d72f7cbc5a
zink: stop leaking inferred resolve surfaces
...
this is tricky since the ownership model here works the other
way around from normal: the resolve surface's lifetime is
determined by the resource's lifetime, meaning that
the surface should be destroyed by the resource instead
of the resource being destroyed by the surface
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31124 >
2024-09-11 21:07:45 +00:00
a3c2be8f85
zink: set resolve layout when inferring scanout resolves
...
this is otherwise broken
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31124 >
2024-09-11 21:07:45 +00:00
db371633e6
panfrost: improve swizzle inversion
...
panfrost_invert_swizzle produces a (one-sided) inverse, that is a
function g such that g(f(x)) = x. However, it is actually used in
some cases where we want a two-sided inverse, where we also have
f(g(x)) = x. An example is its use to pre-swizzle border color
values so that a later pass will un-swizzle them. If the swizzle
is not one-to-one this two sided inverse does not exist. However,
we can do better than the original code, e.g. for an RRR1 swizzle
inverse produced was originally B000, which when applied on the wrong
side results in BBB1 as output, whereas R000 would produce the
desired RRR1 output. Using the first valid component we see, rather
than the last one, is thus usually better.
The "correct" solution is to re-write all the code that uses
an inverse to handle non-unique inverses. But frankly these uses only
crop up in fairly niche cases like tests, and it's probably not worth
spending a lot of effort to deal with these edge cases when this
patch fixes most of them.
Fixes some failing piglit ext_framebuffer_multisample tests.
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31060 >
2024-09-11 19:29:16 +00:00
78a121b8cf
freedreno/ir3: Use nir_lower_mem_access_bit_sizes instead custom lowering
...
- More robust.
- Handles properly UBO cases, needed for proper OpenCL support (rusticl).
- Resolved KHR-GL46.gpu_shader_fp64.fp64.max_uniform_components failure.
Fixes: f5ce806ed7robdclark@freedesktop.org >
Co-authored-by: Rob Clark <robclark@freedesktop.org >
Signed-off-by: David Heidelberg <david@ixit.cz >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30961 >
2024-09-12 01:04:58 +09:00
5db135f66a
panvk: Advertize v10 support
...
Patch panvk_macros.h to add a case for v10 and accept v10 devices.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
f8598e599e
panvk: Add VkEvent support to the CSF backend
...
With its sync object primitive, CSF allows native event support.
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
5544d39f44
panvk: Add a CSF backend for panvk_queue/cmd_buffer
...
With those two components implemented, we can now compile all common
per-arch source files.
Co-developed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
7049d31676
panvk: Add a read-write non-cached memory pool for CSF events
...
Events cross the device boundary and can be read/written by the host.
We could add the necessary cache flush and made those GPU-cached, but
it's way simpler to allocate non-cached memory.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
53fb1d99ca
panvk: Transition to explicit VA assignment on v10+
...
Panthor allows us to control the GPU virtual address space,
which, among other things, will allow us to support sparse
bindings.
Let's tweak the device initialization and BO mapping to
takes this explicit VA management mode into account.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
ca79fc99cd
panvk: Don't pretend we can compile for v9
...
The JM backend is not ready for v9 yet, so let's just drop it from the
list of architecture to build for.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
db1073357a
panvk: Be robust against allocation failures in a command buffer context
...
When an allocation failure happens, the command buffer should be
flagged as invalid, and anything using this memory from the CPU side should
be skipped to avoid segfaults.
For allocations going through memory pools owned by a command buffer
we automate that with panvk_cmd_alloc_xx() macros.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
eb0b1c36f2
panvk: Move CmdPushConstants2KHR to panvk_vX_cmd_push_constant.c
...
Turns out the panvk_cmd_buffer structures for CSF/JM are similar enough
to make CmdPushConstants2KHR() generic. Let's move this function out
of the JM specific code.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
173939ed9b
panvk: Move cmd_bind_shaders() to panvk_vX_shader.c
...
We had planned for a CSF specialization, but it turns out the CSF/JM
are almost the same and we can have something generic implemented in
panvk_vX_shader.c
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
0426c14c7a
panvk: Move Cmd{Bind,Push}DescriptorSets implementation to cmd_desc_state.c
...
Turns out the panvk_cmd_buffer for CSF/JM have enough similarities to
make it possible to have a generic implementation in
panvk_vX_cmd_desc_state.c.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
aedb00ca08
panvk: Set the submit mode to THREADED_ON_DEMAND
...
It's needed for wait-before-submit, which drm_syncobjs don't
support.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
0429cc10d8
panvk: Prepare the panvk_device logic for CSF
...
The tiler heap is handled differently on CSF hardware, and we don't need
to allocate the tiler_heap buffer there.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
be96213634
panvk: Make sure push sets are reset when picked from the CommandPool
...
If we don't do that, we might end up with data from the previous user
of this set.
Fixes: 598a8d9d11boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
7f7e772bdf
panvk: Don't make panvk_device size gen dependent
...
The panvk_device is used in gen-agnostic path, and which means we can't
make its size dependent on PAN_ARCH.
Fixes: 5067921349boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
6527cb742e
panvk: Don't open-code vk_index_type_to_bytes()
...
While at it, store the index_size in bytes instead of bits.
Suggested-by: Alyssa Rosenzweig <alyssa@rosenzweig.io >
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
9f094cdeaf
panvk: Fix valhall_pack_buf_idx()
...
The descriptor lowering pass can add UBO loads to read descriptors, and
these loads already assume a nir_address_format_32bit_index_offset
format.
Bail out if we detect that case in valhall_pack_buf_idx().
Reported-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Fixes: 27beadcbdbboris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
343bff7903
panvk: Drop the NO_MMAP flag when PANVK_DEBUG=trace
...
We have buffers that are not supposed to be CPU-mapped but are useful to
check when tracing is enabled.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
bae5235405
panvk: Kill the set_dirty() macro
...
It's unused and we're not supposed to manually set the dirty bits
anyway.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
a651826b80
panvk: Fix the varying pool label
...
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
2805804dce
pan/desc: Make vertex_count a Midgard field
...
We don't need to pass the vertex count around on Bifrost/Valhall, since
the tiler descriptor is generated outside of emit_fbd() on those gens.
Let's move vertex_count under the midgard struct to make that clear.
This implies maintaining a vertex_count at the panfrost_batch level
in the gallium driver, but that's fine.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:56 +00:00
11fcb23f74
pan/desc: Add a struct for valhall/bifrost to the union in pan_tiler_context
...
Valhall has extra tiler parameters for multilayer rendering that we will
need for the framebuffer descriptor emission. Let's add proper struct
for Valhall and Bifrost instead of assuming a mali_ptr is all we'll ever
need.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:55 +00:00
2149a04de3
pan/kmod: Don't cap VM bind operations to one
...
Looks like some leftovers from a debugging session.
Fixes: 97f6a62f7eboris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:55 +00:00
07b31b2204
pan/cs: Add a facility to restrict register access
...
If the CS register file is used in a stateful mode, some of the registers
will survive the RUN_xxx boundary, and even serve as a context for
subsequent RUN_xxx calls.
Having a way to restrict register access outside specific sections of
code is a handy debug feature, so let's expose the necessary bits to
allow that.
It takes the form of a reg permission callback that gets called every
time a register serves are a source or destination, to make sure this
is allowed in this context.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:55 +00:00
07c463d06b
pan/cs: Add inline specifiers to a bunch of cs_ helpers
...
Just to make things consistent with the rest of the file, the compiler
was probably already inlining those anyway.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:55 +00:00
a87c43def7
pan/cs: Add an optional load/store tracker to make sure we don't omit waits
...
With load/store operations being asynchronous, it's pretty easy to omit
a WAIT(LS) and end up with inconsistent register values or corrupted
stores.
Add an optional load/store tracker to detect such cases early on.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:55 +00:00
b1934057cb
pan/cs: Allow sparse register set passed to loads/stores
...
When it comes to copying stuff around, it becomes handy to be able
to have a discontiguous range of register to load to/store from so
we don't have to move stuff around to make things contiguous in the
register file.
Allow that by making register count the last bit in the mask rather
than counting bits.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:55 +00:00
1a5ce4a96e
pan/cs: Add helpers to emit match (AKA switch-case) blocks
...
We need for the dynamic scoreboard assignment we will do in panvk.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:55 +00:00
8f825a51a8
pan/cs: Accept val=undef when cs_branch_label() condition is ALWAYS
...
The register is not used in that case anyway, so we can pick r0 and
call it a day.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:55 +00:00
f6c3fb11ea
pan/cs: Make sure we're not passed cs_now() to asynchronous instructions
...
FLUSH_CACHE2 is both deferrable and asynchronous. That means we should
always pass an explicit slot to signal instead of picking zero as we do
in cs_now().
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com >
Reviewed-by: Rebecca Mckeever <rebecca.mckeever@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com >
Reviewed-by: John Anthony <john.anthony@arm.com >
Acked-by: Erik Faye-Lund <erik.faye-lund@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30969 >
2024-09-11 12:58:55 +00:00
07b1fd64b7
aco/live_var_analysis: don't set lateKill for p_interp_gfx11 m0 operand
...
There is no danger of overwriting m0 anymore.
Fixes: dfc13fcf9fhttps://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31127 >
2024-09-11 10:36:34 +00:00
1aa7218a1f
mesa: Reset vbo attributes after flushing them to Current in glPopAttrib
...
Fixes: 2fe771f4https://gitlab.freedesktop.org/mesa/mesa/-/issues/11498
Signed-off-by: Sviatoslav Peleshko <sviatoslav.peleshko@globallogic.com >
Reviewed-by: Marek Olšák <marek.olsak@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31107 >
2024-09-11 09:58:52 +00:00
a9d64fa1bd
radv: promote VK_NV_compute_shader_derivatives -> VK_KHR_compute_shader_derivatives
...
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30972 >
2024-09-11 08:37:46 +00:00
b6faf586e6
targets/vdpau: Build vdpau driver into libgallium when building with dri
...
Currently vdpau driver and dri driver are two separate libraries, when
radeonsi is enabled both libraries contain amdgpu winsys. radeonsi
needs shared winsys to ensure sync between OpenGL and VDPAU works.
Build vdpau driver into libgallium to avoid having two instances of amdgpu
winsys.
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31079 >
2024-09-11 08:07:21 +00:00
212d57f7e6
targets/va: Build va driver into libgallium when building with dri
...
Currently va driver and dri driver are two separate libraries, when
radeonsi is enabled both libraries contain amdgpu winsys. radeonsi
needs shared winsys to ensure sync between OpenGL and VAAPI works.
Build va driver into libgallium to avoid having two instances of amdgpu
winsys.
Cc: mesa-stable
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/11598
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31079 >
2024-09-11 08:07:20 +00:00
39f7ed5412
meson/megadriver: Add megadriver_libdir argument
...
To be able to install symlinks to megadriver in different directory.
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31079 >
2024-09-11 08:07:20 +00:00
7ebb2d1dc8
nvk: Only advertise compute_shader_derivatives when NAK is used
...
Fixes: e0e7d8d910https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31125 >
2024-09-10 19:27:10 -05:00
6f82b1ea18
d3d12: Add support for HEVC range extension header parameters
...
Reviewed-By: Jesse Natalie <jenatali@microsoft.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31117 >
2024-09-10 23:06:34 +00:00
92d6989fdc
d3d12: Add support for HEVC 4:4:4 video encode using AYUV format
...
Reviewed-By: Jesse Natalie <jenatali@microsoft.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31117 >
2024-09-10 23:06:34 +00:00
0161b98955
pipe: Add HEVC encode range extension caps, SPS and PPS params
...
Reviewed-By: Jesse Natalie <jenatali@microsoft.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31117 >
2024-09-10 23:06:34 +00:00
ec59a1d1b2
etnaviv: Drop assignment of get_timestamp
...
This hook is only called when PIPE_CAP_QUERY_TIMESTAMP is supported. Let's
drop it as it could confuse other mesa hackers.
Signed-off-by: Christian Gmeiner <cgmeiner@igalia.com >
Reviewed-by: Christopher Michael <cmichael@igalia.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31094 >
2024-09-10 22:37:13 +00:00
e4ee0a2ce1
anv: be consistent about aux usage with modifiers
...
In c1a7d520f3c1a7d520f3nanley.g.chery@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31051 >
2024-09-10 22:09:41 +00:00
3db1a91bfe
zink: delete erroneous kopper assert
...
certain present modes aren't supported by zink,
but that doesn't mean the driver needs to crash if
the driver supports them
cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31121 >
2024-09-10 21:38:41 +00:00
e0e7d8d910
nvk: Advertise VK_NV/KHR_compute_shader_derivatives
...
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31105 >
2024-09-10 20:13:32 +00:00
b5ac20f13f
nak: Shuffle gl_LocalInvocationID for quad derivatives
...
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31105 >
2024-09-10 20:13:32 +00:00
9c4fa79e00
nak: Move subgroup_id and num_subgroups to lower_system_values
...
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31105 >
2024-09-10 20:13:32 +00:00
4602b52a1c
nak: Add a NAK_SUBGROUP_SIZE #define
...
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31105 >
2024-09-10 20:13:32 +00:00
Mike Blumenkrantz