AlexIndustrial/mesa
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
155,425 Commits 1 Branch 0 Tags
ec70291da99333f6d8afa0afc052a9ea30c8e448
Commit Graph

143799 Commits

Author SHA1 Message Date
Icecream95 ec70291da9 panfrost: Stop using sparse_array for batch BOs 
Iterating over a util_sparse_array is very expensive; replace this
with a standard dynarray.

Using the sparse 'nodearray' datastructure instead was tested, but
found to be slower in some cases.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16988>
 2022-06-14 23:44:02 +00:00
Connor Abbott 169e03800d tu: Implement VK_EXT_color_write_enable 
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16936>
 2022-06-14 23:18:17 +00:00
Connor Abbott 94be0dd0b8 tu: Implement extendedDynamicState2LogicOp 
Because this impacts most of the registers in the BLEND draw state, we
make the entire draw state dynamic so that it all gets re-emitted when
the logicOp changes. This also lays the groundwork for
VK_EXT_color_write_enable.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16936>
 2022-06-14 23:18:17 +00:00
Connor Abbott c23eb99357 tu: Fix partial-write LRZ determination 
There were a few problems with this:

- It wasn't considering logic op at all, which is another source of
  reading from the destination.
- It was conditioned on the blend_enable_mask, so it was missing the
  case where there's no blending but some of the outputs were masked
  out.
- It wasn't considering attachments with less than 4 components (for
  example, normals in a typical deferred rendering setup) and would
  always consider them partially written unless the user added extra
  unnecessary components.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16936>
 2022-06-14 23:18:17 +00:00
Boris Brezillon b3b53952b0 dzn: Vulkan 1.1 wants maxPerSetDescriptors >= 1024 
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16926>
 2022-06-14 22:44:42 +00:00
Boris Brezillon 59a6ddd85c dzn: Implement GetDescriptorSetLayoutSupport() 
The 2048 descriptors limit comes from the maximum number of samplers
per heap, but the limit for other descriptors is actually much bigger.
Let's implement GetDescriptorSetLayoutSupport() to reflect that.

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16926>
 2022-06-14 22:44:42 +00:00
Boris Brezillon 1554ece8bd dzn: Add a dzn_desc_type_has_sampler() helper 
Add a dzn_desc_type_has_sampler() helper instead of duplicating
the SAMPLER || COMBINED_IMAGE_SAMPLER test everywhere.

Suggested-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16926>
 2022-06-14 22:44:42 +00:00
Boris Brezillon fb52be9376 dzn: Lower alignment requirements when allocating buffers or single-sample images 
VkMemoryDedicatedAllocateInfo, when present, provides us with extra
information about the memory usage, which allow us to lower the alignment
requirements.

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16926>
 2022-06-14 22:44:42 +00:00
Boris Brezillon 96a7f81072 dzn: Check that no export is requested in the memory allocation path 
We don't support exporting memory objects yet, so let's make sure the
user doesn't request that.

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16926>
 2022-06-14 22:44:42 +00:00
Boris Brezillon 8c23d9db0b dzn: Add dummy {Create,Destroy}SamplerYcbcrConversion() implementations 
We don't support Ycbcr sampler conversion. Add dummy implementations to
make us Vulkan 1.1 compliant.

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16926>
 2022-06-14 22:44:42 +00:00
Boris Brezillon 439b1fc8e8 dzn: Add a dummy GetImageSparseMemoryRequirements2() 
We don't support sparse memory yet, but this function needs to be
present.

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16926>
 2022-06-14 22:44:42 +00:00
Boris Brezillon cd8e322816 dzn: Add a dummy GetDeviceGroupPeerMemoryFeatures() 
We don't support device groups, but Vulkan 1.1 requires a
GetDeviceGroupPeerMemoryFeatures() implementation, so let's just
advertise no peer-memory features.

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16926>
 2022-06-14 22:44:42 +00:00
Boris Brezillon cc7158fe2f dzn: Return empty external properties 
We don't support importing/exporting images yet, so let's zero out the
whole external properties struct, if present.

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16926>
 2022-06-14 22:44:42 +00:00
Boris Brezillon fa79d036f1 dzn: Decorrelate external image info and external image properties 
One can be present without the other.

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16926>
 2022-06-14 22:44:42 +00:00
Boris Brezillon cd4bc4893b microsoft/compiler: Support load_sample_id_no_per_sample 
It's like load_sample_id except it shouldn't force per-sample shading
when not already enabled. In that case, we simply return 0.

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16926>
 2022-06-14 22:44:42 +00:00
Boris Brezillon 65309dab45 microsoft/compiler: s/assert()/unreachable()/ in emit_intrinsic() 
Use an unreachable() statement instead of the incorrect assertion in the
unsupported intrinsic path.

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16926>
 2022-06-14 22:44:42 +00:00
Boris Brezillon 82d3433d2a microsoft/compiler: Set typed_uav_load_additional_formats when appropriate 
This flag should be set to true when the RWTexture is attached a vector,
and we always declare a vec4 right now, so it should always be true.
Might be worth reworking the dxil_module_get_res_type() to use a scalar
when the image only has one component.

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16926>
 2022-06-14 22:44:42 +00:00
Boris Brezillon 303175cfec microsoft/compiler: Images are no longer reprensented by uniform vars 
emit_barrier_impl() was still checking the nir_var_uniform flag to
detect images, which is no longer correct.

Fixes: cfdc7ee066 ("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: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16926>
 2022-06-14 22:44:42 +00:00
Boris Brezillon cdbc8a8c85 spirv: Add a dst/src type to image deref loads/stores coming from OpAtomic{Load,Store} 
nir_to_dxil() uses those types to pick the right operation overload,
and atomic loads/stores are no different from their non-atomic
counterpart apart from the atomicity property, so it makes sense to
pass a type to the deref_{load,store} intrinsic in that case too.

Suggested-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16926>
 2022-06-14 22:44:42 +00:00
Boris Brezillon b78d3ebe72 vulkan/util: Fill VkPhysicalDeviceIDProperties::deviceNodeMask 
This field copy was missing in
vk_get_physical_device_core_1_1_property_ext().

Fixes: 19ff5019b7 ("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: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16926>
 2022-06-14 22:44:42 +00:00
Mike Blumenkrantz b240be28e3 zink: check for pending clears to determine write status of zs attachments 
as @Venemo discovered, zs layouts were being incorrectly set to readonly
in the case where the attachment was only used for an explicit clear,
so ensure that gets taken into account

cc: mesa-stable

fixes (radv):
dEQP-GLES31.functional.stencil_texturing.render.depth24_stencil8_clear
dEQP-GLES31.functional.stencil_texturing.render.depth32f_stencil8_clear

Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17033>
 2022-06-14 20:58:55 +00:00
Jose Maria Casanova Crespo 901f5e6a31 v3dv/ci: increase fraction to 10 on v3dv ci jobs. 
We reduce the v3dv ci jobs time execution from ~20min to
8-11 min.

Reviewed-by: Juan A. Suarez <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17026>
 2022-06-14 20:33:34 +00:00
Mike Blumenkrantz 8c0f3cdfac zink: clamp dynamic renderArea to framebuffer size 
it's illegal to have a renderArea larger than the attachments, so perform
clamping to avoid that scenario

Fixes: c81cd989c8 ("zink: use dynamic rendering (most of the time)")

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16947>
 2022-06-14 20:12:46 +00:00
Mike Blumenkrantz 87c7e75721 zink: fix cubemap lowering bit size 
this isn't always 32

Fixes: 2d745904ca ("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: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17008>
 2022-06-14 19:57:40 +00:00
Mike Blumenkrantz de6af39534 zink: fix cubemap lowering for array samplers 
each sampler is 1 driver location, so use the base variable

Fixes: 2d745904ca ("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: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17008>
 2022-06-14 19:57:40 +00:00
Mike Blumenkrantz 876bfb6bc3 zink: skip some emulated cube code if the extension is present 
cuts down on some cpu overhead

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16944>
 2022-06-14 19:35:42 +00:00
Mike Blumenkrantz d8b4ceb27a zink: rename nonseamless struct members to emulate_nonseamless 
make it clearer that this is separate from the gallium state

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16944>
 2022-06-14 19:35:42 +00:00
Mike Blumenkrantz 08858070b1 zink: use VK_EXT_non_seamless_cube_map when available 
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16944>
 2022-06-14 19:35:42 +00:00
Mike Blumenkrantz 616773e62b zink: hook up VK_EXT_non_seamless_cube_map 
Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16944>
 2022-06-14 19:35:42 +00:00
Mike Blumenkrantz 03d7f31054 zink: don't print VK_EXT_shader_atomic_float warning for nir_intrinsic_image_deref_atomic_exchange 
this is covered by base spec and doesn't require the extension

cc: mesa-stable

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17010>
 2022-06-14 19:21:15 +00:00
Mike Blumenkrantz d8877437d6 zink: avoid uninit values in renderpass state 
cc: mesa-stable

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17010>
 2022-06-14 19:21:15 +00:00
Mike Blumenkrantz 4c908e3d7b zink: only add necessary binds during dmabuf export 
SHARED might only be present, so add it conditionally

Fixes: 4eeabb59f5 ("zink: rebind resources for export as needed")

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17010>
 2022-06-14 19:21:15 +00:00
Mike Blumenkrantz 44cf6f7564 zink: fix generated tcs deletion 
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: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17010>
 2022-06-14 19:21:15 +00:00
Mike Blumenkrantz cf34d3fe05 zink: unset generated tcs pointer from tes on deletion 
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: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17010>
 2022-06-14 19:21:15 +00:00
Mike Blumenkrantz c36c5ff057 zink: use the patch_vertices value from the shader key, not the context 
ctx shouldn't be used here unless necessary

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17010>
 2022-06-14 19:21:15 +00:00
Mike Blumenkrantz 338d986439 zink: fix generated tcs update path 
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: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17010>
 2022-06-14 19:21:15 +00:00
Mike Blumenkrantz 336b5c8823 zink: fix dynamic stride conditional in pipeline creation 
if dynamic stride isn't used then this struct member must be filled

cc: mesa-stable

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17010>
 2022-06-14 19:21:15 +00:00
Louis-Francis Ratté-Boulianne ac34866166 dzn: Allow some non-native formats to be used as vertex inputs 
This requires shader-side lowering, which is handled in
dxil_nir_lower_vs_vertex_conversion().

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15955>
 2022-06-14 17:42:51 +00:00
Louis-Francis Ratté-Boulianne b53f9011b1 microsoft/compiler: Add support for more scaled formats 
Lower 8 and 16 bit scaled formats to integer formats.

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15955>
 2022-06-14 17:42:51 +00:00
Louis-Francis Ratté-Boulianne b0b9b50077 d3d12: Move d3d12_nir_lower_vs_vertex_conversion() to a common place 
So we can re-use it in dozen.

Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15955>
 2022-06-14 17:42:51 +00:00
Jesse Natalie c378bbab75 mesa: Use C11 alignas for cross-platform SSE support 
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16908>
 2022-06-14 15:08:38 +00:00
Erik Faye-Lund 8fbad5e42f gallium: clean up PIPE_ALIGN_STACK-definition 
There's really just one case where this is supported; on GCC for x86.
All other cases do nothing, so let's remove the complexity that is no
longer needed.

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16908>
 2022-06-14 15:08:38 +00:00
Erik Faye-Lund d1349eb500 gallium: remove unused macros 
These aren't used, so let's remove them.

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16908>
 2022-06-14 15:08:38 +00:00
Erik Faye-Lund ea7ba77be9 gallium: remove repeated check 
This is already inside an "#if defined(__GNUC__)" conditional, there's no
need to repeat the __GNUC__-check.

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16908>
 2022-06-14 15:08:38 +00:00
Erik Faye-Lund 1b53fb625e gallium: use c++11 alignas instead of PIPE_ALIGN_VAR 
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16908>
 2022-06-14 15:08:38 +00:00
Erik Faye-Lund 04ca04ba46 gallium: use c11 alignas instead of PIPE_ALIGN_VAR 
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16908>
 2022-06-14 15:08:38 +00:00
Erik Faye-Lund 69d2d38f6e llvmpipe: use c11 alignas instead of PIPE_ALIGN_VAR 
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16908>
 2022-06-14 15:08:38 +00:00
Erik Faye-Lund 3c82e23f9a gallium: reorganize includes 
There's no reason to have standard includes in two different sections of
the header, let's merge them. While we're at it, let's sort the includes
as well.

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16908>
 2022-06-14 15:08:38 +00:00
Erik Faye-Lund d791184b46 gallium: remove stale comment 
This hasn't been true for a long time, let's remove the comment.

Fixes: 0d4898ae80 ("include,gallium: Remove pre-MSVC 2013 compatibility.")
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16908>
 2022-06-14 15:08:37 +00:00
Erik Faye-Lund e3bc78b8e3 util: use c11 alignas instead of rolling our own 
Due to how alignas is defined, it itsn't allowed to use it on a struct,
it needs to be used on the first member instead. So move the declaration
in those cases.

This still leaves the ALIGN16 macro using compiler-specific directives,
because it's a lot of work to untangle the above. This probably deserves
its own MR.

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16908>
 2022-06-14 15:08:37 +00:00