Commit graph

176189 commits

Author SHA1 Message Date
Faith Ekstrand
f0a165799c nvk: Add an addr field to nvk_image_plane
Nothing actually needs the pointer to the memory object so it's better
to not even include it.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
Dave Airlie
ad37394dfe nvk: init image fields for requirements
Fixes a crash in dEQP-VK.memory.requirements.create_info.*

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
Faith Ekstrand
0498eeba32 nvk: Add a buffer alignment helper
This consolodates buffer alignment calculations into one place,
including physical device queries, buffer creation, and buffer memory
requirements queries.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
Faith Ekstrand
6a0cc93bd4 nvk: Reformat nvk_buffer.c
Mostly, using dev instead of device and re-flow the arguments to
nvk_GetDeviceBufferMemoryRequirements().

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
Karol Herbst
5fb3298a95 nouveau/winsys: fix SM value for Ada
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
Emma Anholt
9ffbd53bb7 nvk: Add support for probing as a platform device.
The driver now comes up on gk20a (Jetson TK1).

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
Emma Anholt
880663c25d nvk: Clean up redundant vendor checking for physical device creation.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
Emma Anholt
826486339b nvk: Quiet a compiler warning.
Looks like it's always set to me, but 🤷

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
Emma Anholt
0410054459 nouveau/winsys: Fix an undefined use in the error path.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
Dave Airlie
820c03ebee nvk/query: fix push buffer size for copy pool results.
the immediate takes two slots, so bump this by one,
fixes a crash in zink KHR-GL45.transform_feedback.query_vertex_interleaved_test

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
Faith Ekstrand
4fe57b2d7a nvk: Enable external memory extensions
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
Faith Ekstrand
113b9b00e7 nvk: Support dma-buf export
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
Faith Ekstrand
0399999dec nvk: Support dma-buf import
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
Faith Ekstrand
c370260a8f nouveau/winsys: Add dma-buf import support
This requires a lock and a buffer cache on the nouveau_ws_device.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
Faith Ekstrand
10ffe8e709 nvk: Require dedicated allocations for external images
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
Faith Ekstrand
7f0c06e130 nvk: Dedicated allocations override internal
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
Faith Ekstrand
702326d013 nvk: Add external memory queries
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
Faith Ekstrand
3d247f447e nvk: Add device and driver UUIDs
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
Faith Ekstrand
cccc044f51 nvk: Use abbreviated names in nvk_device_memory.c
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
Faith Ekstrand
14d35c7048 nvk: Use NVIDIA_VENDOR_ID in pdev try_create()
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
George Ouzounoudis
ccfbd94891 nvk: Enable descriptor indexing
These values follow the already existing per-stage and descriptor set
limits.  If we change these to be similar to the reported values of the
proprietary driver, care must be taken as Turing has bigger limits for
uniform buffers than older generations.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
George Ouzounoudis
edb58c6660 nvk: Add nir non-uniform optimization pass
The hardware supports non-uniform access natively so we do not need to
lower most non-uniform access in the shader. Just lower the ubo access
as it should probably be faster when we uniformly load from real hw
ubos in the future.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:06 +00:00
George Ouzounoudis
0dd15199a1 nvk: Handle cases of descriptor bindings with variable counts
We need to compute the allocated descriptor buffer size at descriptor allocation
time. So we need to split to a pre-computed size for all the other bindings
except the one with the variable count.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
George Ouzounoudis
6a74b3e311 nvk: Support VkDescriptorSetVariableDescriptorCountLayoutSupport
On cases with variable descriptor counts we need to calculate the
maximum supported variable count while taking into account the other existing
descriptors and their alignments in the set.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Faith Ekstrand
81bec8b63b nvk: Use the max descriptor alignemtn in GetDescriptorSetLayoutSupport
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Benjamin Lee
354fe8fda0 nvk: Fix segfault when opening DRI device file returns error
Previously, ver was unitialized on this path, causing UB on the
`if (ver)` branch after the out_err label.

Signed-off-by: Benjamin Lee <benjamin@computer.surgery>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Dave Airlie
3572f5cd7e nvk: add support for D32_SFLOAT_S8_UINT
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Faith Ekstrand
62fa4ff5e0 nvk: Take an nvk_image_plane in nouveau_copy_rect_image
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Faith Ekstrand
8d712c2e06 nvk: Add helpers for binding image planes
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Faith Ekstrand
fe022d4e87 nvk: s/device/dev in nvk_image.c
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Dave Airlie
ebcac6b81c nvk: introduce an optimisation loop.
This brings in pretty much the radv opt loop.

So for tess ctrl where this a loop we can unroll, this fixes
some dynamic array access later to be non-dynamic.

Fixes:
dEQP-VK.clipping.user_defined.clip_distance_dynamic_index.vert_tess.1

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Dave Airlie
0b6b59e654 nvk: merge tess info between tcs/tes.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Dave Airlie
136d49054b nvk: fix transform feedback crash when optimiser removes things.
This optimise loop can remove stuff, which causes this to crash,
just exit early.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
George Ouzounoudis
7715aea5aa nvk: Lower io to temporaries for tessellation evaluation nir
This is consistent to what radv does.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
George Ouzounoudis
f561406d38 nvk: Respect tessellation domain origin state
When switching to lower left tessellation domain origin we need to
switch the orientation for triangles/quads. The changes are based on
what anv and radv do.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
George Ouzounoudis
93369f1f96 nvk: Fix cases where execution mode is specified in the tesc shader.
We need to keep some context for the compilation of the tessellation shaders.
This is required in the case where the domain is specified in the
tessellation control shader instead of the tessellation evaluation
shader, as codegen needs the domain information when compiling the
tessellation evaluation shader.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
George Ouzounoudis
544aadc56b nvk: Enable multiview with tessellation shader
This works by default, we just switch it on.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
George Ouzounoudis
76e02c7f77 nvk: Assign locations correctly for arrayed IO
Arrayed IO, encountered in tessellation shaders, should not take up
multiple driver locations.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
George Ouzounoudis
304636297f nvk: Basic tessellation shader support
Based on nvc0 implementation but with added nvidia-headers usage.
Device limits taken from http://www.vulkan.gpuinfo.org/ and minimum
required values.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Thomas H.P. Andersen
ecbcc26a6e nvk: remove duplicated device features
These are already set in the Vulkan 1.2 section

Add status for bufferDeviceAddressCaptureReplayEXT

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Mohamed Ahmed
e41031d8ff nvk: Enable multiplane images and image views
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Faith Ekstrand
39d8d3811d nvk: Advertise KHR_spirv_1_4
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Faith Ekstrand
1ea94caab6 nvk: Fix the NO_PREFETCH assert for CmdDrawIndirect
The NO_PREFETCH bit gets set on the range.  The offset can pretty much
be whatever.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Faith Ekstrand
e2379a2e87 nvk: Advertise shaderStorageImageReadWithoutFormat
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Faith Ekstrand
6334cac288 nvk/heap: Set the right pitch for heap resize copies
Otherwise, once we need to grow past 2^17 B, we attempt to copy two
lines, the hardware errors out, and we get left with a junk heap.
While we're here, re-arrange the code a bit so we keep all the P_FOO
together.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Echo J
729b34be8a nil: Add A4B4G4R4_UNORM format support
DXVK versions before v1.5.5 need this format in some cases

I haven't tried running CTS yet, but GTA San Andreas runs
well now on DXVK v1.5.1 (before this change it crashed) so
I think CTS should be happy enough 🐸

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Dave Airlie
ac4073a4ee nvk: expose EXT_sampler_filter_minmax
The sampler bits of this were already hooked up.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Faith Ekstrand
72263d7589 nvk: Only use view_id for layer in multiview
I can't prove that this fixes anything but it seems right.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Rebecca Mckeever
3e93173cc9 nvk: Move code inside view mask loops to a helper function
Signed-off-by: Rebecca Mckeever <rebecca.mckeever@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00
Rebecca Mckeever
45001c878f nvk: Combine CLEAR_VIEWS and CLEAR_LAYERS MME macros
Check view_mask inside the macro to determine which loop to run. In
emit_clear_rects(), move code out of if/else that is the same in both
branches.

Signed-off-by: Rebecca Mckeever <rebecca.mckeever@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
2023-08-04 21:32:05 +00:00