Pass the robustness state to nir_lower_descriptors() instead of using
the default robustness set at device creation time.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31251>
On Valhall, the nir_lower_explicit_io(ssbo) pass can add
get_ssbo_size() intrinsics. Make sure those are lowered to load_ubo()s
on the targeted buffer descriptor.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31248>
The source/destination was inverted. While at it, make this copy
conditional on tlsinfo.tls.size > 0.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31248>
The cache flush reduction mechanism relies on the flush ID to
conditionally eliminate cache flush requests if another cache flush
happened between the time the flush ID was retrieved and the time
the flush operation happens.
Store the current flush ID at EndCommandBuffer() time, and pass it
back to the submit ioctl().
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31230>
The cs_while() in cs_render_desc_ringbuf_move_ptr() was never meant to
be a while() loop, but we didn't have a cs_if() helper back then.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com>
Reviewed-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31205>
The logic is a bit more involved than with other blocks to keep
cs_if/else() declaration consistent with the rest. This forces us to
keep track of the last end_if() so we can flush it if the next CS
[pseudo-]instruction is not a cs_else().
These changes require quite a bit of code motion to avoid forward
declarations.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com>
Reviewed-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31205>
We don't need a double-linked list to manage our block stack. Use a
single-link list instead, which simplifies things a bit and hopefully
lets the compiler optimize things a bit more.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com>
Reviewed-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31205>
The current block is just the top of the stack, so let's just provide
an helper that returns the current block and drop the cur field.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com>
Reviewed-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31205>
This allows us to check that the block to end is at the top of the
stack.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com>
Reviewed-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31205>
Always pass the loop object around so we can check that the current
block is the loop block, and provide syntactic sugor for unconditional
continue/break statements.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com>
Reviewed-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31205>
If zero instructions are requested, we should return NULL, but there's
no good reason to accept this case in the first place, so let's assert
that num_instrs > 0 instead.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com>
Reviewed-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31205>
assert that a block is no bigger than a chunk allocated by
::alloc_buffer().
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com>
Reviewed-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31205>
If cs_alloc_ins() fails, it returns a dummy instruction slot, which can
only hold one instruction. Make sure we skip the memcpy() if the CS
is invalid to avoid a buffer overflow.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Lars-Ivar Hesselberg Simonsen <lars-ivar.simonsen@arm.com>
Reviewed-by: Louis-Francis Ratté-Boulianne <lfrb@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31205>
The `add_global*_arguments` family of functions affect subprojects, and
are therefore generally discouraged from being used.
Fixes: a42c2293ab ("wsi: Metal support")
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31267>
In commit 41e83b243c (meson: remove dri-search-path, 2024-09-18) the
`dri-search-path` meson option was removed but its usage in the Android
build files was left there, resulting in the following error when
building for Android:
-----------------------------------------------------------------------
FAILED: out/target/product/vsoc_x86_64/obj_x86/MESON_MESA3D/.build.timestamp
...
meson.build:4:0: ERROR: Unknown options: "dri-search-path"
-----------------------------------------------------------------------
Remove the usage of the option when building for Android, and while at
it remove also any mention of if in docs/drivers/llvmpipe.rst
Remove also the MESA3D_LIB_DIR variable which was used only for
dri-search-path.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/31266>
vk_common_QueueWaitIdle() creates a syncobj, does a submit with no
batch buffers what translates to execute trivial_batch_bo and then
waits for syncobj to be signaled when trivial_batch_bo finishes.
On Xe KMD on other hand we can avoid the trivial_batch_bo submission
and instead use the special DRM_IOCTL_XE_EXEC with num_batch_buffer == 0
to get a syncobj to be signaled when the last exec finish execution.
This should free a bit GPU to execute more important workloads.
This will also optimize vkDeviceWaitIdle() that calls QueueWaitIdle().
It have to fallback to vk_common_QueueWaitIdle() when queue is in
VK_QUEUE_SUBMIT_MODE_THREADED mode because vkQueueWaitIdle()
could return but there still stuff in VK/CPU submission queue.
Also it could cause use after free when resources attached to
submission are freed before it is processed, example:
vkCreateFence() or vkCreateSemaphore()
vkQueueSubmit() // with Fence or Semaphore created above
vkQueueWaitIdle() // with the race it returns
vkDestroyFence() or vkDestroySemaphore()
// vk_queue_submit_thread_func() start to process submission above...
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30958>
Split anv_xe_wait_exec_queue_idle() into 2 functions, the first
function creates the syncobj and prepares it to be signaled when the
last workload in queue is completed.
And the second one that calls the first function, then waits for the
syncobj to be signaled and destroy the syncobj.
The main reason for that is that the first function can be reused in
Iris and a future patch will add another user, so lets share it.
No changes in behavior are expected here.
Reviewed-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30958>
Fixes multiple piglit tests, except (u)long/char hadd,
which now fails instead of crashing.
Acked-by: Eric R. Smith <eric.smith@collabora.com>
Signed-off-by: David Heidelberg <david@ixit.cz>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30963>
We currently disable CCS if a 3D Ys/Yf surface uses miptails. However,
ISL generally configures surfaces to be compatible with compression. For
consistency, disable miptails on 3D Ys/Yf surfaces in order to allow
compression.
If drivers prefer to have a more compact layout, they can pass the
ISL_SURF_USAGE_DISABLE_AUX_BIT flag at surface creation time.
Reviewed-by: Rohan Garg <rohan.garg@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30081>
We currently disable CCS if a surface uses more than 11 slots in a
miptail. However, ISL generally configures surfaces to be compatible
with compression. For consistency, reduce the number of slots used in
miptails in order to allow compression.
If drivers prefer to have a more compact layout, they can pass the
ISL_SURF_USAGE_DISABLE_AUX_BIT flag at surface creation time.
Reviewed-by: Rohan Garg <rohan.garg@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30081>
This adds gfxstream-vk as a vulkan driver to Mesa. It
will be used in the following places:
- Android Emulator
- Fuchsia Emulator
- Cloud Android
- AAOS reference hardware
- [a few other places]
meson amd64-build/ -Dvulkan-drivers="gfxstream-experimental" -Dgallium-drivers="" -Dopengl=false
Reviewed-by: Aaron Ruby <aruby@blackberry.com>
Acked-by: Yonggang Luo <luoyonggang@gmail.com>
Acked-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27246>
These functions are now in the Mesa common
code.
Reviewed-by: Aaron Ruby <aruby@blackberry.com>
Acked-by: Yonggang Luo <luoyonggang@gmail.com>
Acked-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27246>
Partial update to vulkan headers and auto generated code to
include the new extension. It will be used to fix issues on
external memory implementation with MoltenVK later on.
References:
Khronos MR 6800
MoltenVK PR 2314
Reviewed-by: Aaron Ruby <aruby@blackberry.com>
Acked-by: Yonggang Luo <luoyonggang@gmail.com>
Acked-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27246>
This simplifies the amount of code imported from Vulkan
docs. It vendors the code to just a subset needed by
the cerealgenerator. The files that are kept are:
- generator.py
- cgenerator.py
- reg.py
- genvk.py
Since these files originate with Khronos, they are
Apache licensed.
Long-term, there are various ideas on how to proceed
with codegen. Probably the above files can be nuked
in the event some of those ideas come to pass.
Reviewed-by: Aaron Ruby <aruby@blackberry.com>
Acked-by: Yonggang Luo <luoyonggang@gmail.com>
Acked-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27246>
While cereal implements an API provided by Khronos Apache 2.0
files, the files were started in AOSP, and can be relicensed to
MIT.
Reviewed-by: Aaron Ruby <aruby@blackberry.com>
Acked-by: Yonggang Luo <luoyonggang@gmail.com>
Acked-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27246>
Re-license files to MIT, by popular demand. These files are have
an origin in AOSP.
Reviewed-by: Aaron Ruby <aruby@blackberry.com>
Acked-by: Yonggang Luo <luoyonggang@gmail.com>
Acked-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27246>
genvk.py itself includes different types of generators:
spirvgenerator.py, rubygenerator.py, etc. cerealgenerator.py
is an additional one. A strategy to pare down the size of
the code would be modify genvk.py to remove unneeded generators,
and just import the cerealgenerator: i.e, importing the subset
of scripts useful to gfxstream.
The removes the notion of upstreaming to vulkan-docs, but does
imply more dedicated codegen strategy long-term.
About 300kLoC can be removed, since previously we just a blanket
copy.
Reviewed-by: Aaron Ruby <aruby@blackberry.com>
Acked-by: Yonggang Luo <luoyonggang@gmail.com>
Acked-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27246>
Works a little better with the guest Linux build..
Reviewed-by: Aaron Ruby <aruby@blackberry.com>
Acked-by: Yonggang Luo <luoyonggang@gmail.com>
Acked-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27246>
When invoking the build in the guest, we don't need
to define this.
Reviewed-by: Aaron Ruby <aruby@blackberry.com>
Acked-by: Yonggang Luo <luoyonggang@gmail.com>
Acked-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27246>
This reverts commit dd5c7b6aeb41270f0b457c18a2a31cf136562bf0.
Reason for revert: b/366153044
Reviewed-by: Aaron Ruby <aruby@blackberry.com>
Acked-by: Yonggang Luo <luoyonggang@gmail.com>
Acked-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27246>
