[PATCH i-g-t 0/2] Fix needing to call igt_display_require() twice to enable writeback output

Arthur Grillo arthurgrillo at riseup.net
Thu Feb 15 19:21:03 UTC 2024


After discussing with Pekka Paalanen about needing to call
igt_display_require() twice[1], I found the issue.

As stated on the docs[2]. To enable DRM_CLIENT_CAP_WRITEBACK_CONNECTORS,
we first need to set DRM_CLIENT_CAP_ATOMIC.

Patch #1 solves the bug and #2 uses the fix where the duplicate call is
done. I couldn't run kms_writeback (VKMS already fails on this test and
my GPU doesn't have writeback outputs), so please test it.

[1]; https://lore.kernel.org/dri-devel/20240212104330.7de30010@eldfell/
[2]: https://docs.kernel.org/gpu/drm-uapi.html?highlight=writeback#c.DRM_CLIENT_CAP_WRITEBACK_CONNECTORS

Cc: Petri Latvala <adrinael at adrinael.net>
Cc: Arkadiusz Hiler <arek at hiler.eu>
Cc: Kamil Konieczny <kamil.konieczny at linux.intel.com>
Cc: Juha-Pekka Heikkila <juhapekka.heikkila at gmail.com>
Cc: Bhanuprakash Modem <bhanuprakash.modem at intel.com>
Cc: Ashutosh Dixit <ashutosh.dixit at intel.com>
Cc: Pekka Paalanen <pekka.paalanen at haloniitty.fi>
To: igt-dev at lists.freedesktop.org

Signed-off-by: Arthur Grillo <arthurgrillo at riseup.net>
---
Arthur Grillo (2):
      lib/igt_kms: Enable DRM_CLIENT_CAP_ATOMIC before enabling DRM_CLIENT_CAP_WRITEBACK_CONNECTORS
      igt: Remove duplicate igt_display_require() calls

 benchmarks/kms_fb_stress.c | 2 --
 lib/igt_kms.c              | 5 +++--
 tests/kms_writeback.c      | 1 -
 3 files changed, 3 insertions(+), 5 deletions(-)
---
base-commit: 9ee96f431b2baf552e95c9d839cdfbcdae2c8b41
change-id: 20240215-duplicate-calls-ae33c2a855ed

Best regards,
-- 
Arthur Grillo <arthurgrillo at riseup.net>



More information about the igt-dev mailing list