Charmaine Lee [Fri, 1 Dec 2017 20:45:30 +0000 (12:45 -0800)]
svga: fix missing format multisample devcap check
In commit
e4048f6cd1, svga_is_dx_format_supported() is supposed to
also check the SVGA3D_DXFMT_MULTISAMPLE bit for multisample
support of a format. Somehow that code is not included in that commit.
This patch fixes it.
Fixes piglit test spec@ext_framebuffer_multisample@formats all_samples.
Reviewed-by: Brian Paul <brianp@vmware.com>
Charmaine Lee [Thu, 30 Nov 2017 16:54:44 +0000 (08:54 -0800)]
svga: fix incorrect multisample support in VGPU9 device
Commit
e4048f6cd1 unintentionally allows multisample support for VGPU9 device.
This patch fixes this regression.
Reviewed-by: Neha Bhende <bhenden@vmware.com>
Charmaine Lee [Wed, 29 Nov 2017 21:27:58 +0000 (13:27 -0800)]
svga: fix the missing devcap for SVGA3D_BC3_UNORM_SRGB
Set the devcap to SVGA3D_DEVCAP_DXFMT_BC3_UNORM_SRGB
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Neha Bhende <bhenden@vmware.com>
Charmaine Lee [Wed, 29 Nov 2017 21:25:30 +0000 (13:25 -0800)]
svga: add a separate function to get dx format capabilities from vgpu10 device
Currently we have one function to get format capabailities and
we convert DX10 devcaps back to DX9. This can be confusing.
Going forward we will have a separate function for dealing with dx formats.
This patch also fixes the depth stencil devcap. Instead of hardcoding
the capabilities for the depth stencil formats, we will inquire the
device for the capabilities. Note: we will still need to explicity set
the SVGA3D_DXFMT_SHADER_SAMPLE bit for SVGA3D_R32_FLOAT_X8X24 and
SVGA3D_R24_UNORM_X8 since this bit is not advertised but supported
by the device.
v2: reapply the patch after svga_is_format_supported is moved to svga_format.c
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Neha Bhende <bhenden@vmware.com>
Charmaine Lee [Wed, 29 Nov 2017 21:09:26 +0000 (13:09 -0800)]
svga: assign a separate function for is_format_supported() for vgpu10 device
This patch adds a new function svga_is_dx_format_supported() to check
for format support in a VGPU10 device.
v2: reapply the patch after svga_is_format_supported is moved to svga_format.c
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Neha Bhende <bhenden@vmware.com>
Brian Paul [Wed, 29 Nov 2017 18:28:10 +0000 (11:28 -0700)]
svga: add some devcap debugging code
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Charmaine Lee [Mon, 27 Nov 2017 21:24:26 +0000 (13:24 -0800)]
svga: fix depth and coverage mask output declaration
Set the component mask to zero for both registers.
Reviewed-by: Brian Paul <brianp@vmware.com>
Charmaine Lee [Fri, 17 Nov 2017 03:46:35 +0000 (19:46 -0800)]
svga: add sample positions for 2 samples
Fixes piglit tests spec@arb_sample_shading@builtin-gl-sample-position 2
spec@arb_texture_multisample@fb-completeness@2
Reviewed-by: Brian Paul <brianp@vmware.com>
Charmaine Lee [Thu, 16 Nov 2017 06:21:22 +0000 (22:21 -0800)]
svga: check sample count devcaps
Check sample count devcaps from the svga device to determine the
supported sample counts.
Reviewed-by: Brian Paul <brianp@vmware.com>
Brian Paul [Tue, 14 Nov 2017 16:36:23 +0000 (09:36 -0700)]
svga: fix 1-element cube map array issue
As with 1D and 2D array textures, if there's only one array element
(one cubemap in this case) we have to issue different shader code.
This fixes a number of Piglit cubemap array tests.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Tue, 14 Nov 2017 16:34:29 +0000 (09:34 -0700)]
svga: simplify array test in svga_init_shader_key_common()
And squash commit a patch to silence a compiler warning (add
default case to the switch statement).
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Charmaine Lee [Mon, 6 Nov 2017 22:13:47 +0000 (14:13 -0800)]
winsys/drm: check for CAPS2/SM41 support if VGPU10 is enabled
No need to check for HW_CAPS2 or SM4_1 support if VGPU10 is not
enabled or is explicitly disabled via the environment variable
SVGA_VGPU10.
Reviewed-by: Deepak Rawat <drawat@vmware.com>
Deepak Rawat [Thu, 2 Nov 2017 00:40:50 +0000 (17:40 -0700)]
winsys/drm: Add support for quality level in surface ioctl
A new argument "quality level" is added in surface define v3 which
represets precision settings for surface. This commit add support
for quality level in DRM_VMW_GB_SURFACE_CREATE_EXT and
DRM_VMW_GB_SURFACE_REF_EXT.
Signed-off-by: Deepak Rawat <drawat@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Charmaine Lee [Mon, 30 Oct 2017 14:30:33 +0000 (07:30 -0700)]
svga: sync svga3d_types.h with upstream changes
Reviewed-by: Brian Paul <brianp@vmware.com>
Charmaine Lee [Fri, 27 Oct 2017 22:41:37 +0000 (15:41 -0700)]
winsys/drm: enable intra_surface_copy if HW_CAP2 is supported
With drm version 2_15, we can inquire for support of HW_CAP2.
If it is supported, we can enable intra_surface_copy support.
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Deepak Rawat <drawat@vmware.com>
Charmaine Lee [Fri, 27 Oct 2017 17:24:46 +0000 (10:24 -0700)]
svga: add git version logging at init time
Before we can log the git version in the host log,
we'll add the git version in the init debug message.
Reviewed-by: Brian Paul <brianp@vmware.com>
Charmaine Lee [Wed, 25 Oct 2017 05:10:20 +0000 (22:10 -0700)]
svga: fix a typo in svga_texture_copy_region()
Trivial.
Charmaine Lee [Tue, 24 Oct 2017 17:57:42 +0000 (10:57 -0700)]
svga: use helper function to do copy region
Use the common helper function svga_texture_copy_region
for copy region command.
Reviewed-by: Brian Paul <brianp@vmware.com>
Charmaine Lee [Thu, 19 Oct 2017 23:03:44 +0000 (16:03 -0700)]
svga: fix cubemap array rendering with backed surface view
This patch fixes the layer index when rendering to a
backed surface view of a cubemap array.
Fixes piglit test fbo-generatemipmap-cubemap array.
Reviewed-by: Brian Paul <brianp@vmware.com>
Charmaine Lee [Mon, 16 Oct 2017 18:31:15 +0000 (11:31 -0700)]
svga: add a helper function to send ResolveCopy command
Reviewed-by: Brian Paul <brianp@vmware.com>
Charmaine Lee [Fri, 13 Oct 2017 22:08:02 +0000 (15:08 -0700)]
svga: sync svga3d header files
This is a squash of what was orginally three commits.
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Neha Bhende <bhenden@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Charmaine Lee [Wed, 11 Oct 2017 21:26:52 +0000 (14:26 -0700)]
svga: add SM4_1 enable debug print
Reviewed-by: Brian Paul <brianp@vmware.com>
Charmaine Lee [Tue, 10 Oct 2017 22:30:21 +0000 (15:30 -0700)]
svga: fix swizzling for texture gather
Texture swizzling for texture gather needs to be done to the selected texels
rather than to the returned vector. This patch has specical cases
for the different swizzles in emit_tg4().
Fixes a lot of piglit texture gather tests.
Reviewed-by: Neha Bhende <bhenden@vmware.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Charmaine Lee [Wed, 4 Oct 2017 23:51:35 +0000 (16:51 -0700)]
svga: fix starting index for system values
Currently, the starting index for system values is assigned to
the next index after the highest index of the tgsi declared input registers.
But the tgsi index might be different from the actual assigned index, hence
this might cause overlap of indices.
With this patch, the shader linker keeps track of the highest index of the
translated input registers, and the next index will be used for the
starting index for system values.
Fixes SHIM errors running arb_copy_image-formats on SM4_1 device.
Reviewed-by: Brian Paul <brianp@vmware.com>
Deepak Rawat [Tue, 19 Sep 2017 22:56:26 +0000 (15:56 -0700)]
winsys/svga: Add support for new surface ioctl, multisample pattern
Kernel driver version 2.15 added new surface ioctl named:
DRM_VMW_GB_SURFACE_CREATE_EXT
DRM_VMW_GB_SURFACE_REF_EXT
The new ioctl has support for 64-bit svga3d_flags if
DRM_VMW_PARAM_SM4_1 is available.
Multisampling surface mob size calculation is added. Also synced the
relevant header update.
svga device modified the surface define command V3 with new parameter
multisampling pattern. Adding support for that in winsys.
Signed-off-by: Deepak Rawat <drawat@vmware.com>
Reviewed-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Sinclair Yeh <syeh@vmware.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Wed, 5 Sep 2018 15:48:35 +0000 (09:48 -0600)]
svga: enable MSAA for SM4_1 device
The SVGA device is deprecating the DX9 MSAA support.
This patch enables MSAA for SM4_1 device by explicitly
setting the SVGA3D_SURFACE_MULTISAMPLE bit.
For SM4_1 device, only 4 samples is supported.
Reviewed-by: Brian Paul <brianp@vmware.com>
Charmaine Lee [Tue, 26 Sep 2017 23:23:58 +0000 (16:23 -0700)]
svga: add sample count to the surface_can_create interface
With this patch, sample count is also taken into account
when determining if a resource can be created.
Reviewed-by: Brian Paul <brianp@vmware.com>
Brian Paul [Wed, 4 Oct 2017 17:21:18 +0000 (11:21 -0600)]
svga: implement support for GL_ARB_texture_query_lod
Just translate the TGSI LODQ intruction to VGPU10 LOD instruction.
All (4) Piglit GL_ARB_texture_query_lod tests pass.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Reviewed-by: Neha Bhende <bhenden@vmware.com>
Neha Bhende [Fri, 25 Aug 2017 12:06:41 +0000 (05:06 -0700)]
svga: Add support for arb_texture_gather
With sm4_1, we can support single channel 2D or CubeMap textures.
This patch exercises this feature.
Tested with piglit
v2: As per Brian's comment
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Tue, 1 Aug 2017 03:12:07 +0000 (21:12 -0600)]
svga: add support for interpolation at sample position
Vs. sampling at the centroid or the fragment center.
Note that this does not fix failures with the Piglit
arb_sample_shading-interpolate-at-sample-position or
arb_sample_shading-ignore-centroid-qualifier.exe tests at this time.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Tue, 1 Aug 2017 03:12:07 +0000 (21:12 -0600)]
svga: clarify sys value -> input register mapping
We translate TGSI system value registers to VGPU10 input registers.
Add a comment and set file = TGSI_FILE_INPUT. That's not stricly
necessary since we map both TGSI_FILE_INPUT and TGSI_FILE_SYSTEM_VALUE
to VGPU10_OPERAND_TYPE_INPUT, but this makes the code a bit more
understandable.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Tue, 1 Aug 2017 03:12:07 +0000 (21:12 -0600)]
svga: add support for FS sample mask output
This, with the previous work for sample position/id query, allows
us to enable per-sample shading for VGPU 10.1.
Note that quite a few Piglit arb_sample_shading tests still do not
pass, but many do.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Tue, 1 Aug 2017 03:12:07 +0000 (21:12 -0600)]
svga: add support for sample id, sample position
Sample ID is just a system value. Sample position must be implemented
with the VGPU10_OPCODE_SAMPLE_POS instruction.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Tue, 1 Aug 2017 03:12:07 +0000 (21:12 -0600)]
svga: implement no-op svga_set_min_samples()
This is part of the per-sample shading feature (PIPE_CAP_SAMPLE_SHADING).
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Charmaine Lee [Wed, 4 Oct 2017 17:05:58 +0000 (11:05 -0600)]
svga: add support for independent blend function per render target
This patch adds support for GL_ARB_draw_buffers_blend extension
for SM4_1 device.
Fixes piglit test fbo-draw-buffers-blend.
This patch is squashed with a subsequent patch which fixed a
regression.
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Neha Bhende <bhenden@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Thu, 25 May 2017 19:46:49 +0000 (13:46 -0600)]
svga: emit shader version as 4.0 or 4.1 depending on device support
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Thu, 25 May 2017 18:45:01 +0000 (12:45 -0600)]
svga: restructure nested if's in emit_src_register()
To make it cleaner for subsequent changes.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Brian Paul [Wed, 24 May 2017 13:38:48 +0000 (07:38 -0600)]
svga: sync VGPU10ShaderTokens.h with upstream changes
This includes new DX 10.1 opcodes and tokens.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
Charmaine Lee [Sat, 6 May 2017 00:00:55 +0000 (17:00 -0700)]
svga: add support for shadow cubemap array
Reviewed-by: Neha Bhende <bhenden@vmware.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Charmaine Lee [Fri, 5 May 2017 21:32:45 +0000 (14:32 -0700)]
svga: add support for rendering to cubemap array
Fixes piglit test arb_texture_cube_map_array-fbo-cubemap-array
Reviewed-by: Neha Bhende <bhenden@vmware.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Charmaine Lee [Fri, 5 May 2017 21:24:53 +0000 (14:24 -0700)]
svga: add support for TXL2 opcode
This patch adds support for cubemap array texture lookup with
explicit LOD.
Fixes piglit test arb_texture_cube_map_array-cubemap-lod
Reviewed-by: Neha Bhende <bhenden@vmware.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Charmaine Lee [Wed, 4 Oct 2017 16:56:30 +0000 (10:56 -0600)]
svga: add support for cubemap array
This patch adds support for cubemap array for SM4_1.
Fixes piglit test arb_texture_cube_map_array-cubemap
Reviewed-by: Neha Bhende <bhenden@vmware.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Brian Paul [Wed, 4 Oct 2017 16:49:42 +0000 (10:49 -0600)]
svga: add have_sm4_1 flag, helper function
Signed-off-by: Brian Paul <brianp@vmware.com>
Marek Olšák [Sat, 1 Sep 2018 05:28:08 +0000 (01:28 -0400)]
gallium/u_inlines: remove the destroy variable in pipe_reference_described
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Marek Olšák [Sat, 1 Sep 2018 05:22:08 +0000 (01:22 -0400)]
gallium/u_inlines: improve pipe_reference_described perf for debug builds
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Marek Olšák [Sat, 1 Sep 2018 05:17:03 +0000 (01:17 -0400)]
gallium/auxiliary: don't dereference counters twice needlessly
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Marek Olšák [Sat, 1 Sep 2018 04:36:14 +0000 (00:36 -0400)]
gallium/u_inlines: normalize naming, use dst & src, style fixes (v2)
v2: update comments
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Tested-by: Dieter Nützel <Dieter@nuetzel-hh.de>
Marek Olšák [Sun, 9 Sep 2018 01:02:18 +0000 (21:02 -0400)]
util: try to fix the Android and MacOS build
Bionic does not have pthread_setaffinity_np.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107869
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Jason Ekstrand [Mon, 10 Sep 2018 17:05:41 +0000 (12:05 -0500)]
anv: Support v3 of VK_EXT_vertex_attribute_divisor
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Jason Ekstrand [Mon, 10 Sep 2018 17:07:51 +0000 (12:07 -0500)]
vulkan: Update the XML and headers to 1.1.84
Acked-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Sergii Romantsov [Fri, 7 Sep 2018 09:43:41 +0000 (12:43 +0300)]
mesa/meson: 32bit xmlconfig linkage
Building of 32bit mesa with meson causes linkage issue:
"undefined reference to `util_get_process_name'"
Fixed by adding link-with mesa_util for xmlconfig primary.
v2: Removed '[]', commit message corrected.
v3: Reverted changes in gbm and glx libraries.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107843
Fixes: 2e1e6511f76370870b5cd "util: extract get_process_name from xmlconfig.c"
Cc: Marek Olšák <marek.olsak@amd.com>
Cc: Dylan Baker <dylan@pnwbakers.com>
Signed-off-by: Sergii Romantsov <sergii.romantsov@globallogic.com>
Reviewed-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
Jose Fonseca [Fri, 7 Sep 2018 12:27:47 +0000 (13:27 +0100)]
Require Visual Studio 2015.
We no longer need or use Visual Studio 2013.
https://ci.appveyor.com/project/jrfonseca/mesa/build/52
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Jose Fonseca [Mon, 10 Sep 2018 09:08:36 +0000 (10:08 +0100)]
util: Make util_context_thread_changed a no-op on Windows.
Spite using thrd_t types, these functions are wed to pthreads, and break
Windows builds, because thrd_current() is not implemented there, as it's
impossible to have an efficient thrd_current() implementation on
Windows.
Trivial.
Erik Faye-Lund [Wed, 5 Sep 2018 13:34:51 +0000 (14:34 +0100)]
virgl: do not map zero-sized resource
When creating textures, we avoid creating backing-store for all
multisampled textures, not just depth buffers.
So we can't try to map them later. That's just going to fail. So
let's take the blit-based code-path that seems to avoid this problem.
This make this piglit test-case no longer crash (although it still
fails):
bin/copyteximage 2D -samples=2 -auto
Signed-off-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Erik Faye-Lund [Wed, 5 Sep 2018 13:17:24 +0000 (14:17 +0100)]
virgl: remove dead code
We don't use the size we calculate in this function, so let's just
drop the calculation
Signed-off-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Erik Faye-Lund [Wed, 5 Sep 2018 12:58:35 +0000 (13:58 +0100)]
virgl: drop needless return-code
We always return TRUE, and we never check the return-value. Let's
just drop the return value instead.
Signed-off-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Erik Faye-Lund [Wed, 5 Sep 2018 11:50:18 +0000 (12:50 +0100)]
virgl: free trans on map-error
When we fail to map memory, we should also free trans to avoid
leaking memory.
Noticed while reading code.
Signed-off-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
Chris Wilson [Fri, 7 Sep 2018 13:32:05 +0000 (14:32 +0100)]
i965: Bump aperture tracking to u64
As a prelude to handling large address spaces, first allow ourselves the
luxury of handling the full 4G.
Reported-by: Andrey Simiklit <asimiklit.work@gmail.com>
Cc: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Mathias Fröhlich [Thu, 6 Sep 2018 14:13:42 +0000 (16:13 +0200)]
etnaviv: Reduce max offset to available hardware bits.
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Signed-off-by: Mathias Fröhlich <Mathias.Froehlich@web.de>
Mathias Fröhlich [Thu, 6 Sep 2018 14:13:42 +0000 (16:13 +0200)]
gallium: New cap PIPE_CAP_MAX_VERTEX_ELEMENT_SRC_OFFSET.
Introduce a new capability for the maximum value of
pipe_vertex_element::src_offset. Initially just every driver
backend returns the value previously set from _mesa_init_constants.
So this shall end up in no functional change.
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Signed-off-by: Mathias Fröhlich <Mathias.Froehlich@web.de>
Dave Airlie [Tue, 28 Aug 2018 22:32:29 +0000 (08:32 +1000)]
virgl: don't send a shader create with no data. (v2)
This fixes the situation where we'd send a shader with just the
header and no data.
piglit/glsl-max-varyings test was causing this to happen, and
the renderer fix was breaking it.
v2: drop fprintf
Fixes: a8987b88ff1d "virgl: add driver for virtio-gpu 3D (v2)"
Reviewed-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Timothy Arceri [Sun, 2 Sep 2018 10:27:40 +0000 (20:27 +1000)]
mesa: enable ARB_vertex_buffer_object in core profile
This extension is required by "Wolfenstein: The Old Blood"
and is exposed in core in the Nvidia binary driver.
All the functions are just alias of the core functions so
there should be nothing more to do.
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Marek Olšák [Sat, 1 Sep 2018 00:58:10 +0000 (20:58 -0400)]
st/mesa: throttle texture uploads if their memory usage goes beyond a limit
This prevents radeonsi from running out of memory. It also increases
texture upload performance by being nice to the kernel memory manager.
Marek Olšák [Sat, 1 Sep 2018 00:52:29 +0000 (20:52 -0400)]
gallium: add PIPE_CAP_MAX_TEXTURE_UPLOAD_MEMORY_BUDGET
Andres Gomez [Fri, 7 Sep 2018 21:40:43 +0000 (00:40 +0300)]
docs: update calendar, add news item and link release notes for 18.2.0
Signed-off-by: Andres Gomez <agomez@igalia.com>
Andres Gomez [Fri, 7 Sep 2018 21:26:57 +0000 (00:26 +0300)]
docs: add sha256 checksums for 18.2.0
Signed-off-by: Andres Gomez <agomez@igalia.com>
(cherry picked from commit
cb1ddf48e223231fc4e2cfdc92fbcdaec673929d)
Andres Gomez [Fri, 7 Sep 2018 21:10:22 +0000 (00:10 +0300)]
docs: update 18.2.0 release notes
Signed-off-by: Andres Gomez <agomez@igalia.com>
(cherry picked from commit
7378180e7aa652f3f95e4b953a552dcaf8fb8408)
Marek Olšák [Fri, 7 Sep 2018 20:42:00 +0000 (16:42 -0400)]
Revert "gallium/os_thread: simplify helper pipe_current_thread_get_time_nano"
This reverts commit
6d477bc5460eec14c6a0d047a0384c9ce5c7609b.
It fixes the Windows build hopefully.
Jason Ekstrand [Wed, 22 Aug 2018 01:43:57 +0000 (20:43 -0500)]
anv: Clamp scissors to the framebuffer boundary
The Vulkan 1.1.81 spec says:
"It is legal for offset.x + extent.width or offset.y + extent.height
to exceed the dimensions of the framebuffer - the scissor test still
applies as defined above. Rasterization does not produce fragments
outside of the framebuffer, so such fragments never have the scissor
test performed on them."
Elsewhere, the Vulkan 1.1.81 spec says:
"The application must ensure (using scissor if necessary) that all
rendering is contained within the render area, otherwise the pixels
outside of the render area become undefined and shader side effects
may occur for fragments outside the render area. The render area
must be contained within the framebuffer dimensions."
Unfortunately, there's some room for interpretation here as to what the
consequences are of having the render area set to exactly the
framebuffer dimensions and having a scissor that is larger than the
framebuffer. Given that GL and other APIs provide automatic clipping to
the framebuffer, it makes sense that applications would assume that
Vulkan does this as well. It costs us very little to play it safe and
just clamp client-provided scissors to the framebuffer dimensions.
Fortunately, the user is required to provide us with at least one
scissor so we don't need to handle the case where they don't.
Fixes: fb2a5ceb3264 "anv: Emit DRAWING_RECTANGLE once at driver..."
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Wed, 22 Aug 2018 01:40:50 +0000 (20:40 -0500)]
anv: Disable the vertex cache when tessellating on SKL GT4
I have no idea if I'm correct about what's going wrong or if this is the
correct fix. However, in my multiple weeks of banging my head on this
hang, a VUE reference counting bug seems to match all the symptoms and
it definitely fixes the hang.
Cc: mesa-stable@lists.freedesktop.org
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107280
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Tue, 21 Aug 2018 22:52:54 +0000 (17:52 -0500)]
anv: Implement a VF cache invalidate workaround
Known to fix nothing whatsoever but it's in the docs.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Jason Ekstrand [Tue, 21 Aug 2018 22:14:38 +0000 (17:14 -0500)]
anv: Re-emit vertex buffers when the pipeline changes
Some of the bits of VERTEX_BUFFER_STATE such as access type, instance
data step rate, and pitch come from the pipeline.
Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Marek Olšák [Thu, 6 Sep 2018 03:13:56 +0000 (23:13 -0400)]
radeonsi: pin the winsys thread to the requested L3 cache (v2)
v2: rebase
Reviewed-by: Brian Paul <brianp@vmware.com>
Marek Olšák [Thu, 6 Sep 2018 03:12:27 +0000 (23:12 -0400)]
gallium/u_threaded: implement set_context_param for thread pinning (v2)
v2: - use set_context_param
- set set_context_param even if the driver doesn't implement it
Reviewed-by: Brian Paul <brianp@vmware.com>
Marek Olšák [Thu, 6 Sep 2018 03:10:57 +0000 (23:10 -0400)]
st/mesa: pin driver threads to a specific L3 cache on AMD Zen (v2)
v2: use set_context_param
Reviewed-by: Brian Paul <brianp@vmware.com>
Marek Olšák [Thu, 6 Sep 2018 02:57:19 +0000 (22:57 -0400)]
gallium: add pipe_context::set_context_param for tuning perf on AMD Zen (v2)
State trackers will not use the new param directly, but will instead use
a helper in MakeCurrent that does the right thing.
v2: rework the interface
Reviewed-by: Brian Paul <brianp@vmware.com>
Marek Olšák [Thu, 6 Sep 2018 03:00:07 +0000 (23:00 -0400)]
gallium/os_thread: simplify helper pipe_current_thread_get_time_nano
Reviewed-by: Brian Paul <brianp@vmware.com>
Marek Olšák [Thu, 6 Sep 2018 01:30:44 +0000 (21:30 -0400)]
gallium/u_cpu_detect: get the number of cores per L3 cache for AMD Zen
Reviewed-by: Brian Paul <brianp@vmware.com>
Marek Olšák [Thu, 6 Sep 2018 00:11:34 +0000 (20:11 -0400)]
gallium/u_cpu_detect: fix parsing the CPU family
According to:
https://support.amd.com/TechDocs/54945_PPR_Family_17h_Models_00h-0Fh.pdf
Also Intel:
https://www.microbe.cz/docs/CPUID.pdf
Reviewed-by: Brian Paul <brianp@vmware.com>
Marek Olšák [Wed, 5 Sep 2018 23:15:56 +0000 (19:15 -0400)]
gallium/u_cpu_detect: fix a race condition on initialization
Reviewed-by: Brian Paul <brianp@vmware.com>
Dylan Baker [Tue, 21 Aug 2018 16:46:46 +0000 (09:46 -0700)]
Replace uses of _mesa_bitcount with util_bitcount
and _mesa_bitcount_64 with util_bitcount_64. This fixes a build problem
in nir for platforms that don't have popcount or popcountll, such as
32bit msvc.
v2: - Fix additional uses of _mesa_bitcount added after this was
originally written
Acked-by: Eric Engestrom <eric.engestrom@intel.com> (v1)
Acked-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Dylan Baker [Fri, 17 Aug 2018 00:20:38 +0000 (17:20 -0700)]
move u_math to src/util
Currently we have two sets of functions for bit counts, one in gallium
and one in core mesa. The ones in core mesa are header only in many
cases, since they reduce to "#define _mesa_bitcount popcount", but they
provide a fallback implementation. This is important because 32bit msvc
doesn't have popcountll, just popcount; so when nir (for example)
includes the core mesa header it doesn't (and shouldn't) link with core
mesa. To fix this we'll promote the version out of gallium util, then
replace the core mesa uses with the util version, since nir (and other
non-core mesa users) can and do link with mesautils.
Acked-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Dylan Baker [Fri, 7 Sep 2018 17:19:33 +0000 (10:19 -0700)]
docs: update calendar, add news item and link release notes for X.Y.Z
Signed-off-by: Dylan Baker <dylan@pnwbakers.com>
Dylan Baker [Fri, 7 Sep 2018 17:10:13 +0000 (10:10 -0700)]
docs/relnotes: Add sha256 sums for mesa 18.1.8
Dylan Baker [Fri, 7 Sep 2018 15:27:26 +0000 (08:27 -0700)]
docs: Add release notes for 18.1.8
Jason Ekstrand [Fri, 7 Sep 2018 05:40:54 +0000 (00:40 -0500)]
i965: Workaround the gen9 hw astc5x5 sampler bug
gen9 hardware has a bug in the sampler cache that can cause GPU hangs
whenever an texture with aux compression enabled is in the sampler cache
together with an ASTC5x5 texture. Because we can't control what the
client binds at any given time, we have two options: resolve the CCS or
decompresss the ASTC. Doing a CCS or HiZ resolve is far less drastic
and will likely have a smaller performance impact.
Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Kristian H. Kristensen <hoegsberg@google.com>
Tested-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Eric Anholt [Fri, 24 Aug 2018 22:45:37 +0000 (15:45 -0700)]
v3d: Fix setup of the VCM cache size.
There were two bugs working together to make things mostly work: I wasn't
dividing the VPM output size available by the size of a batch (vertex),
but I also had the size of the VPM reduced by a factor of 8.
Fixes dEQP-GLES3.functional.vertex_array_objects.all_attributes and it
seems also my intermittent varying failures.
Fixes: 1561e4984eb0 ("v3d: Emit the VCM_CACHE_SIZE packet.")
Eric Anholt [Thu, 6 Sep 2018 20:37:37 +0000 (13:37 -0700)]
v3d: Fix SRC_ALPHA_SATURATE blending for RTs without alpha.
Fixes
dEQP-GLES3.functional.fragment_ops.blend.default_framebuffer.rgb_func_alpha_func.dst.src_alpha_saturate_src_alpha_saturate
and friends with --deqp-egl-config-name=rgb565d0s0
Cc: "18.2" <mesa-stable@lists.freedesktop.org>
Lionel Landwerlin [Fri, 7 Sep 2018 10:55:45 +0000 (11:55 +0100)]
intel/genxml: turn SLM Enable bit into boolean
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Sergii Romantsov [Thu, 6 Sep 2018 09:59:47 +0000 (12:59 +0300)]
i965/tools: 32bit compilation with meson
Building of 32bit mesa with meson causes issue:
"implicit declaration of function ‘__builtin_ia32_clflush’".
Fixed by adding msse2 compilation flag.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107843
Fixes: 314879f7fec0 (i965: Fix asynchronous mappings on !LLC platforms.)
Signed-off-by: Sergii Romantsov <sergii.romantsov@globallogic.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Sergii Romantsov [Thu, 6 Sep 2018 08:07:45 +0000 (11:07 +0300)]
intel: compiler option msse2 and mstackrealign
Seems in case of 32-bit library, usage of msse2 makes
some stack corruption or incorrect instructions.
Usage with mstackrealign fixes that case.
v2: Fixed meson.
v3: Definition of c_sse2_args moved on the top (L.Landwerlin).
Added mstackrealign for Android's mks where msee4.1 is used.
v4: Added for Vulkan also.
v5: Commit message correction.
CC: <mesa-stable@lists.freedesktop.org>
Fixes: 6b05c080f202 (i965: Compile with -msse3)
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107779
Signed-off-by: Sergii Romantsov <sergii.romantsov@globallogic.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Rob Clark [Fri, 7 Sep 2018 11:41:43 +0000 (07:41 -0400)]
freedreno: fix rast->depth_cleap_near/far
Fixes: daa19363def gallium: split depth_clip into depth_clip_near & depth_clip_far
Signed-off-by: Rob Clark <robdclark@gmail.com>
Marek Olšák [Wed, 22 Aug 2018 02:00:11 +0000 (22:00 -0400)]
gallium: enable GL_AMD_depth_clamp_separate on r600, radeonsi
Marek Olšák [Wed, 22 Aug 2018 01:59:23 +0000 (21:59 -0400)]
gallium: split depth_clip into depth_clip_near & depth_clip_far
for AMD_depth_clamp_separate.
Jason Ekstrand [Tue, 4 Sep 2018 18:58:01 +0000 (13:58 -0500)]
anv/pipeline: Only consider double elements which actually exist
The brw_vs_prog_data::double_inputs_read field comes directly from
shader_info::double_inputs which may contain inputs which are not
actually read. Instead of using it directly, AND it with inputs_read
which is only things which are read. Otherwise, we may end up
subtracting too many elements when computing elem_count.
Cc: mesa-stable@lists.freedesktop.org
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=103241
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Jason Ekstrand [Fri, 31 Aug 2018 12:35:17 +0000 (07:35 -0500)]
nir: Drop the vs_inputs_dual_locations option
It was very inconsistently handled; the only things that made use of it
were glsl_to_nir, glspirv, and nir_gather_info. In particular,
nir_lower_io completely ignored it so anyone using nir_lower_io on
64-bit vertex attributes was going to be in for a shock. Also, as of
the previous commit, it's set by every driver that supports 64-bit
vertex attributes. There's no longer any reason to have it be an option
so let's just delete it.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Jason Ekstrand [Fri, 31 Aug 2018 12:55:07 +0000 (07:55 -0500)]
radeonsi/nir: Set vs_inputs_dual_locations and let NIR do the remap
We were going out of our way to disable dual-location re-mapping in NIR
only to then do the remapping in st_glsl_to_nir.cpp. Presumably, this
was so that double_inputs would be correct for the core state tracker.
However, now that we've it to gl_program::DualSlotInputs which is
unaffected by NIR lowering, we can let NIR lower things for us. The one
tricky bit here is that we have to remap the inputs_read bitfield back
to the single-slot convention for the gallium state tracker to use.
Since radeonsi is the only NIR-capable gallium driver that also supports
GL_ARB_vertex_attrib_64bit, we only have to worry about radeonsi when
making core gallium state tracker changes.
Acked-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Jason Ekstrand [Thu, 30 Aug 2018 20:02:25 +0000 (15:02 -0500)]
compiler: Move double_inputs to gl_program::DualSlotInputs
Previously, we had two field in shader_info: double_inputs_read and
double_inputs. Presumably, the one was for all double inputs that are
read and the other is all that exist. However, because nir_gather_info
regenerates these two values, there is a possibility, if a variable gets
deleted, that the value of double_inputs could change over time. This
is a problem because double_inputs is used to remap the input locations
to a two-slot-per-dvec3/4 scheme for i965. If that mapping were to
change between glsl_to_nir and back-end state setup, we would fall over
when trying to map the NIR outputs back onto the GL location space.
This commit changes the way slot re-mapping works. Instead of the
double_inputs field in shader_info, it adds a DualSlotInputs bitfield to
gl_program. By having it in gl_program, we more easily guarantee that
NIR passes won't touch it after it's been set. It also makes more sense
to put it in a GL data structure since it's really a mapping from GL
slots to back-end and/or NIR slots and not really a NIR shader thing.
Tested-by: Alejandro Piñeiro <apinheiro@igalia.com> (ARB_gl_spirv tests)
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Marek Olšák [Thu, 6 Sep 2018 04:09:27 +0000 (00:09 -0400)]
gallium: add PIPE_CAP_RASTERIZER_SUBPIXEL_BITS
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Eric Engestrom [Tue, 17 Jul 2018 14:12:32 +0000 (15:12 +0100)]
meson: disable asserts by default on release builds
By the time Mesa 18.3 comes out (probably December '18), Meson 0.45 will
be 9 months old (March '18), so I think this is reasonable.
(btw, the currently-required Meson 0.44.1 was released less than 12 days
before 0.45, so we're really not bumping by much.)
Currently, the Meson versions in the major distributions are:
Arch: ships 0.47.2
CentOS: 7 ships 0.47.1
Debian: stable ships 0.37.1, so it hasn't been usable in a long time.
everything more recent ships 0.47.2
Fedora: 28 ships 0.45.1
FreeBSD: ships 0.46.1 (ports)
Gentoo: ships 0.46.1
OpenSUSE: 15 ships 0.46
Ubuntu: 18.04 ships 0.45.1
Signed-off-by: Eric Engestrom <eric.engestrom@intel.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>