Marek Olšák [Wed, 7 Jul 2021 02:30:32 +0000 (22:30 -0400)]
ac/llvm: don't return a status from ac_cull_triangle because it's unused
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11754>
Marek Olšák [Tue, 6 Jul 2021 17:18:00 +0000 (13:18 -0400)]
radeonsi: drop smoothing quality to 4xAA for better performance
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11754>
Jason Ekstrand [Thu, 11 Jun 2020 23:23:07 +0000 (18:23 -0500)]
nir: Drop nir_ssa_def::name and nir_register::name
We say that they're for debug only but we don't really have a good
policy around when to set them and when not to. In particular,
nir_lower_system_values and nir_lower_vars_to_ssa which are the chief
producers of SSA values which might reasonably have a name do not bother
to set one. We have some names set from things like BLORP and RADV's
meta shaders but AFAICT, they're setting a name more because it's there
than because they actually care.
Also, most things other than nir_clone and nir_serialize don't bother to
try and preserve them. You can see in the diffstat of this commit
exactly what passes attempt to preserve names. Notably missing from the
list is opt_algebraic which is the single largest source of SSA def
churn and it happily throws names away.
These observations lead me to question whether or not names are actually
useful at all or if they're just taking up space (8B per instruction)
and wasting CPU cycles (to ralloc_strdup on the off chance we do have
one). I don't think I can think of a single time in recent history
where I've been debugging a shader issue and a SSA value name has been
there and been useful. If anything, the few times they are there, they
just throw me off because they mess up the indentation in nir_print.
iris shader-db on my system gets runtime -2.07734% +/- 1.26933% (n=5)
Reviewed-by: Emma Anholt <emma@anholt.net>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5439>
Chia-I Wu [Wed, 7 Jul 2021 22:01:13 +0000 (15:01 -0700)]
vulkan/wsi: fix select_memory_type when all MTs are local
The intention is to pick the system memory for the prime blit dst, but
that is not possible when all memory types are advertised to be local.
This fixes venus over vtest (i.e., unix socket) because the driver
provides no PCI bus info and wsi_device_matches_drm_fd returns false. A
driver might also use can_present_on_device to force prime blit.
Fixes:
469875596a6 ("vulkan/wsi: Fix prime blits to use system memory for the destination")
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11774>
Connor Abbott [Mon, 14 Sep 2020 10:30:12 +0000 (12:30 +0200)]
tu: Update subgroup properties
Everything should be in place for this to actually work. Support a size
of 128, unlike the blob. I've also plumbed through ballot support, so
enable that.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Mon, 28 Jun 2021 11:22:59 +0000 (13:22 +0200)]
ir3/legalize: Fix loop convergence behavior
This prevents the previous commit from being undone by the jump
optimizations in legalize, and fixes another potential case where
instead of a continue we have an if/else at the end of a loop.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Mon, 28 Jun 2021 10:56:15 +0000 (12:56 +0200)]
ir3: Fix convergence behavior for loops with continues
When loops have continue statements, it's expected that when we execute
a divergent continue (i.e. a continue where not all of the threads
active at the start take it) we keep going with the rest of the loop
body and then reconverge at the start of the next iteration. However the
Adreno ISA seems to always take a branch that jumps backwards, assuming
it's the bottom of a loop, so we get a different, undesired convergence
behavior. There's no way I know of to control this behavior in the
instruction set, so we have to instead insert a "continue block" at the
end of the loop where continue statements reconverge which then jumps
back to the top of the loop. Since this doesn't correspond 1:1 with any
NIR block we have to make control flow handling in NIR->ir3 a bit more
complicated, unfortunately.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Mon, 31 May 2021 10:58:26 +0000 (12:58 +0200)]
ir3: Implement nir subgroup intrinsics
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Mon, 31 May 2021 10:21:29 +0000 (12:21 +0200)]
ir3: Handle shared registers in lower_parallelcopy
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Mon, 31 May 2021 10:09:42 +0000 (12:09 +0200)]
ir3: Add subgroup pseudoinstructions
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Fri, 11 Sep 2020 11:17:40 +0000 (13:17 +0200)]
ir3: Support any/all/getone branches
This plumbs through the support in the IR.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Tue, 1 Sep 2020 13:22:14 +0000 (15:22 +0200)]
ir3: Cleanup ir3_legalize jump optimization
Do the optimization parts in their own loop, and be more robust when
detecting the useless jumps.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Fri, 28 May 2021 15:31:48 +0000 (17:31 +0200)]
ir3/sched: Handle branch condition in split_pred()
Before this, if there was a block with multiple things writing p0.x,
it was a tossup whether the right one would be used as the branch
condition. Found by inspection.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Mon, 28 Jun 2021 16:41:41 +0000 (18:41 +0200)]
ir3: Fix infinite loop in scheduler when splitting
When we go to split e.g. a p0.x producer, the only other instructions
ready to schedule are often only p0.x producers. It could happen that
they all have a lower priority than the split instruction. Then we would
immediately schedule the split instruction again, then again try to
schedule one of the other producers, be blocked, and split it, around
and around again, leading to an infinite loop. The following commit
triggered this with
dEQP-GLES3.functional.shaders.discard.dynamic_loop_always on a3xx.
Fixes: d2f4d33 ("freedreno/ir3: new pre-RA scheduler")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Fri, 28 May 2021 14:03:16 +0000 (16:03 +0200)]
ir3: Make MOVMSK use repeat
MOVMSK is a bit of a special case, because it takes multiple cycles (and
therefore reduces the nops needed if it's between some other assigner
and consumer) however weird things happen if you try to start reading
the first component while it isn't finished yet. On balance making it
use repeat seems to result in a fewer special cases.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Fri, 28 May 2021 13:41:56 +0000 (15:41 +0200)]
ir3: Fix shared reg delay
Based on computerator experiments, this is actually 6, including for
movmsk.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Fri, 25 Jun 2021 14:51:22 +0000 (16:51 +0200)]
ir3: Actually allow shared reg moves to be folded
I realized that shared registers were never actually getting folded,
even after adding them to valid_flags, because the move wasn't even
being considered.
I looked at the other uses of is_same_type_mov(), and they should be ok
with this.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Fri, 25 Jun 2021 15:50:47 +0000 (17:50 +0200)]
ir3: Better valid flags for shared regs
Shared registers seem to use the same port as consts, so the same
restrictions for cat2/cat3 apply to them.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Thu, 27 May 2021 15:28:09 +0000 (17:28 +0200)]
ir3: Prevent propagating shared regs out of loops
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Mon, 28 Jun 2021 12:48:08 +0000 (14:48 +0200)]
ir3: Handle unreachable blocks
This fixes a pre-existing bug in ir3, but it showed up even more due to
other changes in this series and it interacts with the logical/physical
CFG split. When both sides of an if end with a jump, a block may become
unreachable via the logical CFG, which can cause problems because it has
no predecessors to figure out the location of live-in non-shared
values. In this case we assume that nir_opt_if has removed any code in
these blocks and just skip processing live-ins for these blocks,
pretending that they aren't live.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Thu, 27 May 2021 14:54:20 +0000 (16:54 +0200)]
ir3: Handle shared register liveness correctly
As explained in the comments added, we need to add extra edges to the
CFG which are ignored except for shared registers. This plumbs through
support for this.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Fri, 11 Sep 2020 11:31:44 +0000 (13:31 +0200)]
ir3/nir: Call nir_lower_subgroups
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Mon, 31 May 2021 12:21:04 +0000 (14:21 +0200)]
tu, ir3: Plumb through support for CS subgroup size/id
The way that the blob obtains the subgroup id on compute shaders is by
just and'ing gl_LocalInvocationIndex with 63, since it advertizes a
subgroupSize of 64. In order to support VK_EXT_subgroup_size_control and
expose a subgroupSize of 128, we'll have to do something a little more
flexible. Sometimes we have to fall back to a subgroup size of 64 due to
various constraints, and in that case we have to fake a subgroup size of
128 while actually using 64 under the hood, by just pretending that the
upper 64 invocations are all disabled. However when computing the
subgroup id we need to use the "real" subgroup size. For this purpose we
plumb through a driver param which exposes the real subgroup size. If
the user forces a particular subgroup size then we lower
load_subgroup_size in nir_lower_subgroups, otherwise we let it through,
and we assume when translating to ir3 that load_subgroup_size means
"give me the *actual* subgroup size that you decided in RA" and give you
the driver param.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Mon, 14 Sep 2020 08:14:55 +0000 (10:14 +0200)]
nir: Add read_invocation_cond_ir3 intrinsic
On qualcomm, we have shared registers similar to SGPR's on AMD. However,
there is no readlane or readfirstlane primitive. shared registers can
only be written to when just one lane is active. This means that we have
to lower readInvocation(val, id) to something like:
if (gl_SubgroupInvocation == id) {
scalar_reg = val;
}
return scalar_reg;
However it's a bit difficult to actually get the value of
gl_SubgroupInvocation in the backend, because for compute it requires
some calculations and we don't have any CSE support in the backend. This
intrinsic lets us turn it into
"readInvocationCond(val, id == gl_SubgroupInvocation)" in NIR at which
point the backend code generation is a lot easier.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Thu, 10 Sep 2020 16:48:04 +0000 (18:48 +0200)]
nir/subgroups: Support > 1 ballot components
Qualcomm has a mode with a subgroup size of 128, so just emitting larger
integer operations and then lowering them later isn't an option. This
makes the pass able to handle the lowering itself, so that we don't have
to go down to 64-thread wavefronts when ballots are used.
(The GLSL and legacy SPIR-V extensions only support a maximum of 64
threads, but I guess we'll cross that bridge when we come to it...)
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Connor Abbott [Fri, 11 Sep 2020 11:07:48 +0000 (13:07 +0200)]
nir/subgroups: Replace lower_vote_eq_to_ballot with lower_vote_eq
Lower it to a vote instead of a ballot. This was only used for AMD, and
in that case they're pretty much the same. However Qualcomm has a vote
builtin, which we want to use instead of ballots.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Acked-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/6752>
Timur Kristóf [Tue, 8 Jun 2021 10:25:36 +0000 (12:25 +0200)]
radv: Set parameter cache oversubscription according to the PC lines.
This matches how RadeonSI also sets this register.
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/11711>
Timur Kristóf [Mon, 7 Jun 2021 21:33:01 +0000 (23:33 +0200)]
radv: Allow enabling vertex grouping, fix NGG info with it disabled.
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/11711>
Timur Kristóf [Mon, 5 Jul 2021 07:05:43 +0000 (09:05 +0200)]
radv: Move radv_optimize_nir_algebraic to a separate function.
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/11711>
Timur Kristóf [Mon, 5 Jul 2021 06:51:09 +0000 (08:51 +0200)]
radv: Add last_vgt_api_stage and use it to simplify some code.
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/11711>
Mike Blumenkrantz [Wed, 7 Jul 2021 19:36:45 +0000 (15:36 -0400)]
zink: ensure sparse allocations aren't marked host-visible
Fixes:
5fee58bf599 ("zink: collapse host_visible and non-coherent alignment alloc cases")
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11768>
Samuel Pitoiset [Thu, 8 Jul 2021 07:09:25 +0000 (09:09 +0200)]
aco: fix shared_atomic_comp_swap if the second source isn't a VGPR
Only VGPRs are valid with DS instructions.
Cc: 21.1 mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11777>
Simon Ser [Thu, 24 Jun 2021 18:32:16 +0000 (20:32 +0200)]
vulkan/wsi/wayland: handle dmabuf params allocation failure
Return VK_ERROR_OUT_OF_HOST_MEMORY if
zwp_linux_dmabuf_v1_create_params fails.
Signed-off-by: Simon Ser <contact@emersion.fr>
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11589>
Simon Ser [Thu, 24 Jun 2021 18:30:25 +0000 (20:30 +0200)]
vulkan/wsi/wayland: fix wsi_wl_image_init error code
If image->buffer cannot be allocated, the value returned by
wsi_create_native_image is returned. However, if we got that far,
that value is VK_SUCCESS.
Fix it and return VK_ERROR_OUT_OF_HOST_MEMORY.
Signed-off-by: Simon Ser <contact@emersion.fr>
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11589>
Simon Ser [Thu, 24 Jun 2021 18:27:42 +0000 (20:27 +0200)]
vulkan/wsi/wayland: remove unnecessary wl_proxy_set_queue call
A wl_proxy inherits its queue from its parent.
display->dmabuf.wl_dmabuf already has its queue correctly set up,
so it's unnecessary to set it again on the child
zwp_linux_buffer_params_v1 proxy.
Signed-off-by: Simon Ser <contact@emersion.fr>
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11589>
Simon Ser [Thu, 24 Jun 2021 18:24:28 +0000 (20:24 +0200)]
vulkan/wsi/wayland: remove swapchain wl_drm wrapper
The sole purpose of this wl_proxy is to set the queue to
chain->display->queue. However, wl_proxy inherit their queue from
their parent, so the original wl_drm proxy already has its queue
set up properly (inherited from wl_registry).
Signed-off-by: Simon Ser <contact@emersion.fr>
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11589>
Qiang Yu [Wed, 7 Jul 2021 07:00:50 +0000 (15:00 +0800)]
st/mesa: fix size miss match for some check
While we shrink some variable from "GLuint" to "ubyte",
need to update the check from "x != ~0U" to "x != 0xff" too.
This fixes the crash for SPECviewperf 13 benchmark medical
case.
Fixes:
d947e3e2c8c2 "st/mesa: decrease the size of st_vertex_program"
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: Qiang Yu <yuq825@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11757>
Emma Anholt [Tue, 6 Jul 2021 19:29:58 +0000 (12:29 -0700)]
nir_to_tgsi: Declare immediates as float on non-native-ints hardware.
Makes the values more legible on i915g, and may keep us from tripping
asserts on nouveau's non-native-integer HW.
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11744>
Emma Anholt [Tue, 6 Jul 2021 19:24:51 +0000 (12:24 -0700)]
nir_to_tgsi: Run copy prop (and thus dce) after lower_bool_to_float.
Fixes some i915g instruction count failures by eliminating some MOVs from
translating b2f32s.
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11744>
Emma Anholt [Tue, 6 Jul 2021 19:41:41 +0000 (12:41 -0700)]
i915g: Fix writemasking of SEQ/SNE/SSG.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11744>
Heinrich Fink [Tue, 6 Jul 2021 18:37:52 +0000 (20:37 +0200)]
softpipe: add missing sentinel to debug option array
Add DEBUG_NAMED_VALUE_END to finalize debug option array (see
lp_screen.c). Otherwise debug_get_flags_option might attempt to read
debug_named_value::name at an offset and SIGSEGV.
Signed-off-by: Heinrich Fink <hfink@snap.com>
Fixes:
991def0edc8 ("softpipe: Convert to comma-separated SOFTPIPE_DEBUG for debug options.")
Reviewed-by: Marcin Ślusarz <marcin.slusarz@intel.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11742>
Erik Faye-Lund [Tue, 6 Jul 2021 14:00:32 +0000 (16:00 +0200)]
ci: fix source-deps for radv on windows
If we're giong to build RADV on Windows, we need to make sure we trigger
the build on all RADV-changes.
Fixes:
d18563ea58b ("ci: Update Windows image to build RADV")
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11731>
Kenneth Graunke [Thu, 1 Jul 2021 21:49:49 +0000 (14:49 -0700)]
iris: Force device local memory for u_upload_mgr buffers
We try to place persistent/coherent buffers from the application in
system memory, because they want the CPU-GPU coherency.
However, our internal u_upload_mgr buffers are also flagged persistent +
coherent, but we absolutely want most of them in device local memory.
Mark had done this correctly in an earlier patch series, but I made a
mistake when refactoring things during upstreaming, and accidentally
put these in SMEM again. This fixes that mistake.
Tested-by: Luis Felipe Strano Moraes <luis.strano@intel.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11681>
Dylan Baker [Wed, 7 Jul 2021 19:23:11 +0000 (12:23 -0700)]
docs: Add calendar entries for 21.2 release candidates.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11767>
Roman Stratiienko [Wed, 16 Jun 2021 15:00:06 +0000 (18:00 +0300)]
AOSP: Do not add '-Wl,--gc-sections' to the linker arguments
With '-Wl,--gc-sections' meson.build cc.has_function() will never fail,
providing wrong input to the build system.
Fixes:
8621bd8d5e67 ("android: Add scripts to build using meson")
Signed-off-by: Roman Stratiienko <r.stratiienko@gmail.com>
Tested-by: Mauro Rossi <issor.oruam@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11637>
Mike Blumenkrantz [Wed, 7 Jul 2021 13:14:04 +0000 (09:14 -0400)]
lavapipe: disable line rasterization ext
need to add some gallium hooks for line drawing mode before this can
be reliably supported
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11762>
Mike Blumenkrantz [Sat, 1 May 2021 15:45:28 +0000 (11:45 -0400)]
nir/format_convert: add ssa version of uint packing
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10619>
Mike Blumenkrantz [Tue, 4 May 2021 18:08:09 +0000 (14:08 -0400)]
nir/format_convert: nir_shift -> nir_shift_imm
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10619>
Alyssa Rosenzweig [Tue, 6 Jul 2021 22:17:02 +0000 (18:17 -0400)]
panfrost: Remove unused midgard-pack.h includes
Now only included from the per-gen file.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Acked-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11745>
Alyssa Rosenzweig [Tue, 6 Jul 2021 22:15:54 +0000 (18:15 -0400)]
panfrost: Remove pan_cmdstream.h
Now empty.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Acked-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11745>
Alyssa Rosenzweig [Tue, 6 Jul 2021 22:14:11 +0000 (18:14 -0400)]
panfrost: Move sample accessor to pan_cmdstream
Not really arch-dependent but technically uses GenXML. This is pretty
trivial anyway.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Acked-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11745>
Alyssa Rosenzweig [Tue, 6 Jul 2021 22:10:06 +0000 (18:10 -0400)]
panfrost: Clean up pan_cmdstream.h
We don't want other files accessing these functions except through the
vtables, since they will soon be architecture dependent.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Acked-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11745>
Alyssa Rosenzweig [Tue, 6 Jul 2021 22:09:43 +0000 (18:09 -0400)]
panfrost: Use vtable for fragment descriptor functions
Arch dependent and not directly from Gallium.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Acked-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11745>
Alyssa Rosenzweig [Tue, 6 Jul 2021 22:08:52 +0000 (18:08 -0400)]
panfrost: Move panfrost_emit_tile_map to pan_job
...where it's actually called. Saves us from using the vtable since this
isn't version-dependent.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Acked-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11745>
Alyssa Rosenzweig [Tue, 6 Jul 2021 22:07:26 +0000 (18:07 -0400)]
panfrost: Move launch_grid to pan_cmdstream
Same issues apply as for draw_vbo.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Acked-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11745>
Alyssa Rosenzweig [Tue, 6 Jul 2021 21:55:21 +0000 (17:55 -0400)]
panfrost: Don't ralloc panfrost_blend_state
Pointless divergence from other CSO creates.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Acked-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11745>
Alyssa Rosenzweig [Tue, 6 Jul 2021 21:53:15 +0000 (17:53 -0400)]
panfrost: Move blend CSO to cmdstream/context
Create is hardware specifc, bind/delete/set_colour is not.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Acked-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11745>
Alyssa Rosenzweig [Tue, 6 Jul 2021 21:48:46 +0000 (17:48 -0400)]
panfrost: Split out prepare_rsd into a vtbl
This needs to be per-gen, but the rest of the caller does not, so let's
split this out. iris makes extensive use of this vtbl pattern for cold
paths like this one.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Acked-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11745>
Alyssa Rosenzweig [Tue, 6 Jul 2021 21:33:30 +0000 (17:33 -0400)]
panfrost: Move most CSO creates to pan_cmdstream.c
These need to be conditionally compiled per-generation.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Acked-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11745>
Alyssa Rosenzweig [Tue, 6 Jul 2021 21:27:07 +0000 (17:27 -0400)]
panfrost: Move draw_vbo to pan_cmdstream.c
This needs to be conditionally compiled per-generation.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Acked-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11745>
Samuel Pitoiset [Tue, 6 Jul 2021 16:06:40 +0000 (18:06 +0200)]
aco: fix emitting a16 for MIMG instructions on GFX10+
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11734>
Samuel Pitoiset [Tue, 6 Jul 2021 16:06:10 +0000 (18:06 +0200)]
aco: fix emitting d16 for MIMG instructions on GFX9+
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11734>
Iago Toral Quiroga [Wed, 7 Jul 2021 09:52:17 +0000 (11:52 +0200)]
v3dv: account for dst offset of copy query results operations
This fixes several CTS tests with this pattern:
dEQP-VK.query_pool.occlusion_query.copy_results*dstoffset
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11758>
Corentin Noël [Wed, 16 Jun 2021 17:31:52 +0000 (19:31 +0200)]
ci: Bump virglrenderer
Use the latest commit from master branch
Signed-off-by: Corentin Noël <corentin.noel@collabora.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Rohan Garg <rohan.garg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11572>
Corentin Noël [Wed, 23 Jun 2021 17:21:15 +0000 (19:21 +0200)]
ci: Re-enable virgl tesselation shader
Partially undoes
c8c7450fc73b8 ("llvmpipe: move coroutines out of noopt case")
as the required features are now implemented in llvmpipe.
Signed-off-by: Corentin Noël <corentin.noel@collabora.com>
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Rohan Garg <rohan.garg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11572>
Alyssa Rosenzweig [Wed, 7 Jul 2021 02:11:31 +0000 (22:11 -0400)]
agx: Ensure we don't overallocate registers
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11751>
Alyssa Rosenzweig [Wed, 7 Jul 2021 01:54:01 +0000 (21:54 -0400)]
agx: Pipe in nir_register
This is kind of lazy...
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11751>
Alyssa Rosenzweig [Sat, 19 Jun 2021 18:34:44 +0000 (14:34 -0400)]
agx: Assign registers locally
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11751>
Alyssa Rosenzweig [Sat, 19 Jun 2021 18:33:30 +0000 (14:33 -0400)]
agx: Count read registers as well
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11751>
Alyssa Rosenzweig [Wed, 7 Jul 2021 02:16:20 +0000 (22:16 -0400)]
agx: Don't choke on registers in the optimizer
Just skip over them.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11751>
Mike Blumenkrantz [Fri, 14 May 2021 22:30:58 +0000 (18:30 -0400)]
zink: store the default variant hash for a program
the first created pipeline with a program is the default variants, and
this is likely to be the most common, so we can store the hash to avoid
needing to recalc it later
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11740>
Mike Blumenkrantz [Fri, 14 May 2021 22:17:23 +0000 (18:17 -0400)]
zink: flag shader modules as default
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11740>
Mike Blumenkrantz [Fri, 14 May 2021 22:15:42 +0000 (18:15 -0400)]
zink: keep a mask of stages present in a gfx program
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11740>
Mike Blumenkrantz [Fri, 14 May 2021 22:10:08 +0000 (18:10 -0400)]
zink: flag all shaders for create during gfx program init
now we can avoid needing extra conditionals in the update hotpath
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11740>
Mike Blumenkrantz [Fri, 14 May 2021 21:57:16 +0000 (17:57 -0400)]
zink: remove zink_shader_module refcounting
this is no longer necessary
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11740>
Mike Blumenkrantz [Wed, 7 Jul 2021 02:09:03 +0000 (22:09 -0400)]
features: mark off VK_EXT_multi_draw for radv
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11750>
Mike Blumenkrantz [Tue, 11 May 2021 17:50:51 +0000 (13:50 -0400)]
zink: consolidate and optimize index buffer handling during draw
this can be reorganized to have more readable handling
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11746>
Mike Blumenkrantz [Wed, 7 Jul 2021 01:41:51 +0000 (21:41 -0400)]
zink: move descriptor update closer to start of draw
this can trigger flushes, so ensure that it doesn't break the cmdbuf
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11746>
Mike Blumenkrantz [Tue, 11 May 2021 16:50:32 +0000 (12:50 -0400)]
zink: update streamout buffer strides inline
loops--
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11746>
Mike Blumenkrantz [Tue, 11 May 2021 16:43:38 +0000 (12:43 -0400)]
zink: use last_vertex_stage pointer to optimize streamout emission during draw
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11746>
Mike Blumenkrantz [Tue, 11 May 2021 16:39:42 +0000 (12:39 -0400)]
zink: store the last vertex stage to the context during bind
make this more readily accessible
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11746>
Timothy Arceri [Sun, 4 Jul 2021 11:20:59 +0000 (21:20 +1000)]
util/radeonsi: add radeonsi workaround for Nuclear Throne
Without radeonsi_sync_compile the game crashes at startup. The
game engine uses a custom global new operator which doesn't
play well with multithreading.
Fixes:
f1cc13727c68 ("radeonsi: enable workarounds for YoYo engine based games")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/5023
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11701>
Mike Blumenkrantz [Wed, 7 Jul 2021 00:38:02 +0000 (20:38 -0400)]
relnotes: EXT_extended_dynamic_state2 for lavapipe
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11736>
Mike Blumenkrantz [Wed, 7 Jul 2021 00:37:55 +0000 (20:37 -0400)]
features: EXT_extended_dynamic_state2 for lavapipe
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11736>
Mike Blumenkrantz [Tue, 6 Jul 2021 18:26:05 +0000 (14:26 -0400)]
lavapipe: implement EXT_extended_dynamic_state2
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11736>
Mike Blumenkrantz [Tue, 6 Jul 2021 16:20:54 +0000 (12:20 -0400)]
lavapipe: only apply pipeline state for depth bias if it's enabled
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11736>
Mike Blumenkrantz [Mon, 17 May 2021 12:44:26 +0000 (08:44 -0400)]
zink: remove unnecessary stall during device-local map case
this stalls after the copy already, no need to also stall before it
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11749>
Mike Blumenkrantz [Mon, 17 May 2021 23:03:08 +0000 (19:03 -0400)]
zink: redo streamout and texture components of memory_barrier hook
this is more consolidated (and legal)
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11748>
Mike Blumenkrantz [Tue, 11 May 2021 15:08:03 +0000 (11:08 -0400)]
zink: optimize buffer rebinds
this leverages all the slot masks to more optimally run through all
the rebinds, also enabling some code consolidation
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11747>
Mike Blumenkrantz [Tue, 11 May 2021 14:53:58 +0000 (10:53 -0400)]
zink: remove barriers from buffer rebinds
these aren't needed anymore
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11747>
Mike Blumenkrantz [Tue, 11 May 2021 14:34:40 +0000 (10:34 -0400)]
zink: make image_bind_count work for buffers
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11747>
Mike Blumenkrantz [Tue, 11 May 2021 14:13:45 +0000 (10:13 -0400)]
zink: make samplerview bind mask apply to buffer resources too
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11747>
Mike Blumenkrantz [Tue, 11 May 2021 14:12:39 +0000 (10:12 -0400)]
zink: add a per-stage bind mask for ssbos
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11747>
Mike Blumenkrantz [Tue, 11 May 2021 14:09:06 +0000 (10:09 -0400)]
zink: add a per-stage mask for ubo binds
optimizing buffer replacement was never so difficult
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11747>
Mike Blumenkrantz [Tue, 11 May 2021 14:20:35 +0000 (10:20 -0400)]
zink: remove some descriptor_refs_dirty checks from resource binding
these no longer provide any optimization
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11692>
Mike Blumenkrantz [Tue, 11 May 2021 13:39:11 +0000 (09:39 -0400)]
zink: remove fb surface resource refs
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11692>
Mike Blumenkrantz [Tue, 11 May 2021 13:24:26 +0000 (09:24 -0400)]
zink: remove fb surface refs
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11692>
Mike Blumenkrantz [Tue, 11 May 2021 13:23:25 +0000 (09:23 -0400)]
zink: remove samplerview refs
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11692>
Mike Blumenkrantz [Tue, 11 May 2021 13:23:06 +0000 (09:23 -0400)]
zink: remove imageview refs from shader images
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11692>
Mike Blumenkrantz [Tue, 11 May 2021 13:32:41 +0000 (09:32 -0400)]
zink: set new batch usage during surface rebinds
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11692>