On GFX10.3, VRS rates need to be copied to the HTILE buffer but in some
situations, like for mips, it's not always possible to enable HTILE.
In this case, we can fallback to our internal HTILE buffer and tweak
the depth/stencil registers to use this HTILE buffer.
This fixes a bunch of VRS crashes on GFX10.3.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26025>
The main change is to use struct radv_vs_prolog_key directly instead of
the compressed representation to simplify an upcoming rework in prolog /
epilog caching. In doing so the state struct pointer was replaced with
an inline struct.
Care was also taken to pre-mask all the states with the active attribute
mask and other masks when it makes sense; this ensures that we don't
accidentally use information not hashed into the key during compilation.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26023>
This was broken as the field was never assigned to. This will also be
dropped from the upcoming prolog/epilog lookup rework, as it adds to
code complexity while the benefit of saving one hash table memory access
seems questionable.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26023>
It's not possible when they aren't all inlined because they need to be
copied to the upload BO and the DGC shader also copies the ones that
come from the indirect layout.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25935>
Some copy operations are poorly supported by the SDMA hardware,
meaning that the built-in packets don't support them, so we will
need to work around that by copying to and from a temporary BO.
The size of the temporary buffer was chosen so that it can fit
at least one full pixel row of the largest possible image.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Tatsuyuki Ishi <ishitatsuyuki@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25831>
The following sequence is valid (although weird) but many other drivers
(including RADV) were broken:
- bind pipeline with some static state
- set state command for that static state (to a bad value)
- bind the same pipeline again
- draw
Fixes new dEQP-VK.dynamic_state.*.double_static_bind.
Cc: mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25954>
This used to be a PS_PARTIAL_FLUSH to fix synchronization issues with
NGG streamout on RDNA2 but it's no longer needed for RDNA3. It's
already synchronized in CmdEndTransformFeedbackEXT().
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25903>
This was useful for experimenting it on RDNA2 and during RNDA3 bringup,
but now the support is rock solid on RDNA3 and it's useless to keep the
RADV_PERFTEST=ngg_streamout option.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25903>
These functions implement the possibility to have the
gang leader wait for its follower.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25796>
Create a version of this function that takes a CS and queue family.
move it to radv_cs.h so it can be called from multiple other files.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25770>
According to RadeonSI, this is required for preemption, user queues,
and we only have to wait for VS after streamout which should be more
performant.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25284>
Primitive restart is also applied to non-indexed draws on AMD GPUs. On
GFX11, DISABLE_FOR_AUTO_INDEX can be set but we will need a different
solution for older GPUs.
This fixes all line related flakes in CI (at least).
Cc: mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25639>
In a scenario like:
CmdBindTransformFeedbackBuffers()
BeginTransformFeedback()
CmdDraw() --> streamout descriptors emitted
EndTransformFeedback() --> streamout descriptors emitted as 0 (disabled)
CmdDraw()
BeginTransformFeedback()
CmdDraw() --> streamout descriptor not re-emitted
EndTransformFeedback()
Fix this by re-emitting streamout descriptors when streamout is
enabled/disabled because a buffer size of 0 acts like a disable bit.
This fixes dEQP-VK.transform_feedback.simple.backward_dependency_indirect*
on NAVI31.
Cc: mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25583>
I don't know if these can be done properly, but for now just don't
emit the standard cp stuff since it hangs the GPU.
"Fixes" dEQP-VK.video.synchronizat*
Cc: mesa-stable
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25170>
Starfield has a lot of empty ExecuteIndirect() calls. This optimizes
them by using the indirect sequence count as predicate.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25060>
Looks like indirect dispatches require an event marker instead of an
event marker with dims. That makes sense somehow given the blocks size
is not known at record time with indirect dispatches.
This allows RGP to report correct block sizes.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24994>
When TCS and TES aren't linked together and TCS exports unused outputs,
the per-patch data offset needs to be adjusted. This is similar to the
LS-HS vertex stride when VS and TCS aren't linked together.
This fixes a bunch of failures by forcing the driver to use TCS epilogs.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24776>
