Marek Olšák [Thu, 8 Sep 2016 17:12:00 +0000 (19:12 +0200)]
tgsi/scan: don't set interp flags for inputs only used by INTERP instructions
radeonsi depends on the interp flags a little bit too much.
This fixes 9 randomly failing tests:
GL45-CTS.shader_multisample_interpolation.render.interpolate_at_centroid.*
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Sun, 11 Sep 2016 23:38:53 +0000 (01:38 +0200)]
radeonsi: fix FP64 UBO loads with indirect uniform block indexing
No known tests.
Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Sun, 11 Sep 2016 19:49:24 +0000 (21:49 +0200)]
winsys/amdgpu: don't assume GTT if the VRAM flag isn't set
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Sun, 11 Sep 2016 15:57:31 +0000 (17:57 +0200)]
radeonsi: clean up CP DMA emit code
Unify the clear and copy paths, clean up the definitions.
It looks more like a rework. It's a preparation for GDS support,
which might or might not come.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Mon, 12 Sep 2016 18:23:55 +0000 (20:23 +0200)]
radeonsi: print the IB and buffer list in VM fault reports
This is a fallout from reworking the debug flags.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Mon, 12 Sep 2016 10:49:34 +0000 (12:49 +0200)]
radeonsi: add sampler view BOs to the BO list last
If si_sampler_view_add_buffer ends up flushing, then the code
in begin_new_cs would previously have added the buffer(s) for
whatever was previously bound to that slot. Now it would add only
the new buffer.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Fri, 9 Sep 2016 23:21:11 +0000 (01:21 +0200)]
radeonsi: export SampleMask from pixel shaders at full rate
Heaven and Valley write gl_SampleMask and not Z.
Use 16_ABGR instead of 32_ABGR if Z isn't written.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Mon, 12 Sep 2016 10:47:13 +0000 (12:47 +0200)]
gallium/radeon: set new r600_resource fields correctly in other places too
This was missed in:
commit
0d2e43fcb1198a6e67c85feadb1ca8c360ddc284
Author: Marek Olšák <marek.olsak@amd.com>
Date: Thu Aug 18 16:30:00 2016 +0200
gallium/radeon: derive buffer placement and flags only at initialization
Tested-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Sat, 10 Sep 2016 08:16:54 +0000 (10:16 +0200)]
ddebug: dump shader buffers and images
this was unimplemented
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Sat, 10 Sep 2016 07:50:42 +0000 (09:50 +0200)]
ddebug: fix a crash in resource_get_handle
broken recently
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Jan Vesely [Tue, 13 Sep 2016 00:47:37 +0000 (20:47 -0400)]
radeon: Don't check DCC on pipe buffers
Fixes segfaults in EG compute since:
commit
21de3be8e62b2b093569a99550e6356ed2f106b4
radeonsi: fix texture format reinterpretation with DCC
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Andy Furniss [Tue, 13 Sep 2016 17:52:31 +0000 (13:52 -0400)]
vl/util: Fix YV12/I420 convert to NV12 U/V reversal
Fix VAAPI YV12/I420 convert to NV12 U/V reversal.
Input order is YVU when this is called.
Signed-off-by: Andy Furniss <adf.lists@gmail.com>
Reviewed-by: Boyuan Zhang <boyuan.zhang@amd.com>
Jason Ekstrand [Thu, 8 Sep 2016 04:33:48 +0000 (21:33 -0700)]
anv/allocator: Use VG_NOACCESS_WRITE in anv_bo_pool_free
Previously, we were relying on the fact that VALGRIND_MEMPOOL_FREE came
later on in the function to prevent "link->bo = bo" from causing an invalid
write. However, in the case where the size requested by the user is very
small (less than sizeof(struct anv_bo)), this isn't sufficient. Instead,
we should call VALGRIND_MEMPOOL_FREE early and then use VG_NOACCESS_WRITE.
We do, however, have to call VALGRIND_MEMPOOL_FREE after reading bo_in
because it may be stored in the bo itself.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Jason Ekstrand [Tue, 13 Sep 2016 05:31:00 +0000 (22:31 -0700)]
intel/isl: Ignore base_array_layer and array_len for 3D storage surfaces
The time we want to restrict the Z range of a 3-D surface is when rendering
to it. For storage surfaces, we always want he full range. However, we
still need to set MinimumArrayElement and RenderTargetViewExtent to
sensible values so we'll just set them to the reasonable defaults we used
before we started respecting the base_array_layer and array_len.
This fixes a bunch of Vulkan CTS regressions caused by
48f195d7c6483ed.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97790
Reviewed-by: Chad Versace <chadversary@chromium.org>
Jose Fonseca [Tue, 13 Sep 2016 16:53:30 +0000 (17:53 +0100)]
appveyor: Update winflexbison download URL.
This particular version got moved into a `old_versions` subdirectory.
Jason Ekstrand [Fri, 9 Sep 2016 04:34:16 +0000 (21:34 -0700)]
i965: Use blorp_copy for all copy_image operations on gen6+
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Chad Versace <chadversary@chromium.org>
Jason Ekstrand [Fri, 9 Sep 2016 04:27:01 +0000 (21:27 -0700)]
i965/blorp: Add a copy_miptrees helper
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Chad Versace <chadversary@chromium.org>
Jason Ekstrand [Fri, 9 Sep 2016 05:18:47 +0000 (22:18 -0700)]
intel/isl: Add support for RGB formats in X and Y-tiled memory
Normally, using a non-linear tiling format helps improve cache locality by
ensuring that neighboring pixels are usually close-by in memory. For RGB
formats, this still sort-of holds, but it can also lead to rather terrible
memory access patterns where a single RGB pixel value crosses a tile
boundary and gets split into two pieces in different 4K pages. It also
makes for some rather awkward calculations because your tile size is no
longer an even multiple of surface element size. For these reasons, we
chose to simply never create tiled RGB images in the Vulkan driver.
The GL driver, however, is not so kind so we need to support it somehow. I
briefly toyed with a couple of different schemes but this is the best one I
could come up with. The fundamental problem is that a tile no longer
contains an integer number of surface elements. I briefly considered a
couple other options but found them wanting:
1) Using floats for the logical tile size. This leads to potential
rounding error problems.
2) When presented with a RGB format, just make the tile 3-times as wide.
This isn't so nice because now our tiles are no longer power-of-two
size. Also, it can force the row_pitch to be larger than needed which,
while not strictly a problem for ISL, causes incompatibility problems
with the way the GL driver chooses surface pitches.
The chosen method requires that you pay attention and not just assume that
your tile_info is in the units you think it is. However, it's nice because
it provides a nice "these are the units" declaration in isl_tile_info
itself. Previously, the tile_info wasn't usable as a stand-alone structure
because you had to also know the format. It also forces figuring out how
to deal with inconsistencies between tiling and format back to the caller
which is good because the two different consumers of isl_tile_info really
want to deal with it differently: Computation of the surface size wants
the fewest number of horizontal tiles possible while get_intratile_offset
is far more concerned with things aligning nicely.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Acked-by: Chad Versace <chadversary@chromium.org>
Jason Ekstrand [Fri, 9 Sep 2016 15:57:14 +0000 (08:57 -0700)]
intel/isl: Allow valign2 for texture-only Y-tiled surfaces on gen7
The restriction that Y-tiled surfaces must have valign == 4 only aplies to
render targets but we were applying it universally. This causes problems
if ISL_FORMAT_R32G32B32_FLOAT is used because it requires valign == 2; this
should be okay because you can't render to that format.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Chad Versace <chadversary@chromium.org>
Jason Ekstrand [Sat, 3 Sep 2016 18:40:09 +0000 (11:40 -0700)]
intel/blorp: Work in terms of logical array layers
When Ivy Bridge introduced array multisampling, someone made the decision
to do lots of stuff throughout the driver in terms of physical array layers
rather than logical array layers. In ISL, we use logical array layers most
of the time and it really makes no sense to use physical array layers in
the blorp API. Every time someone passes physical array layers into blorp
for an array multisampled surface, they're always divisible by the number
of samples and we divide right away.
Eventually, I'd like to rework most of the GL driver internals to use
logical array layers but that's going to be a big project and will probably
happen as part of the ISL conversion. For now, we'll do the conversion in
brw_blorp and let blorp just use the logical layers.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Jason Ekstrand [Sat, 3 Sep 2016 16:49:24 +0000 (09:49 -0700)]
intel/blorp: Increase the presision of coordinate transform calculations
The result of this calculation goes into an fma() in the shader and we
would like it to be as precise as possible. The division in particular
was a source of imprecision whenever dst1 - dst0 was not a power of two.
This prevents regressions in some of the new Vulkan CTS tests for blitting
using a filtering of NEAREST.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Wed, 31 Aug 2016 01:01:27 +0000 (18:01 -0700)]
intel/blorp: Add a swizzle parameter to blorp_clear
While we're here, we also re-arrange the parameters to better match the
parameter order of blorp_blit.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Jason Ekstrand [Tue, 30 Aug 2016 23:55:35 +0000 (16:55 -0700)]
intel/blorp: Make color_write_disable const and optional
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Jason Ekstrand [Tue, 30 Aug 2016 23:32:51 +0000 (16:32 -0700)]
intel/blorp: Add support for clearing R9G9B9E5 surfaces
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Jason Ekstrand [Sat, 27 Aug 2016 19:07:31 +0000 (12:07 -0700)]
intel/blorp: Add support for RGB destinations in copies
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Jason Ekstrand [Tue, 30 Aug 2016 20:13:43 +0000 (13:13 -0700)]
intel/blorp: Add an entrypoint for doing bit-for-bit copies
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Jason Ekstrand [Tue, 30 Aug 2016 19:49:54 +0000 (12:49 -0700)]
intel/blorp: Pull the guts of blorp_blit into a helper
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Jason Ekstrand [Tue, 30 Aug 2016 18:18:39 +0000 (11:18 -0700)]
intel/blorp: Stop using the X/YOffset field of RENDER_SURFACE_STATE
While it can be useful, the field has substantial limtations. In
particular, the bittom 2 or 3 bits is missing so your offset always has to
be a multiple of 4 or 8. While surface alignments usually work out to make
this ok, when you start trying to fake compressed surfaces as uncompressed
(which we will want to do) this falls apart. The easiest solution is to
simply align all offsets to a tile boundary and munge the regions we're
copying to account for the intratile offset.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Jason Ekstrand [Tue, 30 Aug 2016 01:05:11 +0000 (18:05 -0700)]
intel/blorp: Use fake_interleaved_msaa in retile_w_to_y
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Jason Ekstrand [Tue, 30 Aug 2016 00:52:52 +0000 (17:52 -0700)]
intel/blorp: Use isl_get_interleaved_msaa_px_size_sa
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Jason Ekstrand [Mon, 29 Aug 2016 23:34:06 +0000 (16:34 -0700)]
intel/isl: Add a helper for getting the size of an interleaved pixel
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Jason Ekstrand [Mon, 29 Aug 2016 16:48:10 +0000 (09:48 -0700)]
intel/blorp: Handle 3D surfaces in convert_to_single_slice
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Jason Ekstrand [Mon, 29 Aug 2016 16:16:06 +0000 (09:16 -0700)]
intel/isl: Fix an assert in get_intratile_offset_sa
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Jason Ekstrand [Wed, 31 Aug 2016 19:58:54 +0000 (12:58 -0700)]
intel/blorp: Fix the early return condition in convert_to_single_slice
The convert_to_single_slice operation is *mostly* idempotent. The only
non-repeatable thing it does is that, when it sets the intratile offset
fields, it just overwrites them instead of doing a += operation. This is
supposed to be ok because we have an early return at the top that should
make it bail of the surface is already a single slice. Unfortunately, the
if condition has been broken ever since it was first added in
96fa98c18.
This commit fixes the condition and adds an assert to ensure we don't stomp
any non-zero intratile offsets.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Jason Ekstrand [Sun, 28 Aug 2016 05:44:15 +0000 (22:44 -0700)]
intel/blorp: Use the surface format for computing offsets
If we use the view format, it may be an uncompressed view of a compressed
image which throws things off. Since we're computing offsets of images, we
want the actual surface offset anyway.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Jason Ekstrand [Sun, 28 Aug 2016 05:06:11 +0000 (22:06 -0700)]
intel/blorp: Don't assume R8_UINT in convert_to_single_slice
We're going to use it for more than just stencil textures
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Jason Ekstrand [Sun, 28 Aug 2016 04:57:51 +0000 (21:57 -0700)]
intel/blorp: Take a destination swizzle in blorp_blit
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Jason Ekstrand [Sun, 28 Aug 2016 04:48:40 +0000 (21:48 -0700)]
intel/blorp: Take an isl_swizzle instead of a SWIZZLE
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Jason Ekstrand [Sun, 28 Aug 2016 04:39:16 +0000 (21:39 -0700)]
intel/isl: Add an isl_swizzle structure and use it for isl_view swizzles
This should be more compact than the enum isl_channel_select[4] that we
were using before. It's also very convenient because we already had such a
structure in the Vulkan driver we just needed to pull it over.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Kenneth Graunke [Tue, 13 Sep 2016 00:24:35 +0000 (17:24 -0700)]
docs: Add OES_tessellation_shader to the release notes.
Kenneth Graunke [Tue, 13 Sep 2016 00:23:10 +0000 (17:23 -0700)]
docs: Mark OES_tessellation_shader as done.
Ilia Mirkin [Sun, 11 Sep 2016 17:37:30 +0000 (13:37 -0400)]
st/mesa: fix is_scissor_enabled when X/Y are negative
Similar to commit
49c24d8a24 ("i965: fix noop_scissor range issue on
width/height") - take the X/Y into account to determine whether the
scissor covers the whole area or not.
Fixes the recently-added gl-1.0-scissor-depth-clear-negative-xy piglit
test.
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Cc: <mesa-stable@lists.freedesktop.org>
Mauro Rossi [Mon, 12 Sep 2016 23:15:16 +0000 (01:15 +0200)]
android: add support for libmesa_amdgpu_addrlib
Android porting of the following commits:
f1f1ba3 "radeonsi: move sid.h/r600d_common.h to a common place."
69fca64 "amd/addrlib: move addrlib from amdgpu winsys to common code"
This patch fixes android building errors
Reviewed-by: Dave Airlie <airlied@redhat.com>
Dave Airlie [Mon, 12 Sep 2016 01:21:29 +0000 (11:21 +1000)]
u_endian: add android to glibc clause
Tested-by: Mauro Rossi <issor.oruam@gmail.com>
Jason Ekstrand [Mon, 12 Sep 2016 15:18:53 +0000 (08:18 -0700)]
Revert "i965: Drop the maximum 3D texture size to 512 on Sandy Bridge"
This reverts commit
6ba88bce64b343761aabe3a6c7ee285c6020a959. The commit
was erroneous because GL has a separate limit, GL_MAX_FRAMEBUFFER_LAYERS
which guards the number of layers you are allowed to render into.
The GL 4.5 spec says:
"The framebuffer attachment point attachment is said to be framebuffer
attachment complete if [...] all of the following conditions are true:
[...]
If image is a three-dimensional, one- or two-dimensional array, or
cube map array texture and the attachment is layered, the depth or
layer count of the texture is less than or equal to the value of the
implementation-dependent limit MAX_FRAMEBUFFER_LAYERS."
and goes on to say that "framebuffer complete" requires all attachments to
be "framebuffer attachment complete".
On Sandy Bridge, we set GL_MAX_FRAMEBUFFER_LAYERS to 512 so creating a 3D
texture bigger than 512 is fine; you just can't render into all of the
slices at once.
Fixes ES3-CTS.gtf.GL3Tests.npot_textures.npot_tex_image on Sandy Bridge
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Chad Versace <chadversary@chromium.org>
Jason Ekstrand [Mon, 12 Sep 2016 18:46:22 +0000 (11:46 -0700)]
intel/blorp: Handle the 512 layers restriction on Sandy Bridge
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Mon, 12 Sep 2016 15:13:43 +0000 (08:13 -0700)]
intel/isl: Treat 3-D textures as 2-D arrays for rendering
In particular, this means that isl_view::base_array_layer and
isl_view::array_len get applied to 3-D textures but only when rendering.
We were already applying isl_view::base_array_layer for rendering into 3-D
textures so this isn't a huge deviation.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Sirisha Gandikota [Thu, 8 Sep 2016 23:15:19 +0000 (16:15 -0700)]
aubinator: Simplify gen_disasm_create()'s devinfo handling
Copy the whole devinfo structure instead of just few fields (Ken)
Earlier, copied only couple of fields which added more code. So,
simplify code by copying the whole structure.
Signed-off-by: Sirisha Gandikota <Sirisha.Gandikota@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Sirisha Gandikota [Thu, 8 Sep 2016 23:15:18 +0000 (16:15 -0700)]
aubinator: Fix compiler warning
Add 'const' qualifier to gen_field_iterator::p pointer (Ken)
Signed-off-by: Sirisha Gandikota <Sirisha.Gandikota@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Julien Isorce [Sat, 3 Sep 2016 08:10:21 +0000 (09:10 +0100)]
st/va: also honors interlaced preference when providing a video format
This fixes a crash when using the prefered video format with vaapisink
on Nvidia hardwares.
Also caught by the following assert:
nouveau_vp3_video.c:91: Assertion `templat->interlaced' failed.
TEST= gst-launch-1.0 videotestsrc ! video/x-raw, format=NV12 ! vaapisink
Cc: <mesa-stable@lists.freedesktop.org>
Signed-off-by: Julien Isorce <j.isorce@samsung.com>
Tested-by: Víctor Manuel Jáquez Leal <vjaquez@igalia.com>
Tested-by: Boyuan Zhang <boyuan.zhang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Samuel Pitoiset [Sat, 10 Sep 2016 15:10:46 +0000 (17:10 +0200)]
tgsi: document semantics for compute shaders
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Kenneth Graunke [Fri, 8 Jul 2016 18:48:15 +0000 (11:48 -0700)]
mesa: Enable OES/EXT_tessellation_shader for ES 3.1 + ARB_tess drivers.
Drivers which support ARB_tessellation_shader and ES 3.1 now will
expose OES_tessellation_shader and EXT_tessellation_shader as well.
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Acked-by: Ilia Mirkin <imirkin@alum.mit.edu>
Marek Olšák [Fri, 9 Sep 2016 22:29:31 +0000 (00:29 +0200)]
radeonsi: don't preload constants at the beginning of shaders
LLVM can CSE the loads, thus we can always re-load constants before each
use. The decrease in SGPR spilling is huge.
The best improvements are the dumbest ones.
26011 shaders in 14651 tests
Totals:
SGPRS:
1453346 ->
1251920 (-13.86 %)
VGPRS: 742576 -> 728421 (-1.91 %)
Spilled SGPRs: 52298 -> 16644 (-68.17 %)
Spilled VGPRs: 397 -> 369 (-7.05 %)
Scratch VGPRs: 1372 -> 1344 (-2.04 %) dwords per thread
Code Size:
36136488 ->
36001064 (-0.37 %) bytes
LDS: 767 -> 767 (0.00 %) blocks
Max Waves: 219315 -> 222221 (1.33 %)
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Jason Ekstrand [Mon, 12 Sep 2016 17:13:52 +0000 (10:13 -0700)]
intel/blorp: Add a TODO file
This provides a nice little place to share notes on what still needs to be
done and/or would be nice to have in BLORP.
Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Alejandro Piñeiro [Sat, 10 Sep 2016 22:07:30 +0000 (00:07 +0200)]
i965: check for GL_TEXTURE_EXTERNAL_OES at miptree_create_for_teximage
Forgotten on commit "i965: Fix calculation of the image height at start level".
Thanks to Ilia Mirkin for point it.
Fixes the following regressions on Haswell and Broadwell:
ES2-CTS.gtf.GL2ExtensionTests.egl_image_external.TestSimpleUnassociated (crash back to pass)
ES2-CTS.gtf.GL2ExtensionTests.egl_image_external.TestSimple (crash back to fail)
ES2-CTS.gtf.GL2ExtensionTests.egl_image_external.TestVertexShader (crash back to fail)
https://bugs.freedesktop.org/show_bug.cgi?id=97761
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Chuanbo Weng [Tue, 6 Sep 2016 09:28:43 +0000 (17:28 +0800)]
gbm: fix potential NULL deref of mapImage/unmapImage.
The mapImage/unmapImage functions of DRIimage extension can be NULL,
so we should add additional check for them.
Cc: <mesa-stable@lists.freedesktop.org>
Signed-off-by: Chuanbo Weng <chuanbo.weng@intel.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Emil Velikov [Mon, 12 Sep 2016 12:26:40 +0000 (13:26 +0100)]
Remove GL_GLEXT_PROTOTYPES guards from non-ext headers.
A earlier sync with the Khronos headers added _extension_ prototype
guards to all the GLES2/3/31/32 core entry points. Effectively breaking
all the applications that aim to be portable and do not set the define.
The issue has been reported to Khronos (internal bugzilla #14206) and is
being worked on. Until updated/fixed headers are released locally fix
the issue.
The following report is when building weston.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97773
Cc: Armin Krezović <krezovic.armin@gmail.com>
Cc: Emmanuel Gil Peyrot <emmanuel.peyrot@collabora.com>
Cc: Pekka Paalanen <ppaalanen@gmail.com>
Fixes: 6a5504de2fb ("Update Khronos-supplied headers to r33100")
Cc: Dave Airlie <airlied@redhat.com>
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Emil Velikov [Mon, 12 Sep 2016 11:34:10 +0000 (12:34 +0100)]
aubinator: rework print_help()
Rather than using platform specific methods to retrieve the program
name pass it explicitly. The function is called directly from main().
Similarly - basename comes in two versions POSIX (can modify string,
always pass a copy) and GNU (never modifies the string).
Just printout the complete program name, esp. since the program is not
meant to be installed. Thus using $basename is unlikely to work, not to
mention it is misleading.
Reported-by: Timothy Arceri <timothy.arceri@collabora.com>
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Jonathan Gray <jsg@jsg.id.au>
Adam Jackson [Fri, 9 Sep 2016 16:25:35 +0000 (12:25 -0400)]
docs: Note MESA_configless_context as superseded
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Adam Jackson [Fri, 9 Sep 2016 16:25:34 +0000 (12:25 -0400)]
egl: Rename MESA_configless_context bit to KHR_no_config_context
Keep the old name in the extension string, but refer to the KHR
extension internally.
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Adam Jackson [Fri, 9 Sep 2016 16:25:33 +0000 (12:25 -0400)]
egl: QueryContext on a configless context returns zero
MESA_configless_context does not specify the interaction with
QueryContext at all, and the code to generate an error in this case
predates the Mesa extension. Since EGL_NO_CONFIG_{KHR,MESA} are
numerically identical there's no way to distinguish which one the
application asked for, so use the KHR behaviour.
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
Reviewed-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Boyuan Zhang [Tue, 6 Sep 2016 20:39:45 +0000 (16:39 -0400)]
st/va: enable vbr rate control for vaapi encode
This patch enables variable bit-rate for vaapi encoding. According to va.h,
target bit-rate equals to maximum bit-rate multiplies by target_percentage.
Signed-off-by: Boyuan Zhang <boyuan.zhang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Leo Liu [Tue, 6 Sep 2016 17:14:48 +0000 (13:14 -0400)]
vl/rbsp: match initial escaped bits with valid in the buffer
Otherwise the check for the three byte will not make sense.
Signed-off-by: Leo Liu <leo.liu@amd.com>
Timothy Arceri [Sun, 11 Sep 2016 23:24:00 +0000 (09:24 +1000)]
egl: fix gcc warning braces around scalar initializer
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Nicolai Hähnle [Fri, 9 Sep 2016 16:53:22 +0000 (18:53 +0200)]
winsys/radeon: rename nrelocs, crelocs to max_relocs, num_relocs
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Nicolai Hähnle [Fri, 9 Sep 2016 16:50:34 +0000 (18:50 +0200)]
winsys/radeon: don't pre-allocate the relocations array
It's really not necessary. Switch to an exponential resizing strategy.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Nicolai Hähnle [Fri, 9 Sep 2016 13:34:08 +0000 (15:34 +0200)]
winsys/radeon: remove unused radeon_cs_context::priority_usage
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Nicolai Hähnle [Thu, 8 Sep 2016 07:31:53 +0000 (09:31 +0200)]
winsys/amdgpu: remove amdgpu_cs_lookup_buffer
The radeonsi driver doesn't and shouldn't care about the buffer index.
Only the virtual addresses matter.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Nicolai Hähnle [Wed, 7 Sep 2016 16:07:22 +0000 (18:07 +0200)]
winsys/amdgpu: remove unused field domains from amdgpu_cs_buffer
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Nicolai Hähnle [Wed, 7 Sep 2016 10:42:15 +0000 (12:42 +0200)]
winsys/amdgpu: remove initial buffer list allocation
It's really not necessary.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Nicolai Hähnle [Wed, 7 Sep 2016 10:39:30 +0000 (12:39 +0200)]
winsys/amdgpu: extract adding a new buffer list entry into its own function
While at it, try to be a little more robust in the face of memory allocation
failure.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Nicolai Hähnle [Wed, 7 Sep 2016 08:50:14 +0000 (10:50 +0200)]
winsys/amdgpu: use only one fence per BO
The fence that is added to the BO during flush is guaranteed to be
signaled after all the fences that were in the fences array of the BO
before the flush, because those fences are added as dependencies for the
submission (and all this happens atomically under the bo_fence_lock).
Therefore, keeping only the last fence around is sufficient.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Nicolai Hähnle [Tue, 6 Sep 2016 13:12:18 +0000 (15:12 +0200)]
winsys/amdgpu: add do_winsys_deinit function
The idea is to have matching init/deinit functions so that deinit can be
re-used for cleanup in the error path of amdgpu_winsys_create.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Nicolai Hähnle [Tue, 6 Sep 2016 13:10:06 +0000 (15:10 +0200)]
winsys/amdgpu: clean up error paths in amdgpu_winsys_create
No need to call pb_cache_deinit, because the cache hasn't been initialized
at that point.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Nicolai Hähnle [Fri, 9 Sep 2016 10:59:10 +0000 (12:59 +0200)]
gallium/radeon: page alignment for buffers is unnecessary
In some places (e.g. shader program pointers) we require 256 bytes alignment.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Nicolai Hähnle [Tue, 6 Sep 2016 12:50:52 +0000 (14:50 +0200)]
gallium/radeon/winsyses: remove #includes of pb_bufmgr.h
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Topi Pohjolainen [Thu, 8 Sep 2016 05:22:38 +0000 (08:22 +0300)]
i965/rbc: Clarify rational given for shader image resolves
Original commit added documentation explaining lossless compression
case:
commit
56f29911ec9da25c78fbd3d4945d499e65ca4b5a
Author: Topi Pohjolainen <topi.pohjolainen@intel.com>
Date: Tue Feb 2 10:00:41 2016 +0200
i965: Add a flag telling color resolve pass to ignore CCS_E
It, however, easily gives the impression that the sole purpose
of the intel_miptree_resolve_color() is to address lossless
compression. Original intention is to document the lack of
INTEL_MIPTREE_IGNORE_CCS_E flag given for the resolve call.
This patch fixes this along with a typo found spotted further
down.
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Topi Pohjolainen [Mon, 29 Aug 2016 05:57:31 +0000 (08:57 +0300)]
i965/blorp: Use hw generetad primitive copies for layered clears
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Topi Pohjolainen [Mon, 29 Aug 2016 07:16:40 +0000 (10:16 +0300)]
i965/blorp: Sanity check all layers before actual clear
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Topi Pohjolainen [Fri, 26 Aug 2016 15:44:18 +0000 (18:44 +0300)]
intel/blorp: Add plumbing for setting color clear layer count
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Topi Pohjolainen [Fri, 26 Aug 2016 15:37:09 +0000 (18:37 +0300)]
intel/blorp: Allow multiple layers
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Topi Pohjolainen [Fri, 26 Aug 2016 14:37:44 +0000 (17:37 +0300)]
i965/blorp: Instruct vertex fetcher to provide prim instance id
This will indicate target layer (Render Target Array Index) needed
for layered clears.
v2: Use 3DSTATE_VF_SGVS for gen8+
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Topi Pohjolainen [Fri, 26 Aug 2016 06:26:15 +0000 (09:26 +0300)]
i965/rbc: Allocate mcs directly
such as we do for compressed msaa. In case of non-compressed simgle
sampled buffers the allocation of mcs is deferred until there is
actually a clear operation that needs the mcs.
In case of render buffer compression the mcs buffer always needed
and there is no real reason to defer the allocation. By doing it
directly allows to drop quite a bit unnecessary complexity.
Patch leaves brw_predraw_set_aux_buffers() a no-op. Subsequent
patches will re-use it and it seemed cleaner to leave it instead
of removing and re-introducing.
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Topi Pohjolainen [Wed, 31 Aug 2016 08:16:12 +0000 (11:16 +0300)]
isl/gen8+: Allow 1D and 3D auxiliary surfaces
Otherwise once mcs buffer gets allocated without delay for lossless
compression (same as we do for msaa), assert starts to fire in
piglit case: tex3d. The test uses depth of one which is in fact
supported even now.
v2 (Jason): Allow also 1D case as there is nothing in the specs
constraining it either.
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Topi Pohjolainen [Wed, 7 Sep 2016 06:38:11 +0000 (09:38 +0300)]
i965: Add sanity check for non-compressible texture views
v2: Fix missing inline declaration
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Topi Pohjolainen [Wed, 31 Aug 2016 07:08:17 +0000 (10:08 +0300)]
i965/rbc: Consult rb settings for texture surface setup
Once mcs buffer gets allocated without delay for lossless
compression (same as we do for msaa), one gets regression in:
GL45-CTS.texture_barrier_ARB.same-texel-rw
Setting the auxiliary surface for both sampling engine and data
port seems to fix this. I haven't found any hardware documentation
backing this though.
v2 (Jason): Prepare also for the case where surface is sampled with
non-compressible format forcing also rendering without
compression.
v3: Split asserts and decision making.
v4: Detailed comment provided by Jason explaining the need for using
auxiliary buffer for texturing when the same surface is also
used as render target.
Added check for existence of renderbuffer before considering if
underlying miptree matches.
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Topi Pohjolainen [Sun, 4 Sep 2016 08:02:39 +0000 (11:02 +0300)]
i965: Track non-compressible sampling of renderbuffers
v3:
- Actually set the flags when needed instead of falsely
overwriting them (Jason).
- Use more generic name for flag (dropped RENDERBUFFER)
- Consult also shader images
v4:
- Consult only lossless compressd shader images
v5:
- Check the existence of renderbuffer before considering
if it matches the given miptree
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Topi Pohjolainen [Mon, 5 Sep 2016 12:14:11 +0000 (15:14 +0300)]
i965: Replace boolean rb surface state setup argument with flags
And add plumbing to provide it all the way to surface state emitter.
This is not used yet but will be in subsequent patches to carry
additional constraints.
v2 (Jason): Use uint32_t instead of int as the type
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Topi Pohjolainen [Thu, 25 Aug 2016 14:46:09 +0000 (17:46 +0300)]
i965/rbc: Allow integer formats as advertised in isl_format.c
Blorp consults brw_is_color_fast_clear_compatible() to see if any
restrictions apply for fast clear in addition to the capablities
advertised in isl_format.c::format_info[]. On Gen8+ integer formats
are backlisted for plain old fast clear but there is no reason why
lossless compression shouldn't be supported. In fact, lossless
compression of integer formats is already supported for normal
render paths.
This patch prepares for dropping the delayed allocating of the mcs
buffer for lossless compression. Until now the skip of fast clear
also prevented the mcs being allocated and hence the lossless
compression being effectively turned off for integer formats.
Once the mcs buffer is allocated beforehand, the assertion addressed
here would start triggering.
v2: Drop the assert instead of relaxing it (Jason)
Fix typo while at it.
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Alejandro Piñeiro [Sat, 10 Sep 2016 22:07:31 +0000 (00:07 +0200)]
i965: remove unused variable at intel_miptree_create_for_teximage
After commit "i965: Fix calculation of the image height at start level", it is
not needed. This commit removes the "warning: unused variable ‘i’" warning.
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
Thomas Helland [Tue, 16 Aug 2016 20:10:36 +0000 (22:10 +0200)]
mesa/glsl: Move string_to_uint_map into the util folder
This clears the last bits of the usecases of the hash table
located in mesa/program, allowing us to remove it.
V2: Rebase on top of changes to Makefile.sources
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
Thomas Helland [Thu, 18 Aug 2016 19:59:19 +0000 (21:59 +0200)]
glsl: Convert glcpp-parse to the util hash table
And change the include in glcpp.h accordingly.
V2: Whitespace fix
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
Thomas Helland [Tue, 16 Aug 2016 20:10:34 +0000 (22:10 +0200)]
glsl: Convert loop analysis to the util hash table
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
Thomas Helland [Tue, 16 Aug 2016 20:10:33 +0000 (22:10 +0200)]
mesa: Convert symbol table to the util hash table
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
Thomas Helland [Tue, 16 Aug 2016 20:10:32 +0000 (22:10 +0200)]
glsl: Convert varying test to the util hash table
V2: remove now unused ht_count_callback() (Timothy Arceri)
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
Thomas Helland [Tue, 16 Aug 2016 20:10:31 +0000 (22:10 +0200)]
glsl: Convert output read lowering to the util hash table
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
Thomas Helland [Tue, 16 Aug 2016 20:10:30 +0000 (22:10 +0200)]
glsl: Convert interface block lowering to the util hash table
V2: move comment to correct location (Timothy Arceri)
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
Thomas Helland [Tue, 16 Aug 2016 20:10:28 +0000 (22:10 +0200)]
glsl: Convert if lowering to use a set
Also do some minor whitespace cleanups
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
Thomas Helland [Tue, 16 Aug 2016 20:10:27 +0000 (22:10 +0200)]
glsl: Convert linker to the util hash table
We are getting the util hash table through the include in
program/hash_table.h for the moment until we migrate the
string_to_uint_map to a separate file.
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>
Thomas Helland [Tue, 16 Aug 2016 20:10:26 +0000 (22:10 +0200)]
glsl: Convert link_varyings to the util hash table
Signed-off-by: Thomas Helland <thomashelland90@gmail.com>
Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>