nir_lower_vars_to_ssa will split temp structs up anyway. This fixes a bug
where mediump wouldn't be propagated to the split vars.
The effect is tiny, I think just shuffling some code scheduling from
optimizing at different places. Affects Natural Selection 2, Serious Sam
3, 3dmark slingshot, and Lego Legacy.
freedreno shader-db:
total instructions in shared programs: 11315637 -> 11315993 (<.01%)
instructions in affected programs: 24861 -> 25217 (1.43%)
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18466>
nir_lower_vars_to_ssa will split temp arrays up anyway. Fixes a bug where
split arrays wouldn't get their precision qualifier.
Helps mostly Android and skia shaders. Also affects Civ5, Witcher 2, and
Borderlands 2.
freedreno shader-db:
total instructions in shared programs: 11319395 -> 11319355 (<.01%)
instructions in affected programs: 65744 -> 65704 (-0.06%)
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18466>
We elminated OOB loads while renaming vars to SSA. However, if the OOB
load only appeared after some other passes had constant folded, there may
be no renaming work to do, at which point we'd leave the OOB load deref
around without renaming it or deleting it. For vc4, this was quite a
surprise and caused a regression when we stop eliminating some OOB
accesses at the GLSL level.
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18466>
It was swept at the end, but it meant that in shaders with lots of copies
available at the start of lots of if statements, you'd blow up memory
usage.
turnip memory consumption on dEQP-VK.ssbo.layout.random.scalar.75 drops
from 1.4GB to 110MB, and runtime from 19s to 17s.
Fixes: #7361
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18891>
Timestamp read is not in any hot path so there is no down-sides in
share the same function between iris, crocus, anv and hasvk.
Also while at it also dropping the functions to read MMIO from kernel,
the only use is read render timestamp so we don't need it.
v2:
- fix compilaton of ds
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: José Roberto de Souza <jose.souza@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18920>
For secondary buffer we need to emit additional vdm state along
with additional ppp state words to provide depthbias and scissor
indexes. These are indexes into depthbias and scissor tables.
In vkCmdExecuteCommands we append secondary buffer tables to primary
buffer tables, so the depthbias and scissor indexes change and
need fixing. This is why we need to emit these separatly.
Signed-off-by: Rajnesh Kanwal <rajnesh.kanwal@imgtec.com>
Reviewed-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18871>
This adds support for PVR_CMD_STREAM_TYPE_GRAPHICS_DEFERRED type control
stream. The main difference is that this is backed by host memory.
This is mainly needed for secondary command buffers allocated with
VK_COMMAND_BUFFER_USAGE_SIMULTANEOUS_USE_BIT. In this case we need
to copy secondary command buffer's control stream over to primary
stream in vkCmdExecuteCommands.
Signed-off-by: Rajnesh Kanwal <rajnesh.kanwal@imgtec.com>
Reviewed-by: Karmjit Mahil <Karmjit.Mahil@imgtec.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18871>
We were previously ignoring the breaks completelly. This patch
is heavily based on analysis and draft patch from Filip Gawin.
It mostly mirrors the current behavior for ENDIF with minor
differences. If the reader is in a branch leading directly to BRK,
we set the AbortOnRead mask as when encountering ENDIF but jump to
the ENDLOOP. If the reader was before the branch, we save the
AbortOnRead mask as if in normal branch handling and restore it
at the end of the loop.
Besides the single dEQP fix, this also fixes one more dEQP and few
piglits when the loop unrolling is disabled.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/7249
Signed-off-by: Pavel Ondračka <pavel.ondracka@gmail.com>
Reviewed-by: Filip Gawin <filip@gawin.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18900>
assert is a statement in python, not a function. Useing parens with it
leads to madness, because assert takes two arguments in the form `assert
expression: bool, message: str`. With parens though it's tempting to
write `assert(expression, message)`, which results in an assert that is
*always* true, because a non-empty tuple (which is what is written) is
*never* false.
Reviewd-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18917>
Seems like Gitlab did not follow their own documentation:
```
CI_RUNNER_TAGS 8.10 0.5 A comma-separated list of the runner tags.
```
But it would seem like a gitlab update brought json-style lists which
broke:
```
CI_RUNNER_TAGS = '["keywords-gateway", "CI-gateway"]'
```
This commit adds support for both style.
Signed-off-by: Martin Roukala (né Peres) <martin.roukala@mupuf.org>
Reviewed-by: John Brooks <john@fastquake.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18923>
