Iago Toral Quiroga [Tue, 29 Oct 2019 09:12:28 +0000 (10:12 +0100)]
v3d: handle writes to gl_Layer from geometry shaders
When geometry shaders write a value to gl_Layer that doesn't correspond to
an existing layer in the target framebuffer the rendering behavior is
undefined according to the spec, however, there are CTS tests that trigger
this scenario on purpose, probably to ensure that nothing terrible happens.
For V3D, this situation is problematic because the binner uses the layer
index to select the offset to write into the tile state data, and we only
allocate tile state for MAX2(num_layers, 1), so we want to make sure we
don't produce values that would lead to out of bounds writes. The simulator
has an assert to catch this, although we haven't observed issues in actual
hardware it is probably best to play safe.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Thu, 31 Oct 2019 09:46:58 +0000 (10:46 +0100)]
v3d: move layer rendering to a separate helper
This helps with reducing nesting level after adding the loop
to handle layered rendering.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Tue, 29 Oct 2019 09:27:23 +0000 (10:27 +0100)]
v3d: support rendering to multi-layered framebuffers
When doing layered rendering the binning stage will prepare per-tile
lists for each layer in the framebuffer, so we need to make sure
we allocate enough space for them .
We also need to emit the NUMBER_OF_LAYERS packet. This is required
even when the number of layers is only 1, otherwise the simulator
detects buffer overflows in the tile_state BO during some CTS test
cases involving layered FBOs.
When rendering, we need to emit commands for each layer of the
framebuffer separately and make sure we address the correct layers for
each one.
v2: fixed typo in comment (Alejandro)
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Wed, 30 Oct 2019 10:33:44 +0000 (11:33 +0100)]
v3d: do not limit new CL space allocations with branch to 4096 bytes
For layered rendering we need to emit per layer rendering commands
lists so we we can end up requiring a fairly large buffer for this
if the number of layers is large enough.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Tue, 29 Oct 2019 08:32:05 +0000 (09:32 +0100)]
v3d: remove obsolete assertion
OES_geometry_shader introduced the concept of layered framebuffers.
Removing this assertion gets a bunch of CTS tests to pass. We will
also need layered images to implement layered rendering with geometry
shaders.
v2: fix typo in commit message (Alejandro)
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Tue, 15 Oct 2019 06:32:47 +0000 (08:32 +0200)]
v3d: support transform feedback with geometry shaders
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Fri, 11 Oct 2019 10:01:32 +0000 (12:01 +0200)]
v3d: save geometry shader state for blitting
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Fri, 11 Oct 2019 09:40:38 +0000 (11:40 +0200)]
v3d: predicate geometry shader outputs inside non-uniform control flow
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Wed, 9 Oct 2019 09:13:00 +0000 (11:13 +0200)]
v3d: don't try to render if shaders failed to compile
This is the same we do in the compute path to avoid crashes
at draw time.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Wed, 9 Oct 2019 08:26:16 +0000 (10:26 +0200)]
v3d: add support for adjacency primitives
v2: remove obsolete comment (Alejandro)
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Tue, 8 Oct 2019 11:48:00 +0000 (13:48 +0200)]
v3d: we always have at least one output segment
If we program an output size of 0 the simulator asserts. This was
not a problem until now because our VS would always have to
emit fixed function outputs, however, now that it can be paired
with a GS we can end up with a VS shader that no longer emits
any outputs.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Thu, 7 Nov 2019 14:40:45 +0000 (15:40 +0100)]
v3d: compute appropriate VPM memory configuration for geometry shader workloads
Geometry shaders can output many vertices and thus have higher VPM memory
pressure as a result. It is possible that too wide geometry shader dispatches
exceed the maximum available VPM output allocated, in which case we need
to reduce the dispatch width until we can fit the VPM memory requirements.
Supported dispatch widths for geometry shaders are 16, 8, 4, 1.
There is a limit in the number of VPM output sectors that can be used by a
geometry shader that we can meet by lowering the dispatch width at compile
time, however, at draw time we need to revisit this number and, together with
other elements that can contribute to total VPM memory requirements, decide
on a configuration that can fit the program into the available VPM memory.
Ideally, we also want to aim for not using more than half of the available
memory so we that we can run a pair of bin and render programs in parallel.
v2: fixed language in comment and typo in commit log. (Alejandro)
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Mon, 28 Oct 2019 13:23:51 +0000 (14:23 +0100)]
v3d: add 1-way SIMD packing definition
According to the documentation, the 1-way dispatch width is only supported
with geometry shaders.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Thu, 3 Oct 2019 15:08:47 +0000 (17:08 +0200)]
v3d: implement geometry shader instancing
v2:
- Remove unused field uses_iid from v3d_gs_prog_data (Alejandro)
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Tue, 24 Sep 2019 08:34:55 +0000 (10:34 +0200)]
v3d: emit geometry shader state commands
This is good enough to get basic GS workloads working, later patches will
improve this by adding instancing support, proper SIMD configuration, etc.
Notice that most of the TESSELLATION_GEOMETRY_SHADER_PARAMS fields are only
relevant when tessellation shaders are present. We do not support tessellation
yet, but we still need to fill in these tessellation state with default values
since our packing functions require some of these to have non-zero values.
v2:
- Add a comment in the code explaining why we fill in
tessellation fields (Alejandro)
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Thu, 19 Sep 2019 11:36:30 +0000 (13:36 +0200)]
v3d: fix packet descriptions for geometry and tessellation shaders
Every code address starts at bit 3 (addresses must be 64-bit aligned),
with the first 3 bits used to specify threading and NaN propagation
parameters for the shader program.
We generally skip "reserved" bits, however, doing this when the
reserved field is the last in a struct and it is large enough can
make us compute incorrect (smaller) struct sizes which can
lead to corrupt CLs. In particular, the "Tess/Geom Common Params"
struct has a reserved field at the end that is 8-bit, so if we
don't include this we compute a packet size that is 1 byte smaller
than it shold, making the next packet we emit start 1 byte
earlier and therefore leading to incorrect CL data from that point
forward.
The name of one of the fields was not correct.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Mon, 28 Oct 2019 12:24:44 +0000 (13:24 +0100)]
v3d: add initial compiler plumbing for geometry shaders
Most of the relevant work happens in the v3d_nir_lower_io. Since
geometry shaders can write any number of output vertices, this pass
injects a few variables into the shader code to keep track of things
like the number of vertices emitted or the offsets into the VPM
of the current vertex output, etc. This is also where we handle
EmitVertex() and EmitPrimitive() intrinsics.
The geometry shader VPM output layout has a specific structure
with a 32-bit general header, then another 32-bit header slot for
each output vertex, and finally the actual vertex data.
When vertex shaders are paired with geometry shaders we also need
to consider the following:
- Only geometry shaders emit fixed function outputs.
- The coordinate shader used for the vertex stage during binning must
not drop varyings other than those used by transform feedback, since
these may be read by the binning GS.
v2:
- Use MAX3 instead of a chain of MAX2 (Alejandro).
- Make all loop variables unsigned in ntq_setup_gs_inputs (Alejandro)
- Update comment in IO owering so it includes the GS stage (Alejandro)
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Mon, 28 Oct 2019 09:29:15 +0000 (10:29 +0100)]
v3d: remove unused variable
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Tue, 24 Sep 2019 08:33:30 +0000 (10:33 +0200)]
v3d: enable debug options for geometry shader dumps
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Mon, 23 Sep 2019 09:14:52 +0000 (11:14 +0200)]
v3d: add debug assert
While lowering vpm outputs we look for the NIR variables matching
particular store output instructions and we expect to find a match,
so assert on that.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Iago Toral Quiroga [Tue, 24 Sep 2019 08:43:41 +0000 (10:43 +0200)]
v3d: add missing plumbing for VPM load instructions
We will need to use LDVPMG_IN specifically to read VPM inputs
in geometry shaders.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Eric Anholt [Wed, 27 Nov 2019 18:43:54 +0000 (10:43 -0800)]
turnip: Lower usub_borrow.
Fixes dEQP-VK.glsl.builtin.function.integer.usubborrow.uvec2_mediump_fragment.
Reviewed-by: Jonathan Marek <jonathan@marek.ca>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2986>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2986>
Caio Marcelo de Oliveira Filho [Thu, 12 Dec 2019 21:25:33 +0000 (13:25 -0800)]
intel/fs: Lower 64-bit MOVs after lower_load_payload()
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3070>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3070>
Bas Nieuwenhuizen [Thu, 12 Dec 2019 11:10:58 +0000 (12:10 +0100)]
amd/common: Always use addrlib for HTILE tc-compat.
Even without depth+stencil addrlib can (correctly!) decide to
disable tc compatible HTILE.
One example is 8x sampling with 32-bit depth on Stoney. The row size
on Stoney is 1024, while the tile size is 2048, which results in
tile splits which are not supported with tc-compat.
On Stoney, this fixes
dEQP-VK.glsl.builtin_var.fragdepth.*_list_d32_sfloat_multisample_8
CC: <mesa-stable@lists.freedesktop.org>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3054>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3054>
Bas Nieuwenhuizen [Wed, 11 Dec 2019 15:04:58 +0000 (16:04 +0100)]
amd/common: Fix tcCompatible degradation on Stoney.
addrlib sometimes returns smaller sizes for tcCompat as it does
not seem to take into account the depth+stencil matching config
gymnastics with tcCompat.
This fixes
dEQP-VK.pipeline.render_to_image.core.2d_array.huge.height.r8g8b8a8_unorm_d32_sfloat_s8_uint
CC: <mesa-stable@lists.freedesktop.org>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3054>
Denis Pauk [Sat, 14 Dec 2019 07:58:25 +0000 (09:58 +0200)]
docs/features: mark GL_ARB_texture_compression_bptc as done for llvmpipe, softpipe, swr
Signed-off-by: Denis Pauk <pauk.denis@gmail.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
CC: Marek Olšák <maraeo@gmail.com>
CC: Rhys Perry <pendingchaos02@gmail.com>
CC: Bruce Cherniak <bruce.cherniak@intel.com>
CC: Matt Turner <mattst88@gmail.com>
Denis Pauk [Sat, 14 Dec 2019 07:54:48 +0000 (09:54 +0200)]
gallium/swr: Enable support bptc format.
Reuse Code from:
f69bc797e1 gallium/auxiliary: Add helper support for bptc format compress/decompress
Signed-off-by: Denis Pauk <pauk.denis@gmail.com>
Reviewed-by: Bruce Cherniak <bruce.cherniak@intel.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
CC: Marek Olšák <maraeo@gmail.com>
CC: Tim Rowley <timothy.o.rowley@intel.com>
Rob Clark [Sat, 14 Dec 2019 17:09:08 +0000 (09:09 -0800)]
freedreno/a6xx: fix OUT_REG() vs growable cmdstream
BEGIN_RING() could decide we can't fit the next packet in the current
cmdstream segment, and grow a new segment. So we need to grab ring->cur
*after* BEGIN_RING(), otherwise we are writing cmdstream past the end of
the previous segment.
Fixes: bdd98b892f3 ("freedreno: New struct packing macros")
Signed-off-by: Rob Clark <robdclark@chromium.org>
Erico Nunes [Sat, 9 Nov 2019 12:50:52 +0000 (13:50 +0100)]
lima: split draw calls on 64k vertices
The Mali400 only supports draws with up to 64k vertices per command.
To handle this, break the draw_vbo call into multiple commands.
Indexed drawing is left to a separate code path.
This implementation was ported from vc4_draw_vbo.
Signed-off-by: Erico Nunes <nunes.erico@gmail.com>
Reviewed-by: Vasily Khoruzhick <anarsoul@gmail.com>
Reviewed-by: Andreas Baierl <ichgeh@imkreisrum.de>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2445>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2445>
Erico Nunes [Tue, 12 Nov 2019 13:56:33 +0000 (14:56 +0100)]
vc4: move the draw splitting routine to shared code
This can also be useful for other hardware which has similar limitations
on vertex count per single draw.
The Mali400 has a similar limitation and can reuse this.
Signed-off-by: Erico Nunes <nunes.erico@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2445>
Erico Nunes [Sat, 9 Nov 2019 12:50:07 +0000 (13:50 +0100)]
lima: refactor indexed draw indices upload
As of this commit this is just a refactor in preparation to enable
support for more than 64k vertices.
To support splitting the draw_vbo call, indices shouldn't be re-uploaded
every time.
Signed-off-by: Erico Nunes <nunes.erico@gmail.com>
Reviewed-by: Vasily Khoruzhick <anarsoul@gmail.com>
Reviewed-by: Andreas Baierl <ichgeh@imkreisrum.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2445>
Erico Nunes [Wed, 23 Oct 2019 22:27:22 +0000 (00:27 +0200)]
lima: allocate separate bo to store varyings
The current strategy using the suballocator with fixed size doesn't
scale and causes some programs with large number of vertices (like some
glmark2 scenes) to crash.
Change it to dynamically allocate a separate bo to accomodate for
arbitrary number of vertices.
This also fixes the buffer read/write flags for gp.
Signed-off-by: Erico Nunes <nunes.erico@gmail.com>
Reviewed-by: Vasily Khoruzhick <anarsoul@gmail.com>
Reviewed-by: Andreas Baierl <ichgeh@imkreisrum.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2445>
Erico Nunes [Sat, 7 Dec 2019 03:38:03 +0000 (04:38 +0100)]
gallium/util: add alignment parameter to util_upload_index_buffer
At least on Mali Utgard, index buffers need to be aligned on 0x40.
To avoid duplicating this, add an alignment parameter.
Keep the previous default for the other existing users.
Signed-off-by: Erico Nunes <nunes.erico@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2445>
Kenneth Graunke [Fri, 13 Dec 2019 05:26:12 +0000 (21:26 -0800)]
drirc: Final Fantasy VIII: Remastered needs allow_higher_compat_version
This gets it running on i965 with Mesa master. (The game won't start
without GL 3.3 compatibility, but uses 1.20 with GL_EXT_gpu_shader4
for shaders.)
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3076>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3076>
Timothy Arceri [Fri, 13 Dec 2019 10:58:28 +0000 (21:58 +1100)]
st/glsl_to_nir: fix SSO validation regression
Fixes: b77907edb554 ("st/glsl_to_nir: use nir based program resource list builder")
Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/2216
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Alyssa Rosenzweig [Fri, 13 Dec 2019 22:13:14 +0000 (17:13 -0500)]
ci: Remove T760/T860 from CI temporarily
I feel really bad about this but this one test is flaking. I don't want
to do a mass revert (and bisection is extremely difficult with
nondeterministic/Heisenbugs), but it's Friday night and master needs to
pass. This commit should be reverted asap (once the flake is solved)
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Rafael Antognolli [Tue, 19 Nov 2019 23:00:06 +0000 (15:00 -0800)]
iris: Implement WA for push constants.
v2: Apply WA to gen11+ instead of gen12+ (Jordan).
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Andreas Baierl [Mon, 9 Dec 2019 11:42:30 +0000 (12:42 +0100)]
lima/parser: Add texture descriptor parser
Signed-off-by: Andreas Baierl <ichgeh@imkreisrum.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2980>
Andreas Baierl [Fri, 6 Dec 2019 08:30:14 +0000 (09:30 +0100)]
lima/parser: Add RSW parsing
Signed-off-by: Andreas Baierl <ichgeh@imkreisrum.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2980>
Andreas Baierl [Thu, 5 Dec 2019 16:39:01 +0000 (17:39 +0100)]
lima/parser: Some fixes and cleanups
Signed-off-by: Andreas Baierl <ichgeh@imkreisrum.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2980>
Rafael Antognolli [Thu, 12 Dec 2019 21:55:11 +0000 (13:55 -0800)]
vulkan/overlay: Update docs.
Add mention to overlay control socket.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Rafael Antognolli [Thu, 12 Dec 2019 14:45:51 +0000 (06:45 -0800)]
vulkan/overlay: Add basic overlay control script.
This can be used to start/stop statistics capturing from the command
line.
v3:
- Install script (Lionel)
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Rafael Antognolli [Fri, 6 Dec 2019 20:18:19 +0000 (12:18 -0800)]
vulkan/overlay: Add a command to start capturing data to a file.
By default, if an output_file is specified, the overlay layer will start
capturing data immediately. After this commit, when a control socket is
used, the capture starts disabled by default, and is only enabled when a
command ":capture=1;" is received.
when the capture is enabled, we might have already accumulated some
stats. To avoid capturing such noise, we discard and reset the fps and
stats, updating the display and capturing only data from that point on.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Rafael Antognolli [Fri, 6 Dec 2019 21:44:19 +0000 (13:44 -0800)]
vulkan/overlay: Add support for a control socket.
Add support for socket from which the overlay layer can receive
commands. This control socket can be useful to allow setting options
once the application is already running. For instance, triggering the
capture of fps data at a certain point.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Rafael Antognolli [Fri, 6 Dec 2019 22:38:07 +0000 (14:38 -0800)]
vulkan/overlay: Add a control socket.
v2: Use a socket instead of named pipe.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Rafael Antognolli [Wed, 11 Dec 2019 23:01:11 +0000 (15:01 -0800)]
util/os_socket: Add socket related functions.
v3:
- Add os_socket.c/h into Makefile.sources (Lionel)
- Add empty non-linux implementation to public functions.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Eric Engestrom [Fri, 13 Dec 2019 17:16:17 +0000 (17:16 +0000)]
anv: drop unused #include
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Acked-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Eric Engestrom [Fri, 13 Dec 2019 17:12:48 +0000 (17:12 +0000)]
util/simple_mtx: don't set the canary when it can't be checked
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Eric Engestrom [Fri, 13 Dec 2019 17:01:39 +0000 (17:01 +0000)]
intel/compiler: replace `0` pointer with `NULL`
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Eric Engestrom [Fri, 13 Dec 2019 17:01:17 +0000 (17:01 +0000)]
intel/compiler: add ASSERTED annotation to avoid "unused variable" warning
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Caio Marcelo de Oliveira Filho <caio.oliveira@intel.com>
Kenneth Graunke [Fri, 13 Dec 2019 19:03:03 +0000 (11:03 -0800)]
iris: Alphabetize source files after iris_perf.c was added
Rob Clark [Thu, 12 Dec 2019 23:30:49 +0000 (15:30 -0800)]
freedreno/ir3: add iterator macros
So many open coded list iterators were getting annoying.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Rob Clark [Fri, 22 Nov 2019 19:13:19 +0000 (11:13 -0800)]
freedreno/ir3: add scheduler traces
Add some infrastructure to trace scheduler decisions. The next patch
will add some more traces, just splitting this out to reduce clutter.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Rob Clark [Wed, 11 Dec 2019 23:52:32 +0000 (15:52 -0800)]
freedreno/ir3: add last-baryf shaderdb stat
Sometimes sched changes that are a win in terms of instruction count
and/or register pressure, are worse in real life, due to keeping varying
storage locked for too long. Add a shader-db stat to give this more
visibility.
Signed-off-by: Rob Clark <robdclark@chromium.org>
Alejandro Piñeiro [Fri, 13 Dec 2019 13:57:37 +0000 (14:57 +0100)]
nir/opt_peephole_select: remove unused variables
To avoid "unused variable" warnings.
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Alyssa Rosenzweig [Mon, 9 Dec 2019 21:02:17 +0000 (16:02 -0500)]
panfrost: Report GPU name in es2_info
We can prettify the ID.
Closes #2093
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Mon, 9 Dec 2019 21:02:03 +0000 (16:02 -0500)]
panfrost: Add panfrost_model_name helper
This gives us a string representation of a GPU ID.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Mon, 9 Dec 2019 20:54:09 +0000 (15:54 -0500)]
panfrost: Move property queries to _encoder
We'll want these in non-Gallium devices.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Mon, 9 Dec 2019 20:41:52 +0000 (15:41 -0500)]
panfrost: Move nir_undef_to_zero to Midgard compiler
Nothing Gallium about it.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Fri, 13 Dec 2019 15:22:06 +0000 (10:22 -0500)]
pandecode: Add cast
Fixes minor coverity warning about the format specifier.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Mon, 9 Dec 2019 16:02:15 +0000 (11:02 -0500)]
panfrost: Pass size to panfrost_batch_get_scratchpad
We'll compute the size with the new scratchpad helpers.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Mon, 9 Dec 2019 16:18:47 +0000 (11:18 -0500)]
panfrost: Calculate maximum stack_size per batch
We'll need this so we can allocate a stack for the batch large enough
for all the jobs within it.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Fri, 13 Dec 2019 15:13:24 +0000 (10:13 -0500)]
pan/midgard: Handle misc. cppcheck warnings
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Fri, 13 Dec 2019 15:09:16 +0000 (10:09 -0500)]
pan/midgard: Remove unused ld/st packing hepers
Identified by cppcheck.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Fri, 13 Dec 2019 15:07:44 +0000 (10:07 -0500)]
panfrost: Handle minor cppcheck issues
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Mon, 9 Dec 2019 16:00:42 +0000 (11:00 -0500)]
panfrost: Emit SFBD/MFBD after a batch, instead of before
The size of the scratchpad (as well as some tiler details) depend on the
contents of the batch, so we need to wait to defer filling out the FBD
until after all draws are queued.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Mon, 9 Dec 2019 16:18:21 +0000 (11:18 -0500)]
panfrost: Route stack_size from compiler
We'll need it in pan_context.c
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Jonathan Marek [Sun, 8 Dec 2019 23:16:34 +0000 (18:16 -0500)]
etnaviv: add missing vs_needs_z_div handling to NIR backend
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Jonathan Marek [Sun, 8 Dec 2019 16:52:32 +0000 (11:52 -0500)]
etnaviv: add missing formats
Add missing texture/render formats supported by hardware.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Jonathan Marek [Sun, 8 Dec 2019 16:20:46 +0000 (11:20 -0500)]
etnaviv: remove swizzle from format table
The only format that needs swizzle is R8 emulated with L8, so we can get
rid of the SWIZ(X, Y, Z, W) everywhere.
Note: R8G8 also had a swizzle, but it wasn't necessary.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Jonathan Marek [Sun, 8 Dec 2019 16:54:31 +0000 (11:54 -0500)]
etnaviv: disable integer vertex formats on pre-HALTI2 hardware
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Jonathan Marek [Sun, 20 Oct 2019 06:10:43 +0000 (02:10 -0400)]
etnaviv: update INT_FILTER choice for GLES3 formats
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Jonathan Marek [Mon, 12 Aug 2019 18:19:59 +0000 (14:19 -0400)]
etnaviv: set output mode and saturate bits
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Jonathan Marek [Sun, 20 Oct 2019 06:21:43 +0000 (02:21 -0400)]
etnaviv: sRGB render target support
Note: no srgb render target support before HALTI3
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Jonathan Marek [Sat, 10 Aug 2019 18:38:32 +0000 (14:38 -0400)]
etnaviv: remove sRGB formats from format table
This supports all sRGB formats, without having them in the format table.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Tomasz Pyra [Thu, 12 Dec 2019 14:38:43 +0000 (15:38 +0100)]
gallium/swr: Fix arb_transform_feedback2
Added support for pause/resume transform feedback.
Fixed DrawTransformFeedback.
Reviewed-by: Jan Zielinski <jan.zielinski@intel.com>
Reviewed-by: Krzysztof Raszkowski <krzysztof.raszkowski@intel.com>
Samuel Pitoiset [Fri, 29 Nov 2019 14:12:30 +0000 (15:12 +0100)]
radv: handle unaligned vertex fetches on GFX6/GFX10
The Vulkan spec doesn't have any words for vertex attributes alignment.
Fixes a test failure on GFX6 and a GPU hang on GFX10 with:
dEQP-VK.spirv_assembly.instruction.spirv1p4.entrypoint.tess_con_pc_entry_point
vkpipeline-db results on GFX10:
Totals from affected shaders:
SGPRS: 463772 -> 472972 (1.98 %)
VGPRS: 343208 -> 343752 (0.16 %)
Spilled SGPRs: 323 -> 336 (4.02 %)
Spilled VGPRs: 0 -> 0 (0.00 %)
Code Size:
13806200 ->
14164472 (2.60 %) bytes
Max Waves: 84021 -> 83755 (-0.32 %)
Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/2161
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Lionel Landwerlin [Mon, 20 May 2019 06:56:18 +0000 (07:56 +0100)]
i965/iris: perf-queries: don't invalidate/flush 3d pipeline
Our current implementation of performance queries is fairly harsh
because it completely flushes and invalidates the 3d pipeline caches
at the beginning and end of each query. An argument can be made that
this is how performance should be measured but it probably doesn't
reflect what the application is actually doing and the actual cost of
draw calls.
A more appropriate approach is to just stall the pipeline at
scoreboard, so that we measure the effect of a draw call without
having the pipeline in a completely pristine state for every draw
call.
v2: Use end of pipe PIPE_CONTROL instruction for Iris (Ken)
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Lionel Landwerlin [Thu, 12 Dec 2019 10:14:09 +0000 (12:14 +0200)]
intel/perf: drop batchbuffer flushing at query begin
This was initially intended to fix issues with the query timings going
occassionally high.
It turns out there was a bug in the attribution of OA reports to our
context when parsing the OA data. This led to reports flagged with
other context IDs to be included in our queries results.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Iago Toral Quiroga [Thu, 12 Dec 2019 10:19:23 +0000 (11:19 +0100)]
v3d: actually root the first BO in a command list in the job
We were passing cl->bo, which is NULL, so v3d_job_add_bo was a no-op.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Christian Gmeiner [Tue, 10 Dec 2019 16:15:35 +0000 (17:15 +0100)]
etnaviv: drop compiled_rs_state forward declaration
Signed-off-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Christian Gmeiner [Tue, 10 Dec 2019 08:54:18 +0000 (09:54 +0100)]
etnaviv: remove not used etna_bits_ones(..)
Signed-off-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Vinson Lee [Thu, 12 Dec 2019 01:34:30 +0000 (17:34 -0800)]
swr: Fix build with llvm-10.0.
Fix build error after llvm-10.0 commit ("
1b2842bf902a [Alignment][NFC]
CreateMemSet use MaybeAlign").
../src/gallium/drivers/swr/swr_shader.cpp: In member function ‘void (* BuilderSWR::CompileGS(swr_context*, swr_jit_gs_key&))(HANDLE, HANDLE, SWR_GS_CONTEXT*)’:
../src/gallium/drivers/swr/swr_shader.cpp:738:65: error: no matching function for call to ‘BuilderSWR::MEMSET(llvm::Value*&, llvm::Constant*, int, long unsigned int)’
MEMSET(pStream, C((char)0), VERTEX_COUNT_SIZE + CONTROL_HEADER_SIZE, sizeof(float) * KNOB_SIMD_WIDTH);
^
In file included from ../src/gallium/drivers/swr/rasterizer/jitter/builder.h:163:0,
from ../src/gallium/drivers/swr/swr_shader.cpp:43:
src/gallium/drivers/swr/rasterizer/jitter/gen_builder.hpp:51:11: note: candidate: llvm::CallInst* SwrJit::Builder::MEMSET(llvm::Value*, llvm::Value*, uint64_t, llvm::MaybeAlign, bool, llvm::MDNode*, llvm::MDNode*, llvm::MDNode*)
CallInst* MEMSET(Value *Ptr, Value *Val, uint64_t Size, MaybeAlign Align, bool isVolatile = false, MDNode *TBAATag = nullptr, MDNode *ScopeTag = nullptr, MDNode *NoAliasTag = nullptr)
^
src/gallium/drivers/swr/rasterizer/jitter/gen_builder.hpp:51:11: note: no known conversion for argument 4 from ‘long unsigned int’ to ‘llvm::MaybeAlign’
In file included from ../src/gallium/drivers/swr/rasterizer/jitter/builder.h:163:0,
from ../src/gallium/drivers/swr/swr_shader.cpp:43:
src/gallium/drivers/swr/rasterizer/jitter/gen_builder.hpp:56:11: note: candidate: llvm::CallInst* SwrJit::Builder::MEMSET(llvm::Value*, llvm::Value*, llvm::Value*, llvm::MaybeAlign, bool, llvm::MDNode*, llvm::MDNode*, llvm::MDNode*)
CallInst* MEMSET(Value *Ptr, Value *Val, Value *Size, MaybeAlign Align, bool isVolatile = false, MDNode *TBAATag = nullptr, MDNode *ScopeTag = nullptr, MDNode *NoAliasTag = nullptr)
^
src/gallium/drivers/swr/rasterizer/jitter/gen_builder.hpp:56:11: note: no known conversion for argument 4 from ‘long unsigned int’ to ‘llvm::MaybeAlign’
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Jan Zielinski <jan.zielinski@intel.com>
Jonathan Marek [Thu, 12 Dec 2019 22:05:22 +0000 (17:05 -0500)]
turnip: implement subpass input attachments
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Eric Anholt <eric@anholt.net>
Jonathan Marek [Thu, 12 Dec 2019 19:02:49 +0000 (14:02 -0500)]
turnip: CmdClearAttachments fixes
Partial depth/stencil clear and skipping unused attachments.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Eric Anholt <eric@anholt.net>
Jonathan Marek [Fri, 6 Dec 2019 01:53:34 +0000 (20:53 -0500)]
turnip: subpass rework
A renderpass is a tile load/store cycle.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Eric Anholt <eric@anholt.net>
Jonathan Marek [Thu, 12 Dec 2019 22:13:55 +0000 (17:13 -0500)]
turnip: add dirty bit for push constants
Fixes push constants not updating in some cases.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Eric Anholt <eric@anholt.net>
Jonathan Marek [Thu, 12 Dec 2019 22:03:26 +0000 (17:03 -0500)]
turnip: no 8x msaa on 128bpp formats
We don't have an entry for cpp 128 in the tile_alignment table, but I don't
think the HW supports this at all (blob driver just doesn't have 8x msaa).
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Eric Anholt <eric@anholt.net>
Jonathan Marek [Thu, 12 Dec 2019 22:01:52 +0000 (17:01 -0500)]
turnip: fix VK_IMAGE_ASPECT_STENCIL_BIT image view
Use a special format which allows sampling the stencil and set the correct
swizzle.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Eric Anholt <eric@anholt.net>
Jonathan Marek [Thu, 12 Dec 2019 22:00:13 +0000 (17:00 -0500)]
turnip: set FRAG_WRITES_SAMPMASK bit
GPU hangs if SAMPMASK_REGID is used without this bit.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Eric Anholt <eric@anholt.net>
Jonathan Marek [Thu, 12 Dec 2019 21:58:56 +0000 (16:58 -0500)]
turnip: set load_layer_id to zero
We don't have layered rendering and ir3 doesn't support this intrinsic, so
just set it to zero for now.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Eric Anholt <eric@anholt.net>
Jonathan Marek [Thu, 12 Dec 2019 21:58:03 +0000 (16:58 -0500)]
turnip: update tile_align_w/tile_align_h
It looks like the actual tile alignment requirement is less than 32x32, but
in some cases input attachment texture needs 64 alignment.
Reduced the h alignment to 16 to compensate and it seems to work fine.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Eric Anholt <eric@anholt.net>
Jonathan Marek [Thu, 12 Dec 2019 21:55:15 +0000 (16:55 -0500)]
turnip: fix tile layout logic
Use DIV_ROUND_UP and stop trying to increase the tile_count width/height
once tile_align_w/tile_align_h are reached.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Eric Anholt <eric@anholt.net>
Jonathan Marek [Thu, 12 Dec 2019 21:51:39 +0000 (16:51 -0500)]
turnip: fix hw binning render area
Fix a mistake in the y2 coordinate.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Eric Anholt <eric@anholt.net>
Jonathan Marek [Thu, 12 Dec 2019 18:59:45 +0000 (13:59 -0500)]
freedreno/registers: add a6xx texture format for stencil sampler
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Eric Anholt <eric@anholt.net>
Jonathan Marek [Thu, 12 Dec 2019 18:58:28 +0000 (13:58 -0500)]
freedreno/ir3: add GLSL_SAMPLER_DIM_SUBPASS to tex_info
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Eric Anholt <eric@anholt.net>
Jonathan Marek [Fri, 6 Dec 2019 01:58:58 +0000 (20:58 -0500)]
turnip: fix incorrectly failing assert
pColorBlendState is allowed to be NULL if subpass has >0 color attachments
but they are all unused.
Signed-off-by: Jonathan Marek <jonathan@marek.ca>
Reviewed-by: Eric Anholt <eric@anholt.net>
Alyssa Rosenzweig [Mon, 9 Dec 2019 14:06:51 +0000 (09:06 -0500)]
panfrost: Query core count and thread tls alloc
This is supported only on newer kernels.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Alyssa Rosenzweig [Mon, 9 Dec 2019 14:00:49 +0000 (09:00 -0500)]
panfrost: Factor out panfrost_query_raw
We would like to query properties other than product ID.
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Timothy Arceri [Fri, 6 Dec 2019 10:57:16 +0000 (21:57 +1100)]
st/glsl_to_nir: use nir based program resource list builder
Here we use the NIR based builder to add everything to the resource
list execpt for SSO packed varyings. Since the details of those
varyings get lost during packing we leave the special handing to
the GLSL IR pass for now. In order to do this we add some bools
to the build resource list functions.
Using the NIR based resource list builder gets us a step closer to
using a native NIR based linker. It should also be faster than the
GLSL IR builder, one because the NIR optimisations should mean we
add less entries due to better optimisations, and two because nir
gives us better lists to work with and we don't need to walk the
entire IR to find the resources.
Ack-by: Alejandro Piñeiro <apinheiro@igalia.com>