Commit graph

2437 commits

Author SHA1 Message Date
Dave Airlie
82a2f10529 radv/gfx10: set the pgm rsrc3/4 regs using index sh reg set
This is ported from AMDVLK, it's probably not requires unless
we want to use "real time queues", but it might be nice to just have
in place.

Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-18 10:24:26 +10:00
Dave Airlie
de524b2c37 radv: use correct register setter for ngg hw addr
this shouldn't matter, but it's good to be correct.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
2019-07-18 10:17:37 +10:00
Bas Nieuwenhuizen
f1a8967344 radv: Only save the descriptor set if we have one.
After reset, if valid does not contain the relevant bit the descriptor
can be != NULL but still not be valid.

CC: <mesa-stable@lists.freedesktop.org>
Reviewed-by: Dave Airlie <airlied@redhat.com>
2019-07-18 00:49:43 +02:00
Samuel Pitoiset
24b1b1f574 radv: add an option for disabling NGG on GFX10
Will be useful for testing the legacy path.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-17 15:43:36 +02:00
Samuel Pitoiset
07ff367442 radv/gfx10: implement VK_EXT_post_depth_coverage
I did implement this extension a while ago but it didn't work
on pre GFX10 for some reasons. Now all CTS pass.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-17 08:32:39 +02:00
Samuel Pitoiset
ed53d2c4be radv/gfx10: disable the TC compat zrange workaround
Unnecessary.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-17 08:32:36 +02:00
Samuel Pitoiset
edf1af696f radv/gfx10: fallback to the legacy path if tess and extreme geometry
This is unsupported and hangs.

This fixes GPU hangs with
dEQP-VK.tessellation.geometry_interaction.limits.output_required_*.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-17 08:32:33 +02:00
Samuel Pitoiset
ae4b1fc095 radv/gfx10: always build the GS copy shader but uses it on-demand
It should be possible to build it on-demand too but it requires
more work. On GFX10, the GS copy shader is required when tess
is enabled with extreme geometry.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-17 08:32:30 +02:00
Mauro Rossi
3630988b1d android: radv/gfx10: generate gfx10_format_table.h
This patch adds the missing building rules for Android,
to avoid following building errors:

In file included from external/mesa/src/amd/vulkan/radv_debug.c:35:
In file included from external/mesa/src/amd/vulkan/radv_debug.h:27:
external/mesa/src/amd/vulkan/radv_private.h:95:10:
fatal error: 'gfx10_format_table.h' file not found
         ^~~~~~~~~~~~~~~~~~~~~~
1 error generated.

In file included from external/mesa/src/amd/vulkan/radv_android.c:31:
external/mesa/src/amd/vulkan/radv_private.h:95:10:
fatal error: 'gfx10_format_table.h' file not found
         ^~~~~~~~~~~~~~~~~~~~~~
1 error generated.

Fixes: 3dc5ec5d16 ("radv/gfx10: generate gfx10_format_table.h")
Signed-off-by: Mauro Rossi <issor.oruam@gmail.com>
Acked-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
2019-07-16 21:31:24 +02:00
Samuel Pitoiset
afa102d65b radv: add radv_emit_streamout_{begin,end} helpers
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
2019-07-16 11:17:00 +02:00
Samuel Pitoiset
17464d205c radv: pass output values to radv_emit_stream_output()
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
2019-07-16 11:16:58 +02:00
Samuel Pitoiset
4dcdc4cdc5 radv: allow to select DST_SEL with RELEASE_MEM
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
2019-07-16 11:16:57 +02:00
Samuel Pitoiset
3c6d6bd71f radv: allow to emit PS_DONE/CS_DONE with RELEASE_MEM
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
2019-07-16 11:16:55 +02:00
Samuel Pitoiset
219dc1b25c radv: restore an assertion in handle_vs_outputs()
The NGG GS epilogue no longers call that function so the assertion
is just useless now.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
2019-07-16 11:16:53 +02:00
Samuel Pitoiset
68603b767f radv/gfx10: emit ES outputs of TES when it's not NGG
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
2019-07-16 11:16:51 +02:00
Samuel Pitoiset
b0f7a6e981 radv: update LATE_ALLOC_VS.LIMIT
Mirror RadeonSI.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
2019-07-16 10:10:22 +02:00
Samuel Pitoiset
27d91062a8 radv/gfx10: support pixel shaders without exports
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
2019-07-16 10:10:21 +02:00
Samuel Pitoiset
1b2bfeaaaa radv: fix gathering clip/cull distance masks for GS
For NGG, the driver relies on the VS outinfo struct.

This fixes
dEQP-VK.clipping.user_defined.clip_*_vert_tess_geom_*

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Dave Airlie <airlied@redhat.com>
2019-07-16 10:09:37 +02:00
Samuel Pitoiset
361d549f87 Revert "radv/gfx10: don't set array pitch field on images"
It introduces too many regressions.

This reverts commit 6d50dcd80f.
2019-07-16 09:37:56 +02:00
Samuel Pitoiset
994253b400 radv/gfx10: add missing conversions for 16-bit exports
This fixes
dEQP-VK.spirv_assembly.instruction.graphics.16bit_storage.input_output_*

Found with RADV_DEBUG=checkir

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-16 08:12:34 +02:00
Samuel Pitoiset
d8844533af radv: remove unused code in radv_export_param()
It was hack for geometry shaders.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-16 08:12:20 +02:00
Dave Airlie
6d50dcd80f radv/gfx10: don't set array pitch field on images
Setting this seems to be broken, amdvlk only sets it for quilted
textures which I'm not sure what those are.

Fixes dEQP-VK.glsl.texture_functions.query.texturesize*3d*

Fixes: bf11f1c3a4 ("radv/gfx10: add gfx10_make_texture_descriptor")
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-16 10:41:27 +10:00
Samuel Pitoiset
ed12be1b8f radv/gfx10: enable OC_LDS_EN for NGG GS if the ES stage is TES
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-15 20:05:21 +02:00
Samuel Pitoiset
b650f3d197 radv/gfx10: export the PrimitiveID for ES stages (VS or TES)
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-15 11:30:10 +02:00
Samuel Pitoiset
8175f6269b radv/gfx10: declare an external symbol for the ESGS ring
It will be used for stream output but for now only declares it
if VS and if the PrimitiveID needs to be exported.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-15 11:30:08 +02:00
Samuel Pitoiset
f0a90eddb6 radv/gfx10: allocate ESGS ring space for exporting PrimitiveID
Only VS needs that. We shouldn't hardcode these values but
that's complicated to not do that for now.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-15 11:30:05 +02:00
Samuel Pitoiset
4478f14327 radv/gfx10: fix crash when emitting NGG GS prologue
ac_nir_context is initialized after the driver emits the NGG GS
prologue so it's likely to crash.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-15 08:51:53 +02:00
Bas Nieuwenhuizen
1f58b6ffef radv/gfx10: Fix DCC clears.
Looks like if the reg clear bit is set, the hwardware does not use the 0/1
clears for textures.

Reviewed-by: Dave Airlie <airlied@redhat.com>
2019-07-14 07:30:04 +00:00
Samuel Pitoiset
f239e22813 radv/gfx10: enable 1D textures
Mirror RadeonSI. This also fixes crashes in addrlib.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-12 18:25:45 +02:00
Samuel Pitoiset
b393b2ce95 radv/gfx10: emit DISABLE_CONSERVATIVE_ZPASS_COUNTS
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-12 17:47:12 +02:00
Samuel Pitoiset
8cc4e4a81e radv/gfx10: init more registers in the graphics preamble
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-12 17:47:12 +02:00
Samuel Pitoiset
e68b55f5e3 radv/gfx10: set HS/GS/CS.WGP_MODE
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-12 17:47:12 +02:00
Samuel Pitoiset
5d5e26230a radv/gfx10: emit GE_PC_ALLOC
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-12 17:47:11 +02:00
Samuel Pitoiset
df062afa03 radv/gfx10: enable vertex shaders without export parameters
GFX10 allows this.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-12 17:47:11 +02:00
Samuel Pitoiset
3f76c0f47c radv/gfx10: launch 2 compute waves per CU before going onto the next CU
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-12 17:47:11 +02:00
Samuel Pitoiset
e631d65fc6 radv: use ac_get_compute_resource_limits()
No behaviour change.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-12 17:47:11 +02:00
Samuel Pitoiset
37aefb2be1 radv/gfx10: invalidate everything in L2 when shaders read data
This includes metadata as well. On GFX10, we have to invalidate
the L2 metadata cache when shaders read DCC.

Note that we still have to implement GFX10 coherency by
introducing INV_L2_METATADA but for now just flush L2.

This fixes a corruption with DCC and Talos.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-12 14:08:12 +02:00
Samuel Pitoiset
4e38322dd8 radv/gfx10: fix wrong emission of GE_CNTL
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-12 12:15:08 +02:00
Samuel Pitoiset
219d6939df radv: add more assertions to make sure packets are correctly emitted
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-12 12:15:06 +02:00
Samuel Pitoiset
958ee4c21a radv: report shader stage name when dumping LLVM IR
For debugging purposes.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-12 08:19:53 +02:00
Samuel Pitoiset
2b6a089813 radv: tidy up radv_get_shader_name() and add NGG stages
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-12 08:19:53 +02:00
Samuel Pitoiset
ffd6a979bf radv/gfx10: update OVERWRITE_COMBINER_{MRT_SHARING,WATERMARK}
DCC related, mirror RadeonSI.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Dave Airlie <airlied@redhat.com
2019-07-12 08:19:53 +02:00
Samuel Pitoiset
c6fa4de15d radv/gfx10: do not set alignment on the ngg_emit pointer
This is invalid and this fixes a crash in LLVM.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-12 08:19:53 +02:00
Samuel Pitoiset
df0a23ad1e radv/gfx10: fix exporting clip/cull distances for GS
This fixes dEQP-VK.clipping.user_defined.clip_distance.*geom*.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-12 08:19:53 +02:00
Samuel Pitoiset
edcd2bc833 radv/gfx10: fix exporting the subpass view index for GS
This fixes dEQP-VK.multiview.*geometry*.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-12 08:19:20 +02:00
Samuel Pitoiset
a7b7e94085 radv: only enable the GS copy shader stage if GS is enabled
Ooops.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
2019-07-11 21:44:44 +02:00
Samuel Pitoiset
cd403a931f radv/gfx10: enable geometry shaders
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
2019-07-11 15:46:02 +02:00
Bas Nieuwenhuizen
0a8ef756d3 radv/gfx10: Fix NGG GS output mask handlings for LDS indexing.
In emit_vertex we optimize storage if the output mask does not
have all bits set. Do the same in the epilogue so the indices
actually match up.

Fixes dEQP-VK.geometry.input.basic_primitive.points because it
outputs PSIZE with an output mask of 1, which cause the generic
attribute for the color to be loaded from the wrong indices.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
2019-07-11 15:45:59 +02:00
Bas Nieuwenhuizen
f5982917ff radv/gfx10: Simplify output mask handling for NGG GS.
We only ever get in this function for a NGG GS proper.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
2019-07-11 15:45:58 +02:00
Bas Nieuwenhuizen
7515f41c78 radv/gfx10: Do GS prologue outside of gs_threads if.
Mirror radeonsi.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
2019-07-11 15:45:56 +02:00