Alyssa Rosenzweig [Wed, 27 May 2020 21:58:41 +0000 (17:58 -0400)]
pan/bi: Remove schedule_barrier
Legacy from Midgard.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5260>
Alyssa Rosenzweig [Tue, 5 May 2020 22:20:08 +0000 (18:20 -0400)]
pan/bi: Add helper to measure clause size
Useful for branching.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5260>
Alyssa Rosenzweig [Tue, 5 May 2020 21:58:16 +0000 (17:58 -0400)]
pan/bi: Add bi_layout.c for clause layout helpers
Figuring out what "shapes" of clauses are kosher happens during
scheduling, not packing, but shouldn't distract the scheduler. So let's
add a new file for these sorts of questions.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5260>
Alyssa Rosenzweig [Tue, 5 May 2020 21:29:24 +0000 (17:29 -0400)]
pan/bi: Remove more artefacts of 2-pass scheduling
A clause is, by definition, already scheduled.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5260>
Alyssa Rosenzweig [Tue, 5 May 2020 20:15:44 +0000 (16:15 -0400)]
pan/bi: Add MUL.i32 to disasm
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5260>
Alyssa Rosenzweig [Tue, 5 May 2020 20:15:36 +0000 (16:15 -0400)]
pan/bi: Disassemble pos=0xe
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5260>
Alyssa Rosenzweig [Tue, 5 May 2020 20:15:16 +0000 (16:15 -0400)]
pan/bi: Document constant count invariant
constants + instructions <= 13
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5260>
Alyssa Rosenzweig [Tue, 5 May 2020 18:34:58 +0000 (14:34 -0400)]
pan/bi: Move bi_flip_ports out of port assignment
It's more of a packing fixup than anything scheduler-y, and port
assignment will soon be the domain of the scheduler.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5260>
Alyssa Rosenzweig [Tue, 5 May 2020 18:31:20 +0000 (14:31 -0400)]
pan/bi: Add FILE* argument to bi_print_registers
In case we need it in general IR printing.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5260>
Alyssa Rosenzweig [Tue, 5 May 2020 18:30:06 +0000 (14:30 -0400)]
pan/bi: Drop `struct` from bi_registers
It's a full-fledged part of the IR now.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5260>
Alyssa Rosenzweig [Tue, 5 May 2020 18:28:53 +0000 (14:28 -0400)]
pan/bi: Move bi_registers to bi_bundle
Make it a part of the IR itself.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5260>
Alyssa Rosenzweig [Tue, 5 May 2020 18:23:41 +0000 (14:23 -0400)]
pan/bi: Move bi_registers to common IR structures
Port assignments are critical to scheduling, this can't just live in
bi_pack.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5260>
Alyssa Rosenzweig [Tue, 5 May 2020 18:17:58 +0000 (14:17 -0400)]
pan/bi: Remove comment about old scheduler design
I've realized it really has to be 1-pass to be sane.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5260>
Alyssa Rosenzweig [Mon, 4 May 2020 23:46:32 +0000 (19:46 -0400)]
pan/bi: Remove FMA? parameter from get_src
We can lower away zeroes a bit earlier.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5260>
Alyssa Rosenzweig [Fri, 29 May 2020 20:06:39 +0000 (16:06 -0400)]
panfrost: Preload gl_FragCoord on Bifrost
It's a precoloured register but we do need to specify in the cmdstream
that we want the preloading to happen.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5267>
Alyssa Rosenzweig [Fri, 29 May 2020 20:06:10 +0000 (16:06 -0400)]
panfrost: Set reads_frag_coord as a sysval
In addition to parsing out the varying. This is needed so it works on
Bifrost as well.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5267>
Alyssa Rosenzweig [Fri, 29 May 2020 20:05:22 +0000 (16:05 -0400)]
panfrost: Don't generate gl_FragCoord varying on Bifrost
It's treated as a sysval there, so that's silly.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5267>
Rob Clark [Mon, 18 May 2020 21:14:51 +0000 (14:14 -0700)]
freedreno/a6xx: fix vsc assert
Fixes a debug build assert seeing with an android app. Not quite sure
which path was passing us draw_info w/ instance_count==0. But we should
just treat non-instanced draws as having a single instance.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5091>
Kristian H. Kristensen [Thu, 21 May 2020 08:26:02 +0000 (01:26 -0700)]
freedreno/a6xx: Program VFD_DEST_CNTL from program stateobj
This only depends on the generated shader.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5140>
Kristian H. Kristensen [Thu, 21 May 2020 08:38:44 +0000 (01:38 -0700)]
freedreno/a6xx: Create stateobj for VFD_DECODE
This now only depends on vertex state and we can create it once
up front in pctx->create_vertex_elements_state().
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5140>
Kristian H. Kristensen [Thu, 21 May 2020 05:50:20 +0000 (22:50 -0700)]
freedreno/a6xx: Decouple VFD_FETCH and VFD_DECODE
We used to output a VFD_FETCH entry for each VFD_DECODE, but we can
instead output just one VFD_FETCH per VBO and point multiple
VFD_DECODE entries at the same VFD_FETCH entry. There's typically
fewer VBOs than vertex elements so this is a small win in itselfs, but
more importantly, the VFD_DECODE state now only depends on program
state.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5140>
Kristian H. Kristensen [Thu, 21 May 2020 05:34:59 +0000 (22:34 -0700)]
freedreno/a6xx: Move per element offset to VFD_DECODE
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5140>
Eric Anholt [Sat, 23 May 2020 03:00:07 +0000 (20:00 -0700)]
ci: Rename x86_cross_arm_test to just arm_test.
This gets us back to the behavior we used to have for freedreno: clicking
play on arm_test gets you testing of the ARM platforms that aren't under
arm-build (the LAVA runners).
Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5247>
Eric Anholt [Sat, 23 May 2020 02:58:50 +0000 (19:58 -0700)]
ci: Don't build an arm_test container now that the last user is gone.
db410c and cheza used to use it, and now both are on baremetal.
Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5247>
Eric Anholt [Fri, 15 May 2020 18:22:37 +0000 (11:22 -0700)]
ci: Switch cheza (freedreno a630) testing to baremetal.
Now that we have scripts in place to do baremetal testing of cheza, switch
it over. As of this writing, we have 5 chezas for baremetal and 4 for the
old docker CI setup (just 2 fewer than we originally had before this work,
since some had had filesystem failures and I switched those first), and
once we are sure of this we can backport to stable branch CI and move the
rest of them to baremetal.
I've run a lot of jobs through the baremetal scripts as I worked on
sorting out vulkan CTS stability, so I feel good about the stability of
the GLES CTS here.
The options job is now split out to separate jobs, as we don't currently
have a way to stack multiple sets deqp runs with different env vars in a
single baremetal run, and just chaining cros_servo.sh invocations runs
into a lack of cleanup of the serial-watching scripts which we rely on
container exit sorting out for us. This means a little less than 2x the
artifacts downloads we had before for a630 and a few more container
instantiations.
Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5247>
Eric Anholt [Fri, 15 May 2020 16:57:25 +0000 (09:57 -0700)]
ci: Add scripts for controlling bare-metal chezas.
This will let us:
- deploy kernels for testing code depending on new kernel featuers
- Ensure a pristine state in the HW before starting our tests
- Avoid disk rot on the chezas taking them out (we'd lost 3/9 in a few
months).
Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5247>
Eric Anholt [Mon, 6 Apr 2020 22:55:30 +0000 (15:55 -0700)]
ci: Build a cheza kernel.
This is a set of kernel options I've come up with mostly cribbing from
chrome os's kernel config snippet. We also build an lzma kernel, as
uncompressed kernel is big but lzma is the only compression supported by
the bootloader. With that image, we have to pack it into a FIT formatted
image+dtb blob.
CONFIG_SUNRPC_DEBUG is added so that you can set "nfsrootdebug" to figure
out what's going wrong with your nfs mount (mine were "both the tcp and
nfsvers options were required, and don't try to use 'default' as the root
path to defer to DHCP's answer because otherwise you get
/tftpboot/default, just use an empty root path which doesn't prepend
/tftpboot.")
Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5247>
Eric Anholt [Tue, 7 Apr 2020 00:00:14 +0000 (17:00 -0700)]
ci: Disable the firmware loader user helper option in arm64 kernels.
We won't have a user helper, so don't block for 60 seconds for it to show
up. Speeds up debug of new kernel builds.
Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5247>
Samuel Pitoiset [Thu, 21 May 2020 15:14:12 +0000 (17:14 +0200)]
radv/aco: enable VK_KHR_subgroup_extended_types on GFX8+
Should be working now.
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/5148>
Samuel Pitoiset [Thu, 28 May 2020 07:09:49 +0000 (09:09 +0200)]
aco: sign-extend input/indentity for 32-bit reduce ops on GFX10
Because some 16-bit instructions are already VOP3 on GFX10, we use
the 32-bit variants to remove the temporary VGPR and to use DDP with
the arithmetic instructions.
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/5148>
Samuel Pitoiset [Mon, 18 May 2020 15:01:46 +0000 (17:01 +0200)]
aco: allow gfx10_wave64_bpermute with 8-bit/16-bit input
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/5148>
Samuel Pitoiset [Mon, 18 May 2020 15:03:21 +0000 (17:03 +0200)]
aco: allocate a temp VGPR for some 8-bit/16-bit reduction ops 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/5148>
Samuel Pitoiset [Mon, 18 May 2020 15:38:17 +0000 (17:38 +0200)]
aco: implement 8-bit/16-bit reductions on GFX10
Some 16-bit instructions are VOP3 on GFX10 and we have to emit a
32-bit DPP mov followed by the ALU instruction.
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/5148>
Samuel Pitoiset [Fri, 29 May 2020 08:02:39 +0000 (10:02 +0200)]
aco: fix register allocation for subdword instructions on GFX10
Cc: 20.1 <mesa-stable@lists.freedesktop.org>
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5148>
Bas Nieuwenhuizen [Mon, 25 May 2020 13:53:59 +0000 (15:53 +0200)]
frontend/dri: Implement mapping individual planes.
It is kinda surprising that
image2 = fromPlanar(image, 2, NULL)
mapImage(..., image2, ...)
does not map the third plane.
This implements that behavior in the case where the DRI frontend
lowers the multi-planar textures.
In the case it doesn't this would need driver support. AFAIU at
least etnaviv is impacted, and while it looks possible, I don't
have the etnaviv knowledge to implement it.
Instead of silently returning weird results (either always plane 0
or possibly something interleaved) this adds an error return on
mapping multi-planar textures otherwise.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5200>
Vinson Lee [Wed, 27 May 2020 23:19:25 +0000 (16:19 -0700)]
zink: Check fopen result.
Fix warning reported by Coverity.
Dereference null return value (NULL_RETURNS)
dereference: Dereferencing a pointer that might be NULL fp when calling
fwrite.
Fixes: 8d46e35d16e3 ("zink: introduce opengl over vulkan")
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5235>
Samuel Pitoiset [Wed, 20 May 2020 06:41:46 +0000 (08:41 +0200)]
radv/aco: enable VK_EXT_subgroup_size_control
ACO should already support Wave32 on GFX10 with all shader stages
and CTS pass. RADV currently only allows Wave32 with the compute
shader stage.
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/5056>
Rob Clark [Wed, 27 May 2020 20:50:05 +0000 (13:50 -0700)]
freedreno/a6xx: document LRZ flag buffer
Doesn't seem to be a big win, although I could still be missing
something in my implementation. But might as well add the
documentation.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5217>
Rob Clark [Tue, 26 May 2020 20:58:22 +0000 (13:58 -0700)]
freedreno/a6xx: LRZ fix for alpha-test
Similarly to stencil-test, if alpha-test is enabled, we don't know
necessarily whether the fragment will pass.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/3045
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5217>
Neha Bhende [Wed, 27 May 2020 15:15:29 +0000 (20:45 +0530)]
util: Initialize pipe_shader_state for passthrough and transform shaders
mesa/st is initializing pipe_shader_state for user define shaders.
This patch intialized pipe_shader_state for all passthough
and transform shaders.
This fixes crashes for several opengl apps. Issue is found in vmware
internal testing
Fixes: f01c0565bb9 ("draw: free the NIR IR.")
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5240>
Chris Wilson [Wed, 27 May 2020 21:19:52 +0000 (22:19 +0100)]
iris: Rename iris_seqno to iris_fine_fence
Rename iris_seqno to iris_fine_fence, borrowed from si_fine_fence, to
avoid introducing any confusion with any other seqno used for tracking
pipelines.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5233>
Gert Wollny [Tue, 19 May 2020 15:52:31 +0000 (17:52 +0200)]
nir: lower_tex: Don't normalize coordinates for TXF with RECT
v2: remove the option to actually request normalization and its
application in Intel < Gen6 (Jason)
v3: Also don't lower for query operations (Jason)
Fixes: 1ce8060c25c7f2c7a54159fab6a6974c0ba182a8
nir/lower_tex: support for lowering RECT textures
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5105>
Eric Anholt [Tue, 26 May 2020 20:28:30 +0000 (13:28 -0700)]
ci: Quick exit qpa extraction for non-matching qpas.
When you're bringing up a new driver in CI with significant number of
failures (or when a CI run breaks a driver), the QPA extraction can easily
take the whole job timeout as we go about processing each QPA (100 of them
in my early VK CI fails) per unexpected result we're saving (50), which
involves reading and each line of the file in shell. By quickly filtering
out the QPA files not including our test, we can save all that shell
overhead, bringing QPA extract time down to a couple of minutes.
Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Reviewed-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5225>
Eric Anholt [Fri, 22 May 2020 22:48:55 +0000 (15:48 -0700)]
ci: Move baremetal DEQP_NO_SAVE_RESULTS setup to the yml.
I'm going to want it unset (artifacts enabled) for the cheza jobs.
Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5225>
Eric Anholt [Tue, 26 May 2020 17:53:05 +0000 (10:53 -0700)]
ci: Add DEQP_EXPECTED_RENDERER support for VK tests.
I used this to debug what was going on with freedreno VK in CI.
Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5225>
Eric Anholt [Tue, 26 May 2020 18:18:33 +0000 (11:18 -0700)]
ci: Auto-detect the architecture for VK ICD filenames.
Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5225>
Eric Anholt [Tue, 26 May 2020 18:07:41 +0000 (11:07 -0700)]
ci: Drop old comment about enabling --deqp-watchdog.
The parallel deqp runner does its own 60s watchdog.
Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5225>
Eric Anholt [Tue, 26 May 2020 17:15:19 +0000 (10:15 -0700)]
ci: Drop double ".txt" suffix on the unexpected results file.
Just a cosmetic fix in reviewing logs.
Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5225>
Samuel Pitoiset [Thu, 28 May 2020 08:21:27 +0000 (10:21 +0200)]
spirv,radv,anv: implement no-op VK_GOOGLE_user_type
This extension only allows HLSL shader compilers to optionally embed
unambiguous type information which can be safely ignored by the driver.
This fixes a crash with the recent Vulkan backend of Path Of Exile
(it uses the extension without checking if it's supported).
Cc: <mesa-stable@lists.freedesktop.org>
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Tested-by: Edmondo Tommasina <edmondo.tommasina@gmail.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5237>
Rhys Perry [Mon, 4 May 2020 13:11:00 +0000 (14:11 +0100)]
aco: fix 64-bit shared_atomic_exchange
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4880>
Rhys Perry [Fri, 1 May 2020 10:04:27 +0000 (11:04 +0100)]
aco: don't reorder barriers in the scheduler
Unless we're reordering it around a barrier of the same type
No shader-db changes.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4880>
Rhys Perry [Thu, 30 Apr 2020 15:10:42 +0000 (16:10 +0100)]
aco: preserve more fields when combining additions into SMEM
Totals from 11 (0.01% of 127638) affected shaders:
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Fixes: 93c8ebfa780 ('aco: Initial commit of independent AMD compiler')
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4880>
Rhys Perry [Wed, 29 Apr 2020 16:47:50 +0000 (17:47 +0100)]
aco: check instruction format before waiting for a previous SMEM store
Totals from 7 (0.01% of 127638) affected shaders:
CodeSize: 40336 -> 40320 (-0.04%)
Instrs: 7807 -> 7803 (-0.05%)
Cycles: 118588 -> 118344 (-0.21%); split: -0.23%, +0.02%
SMEM: 331 -> 339 (+2.42%)
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Fixes: 1749953ea3 ('aco/gfx10: Wait for pending SMEM stores before loads')
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4880>
Rhys Perry [Fri, 22 May 2020 14:42:39 +0000 (15:42 +0100)]
aco: consider SDWA during value numbering
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Fixes: 23ac24f5b1fdde73cf8ec1ef6cbe08d73d6776f5
('aco: add missing conversion operations for small bitsizes')
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5164>
Rhys Perry [Fri, 22 May 2020 14:42:12 +0000 (15:42 +0100)]
aco: fix interaction with 3f branch workaround and p_constaddr
The offset was incorrect if we inserted a nop before the p_constaddr.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Fixes: 93c8ebfa ('aco: Initial commit of independent AMD compiler')
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5164>
Michel Dänzer [Tue, 26 May 2020 09:04:19 +0000 (11:04 +0200)]
gitlab-ci: Pull in GCC 9 from Debian testing in x86_test-gl/vk images
The GCC 8 packages from buster are no longer compatible with libc6 from
testing. We could use the GCC 8 packages from testing instead, but this
is easier.
v2:
* Update piglit-quick_gl test results, due to the piglit issue fixed by
https://gitlab.freedesktop.org/mesa/piglit/-/merge_requests/294
Reviewed-by: Eric Anholt <eric@anholt.net> # v1
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5186>
Michel Dänzer [Sat, 16 May 2020 15:17:23 +0000 (17:17 +0200)]
gitlab-ci: x86_test-base image as common base for x86_test-gl/vk
Making use of the relatively recent FDO_BASE_IMAGE feature of the
templates, the x86_test-base image contents are shared as a separate
layer by the x86_test-gl/vk images (meaning the former only needs to be
downloaded once for either or both of the latter). This should be more
efficient in terms of overall network bandwidth and storage, in
particular if the base image changes less often than the -gl/vk ones.
v2:
* List x86_test-base in needs: along with x86_test-gl/vk (see parent
commit)
* Always put $STABLE/TESTING_EPHEMERAL on separate lines, will make it
easier to add any non-ephemeral packages
Reviewed-by: Eric Anholt <eric@anholt.net> # v1
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5186>
Michel Dänzer [Sat, 16 May 2020 10:07:42 +0000 (12:07 +0200)]
gitlab-ci: Also list arm/x86_build in needs: of test jobs
Without this, the test jobs may spuriously run if the arm/x86_build
jobs fail.
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5186>
Caio Marcelo de Oliveira Filho [Wed, 27 May 2020 15:05:41 +0000 (08:05 -0700)]
intel: Add helper to calculate GPGPU_WALKER::RightExecutionMask
Suggested by Jason.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5142>
Caio Marcelo de Oliveira Filho [Thu, 21 May 2020 06:32:18 +0000 (23:32 -0700)]
iris, i965: Update limits for ARB_compute_variable_group_size
The CS compiler now produces multiple SIMD variants, so the previous
trade-off between "always using SIMD32" and "having a smaller max
invocations" is now gone.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5142>
Caio Marcelo de Oliveira Filho [Thu, 21 May 2020 06:55:23 +0000 (23:55 -0700)]
iris, i965: Drop max_variable_local_size
This was used to decide which SIMD width to generate code for
ARB_compute_variable_group_size. Now that compiler will generate
multiple SIMD widths, this information is unused.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5142>
Caio Marcelo de Oliveira Filho [Thu, 21 May 2020 08:56:54 +0000 (01:56 -0700)]
intel/fs: Generate multiple CS SIMD variants for variable group size
This will make the GL drivers pick the right SIMD variant for a given
group size set during dispatch. The heuristic implemented in
brw_cs_simd_size_for_group_size() is the same as in brw_compile_cs().
The cs_prog_data::simd_size field was removed. The generated SIMD
sizes are marked in a bitmask, which is already used via
brw_cs_simd_size_for_group_size() by the drivers.
When in variable group size, it is OK if larger SIMD shader spill,
since we'd need it for the cases where the smaller one can't hold all
the invocations.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5142>
Caio Marcelo de Oliveira Filho [Thu, 21 May 2020 07:17:27 +0000 (00:17 -0700)]
anv: Use new helper functions to pick SIMD variant for CS
Also combine the existing individual anv helpers into a single one for
all CS related parameters.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5142>
Caio Marcelo de Oliveira Filho [Thu, 21 May 2020 07:09:19 +0000 (00:09 -0700)]
iris: Use new helper functions to pick SIMD variant for CS
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5142>
Caio Marcelo de Oliveira Filho [Thu, 21 May 2020 09:34:52 +0000 (02:34 -0700)]
iris: Set CS KernelStatePointer at dispatch
There's an update for INTERFACE_DESCRIPTOR_DATA at dispatch, so we can
just move the KSP assignment there. This flexibility will later allow
variable group size to pick the right SIMD variant.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5142>
Caio Marcelo de Oliveira Filho [Thu, 21 May 2020 07:51:44 +0000 (00:51 -0700)]
i965: Use new helper functions to pick SIMD variant for CS
Also expand the existing i965 helper to return the other CS related
paramters.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5142>
Caio Marcelo de Oliveira Filho [Thu, 21 May 2020 09:26:21 +0000 (02:26 -0700)]
intel/fs: Add helper to get prog_offset and simd_size
This indirection will be used by the variable group size case in a
later change.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5142>
Caio Marcelo de Oliveira Filho [Tue, 19 May 2020 17:08:12 +0000 (10:08 -0700)]
intel/fs: Support INTEL_DEBUG=no8,no32 in compute shaders
The "no32" flag will have precedence over "do32", like is done for FS.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5142>
Caio Marcelo de Oliveira Filho [Tue, 19 May 2020 21:37:44 +0000 (14:37 -0700)]
intel/fs: Remove min_dispatch_width spilling decision from RA
Move the decision one level up, let brw_compile_*() functions use the
spilling information to decide whether or not a certain width
compilation can spill (passed via run_*() functions).
The min_dispatch_width was used to compare with the dispatch_width and
decide whether "a previous shader is already available, so don't
accept spill".
This is replaced by:
- Not calling run_*() functions if it is know beforehand a smaller width
already spilled -- since the larger width will spill and fail;
- Explicitly passing whether or not a shader is allowed to spill. For
the cases where the smaller width is available and haven't spilled,
the larger width will be compiled but is only useful if it won't
spill.
Moving the decision to this level will be useful later for variable
group size, which is a case where we want all the widths to be allowed
to spill.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5142>
Eric Engestrom [Wed, 27 May 2020 20:27:17 +0000 (22:27 +0200)]
docs: update calendar, add news item, and link releases notes for 20.1.0
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5234>
Eric Engestrom [Wed, 27 May 2020 21:46:48 +0000 (23:46 +0200)]
docs: Add release notes for 20.1.0
Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5234>
Mike Blumenkrantz [Wed, 27 May 2020 19:28:42 +0000 (15:28 -0400)]
zink: always use logical eq ops in ntv with 1bit inputs
integer and float compare ops cannot take boolean types, so the bit size
of the inputs should be checked here so that we can swap to the logical
equality functions if we're being passed a bool value
resolves tons of validator errors in glsl piglit tests
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5231>
Vinson Lee [Wed, 27 May 2020 00:26:47 +0000 (17:26 -0700)]
pan/bi: Initialize struct fma_op_info member extended.
Fix warning reported by Coverity Scan.
Uninitialized scalar variable (UNINIT)
uninit_use: Using uninitialized value info. Field info.extended is
uninitialized.
Fixes: 8c79c710d4e1 ("pan/bi: Identify extended FMA opcodes")
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5224>
Erico Nunes [Sun, 17 May 2020 13:56:42 +0000 (15:56 +0200)]
lima/ppir: use a ready list in node_to_instr
After the recent optimizations in ppir lowering that increase options
for combining, node_to_instr now may have multiple options of nodes to
insert and needs to decide which is better.
For example, if an instruction uses both a varying and a texture, there
are two nodes nodes that can be inserted to the load varying slot in the
same instruction (ld_var and ld_coords). It is much more advantageous to
pipeline the load texture coords since that enables the higher precision
path for texture coordinates. However, with the current recursive
expansion, this cannot be influenced.
This simple ready list implementation in node_to_instr allows it to
choose the next node to expand based on a priority score, rather than
relying on the random order coming from the recursive expansion.
Other than preferring nodes with pipeline output (which covers ld_coords
vs ld_var), nodes using later slots in the pipeline are now expanded
first, allowing node_to_instr to make all of the earlier (pipelineable)
nodes available in the ready list so the best one can be chosen when
picking nodes for the earlier slots.
Fixes: 632a921bd0d lima/ppir: optimize tex loads with single successor
Signed-off-by: Erico Nunes <nunes.erico@gmail.com>
Reviewed-by: Vasily Khoruzhick <anarsoul@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5092>
Alyssa Rosenzweig [Wed, 27 May 2020 18:53:05 +0000 (14:53 -0400)]
pan/bi: Suppress inf/nan for now
This is a (hopefully temporary) hack. The blob does it for ES2 at any
rate.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Wed, 27 May 2020 16:32:28 +0000 (12:32 -0400)]
pan/bi: Add CSEL.16 packing tests
Passing but let's increase coverage.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Wed, 27 May 2020 15:52:57 +0000 (11:52 -0400)]
pan/bi: Pack compact vertex texturing
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Wed, 27 May 2020 15:43:37 +0000 (11:43 -0400)]
pan/bi: Add f16 TEXC.vtx op
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Wed, 27 May 2020 15:41:42 +0000 (11:41 -0400)]
pan/bi: Allow vertex txl with lod=0 as compact
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Wed, 27 May 2020 15:33:26 +0000 (11:33 -0400)]
pan/bi: Document compute_lod bit for compact tex
At least I assume this works this way.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Wed, 27 May 2020 15:28:11 +0000 (11:28 -0400)]
pan/bi: Also add compact vertex texturing
This implies lod=0.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Wed, 27 May 2020 15:24:49 +0000 (11:24 -0400)]
pan/bi: Add TEX.vtx opcode for vertex texturing
Always has an LOD.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Wed, 27 May 2020 20:38:47 +0000 (16:38 -0400)]
pan/decode: Decode Bifrost shader flags
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Tue, 26 May 2020 23:48:25 +0000 (19:48 -0400)]
panfrost: Set MALI_BIFROST_EARLY_Z as necessary
Fixes blending.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Wed, 27 May 2020 20:29:01 +0000 (16:29 -0400)]
panfrost: Identify MALI_BIFROST_EARLY_Z flag
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Wed, 27 May 2020 20:26:29 +0000 (16:26 -0400)]
panfrost: Add defines for bifrost unk1 flags
Instead of open-coding.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Wed, 27 May 2020 20:07:00 +0000 (16:07 -0400)]
panfrost: Document Midgard Inf/NaN suppress bit
We should probably not be setting this..
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Wed, 27 May 2020 14:58:17 +0000 (10:58 -0400)]
panfrost: Ensure nonlinear strides are 16-aligned
To match how they are encoded.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Fixes: bde19c0e7ba ("panfrost: Fix tiled texture "stride"s on Bifrost")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Tue, 26 May 2020 23:48:29 +0000 (19:48 -0400)]
panfrost: Identify Bifrost texture format swizzle
We don't force w=1 for Bifrost textures. We already compose this into
the swizzle as necessary, so we can just ignore this field I think. But
let's identify it so we don't forget what it is.
The blob uses it to force w=1 for <= 3-channel formats (0x10), as well
as a flag to swap r/b for BGRA (0x4). There are probably other flags
here but it doesn't.. really matter to us.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Tue, 26 May 2020 22:50:29 +0000 (18:50 -0400)]
panfrost: Set unk2 to accomodate blending
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Tue, 26 May 2020 22:48:04 +0000 (18:48 -0400)]
panfrost: Share MRT blend flag calculation with Bifrost
As far as I know the field is the same.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Tue, 26 May 2020 22:21:57 +0000 (18:21 -0400)]
panfrost: Force Z/S tiling on Bifrost
Like we do on SFBD since we don't know the format bits yet.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Tue, 26 May 2020 22:20:56 +0000 (18:20 -0400)]
panfrost: Tweak Bifrost colour buffer magic
For tiled or linear.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Tue, 26 May 2020 22:19:19 +0000 (18:19 -0400)]
panfrost: Tweak zsbuf magic numbers for Bifrost
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Tue, 26 May 2020 22:16:51 +0000 (18:16 -0400)]
panfrost: Adjust null_rt for Bifrost
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
Alyssa Rosenzweig [Mon, 4 May 2020 16:48:50 +0000 (12:48 -0400)]
panfrost: Fix Bifrost blending with depth-only FBO
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5232>
James Zhu [Tue, 26 May 2020 13:28:41 +0000 (09:28 -0400)]
ac/gpu_info: Correct Acturus cu bitmap
The cu bitmap in amd gpu info structure is
4x4 size array, and it's usually suitable for Vega
ASICs which has 4*2 SE/SH layout.
But for Arcturus, SE/SH layout is changed to 8*1.
To mostly reduce the impact, we make it compatible
with current bitmap array as below:
SE4,SH0 --> cu_bitmap[0][1]
SE5,SH0 --> cu_bitmap[1][1]
SE6,SH0 --> cu_bitmap[2][1]
SE7,SH0 --> cu_bitmap[3][1]
Signed-off-by: James Zhu <James.Zhu@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5212>
Danylo Piliaiev [Thu, 14 May 2020 12:29:36 +0000 (15:29 +0300)]
intel/fs: Work around dual-source blending hangs in combination with SIMD16
It was found that dual-source blending hangs with SIMD16 dispatch in some
specific but unknown situation. Which in the wild happen when rgba
anti-aliasing is enabled for fonts.
Cc: <mesa-stable@lists.freedesktop.org>
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/2183
Signed-off-by: Danylo Piliaiev <danylo.piliaiev@globallogic.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5037>
Erik Faye-Lund [Mon, 25 May 2020 10:06:04 +0000 (12:06 +0200)]
zink: use general-layout when blitting to/from same resource
This avoids a validator warning when for instance generating mipmaps.
Fixes: d2bb63c8d4c ("zink: Use optimal layout instead of general. Reduces valid layer warnings. Fixes RADV image noise.")
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5199>
Pierre-Eric Pelloux-Prayer [Fri, 22 May 2020 16:38:24 +0000 (18:38 +0200)]
radeonsi/drirc: enable zerovram option for 7 Days to Die
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/2686
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/5168>
projects / mesa.git / log