Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH 0/2] Fixes for MI_REPORT_PERF_COUNT
@ 2024-12-19  0:23 Umesh Nerlige Ramappa
  2024-12-19  0:23 ` [PATCH 1/2] xe/oa: Fix query mode of operation for OAR/OAC Umesh Nerlige Ramappa
                   ` (9 more replies)
  0 siblings, 10 replies; 18+ messages in thread
From: Umesh Nerlige Ramappa @ 2024-12-19  0:23 UTC (permalink / raw)
  To: intel-xe, Ashutosh Dixit, Lucas De Marchi; +Cc: jose.souza

OA programming sequence for query mode or MI_REPORT_PERF_COUNT requires
modifying some HW registers in the same hw context as the user exec
queue. User passes the exec_queue to the OA interface and OA
implementation submits an MI_LOAD_REGISTER_IMM to this queue to modify
the registers.

The OA implementation submits a batch mapped in GGTT to the user exec
queue and hence, some plumbing is added into relevant code to enable
that (as per suggestions from Matthew Brost).

v2: review rework
v3:
- review rework
- original patches squashed for porting to stable
- code cleanup

opens:
- Should the cleanup patch be squashed as well and sent to stable? I have kept
  it separate hoping the build does not complain about dead code!!
- Applying these patches to stable is not clean since there is a dependency on
  another patch - "drm/xe: Force write completion of MI_STORE_DATA_IMM". Not
  sure if dependent patches need to be sent to stable as well.

Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>

Umesh Nerlige Ramappa (2):
  xe/oa: Fix query mode of operation for OAR/OAC
  xe/oa: Drop the unused logic to parse context image

 drivers/gpu/drm/xe/xe_oa.c              | 223 +++++-------------------
 drivers/gpu/drm/xe/xe_oa_types.h        |   3 -
 drivers/gpu/drm/xe/xe_ring_ops.c        |   5 +-
 drivers/gpu/drm/xe/xe_sched_job_types.h |   2 +
 4 files changed, 53 insertions(+), 180 deletions(-)

-- 
2.34.1


^ permalink raw reply	[flat|nested] 18+ messages in thread
* [PATCH 0/2] Fixes for MI_REPORT_PERF_COUNT
@ 2024-12-17  0:58 Umesh Nerlige Ramappa
  2024-12-17 17:05 ` Souza, Jose
  0 siblings, 1 reply; 18+ messages in thread
From: Umesh Nerlige Ramappa @ 2024-12-17  0:58 UTC (permalink / raw)
  To: Ashutosh Dixit, intel-xe; +Cc: matthew.brost

OA programming sequence for query mode or MI_REPORT_PERF_COUNT requires
modifying some HW registers in the same hw context as the user exec
queue. User passes the exec_queue to the OA interface and OA
implementation submits an MI_LOAD_REGISTER_IMM to this queue to modify
the registers.

The OA implementation submits a batch mapped in GGTT to the user exec
queue and hence, some plumbing is added into relevant code to enable
that (as per suggestions from Matthew Brost).

v2: review rework

Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>

Umesh Nerlige Ramappa (2):
  xe: Allow a GGTT mapped batch to be submitted to user exec queue
  xe/oa: Use MI_LOAD_REGISTER_IMMEDIATE to enable OAR/OAC

 drivers/gpu/drm/xe/xe_oa.c              | 71 ++++++++++++++++++-------
 drivers/gpu/drm/xe/xe_ring_ops.c        |  5 +-
 drivers/gpu/drm/xe/xe_sched_job_types.h |  2 +
 3 files changed, 57 insertions(+), 21 deletions(-)

-- 
2.34.1


^ permalink raw reply	[flat|nested] 18+ messages in thread
* [PATCH 0/2] Fixes for MI_REPORT_PERF_COUNT
@ 2024-11-27  0:31 Umesh Nerlige Ramappa
  0 siblings, 0 replies; 18+ messages in thread
From: Umesh Nerlige Ramappa @ 2024-11-27  0:31 UTC (permalink / raw)
  To: intel-xe, Ashutosh Dixit, matthew.brost

OA programming sequence for query mode or MI_REPORT_PERF_COUNT requires
modifying some HW registers in the same hw context as the user exec
queue. User passes the exec_queue to the OA interface and OA
implementation submits an MI_LOAD_REGISTER_IMM to this queue to modify
the registers.

The OA implementation submits a batch mapped in GGTT to the user exec
queue and hence, some plumbing is added into relevant code to enable
that (as per suggestions from Matthew Brost).

Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>

Umesh Nerlige Ramappa (2):
  xe: Allow a GGTT mapped batch to be submitted to user exec queue
  xe/oa: Use MI_LOAD_REGISTER_IMMEDIATE to enable OAR/OAC

 drivers/gpu/drm/xe/xe_oa.c              | 68 ++++++++++++++++++-------
 drivers/gpu/drm/xe/xe_ring_ops.c        |  5 +-
 drivers/gpu/drm/xe/xe_sched_job.c       |  1 +
 drivers/gpu/drm/xe/xe_sched_job_types.h |  2 +
 4 files changed, 57 insertions(+), 19 deletions(-)

-- 
2.34.1


^ permalink raw reply	[flat|nested] 18+ messages in thread

end of thread, other threads:[~2024-12-20  0:19 UTC | newest]

Thread overview: 18+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2024-12-19  0:23 [PATCH 0/2] Fixes for MI_REPORT_PERF_COUNT Umesh Nerlige Ramappa
2024-12-19  0:23 ` [PATCH 1/2] xe/oa: Fix query mode of operation for OAR/OAC Umesh Nerlige Ramappa
2024-12-19 23:15   ` Dixit, Ashutosh
2024-12-20  0:19     ` Umesh Nerlige Ramappa
2024-12-19  0:23 ` [PATCH 2/2] xe/oa: Drop the unused logic to parse context image Umesh Nerlige Ramappa
2024-12-19  2:31 ` ✓ CI.Patch_applied: success for Fixes for MI_REPORT_PERF_COUNT (rev3) Patchwork
2024-12-19  2:31 ` ✓ CI.checkpatch: " Patchwork
2024-12-19  2:32 ` ✓ CI.KUnit: " Patchwork
2024-12-19  2:50 ` ✓ CI.Build: " Patchwork
2024-12-19  2:53 ` ✓ CI.Hooks: " Patchwork
2024-12-19  2:54 ` ✓ CI.checksparse: " Patchwork
2024-12-19  3:27 ` ✓ Xe.CI.BAT: " Patchwork
2024-12-19 23:44 ` ✗ Xe.CI.Full: failure " Patchwork
  -- strict thread matches above, loose matches on Subject: below --
2024-12-17  0:58 [PATCH 0/2] Fixes for MI_REPORT_PERF_COUNT Umesh Nerlige Ramappa
2024-12-17 17:05 ` Souza, Jose
2024-12-17 17:13   ` Souza, Jose
2024-12-17 18:51     ` Umesh Nerlige Ramappa
2024-11-27  0:31 Umesh Nerlige Ramappa

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox