Mesa (master): 31 new commits
Marek Olšák
mareko at kemper.freedesktop.org
Wed Oct 26 11:04:46 UTC 2016
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=1ac40173c2abf402150ffa245d03664b159e89e0
Author: Marek Olšák <marek.olsak at amd.com>
Date: Tue Oct 18 23:47:01 2016 +0200
configure.ac: simplify EGL requirements for drivers dependent on EGL
Reviewed-by: Emil Velikov <emil.velikov at collabora.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=b687f766fddb7b39479cd9ee0427984029ea3559
Author: Marek Olšák <marek.olsak at amd.com>
Date: Tue Oct 25 13:10:49 2016 +0200
st/mesa: allow multiple concurrent waiters in ClientWaitSync
so->fence can be unreferenced by one thread while another thread is
somewhere in ClientWaitSync and expecting so->fence to be non-NULL.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=98172
Cc: 12.0 13.0 <mesa-stable at lists.freedesktop.org>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=f240ad98bc05281ea7013d91973cb5f932ae9434
Author: Marek Olšák <marek.olsak at amd.com>
Date: Tue Oct 25 12:58:03 2016 +0200
st/mesa: unduplicate st_check_sync code
It's the same as st_client_wait_sync. Discovered by Michel.
This is needed to make the following fix simpler.
Cc: 12.0 13.0 <mesa-stable at lists.freedesktop.org>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=ad45dce4a2896639a28a91f4250e40dae9d43b1f
Author: Marek Olšák <marek.olsak at amd.com>
Date: Mon Oct 24 23:29:50 2016 +0200
radeonsi: remove si_resource_create_custom
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=29144d0f34d9325a3549e4ed0feecc0577c70358
Author: Marek Olšák <marek.olsak at amd.com>
Date: Mon Oct 24 23:26:39 2016 +0200
gallium/radeon: stop using PIPE_BIND_CUSTOM
it has no effect whatsoever
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=e3c3a7fadaf87f68b5d8053d21d8c986a90dec30
Author: Marek Olšák <marek.olsak at amd.com>
Date: Mon Oct 24 02:21:38 2016 +0200
r600g: remove a redundant buffer_create helper
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=3dc78c33a95fc50191cfe0713865197338ef65b0
Author: Marek Olšák <marek.olsak at amd.com>
Date: Mon Oct 24 18:29:43 2016 +0200
gallium/radeon: remove unused r600_cmask_info members
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=d18bf0b9441934c5119351b19e628f9e8df0dd3a
Author: Marek Olšák <marek.olsak at amd.com>
Date: Mon Oct 24 18:22:31 2016 +0200
gallium/radeon: don't force the same tiling parameters for FMASK
GCN can use a completely different tile mode for FMASK.
FMASK allocation now skips one unrelated amdgpu_surface_init codepath as
hinted by the assertion.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=ecf045b4f7c83e2ea070e490d9475c2132308ee4
Author: Marek Olšák <marek.olsak at amd.com>
Date: Mon Oct 24 18:01:52 2016 +0200
winsys/amdgpu: allocate FMASK properly
I expect no change in behavior, because r600_texture.c forces the same
tile mode as the base texture has.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=24faeb94bea35d750132fb27be4bad94caf064f3
Author: Marek Olšák <marek.olsak at amd.com>
Date: Mon Oct 24 17:57:19 2016 +0200
gallium/radeon: print tiling index when printing texture info
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=37659071b870d3456df5c5a4f996c9fd4f1930d9
Author: Marek Olšák <marek.olsak at amd.com>
Date: Mon Oct 24 01:31:05 2016 +0200
gallium/radeon: don't do (fmask.size && cmask.size)
fmask implies that cmask is present too.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=2664351dfeeba2c1d0de272cdf6d5fd940a367e9
Author: Marek Olšák <marek.olsak at amd.com>
Date: Sun Oct 23 21:28:29 2016 +0200
gallium/radeon: re-order radeon_surf::dcc and htile members
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=2a2e53757796b3fed3119cb033e5cf5144843850
Author: Marek Olšák <marek.olsak at amd.com>
Date: Sun Oct 23 21:26:43 2016 +0200
gallium/radeon: rename bo_size -> surf_size, bo_alignment -> surf_alignment
these names were misleading.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=67a44c97afb72812639039eb4594592c91c9ead5
Author: Marek Olšák <marek.olsak at amd.com>
Date: Sun Oct 23 21:03:40 2016 +0200
gallium/radeon: remove flags specific to libdrm_radeon from winsys interface
These just say whether libdrm can assume that the latest radeon_surface
definition is used by Mesa.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=7a706ad25cefc666b54ddf778d047691a575b689
Author: Marek Olšák <marek.olsak at amd.com>
Date: Sun Oct 23 16:43:22 2016 +0200
gallium/radeon: remove r600_htile_info
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=7e73ff87c0255a8e0498a47991b640cdece35928
Author: Marek Olšák <marek.olsak at amd.com>
Date: Sun Oct 23 16:09:58 2016 +0200
gallium/radeon: remove unnecessary fields from radeon_surf_level
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=d5c7ea3b83168d8fd77ed4bd834901209e1d47da
Author: Marek Olšák <marek.olsak at amd.com>
Date: Sun Oct 23 15:51:46 2016 +0200
gallium/radeon: decrease the size of radeon_surf
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=e9590d9092fbcd551f5597149c64016b6f861233
Author: Marek Olšák <marek.olsak at amd.com>
Date: Sun Oct 23 20:17:32 2016 +0200
gallium/radeon: pass pipe_resource and other params to surface_init directly
This removes input-only parameters from the radeon_surf structure.
Some of the translation logic from pipe_resource to radeon_surf is moved to
winsys/radeon.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=8b94976df9996c222e5510f491fc865dae3bc412
Author: Marek Olšák <marek.olsak at amd.com>
Date: Sun Oct 23 20:11:17 2016 +0200
radeon/vce: use nblk_y instead of npix_y
npix_y will be removed. level[0].npix_y will be removed too. nblk_y should
be the same as npix_y if the block height == 1. However, nblk_y is aligned
to the tile size, so it can be greater than npix_y.
If that's a problem, we'll have to save the input height of surface_init
and use that.
Reviewed-by: Christian König <christian.koenig at amd.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=ba174b8dfff0777075f62a9188124d0bb22a5f17
Author: Marek Olšák <marek.olsak at amd.com>
Date: Sun Oct 23 20:07:39 2016 +0200
gallium/radeon: define RADEON_SURF_MODE_* as enums
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=b5118fe0543400295be5b1a65a5001882498831e
Author: Marek Olšák <marek.olsak at amd.com>
Date: Sun Oct 23 19:55:19 2016 +0200
gallium/radeon: stop using some input fields from radeon_surface
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=28d237d63d8161124e62b5dab6f864de494cd652
Author: Marek Olšák <marek.olsak at amd.com>
Date: Sun Oct 23 19:18:46 2016 +0200
gallium/radeon: fold r600_setup_surface into r600_init_surface
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=b0d8a717a7d3fbc857dea415f9152931dbb59b62
Author: Marek Olšák <marek.olsak at amd.com>
Date: Fri Oct 21 13:30:31 2016 +0200
winsys/amdgpu: remove unused definitions
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=81a95946dab730c1cfcf04ac1373cea5f7908f65
Author: Marek Olšák <marek.olsak at amd.com>
Date: Thu Oct 20 22:57:03 2016 +0200
gallium/radeon: fold radeon_winsys::surface_best into radeon/winsys
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=dc6bbe2dd0eb9ba421c8454b3b6179a6b0452b40
Author: Marek Olšák <marek.olsak at amd.com>
Date: Mon Oct 3 15:37:19 2016 +0200
gallium/radeon: use r600_gfx_write_event_eop everywhere
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=462e3cdf3b92464dab57da2b9e00986f8e6091be
Author: Marek Olšák <marek.olsak at amd.com>
Date: Mon Oct 3 15:32:36 2016 +0200
gallium/radeon: make r600_gfx_write_fence more generic
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=edf56fb428b1cda51264270881dda3af8ec01a04
Author: Marek Olšák <marek.olsak at amd.com>
Date: Mon Oct 3 15:13:56 2016 +0200
gallium/radeon: fix a ZPASS comment, EVENT_WRITE_EOP fixups
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=d883c83ba99c5459a8bb9183420124518b489ff8
Author: Marek Olšák <marek.olsak at amd.com>
Date: Mon Oct 24 19:32:31 2016 +0200
radeonsi: enable SDMA on Carrizo and all CIK chips again
SDMA might be fixed by:
"winsys/amdgpu: fix radeon_surf::macro_tile_index for imported textures"
Reviewed-by: Bas Nieuwenhuizen <bas at basnieuwenhuizen.nl>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=6ec3b2a4b1d41b83a4721d06b42c49f55e695cbf
Author: Marek Olšák <marek.olsak at amd.com>
Date: Mon Oct 24 19:05:10 2016 +0200
winsys/amdgpu: fix radeon_surf::macro_tile_index for imported textures
Maybe this is why SDMA has been broken for many amdgpu users?
SDMA is the only block which is used with imported textures and relies
on this variable. DB also uses it, but it doesn't get imported textures,
so it's unaffected.
I do get SDMA failures on Tonga before this patch if R600_DEBUG=testdma
is changed to use imported textures.
Cc: 11.2 12.0 13.0 <mesa-stable at lists.freedesktop.org>
Reviewed-by: Bas Nieuwenhuizen <bas at basnieuwenhuizen.nl>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=dce05b342355eac9296ee7110385b16d6edb059d
Author: Marek Olšák <marek.olsak at amd.com>
Date: Mon Oct 24 21:16:11 2016 +0200
gallium/radeon: make sure the address of separate CMASK is aligned properly
This should fix random GPU hangs on Hawaii and Fiji.
Cc: 11.2 12.0 13.0 <mesa-stable at lists.freedesktop.org>
Reviewed-by: Bas Nieuwenhuizen <bas at basnieuwenhuizen.nl>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=8a21f52d73936e23a314a288a36782a698c7c1b9
Author: Marek Olšák <marek.olsak at amd.com>
Date: Sun Oct 23 15:29:18 2016 +0200
gallium/radeon: fix incorrect bpe use in si_set_optimal_micro_tile_mode
Oh my god, I wonder what catastrophic issues this was causing on SI.
Cc: 13.0 <mesa-stable at lists.freedesktop.org>
Reviewed-by: Bas Nieuwenhuizen <bas at basnieuwenhuizen.nl>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle at amd.com>
More information about the mesa-commit
mailing list