Alejandro Piñeiro [Wed, 26 May 2021 21:21:31 +0000 (23:21 +0200)]
v3d/simulator: use the proper register when waiting on a CSD submit
Until now we were waiting until having a dispatch current and/or
queued. But that would only wait for all shaders to have started, it
won't wait for them to have finished.
With this commit we wait until the NUM_COMPLETED_JOBS (that in spite
of that name, it is about dispatches) field got increased.
This is in general safest, and it is needed after the latest simulator
update to get CTS tests like the following ones working:
dEQP-VK.compute.basic.copy_ssbo_multiple_invocations
dEQP-VK.compute.basic.copy_ssbo_single_invocation
dEQP-VK.compute.basic.ssbo_rw_single_invocation
dEQP-VK.compute.basic.ssbo_unsized_arr_single_invocation
dEQP-VK.compute.basic.ubo_to_ssbo_multiple_invocations
dEQP-VK.compute.basic.ubo_to_ssbo_single_invocation
v2 (from Juan feedback):
* Clarify JOBS vs DISPATCHES
Reviewed-by: Juan A. Suarez <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11039>
Alejandro Piñeiro [Tue, 25 May 2021 21:52:48 +0000 (23:52 +0200)]
v3d/simulator: wait for cache flushes
Current code just assumes that flushes are instant, as simulator
doesn't really model the caches. So right now we have just an assert
that the flush has been done.
But that can change on the future, so let's change the assert for a
wait.
Note that for the l1t case we are writing on the field TMUWCF. So I
understand that then we need to wait for TMUWCF_SET, even if the
previous code was using L2TFLS_SET.
This also happpens on the kernel side. We need to check if this was a
typo on the kernel side.
v2 (from Juan feedback)
* Add comment about the TMUWCF vs L2TFLS difference between this
commit and the kernel.
Reviewed-by: Juan A. Suarez <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11039>
Alejandro Piñeiro [Sun, 23 May 2021 21:09:58 +0000 (23:09 +0200)]
v3d/simulator: add a cache flush mode enum
Makes the write to the l2t cache control more readable (without magic
numbers).
Reviewed-by: Juan A. Suarez <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11039>
Alejandro Piñeiro [Fri, 7 May 2021 23:11:42 +0000 (01:11 +0200)]
v3d/simulator: capture hub interrupts
So far we were not capturing any HUB interrupt, just core. This could
be a problem if any is fired, as we could enter on an infinite
loop. With this commit we start to capture them. So we split v3d_isr
into core and hub interrupt handling.
As reference we capture the same HUB interrupts that we capture on the
v3d kernel support.
It is worth to note that all those are mostly untested. Now with both
opengl/vulkan driver being stable we were not able to raise those
interrupts.
v2 (Juan feedback):
* Just one V3D_VERSION >= 41 block, more readable
* Assert that the core is 0 at v3d_isr_core (we don't handle
multi-core right now).
Reviewed-by: Juan A. Suarez <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11039>
Alejandro Piñeiro [Wed, 28 Apr 2021 22:57:25 +0000 (00:57 +0200)]
broadcom/compiler: use proper type field for atomic operations
We were using the num_components to infer it, but in the end it is
VEC2 for CMPXCHG and 32BIT for anything else.
This doesn't affect any test with the real hw, but fixes an assert
with the last version of the simulator.
Reviewed-by: Juan A. Suarez <jasuarez@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11039>
Tomeu Vizoso [Fri, 28 May 2021 09:41:20 +0000 (11:41 +0200)]
ci/lava: Disable CPU frequency scaling
Lock CPU frequency scaling to max to speed up test execution and lower
the variation of frame times from performance replay jobs.
Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Corentin Noël <corentin.noel@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11101>
Tomeu Vizoso [Fri, 28 May 2021 09:20:50 +0000 (11:20 +0200)]
Revert "ci/freedreno: Skip Portal 2 trace on a630, due to flakiness"
This reverts commit
e381bc0e6711de83d393ef3cc0b6b795772c1813.
Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Corentin Noël <corentin.noel@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11058>
Samuel Pitoiset [Fri, 28 May 2021 12:56:04 +0000 (14:56 +0200)]
ac/rgp: bump the SQTT file minor version to 5
To match latest RGP spec. Captures generated by RADV still work
with latest RGP (v1.10).
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/11063>
Samuel Pitoiset [Fri, 28 May 2021 12:55:40 +0000 (14:55 +0200)]
ac/rgp: mark SQTT_FILE_CHUNK_TYPE_ISA_DATABASE as deprecated
This is now deprecated and reserved for future uses.
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/11063>
Alyssa Rosenzweig [Tue, 1 Jun 2021 00:08:50 +0000 (20:08 -0400)]
asahi: Allocate slices for mipmapping
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Tue, 1 Jun 2021 00:08:57 +0000 (20:08 -0400)]
asahi: Set levels in texture descriptor
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Tue, 1 Jun 2021 00:08:37 +0000 (20:08 -0400)]
asahi: Add mipmapping state to the XML
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Mon, 31 May 2021 23:21:49 +0000 (19:21 -0400)]
asahi: Abort on blit()
Unimplemented.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Mon, 31 May 2021 22:59:02 +0000 (18:59 -0400)]
asahi: Make data_valid a bitset to save memory
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Mon, 31 May 2021 20:35:36 +0000 (16:35 -0400)]
asahi: Allow half-float vertex buffers
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Mon, 31 May 2021 20:35:21 +0000 (16:35 -0400)]
asahi: Flesh out the formats table
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Mon, 31 May 2021 20:34:23 +0000 (16:34 -0400)]
asahi: Fake CAPs for ES3 with AGX_MESA_DEBUG=deqp
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Mon, 31 May 2021 20:10:40 +0000 (16:10 -0400)]
asahi: Lift streamout scaffolding from Panfrost
Trying to fake ES3 for dEQP-GLES3.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Mon, 31 May 2021 20:10:02 +0000 (16:10 -0400)]
asahi: Add "hacks for dEQP" flag
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Mon, 31 May 2021 19:46:18 +0000 (15:46 -0400)]
asahi: Add ETC2 formats to table
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Mon, 31 May 2021 19:37:40 +0000 (15:37 -0400)]
asahi: Respect render target format swizzle
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Mon, 31 May 2021 19:34:40 +0000 (15:34 -0400)]
asahi: Use pixel table in is_format_supported
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Mon, 31 May 2021 19:29:20 +0000 (15:29 -0400)]
asahi: Scaffold format table
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Mon, 31 May 2021 19:08:55 +0000 (15:08 -0400)]
asahi: Hide pixel formats behind an opaque type
Convenient for mapping a format table.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Mon, 31 May 2021 18:37:06 +0000 (14:37 -0400)]
asahi: Add format enums
Split it up into the underlying parts. Now we can decode every format
accessible in Metal.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Mon, 31 May 2021 18:36:56 +0000 (14:36 -0400)]
asahi: Print unknown enum values
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Sun, 30 May 2021 22:54:40 +0000 (04:24 +0530)]
asahi: Align strides to 16 bytes
Apparently required for texturing.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Sun, 30 May 2021 22:39:51 +0000 (04:09 +0530)]
asahi: Note that "render target" lacks an sRGB bit
sRGB is handled through other mechanisms.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Mon, 31 May 2021 22:08:14 +0000 (18:08 -0400)]
asahi: Handle linear display targets as well as tiled
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Sun, 30 May 2021 22:03:05 +0000 (03:33 +0530)]
asahi: Respect linear strides
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Sun, 30 May 2021 22:01:43 +0000 (03:31 +0530)]
asahi: Identify line stride in texture/RT XML
The off-by-four is concerning.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Sun, 30 May 2021 21:55:34 +0000 (03:25 +0530)]
asahi: Translate layouts for texture and RTs
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Sun, 30 May 2021 21:49:35 +0000 (03:19 +0530)]
asahi: Add layout enum to XML
Used for textures and render targets. There are more values here, probably for
other tiling schemes and compression.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Sun, 30 May 2021 21:45:10 +0000 (03:15 +0530)]
asahi: Use dt_stride for line_stride where needed
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Sun, 30 May 2021 21:31:07 +0000 (03:01 +0530)]
asahi: Determine tiling vs linear for internal textures
Logic lifted from Panfrost.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Sun, 30 May 2021 17:18:37 +0000 (22:48 +0530)]
asahi: Implement wide lines
Identify line width field and route through the Gallium line width.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Alyssa Rosenzweig [Fri, 28 May 2021 02:51:09 +0000 (22:51 -0400)]
asahi: Fix off-by-one in viewport scissoring
As discovered during the stream. Typo pointed out by someone in the Twitch chat.
Fixes:
f285a1220ea ("asahi: Implement scissors and scissor to viewport")
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11086>
Rob Clark [Mon, 31 May 2021 20:24:04 +0000 (13:24 -0700)]
freedreno/headergen2: Fix compile warnings with CP_DRAW_INDIRECT_MULTI
Using stripes to deal with the different packet layout variants resulted
in redefining "register" offsets with different values, so use "prefix"
to add a suffix to disambiguate.
drivers/gpu/drm/msm/adreno/adreno_pm4.xml.h:1066: warning: "REG_A6XX_CP_DRAW_INDIRECT_MULTI_INDIRECT" redefined
1066 | #define REG_A6XX_CP_DRAW_INDIRECT_MULTI_INDIRECT 0x00000006
|
drivers/gpu/drm/msm/adreno/adreno_pm4.xml.h:1057: note: this is the location of the previous definition
1057 | #define REG_A6XX_CP_DRAW_INDIRECT_MULTI_INDIRECT 0x00000003
|
(Admittedly it isn't really a "prefix" but that was the field in the
schema available to use, and REG_INDEXED_CP_DRAW_INDIRECT_MULTI_STRIDE
sounds somewhat more funny.)
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10944>
Rob Clark [Mon, 24 May 2021 18:00:28 +0000 (11:00 -0700)]
freedreno/afuc: Use emulator to extract jmptbl
This runs through the SQE bootstrap code to extract the packet-table,
rather than relying on heuristics. As a bonus, it can detect the start
of the LPAC fw in a660+ fw so that we can properly decode the LPAC fw
and packet-table.
Note that this decodes the jmptable as normal instructions, which is a
change in behavior from the previous heuristic based jmptbl extraction.
Not sure if that is a good or bad thing.
For a5xx, for now the legacy heuristic based jmptable decoding is
preserved, at least until enough control regs are figured out.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10944>
Rob Clark [Thu, 27 May 2021 21:37:59 +0000 (14:37 -0700)]
freedreno/ci: Add real packet-table loading for afuc test
When we start running the bootstrap code thru the emulator we will need
the packet-table loading to actually happen. So add this.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10944>
Rob Clark [Thu, 27 May 2021 19:18:53 +0000 (12:18 -0700)]
freedreno/afuc: Add emulator support to run bootstrap
Run until the packet-table is populated, so the disassembler can use
this to know the offsets of various pm4 packet handlers without having
to rely on heuristics.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10944>
Rob Clark [Thu, 27 May 2021 18:35:21 +0000 (11:35 -0700)]
freedreno/afuc: Split out helpers to parse labels and packet-table
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10944>
Rob Clark [Thu, 27 May 2021 18:04:42 +0000 (11:04 -0700)]
freedreno/afuc: Extract full gpu-id
Some of the a6xx gens will require some control reg initialization, and
go into an infinite loop if they don't see the values they expect, so
we'll need to extract the compute gpu-id.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10944>
Rob Clark [Sun, 23 May 2021 17:46:07 +0000 (10:46 -0700)]
freedreno/registers: Add a few a6xx regs and notes
A few things I noticed while playing with the emulator.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10944>
Rob Clark [Sun, 16 May 2021 20:09:29 +0000 (13:09 -0700)]
freedreno/afuc: Add emulator mode to afuc-disasm
This is an (at least somewhat complete) logical emulator of the a6xx SQE
that lets us step through firmware execution (bootstrap, cmdstream pkt
handling, etc). It lets us poke at various fw visible state and run
through pm4 packet(s) to better understand what the fw is doing when it
handles various packets.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10944>
Rob Clark [Fri, 21 May 2021 15:39:08 +0000 (08:39 -0700)]
freedreno/afuc: Add pipe reg name decoding
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10944>
Rob Clark [Fri, 21 May 2021 00:12:35 +0000 (17:12 -0700)]
freedreno/afuc: Clean up special regs
Allow for different mnemonics depending on whether they are used as
source or destination register, to better reflect what they do.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10944>
Rob Clark [Mon, 17 May 2021 01:43:20 +0000 (18:43 -0700)]
freedreno/afuc: Split out utils
With disasm emulator mode, we'll start wanting some things that are
duplicationg what the assembler does, so just split out all the rnndb
bits into shared utils.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10944>
Rob Clark [Sun, 16 May 2021 19:15:31 +0000 (12:15 -0700)]
freedreno/afuc: Split out instruction decode helper
Split the giant switch/decode out into a helper function so that we can
re-use it for emulator mode.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10944>
Rob Clark [Fri, 21 May 2021 20:58:13 +0000 (13:58 -0700)]
freedreno: Move pkt parsing helpers to common
I'll be needing these in afuc as well.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10944>
Rob Clark [Fri, 21 May 2021 19:16:37 +0000 (12:16 -0700)]
freedreno/tu+drm: Extract out pm4 pkt header helpers
I'm going to need these in a 3rd place, so let's deduplicate first.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10944>
Andres Gomez [Mon, 31 May 2021 18:41:58 +0000 (21:41 +0300)]
ci: replace glslangValidator with glslang-tools
Signed-off-by: Andres Gomez <agomez@igalia.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11096>
Andres Gomez [Mon, 31 May 2021 17:50:30 +0000 (20:50 +0300)]
ci: remove glslangValidator installation from the VK test container
We already include glslang-tools.
Signed-off-by: Andres Gomez <agomez@igalia.com>
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11095>
Mike Blumenkrantz [Sun, 16 May 2021 14:16:41 +0000 (10:16 -0400)]
compiler/spirv: expand_to_vec4 -> nir_pad_vec4
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10821>
Rhys Perry [Mon, 31 May 2021 14:35:03 +0000 (15:35 +0100)]
radv: make attrib_end variable in radv_flush_vertex_descriptors 32-bit
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Fixes:
1e9dc0474ef ("radv: make radv_pipeline::attrib_ends 32bit")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11089>
Danylo Piliaiev [Mon, 31 May 2021 08:41:22 +0000 (11:41 +0300)]
freedreno: reduce the upper bound of IB size by one
Going beyond 0x100000 results in hangs, however I found that the
last 0x100000 packet just doesn't get executed. Thus the real limit is
0x0FFFFF. At least this is true for a6xx.
This could be tested by appending nops to the cmdstream and placing
e.g. CP_INTERRUPT at the end, at any position other than being
0x100000 packet it results in a hang.
Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10786>
Danylo Piliaiev [Thu, 13 May 2021 13:50:45 +0000 (16:50 +0300)]
turnip: place a limit on the growth of BOs
There is a limit on IB size, which on freedreno is set to 0x100000.
Going beyond it results in hangs, however I found that the last
0x100000 packet just doesn't get executed. Thus the real limit is
0x0FFFFF.
This could be tested by appending nops to the cmdstream and placing
e.g. CP_INTERRUPT at the end, at any position other than being
0x100000 packet it results in a hang.
Fixes:
dEQP-VK.api.command_buffers.record_many_draws_secondary_2
dEQP-VK.api.command_buffers.record_many_draws_primary_2
However these tests could trigger hangcheck timeouts.
Also this fixes hangs when opening captures of games in RenderDoc.
Signed-off-by: Danylo Piliaiev <dpiliaiev@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10786>
Georg Lehmann [Fri, 28 May 2021 12:53:25 +0000 (14:53 +0200)]
v3dv: use VKAPI_ATTR and VKAPI_CALL.
Closes #4852
Signed-off-by: Georg Lehmann <dadschoorse@gmail.com>
Tested-by: Roman Stratiienko <r.stratiienko@gmail.com>
Acked-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11062>
Neil Roberts [Mon, 31 May 2021 14:32:57 +0000 (16:32 +0200)]
kmsro: Fix confusing comma expression
It looks like this was a copy-and-paste mistake in
827e0d6654a2 where
the initialiser was moved from being a struct initialiser to a
standalone statement. Some of them were fixed with an unrelated change
in
187218395d7c but not all of them. This shouldn’t make any practical
difference to the compiled code.
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11090>
Mike Blumenkrantz [Sat, 29 May 2021 03:01:12 +0000 (23:01 -0400)]
Revert "zink: call tc_driver_internal_flush_notify() on flush"
This reverts commit
4d62fca0877104080d4dee8656d09c67e93c8d0e.
this is broken beyond space and time but only reveals its true nature
when used outside of a unit test
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11080>
Mike Blumenkrantz [Tue, 26 Jan 2021 20:58:53 +0000 (15:58 -0500)]
zink: stop using util_set_vertex_buffers_mask()
we're already iterating this, so there's no point in double iterating
also no point in doing the extra refcounting
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10997>
Mike Blumenkrantz [Tue, 26 Jan 2021 20:58:29 +0000 (15:58 -0500)]
zink: ref vertex buffers during set_vertex_buffers
now we don't need to do it during bind
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10997>
Mike Blumenkrantz [Tue, 26 Jan 2021 15:37:32 +0000 (10:37 -0500)]
zink: update scissor only when necessary
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10997>
Mike Blumenkrantz [Tue, 26 Jan 2021 15:31:24 +0000 (10:31 -0500)]
zink: only update viewport state when necessary
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10997>
Mike Blumenkrantz [Tue, 26 Jan 2021 15:27:32 +0000 (10:27 -0500)]
zink: only rebind vertex buffers when necessary
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10997>
Mike Blumenkrantz [Sat, 30 Jan 2021 14:06:25 +0000 (09:06 -0500)]
zink: handle rebinds for vertex buffers
now that the tracking info is there, it's easy to make this happen
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10997>
Mike Blumenkrantz [Tue, 26 Jan 2021 15:27:50 +0000 (10:27 -0500)]
zink: only rebind pipelines when necessary
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10997>
Mike Blumenkrantz [Tue, 26 Jan 2021 15:19:51 +0000 (10:19 -0500)]
zink: mark some draw functions inline
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10997>
Mike Blumenkrantz [Tue, 26 Jan 2021 19:52:40 +0000 (14:52 -0500)]
zink: remove barriers/refs from descriptor cache
these are applied on bind now so there's no need to track them
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10997>
Mike Blumenkrantz [Tue, 26 Jan 2021 14:56:04 +0000 (09:56 -0500)]
zink: reapply program refs automatically
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10997>
Mike Blumenkrantz [Tue, 26 Jan 2021 14:37:17 +0000 (09:37 -0500)]
zink: reapply resource/surface refs after app flushes
instead of doing this during descriptor update, we can automatically handle
this rollover
Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/10997>
Iago Toral Quiroga [Thu, 27 May 2021 07:16:18 +0000 (09:16 +0200)]
v3dv: implement VK_KHR_device_group
We only support one device group with a single device, so the
implementation is trivial.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11037>
Iago Toral Quiroga [Thu, 27 May 2021 08:17:08 +0000 (10:17 +0200)]
v3dv: implement interactions of VK_KHR_device_group with VK_KHR_swapchain
There are some interactions between these two extensions that need to be
implemented when both are supported. Particularly:
1. Applications can create images that will be bound to swapchain memory
by passing a VkImageSwapchainCreateInfoKHR in the pNext chain
of VkImageCreateInfo. In this case we need to make sure that the
created image takes some of its parameters from the underlying
swapchain.
2. Applications can bind memory from a swapchain image to a VkImage
by passing a VkBindImageMemorySwapchainInfoKHR in the pNext chain
of VkBindImageMemoryInfo.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11037>
Iago Toral Quiroga [Thu, 27 May 2021 08:00:37 +0000 (10:00 +0200)]
v3dv: create a helper for image creation
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11037>
Iago Toral Quiroga [Thu, 27 May 2021 07:06:00 +0000 (09:06 +0200)]
v3dv: implement vkCmdDispatchBase
This was added with VK_KHR_device_group and allows users to specify
a base offset that will be automatically added to gl_WorkGroupID.
Unfortunately, V3D doesn't support this natively, so we need to add
the base to the workgroup id generated by hardware manually. For this,
we inject add instructions that source from a QUNIFORM that will
retrieve the actual dispatch base from the compute job when it is
dispatched.
Since a compute shader can be dispatched with CmdDispatch and/or
CmdDispatchBase, we always need to add these additional add
instructions and use a base of (0,0,0) for regular dispatches.
Since we don't support any version of OpenGL with this dispatch
base functionality we can avoid the extra instructions there.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11037>
Samuel Pitoiset [Thu, 27 May 2021 07:26:24 +0000 (09:26 +0200)]
radv: enable RADV_DEBUG=invariantgeom for Monster Hunter World
DXVK 1.8.1 marks position as always invariant but the DX12 version
of the game has the same issue and it's not yet fixed on the
vkd3d-proton side.
Fixes some Z-fighting on GFX10.3.
Cc: 21.1 mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11029>
Alyssa Rosenzweig [Fri, 28 May 2021 02:51:09 +0000 (22:51 -0400)]
asahi: Implement scissors and scissor to viewport
Construct a scissor descriptor correpsonding to the intersection of the
framebuffer, the viewport, and the region selected for scissoring by the app.
Use the intersected scissor in the "clip tile" fields in the viewport. Select
this scissor descriptor from the command stream.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11084>
Alyssa Rosenzweig [Sun, 30 May 2021 20:25:19 +0000 (01:55 +0530)]
asahi: Update viewport descriptor depth fields
Acts like a translate/scale pair more generally, not specific to near/far
planes.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11084>
Alyssa Rosenzweig [Sun, 30 May 2021 20:13:15 +0000 (01:43 +0530)]
asahi: Enable depth culling
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11084>
Alyssa Rosenzweig [Sun, 30 May 2021 20:12:43 +0000 (01:42 +0530)]
asahi: Add unknown bits seen with the GL driver
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11084>
Alyssa Rosenzweig [Sun, 30 May 2021 20:11:37 +0000 (01:41 +0530)]
asahi: Expose PIPE_CAP_CLIP_HALFZ
Use the Zink lowering pass to handle the non-halfz case. Metal, like Vulkan,
uses half-z (and Metal is not configurable, making r/e tricky).
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11084>
Alyssa Rosenzweig [Fri, 28 May 2021 03:11:18 +0000 (23:11 -0400)]
agx: Add scissor upload BO
Not sure what the proper data structure for this is yet, but this will
hold over until we start optimizing for memory usage.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11084>
Alyssa Rosenzweig [Sun, 30 May 2021 19:58:25 +0000 (01:28 +0530)]
asahi: Skip draws if the scissor culls everything
We can't pack the scissor descriptor for these, and there would be no rendering
anyway, so detect this condition and skip the draw.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11084>
Alyssa Rosenzweig [Sun, 30 May 2021 20:34:41 +0000 (02:04 +0530)]
asahi: Mark scissor dirty if rast->scissor changes
Although there is a scissor enable bit in the hardware rasterizer state, we
cannot rely on it alone as we also "scissor" to the viewport.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11084>
Alyssa Rosenzweig [Fri, 28 May 2021 02:51:00 +0000 (22:51 -0400)]
asahi: Track scissor states
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11084>
Alyssa Rosenzweig [Fri, 28 May 2021 02:44:55 +0000 (22:44 -0400)]
asahi: Dirty track viewport descriptor
Mitigates the extra CPU cost from packing in the previous commit, and
avoids the redundant memcpy.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11084>
Alyssa Rosenzweig [Fri, 28 May 2021 02:42:48 +0000 (22:42 -0400)]
asahi: Defer viewport pack
Nontrivial interaction between viewport state and scissor state, so
defer until draw time to deal with it.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11084>
Alyssa Rosenzweig [Fri, 28 May 2021 02:33:26 +0000 (22:33 -0400)]
asahi: Add scissor enable bit
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11084>
Alyssa Rosenzweig [Fri, 28 May 2021 02:31:08 +0000 (22:31 -0400)]
asahi: Add "set scissor" command
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11084>
Alyssa Rosenzweig [Fri, 28 May 2021 02:29:02 +0000 (22:29 -0400)]
asahi: Fix scissor descriptor definition
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11084>
Daniel Stone [Sun, 30 May 2021 21:59:13 +0000 (22:59 +0100)]
docs: Even more gratutious nitpicks
Signed-off-by: Daniel Stone <daniels@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11085>
Alyssa Rosenzweig [Tue, 25 May 2021 00:38:41 +0000 (20:38 -0400)]
agx: Zero r0l before first use of control flow
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11023>
Alyssa Rosenzweig [Mon, 24 May 2021 02:50:48 +0000 (22:50 -0400)]
agx: Add break/continue support
Following Dougall's notes closely.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11023>
Alyssa Rosenzweig [Wed, 26 May 2021 23:51:46 +0000 (19:51 -0400)]
agx: Implement loops in the simplest way
Again, optimizations are possible, but for now go for conformance.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11023>
Alyssa Rosenzweig [Mon, 24 May 2021 02:51:21 +0000 (22:51 -0400)]
agx: Optimize out empty else blocks
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11023>
Alyssa Rosenzweig [Sun, 23 May 2021 21:26:50 +0000 (17:26 -0400)]
agx: Implement emit_if the simplest way
Lots of optimizations are possible from here.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11023>
Alyssa Rosenzweig [Tue, 25 May 2021 00:15:03 +0000 (20:15 -0400)]
agx: Fix up branch offsets at pack time
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11023>
Alyssa Rosenzweig [Wed, 26 May 2021 23:51:36 +0000 (19:51 -0400)]
agx: Model jump instructions
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11023>
Alyssa Rosenzweig [Sun, 23 May 2021 21:18:20 +0000 (17:18 -0400)]
agx: Pack control flow instructions
Nearly ALU.
Signed-off-by: Alyssa Rosenzweig <alyssa@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11023>
Alyssa Rosenzweig [Tue, 25 May 2021 00:52:16 +0000 (20:52 -0400)]
agx: Add push_exec alias
Use the same canonical form as Metal, so the disassembler can alias it
back.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/11023>
projects / platform / upstream / mesa.git / log