Emil Velikov [Mon, 13 Feb 2017 19:23:39 +0000 (19:23 +0000)]
docs/submittingpatches.html: add #backports section
Provide information about merge conflicts resolution and sending
backports.
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Emil Velikov [Mon, 13 Feb 2017 19:23:38 +0000 (19:23 +0000)]
docs/submittingpatches.html: rework the #criteria section
Reword the section to focus on what is allowed, using a more brief, yet
descriptive wording.
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Emil Velikov [Thu, 16 Feb 2017 14:09:41 +0000 (14:09 +0000)]
travis: bring the scons build on par with AppVeyor
Namely, always build with LLVM and run the check target.
Cc: Rhys Kidd <rhyskidd@gmail.com>
Cc: Eric Anholt <eric@anholt.net>
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Ben Crocker [Thu, 19 Jan 2017 01:44:09 +0000 (20:44 -0500)]
gallivm: Reenable PPC VSX (v3)
Reenable the PPC64LE Vector-Scalar Extension for LLVM versions >= 3.8.1,
now that LLVM bug 26775 and its corollary, 25503, are fixed.
Amendment: remove extraneous spaces in macro def & invocations.
We would prefer a runtime check, e.g. via an LLVMQueryString
(analogous to glGetString, eglQueryString) or LLVMGetVersion API,
but no such API exists at this time.
Signed-off-by: Ben Crocker <bcrocker@redhat.com>
[Emil Velikov: remove LLVM_VERSION macro]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Ben Crocker [Fri, 10 Feb 2017 23:08:07 +0000 (18:08 -0500)]
gallivm: Override getHostCPUName() "generic" w/ "pwr8" (v4)
If llvm::sys::getHostCPUName() returns "generic", override
it with "pwr8" (on PPC64LE).
This is a work-around for a bug in LLVM: a table entry for "POWER8NVL"
is missing, resulting in (big-endian) "generic" being returned on
little-endian Power8NVL systems. The result is that code that
attempts to load the least significant 32 bits of a 64-bit quantity in
memory loads the wrong half.
This omission should be fixed in the next version of LLVM (4.0),
but this work-around should be left in place in case some
future version of POWER<n> also ends up unrepresented in LLVM's table.
This workaround fixes failures in the Piglit arb_gpu_shader_fp64 conversion
tests on POWER8NVL processors.
(V4: add similar comment in the code.)
Signed-off-by: Ben Crocker <bcrocker@redhat.com>
Cc: 12.0 13.0 17.0 <mesa-stable@lists.freedesktop.org>
Acked-by: Emil Velikov <emil.velikov@collabora.com>
Ben Crocker [Tue, 14 Feb 2017 21:16:01 +0000 (16:16 -0500)]
gallivm: Improve debug output (V2)
Improve debug output from gallivm_compile_module and
lp_build_create_jit_compiler_for_module, printing the
-mcpu and -mattr options passed to LLC.
V2: enclose MAttrs debug_printf block and llc -mcpu debug_printf
in "if (gallivm_debug & <flags>)..."
Signed-off-by: Ben Crocker <bcrocker@redhat.com>
Cc: 12.0 13.0 17.0 <mesa-stable@lists.freedesktop.org>
Reviewed-by: Roland Scheidegger <sroland@vmware.com> (v2)
[Emil Velikov: rebase]
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Marek Olšák [Mon, 20 Feb 2017 16:54:12 +0000 (17:54 +0100)]
gallium/u_suballoc: update comments
as requested by Brian. Trivial.
Jonathan Gray [Sat, 18 Feb 2017 07:56:58 +0000 (18:56 +1100)]
util/build-id: define ElfW and NT_GNU_BUILD_ID if needed
Define ElfW() and NT_GNU_BUILD_ID if needed as these defines are not
present on at least OpenBSD and FreeBSD. Fixes the build on OpenBSD.
Fixes:
d4fa083e11f ("util: Add utility build-id code.")
Signed-off-by: Jonathan Gray <jsg@jsg.id.au>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Mauro Rossi [Sat, 18 Feb 2017 22:10:24 +0000 (23:10 +0100)]
android: define HAVE_DL_ITERATE_PHDR for build-id code
Required due to d4fa083 "util: Add utility build-id code."
to avoid following build error and warnings:
external/mesa/src/intel/vulkan/anv_device.c:60:32: error: incompatible integer to pointer conversion initializing 'const struct build_id_note *' with an expression of type 'int' [-Werror,-Wint-conversion]
const struct build_id_note *note = build_id_find_nhdr("libvulkan_intel.so");
^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
external/mesa/src/intel/vulkan/anv_device.c:64:19: warning: implicit declaration of function 'build_id_length' is invalid in C99 [-Wimplicit-function-declaration]
unsigned len = build_id_length(note);
^
external/mesa/src/intel/vulkan/anv_device.c:68:4: warning: implicit declaration of function 'build_id_read' is invalid in C99 [-Wimplicit-function-declaration]
build_id_read(note, uuid, VK_UUID_SIZE);
^
3 warnings and 1 error generated.
[ 40% 1438/3588] target C: libmesa_vulkan_common_32 <= external/mesa/src/intel/vulkan/anv_image.c
ninja: build stopped: subcommand failed.
build/core/ninja.mk:148: recipe for target 'ninja_wrapper' failed
make: *** [ninja_wrapper] Error 1
Fixes:
d4fa083e11f ("util: Add utility build-id code.")
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Acked-by: Emil Velikov <emil.velikov@collabora.com>
Mauro Rossi [Sat, 18 Feb 2017 21:23:18 +0000 (22:23 +0100)]
android: glsl: build shader cache sources
Fixes the following building errors:
external/mesa/src/compiler/glsl/linker.cpp:4642: error: undefined reference
to 'shader_cache_read_program_metadata(gl_context*, gl_shader_program*)'
external/mesa/src/mesa/program/ir_to_mesa.cpp:3135: error: undefined reference
to 'shader_cache_write_program_metadata(gl_context*, gl_shader_program*)'
clang++: error: linker command failed with exit code 1
...
external/mesa/src/mesa/program/ir_to_mesa.cpp:3135: error: undefined reference
to 'shader_cache_write_program_metadata(gl_context*, gl_shader_program*)'
external/mesa/src/compiler/glsl/linker.cpp:4642: error: undefined reference
to 'shader_cache_read_program_metadata(gl_context*, gl_shader_program*)'
clang++: error: linker command failed with exit code 1 (use -v to see invocation)
ninja: build stopped: subcommand failed.
build/core/ninja.mk:148: recipe for target 'ninja_wrapper' failed
make: *** [ninja_wrapper] Error 1
Fixes:
9f8dc3bf03e ("utils: build sha1/disk cache only with
Android/Autoconf")
Acked-by: Emil Velikov <emil.velikov@collabora.com>
Mauro Rossi [Wed, 8 Feb 2017 07:14:05 +0000 (08:14 +0100)]
android: radeonsi: fix sid_table.h generated header include path
generated-sources-dir-for macro replaces intermediates-dir-for
and LOCAL_MODULE_CLASS is defined as required by new macro,
in order to avoid the following building error:
external/mesa/src/gallium/drivers/radeonsi/si_debug.c:29:10: fatal error: 'sid_tables.h' file not found
^
1 error generated.
Fixes:
730574c58e8 ("android: ac/debug: move sid_tables.h generation and
IB decode to amd/common")
Acked-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Acked-by: Emil Velikov <emil.velikov@collabora.com>
Emil Velikov [Mon, 20 Feb 2017 11:56:39 +0000 (11:56 +0000)]
docs: add news item and link release notes for 13.0.5
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
Emil Velikov [Mon, 20 Feb 2017 11:54:07 +0000 (11:54 +0000)]
docs: add sha256 checksums for 13.0.5
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
(cherry picked from commit
112e75f51bc7c82bde4722485115c99734e62064)
Emil Velikov [Mon, 20 Feb 2017 11:43:27 +0000 (11:43 +0000)]
docs: add release notes for 13.0.5
Signed-off-by: Emil Velikov <emil.velikov@collabora.com>
(cherry picked from commit
71f3ff57fa67ef72630821f4fa13a17e264d7ccb)
Dave Airlie [Sun, 19 Feb 2017 05:27:47 +0000 (15:27 +1000)]
vulkan/wsi/x11: add support to detect if we can support rendering (v3)
This adds support to radv_GetPhysicalDeviceXlibPresentationSupportKHR
and radv_GetPhysicalDeviceXcbPresentationSupportKHR to check if the
local device file descriptor is compatible with the descriptor
retrieved from the X server via DRI3.
This will stop radv binding to an X server until we have prime
support in place. Hopefully apps use this API before trying
to render things.
v2: drop unneeded function, don't leak memory. (jekstrand)
v3: also check in surface_get_support callback.
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Dave Airlie [Sun, 19 Feb 2017 22:44:58 +0000 (08:44 +1000)]
Revert "radv: detect command buffers that do no work and drop them (v2)"
This just keeps popping up minor problems and regressions we should
revisit in a more sustainable manner later.
This also reverts:
Revert "radv: query cmds should mark a cmd buffer as having draws."
Revert "radv: also fixup event emission to not get culled."
This reverts commit
d1640e79328af4a63c056e3ccab299a3bffbede7.
This reverts commit
8b47b97215af7157bc15676167cab73aa5a61a76.
This reverts commit
b4b19afebeac84ddce3f1f8b7ffe0da6474fe29a.
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Bas Nieuwenhuizen [Sun, 19 Feb 2017 13:00:25 +0000 (14:00 +0100)]
radv: Handle VK_REMAINING_ARRAY_LAYERS in fast clear eliminate.
Signed-off-by: Bas Nieuwenhuizen <basni@google.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Marek Olšák [Fri, 17 Feb 2017 11:55:24 +0000 (12:55 +0100)]
gallium/u_index_modify: don't add PIPE_TRANSFER_UNSYNCHRONIZED unconditionally
It's OK for r300g (because r300g can't write to buffers via the GPU), but
not later hardware. This issue was spotted randomly.
Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Wed, 15 Feb 2017 16:24:38 +0000 (17:24 +0100)]
radeonsi: fix UNSIGNED_BYTE index buffer fallback with non-zero start (v2)
start can only be non-zero with MultiDrawElements, which is unlikely
to occur with UNSIGNED_BYTE indices.
v2: Also fix the util_shorten_ubyte_elts_to_userptr call.
Tested with the new piglit.
Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Dave Airlie [Sun, 19 Feb 2017 06:33:14 +0000 (16:33 +1000)]
radv: handle layered fast clears.
This iterates the fast clear flush across the layers in the
specified range.
It also moves the compute resolve flush into the function
and builds the range in there.
This fixes:
dEQP-VK.geometry.layered.* regressions since fast clears.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Dave Airlie [Sun, 19 Feb 2017 06:33:13 +0000 (16:33 +1000)]
radv: pass subresourceRange by pointer.
This struct is 5 dwords, we should really just pass a pointer
to it.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Dave Airlie [Sun, 19 Feb 2017 06:38:16 +0000 (16:38 +1000)]
radv: fix typo in a2b10g10r10 fast clear calculation.
This fixes:
dEQP-VK.renderpass.formats.a2b10g10r10_unorm_pack32*
regressions.
Fixes:
f22836dbdd radv: Add CPU color packing for VK_FORMAT_A2B10G10R10_UNORM_PACK32.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Bas Nieuwenhuizen [Sun, 19 Feb 2017 02:35:02 +0000 (03:35 +0100)]
radv: Invert ring SGPR check.
I assume this wants to check if all pipelines use the same SGPR for
the rings.
Signed-off-by: Bas Nieuwenhuizen <basni@google.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Acked-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Bas Nieuwenhuizen [Sun, 19 Feb 2017 00:16:19 +0000 (01:16 +0100)]
radv: Clamp framebuffer dimensions to min. attachment dimensions.
Even though the preferred stance is not to fix incorrect applications
via the driver, this prevents some nasty GPU hangs.
Signed-off-by: Bas Nieuwenhuizen <basni@google.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Marek Olšák [Thu, 16 Feb 2017 17:01:53 +0000 (18:01 +0100)]
gallium: remove TGSI_OPCODE_CLAMP
Not used and not widely supported. Use MIN+MAX instead.
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Thu, 16 Feb 2017 21:52:53 +0000 (22:52 +0100)]
ac/llvm: use min+max instead of AMDGPU.clamp on LLVM 5.0
It selects v_med3_f32, which has the same rate & size.
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Thu, 16 Feb 2017 21:41:16 +0000 (22:41 +0100)]
radeonsi: stop using TGSI_OPCODE_CLAMP by moving it amd/common
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Thu, 16 Feb 2017 16:11:30 +0000 (17:11 +0100)]
tgsi/lowering: stop using TGSI_OPCODE_CLAMP
v2: do it correctly
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Thu, 16 Feb 2017 15:59:05 +0000 (16:59 +0100)]
st/mesa: stop using TGSI_OPCODE_CLAMP
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Wed, 15 Feb 2017 19:44:24 +0000 (20:44 +0100)]
radeonsi: use R600_RESOURCE_FLAG_UNMAPPABLE where it's desirable
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Wed, 15 Feb 2017 19:41:01 +0000 (20:41 +0100)]
gallium/radeon: add R600_RESOURCE_FLAG_UNMAPPABLE
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Wed, 15 Feb 2017 19:07:53 +0000 (20:07 +0100)]
gallium/radeon: change r600_aligned_buffer_create to take flags, not bind
All call sites set bind = 0. The next commit will use this.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Wed, 15 Feb 2017 18:50:15 +0000 (19:50 +0100)]
radeonsi: upload constants into VRAM instead of GTT
This lowers lgkm wait cycles by 30% on VI and normal conditions.
The might be a measurable improvement when CE is disabled (radeon)
or under L2 thrashing.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Wed, 15 Feb 2017 17:49:11 +0000 (18:49 +0100)]
gallium/radeon: use TCC line size as alignment in other places
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Wed, 15 Feb 2017 17:32:34 +0000 (18:32 +0100)]
radeonsi: use a clever alignment for index buffer uploads
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Wed, 15 Feb 2017 17:29:34 +0000 (18:29 +0100)]
radeonsi: use a clever alignment for descriptor uploads
Non-VBO descriptors won't be smaller than the cache line, so simply use
the cache line size.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Wed, 15 Feb 2017 17:22:27 +0000 (18:22 +0100)]
radeonsi: use a clever alignment for constant buffer uploads
This results in a very tiny decrease in lgkm wait cycles.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Wed, 15 Feb 2017 17:36:21 +0000 (18:36 +0100)]
radeonsi: move index buffer flushing into a non-upload indexed case
The other codepaths don't need this.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Thu, 16 Feb 2017 10:39:01 +0000 (11:39 +0100)]
radeonsi: use SI_MAX_ATTRIBS where it should be used
for consistency; no change in behavior
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Tue, 14 Feb 2017 21:14:55 +0000 (22:14 +0100)]
radeonsi: sort members of si_shader_key::part
and improve some comments
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Tue, 14 Feb 2017 21:08:32 +0000 (22:08 +0100)]
radeonsi: have separate LS and ES main shader parts in the shader selector
This might reduce the on-demand compilation if the initial VS/LS/ES
determination is wrong.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Tue, 14 Feb 2017 21:06:51 +0000 (22:06 +0100)]
radeonsi: don't compile pure monolithic shaders asynchronously
there is no point, we have to wait anyway.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Fri, 10 Feb 2017 00:15:21 +0000 (01:15 +0100)]
radeonsi: allow unaligned vertex buffer offsets and strides on CIK-VI
So that we can disable u_vbuf for GL core profiles.
This is a v2 of the previous VI-only patch.
It requires SH_MEM_CONFIG.ALIGNMENT_MODE = UNALIGNED on CIK-VI.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Thu, 16 Feb 2017 00:20:32 +0000 (01:20 +0100)]
radeonsi: remove the fix_size3 workaround
not needed with the shader fallback
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Thu, 16 Feb 2017 00:17:48 +0000 (01:17 +0100)]
radeonsi: add a workaround for clamping unaligned RGB 8 & 16-bit vertex loads
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Wed, 15 Feb 2017 23:47:48 +0000 (00:47 +0100)]
radeonsi: make fix_fetch an array of uint8_t
so that we can add 3-component fallbacks.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Wed, 15 Feb 2017 14:07:03 +0000 (15:07 +0100)]
vl: fix a buffer leak in the bicubic filter by using an uploader
there's no error checking, because the previous code didn't do it either.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Thu, 16 Feb 2017 12:29:16 +0000 (13:29 +0100)]
gallium/hud: create files after graphs are created to get final names
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Edmondo Tommasina <edmondo.tommasina@gmail.com>
Marek Olšák [Wed, 15 Feb 2017 19:55:15 +0000 (20:55 +0100)]
gallium/u_suballoc: allow setting pipe_resource::flags
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Wed, 15 Feb 2017 00:47:25 +0000 (01:47 +0100)]
gallium/u_suballoc: use clear_buffer if available
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Wed, 15 Feb 2017 14:04:51 +0000 (15:04 +0100)]
gallium/util: correctly unref a buffer in u_prim_restart
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Wed, 15 Feb 2017 13:53:20 +0000 (14:53 +0100)]
gallium/util: remove unused u_index_modify helpers
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Wed, 15 Feb 2017 13:50:16 +0000 (14:50 +0100)]
gallium/util: remove unused helper util_draw_texquad
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Marek Olšák [Fri, 17 Feb 2017 10:20:00 +0000 (11:20 +0100)]
gallium/docs: remove documentation of non-existent instructions
trivial
Jason Ekstrand [Fri, 17 Feb 2017 22:18:27 +0000 (14:18 -0800)]
anv/TODO: Check off Storage Image Without Format
The code for this landed a few days ago.
Marek Olšák [Fri, 17 Feb 2017 13:26:07 +0000 (14:26 +0100)]
ac/llvm: fix various findMSB bugs
sffbh needs to be suffixed with ".i32"
Reviewed-by: Dave Airlie <airlied@redhat.com>
Jose Maria Casanova Crespo [Fri, 17 Feb 2017 13:06:46 +0000 (14:06 +0100)]
glsl: link error if unsized array not-last in ssbo
If an unsized declared array is not the last in an SSBO
and an implicit size can not be defined on linking time,
the linker should raise an error instead of reaching
an assertion on GL.
This reverts part of commit
3da08e166415a745139c1127040a24e8a45dc553
getting back to the behavior of commit
5b2675093e863a52b610f112884ae12d42513770
The original patch was correct for GLES that should produce
a compile-time error but the linker error is still necessary
in desktop GL.
Fixes the following piglit tests:
tests/spec/arb_shader_storage_buffer_object/non_integral_size_array_member.shader_test
tests/spec/arb_shader_storage_buffer_object/unsized_array_member.shader_test
Reviewed-by: Dave Airlie <airlied@redhat.com>
Reviewed-by: Iago Toral Quiroga <itoral@igalia.com>
Signed-off-by: Jose Maria Casanova Crespo <jmcasanova@igalia.com>
Lionel Landwerlin [Thu, 16 Feb 2017 13:06:07 +0000 (13:06 +0000)]
i965/fs: fix uninitialized memory access
Found while running shader-db under valgrind.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Cc: "13.0 17.0" <mesa-stable@lists.freedesktop.org>
Timothy Arceri [Sat, 30 Apr 2016 05:26:43 +0000 (15:26 +1000)]
glsl: disable on disk shader cache when running as another user
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Alejandro Piñeiro [Sat, 11 Feb 2017 16:21:35 +0000 (17:21 +0100)]
mesa/formatquery: use consistent local function names
Reviewed-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>
Bas Nieuwenhuizen [Thu, 16 Feb 2017 20:23:58 +0000 (21:23 +0100)]
radv: Use different allocator for descriptor set vram.
This one only keeps allocated memory in the list, and list nodes
in the descriptor sets. Thsi doesn't need messing around with
max_sets, and we get automatic merging of free regions.
Signed-off-by: Bas Nieuwenhuizen <basni@google.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Bas Nieuwenhuizen [Thu, 16 Feb 2017 19:52:24 +0000 (20:52 +0100)]
radv: Never try to create more than max_sets descriptor sets.
We only use the freed ones after all free space has been used. If
the app only allocates small descriptor sets, we might go over
max_sets before the memory is full.
Signed-off-by: Bas Nieuwenhuizen <basni@google.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
CC: <mesa-stable@lists.freedesktop.org>
Fixes:
f4e499ec79147f4172f3669ae9dafd941aaeeb65
Samuel Iglesias Gonsálvez [Fri, 10 Feb 2017 13:06:43 +0000 (14:06 +0100)]
i965/fs: fix 32-bit data type to int64 conversion on BSW/BXT
The 32-bit to 64-bit conversions need to have the 32-bit
data source elements aligned to 64-bit but only with doubles as
destination type.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99660
Signed-off-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>
Tested-by: Mark Janes <mark.a.janes@intel.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Timothy Arceri [Fri, 17 Feb 2017 05:27:53 +0000 (16:27 +1100)]
glsl: fix scons builds with shader cache
For now its disabled for scons so wrap glsl cache calls in a
define conditional.
Timothy Arceri [Fri, 17 Feb 2017 04:51:30 +0000 (15:51 +1100)]
util/disk_cache: fix typo in function stub
Jason Ekstrand [Wed, 15 Feb 2017 06:22:51 +0000 (22:22 -0800)]
i965/fs: Remove hand-coded 64-bit packing optimizations
The optimization in unpack_64 is clearly subsumed with the opt_algebraic
optimizations in the previous commit. The pack optimization may not be
quite handled by opt_algebraic but opt_algebraic should get the really
bad cases. Also, it's been broken since it was merged and we've never
noticed so it must not be doing anything.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Wed, 15 Feb 2017 06:22:36 +0000 (22:22 -0800)]
nir/algebraic: Optimize 64bit pack/unpack
This reduces the instruction count in some fp64 and int64 piglit tests
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Wed, 15 Feb 2017 06:20:32 +0000 (22:20 -0800)]
nir: Rename lower_double_pack to lower_64bit_pack
There's nothing "double" about it other than, perhaps, the fact that it
packs two 32-bit values.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Wed, 15 Feb 2017 06:15:16 +0000 (22:15 -0800)]
nir: Combine the int and double [un]pack opcodes
NIR is a typeless IR and the two opcodes, when considered bitwise, do
exactly the same thing. There's no reason to have two versions.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Wed, 15 Feb 2017 07:00:51 +0000 (23:00 -0800)]
i965/fs: Fix the inline nir_op_pack_double optimization
We can only do the optimization if the source *is* SSA.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Cc: "13.0 17.0" <mesa-stable@lists.freedesktop.org>
George Kyriazis [Thu, 16 Feb 2017 02:55:55 +0000 (20:55 -0600)]
swr: remove unneeded extern "C"
the guards have been added to the header files that needed them.
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
George Kyriazis [Thu, 16 Feb 2017 02:54:43 +0000 (20:54 -0600)]
gallium: add extern "C" guards
Added extern "C" __cplusplus guards on headers that did not have them.
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Timothy Arceri [Tue, 7 Feb 2017 01:10:19 +0000 (12:10 +1100)]
util/disk_cache: check cache exists before calling munmap()
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Mon, 6 Feb 2017 22:49:47 +0000 (09:49 +1100)]
util/disk_cache: add support for removing old versions of the cache
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Mon, 6 Feb 2017 01:56:08 +0000 (12:56 +1100)]
util/disk_cache: allow drivers to pass a directory structure
In order to avoid costly fallback recompiles when cache items are
created with an old version of Mesa or for a different gpu on the
same system we want to create directories that look like this:
./{TIMESTAMP}_{LLVM_TIMESTAMP}/{GPU_ID}
Note: The disk cache util will take a single timestamp string, it is
up to the backend to concatenate the llvm string with the mesa string
if applicable.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Mon, 6 Feb 2017 10:23:31 +0000 (21:23 +1100)]
mesa: remove cache creation from _mesa_initialize_context()
We will change the way we create the cache directory in the following
patches.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Thu, 16 Feb 2017 23:16:16 +0000 (10:16 +1100)]
st/mesa/glsl: build string of dri options and use as input to building sha for shaders
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Wed, 8 Jun 2016 00:18:33 +0000 (10:18 +1000)]
glsl: reserve parameter storage on cache restore
Since we know how big the list will be we can allocate the storage
upfront.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Thu, 2 Jun 2016 04:13:26 +0000 (14:13 +1000)]
glsl: don't try to load/store buffer object values in the cache
Also add an assert to catch buffer overflows.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Mon, 20 Jun 2016 01:09:34 +0000 (11:09 +1000)]
glsl: don't reprocess or clear UBOs on cache fallback
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Sat, 23 Apr 2016 12:34:38 +0000 (22:34 +1000)]
glsl: skip more uniform initialisation when doing fallback linking
We already pull these values from the metadata cache so no need to
recreate them.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Wed, 27 Apr 2016 05:41:19 +0000 (15:41 +1000)]
glsl: don't lose uniform values when falling back to full compile
Here we skip the recreation of uniform storage if we are relinking
after a cache miss. This is improtant because uniform values may
have already been set by the application and we don't want to reset
them.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Sun, 20 Nov 2016 06:48:27 +0000 (17:48 +1100)]
glsl: don't reference shader prog data during cache fallback
We already have a reference.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Sun, 20 Nov 2016 01:47:00 +0000 (12:47 +1100)]
mesa/glsl: add cache_fallback flag to gl_shader_program_data
This will allow us to skip certain things when falling back to
a full recompile on a cache miss such as avoiding reinitialising
uniforms.
In this change we use it to avoid reading the program metadata
from the cache and skipping linking during a fallback.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Mon, 13 Feb 2017 00:53:30 +0000 (11:53 +1100)]
glsl: add api and glsl version to hash generation for shaders
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Wed, 13 Apr 2016 00:28:04 +0000 (10:28 +1000)]
glsl: cache uniform values
These may be lowered constant arrays or uniform values that we set before linking
so we need to cache the actual uniform values.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Wed, 13 Apr 2016 05:42:04 +0000 (15:42 +1000)]
glsl: make uniform values helper available for use elsewhere
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Wed, 30 Mar 2016 23:37:27 +0000 (10:37 +1100)]
glsl: cache some more image metadata
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Sat, 26 Mar 2016 02:37:30 +0000 (13:37 +1100)]
glsl: add support for caching atomic buffers
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Wed, 23 Mar 2016 02:50:08 +0000 (13:50 +1100)]
glsl: add shader cache support for buffer blocks
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Tue, 12 Apr 2016 05:10:24 +0000 (15:10 +1000)]
glsl: store subroutine remap table in shader cache
V2: use new helpers to store/restore table entries.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Tue, 22 Mar 2016 10:54:27 +0000 (21:54 +1100)]
glsl: add support for caching subroutines
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Mon, 23 Jan 2017 21:34:23 +0000 (08:34 +1100)]
glsl: add support for caching shaders with xfb qualifiers
For now this disables the shader cache when transform feedback is
enabled via the GL API as we don't currently allow for it when
generating the sha for the shader.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Sat, 19 Mar 2016 03:42:46 +0000 (14:42 +1100)]
glsl: add shader cache support for samplers
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Fri, 18 Mar 2016 10:51:49 +0000 (21:51 +1100)]
glsl: add basic support for resource list to shader cache
This initially adds support for simple uniforms and varyings.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Tue, 12 Apr 2016 07:20:49 +0000 (17:20 +1000)]
glsl: fix uniform remap table cache when explicit locations used
V2: don't store pointers use an enum instead to flag what should be
restored. Also do the work in a helper that we will later use for
the subroutine remap table.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Carl Worth [Thu, 13 Nov 2014 07:57:15 +0000 (23:57 -0800)]
glsl: Serialize three additional hash tables with program metadata
The three additional tables are AttributeBindings, FragDataBindings,
and FragDataIndexBindings.
The first table (AttributeBindings) was identified as missing by
trying to test the shader cache with a program that called
glGetAttribLocation.
Many thanks to Tapani Pälli <tapani.palli@intel.com>, as it was review
of related work that he had done previously that pointed me to the
necessity to also save and restore FragDataBindings and
FragDataIndexBindings.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Sun, 12 Feb 2017 22:34:54 +0000 (09:34 +1100)]
glsl: use correct shader source in case of cache fallback
The scenario is:
glShaderSource
glCompileShader <-- deferred due to cache hit of shader
glShaderSource <-- with new source code
glAttachShader
glLinkProgram <-- no cache hit for program
At this point we need to compile the original source when we
fallback.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Mon, 23 Jan 2017 21:39:13 +0000 (08:39 +1100)]
glsl: make use of on disk shader cache
The hash key for glsl metadata is a hash of the hashes of each GLSL
source string.
This commit uses the put_key/get_key support in the cache put the SHA-1
hash of the source string for each successfully compiled shader into the
cache. This allows for early, optimistic returns from glCompileShader
(if the identical source string had been successfully compiled in the past),
in the hope that the final, linked shader will be found in the cache.
This is based on the intial patch by Carl.
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Timothy Arceri [Thu, 14 Apr 2016 01:08:32 +0000 (11:08 +1000)]
glsl: add initial implementation of shader cache
This uses disk_cache.c to write out a serialization of various
state that's required in order to successfully load and use a
binary written out by a drivers backend, this state is referred to as
"metadata" throughout the implementation.
This initial version is intended to work with all stages beside
compute.
This patch is based on the initial work done by Carl.
V2: extend the file's doxygen comment to cover some of the
design decisions.
V3:
- skip cache for fixed function shaders
- add int64 support
- fix glsl IR program parameter caching/restore and cache the
parameter values which are used by gallium backends.
- use new link status enum
V4:
- add compute program support
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>