The lane select is in bit 28, this is covered by the "16-bit swizzle" mode.
However, the source type isn't inferred from the name in valhall.py, so
explicitly annotate the source as 16-bit.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14833>
The lanes are at bit 28 and bit 26 respectively. This matches the 16-bit "swizzle" encoding. In general the handling of widens/swizzles/lane/lanes on Valhall is rather confused but... one problem at a time.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14833>
Per primitive & attachment shading rate support added.
v2: Rebase on KHR_dynamic_rendering
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13739>
For instance, the current code in genX_cmd_buffer.c assumes that the
depth/stencil attachments & resolves will be at the end of all
attachments, but that won't be the case anymore with fragment rate
shading.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13739>
Those surfaces are used as attachment to rendering passes and describe
the rate of coarse pixel shading for the pass.
v2: Move CPB_BIT tile filtering to isl_gfx125_filter_tiling() (Nanley)
v3: Drop unused macro (Nanley)
s/isl_to_gen/isl_encode/ (Nanley)
Remove pitch alignment 128B constraint already covered by tiling (Nanley)
Move some asserts together (Nanley)
v4: Disable miptail for now (Nanley)
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13739>
DG2 introduces per primitive coarse pixel settings (in stages
preceding the PS shader) and also a control surface specifying the
rate at through the resulting surface.
v2: update comment (Caio)
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13739>
Recursive "loops" tend to be more difficult to follow
and understand. Additionally iterative approach should be
nicer for compiler. (Less to allocate on stack and easier to optimize)
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13226>
Add text describing why HierarchicalDepthBufferEnable must be set along
with SeparateStencilBufferEnable.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14825>
The assert for the TiledSurface field caught a programming error, but
with a segfault instead of the usual route of assert-failing. We only
set this field when we have a depth surface, but we also need to set it
when one isn't provided. Fix this issue and drop the assert.
Fixes: b77d694223 ("intel/isl: Allow HiZ with Tile4/64 surfaces")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/5950
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14825>
For SNB and prior, assert that the surface is Y-tiled and use constants
when configuring the tiling parameters. This makes a follow-on commit
clearer.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14825>
There should be no visible functional change. Although an unrelated
change in the codegen replaced vn_info_extension_spec_version by
vn_info_extension_get. We have to adapt to that.
Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org>
Reviewed-by: Ryan Neph <ryanneph@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14782>
I was tired of swapping gpus around just to check shader-db results of MRs
for these. I put it in src/amd since it doesn't make sense in either of
r300 or r600.
Reviewed-by: Matt Turner <mattst88@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14809>
We already had them disabled for hwtcl, but in the swtcl case gallivm's
param query would return (nir) support even though nir-to-tgsi couldn't
handle it because TGSI doesn't do fp16/int16.
Fixes: 7d2ea9b0ed ("r300: Request NIR shaders from mesa/st and use NIR-to-TGSI.")
Reviewed-by: Matt Turner <mattst88@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14809>
With latest Fossilize everything should now be captured correctly
but without this, all Fossilize databases that need
VK_EXT_custom_border_color would just crash.
Cc: mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14821>
We need to check the source of the copy, not the destination. That means
this we need to move this check inside the ifs, where we know that the
source is a copy.
Fixes: 590efd180b ("Prevent propagating shared regs out of loops")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14412>
The old code was both too aggressive and not aggressive enough in
inserting (jp), and it wasn't based at all on a solid understanding of
how the hardware operates. It inserted an extra unnecessary (jp) at the
beginning of an if statement right after the conditional branch, which
is unnecessary. On the other hand, the only case where it didn't
insert a (jp) was a block with one predecessor that has only one
successor. We usually don't emit these kinds of blocks, or if we do then
one of the blocks is empty, but there is a case where we *do* and the
difference actually matters: for something like
while (true) {
if (...) {
// do stuff
...
break;
}
}
The instructions inside the if could be moved below the loop, except
that they are supposed to execute before control flow is merged if the
loop trip count is nonuniform. The subgroup reduce/scan macro does
exactly this, and relies on the control flow being correct. We have to
insert a (jp) after the loop, which this code wasn't doing, breaking the
scan macro.
Since we're now using the physical edges in a non-trivial way, we have
to preserve them better when we modify control flow in ir3_legalize.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14412>
Qualcomm doesn't natively support shuffle, but it does natively support
relative shuffles where the delta is a constant. Therefore we'll expose
emulated support for both. Add support for this emulation of
subgroupShuffle() to NIR.
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Reviewed-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14412>
This option only applies to relative shuffles (up/down/xor), and in a
moment we're going to add an option to lower normal shuffles, so rename
it.
While we're here, rename lower_shuffle() to lower_to_shuffle() for
similar reasons.
Reviewed-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14412>
We recently had a bug of forgeting to add the buf->bo_offset. Just make
the easiest field to get be the bo->iova + buf->bo_offset already. Plus,
a little less work at emit time.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/14816>