Faith Ekstrand [Tue, 31 Jan 2023 02:11:56 +0000 (20:11 -0600)]
nvk: Invalidate texture header and sampler caches each submit
It's possible descriptors have been added or removed. We could track
it, but there's really no point. Per-submit shouldn't be terrible.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:56 +0000 (20:11 -0600)]
nvk: Rework TLS/SLM and image/sampler table handling
When Dave originally added SLM, he did so with a preamble pushbuf which
sets up the SLM area prior to executing any actual command buffers.
This has the advantage that the SLM BO is never directly referenced in
any command buffer and we can swap it out even after command buffers
have been built. When I added image/sampler tables, I sort-of went
along with it but then did something different for 3D.
Thanks to buffer reference counting, we can employ a fairly simple
locking scheme. Each of the globals: images, samplers, and TSL has a
lock which gets taken whenever the table is modified. When we go to do
a submit, we take each lock and take a queue-local reference to the BO
and grab a copy of its size parameter under the lock. If anything has
changed, we update the queue-local preamble pushbuf. Because the queue
also holds a reference, we can safely proceed to submit command buffers
which reference those global BOs without the individual global locks
taken. If SLM or one of the descriptor tables resizes while we're
mid-submit, our reference to the BO will remain valid until either the
destruction of the queue or the next submit.
The one small bit of threaded cleverness (one always has to be careful
with that) is that the SLM size can be checked without taking the lock
because the SLM size only ever increases.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:56 +0000 (20:11 -0600)]
nvk: Drop most buffer tracking
If it's backed by an nvk_device_memory, we don't need to track it since
we always add all of those to every submit so that bindless works.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:56 +0000 (20:11 -0600)]
nvk/dispatch: Rename push to p
I took to using this convention in nvk_cmd_draw.c because it's shorter.
Let's use it other places too.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:56 +0000 (20:11 -0600)]
nvk/blit: Rename push to p
I took to using this convention in nvk_cmd_draw.c because it's shorter.
Let's use it other places too.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:56 +0000 (20:11 -0600)]
nvk/copy: Rename push to p
I took to using this convention in nvk_cmd_draw.c because it's shorter.
Let's use it other places too.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:56 +0000 (20:11 -0600)]
nvk: Inline nouveau_copy_linear
It has exactly one caller so the helper's not really doing us much.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Karol Herbst [Mon, 29 Aug 2022 13:59:44 +0000 (15:59 +0200)]
nvk: clean up bo mappings
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Karol Herbst [Mon, 29 Aug 2022 13:48:44 +0000 (15:48 +0200)]
nouveau/ws: add a bo unmap helper function
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Karol Herbst [Wed, 24 Aug 2022 11:38:18 +0000 (13:38 +0200)]
nouveau/ws: add API to query if the context was killed
Signed-off-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:56 +0000 (20:11 -0600)]
nvk: Fix dynamic buffer descriptor copies
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:56 +0000 (20:11 -0600)]
nvk: Set up clip and cull distances
Closes: https://gitlab.freedesktop.org/nouveau/mesa/-/issues/12
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:56 +0000 (20:11 -0600)]
nil: Fix nonnormalized coordinates
Also, replace our TODO with a real comment explaining how nonnormalized
coordinate select works between image and texture header.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:56 +0000 (20:11 -0600)]
nil: Fix max mip level
The MAX_MIP_LEVEL field needs the actual number from the image because
it's used to compute the stride between layers. The restrictions on mip
level come entirely from the RES_VIEW fields.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:56 +0000 (20:11 -0600)]
nil/format: Depth/stencil formats appear as red
Previously, it was treated as luminance and the depth or stencil
component was broadcast to all channels except alpha. In Vulkan, depth
and stencil textures are treated as red. If we want it treated
differntly by some other API, that driver can do that with a swizzle.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:56 +0000 (20:11 -0600)]
nvk: Fix border color alpha
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: VkSamplerCreateInfo::mipLodBias is signed
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Assert on CmdExecuteCommands
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Fix max anisotropy
Now that we're using the bitpack helpers, we don't need the extra shift
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Don't advertise tiling on non-power-of-two formats
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Use a UINT format to clear non-renderable images
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Implement vkCmdClear*Image directly
Instead of invoking the meta code, implement it ourselves so we get the
HW clear path. We could probably put these implementations somewhere
common since they only set up a few things and invoke Begin/EndRendering
but there's not that much point for now.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Add image_view_init/finish functions
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Use hardware clears for attachment clears
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nouveau/parser: Add array and float tags for clear values
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Mauro Rossi [Tue, 23 Aug 2022 08:36:11 +0000 (10:36 +0200)]
vulkan/meta: fix gnu-empty-initializer build error
Fixes the following building error:
../src/vulkan/runtime/vk_meta_clear.c:213:42: error: use of GNU empty initializer extension [-Werror,-Wgnu-empty-initializer]
struct vk_meta_clear_push_data push = { };
^
1 error generated.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Save/restore push constants around meta ops
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Wire up early z and post depth coverage
Was crawling through header bits and found that we weren't plumbing
these through.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Dave Airlie [Tue, 16 Aug 2022 04:27:40 +0000 (14:27 +1000)]
nvk: add basic primitive restart
fixes crucible
func.draw-index16-restart
func.draw-index32-restart
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Yusuf Khan [Tue, 23 Aug 2022 21:28:05 +0000 (16:28 -0500)]
nvk: support GetImageSparseMemoryRequirements2
KHR_get_memory_requirements2 was enabled but GetImageMemoryRequirem
ents2 was not. This stubs out GetImageMemoryRequirements2 to fix it.
Signed-off-by: Yusuf Khan <yusisamerican@gmail.com>
v2: remove the v1 version of the command as its provided by the
common runtime
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Yusuf Khan [Tue, 23 Aug 2022 14:28:39 +0000 (09:28 -0500)]
nvk: implement GetDeviceMemoryCommitment
vulkan spec says:
If the memory object is allocated from a heap with the
VK_MEMORY_PROPERTY_LAZILY_ALLOCATED_BIT bit set, that object’s
backing memory may be provided by the implementation lazily. The
actual committed size of the memory may initially be as small as
zero (or as large as the requested size), and monotonically
increases as additional memory is needed.
As far as I can tell we ignore the bit meaning that we allocate the
requested size so return that as the size.
Signed-off-by: Yusuf Khan <yusisamerican@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Allocate descriptors for input attachments
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Assert samples == 1
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Re-flow a couple function prototypes
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Align descriptor buffers to NVK_MIN_UBO_ALIGNMENT
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Add nvk_queue_init/finish() helpers
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nil: Add a PTE kind for Z32_FLOAT
This values was found by trial-and-error poking at the hardware until it
stopped complaining. It's not clear if we need separate values for
compressed vs. not. This appears to work on Turing regardless of what
we set for SET_Z_COMPRESSION.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Return OUT_OF_DEVICE_MEMORY if bo_new fails
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Set GEOMETRY_SHADER_SELECTS_LAYER properly
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Call nir_lower_input_attachments
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Don't destroy NULL descriptor pool BOs
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Advertise depth/stencil support
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Flip the front-face setting
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Enable two-sided stencil
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:55 +0000 (20:11 -0600)]
nvk: Fix viewport Z scale
Because Z needs to end up in [0, 1] and not [-1, 1], we don't want to
divide by 2.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nvk: Fix depth/stencil render pass clears
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nvk: Do internal dedicated allocations for ZS images
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nvk: Add internal helpers for device memory allocation
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nouveau: Add a function to allocate a tiled buffer
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nil: Compute PTE kinds and tile modes for images
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nvk: Disable sparse Z on Maxwell+
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nvk: Depth buffers don't allow Z-tiling
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
vulkan/meta: Make stencil reference dynamic for clears
We call CmdSetStencilReference() but don't set it to dynamic in the
pipeline.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nvk: Clear dynamic state dirty after flushing it all
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nvk: Advertise more format features
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nil: Add more format support helpers
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nil: Re-introduce the format capabilities
The bits are already in the table we copied from the nouveau GL drivers,
we just need to set up some enums and plumb it through. With that,
we're now properly advertising which formats are renderable.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nil: Rename rendering to color_target
Also, add a helper for depth/stencil formats and use
nil_format_supports_color_targets() helper to compute Vulkan format
features rather than checking for a non-zero color format.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nil: Rename rt to czt in the format info struct
It's used for both depth/stencil and color so rt isn't quite the name we
want. Also, add a comment.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nil: Use nvidia headers for ZS format enums
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nil: Use the nvidia headers for render target format enums
Two notable changes here:
1. BITMAP no longer exists
2. B10G10R10A2_UINT was wrongly mapped to RGB10_A2_UINT before
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nil: Use nvidia headers for texture format enums
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nil: Properly depend on nouveau winsys and nvidia-headers
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nouveau: Add class information to nouveau_ws_device
Annoyingly, with the current uAPI, this requires creating a dummy
context and throwing it away. Fortunately, we should only do this once
on driver load so it's really not that big a problem.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nvk: Depth bounds need fui()
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nouveau/parser: Dump more fields as float
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nvk: Clean up try_create_physical_device error handling
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nvk: Advertise vertex buffer format featues
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nvk: Move the vertex format table into nvk_format.h
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nvk: Re-initialize dynamic_graphics_state.vi when recycling
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nvk/meta: Save and restore VI state
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nvk: Drop sample locations structs
IDK if NVIDIA hardware even supports setting sample locations. I've not
been able to find the packets. In any case, if and when we implement
the extension, we can wire it up then.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nvk: Implement copies for D24_UNORM_S8_UINT images
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:54 +0000 (20:11 -0600)]
nvk/copy: Mape bpp part of nouveau_copy_buffer
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Dave Airlie [Tue, 16 Aug 2022 20:22:17 +0000 (06:22 +1000)]
nvk: handle NULL to destroy descriptor pool
Fixes dEQP-VK.api.null_handle.destroy_descriptor_pool
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Dave Airlie [Wed, 17 Aug 2022 03:09:50 +0000 (13:09 +1000)]
nvk: lower io to temps to avoid output reads in vertex shaders
fixes a bunch of these
nv50_ir_lowering_nvc0.cpp:2878: void nv50_ir::NVC0LoweringPass::handleLDST(nv50_ir::Instruction*): Assertion `prog->getType() == Program::TYPE_TESSELLATION_CONTROL' failed.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Dave Airlie [Tue, 16 Aug 2022 05:27:47 +0000 (15:27 +1000)]
nvk: assign vertex locations according to input attrib index
This copies what lavapipe does. is there a better plan?
Fixes
dEQP-VK.draw.renderpass.simple_draw.simple_draw_triangle_list
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:53 +0000 (20:11 -0600)]
vulkan: Allow scissors or viewports to be set without counts
This is unlikely but can happen if you have the following sequence:
1. vkCmdSetScissors()
2. No pipeline bind
3. vkCmdClearImage() which causes a meta save
4. Meta restore with vk_cmd_set_dynamic_graphics_state()
In that case, we don't have scissor counts but need to restore the
scissors set with `vkCmdSetScissors()` before the meta save. We can
safely copy all of them, it's just more memory traffic than maybe we'd
like. Fortunately, this can only happen at the start of a command
buffer and only with a fairly silly sequence of commands.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:53 +0000 (20:11 -0600)]
nvk: Track and reference all device memory objects
This way bindless will work properly
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:53 +0000 (20:11 -0600)]
nvk: Don't assert when there are no attachments
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:53 +0000 (20:11 -0600)]
vulkan/meta_clear: Populate VkRenderingInfo::renderArea
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:53 +0000 (20:11 -0600)]
nvk: Implement nir_intrinsic_load_frag_coord
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:53 +0000 (20:11 -0600)]
nvk: Use the new headers for samplers
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:53 +0000 (20:11 -0600)]
nvk: Bind texture and sampler header pools for 3D
No idea if this has to be done separately for 3D vs. compute but it
shouldn't hurt anything to do it for both.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:53 +0000 (20:11 -0600)]
nvk: Zero client memory objects
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:53 +0000 (20:11 -0600)]
nvk: Use meta for CmdClear*Image
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:53 +0000 (20:11 -0600)]
nil: Switch to using the new headers for TIC entries
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:53 +0000 (20:11 -0600)]
nvk: Use max alignment for descriptor pool sizes
Since we only know the number of each descriptor type and not their
groupings, we don't know how alignments will compound. Figure out the
maximum alignment and then use that for everything to ensure that, no
matter how they're combined, we always have enough room.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Dave Airlie [Mon, 15 Aug 2022 05:56:47 +0000 (15:56 +1000)]
nvk: add indexed draw support.
This passes the
dEQP-VK.draw.renderpass.basic_draw.draw_indexed* tests for me on Turing
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Dave Airlie [Mon, 15 Aug 2022 01:56:28 +0000 (11:56 +1000)]
nvk: add some limits/features from binary driver.
These are just copied from nvidia vulkaninfo for my tu104.
I've left geom/tess stuff out for now.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Dave Airlie [Tue, 16 Aug 2022 02:27:13 +0000 (12:27 +1000)]
nvk: fix depth emission ordering.
dEQP-VK.api.image_clearing.core.clear_depth_stencil_image.single_layer.x8_d24_unorm_pack32_200x180
was hitting
nouveau_push.h:233: nvk_push_val: Assertion `last_method == idx' failed.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Dave Airlie [Tue, 16 Aug 2022 02:16:03 +0000 (12:16 +1000)]
nvk: increase graphics cpu push buffer
dEQP-VK.pipeline.monolithic.image.suballocation.sampling_type.separate.view_type.1d.format.r8g8_uint.count_4.size.512x1
was hitting 98 dwords
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Dave Airlie [Tue, 16 Aug 2022 02:15:08 +0000 (12:15 +1000)]
nvk: fix pipeline pushbuf sizing
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:53 +0000 (20:11 -0600)]
nvk: add linear image creation support.
the dEQP-VK.draw.* tests use a tiled + linear image and copy between
them.
This fixes the result gathering, not the actual result.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Dave Airlie [Mon, 15 Aug 2022 00:18:59 +0000 (10:18 +1000)]
nvk: add initial 8/16 byte clears.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:53 +0000 (20:11 -0600)]
nouveau: Fix pushbuf ref reset for user command buffers
Copy+paste error. Also, we should do the refs reset before potentialy
doing a full reset, otherwise the refs reset will assert.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:53 +0000 (20:11 -0600)]
nvk: Enable all the dynamic state features
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Wed, 26 Jul 2023 19:22:47 +0000 (14:22 -0500)]
nvk: Stop using vk_cmd_set_dynamic_graphics_state in meta_end()
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:53 +0000 (20:11 -0600)]
nvk: Use vk_meta for CmdClearAttachments
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>
Faith Ekstrand [Tue, 31 Jan 2023 02:11:53 +0000 (20:11 -0600)]
nvk: Expose a bind_vertex_buffer helper
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24326>