot, so ensure that the xfb outputs begin at the appropriate place cc: mesa-stable Reviewed-by: Adam Jackson <ajax@redhat.com> Part-of: <mesa/mesa!16669> (cherry picked from commit e9d28cbe)
-
the previous conditional here was stupid and wrong: it should be comparing to see whether the writemask is the full mask of the type's size cc: mesa-stable Reviewed-by: Adam Jackson <ajax@redhat.com> Part-of: <mesa/mesa!16669> (cherry picked from commit 31ba19ff)
-
If the executables are still hanging out, anv_GetPipelineExecutableStatisticsKHR will try to dereference NULL pointers in pipeline->shaders[MESA_SHADER_FRAGMENT]. At least in terms of fossil-db output, this matches the behavior from before 73b3efcd. Fixes: 73b3efcd ("anv: Handle the null FS optimization after compiling shaders") Closes: #6590 Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com> Part-of: <mesa/mesa!16898> (cherry picked from commit 65d6708b)
-
With the following structures : struct StructA { uint64_t value0; uint8_t value1; }; struct TopStruct { struct StructA a; uint8_t value3; }; Currently offsetof(struct TopStruct, value3) = 9. While the same code on the CPU gives offsetof(struct TopStruct, value3) = 16. This is impacting OpenCL kernels we're trying to use to build acceleration structures. v2: Add comment/link to some description of the alignment/size computation Cc: mesa-stable Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com> Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com> Part-of: <mesa/mesa!16940> (cherry picked from commit 13362019)
-
the 'notemplates' debug mode is somewhat misleading since there's no uncached+notemplates mechanism, meaning that if the descriptor cache explodes it'll still use templates for updating in the fallback path Fixes: 4e376891 ("zink: add ZINK_DESCRIPTORS env var to explicitly set a mode") Reviewed-by: Dave Airlie <airlied@redhat.com> Part-of: <mesa/mesa!16927> (cherry picked from commit ee1a0a07)
-
- Jun 13, 2022
-
-
Dylan Baker authored
-