All of lore.kernel.org
 help / color / mirror / Atom feed
* [Intel-xe] [PATCH v2 0/7] Cpu page-table updates and fixes
@ 2023-03-15 15:55 Thomas Hellström
  2023-03-15 15:55 ` [Intel-xe] [PATCH v2 1/7] drm/xe: Use a define to set initial seqno for fences Thomas Hellström
                   ` (10 more replies)
  0 siblings, 11 replies; 15+ messages in thread
From: Thomas Hellström @ 2023-03-15 15:55 UTC (permalink / raw)
  To: intel-xe; +Cc: matthew.auld

There were flaws in the engine idle computation used to determine whether
we could use CPU page-table updates. Fix those

It also turns out fixing those triggers CPU-page-table updates in the
migrate test, which didn't really support that, so also fix the migrate test
to handle that and to run the test twice, once with CPU- and once with
GPU page-table updates, and to also provide some timings for
non-synced updates. (The timings avoid the GPU submission latency
to better reflect the case where we are able to pipeline GPU
page-table updates).

Provide some coverage for the fence seqno 32-bit wrapping
case by setting the initial fence seqno to a small negative number.

Finally use the new xe_engine_is_idle() function to fix the memory
ping-ponging issue that can be triggered by the xe_noexec_ping_pong igt
currently on review.

v2:
- Use __dma_fence_is_later() to compare fence seqnos.
- Add Patch drm/xe/vm: Defer vm rebind until next exec if nothing to execute

Thomas Hellström (7):
  drm/xe: Use a define to set initial seqno for fences
  drm/xe/migrate: Update cpu page-table updates
  drm/xe/tests: Support CPU page-table updates in the migrate test
  drm/xe: Introduce xe_engine_is_idle()
  drm/xe: Use a small negative initial seqno
  drm/xe/tests: Test both CPU- and GPU page-table updates with the
    migrate test
  drm/xe/vm: Defer vm rebind until next exec if nothing to execute

 drivers/gpu/drm/xe/tests/xe_migrate.c | 35 +++++++++++++++--
 drivers/gpu/drm/xe/tests/xe_test.h    |  1 +
 drivers/gpu/drm/xe/xe_engine.c        | 28 +++++++++++++-
 drivers/gpu/drm/xe/xe_engine.h        |  2 +
 drivers/gpu/drm/xe/xe_exec.c          |  1 +
 drivers/gpu/drm/xe/xe_hw_fence.c      |  4 +-
 drivers/gpu/drm/xe/xe_hw_fence.h      |  3 ++
 drivers/gpu/drm/xe/xe_lrc.c           |  3 ++
 drivers/gpu/drm/xe/xe_migrate.c       | 55 ++++++++++++++-------------
 drivers/gpu/drm/xe/xe_vm.c            | 18 +++++++++
 drivers/gpu/drm/xe/xe_vm.h            | 17 +++++++++
 drivers/gpu/drm/xe/xe_vm_types.h      |  5 +++
 12 files changed, 137 insertions(+), 35 deletions(-)

-- 
2.39.2


^ permalink raw reply	[flat|nested] 15+ messages in thread

end of thread, other threads:[~2023-03-16 16:45 UTC | newest]

Thread overview: 15+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-03-15 15:55 [Intel-xe] [PATCH v2 0/7] Cpu page-table updates and fixes Thomas Hellström
2023-03-15 15:55 ` [Intel-xe] [PATCH v2 1/7] drm/xe: Use a define to set initial seqno for fences Thomas Hellström
2023-03-16 16:09   ` Matthew Brost
2023-03-16 16:45     ` Thomas Hellström
2023-03-15 15:55 ` [Intel-xe] [PATCH v2 2/7] drm/xe/migrate: Update cpu page-table updates Thomas Hellström
2023-03-15 15:55 ` [Intel-xe] [PATCH v2 3/7] drm/xe/tests: Support CPU page-table updates in the migrate test Thomas Hellström
2023-03-15 15:55 ` [Intel-xe] [PATCH v2 4/7] drm/xe: Introduce xe_engine_is_idle() Thomas Hellström
2023-03-15 15:55 ` [Intel-xe] [PATCH v2 5/7] drm/xe: Use a small negative initial seqno Thomas Hellström
2023-03-15 15:55 ` [Intel-xe] [PATCH v2 6/7] drm/xe/tests: Test both CPU- and GPU page-table updates with the migrate test Thomas Hellström
2023-03-15 15:55 ` [Intel-xe] [PATCH v2 7/7] drm/xe/vm: Defer vm rebind until next exec if nothing to execute Thomas Hellström
2023-03-15 23:23   ` Matthew Brost
2023-03-15 15:58 ` [Intel-xe] ✓ CI.Patch_applied: success for Cpu page-table updates and fixes (rev2) Patchwork
2023-03-15 15:59 ` [Intel-xe] ✓ CI.KUnit: " Patchwork
2023-03-15 16:03 ` [Intel-xe] ✓ CI.Build: " Patchwork
2023-03-15 16:14 ` [Intel-xe] ○ CI.BAT: info " Patchwork

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.