Make gpgpu fill tests libdrm independent
Get series with b4
b4 am [email protected]Versions newest first · linked by author and title
- v7 Latest8 patches 1 replyIn review
- v68 patches 2 repliesSuperseded
- v58 patches 2 repliesSuperseded
Earlier versions · 4
| Version | Posted | Patches | Replies | Status |
|---|---|---|---|---|
| v4 | 8 patches | 0 replies | Superseded | |
| v3 | 7 patches | 5 replies | Superseded | |
| v2 | 7 patches | 0 replies | Superseded | |
| v1 | 6 patches | 2 replies | Superseded |
Patches in v7 No applies recorded yet
| # | Subject and files | Size | Replies | Applied |
|---|---|---|---|---|
| 1/8 | lib/intel_bufops: Add bufops reference and adapt stride requirementlib/ lib/ | +50 | 0 replies | |
| 2/8 | lib/rendercopy_bufmgr: Pass alignment during buffer initializationlib/ | +2 | 0 replies | |
| 3/8 | lib/intel_batchbuffer: Introduce intel_bblib/ lib/ | +578 | 0 replies | |
| 4/8 | lib/gpu_cmds: Add gpgpu pipeline functions based on intel_bblib/ lib/ | +617 | 0 replies | |
| 5/8 | lib/gpgpu_fill: libdrm-free gpgpu pipeline creationlib/ lib/ | +221 | 0 replies | |
| 6/8 | lib/intel_batchbuffer: Introduce temporary igt_fillfunc_v2_tlib/ lib/ | +35 | 0 replies | |
| 7/8 | tests/gem_gpgpu_fill: Remove libdrm dependencytests/ | +99 | 0 replies | |
| 8/8 | HAX: run gpgpu_fill in BAT onlytests/ | +1 | 0 replies |
Reviews trailers per patch
None of the 8 patches has a Reviewed-by or Acked-by of their own; the series has one for all of them.
| Patch | Chris Wilson |
|---|---|
| Series | R-b |
| 1/8 | |
| 2/8 | |
| 3/8 | |
| 4/8 | |
| 5/8 | |
| 6/8 | |
| 7/8 | |
| 8/8 |
Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.
Cover letter [igt-dev] [PATCH i-g-t v7 0/8] Make gpgpu fill tests libdrm independent
I've introduced intel_bb which is very simplified version of libdrm powered intel_batchbuffer. Currently it supports simple relocation but it is enough to cover all gpgpu fill needs. Other (fenced) relocation will be added when I will have to face with such pipelines / batchbuffers.
As commands created in gpu_cmds.c are coupled more than I thought on the beginning (media fill also use them) "_v2" suffix was introduced. When all tests will be rewritten to "_v2" old functions can be removed and "_v2" can be deleted.
v2: changes according to the review:
- make intel_bb api more consistent and universal
- add intel_bb api documentation
- add alignment field in buf_ops buffer initalization for linear buffers
v3: changes:
- add all gens pipelines
- fix compiling issues on non-x86 archs
- add ctx (suggested by knr)
v4: changes:
- add indexing tree and render target write flag
v5: changes:
- randomizing addresses to avoid relocations
- fixing bug in gen8+ pipeline setup (Chris)
v6: changes:
- fixing invalid realloc size
- change invalid WB mapping to WC
v7: changes:
- IS_BROADWELL() to IS_GEN8(), this is likely historical issue so lets CI verify
Zbigniew Kempczyński (8):
lib/intel_bufops: Add bufops reference and adapt stride requirement lib/rendercopy_bufmgr: Pass alignment during buffer initialization lib/intel_batchbuffer: Introduce intel_bb
lib/gpu_cmds: Add gpgpu pipeline functions based on intel_bb lib/gpgpu_fill: libdrm-free gpgpu pipeline creation lib/intel_batchbuffer: Introduce temporary igt_fillfunc_v2_t tests/gem_gpgpu_fill: Remove libdrm dependency
HAX: run gpgpu_fill in BAT only
lib/gpgpu_fill.c | 187 +++++++++ lib/gpgpu_fill.h | 34 ++ lib/gpu_cmds.c | 563 ++++++++++++++++++++++++++ lib/gpu_cmds.h | 54 +++ lib/intel_batchbuffer.c | 504 +++++++++++++++++++++++ lib/intel_batchbuffer.h | 109 +++++ lib/intel_bufops.c | 55 ++- lib/intel_bufops.h | 7 +- lib/rendercopy_bufmgr.c | 4 +- tests/i915/gem_gpgpu_fill.c | 126 ++++-- tests/intel-ci/fast-feedback.testlist | 161 +------- 11 files changed, 1603 insertions(+), 201 deletions(-)
Signature · 6 lines
2.26.0 _______________________________________________ igt-dev mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/igt-dev