Unfortunately it's not as simple as passing the indirect draw count
buffer to ExecuteIndirect. The compute job that populate the execute
buffer also needs to know the number of entries that need to be
patched. Instead of transitioning the indirect count buffer from
GENERIC_READ to INDIRECT_ARGUMENT we just keep at as a read-only
resource and copy the draw_count value to the exec buffer in the
compute job.
Acked-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15914>
The key is created on stack, so as soon as the function returns this key
is lost, so the inserted key in the hashtable is invalid.
Rather, insert a duplicated version on heap.
This fixes a stack-buffer-overflow when running some Vulkan CTS tests.
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16083>
pipe_context shouldn't have functions that return values because this
prevent multithreading.
Move this hook to pipe_screen.
Fixes: 606e42027e ("gallium: add hook on getting canonical format")
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16078>
The core provides generic render-pass -> dynamic-rendering wrappers,
so let's rely on them instead of implementing our own logic.
Suggested-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15913>
This is how D3D12 works, so nothing really complicated here. We just
add a bunch of states to the render state, patch the pipeline building
logic to support dynamic rendering, and add new functions to
Suggested-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15913>
Let the caller extract the image view from the attachment index, so
we can make this function framebuffer-agnostic and re-use it in the
dynamic rendering path.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15913>
We don't need the first_vertex argument when turning non-indexed
triangle-fan draws into indexed triangle-list ones, the firstVertex
argument can simply be passed as a vertexOffset.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15912>
these are translated into memory+control barriers in nir, and only
the control barrier needs to be handled
these semantics match what glslang does, so they must be right
cc: mesa-stable
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15959>
this is semi transposed from the dri version of kopper and mostly ignores
what wgl tries to do in order to let zink drive everything
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16068>
The Iris code that deals with implicit tracking is protected by
bufmgr->bo_deps_lock. Before this patch, we hold this lock during
update_batch_syncobjs() but don't keep it held until we actually
submit the batch in the execbuf ioctl. This can lead to the following
race condition:
- Context C1 generates a batch B1 that signals syncobj S1.
- Context C2 generates a batch B2 that depends on something that B1
from C1 is using, so we mark B2 as having to wait syncobj S1.
- C2 calls submit_batch() before C1 does it.
- The Kernel detects it was told to wait on syncobj S1 that was
never even submitted, so it returns EINVAL to the execbuf ioctl.
- We run abort() at the end of _iris_batch_flush().
- If DEBUG is defined, we also print:
iris: Failed to submit batchbuffer: Invalid argument
I couldn't figure out a way to reproduce this issue with real
workloads, but I was able to write a small reproducer to trigger this.
Basically it's a little GL program that has lots of contexts running
in different threads submitting compute shaders that keep using the
same SSBOs. I'll submit this as a piglit test. Edit: Tapani found a
dEQP test case which fails intermintently without this fix, so I'm not
sure a new Piglit is worth it now.
The solution itself is quite simple: just keep bo_deps_lock held all
the way from update_batch_syncobjs() until ioctl(). In order to make
that easier we just call update_batch_syncobjs() a little later. We
have to drop the lock as soon as the ioctl returns because removing
the references on the buffers would trigger other functions to try to
grab the lock again, leading to deadlocks.
Thanks to Kenneth Graunke for pointing out this issue.
This has also been confirmed to fix a dEQP test that was giving
intermittent failures:
dEQP-EGL.functional.sharing.gles2.multithread.random.images.copyteximage2d.12
v2: Move decode_batch() out, just to be safe (Jason).
v3: Do it all after assembling validation_list (Ken).
Cc: mesa-stable
Fixes: 89a34cb845 ("iris: switch to explicit busy tracking")
Tested-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14964>
With !8044 all TGSI drivers will end up going through the nir-to-tgsi
path, so make the switch now that CI is happy (which will also make sure
that future NTT work doesn't break virgl).
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13295>
With nir-to-tgsi, virgl saw a case where a previous declaration of array
.x and scalar .y (turning into an array with .xy) ended up being a
declaration of scalar .x and array .y (turning into a scalar with .xy).
Make sure we extend the declared array length as well.
One might think that the fix would be to union the .first/.last between
the two declarations being merged, but note that ureg_DECL_output() passes
in the current nr_output_regs as the index, so the .last would end up
getting extended for those callers (such as nir_to_tgsi fs outputs) every
time you merged.
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13295>
We can't declare input arrays because mesa/st lowers NIR VS output
declarations to elements no matter what, and virgl has depended on
matching array sizes of declarations between producers and consumers. So,
we have to lower it away (which is fine because hardware drivers will
generally be lowering anyway).
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13295>
The output may be an array, e.g. with a TCS shader, so if the
value is written to a temporaray first remove the indirect
indicator for that write.
Fixes: 36f12c85c8
virgl: Extend integer write out output fix to all non-move integers ops
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13295>
