Commit graph

2087 commits

Author SHA1 Message Date
Samuel Pitoiset
3ecf9a17f8 radv: stop abusing dirty_dynamic when binding a NULL fragment shader
RADV_DYNAMIC_RASTERIZATION_SAMPLES triggers the following states:
- FS (not needed when FS is NULL)
- MSAA (already triggered)
- BINNING (use radv_get_ps_iter_samples() on GFX9)
- OCCLUSION_QUERY (doesn't use the PS)
- DB_SHADER_CONTROL (already triggered)
- RAST_SAMPLES (use radv_get_ps_iter_samples())
- NGGC (doesn't use the PS)

So this can be simplified to BINNING (gfx9) | RAST_SAMPLES.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36912>
2025-08-25 12:17:38 +00:00
Samuel Pitoiset
19c712c8ef radv: rename rast_prim to vgt_outprim_type everywhere
To avoid confusion between the primitive topology and the output
rasterized primitive.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36912>
2025-08-25 12:17:38 +00:00
Samuel Pitoiset
5dec043cee radv: dirty the rast samples state when VRS is forced to 1x1
Only these two states use force_sample_iter_shading_rate.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36912>
2025-08-25 12:17:37 +00:00
Samuel Pitoiset
8120d14ee0 radv: dirty RADV_CMD_DIRTY_xx states when binding sample shading state
This reduces the number of states that are re-emitted but the logic
is mostly duplicated because sample shading can be set from the
fragment shader or the graphics pipeline. Could be refactored
eventually.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36912>
2025-08-25 12:17:34 +00:00
Samuel Pitoiset
4afff72e91 radv: dirty the rast sample states for VRS att/OOO rast
These fields are only used by radv_emit_rast_samples_state().

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36912>
2025-08-25 12:17:32 +00:00
Samuel Pitoiset
4dd05c9a91 radv: remove dead code when setting dynamic primitive topology
This is already handled slightly above in the same function. Also
state->dirty isn't for RADV_DYNAMIC_xxx and there is no corresponding
RADV_CMD_DIRTY_xxx either.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36912>
2025-08-25 12:17:31 +00:00
Samuel Pitoiset
dbcc9aa942 radv: remove radv_cmd_buffer_flush_dynamic_state()
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36912>
2025-08-25 12:17:31 +00:00
Samuel Pitoiset
12f8508363 radv: add a new dirty bit for the VGT prim state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36912>
2025-08-25 12:17:31 +00:00
Samuel Pitoiset
4096009ea1 radv: add a new dirty bit for the patch control points state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36912>
2025-08-25 12:17:30 +00:00
Samuel Pitoiset
0c3680bea3 radv: add a new dirty bit for the tess domain origin state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36912>
2025-08-25 12:17:29 +00:00
Samuel Pitoiset
99b287bde5 radv: add missing L2 invalidate cache flush for non-coherent images
Images aren't always coherent with L2 and AMD generations have
different rules, see radv_image_is_l2_coherent() for the full picture.

This fixes a rendering issue on GFX9 because depth/stencil images
aren't coherent, but this also affects color images.

This also fixes a cache coherency issue with an ongoing extension.

Cc: mesa-stable
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/12274
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36815>
2025-08-25 07:28:50 +00:00
Samuel Pitoiset
772f185fc0 radv: add a new dirty bit for the scissor state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36901>
2025-08-22 08:12:05 +00:00
Samuel Pitoiset
54b79d0f50 radv: add a new dirty bit for the sample locations state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36901>
2025-08-22 08:12:04 +00:00
Samuel Pitoiset
dc3371332e radv: add a new dirty bit for the blend constants state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36901>
2025-08-22 08:12:04 +00:00
Samuel Pitoiset
1be9cd43a8 radv: add a new dirty bit for the VS prolog state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36901>
2025-08-22 08:12:03 +00:00
Samuel Pitoiset
cc7b5ab0c9 radv: dirty more states when rendering begins
radv_get_ps_iter_samples() is used by :
- binning state
- rast samples state
- FS state
- MSAA state

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36901>
2025-08-22 08:12:02 +00:00
Samuel Pitoiset
80e8cb196f radv: dirty the cb render state when rendering begins
RADV_DYNAMIC_COLOR_BLEND_ENABLE trigger the cb render state, the db
shader control state and PS epilogs. Though, it's unecessary to re-emit
the db shader control and PS epilogs are already triggered when the
fb is dirty.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36901>
2025-08-22 08:12:02 +00:00
Samuel Pitoiset
1db3d4f182 radv: dirty the depth stencil state when rendering begins
RADV_DYNAMIC_STENCIL_TEST_ENABLE only triggers
RADV_CMD_DIRTY_DEPTH_STENCIL_STATE but it's better to dirty the RADV
state than the dynamic one.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36901>
2025-08-22 08:12:01 +00:00
Samuel Pitoiset
7694ad969e radv: add a new dirty bit for the depth bias state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36901>
2025-08-22 08:12:00 +00:00
Samuel Pitoiset
cbe42a4c43 radv: rename RADV_CMD_DIRTY_TESS_STATE to RADV_CMD_DIRTY_TCS_TES_STATE
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36875>
2025-08-21 18:36:35 +00:00
Samuel Pitoiset
3d1efbce98 radv: add a new dirty bit for the rast samples state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36875>
2025-08-21 18:36:35 +00:00
Samuel Pitoiset
ea6cb5e7bf radv: add a new dirty bit for the FSR state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36875>
2025-08-21 18:36:35 +00:00
Samuel Pitoiset
604296a282 radv: add a new dirty bit for the ngg culling state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36875>
2025-08-21 18:36:34 +00:00
Samuel Pitoiset
d9d9052c56 radv: validate dynamic states for the db shader control state earlier
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36875>
2025-08-21 18:36:34 +00:00
Samuel Pitoiset
02ecb70e7f radv: validate dynamic states for the occlusion query state earlier
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36875>
2025-08-21 18:36:33 +00:00
Samuel Pitoiset
81af123ab1 radv: optimize re-emitting the occlusion query state on GFX12
The number of samples is only needed on older generations.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36875>
2025-08-21 18:36:33 +00:00
Samuel Pitoiset
5db8a16053 radv: add a new dirty bit for the binning state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36875>
2025-08-21 18:36:32 +00:00
Samuel Pitoiset
c4c44187f9 radv: only re-emit needed states when PS inner coverage changes
It's emitted as part of the MSAA state but re-emitting the NGG culling
state isn't necessary because ps.reads_fully_covered isn't used.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36875>
2025-08-21 18:36:32 +00:00
Samuel Pitoiset
fae90ba171 radv: do not emit few RADV_CMD_DIRTY_xxx based on dynamic states
Like RADV_CMD_DIRTY_CB_RENDER_STATE is dirty when the custom blend mode
is updated. Shouldn't fix anything in practice because dynamic_states
is unlikely 0 but better that way.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36875>
2025-08-21 18:36:31 +00:00
Samuel Pitoiset
d40e841cc4 radv: dirty some states from graphics pipeline earlier
Some checks are pending
macOS-CI / macOS-CI (dri) (push) Waiting to run
macOS-CI / macOS-CI (xlib) (push) Waiting to run
This might actually fixes a couple of things because needed dynamic
states are computed before radv_emit_graphics_pipeline(), so dirtying
them too late doesn't make much sense.

This doesn't fix anything known.

Cc: mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36900>
2025-08-21 15:45:48 +00:00
Samuel Pitoiset
5024c02d45 radv: precompute the depth clip enable
This should avoid re-emitting some states if it doesn't actually change.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36852>
2025-08-21 08:23:04 +00:00
Samuel Pitoiset
2b5844df0e radv: precompute the depth clamp mode
This should avoid re-emitting the state if it doesn't actually change.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36852>
2025-08-21 08:23:04 +00:00
Samuel Pitoiset
413f781234 radv: add a new dirty bit for the viewport state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36852>
2025-08-21 08:23:03 +00:00
Samuel Pitoiset
2733b2953e radv: emit depth clamp enable as part of the viewport state
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36852>
2025-08-21 08:23:02 +00:00
Samuel Pitoiset
9c6f37c533 radv: get the depth clamp mode earlier when emitting viewports
Outside of the loop is also faster.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36852>
2025-08-21 08:23:01 +00:00
Marek Olšák
ed246aafd8 util/set: set _mesa_set_init return type to void
it always returns true because it no longer allocates anything

Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36728>
2025-08-21 06:13:48 +00:00
Marek Olšák
c12118decf radv,zink,st/mesa: use _mesa_set_fini instead of ralloc_free
This is the correct way to free the set.

Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36728>
2025-08-21 06:13:47 +00:00
Samuel Pitoiset
baaf5d643a radv: emit inlined push constants with buffered SH regs on GFX12
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36570>
2025-08-19 18:01:23 +00:00
Samuel Pitoiset
c710eaa443 radv: emit descriptor pointers with buffered SH regs on GFX12
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36570>
2025-08-19 18:01:22 +00:00
Samuel Pitoiset
95d2f009a9 radv: emit compute pipeline with buffered SH regs on GFX12
This also includes RT, task shaders and DGC IES for compute.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36570>
2025-08-19 18:01:21 +00:00
Samuel Pitoiset
bbf8338443 radv: rework the helper to emit buffered regs on GFX12
Also reserve enough space if needed.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36570>
2025-08-19 18:01:21 +00:00
Samuel Pitoiset
1f26f93aa7 radv: emit relocation for task shaders at the same place as other stages
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36570>
2025-08-19 18:01:21 +00:00
Samuel Pitoiset
2c26c5deb7 radv: merge two similar loops in lookup_ps_epilog()
Cleanup.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36778>
2025-08-19 11:20:52 +00:00
Samuel Pitoiset
45c91edd18 radv: fix color attachment remapping with fast-GPL/ESO
If vkCmdSetRenderingAttachmentLocations() isn't setting all color
attachment locations (ie. MAX_RTS), the remapping might be wrong
because MESA_VK_ATTACHMENT_UNUSED is used to trim the unused locations

Found by inspection while implementing a new extension.

Cc: mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36778>
2025-08-19 11:20:52 +00:00
Samuel Pitoiset
282186d4aa radv: add a function that uploads push constants
Similar to indirect descriptor sets logic.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36792>
2025-08-18 07:25:34 +00:00
Samuel Pitoiset
4e8728c4f6 radv: rework emitting push constants for less CPU overhead
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36792>
2025-08-18 07:25:34 +00:00
Samuel Pitoiset
81ec36edb8 radv: determine if push constants need to be uploaded earlier
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36792>
2025-08-18 07:25:34 +00:00
Samuel Pitoiset
47fd1803b2 radv: use radv_shader_need_indirect_descriptor_sets() more
While we are at it, move it to the appropriate header file.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36792>
2025-08-18 07:25:33 +00:00
Samuel Pitoiset
8f0ff009c0 radv: do not emit inlined SGPRs twice for merged shaders
This is wasteful.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36792>
2025-08-18 07:25:33 +00:00
Samuel Pitoiset
4bd0bf7e19 radv: invalidating push constants for compute<->rt during dispatches
It's similar but a bit cleaner.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36792>
2025-08-18 07:25:32 +00:00