The vma_samplers vma heap is initialized unconditionally. Don't use
device->physical->indirect_descriptors as a condition on whether to
free it or not.
From my TGL machine:
==373617== 32 bytes in 1 blocks are definitely lost in loss record 1 of 1
==373617== at 0x48459F3: calloc (in /usr/libexec/valgrind/vgpreload_memcheck-amd64-linux.so)
==373617== by 0x6926DC0: util_vma_heap_free (vma.c:339)
==373617== by 0x6925ED3: util_vma_heap_init (vma.c:53)
==373617== by 0x5334EDA: anv_CreateDevice (anv_device.c:3404)
==373617== by 0x685593A: vk_tramp_CreateDevice (vk_dispatch_trampolines.c:78)
==373617== by 0x48A6D56: terminator_CreateDevice (loader.c:5833)
==373617== by 0x9C2293F: vulkan_layer_chassis::CreateDevice(VkPhysicalDevice_T*, VkDeviceCreateInfo const*, VkAllocationCallbacks const*, VkDevice_T**) (chassis.cpp:497)
==373617== by 0x48B0690: loader_create_device_chain (loader.c:4937)
==373617== by 0x48B1327: loader_layer_create_device (loader.c:4317)
==373617== by 0x48B8D79: vkCreateDevice (trampoline.c:1004)
==373617== by 0x10CC7A: MyApp::MyApp(int, bool) (sparse.cpp:608)
==373617== by 0x1201E8: main (sparse.cpp:6025)
Fixes: 7c76125db2 ("anv: use 2 different buffers for surfaces/samplers in descriptor sets")
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28303>
Looks like `OCL_ICD_VENDORS` is meant to be a directory containing
`.icd` files, but we were giving it a path to a driver. This meant that
rusticl wouldn't get picked up when using devenv.
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28231>
Extend the vkGetPhysicalDeviceFormatProperties2 cache to include
VkFormatProperties3 from the pNext chain. VkFormatProperties3 was
observed being always attached for DXVK and thus skipping the cache
if not handled.
Signed-off-by: Juston Li <justonli@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28194>
This was missing, this is implemented in common code.
Signed-off-by: Joshua Ashton <joshua@froggi.es>
Acked-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Acked-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28275>
This was missing, this is implemented in common code.
Signed-off-by: Joshua Ashton <joshua@froggi.es>
Acked-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28275>
try_merge_merge_set() expects equal_anc_out to be Temp() in the beginning,
so we need to reset it in case it's used again.
Fixes compilation of metro_exodus/163b3b895730d37b with
VK_PIPELINE_CREATE_2_DISABLE_OPTIMIZATION_BIT_KHR.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Fixes: 18ba93e673 ("aco/cssa: rewrite lower_to_cssa pass")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28248>
A seperate meson project with mesa as a subproject was required to
compile the generate_rd executable for replaying rddecompiler
generated source. That approach was overly complicated, now the
generate_rd executable is now compiled as a part of building the
freedreno tools with a blank generate-rd.cc file that can be replaced
with the source generated by rddecompiler.
Signed-off-by: Mark Collins <mark@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28253>
It's useful to clear buffers at the start of sequences to view the
delta, this adds that functionality to wrbufs with a fixed clear
value of 0xDEADBEEF.
Signed-off-by: Mark Collins <mark@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28253>
It's possible for large allocations to hit the maximum address
space size especially with a fake AS, these failures are silent
and can cause a hard to debug segfault later down the line.
Signed-off-by: Mark Collins <mark@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28253>
We should be careful to not read past the end of any buffers when
dumping wrbufs, this clamps the size to the size of the buffer with
a warning.
Signed-off-by: Mark Collins <mark@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28253>
When computing layout for large extents and array size, the computations
can overflow the 32-bit variables holding these values. This can end up
underreporting memory requirements for a given resource, making the
application think that such a resource is allocatable.
The size-holding variables in the fdl_layout struct have their type
upgraded to uint64_t, as does the total size variable in tu_image. This
avoids problems in corner-case tests in the Vulkan CTS, but this code
should be improved further to avoid overflows and narrowing conversions.
Fixes a quartet of tests under
dEQP-VK.pipeline.monolithic.render_to_image.core.2d_array.huge.width_height_layers.
Signed-off-by: Zan Dobersek <zdobersek@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28050>
brw_fs_opt_eliminate_find_live_channel eliminates FIND_LIVE_CHANNEL
outside of control flow. None of our optimization passes generate
additional cases of that instruction, so once it's gone, we shouldn't
ever have to run the pass again. Moving it out of the loop should
save a bit of CPU time.
While we're at it, also clean adjacent BROADCAST instructions that
consume the result of our FIND_LIVE_CHANNEL. Without this, we have
to perform copy propagation to get the MOV 0 immediate into the
BROADCAST, then algebraic to turn it into a MOV, which enables more
copy propagation...not to mention CSE gets involved. Since this
FIND_LIVE_CHANNEL + BROADCAST pattern from emit_uniformize() is
really common, and it's trivial to clean up, we can do that. This
lets the initial copy prop in the loop see MOV instead of BROADCAST.
Zero impact on fossil-db, but less work in the optimization loop.
Together with the previous patches, this cuts compile time in
Borderlands 3 on Alchemist by -1.38539% +/- 0.1632% (n = 24).
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28286>
It's a logical opcode which is lowered to a send-from-GRF later. That
lowering code is responsible for ensuring the sources are set up in a
proper SEND payload.
This was preventing copy propagation of surface handles which started
out as scalars, were splatted out to full-SIMD values with NoMask, then
actually consumed as only component 0 (scalar again), because we thought
that scalar values were not allowed.
fossil-db on Alchemist shows improvements in q2rtx but no other titles:
Totals:
Instrs: 161310436 -> 161310152 (-0.00%)
Cycles: 14370605159 -> 14370601066 (-0.00%)
Totals from 17 (0.00% of 652298) affected shaders:
Instrs: 16097 -> 15813 (-1.76%)
Cycles: 185508 -> 181415 (-2.21%)
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28286>
We don't necessarily want to split up MOVs for 64-bit addresses into
2x 32-bit MOVs right away, as this makes things like copy propagating
the whole address around harder. We should do this late, once, while
still doing other algebraic optimizations earlier.
fossil-db results for Alchemist show tiny improvements:
Totals:
Instrs: 161310502 -> 161310436 (-0.00%); split: -0.00%, +0.00%
Cycles: 14370605606 -> 14370605159 (-0.00%); split: -0.00%, +0.00%
Totals from 33 (0.01% of 652298) affected shaders:
Instrs: 15053 -> 14987 (-0.44%); split: -0.64%, +0.20%
Cycles: 196947 -> 196500 (-0.23%); split: -0.25%, +0.02%
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28286>
Pci revision was included in the shader cache key because it can
enable platform workarounds. While some platform workarounds exist in
the compiler, none are dependent on the silicon stepping.
Many platforms differ only in the pci revision id, causing needless
duplication in cache entries between platforms.
When a platform ships publicly with stepping-specific compiler
workarounds, pci id must be incorporated into the shader cache key.
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28085>
This is to make it more intuitive and also consistent
with last_writer_idx which does allow constant operands.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28046>
If the presentation fence was signalled and we still hold
max_acquires or more images, then clear the timeout to avoid
a possible deadlock.
With that we avoid the validation error
VUID-vkAcquireNextImageKHR-surface-07783
triggered by piglit
spec@!opengl 1.0@gl-1.0-drawbuffer-modes
and others.
v2: clear timeout only if we have acquired more images than the
reported max and add some comment why the timeout is cleared
(Mike).
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28245>
Those opcodes are vestige of the old command format.
This implement handling of them and fix issues when analysing command
buffers that use thoses.
Signed-off-by: Mary Guillemard <mary.guillemard@collabora.com>
Reviewed-by: Karol Herbst <kherbst@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28277>
Emitting UNDEF is only necessary when the instructions we generate to
produce the NIR def are considered partial writes. By adding a simple
check (adapted from fs_inst::is_partial_write()), we can avoid creating
loads of unnecessary UNDEFs that we have to clean up later.
Our first dead code elimination pass does get rid of them pretty
quickly, but this should save memory and time during our first
split_virtual_grfs and dead_code_elimination passes.
This generates roughly 30% fewer instructions at the beginning.
Improves compilation time of shaders:
- Rise of the Tomb Raider: -3.51563% +/- 0.103951% (n=7)
- Borderlands 3: -3.64422% +/- 0.300951% (n=7).
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28169>
Outside SIMD1 instructions, a destination stride of zero doesn't make
any sense. When such strides exist, they would be fixed by the FS
generator. Currently the only place that intentionally generates such a
stride is setup_barrier_message_payload_gfx125, and this commit changes
that.
The existence of a zero stride that won't really be a zero stride causes
a variety of problems with other optimization passes. Those passes don't
know that 0 actually means 1, and they make incorrect assumptions about
sizes written, etc.
The assertion helped catch many bugs in some other work in progress that
tries to store convergent values in SIMD8 registers regardless of the
dispatch width. That code would accidentally generate destination
strides of zero.
v2: Check stride differently depending on register file. Suggested by
Caio.
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28256>
