mesa.git
23 months agogitlab-ci: Add ppc64el and s390x cross-build jobs
Michel Dänzer [Thu, 30 Jan 2020 17:21:15 +0000 (18:21 +0100)]
gitlab-ci: Add ppc64el and s390x cross-build jobs

Using LLVM 8 for ppc64el and 7 for s390x (which hits some coroutine
related issues with LLVM 8).

There are some test failures we need to ignore for now. Also, the
timeout needs to be bumped from the default 30s for some tests, because
they can take longer under emulation.

Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3643>

23 months agogitlab-ci: Merge ccache and libxml2-utils into main apt-get install
Michel Dänzer [Fri, 31 Jan 2020 15:48:36 +0000 (16:48 +0100)]
gitlab-ci: Merge ccache and libxml2-utils into main apt-get install

The motivation for this is that we want to make use of the meson cross
files in this script, which have the ccache compiler paths.

We need to remove the ccache directory at the end, it would just waste
space in the image for no benefit.

Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3643>

23 months agogitlab-ci: Pass -j4 to make
Michel Dänzer [Fri, 31 Jan 2020 16:40:54 +0000 (17:40 +0100)]
gitlab-ci: Pass -j4 to make

Might speed up x86_build docker image build a little.

Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3643>

23 months agogitlab-ci: Update to latest ci-templates HEAD
Michel Dänzer [Fri, 31 Jan 2020 15:57:31 +0000 (16:57 +0100)]
gitlab-ci: Update to latest ci-templates HEAD

Among other things, this increases robustness when copying a docker
image from the main Mesa project to a forked project, avoiding spurious
image rebuilds from scratch.

Also drop the comment about .gitlab-ci/lava-gitlab-ci.yml, it doesn't
include the templates anymore.

Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3643>

23 months agoradeonsi/ngg: add VGT_FLUSH when enabling fast launch
Pierre-Eric Pelloux-Prayer [Thu, 30 Jan 2020 11:16:46 +0000 (12:16 +0100)]
radeonsi/ngg: add VGT_FLUSH when enabling fast launch

Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/2418
Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/2426
Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/2434
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3675>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3675>

23 months agoutil/disk_cache: check for write() failure in the zstd path
Eric Engestrom [Sun, 2 Feb 2020 17:15:09 +0000 (17:15 +0000)]
util/disk_cache: check for write() failure in the zstd path

CoverityID: 1458074
Fixes: a8d941091f72923561a6 ("util: Use ZSTD for shader cache if possible")
Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3672>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3672>

23 months agodri: delete gen-symbol-redefs.py
Eric Engestrom [Sun, 2 Feb 2020 15:11:16 +0000 (15:11 +0000)]
dri: delete gen-symbol-redefs.py

Introduced in ba10d79cca8d93c9f366 but it looks like it was never wired
into anything.

Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3669>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3669>

23 months agoanv: implement gen12 post sync pipe control workaround
Lionel Landwerlin [Wed, 15 Jan 2020 12:09:26 +0000 (14:09 +0200)]
anv: implement gen12 post sync pipe control workaround

Same as Skylake.

v2: Restrict to A0

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3405>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3405>

23 months agoanv: implement gen9 post sync pipe control workaround
Lionel Landwerlin [Wed, 15 Jan 2020 12:07:29 +0000 (14:07 +0200)]
anv: implement gen9 post sync pipe control workaround

We've been missing this workaround for a while and since it's required
for Gen12, let's implement it for Gen9 first.

v2: Update comment for Gen9.

v3: Fix clearing of bits... (Lionel)

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Cc: <mesa-stable@lists.freedesktop.org>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3405>

23 months agoiris: implement gen12 post sync pipe control workaround
Lionel Landwerlin [Wed, 15 Jan 2020 12:06:07 +0000 (14:06 +0200)]
iris: implement gen12 post sync pipe control workaround

Like Skylake, Gen12 requires a workaround for PIPE_CONTROLs using a
post-sync operation.

v2: Restrict to A0

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3405>

23 months agofreedreno: allow ctx->batch to be NULL
Rob Clark [Tue, 4 Feb 2020 21:46:56 +0000 (13:46 -0800)]
freedreno: allow ctx->batch to be NULL

This was mostly true already, now that we use `fd_context_batch()` for
first access to batch in draw/clear/grid paths.  So we can drop the old
code in `batch_flush()` that tried to prevent `ctx->batch` from being
NULL.

Fixes a crash with a large number of tabs in chromium.

Cc: "20.0" mesa-stable@lists.freedesktop.org
Signed-off-by: Rob Clark <robdclark@chromium.org>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3700>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3700>

23 months agofreedreno: Allow UBWC on textures with multiple mipmap levels.
Eric Anholt [Thu, 21 Nov 2019 23:09:02 +0000 (15:09 -0800)]
freedreno: Allow UBWC on textures with multiple mipmap levels.

This is a backport of Jonathan Marek's UBWC work on turnip to GL.

Performance highlights from our trace set (320 frames sampled)
traces/glmark2/texture-texture-filter=mipmap.rdc:    +9.1% +/-   2.2%
traces/android/trex.rdc:                             +8.7% +/-   0.4%
traces/glmark2/desktop-effect=shadow:windows=4.rdc:  +4.2% +/-   2.5%

Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3059>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3059>

23 months agofreedreno: Disable UBWC on Z24S8 if not TEXTURE_2D.
Eric Anholt [Thu, 12 Dec 2019 22:30:29 +0000 (14:30 -0800)]
freedreno: Disable UBWC on Z24S8 if not TEXTURE_2D.

Fixes two of our three remaining GLES CTS failures, and avoids more
regressions once we enable UBWC mipmaps.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3059>

23 months agofreedreno: Blit all array levels when uncompressing UBWC.
Eric Anholt [Thu, 12 Dec 2019 22:20:49 +0000 (14:20 -0800)]
freedreno: Blit all array levels when uncompressing UBWC.

Fixes regressions in GLES CTS's format_reintepret once we enable UBWC
with mipmaps.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3059>

23 months agofreedreno: Swap the whole resource layout in shadowing.
Eric Anholt [Thu, 12 Dec 2019 20:49:12 +0000 (12:49 -0800)]
freedreno: Swap the whole resource layout in shadowing.

Let's not have to worry about whether this unusual code path gets
updated whenever we adjust what is in the layout struct.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3059>

23 months agofreedreno/a6xx: Disable the core layer-size setup.
Eric Anholt [Thu, 12 Dec 2019 00:30:43 +0000 (16:30 -0800)]
freedreno/a6xx: Disable the core layer-size setup.

This was getting in the way of UBWC mipmap handling.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3059>

23 months agofreedreno: Rename the UBWC layer size field and store it as bytes.
Eric Anholt [Mon, 25 Nov 2019 19:49:05 +0000 (11:49 -0800)]
freedreno: Rename the UBWC layer size field and store it as bytes.

This makes the field description match its usage in the code, matches tu's
usage of the field, and avoids storing values in surprising units.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3059>

23 months agofreedreno: Include the layer size in layout debug.
Eric Anholt [Thu, 12 Dec 2019 21:54:43 +0000 (13:54 -0800)]
freedreno: Include the layer size in layout debug.

It's been many of my bugs so far.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3059>

23 months agofreedreno: Move the layout debug under FD_MESA_DEBUG=layout.
Eric Anholt [Thu, 12 Dec 2019 21:48:31 +0000 (13:48 -0800)]
freedreno: Move the layout debug under FD_MESA_DEBUG=layout.

I keep wanting to turn this on while debugging layout stuff, and I
suspect krh and robclark could use it too.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3059>

23 months agoradv: Do not set SX DISABLE bits for RB+ with unused surfaces.
Bas Nieuwenhuizen [Thu, 30 Jan 2020 16:58:55 +0000 (17:58 +0100)]
radv: Do not set SX DISABLE bits for RB+ with unused surfaces.

The extra bits in CB_SHADER_MASK break dual source blending in
SkQP on a Stoney device. However:

- As far as I can tell, some other dual source blend tests are passing
  before and after the change.
- A hacked around skqp passes on my Vega desktop and Raven laptop
- Getting Skqp to give any useful info or to run it outside of Android
  on ChromeOS is proving difficult.

I have confirmed 3 strategies that seem to work:
- The old radv behavior of setting CB_SHADER_MASK to 0xF
- AMDVLK: CB_SHADER_MASK = 0xFF, and the 3 RB+ regs
  are 0.
- radeonsi: CB_SHADER_MASK = 0xFF, but does not set DISABLE
  bits in SX_BLEND_OPT_CONTROL for CB 1-7.

Let us use the radeonsi solution as that solution also seems like the correct
thing to do for holes. I have tested on my Raven laptop that setting the high
surfaces to not disabled and downconvert to 32_R does not imply a performance
penalty.

Fixes: e9316fdfd48 "radv: fix setting CB_SHADER_MASK for dual source blending"
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3670>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3670>

23 months agomesa: implement missing display list functions while switching to the template
Marek Olšák [Wed, 29 Jan 2020 01:18:40 +0000 (20:18 -0500)]
mesa: implement missing display list functions while switching to the template

The vbo_init_tmp.h template tells us which functions are unimplemented.

Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3611>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3611>

23 months agovbo: move reusable code from vbo_attrib_tmp.h into vbo_util.h
Marek Olšák [Wed, 29 Jan 2020 00:48:50 +0000 (19:48 -0500)]
vbo: move reusable code from vbo_attrib_tmp.h into vbo_util.h

Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3611>

23 months agovbo: use the template for save GLvertexformat initialization
Marek Olšák [Tue, 28 Jan 2020 22:25:40 +0000 (17:25 -0500)]
vbo: use the template for save GLvertexformat initialization

Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3611>

23 months agovbo: use the template for noop GLvertexformat initialization
Marek Olšák [Tue, 28 Jan 2020 22:25:40 +0000 (17:25 -0500)]
vbo: use the template for noop GLvertexformat initialization

Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3611>

23 months agovbo: move GLvertexformat initialization into a template header file for reuse
Marek Olšák [Tue, 28 Jan 2020 22:24:48 +0000 (17:24 -0500)]
vbo: move GLvertexformat initialization into a template header file for reuse

Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3611>

23 months agofreedreno/perfcntrs: fix fd leak
Eric Engestrom [Sun, 2 Feb 2020 17:04:29 +0000 (17:04 +0000)]
freedreno/perfcntrs: fix fd leak

CoverityID: 11105681458071
Fixes: 5a13507164a26fc796f0 ("freedreno/perfcntrs: add fdperf")
Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3671>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3671>

23 months agoutil: Drop unpacking from int signed to unsigned and vice versa.
Eric Anholt [Mon, 11 Nov 2019 23:52:53 +0000 (15:52 -0800)]
util: Drop unpacking from int signed to unsigned and vice versa.

After all the previous cleanups, it's clear that the callers only ever
ask for SINT->SINT or UINT->UINT.  Cuts 20k of compiled text from
gallium drivers.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2744>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2744>

23 months agogallium: Refactor some single-pixel util_format_read/writes.
Eric Anholt [Fri, 8 Nov 2019 22:56:07 +0000 (14:56 -0800)]
gallium: Refactor some single-pixel util_format_read/writes.

We can use the new row helpers to cut down on the noise.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2744>

23 months agogallium: Add and use a helper for packing uc from a color_union.
Eric Anholt [Fri, 8 Nov 2019 22:49:44 +0000 (14:49 -0800)]
gallium: Add and use a helper for packing uc from a color_union.

The same pattern kept coming up, and we don't need to hit
util_format_write_4* to do it when we have util_format_pack_rgba().

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2744>

23 months agosoftpipe: Refactor pipe_get/put_tile_rgba_* paths.
Eric Anholt [Fri, 8 Nov 2019 21:39:09 +0000 (13:39 -0800)]
softpipe: Refactor pipe_get/put_tile_rgba_* paths.

We always want the same behavior of choosing which unpack to do to
generate our 4x32-bit RGBA values, so just sink that choice down below
the pipe_get/put_tile API.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2744>

23 months agosoftpipe: Drop the raw_to* part of the tile cache interface.
Eric Anholt [Fri, 8 Nov 2019 21:49:20 +0000 (13:49 -0800)]
softpipe: Drop the raw_to* part of the tile cache interface.

Nothing else uses it, so make it static.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2744>

23 months agogallium/util: Remove pipe_get_tile_z/put_tile_z.
Eric Anholt [Fri, 8 Nov 2019 21:20:18 +0000 (13:20 -0800)]
gallium/util: Remove pipe_get_tile_z/put_tile_z.

The previous caller wasn't using it as tiled, just row-at-a-time, and
didn't want the clipping (since copytexsubimage comes in clipped).  If
someone wanted these functions again in the future, they should be
rewritten on u_format_pack/unpack.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2744>

23 months agomesa/st: Use direct util_format_pack/unpack instead of u_tile.
Eric Anholt [Fri, 8 Nov 2019 21:21:18 +0000 (13:21 -0800)]
mesa/st: Use direct util_format_pack/unpack instead of u_tile.

We're doing a row at a time, and don't need u_tile's clipping.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2744>

23 months agoutil: Make helper functions for pack/unpacking pixel rows.
Eric Anholt [Fri, 8 Nov 2019 20:30:02 +0000 (12:30 -0800)]
util: Make helper functions for pack/unpacking pixel rows.

Almost all users of the unpack functions don't have strides to plug in
(and many are only doing one pixel!), and this will help simplify
them.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2744>

23 months agoclover: add trivial clCreateCommandQueueWithProperties implementation
Karol Herbst [Mon, 12 Mar 2018 10:04:38 +0000 (11:04 +0100)]
clover: add trivial clCreateCommandQueueWithProperties implementation

It's not adding 2.0 features, but it's enough to run the 2.0 CTS on top of
clover and probably most CL applications using it.

We just fail if we hit unknown properties and that's probably good enough
until we implement the other bits properly.

Signed-off-by: Karol Herbst <kherbst@redhat.com>
Acked-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Pierre Moreau <dev@pmoreau.org>
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2370>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/2370>

23 months agogallium/osmesa: Try to fix the test for big-endian.
Eric Anholt [Wed, 15 Jan 2020 00:50:41 +0000 (16:50 -0800)]
gallium/osmesa: Try to fix the test for big-endian.

Our packed expected values will be byte-swapped for the (mostly) array
formats we're testing.

Reviewed-by: Danylo Piliaiev <danylo.piliaiev@globallogic.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3216>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3216>

23 months agogallium/osmesa: Fill out other format tests.
Eric Anholt [Wed, 15 Jan 2020 00:17:39 +0000 (16:17 -0800)]
gallium/osmesa: Fill out other format tests.

Move expected values/bpp into the test params, add more formats now that
we've fixed context setup so that they work.

Reviewed-by: Danylo Piliaiev <danylo.piliaiev@globallogic.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3216>

23 months agogallium/osmesa: Fix MakeCurrent of non-8888 contexts.
Eric Anholt [Wed, 15 Jan 2020 00:35:21 +0000 (16:35 -0800)]
gallium/osmesa: Fix MakeCurrent of non-8888 contexts.

OSMesa is weird and you only get the type (byte/ubyte/565/etc.) at
MakeCurrent time, having only a channel order at CreateContext time.  The
code was setting up a visual at CreateContext time, and then at
MakeCurrent it would fail to validate against the visual.

Reviewed-by: Danylo Piliaiev <danylo.piliaiev@globallogic.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3216>

23 months agogallium/osmesa: Fix a typo in the unit test's test names.
Eric Anholt [Wed, 15 Jan 2020 00:20:53 +0000 (16:20 -0800)]
gallium/osmesa: Fix a typo in the unit test's test names.

Reviewed-by: Danylo Piliaiev <danylo.piliaiev@globallogic.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3216>

23 months agoosmesa/tests: Cover OSMESA_RGB GL_UNSIGNED_BYTE case
Danylo Piliaiev [Mon, 30 Dec 2019 13:51:12 +0000 (15:51 +0200)]
osmesa/tests: Cover OSMESA_RGB GL_UNSIGNED_BYTE case

Signed-off-by: Danylo Piliaiev <danylo.piliaiev@globallogic.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3216>

23 months agost/mesa: Handle the rest renderbuffer formats from OSMesa
Danylo Piliaiev [Fri, 27 Dec 2019 13:12:24 +0000 (15:12 +0200)]
st/mesa: Handle the rest renderbuffer formats from OSMesa

Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/2189
Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/989
Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/2036
CC: <mesa-stable@lists.freedesktop.org>
Signed-off-by: Danylo Piliaiev <danylo.piliaiev@globallogic.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3216>

23 months agoutil/os_socket: fix header unavailable on windows
Eric Engestrom [Tue, 4 Feb 2020 16:23:59 +0000 (16:23 +0000)]
util/os_socket: fix header unavailable on windows

Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/2464
Fixes: e62c3cf350a8b169e640 ("util/os_socket: Include unistd.h to fix build error")
Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Reviewed-by: Krzysztof Raszkowski <krzysztof.raszkowski@intel.com>
23 months agoi965: Do not set front_buffer_dirty if there is no front buffer
Danylo Piliaiev [Wed, 29 Jan 2020 12:34:50 +0000 (14:34 +0200)]
i965: Do not set front_buffer_dirty if there is no front buffer

Otherwise there will be a warning:
 "libEGL warning: FIXME: egl/x11 doesn't support front buffer rendering."

Happens with EGL_KHR_surfaceless_context:

 eglMakeCurrent(egl_display, EGL_NO_SURFACE, EGL_NO_SURFACE, egl_context)
 eglMakeCurrent(egl_display, egl_surface, egl_surface, egl_context)
 glFlush() // Here will be a warning

Cc: <mesa-stable@lists.freedesktop.org>
Closes: https://gitlab.freedesktop.org/mesa/mesa/issues/1525
Signed-off-by: Danylo Piliaiev <danylo.piliaiev@globallogic.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3628>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3628>

23 months agogitlab-ci: Switch kernel for LAVA jobs to 5.5
Tomeu Vizoso [Fri, 31 Jan 2020 07:04:57 +0000 (08:04 +0100)]
gitlab-ci: Switch kernel for LAVA jobs to 5.5

All fixes we were carrying in our branch have been merged already.

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3692>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3692>

23 months agopanfrost: Use size0 when calculating the offset to a depth level
Alyssa Rosenzweig [Tue, 4 Feb 2020 09:31:28 +0000 (10:31 +0100)]
panfrost: Use size0 when calculating the offset to a depth level

Previously, we were using cubemap_stride and sometimes overwriting other
BOs such as shaders.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3692>

23 months agopanfrost: Only clamp the LOD to disable mipmapping when needed
Tomeu Vizoso [Tue, 4 Feb 2020 07:57:52 +0000 (08:57 +0100)]
panfrost: Only clamp the LOD to disable mipmapping when needed

Otherwise, we may be incrementing max_lod over the limit, causing a
DATA_INVALID_FAULT.

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3692>

23 months agopanfrost: Fix decoding of tiled 3D textures
Tomeu Vizoso [Tue, 4 Feb 2020 07:29:50 +0000 (08:29 +0100)]
panfrost: Fix decoding of tiled 3D textures

From decoding cmd streams generated by the blob, the pointers in the
payload don't seem to include those that refer to different depth levels
when the texture is in tiled format.

Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3692>

23 months agost/mesa: use uint-result for sampling stencil buffers
Erik Faye-Lund [Mon, 3 Feb 2020 17:22:31 +0000 (18:22 +0100)]
st/mesa: use uint-result for sampling stencil buffers

Otherwise, we end up mismatching the result-type and the sampler-type.

Fixes: 642125edd97 ("st/mesa: use uint-samplers for sampling stencil buffers")
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3680>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3680>

23 months agopan/midgard: Remove unused variable
Alyssa Rosenzweig [Mon, 3 Feb 2020 19:35:32 +0000 (14:35 -0500)]
pan/midgard: Remove unused variable

../src/panfrost/midgard/mir.c: In function ‘mir_bytemask_of_read_components_index’:
../src/panfrost/midgard/mir.c:471:18: warning: unused variable ‘mask’ [-Wunused-variable]
  471 |         uint16_t mask = 0;

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3684>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3684>

23 months agopan/midgard: Check for null consts
Alyssa Rosenzweig [Mon, 3 Feb 2020 19:32:22 +0000 (14:32 -0500)]
pan/midgard: Check for null consts

Valid shaders shouldn't hit this, but Coverity doesn't know that.

CID 1458029:    (FORWARD_NULL)
Passing null pointer "consts" to "print_scalar_field", which dereferences it.

Tomeu: Fix name of variable

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3684>

23 months agopanfrost: Avoid overlapping copy
Alyssa Rosenzweig [Mon, 3 Feb 2020 19:28:18 +0000 (14:28 -0500)]
panfrost: Avoid overlapping copy

CID 1457486:  Memory - corruptions  (OVERLAPPING_COPY)

Assigning "(*attr).extra_flags = (*attr).size = 0U" to "(*attr).stride",
which have overlapping memory +locations.

Coverity.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3684>

23 months agoetnaviv: Destroy rsc->pending_ctx set in etna_resource_destroy()
Marek Vasut [Thu, 30 Jan 2020 00:40:31 +0000 (01:40 +0100)]
etnaviv: Destroy rsc->pending_ctx set in etna_resource_destroy()

Destroy rsc->pending_ctx set in etna_resource_destroy(), otherwise
the memory is allocated, never free'd, and becomes unreachable. This
fixes a memory leak.

Fixes: 9e672e4d20fb ("etnaviv: keep references to pending resources")
Cc: <mesa-stable@lists.freedesktop.org>
Signed-off-by: Marek Vasut <marex@denx.de>
Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3633>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3633>

23 months agoturnip: Be explicit about converting vk compare func to a6xx
Kristian H. Kristensen [Mon, 3 Feb 2020 20:52:47 +0000 (12:52 -0800)]
turnip: Be explicit about converting vk compare func to a6xx

Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3686>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3686>

23 months agonir: Remove always-true assert
Kristian H. Kristensen [Mon, 3 Feb 2020 20:52:19 +0000 (12:52 -0800)]
nir: Remove always-true assert

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3686>

23 months agoglsl: Use 'using' to be explicit about visitor overloads
Kristian H. Kristensen [Mon, 3 Feb 2020 20:43:19 +0000 (12:43 -0800)]
glsl: Use 'using' to be explicit about visitor overloads

Clang has a warning about overloading virtuals that triggers when a
derived class defines a virtual function that's an overload of
function in the base class.  This kind of thing:

  struct chart; // let's pretend this exists
  struct Base
  {
      virtual void* get(char* e);
  };

  struct Derived: public Base {
      virtual void* get(chart* e); // typo, we wanted to override the same function
  };

The solution is to use

  using Base::get;

to be explicit about the intention to reuse the base class virtual.
We hit this a lot with out glsl ir_hierarchical_visitor visitor
pattern, so let's adds some 'using' to calm down the compiler.

See-also: https://stackoverflow.com/questions/18515183/c-overloaded-virtual-function-warning-by-clang)
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3686>

23 months agospirv/opencl: Cast opcode up front to avoid warnings
Kristian H. Kristensen [Mon, 3 Feb 2020 20:11:45 +0000 (12:11 -0800)]
spirv/opencl: Cast opcode up front to avoid warnings

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3686>

23 months agofreedreno/fdperf: Cast away some ignored return values
Kristian H. Kristensen [Mon, 3 Feb 2020 19:49:15 +0000 (11:49 -0800)]
freedreno/fdperf: Cast away some ignored return values

This is developer tool, it can crash and burn if it fails to allocate.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3686>

23 months agonir: Make unroll pragma work on clang
Kristian H. Kristensen [Mon, 3 Feb 2020 19:48:18 +0000 (11:48 -0800)]
nir: Make unroll pragma work on clang

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3686>

23 months agonir: Delete unused is_var_constant() helper
Kristian H. Kristensen [Mon, 3 Feb 2020 19:47:12 +0000 (11:47 -0800)]
nir: Delete unused is_var_constant() helper

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3686>

23 months agoRevert "gitlab-ci: disable a630 tests as mesa-cheza is down"
Fritz Koenig [Mon, 3 Feb 2020 21:45:16 +0000 (21:45 +0000)]
Revert "gitlab-ci: disable a630 tests as mesa-cheza is down"

This reverts commit f38851d84c583b1c62ea95edbc42eb5e2ad14fa8

Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3687>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3687>

23 months agoclover: Use explicit conversion from llvm::StringRef to std::string
Jan Vesely [Thu, 30 Jan 2020 06:23:54 +0000 (01:23 -0500)]
clover: Use explicit conversion from llvm::StringRef to std::string

Fixes build after llvm 777180a32b61070a10dd330b4f038bf24e916af1
("[ADT] Make StringRef's std::string conversion operator explicit")

CC: <mesa-stable@lists.freedesktop.org>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
23 months agozink: disallow depth-stencil blits with format-change
Erik Faye-Lund [Fri, 1 Nov 2019 12:35:55 +0000 (13:35 +0100)]
zink: disallow depth-stencil blits with format-change

The Vulkan spec says this about vkCmdBlitImage:

   "No format conversion is supported between depth/stencil images. The
    formats must match."

So yeah, let's stop trying to do this.

Reviewed-by: Dave Airlie <airlied@redhat.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3681>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3681>

23 months agozink: be more careful about the mask-check
Erik Faye-Lund [Fri, 1 Nov 2019 12:21:59 +0000 (13:21 +0100)]
zink: be more careful about the mask-check

We currently disallow blits that we can support. Let's be more accurate
when checking the mask.

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3681>

23 months agopanfrost: Fix the damage box clamping logic
Boris Brezillon [Fri, 31 Jan 2020 12:42:19 +0000 (13:42 +0100)]
panfrost: Fix the damage box clamping logic

When the rendering are is not covering the whole FBO, and the biggest
damage rect is empty, we can have damage.max{x,y} > damage.min{x,y},
which leads to invalid reload boxes.

Fixes: 65ae86b85422 ("panfrost: Add support for KHR_partial_update()")
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3676>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3676>

23 months agopan/midgard: Stop leaking instruction objects in mir_schedule_alu()
Boris Brezillon [Fri, 31 Jan 2020 09:46:04 +0000 (10:46 +0100)]
pan/midgard: Stop leaking instruction objects in mir_schedule_alu()

Allocate those instructions with ralloc() instead of using mem_dup.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3676>

23 months agopan/midgard: Don't check 'branch && branch->writeout' twice in mir_schedule_alu()
Boris Brezillon [Fri, 31 Jan 2020 09:41:35 +0000 (10:41 +0100)]
pan/midgard: Don't check 'branch && branch->writeout' twice in mir_schedule_alu()

There's a writeout bool storing the result of this test. Use it instead
of duplicating the test.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3676>

23 months agopan/midgard: Lower bitfield extract to shifts
Boris Brezillon [Fri, 31 Jan 2020 08:25:07 +0000 (09:25 +0100)]
pan/midgard: Lower bitfield extract to shifts

Let's lower bitfield extract to shifts until we figure out if it can be
natively supported.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3676>

23 months agopan/midgard: Make sure we pass the right RT id to emit_fragment_store()
Boris Brezillon [Fri, 31 Jan 2020 08:22:50 +0000 (09:22 +0100)]
pan/midgard: Make sure we pass the right RT id to emit_fragment_store()

nir_intrinsic_base() is assigned nir_variable.data.driver_location,
which is assigned a unique ID based on the variable position in the
shader variable list. There's no guarantee that this position will
match the RT id we want to pass to emit_fragment_store().

Add a search_var() helper to retrieve a nir_variable based on its driver
location, so we can pass the right RT value to emit_fragment_store().

We also make sure the shader output is color, since emit_fragment_store()
is not ready for depth/stencil stores yet.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3676>

23 months agopan/midgard: Add an enum to describe the render targets
Boris Brezillon [Fri, 31 Jan 2020 07:58:22 +0000 (08:58 +0100)]
pan/midgard: Add an enum to describe the render targets

We are about to add a special ZS render target, let's add a enum so we
can easily know which render target we're dealing with.

Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3676>

23 months agoutil/os_socket: Include unistd.h to fix build error
Bernd Kuhls [Fri, 31 Jan 2020 17:48:06 +0000 (18:48 +0100)]
util/os_socket: Include unistd.h to fix build error

Fixes

In file included from ../src/util/os_socket.c:8:
../src/util/os_socket.h:26:1: error: unknown type name ‘ssize_t’; did you mean ‘size_t’?
 ssize_t os_socket_recv(int socket, void *buffer, size_t length, int flags);

seen with gcc version 8.3.0 (Buildroot 2019.11) and uClibc 1.0.32.

Reviewed-by: Eric Engestrom <eric@engestrom.ch>
Fixes: ef5266ebd50e7fa65c56 ("util/os_socket: Add socket related functions.")
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3659>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3659>

23 months agogitlab-ci: disable a630 tests as mesa-cheza is down
Eric Engestrom [Mon, 3 Feb 2020 09:58:41 +0000 (09:58 +0000)]
gitlab-ci: disable a630 tests as mesa-cheza is down

Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3677>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3677>

23 months agonv50: report max lod bias of 15.0
Ilia Mirkin [Sun, 2 Feb 2020 00:48:07 +0000 (19:48 -0500)]
nv50: report max lod bias of 15.0

The blob returns 15, the state creation code clamps it to 15, but since
the dawn of time we've returned 4.0. Setting this to 15 also fixes

GTF-GL33.gtf21.GL3Tests.texture_lod_bias.texture_lod_bias_clamp_m_le_M

which is sensitive to these limits.

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
23 months agoegl: put full path to libEGL_mesa.so in GLVND json
Eric Engestrom [Fri, 6 Dec 2019 16:17:12 +0000 (16:17 +0000)]
egl: put full path to libEGL_mesa.so in GLVND json

This is useful when installing to a non-standard path.

glvnd_icd.py copied & adapted from src/intel/vulkan/anv_icd.py

Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Acked-by: Matt Turner <mattst88@gmail.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3038>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3038>

23 months agoradv: Allow non-dedicated linear images and buffer.
Bas Nieuwenhuizen [Mon, 27 Jan 2020 15:50:23 +0000 (16:50 +0100)]
radv: Allow non-dedicated linear images and buffer.

Requested for virtualized Vulkan as they need to export memory to
map it.

Since radeonsi and the kernel assume an image without metadata is
linear, this should work just fine.

Reviewed-by: Chia-I Wu <olvaffe@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3583>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3583>

23 months agopan/midgard: Implement mixed-type constant packing
Alyssa Rosenzweig [Fri, 31 Jan 2020 13:11:44 +0000 (08:11 -0500)]
pan/midgard: Implement mixed-type constant packing

Lot of churn but mostly just specializes types per source instead of per
instruction.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3653>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3653>

23 months agopan/midgard: Break out one-src read_components
Alyssa Rosenzweig [Fri, 31 Jan 2020 13:11:13 +0000 (08:11 -0500)]
pan/midgard: Break out one-src read_components

For constant packing, this is interesting to break down further.

Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3653>

23 months agopanfrost: Fix non-debug builds
Icecream95 [Sat, 1 Feb 2020 00:50:35 +0000 (13:50 +1300)]
panfrost: Fix non-debug builds

For non-debug builds, where assertions are compiled out, GCC complains
about the end of the non-void function panfrost_translate_channel_width
being reached.

Fixes: 226c1efe9a8 ("panfrost: Add more info to some assertions")
Reported-by: Piotr Oniszczuk
Suggested-by: Boris Brezillon <boris.brezillon@collabora.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3665>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3665>

23 months agoanv/blorp: Use the correct size for vkCmdCopyBufferToImage
Jason Ekstrand [Fri, 31 Jan 2020 22:08:31 +0000 (16:08 -0600)]
anv/blorp: Use the correct size for vkCmdCopyBufferToImage

Now that we're using an uncompressed format for the buffer, we have to
scale down the dimensions we pass into BLORP when doing buffer->image
copies.

Fixes: dd92179a72 "anv: Canonicalize buffer formats for image/buffer..."
Closes: #2452
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3664>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3664>

23 months agoVERSION: bump after 20.0 branch point
Eric Engestrom [Sun, 2 Feb 2020 06:54:14 +0000 (06:54 +0000)]
VERSION: bump after 20.0 branch point

Closes: #2457
23 months agolima: Fix build with GCC 10.
Vinson Lee [Fri, 31 Jan 2020 06:19:32 +0000 (22:19 -0800)]
lima: Fix build with GCC 10.

This patch fixes this build error with GCC 10.

/usr/bin/ld: src/gallium/drivers/lima/liblima.a(lima_context.c.o):src/gallium/drivers/lima/lima_util.h:32: multiple definition of `lima_dump_command_stream'; src/gallium/drivers/lima/liblima.a(lima_screen.c.o):src/gallium/drivers/lima/lima_util.h:32: first defined here
/usr/bin/ld: src/gallium/drivers/lima/liblima.a(lima_resource.c.o):src/gallium/drivers/lima/lima_util.h:32: multiple definition of `lima_dump_command_stream'; src/gallium/drivers/lima/liblima.a(lima_screen.c.o):src/gallium/drivers/lima/lima_util.h:32: first defined here
/usr/bin/ld: src/gallium/drivers/lima/liblima.a(lima_draw.c.o):src/gallium/drivers/lima/lima_util.h:32: multiple definition of `lima_dump_command_stream'; src/gallium/drivers/lima/liblima.a(lima_screen.c.o):src/gallium/drivers/lima/lima_util.h:32: first defined here
/usr/bin/ld: src/gallium/drivers/lima/liblima.a(lima_bo.c.o):src/gallium/drivers/lima/lima_util.h:32: multiple definition of `lima_dump_command_stream'; src/gallium/drivers/lima/liblima.a(lima_screen.c.o):src/gallium/drivers/lima/lima_util.h:32: first defined here
/usr/bin/ld: src/gallium/drivers/lima/liblima.a(lima_submit.c.o):src/gallium/drivers/lima/lima_util.h:32: multiple definition of `lima_dump_command_stream'; src/gallium/drivers/lima/liblima.a(lima_screen.c.o):src/gallium/drivers/lima/lima_util.h:32: first defined here
/usr/bin/ld: src/gallium/drivers/lima/liblima.a(lima_util.c.o):src/gallium/drivers/lima/lima_util.h:32: multiple definition of `lima_dump_command_stream'; src/gallium/drivers/lima/liblima.a(lima_screen.c.o):src/gallium/drivers/lima/lima_util.h:32: first defined here
/usr/bin/ld: src/gallium/drivers/lima/liblima.a(lima_texture.c.o):src/gallium/drivers/lima/lima_util.h:32: multiple definition of `lima_dump_command_stream'; src/gallium/drivers/lima/liblima.a(lima_screen.c.o):src/gallium/drivers/lima/lima_util.h:32: first defined here

Fixes: d71cd245d744 ("lima: Rotate dump files after each finished pp frame")
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Qiang Yu <yuq825@gmail.com>
23 months agofreedreno/ir3: fix a dirty lie
Rob Clark [Mon, 27 Jan 2020 16:35:20 +0000 (08:35 -0800)]
freedreno/ir3: fix a dirty lie

Lies, damn lies, and leftover hacks!

We no longer hard-code these two, so fix the disasm to print the correct
values.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Tested-by: Marge Bot <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>

23 months agofreedreno/ir3: simplify split from collect
Rob Clark [Mon, 27 Jan 2020 16:29:35 +0000 (08:29 -0800)]
freedreno/ir3: simplify split from collect

In some cases we need to split components out from what was already a
collect.  That was making it hard to DCE unused components of the
collect.  (Ie. unused components of fragcoord, etc)

So just detect this case and skip the chained collect+split.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>

23 months agofreedreno/ir3: create fragcoord instructions in input block
Rob Clark [Mon, 27 Jan 2020 16:26:04 +0000 (08:26 -0800)]
freedreno/ir3: create fragcoord instructions in input block

This was somehow working to create the instructions in a random block,
and use the value in other blocks, by dumb luck.  But two-pass-RA's
better choice of register assignment causes a couple dEQPs to start
failing without this fix:

dEQP-GLES3.functional.shaders.metamorphic.bubblesort_flag.variant_1
dEQP-GLES3.functional.shaders.metamorphic.bubblesort_flag.variant_2

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>

23 months agofreedreno/ir3: remove unused tex arg harder
Rob Clark [Fri, 24 Jan 2020 00:45:29 +0000 (16:45 -0800)]
freedreno/ir3: remove unused tex arg harder

Just killing the SSA link isn't enough.  It confuses RA, legalize,
and postsched to see a bogus unused reg.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>

23 months agofreedreno/ir3: add RA sanity check
Rob Clark [Wed, 15 Jan 2020 22:52:43 +0000 (14:52 -0800)]
freedreno/ir3: add RA sanity check

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>

23 months agofreedreno/a6xx: fix lrz overflow
Rob Clark [Sun, 12 Jan 2020 01:30:24 +0000 (17:30 -0800)]
freedreno/a6xx: fix lrz overflow

Running the complete deqp_gles2 seems to trigger an overflow in lrz
cmdstream.  We skip the blit clear fast-path if there have been any
draws (so mid-batch clears of any attached buffer hit the 3d pipe).
Which means it is safe to simply discard any lrz clear rendering.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>

23 months agofreedreno/ir3: two pass register allocation
Rob Clark [Tue, 22 Oct 2019 21:00:09 +0000 (14:00 -0700)]
freedreno/ir3: two pass register allocation

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>

23 months agofreedreno/ir3: don't precolor unused inputs
Rob Clark [Sat, 18 Jan 2020 21:44:26 +0000 (13:44 -0800)]
freedreno/ir3: don't precolor unused inputs

This apparently can happen with gs/tess.  And will cause problems with
two-pass-ra, so lets just skip them.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>

23 months agofreedreno/ir3: add is_tex_or_prefetch()
Rob Clark [Sat, 18 Jan 2020 19:03:38 +0000 (11:03 -0800)]
freedreno/ir3: add is_tex_or_prefetch()

Some of the aspects of tex prefetch are in common with normal tex
instructions, such as having a wrmask to control which components
are written.  Add a helper for this.

This should result in actually using the prefetch wrmask to avoid
fetching unneeded components.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>

23 months agofreedreno/ir3: number instructions from one
Rob Clark [Wed, 15 Jan 2020 22:23:23 +0000 (14:23 -0800)]
freedreno/ir3: number instructions from one

ra_block_compute_live_ranges() treats zero as "not yet defined", so
probably best to not let this be a valid instruction #

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>

23 months agofreedreno/ir3: post-RA sched pass
Rob Clark [Fri, 13 Dec 2019 22:09:39 +0000 (14:09 -0800)]
freedreno/ir3: post-RA sched pass

After RA, we can schedule to increase parallelism (reduce nop's) without
worrying about increasing register pressure.  This pass lets us cut down
the instruction count ~10%, and prioritize bary.f, kill, etc, which
would tend to increase register pressure if we tried to do that before
RA.

It should be more useful if RA round-robin'd register choices.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>

23 months agofreedreno/ir3: fix kill scheduling
Rob Clark [Thu, 23 Jan 2020 18:26:27 +0000 (10:26 -0800)]
freedreno/ir3: fix kill scheduling

kill (and other cat0/flow instructions) do not have a dst register.
Which was mostly harmless before, other than RA thinking it would need
a free register to write.  (But nothing consumed it, so the value would
be immediately dead.)  But this would cause more problems with postsched
which would see a bogus dependency.

Also, post-RA sched *does* need to see the dependency on the predicate
register.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>

23 months agofreedreno/ir3/ra: make use()/def() functions instead of macros
Rob Clark [Tue, 22 Oct 2019 23:05:41 +0000 (16:05 -0700)]
freedreno/ir3/ra: make use()/def() functions instead of macros

Originally these were nested functions, which worked nicely, giving us
the function of a local macro that was actual 'c' syntax (ie. not token
pasted macro).  But these were converted to macros because clang doesn't
let us have nice gcc extensions.

Extract these back out into functions, before adding more things and
making the macros even more cumbersome.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>

23 months agofreedreno/ir3: a bit more optmsgs debug
Rob Clark [Thu, 19 Dec 2019 21:58:32 +0000 (13:58 -0800)]
freedreno/ir3: a bit more optmsgs debug

Also dump where arrays are allocated.  This was useful for debugging.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>

23 months agofreedreno/ir3: move atomic fixup after RA
Rob Clark [Fri, 10 Jan 2020 22:07:03 +0000 (14:07 -0800)]
freedreno/ir3: move atomic fixup after RA

A post-RA sched pass will move the extra mov's to the wrong place, so
rework the fixup so it can run after RA (and therefore after postsched)

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>

23 months agofreedreno/ir3: move block-scheduling into legalize
Rob Clark [Wed, 18 Dec 2019 22:04:36 +0000 (14:04 -0800)]
freedreno/ir3: move block-scheduling into legalize

We want to do this only once.  If we have post-RA sched pass, then we
don't want to do it pre-RA.  Since legalize is where we resolve the
branch/jumps, we might as well move this into legalize.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>

23 months agofreedreno/ir3: move nop padding to legalize
Rob Clark [Wed, 18 Dec 2019 19:57:41 +0000 (11:57 -0800)]
freedreno/ir3: move nop padding to legalize

This way we can deal with it in one place, *after* all the blocks have
been scheduled.  Which will simplify life for a post-RA sched pass.

This has the benefit of already taking into account nop's that legalize
has to insert for non-delay related reasons.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>

23 months agofreedreno/ir3: split out delay helpers
Rob Clark [Wed, 18 Dec 2019 19:10:12 +0000 (11:10 -0800)]
freedreno/ir3: split out delay helpers

We're going to want these also for a post-RA sched pass.  And also to
split nop stuffing out into it's own pass.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>

23 months agofreedreno/ir3: fix crash when no non-input instructions
Rob Clark [Thu, 19 Dec 2019 17:48:39 +0000 (09:48 -0800)]
freedreno/ir3: fix crash when no non-input instructions

This scenario can come up with block-sched and nop-sched moved to after
RA.  So lets fix it first to keep things bisectable.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>

23 months agofreedreno/ir3: cleanup after lower_locals_to_regs
Rob Clark [Tue, 17 Dec 2019 19:57:14 +0000 (11:57 -0800)]
freedreno/ir3: cleanup after lower_locals_to_regs

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3569>