Commit Graph

147067 Commits

Author SHA1 Message Date
Caio Oliveira 8fc6a11f0e intel/blorp: Add option to emit packets that disable Mesh
If a driver doesn't support Mesh, don't emit anything.

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13660>
2021-11-04 14:41:06 -07:00
Caio Oliveira ecba8178bd intel/dev: Add an intel_device_info::has_mesh_shading bit
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13660>
2021-11-04 14:41:06 -07:00
Marcin Ślusarz bba26939b1 intel/decoder: Dump Task/Mesh shaders
Signed-off-by: Marcin Ślusarz <marcin.slusarz@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Acked-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13657>
2021-11-04 21:01:13 +00:00
Caio Oliveira 3567d47f3e intel/genxml: Inline the BODY structs into the instructions
Follows the convention used in other instructions.

Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Acked-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13657>
2021-11-04 21:01:13 +00:00
Caio Oliveira 3fe2e862b5 intel/genxml: Add Mesh Shading structures
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Acked-by: Jordan Justen <jordan.l.justen@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13657>
2021-11-04 21:01:13 +00:00
Jesse Natalie b34fed64fa u_prim_restart: Fix index scanning with start offset
Reviewed-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13681>
2021-11-04 20:39:25 +00:00
Mike Blumenkrantz bc345281ab aux/primconvert: handle singular incomplete restarts
if no restart indices are found, this draw must be discarded to avoid
crashing later on

Fixes: 583070748c ("util/primconvert: handle rewriting of prim-restart draws with unsupported primtype")

Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13630>
2021-11-04 20:12:32 +00:00
Emma Anholt 1e869e3fb4 freedreno/a5xx+: Fix missing LA formats.
GL_ARB_texture_buffer_object uses these formats, and we expose it.  Since
we didn't have the formats in the table, we we were using bad HW
texture/color formats for them.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13666>
2021-11-04 19:07:54 +00:00
Emma Anholt 0e4fcda7e0 freedreno/a6xx: Don't try to generate mipmaps for SNORM with our blitter.
Since we're casting to unorm, the linear filtering will give bad results.

Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13666>
2021-11-04 19:07:54 +00:00
Jason Ekstrand 953a4ca6fe intel: Add has_bit6_swizzle to devinfo
There's no good reason to have this rather complex check in three
drivers.

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13636>
2021-11-04 18:51:04 +00:00
Marek Olšák a0dc303b45 vbo: utilize structure padding to optimize indirection cold->prims[0].begin
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13589>
2021-11-04 18:24:15 +00:00
Marek Olšák 3f997bccc6 radeonsi: increase tc_max_cpu_storage_size
Viewperf benefits. The number is only slightly above the size we need.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13603>
2021-11-04 17:36:26 +00:00
Marek Olšák 74adf22a0a radeonsi: fix a typo preventing a fast depth-stencil clear
Fixes: 9defe8aca9 - radeonsi: implement fast Z/S clears using clear_buffer on HTILE

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13603>
2021-11-04 17:36:26 +00:00
Marek Olšák c0f723ce2b radeonsi: allow and finish TC-compatible MSAA HTILE
This improves perf for Catia by 4%.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13603>
2021-11-04 17:36:26 +00:00
Marek Olšák 3baeaac64b radeonsi: rename stencil_cleared_level_mask -> stencil_cleared_level_mask_once
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13603>
2021-11-04 17:36:26 +00:00
Marek Olšák b1b491cdbb radeonsi: add a faster clear path for glClearTexImage
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13603>
2021-11-04 17:36:26 +00:00
Marek Olšák 5d3aea49b8 radeonsi: fix 2 issues with depth_cleared_level_mask
- Unset depth_cleared_level_mask for non-clear blits. Set the flag after
  the clear, so that we don't have to check blitter_running.
- Set depth_cleared_level_mask only when we set depth_clear_value.

Fixes: ff8a930cf7 - radeonsi: add _once suffix to depth_cleared_level_mask

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13603>
2021-11-04 17:36:26 +00:00
Tapani Pälli 4885e63a6d vulkan/wsi: implement missing wsi_register_device_event
These changes implement vkRegisterDeviceEventEXT and detection of
monitor hotplug. Wsi launches a thread that listens to udev events and
signals the appropriate device fences when hotplug hapens.

v2: use wsi fences instead of syncobj api (Jason Ekstrand)
v3: refactor + cleanups, create thread on demand (Samuel Pitoiset)
v4: bring back syncobj support from initial version for radv
v5: make libudev dependency optional, check for poll errors (Simon Ser)
v6: change matching mechanism to use udev device node instead of path
v7: remove the matching mechanism
v8: fix a race with thread creation + use single mutex + other cleanups
    (Jason Ekstrand)

Fixes:
   dEQP-VK.wsi.display_control.register_device_event

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12305>
2021-11-04 16:57:29 +00:00
Tapani Pälli 9f6764953b anv: setup syncobj fd via wsi_device_setup_syncobj_fd
Patch moves initialization of variable so that we have fd when calling
wsi initialization.

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12305>
2021-11-04 16:57:29 +00:00
Tapani Pälli 01fb24d50e radv: setup syncobj fd via wsi_device_setup_syncobj_fd
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12305>
2021-11-04 16:57:29 +00:00
Tapani Pälli 73f21ea2e1 vulkan/wsi: provide api for drivers to setup syncobj fd
Drivers that import sync_fd to the wsi fences can use this to set file
descriptor for syncobj related calls. This fixes permission errors when
registering display/device events and importing sync_fd from driver
side.

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Jason Ekstrand <jason@jlekstrand.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12305>
2021-11-04 16:57:29 +00:00
Mike Blumenkrantz 92215d8da8 zink: add khr46 to ci
this blocks out all the very long tests and marks failures as needed
to improve the coverage of ci

Acked-by: Daniel Stone <daniels@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13656>
2021-11-04 16:42:04 +00:00
Mike Blumenkrantz f5f2426ffd zink: remove lazy ci job
the push descriptor coverage for lavapipe should be okay in ci now, and
that was the point of adding this job

Acked-by: Daniel Stone <daniels@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13656>
2021-11-04 16:42:04 +00:00
Joshua Ashton 7d64f0dd16 nvc0: Fix uninitialized width/height/depth warning.
This can happen if view->resource is false.

Fixes a warning in GCC 9+ that's been bugging me for a very long time when building Mesa.

Signed-off-by: Joshua Ashton <joshua@froggi.es>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12955>
2021-11-04 15:31:09 +00:00
Marek Olšák 8b287c3f92 gallium/u_blitter: work around broken sample shading in llvmpipe and zink
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13602>
2021-11-04 15:06:09 +00:00
Marek Olšák eb34716c1f gallium/u_blitter: do MSAA copies in 1 pass using sample shading
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13602>
2021-11-04 15:06:09 +00:00
Marek Olšák 6d483fed85 gallium/u_blitter: disable sample shading for all blits
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13602>
2021-11-04 15:06:09 +00:00
Marek Olšák 7ce3f8e639 gallium/util: fix util_can_blit_via_copy_region with unbound render condition
It returned false when a render condition was not bound, but it should
have returned true.

The bool stuff is random and incomplete, but that's life.

Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13602>
2021-11-04 15:06:09 +00:00
Mike Blumenkrantz 5d1b81d8ac zink: clamp PIPE_SHADER_CAP_MAX_INPUTS for xfb
vertex shader stages that can produce xfb must have
their input size clamped to the compiler define MAX_VARYING
to successfully be able to export an xfb output for each input

fixes KHR-GL46.geometry_shader.limits.max_input_components

cc: mesa-stable

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13632>
2021-11-04 14:51:10 +00:00
Mike Blumenkrantz 46e167028d zink: do a better job conserving locations for packed xfb outputs
if an entire vec4 is exported to xfb, mark it as an explicit xfb buffer
whenever possible to avoid blowing out the location limit

Reviewed-by: Dave Airlie <airlied@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13632>
2021-11-04 14:51:10 +00:00
Pierre-Eric Pelloux-Prayer ff981a7e79 drirc: add options for BETA CAE Ansa application.
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13364>
2021-11-04 14:16:55 +00:00
Pierre-Eric Pelloux-Prayer f5dc334b6d drirc: add mesa_extension_override option
This allows specific per-application override.
The existing MESA_EXTENSION_OVERRIDE env variable is kept.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13364>
2021-11-04 14:16:55 +00:00
Pierre-Eric Pelloux-Prayer 50c983402e mesa/init: replace call_once with manual implementation
This will be useful to add parameters to one_time_init().

_MTX_INITIALIZER_NP and Windows don't play nice together,
so I had to keep a call_once() to initialize the mutex.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13364>
2021-11-04 14:16:55 +00:00
Pierre-Eric Pelloux-Prayer 596597f4a4 mesa: don't use dummy_true for some MESA extensions
Otherwise we can't use MESA_EXTENSION_OVERRIDE, or rather: disabling one
extension using dummy_true will disable all others.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13364>
2021-11-04 14:16:55 +00:00
Pierre-Eric Pelloux-Prayer 8a03e25977 mesa: print a warning when an extension can't be disabled
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13364>
2021-11-04 14:16:55 +00:00
Iago Toral Quiroga aa5a0e1dad broadcom/compiler: copy packing when converting add to mul
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13675>
2021-11-04 13:57:39 +00:00
Timur Kristóf b59614619b radv: Use MESA_VULKAN_SHADER_STAGES to make room for mesh/task.
Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13440>
2021-11-04 13:32:07 +00:00
Pierre-Eric Pelloux-Prayer cd6e9ad36a llvmpipe: add missing NIR alu-op handling
nir_op_bcsel implemented based on ac_nir_to_llvm.c emit_bcsel function.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13246>
2021-11-04 13:59:00 +01:00
Pierre-Eric Pelloux-Prayer a21c0f531e mesa: enable force_direct_glx_context for DiscoveryStudio2020
This app uses Qt4.8 which uses an indirect context to save a GL scene to
a pixmap.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13246>
2021-11-04 13:59:00 +01:00
Pierre-Eric Pelloux-Prayer fc3ef76eec glx/drirc: add a force_direct_glx_context option
Some applications may request an indirect context but this feature is
disabled by default on Xorg and thus context creation will fail.

This commit adds a drirc setting to force the creation of direct glx
context, regardless of what the app is requesting.

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13246>
2021-11-04 13:59:00 +01:00
Pierre-Eric Pelloux-Prayer 9b09655a58 vbo/dlist: free copied.buffer if no vertices were copied
Other parts of the code asserts that copied.buffer is NULL if there are
no vertices to copy.

Cc: mesa-stable
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13246>
2021-11-04 13:59:00 +01:00
Marek Olšák 118f2952b6 driconf: set vblank_mode=0 for viewperf2020
It doesn't disable vsync. Reported internally.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13612>
2021-11-04 11:04:35 +00:00
Pierre-Eric Pelloux-Prayer dbf602a6b3 ac/surface: don't validate DCC settings if DCC isn't possible
Cc: mesa-stable
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13550>
2021-11-04 09:38:46 +01:00
Pierre-Eric Pelloux-Prayer bc6d22b920 radeonsi: fix ps_uses_fbfetch value
si_update_ps_colorbuf0_slot used blitter_running as a way to detect
recursive calls.
Unfortunately this catch too many cases; for instance a backtrace
like:
  #0 si_update_ps_colorbuf0_slot
  #1 si_set_framebuffer_state
  #2 do_blits
  [...]
  #5 si_blit
  #6 si_copy_region_with_blit

Would end-up not updating ps_uses_fbfetch; so if the new fb_state is
something like:
  cbufs = {0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0, 0x0}, zsbuf = 0x55b8987545e0}

We can have ps_uses_fbfetch=true but cbufs[0] = NULL, which causes a
crash later in si_ps_key_update_framebuffer.

This commit fixes intermittent crashes in KHR-GL46.stencil_texturing.functional.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13550>
2021-11-04 09:38:42 +01:00
Pierre-Eric Pelloux-Prayer d86d602ed0 radeonsi/sdma: fix bogus assert
src can use dcc even for non sdma v5 variants because si_decompress_dcc
is called in si_sdma_copy_image.

Fixes: 46c95047bd ("radeonsi: implement si_sdma_copy_image for gfx7+")
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13550>
2021-11-04 09:38:41 +01:00
Pierre-Eric Pelloux-Prayer 84d4bda8e5 ac/surface: use a less strict condition in is_dcc_supported_by_L2
While Mesa chooses to always use independent_128B_blocks, other drivers
can make different choices.

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13550>
2021-11-04 09:38:27 +01:00
Pierre-Eric Pelloux-Prayer dc56301f78 radeonsi: treat nir_intrinsic_load_constant as a VMEM operation
This is used by variable indexing of constant arrays, to build
code like this:

   s_add_u32 s6, s6, const_data@rel32@lo+4
   s_addc_u32 s7, s7, const_data@rel32@hi+12
   [...]
   global_load_dword v4, v4, s[6:7

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/5118
Fixes: 8288882965 ("radeonsi: set MEM_ORDERED optimally")
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13550>
2021-11-04 09:38:20 +01:00
Joshua Ashton af8fa2644e radv: Split off cmd_buffer variant of descriptor set template updates
Assumes cmd_buffer != NULL for this path to eliminate the checks in the generic code.

Benchmarks:

I made a microbenchmark based on Bas' vulkan microbench suite.

https://gitlab.freedesktop.org/bnieuwenhuizen/vulkan_microbench/-/merge_requests/1

In this benchmark, this improves vkDescriptorTemplateUpdate performance consistently by 36%.

-------------------------------------------------------------------
Benchmark                         Time             CPU   Iterations
-------------------------------------------------------------------
DescriptorTemplateUpdate       81.2 ns         81.2 ns      8573169

->

-------------------------------------------------------------------
Benchmark                         Time             CPU   Iterations
-------------------------------------------------------------------
DescriptorTemplateUpdate       52.9 ns         52.9 ns     13306065

Signed-off-by: Joshua Ashton <joshua@froggi.es>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13342>
2021-11-04 05:50:54 +00:00
Joshua Ashton 4865cb6514 radv: Split off cmd_buffer variant of descriptor set updates
Assumes cmd_buffer != NULL for this path to eliminate the checks in the generic code.

Signed-off-by: Joshua Ashton <joshua@froggi.es>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13342>
2021-11-04 05:50:48 +00:00
Joshua Ashton 7b0826db56 radv: Always inline descriptor writes
Improves performance, see next commit for benchmarks.

Signed-off-by: Joshua Ashton <joshua@froggi.es>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13342>
2021-11-04 05:50:47 +00:00