| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This should shut up some Valgrind errors during pre-regalloc
scheduling. The errors were harmless since they could only have led
to the estimation of the bank conflict penalty of an instruction
pre-regalloc, which is inaccurate at that point of the program
compilation, but no less accurate than the intended "return 0"
fall-back path. The scheduling pass is normally re-run after regalloc
with a well-defined grf_used value and accurate bank conflict
information.
Fixes: acf98ff933d "intel/fs: Teach instruction scheduler about GRF bank conflict cycles."
Reported-by: Eero Tamminen <[email protected]>
Reviewed-by: Ian Romanick <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
obtain vector storage.
The weight_vector_type constructor was inadvertently assuming C++17
semantics of the new operator applied on a type with alignment
requirement greater than the largest fundamental alignment.
Unfortunately on earlier C++ dialects the implementation was allowed
to raise an allocation failure when the alignment requirement of the
allocated type was unsupported, in an implementation-defined fashion.
It's expected that a C++ implementation recent enough to implement
P0035R4 would have honored allocation requests for such over-aligned
types even if the C++17 dialect wasn't active, which is likely the
reason why this problem wasn't caught by our CI system.
A more elegant fix would involve wrapping the __SSE2__ block in a
'__cpp_aligned_new >= 201606' preprocessor conditional and continue
taking advantage of the language feature, but that would yield lower
compile-time performance on old compilers not implementing it
(e.g. GCC versions older than 7.0).
Fixes: af2c320190f3c731 "intel/fs: Implement GRF bank conflict mitigation pass."
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104226
Reported-by: Józef Kucia <[email protected]>
Reviewed-by: Ian Romanick <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit 9702fac68e8bd07be8871f7925d7f9fb98da3699, which
hangs vulkancts and crucible on all platforms.
The patch is being reverted because it disables continuous integration
testing. The patch from bug 104359 does not apply to master.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104359
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This fixes vmfaults seen on vega with:
dEQP-VK.pipeline.multisample_interpolation.sample_interpolate_at_single_sample_.128_128_1.samples_1
These were caused by the don't allocate cmask but it was just accidental.
The actual problem was the shader was trying to get the sample positions from
a buffer, but the buffer was never getting configured to contain them, as the
previous shader never needed them.
Reviewed-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
Fixes: 1171b304f3 (radv: overhaul fragment shader sample positions.)
Signed-off-by: Dave Airlie <[email protected]>
|
|
|
|
| |
Signed-off-by: Emil Velikov <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Emil Velikov <[email protected]>
(cherry picked from commit f66496d291881f1eaca2ee5d326367fb73537541)
|
|
|
|
|
| |
Signed-off-by: Emil Velikov <[email protected]>
(cherry picked from commit 4f5e85e9e97de4ae6e3d779ff42bf392c4739234)
|
|
|
|
|
|
|
|
| |
Found by inspection.
Cc: 17.3 <[email protected]>
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This fixes a varying packing issue when using transform feedback in
GL_SEPARATE_ATTRIBS mode. By time we get to linking, we already
know that the number of feedback attributes is under the
GL_MAX_TRANSFORM_FEEDBACK_SEPARATE_ATTRIBS limit so packing isn't
as critical. In fact, packing/splitting vec3 attributes can cause
trouble because splitting effectively creates another TFB output
which can exceed device limits. So, disable vec3 packing when it's
not needed to avoid that issue.
Fixes the Piglit ext_transform_feedback-separate test on VMware
driver.
Reviewed-by: Timothy Arceri <[email protected]>
|
|
|
|
|
|
|
| |
Handle comparing the packing class using the same method as we do
for var->data.is_xfb_only
Reviewed-by: Timothy Arceri <[email protected]>
|
|
|
|
|
|
| |
And change *components to components[] as a reminder that it's an array.
Reviewed-by: Timothy Arceri <[email protected]>
|
|
|
|
|
|
|
| |
In some cases, I think loop code is easier to read without continue
statements.
Reviewed-by: Timothy Arceri <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
The mix of bitwise operators with * and + to compute the packing_class
values was a little weird. Just use bitwise ops instead.
v2: add assertion to make sure interpolation bits fit without collision,
per Timothy. Basically, rewrite function to be simpler.
Reviewed-by: Timothy Arceri <[email protected]>
|
|
|
|
|
|
| |
The use of break/continue was kind of weird/confusing.
Reviewed-by: Timothy Arceri <[email protected]>
|
|
|
|
| |
Reviewed-by: Timothy Arceri <[email protected]>
|
|
|
|
| |
Reviewed-by: Timothy Arceri <[email protected]>
|
|
|
|
| |
Reviewed by: Timothy Arceri <[email protected]>
|
|
|
|
|
|
|
|
| |
Cc: Emil Velikov <[email protected]>
Cc: Juan A. Suarez Romero <[email protected]>
Signed-off-by: Andres Gomez <[email protected]>
Reviewed-by: Juan A. Suarez <[email protected]>
Reviewed-by: Emil Velikov <[email protected]>
|
|
|
|
|
|
| |
Fixes: bb1e6ff161c ("spirv: Add a prepass to set types on vtn_values")
Reviewed-by: Emil Velikov <[email protected]>
Signed-off-by: Juan A. Suarez Romero <[email protected]>
|
|
|
|
|
|
|
| |
Push this format to the pipe driver unchanged.
Signed-off-by: Lucas Stach <[email protected]>
Reviewed-by: Philipp Zabel <[email protected]>
|
|
|
|
|
|
|
| |
When walking over all the cases in a OpSwitch, take in account the bitsize
of the literals to avoid getting wrong cases.
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
|
|
| |
Signed-off-by: Tapani Pälli <[email protected]>
Suggested-by: Darius Spitznagel <[email protected]>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104288
Acked-by: Kenneth Graunke <[email protected]>
|
|
|
|
|
|
|
|
| |
Vulkan spec doesn't specify that VK_REMAINING_ARRAY_LAYERS is allowed
in the passed VkClearRect struct.
Signed-off-by: Samuel Iglesias Gonsálvez <[email protected]>
Reviewed-by: Jason Ekstrand <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Thanks to Karol Herbst for the debugging / tracing work that led to this
change.
Move to using lane 0 as the "work" lane for the texture. It is unclear
why this helps, as that computation should be identical to doing it in
the "correct" lane with the properly adjusted quadops.
In order to be able to use the lane 0 result, we also have to ensure
that lane 0 contains the proper array/indirect/shadow values.
This applies to Fermi and Kepler. Maxwell+ may or may not need fixing,
but that lowering logic is separate.
Fixes KHR-GL45.texture_cube_map_array.sampling
Signed-off-by: Ilia Mirkin <[email protected]>
|
|
|
|
|
| |
Most piglit textures happened to work out by RGBW not changing in that
bit, but it did cause failures in RGBA16F fbo-generatemipmap-formats.
|
|
|
|
|
| |
I wrote this early in driver development, and our UIF handling is much
better now.
|
|
|
|
|
| |
The mb_tile_layout table was just the utile_w/h times two, so reuse the
utile code instead.
|
|
|
|
|
|
| |
Apparently gallium's u_blitter wants depth from at least the .z component,
and other swizzling appears to apply on top of that. Fixes
fbo-generatemipmap-formats failures with depth formats.
|
|
|
|
| |
This matches freedreno's behavior.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
There may be some more RCL work to be done (I think I need to split my Z/S
stores when doing separate stencil), but this gets piglit's "texwrap
GL_ARB_depth_buffer_float" working.
v2: Unwrap the z32f_wrapper before calling the helper, rather than having
the helper have a callback.
v3: Rebase on Rob Clark's u_transfer_helper instead
|
|
|
|
|
|
|
| |
Mainly for testing, FD_MESA_DEBUG=hiprio will force high priority
contexts.
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
|
|
|
| |
For devices (and kernels) which support different priority ringbuffers,
expose context priority support.
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
|
|
|
|
| |
Signed-off-by: Rob Clark <[email protected]>
Reviewed-by: Roland Scheidegger <[email protected]>
Reviewed-by: Marek Olšák <[email protected]>
Reviewed-by: Andres Rodriguez <[email protected]>
Reviewed-by: Wladimir J. van der Laan <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
| |
We still have gpu hangs on Cannonlake when using push constants, so
disable them for now until we have a proper fix for these hangs.
v2: Add warning message when creating context too.
Signed-off-by: Rafael Antognolli <[email protected]>
Cc: Ben Widawsky <[email protected]>
Cc: Kenneth Graunke <[email protected]>
Reviewed-by: Ben Widawsky <[email protected]>
|
|
|
|
|
|
|
| |
F1 2017 looks good now.
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
|
|
|
|
|
|
|
|
|
| |
This is not because the vertex stage needs some push constants
that other stages need them too. This should reduce the number
of loaded SGPRs in some situations.
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
|
|
|
|
|
| |
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
|
|
|
|
|
|
|
| |
pipeline->layout can't be NULL now.
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
|
|
|
|
|
|
|
| |
pipeline->layout can't be NULL now.
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
|
|
|
|
|
|
|
| |
The spec requires it.
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
|
|
|
|
|
|
|
|
|
| |
They are dummy objects but the spec requires layout to not be
NULL, this just makes sure we are creating valid pipeline layout
objects. This will allow us to remove some useless checks.
Signed-off-by: Samuel Pitoiset <[email protected]>
Reviewed-by: Bas Nieuwenhuizen <[email protected]>
|
|
|
|
|
|
|
|
|
| |
It uses slightly more memory (though still bounded by the number
of mapped ranges), but gives less quadratic behavior.
Cuts 4 minutes from the runtime of the CTS *.sparse.* tests.
Reviewed-by: Eric Engestrom <[email protected]>
|
|
|
|
|
|
|
|
|
| |
Need to do some gymnastics to copy the parameter from the indirect
parameters buffer to uniform so shader sees the correct base-vertex-id.
Fixes ./bin/arb_draw_indirect-vertexid on a5xx and probably a4xx too.
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
|
|
|
|
| |
For dealing with indirect-draw + gl_VertexID, we'll introduce another
case where we need to use CP_MEM_TO_MEM. Rather than adding more
if(a5xx)/else make this a ctx vfunc.
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Cmdstream traces from blob make it clear that the blob driver dev's
*think* a5xx has a real (non-zero-based) vtxid. But reality claims
differently.
Fixes ./bin/gl-3.2-basevertex-vertexid and probably others.
This means draw-indirect is going to need some gymnastics to copy
base-vertex into uniform. (a4xx probably needs that too.)
Signed-off-by: Rob Clark <[email protected]>
|
|
|
|
|
|
|
|
|
| |
If we aren't binding an image, clear the compressed flags.
This fixes a segfault seen with an apitrace.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104331
Signed-off-by: Dave Airlie <[email protected]>
|
|
|
|
|
|
|
|
|
|
| |
When calculating buffer offsets for client buffers account for info.index_bias.
Fixes the follow piglit tests:
arb_draw_elements_base_vertex-drawelements-user_varrays
arb_draw_elements_base_vertex-negative-index-user_varrays
Reviewed-By: Bruce Cherniak <[email protected]>
|
|
|
|
|
|
| |
This fixes a crash on r600/r700.
Signed-off-by: Dave Airlie <[email protected]>
|