LRZ fast-clear works on all gens, however blob disables it on
gen1 and gen2. We also elect to disable fast-clear on these gens
because for close to none gains it adds complexity and seem to work
a bit differently from gen3+. Which creates at least one edge case:
if first draw which uses LRZ fast-clear doesn't lock LRZ direction
the fast-clear value is undefined.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6829
Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17599>
We're getting several more 630s in the lab, but need a bit of time to swap
out some broken old ones and stabilize the new ones. Fritz thinks this
should be done in an hour or so, but I want to turn off the CI for main so
that we don't block anyone else.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17641>
We're about to make it so that the compiler warns/errors if you use the
wrong iterator macro. Fix up a bunch of places where someone used the
wrong one before we break anything.
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17630>
In addition to hopefully generating shorter code, this optimizes out a
comparison of a mediump-cast value in
dEQP-GLES2.functional.shaders.algorithm.rgb_to_hsl_fragment passed
through ANGLE, and allows the test to pass. We believe it to be a
test bug, but emitting better code like apparently everyone else does
is also a fine result.
No change on GLES gfxbench shaders.
Fixes: #6585
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17546>
RADV and PowerVR use the same implementation.
Turnip does use a slightly modified version but the helper only has one
use -> just inline it and get rid of turnip's vk_format.h.
Signed-off-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Reviewed-by: Jason Ekstrand <jason.ekstrand@collabora.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17515>
It's everyone's favourite day, infrastructure maintenance Friday.
This includes manual disables for a618-vk and zink-anv-tgl, because
apparently the disable-on-variable rules don't carry through to those
jobs for ... some reason.
Signed-off-by: Daniel Stone <daniels@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17553>
Unlike image writes, buffer writes may access the same memory in
different ways, which we've seen in the past can cause problems. Use an
incoherent access to force flush/invalidate between accesses to the same
buffer, unless we know the barrier applies to images only.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17193>
The if statement we insert would insert a new block before the end block
(and remove the old pre-end-block). If the new block ended up later in
the HT due to its pointer's hash value, you'd emit another copy of the if
statement after the last one. I saw this happen up to 4 times in testing.
The worst case would be if all those additions and removals ended up
reallocating the HT, at which point we might use-after-free.
Fixes inconsistent shader-db results with geometry shaders.
Cc: mesa-stable.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17501>
driFetchDrawable is only ever called from the MakeCurrent path, which
means it has to handle the case of pre-GLX-1.3 Windows being named as
the drawable. When it finds the drawable in the hash, it increments its
refcount before returning it, so for a GLXWindow it would be 2 on first
return, one from glXCreateWindow and one from glXMakeCurrent. But when
it does not find the drawable and creates one for the naked Window, the
reference count on first return would only be 1. As a result, if this
context was then ever bound to a different drawable, the old Window's
DRI drawable state (like the back buffer) would be destroyed.
Fixes piglit's glx-multi-window-single-context and glx-make-current for
a variety of drivers.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6713
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17479>
And also handle tess. In all cases, we want to use the VS lowering pass
on the last geometry stage. We don't make a special exception for GS
like other drivers, because GS gets lowered into a quasi-VS.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17341>
If we didn't EmitPrimitive() then the shadow (old) outputs would not
get copied to the emit temps (to eventually be copied back to the real
outputs. This isn't so bad except that means the realy vertex_flags
output has an undefined value.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17341>
Fixed:
- VK_QUERY_TYPE_PRIMITIVES_GENERATED_EXT was able to stop prim counter
when pipeline stats query is running.
- This may have happened when prim gen query was in secondary cmdbuf.
- VK_QUERY_TYPE_PRIMITIVES_GENERATED_EXT counting geometry in each tile.
- VK_QUERY_TYPE_PRIMITIVES_GENERATED_EXT counting geometry in each tile
when pipeline stats query is started inside prim gen query and inside
a renderpass.
The matter of VK_QUERY_TYPE_PRIMITIVES_GENERATED_EXT and pipeline stats
interaction is solved by tracking whether pipeline stats query is
running both on CPU (for non secondary cmdbuf case) and on GPU (for
secondary cmdbuf).
Note, prim gen query is not allowed with secondary command buffers, so
only pipeline stats query is tracked on gpu.
See https://gitlab.khronos.org/vulkan/vulkan/-/issues/3142
Counting geometry per each tile is solved by:
- Conditionally executing START/STOP_PRIMITIVE_CTRS to not run in tiling
pass. Solves the case when prim gen query is inside a renderpass.
- Stop prim counters before executing `draw_cs` and restarting them
afterwards. Solves prim gen query being outside a renderpass.
Fixes GL CTS tests with Zink + `TU_DEBUG=gmem`:
GTF-GL46.gtf30.GL3Tests.transform_feedback.transform_feedback_max_separate
GTF-GL46.gtf40.GL3Tests.transform_feedback2.transform_feedback2_basic
GTF-GL46.gtf40.GL3Tests.transform_feedback2.transform_feedback2_framebuffer
GTF-GL46.gtf40.GL3Tests.transform_feedback3.transform_feedback3_streams_overflow
GTF-GL46.gtf40.GL3Tests.transform_feedback3.transform_feedback3_streams_queried
GTF-GL46.gtf40.GL3Tests.transform_feedback2.transform_feedback2_states
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6602
Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17163>
For a5xx+ renamed:
- RST_PIX_CNT -> START_FRAGMENT_CTRS
- RST_VTX_CNT -> STOP_FRAGMENT_CTRS
- TILE_FLUSH -> START_COMPUTE_CTRS
- STAT_EVENT -> STOP_COMPUTE_CTRS
I'm not sure about a5xx itself but I'll take a chance of it being
similar to a6xx in this regard.
Knowing this emit_begin_stat_query/emit_end_stat_query can now emit
only events that are needed for the pool's flags.
Also primitive generated query clearly doesn't need fragment and
compute counters.
Passes tests:
dEQP-VK.query_pool.statistics_query.*
dEQP-VK.transform_feedback.primitives_generated_query.*
Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17163>
This fixes tests using imageLoad/imageStore on texture
created using glEGLImageTargetTexture2DOES.
Before this change, the format was guessed as GL_RGBA,
which would be rejected by _mesa_get_shader_image_format.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16662>
New tests, dEQP line rasterization test fix that lets Intel pass.
Clears out bogus xfails from 1.3.2.0 uprev on a630, which I suspect were
"we lost the device twice on a full run once, and those fails got pasted
in without checking if it happened a full run again" (since we haven't
seen them in other full run attempts).
Also clears out the a630 vk asan xfails (essentially all tests run) by
turning off leak detection which was just catching leaks in vkcts.
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17304>
Before, we needed CP post-sched to copy-propagate references to NIR
registers produced by out-of-ssa. Now that we're in SSA, this pass ends
up not doing anything useful, and actually gets in the way by occasionally
creating a cycle in the DAG.
The entire shader-db impact is:
instructions HURT: shaders/closed/steam/tropico-5/78.shader_test FRAG: 238 -> 242 (1.68%)
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17320>
Not only runs less testing when only one driver is impacted, but also
makes sure zink+turnip is turned off when the farm is.
Reviewed-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Acked-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17287>
On a650+ we use the new Z24UINT_S8UINT format to sample the stencil
aspect of D24S8, which returns stencil in the second component and also
uses the second integer component for the border color. However Vulkan
mandates that the first component is used for the stencil border color.
There's no workaround we know of, so we have to fall back to the old
behavior where there is a workaround. If we know the format, we can
fixup the border color ourselves though.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17177>
From the API's point of view, border color replacement looks like this:
--------------------
| API Border Color |
--------------------
|
----------- | ---------------- ----------
| API fmt |-------->| User Swizzle |----->| Shader |
----------- ---------------- ----------
From the HW point of view, it looks like this:
-------------------
| HW Border Color |
-------------------
|
---------- ----------- | --------------- ----------
| HW fmt |-----| HW swap |------>| Tex Swizzle |----->| Shader |
---------- ----------- --------------- ----------
When the HW fmt + HW swap isn't enough to represent an API format, we
need to prepend our own swizzle to the the user's swizzle, so the tex
swizzle is a "format swizzle" composed with the user swizzle. However,
we don't want this format swizzle to be applied to border colors, so
there's a workaround in freedreno which is meant to undo the format
swizzle so that the HW border color with the format swizzle applied
equals the API border color, and everything is ok. However, on a6xx at
least it was incorrectly undoing the entire tex swizzle. This broke
border color with a user swizzle, because it was now effectively not
getting applied for the border color. It also made it seem like the user
swizzle is required for the workaround, which would have implications
for VK_EXT_border_color_swizzle with turnip, but it's not.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17177>
On-GPU LRZ direction tracking allows LRZ to support secondary cmdbufs,
reusing LRZ between renderpasses, and in future to support LRZ when
VK_KHR_dynamic_rendering is used.
With on-gpu tracking we have to be careful keeping LRZ state in sync
with underlying depth image, which means we should invalidate LRZ
when underlying image is changed or the view of image is different
from previous renderpass.
All of this resulted in LRZ logic being thinly spread through the code,
making it hard to understand. So most of it was moved to tu_lrz.c.
For more details on past and new LRZ features see comment at the
top of tu_lrz.c.
Note about blob:
- Blob is much more happy to do LRZ_FLUSH, it flushes at the start
of the renderpass, after binning, and at the end of the renderpass.
- Blob seem not to care about changes in depth image done via
vkCmdCopyImage.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6347
Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16251>
We have a lot of spilling coverage in a618 pre-merge, don't do it all (~2
minutes) here. Also, force-gmem touch testing should probably test less than
the default run does!
This should help make up for having added the tu-zink run.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17125>
The libvulkan-dev was needed for building zink, which ended up turning on
the vulkan tests in piglit. Split them out here.
Reviewed-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17125>
Fix defect reported by Coverity Scan.
Dereference before null check (REVERSE_INULL)
check_after_deref: Null-checking rop_reads_dst suggests that it may be
null, but it has already been dereferenced on all paths leading to the
check.
Fixes: 94be0dd0b8 ("tu: Implement extendedDynamicState2LogicOp")
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17099>
SSBO access works very differently from UBO access. Straddling
loads/stores isn't an issue, loads/stores instead must be aligned to the
element size and can have up to 4 components.
We support 16-bit access with SSBOs on a650+, and sometimes the
vectorizer tries to create a misaligned 32-bit access when combining
32-bit and 16-bit accesses. The UBO-focused logic didn't reject this,
which is now fixed. This fixes a number of VK-CTS regressions on a650+.
Fixes: bf49d4a084 ("freedreno/ir3: Enable load/store vectorization for SSBO access, too.")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17040>
We had it set up for arm64 asan already, do it for everyone else too. In
cleaning up the duplication, this fixes a pasteo in rpi3 which had the
"artifacts: false" on the wrong job, causing it to do a slow download of
the mesa build from gitlab.
Doing this required also moving the ".use-debian/arm_test" in as well, so
that its "needs:" didn't overwrite ours if it appeared after us in the
consumer's "extends:"
Should save about 20 seconds on rpi3 jobs.
Reviewed-by: Juan A. Suarez <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17146>
This needs to be accurate so that when we split and then schedule a new
a0.x/a1.x/p0.x write we will eventually make progress. It wasn't taking
the kill_path into account which could create an infinite loop as we
keep scheduling writes whose uses are blocked because they are memory
instructions not on the kill_path.
Closes: #6413
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16635>
Even if there is libdrm we shouldn't use it if KGSL is selected.
Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Hyunjun Ko <zzoon@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17173>
This should help with "marge got stuck for an hour and all I got was this
failed job with no results/" when a system intermittently wedges.
This replaces the BM_POE_TIMEOUT ("did we get something on serial in the
last 3 minutes?") that rpi had, in favor of checking that the whole test
job gets through in 20 minutes.
Acked-by: Juan A. Suarez <jasuarez@igalia.com>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17096>
Per Vulkan spec only "Draw" commands should be counted towards
occlusion query.
Apparently RB_SAMPLE_COUNT_CONTROL::UNK0 bool controls whether
sample counting is enabled, so we could use it to disable
sample counting for 3d blits which are sometimes used for
clear/copy/blit/gmem-store/resolve operations.
Fixes GL CTS tests running through Zink:
dEQP-GLES3.functional.occlusion_query.depth_clear
dEQP-GLES3.functional.occlusion_query.depth_clear_stencil_clear
dEQP-GLES3.functional.occlusion_query.scissor_depth_clear_stencil_clear
dEQP-GLES3.functional.occlusion_query.scissor_stencil_clear
dEQP-GLES3.functional.occlusion_query.stencil_clear
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6559
Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17138>
We used to use WZYX and apply swizzles. Because swizzles apply for
border colors as well, the gallium driver un-swizzled the border colors
to cancel out swizzles. That did not work for turnip because turnip
advertises customBorderColorWithoutFormat and does not know when to
un-swizzle.
This change replaces WZYX by XYZW and removes the swizzles.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/6516
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16647>
A single compiled_shaders instance could be reused by several
pipelines, but strings from disasm info could be stolen only once.
So now we have to copy them.
Fixes crashes when using RenderDoc.
Fixes: 05329d7f9a
("tu: Implement pipeline caching with shared Vulkan cache")
Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17053>
These are normally only set once because it's constant across the entire
renderpass, but they're trashed by the 3d store path because it needs to
store to CCU instead of GMEM. Therefore we need to save/restore them. Do
it in a way compatible with #5181.
Fixes: b157a5d ("tu: Implement non-aligned multisample GMEM STORE_OP_STORE")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17058>
Even though image views for attachments must use the identity swizzle,
there are cases where we have to add in our own swizzle, in particular
for D24S8 when the view is depth-only/stencil-only. Therefore we have to
reset it to the identity, similar to what we do with input attachments.
Fixes: b157a5d ("tu: Implement non-aligned multisample GMEM STORE_OP_STORE")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17058>
This is enough for zink to expose ARB_texture_buffer_object_rgb32 and
therefore GL 4.0. We could enable sampled images with a few more
workarounds, but the blob doesn't bother and there isn't any need at the
moment.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16980>
Because this impacts most of the registers in the BLEND draw state, we
make the entire draw state dynamic so that it all gets re-emitted when
the logicOp changes. This also lays the groundwork for
VK_EXT_color_write_enable.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16936>
There were a few problems with this:
- It wasn't considering logic op at all, which is another source of
reading from the destination.
- It was conditioned on the blend_enable_mask, so it was missing the
case where there's no blending but some of the outputs were masked
out.
- It wasn't considering attachments with less than 4 components (for
example, normals in a typical deferred rendering setup) and would
always consider them partially written unless the user added extra
unnecessary components.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16936>