mbox series

[v4,0/1] drm/i915/guc: Don't update engine busyness stats too frequently

Message ID 20220623023157.211650-1-alan.previn.teres.alexis@intel.com (mailing list archive)
Headers show
Series drm/i915/guc: Don't update engine busyness stats too frequently | expand

Message

Alan Previn June 23, 2022, 2:31 a.m. UTC
This change ensures we don't resample GuC busyness stat
counters too soon after the last sample.

Prior receipts of rvb's:
  Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>

Changes from prior revs:
  v3:  - Added Umesh's Rvb into patch
  v2:  - Align the name of last_stat_jiffies (Tvrtko)
       - Use 32-bit jiffes (Tvrtko)
       - use time_after() macro (Tvrtko)
       - change from ">> 1" to "/ 2" for ping delay halving (Tvrtko)
  v1:  - Move the location of the new logic to higher up
         the callstack in intel_guc_busyness_park (Umesh).
       - Fix typo threshold of jiffies-since-last from double
         to half of ping_delay (Umesh)

Alan Previn (1):
  drm/i915/guc: Don't update engine busyness stats too frequently

 drivers/gpu/drm/i915/gt/uc/intel_guc.h            |  8 ++++++++
 drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c | 13 +++++++++++++
 2 files changed, 21 insertions(+)


base-commit: ac17a5249380aaabe5d1eaebd9b3a2eedc08ccdc

Comments

Alan Previn June 27, 2022, 4:56 p.m. UTC | #1
WRT possible new issues below, the patches i made does not impact any display or even execution operation.
Unfortunately, i cant seem to access the "possible regression" failure below. Additionally, GuC is not supported on SKL.
thus we can safely ignore these.


On Mon, 2022-06-27 at 12:47 +0000, Patchwork wrote:
Patch Details
Series: drm/i915/guc: Don't update engine busyness stats too frequently
URL:    https://patchwork.freedesktop.org/series/105525/
State:  failure
Details:

...
...
Possible new issues

Here are the unknown changes that may have been introduced in Patchwork_105525v1_full:

IGT changes
Possible regressions

  *   igt@kms_cursor_legacy@flip-vs-cursor@varying-size:
     *   shard-skl: NOTRUN -> FAIL<https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_105525v1/shard-skl7/igt@kms_cursor_legacy@flip-vs-cursor@varying-size.html> +3 similar issues

Suppressed

The following results come from untrusted machines, tests, or statuses.
They do not affect the overall result.

  *   {igt@kms_cursor_crc@cursor-offscreen@pipe-d-hdmi-a-1-32x10}:
     *   {shard-dg1}: NOTRUN -> SKIP<https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_105525v1/shard-dg1-15/igt@kms_cursor_crc@cursor-offscreen@pipe-d-hdmi-a-1-32x10.html> +15 similar issues