https://intel-gfx-ci.01.org/tree/drm-tip/IGT_4882/shard-glk4/igt@kms_plane_multiple@atomic-pipe-a-tiling-yf.html Starting subtest: atomic-pipe-A-tiling-yf (kms_plane_multiple:2399) igt_debugfs-CRITICAL: Test assertion failure function igt_assert_crc_equal, file ../lib/igt_debugfs.c:419: (kms_plane_multiple:2399) igt_debugfs-CRITICAL: Failed assertion: !mismatch (kms_plane_multiple:2399) igt_debugfs-CRITICAL: Last errno: 25, Inappropriate ioctl for device Subtest atomic-pipe-A-tiling-yf failed.
This is not seen on the legacy path, so marking as an atomic issue.
The CI Bug Log issue associated to this bug has been updated. ### New filters associated * all machines: igt@kms_plane_multiple@atomic-pipe-[abc]-* - fail - Failed assertion: !mismatch, Last errno: 25, Inappropriate ioctl for device (No new failures associated)
A CI Bug Log filter associated to this bug has been updated: {- all machines: igt@kms_plane_multiple@atomic-pipe-[abc]-* - fail - Failed assertion: !mismatch, Last errno: 25, Inappropriate ioctl for device -} {+ all machines: igt@kms_plane_multiple@atomic-pipe-[abc]-* / igt@kms_universal_plane@universal-plane-pipe-* - fail - Failed assertion: !mismatch, Last errno: 25 +} No new failures caught with the new filter
*** Bug 110114 has been marked as a duplicate of this bug. ***
This tests sets all the non-primary planes to be 256x256 (cursor to 128x128) and randomizes their positions, but with a limitation that they can't overhang the visible area. Then for that data it prepares primary plane, spanning the whole screen, also filled with the same blue color and black holes for each of the other planes, so that the holes would be covered by the overlay planes. The final result, if blended correctly, should be a solid color on the whole pipe. We compare the CRC against a CRC we gathered from solid primary plane with nothing overlayed. This test failing may mean there are inaccuracies in blending planes. User impact: small, as not much software is using multiple planes, and we have other tests that seem to be fine with 2 planes / cursor. Recommended next steps: extra logging for plane positions and maybe manual verification/Chamelium capturing the broken frames for manual inspection.
A CI Bug Log filter associated to this bug has been updated: {- all machines: igt@kms_plane_multiple@atomic-pipe-[abc]-* / igt@kms_universal_plane@universal-plane-pipe-* - fail - Failed assertion: !mismatch, Last errno: 25 -} {+ all machines: igt@kms_plane_multiple@atomic-pipe-[abc]-* / igt@kms_universal_plane@universal-plane-pipe-* - fail - Failed assertion: !mismatch +} No new failures caught with the new filter
Seems like the reproduction rate went down. We have seen it ~10 times each drmtip run (https://intel-gfx-ci.01.org/#idle-runs) but drmtip_260 and _261 are clean. drmtip_262 and _263 have seen it only one.
In case of ICL we run the test all 7 planes enabled, we may run out of bandwidth. This causes fifo underruns and mismatches.
The bug hasn't appeared lately in shards. Closing this one now. Let's reopen if the issue comes back.
Yeah, it used to happen once every run, and now nothing for the past 72 runs. Closing!
The CI Bug Log issue associated to this bug has been archived. New failures matching the above filters will not be associated to this bug anymore.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.