[PATCH v4 0/3] drm/xe: Fix races on fdinfo

[Lucas De Marchi]

The current reading of engine utilization has same races. This should
fix most of them while also drastically reducing the update rate needed
on "normal apps".

I left tests/xe_drm_fdinfo --r utilization-single-full-load-destroy-queue
running on 2 systems and saw no failures after 100 iterations about
execution cycles being 0.

As for the patch to cache hwe, I'm still not sure it's worth. Tracing
show_run_ticks() doesn't show noticeable performance improvement.

v4:
  - Remove patches already applied for tracepoint, doc fix and where the
    timestamp is accumulate
  - Move wait to the beginning, to avoid impacting the read of gpu
    timestamp vs pceu
  - Refactor patch to extract force wake getting: this should make it
    simpler to eventually cache hwe
  - Add patch to cache hwe

Lucas De Marchi (3):
  drm/xe: Wait on killed exec queues
  drm/xe: Sample gpu timestamp closer to exec queues
  drm/xe: Cache hwe used for timestamp

 drivers/gpu/drm/xe/xe_device.c       |  2 +
 drivers/gpu/drm/xe/xe_device_types.h | 10 ++++
 drivers/gpu/drm/xe/xe_drm_client.c   | 90 ++++++++++++++++++++--------
 drivers/gpu/drm/xe/xe_drm_client.h   |  2 +
 drivers/gpu/drm/xe/xe_exec_queue.c   |  6 ++
 5 files changed, 86 insertions(+), 24 deletions(-)

-- 
2.47.0