gnome-shell crashed (out of memory) with SIGSEGV in crocus_begin_query() from crocus_begin_query() from crocus_end_query() from crocus_end_query() from tc_call_end_query()

Status changed to 'Confirmed' because the bug affects multiple users.

Upstream says the crash is caused by a leak and reviewing the Mesa code seems to support that theory. The fix is:

https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17731

Only fixed in Mesa 23 right now but I would expect to see it in 22.x

I find it confusing to move a bug 'fix committed' when it's only in upstream master. Kinetic will get 22.2.x, and it'll have this (hasn't branched yet).

I agree... it's just desktop team policy requested by seb128.

I misinterpreted mesa's branching structure. The commit is indeed in 22.2 (not tagged yet).

This is now the top gnome-shell crasher in jammy, which is impressive for the quality of gnome-shell because the crash only occurs on Intel 4th gen and older.

An upload of mesa to jammy-proposed has been rejected from the upload queue for the following reason: "Looks like there's unintentional changes in src/mesa/formats.csv".

Hello errors.ubuntu.com, or anyone else affected,

Accepted mesa into jammy-proposed. The package will build now and be available at https://launchpad.net/ubuntu/+source/mesa/22.0.5-0ubuntu0.2 in a few hours, and then in the -proposed repository.

Please help us by testing this new package. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation on how to enable and use -proposed. Your feedback will aid us getting this update out to other Ubuntu users.

If this package fixes the bug for you, please add a comment to this bug, mentioning the version of the package you tested, what testing has been performed on the package and change the tag from verification-needed-jammy to verification-done-jammy. If it does not fix the bug for you, please add a comment stating that, and change the tag to verification-failed-jammy. In either case, without details of your testing we will not be able to proceed.

Further information regarding the verification process can be found at https://wiki.ubuntu.com/QATeam/PerformingSRUVerification . Thank you in advance for helping!

N.B. The updated package will be released to -updates after the bug(s) fixed by this package have been verified and the package has been in -proposed for a minimum of 7 days.

All autopkgtests for the newly accepted mesa (22.0.5-0ubuntu0.2) for jammy have finished running.
The following regressions have been reported in tests triggered by the package:

mutter/42.5-0ubuntu1 (arm64)

Please visit the excuses page listed below and investigate the failures, proceeding afterwards as per the StableReleaseUpdates policy regarding autopkgtest regressions [1].

https://people.canonical.com/~ubuntu-archive/proposed-migration/jammy/update_excuses.html#mesa

[1] https://wiki.ubuntu.com/StableReleaseUpdates#Autopkgtest_Regressions

Thank you!

vanvugt: any way to verify this fix?

I've only been monitoring crash reports to verify the fix worked in kinetic and later. Although anecdotally I recall one or two users of Haswell or older saying it happened frequently for them. I guess we could try such an older system and monitor for the gnome-shell memory leak that should precede the crash.

(Sorry for the delay - it was a long end of year break)

or just skip this and accept the kinetic backport which has the fix via upstream

I have promised to actually test this on some older hardware but I want to get through the past month's bug mail first...

Tested jammy with crocus on an i5-4300U, both 22.0.5-0ubuntu0.1 and 22.0.5-0ubuntu0.2

gnome-shell's RSS grows but seems to plateau similarly with both.

Based on upstream's comments, I started counting i915.gem instances in /proc/PID/smaps and found that both 22.0.5-0ubuntu0.1 and 22.0.5-0ubuntu0.2 seem to leak them equally. Never reached the point of crashing but the evidence is that mesa 22.0.5-0ubuntu0.2 makes gnome-shell leak almost as much as 22.0.5-0ubuntu0.1 ... That said, I expect most graphics leaks to be hardware independent and the fault of gnome-shell/mutter, not mesa.

Looking at the upstream fix again I realize it's not fixing all the leaks that's important here, but avoiding crashes when they occur. And for that we are missing the correct patch:

https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17615/diffs

It's not in mesa_22.0.5-0ubuntu0.2.diff.gz

but it is in 22.2.x which we need in proposed asap anyway, so I'll drop this from the current upload so that it can be fast-tracked through, as #1998893 is even more urgent.

or, just accept the current upload and revisit it again with the 22.2 backport if needed, but in any case I don't think this bug should block either one from getting into updates

Yeah kinetic doesn't have this bug so we can probably fix jammy via the backport without knowing which commits are most important.

Hello errors.ubuntu.com, or anyone else affected,

Accepted mesa into jammy-proposed. The package will build now and be available at https://launchpad.net/ubuntu/+source/mesa/22.0.5-0ubuntu0.3 in a few hours, and then in the -proposed repository.

Please help us by testing this new package. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation on how to enable and use -proposed. Your feedback will aid us getting this update out to other Ubuntu users.

If this package fixes the bug for you, please add a comment to this bug, mentioning the version of the package you tested, what testing has been performed on the package and change the tag from verification-needed-jammy to verification-done-jammy. If it does not fix the bug for you, please add a comment stating that, and change the tag to verification-failed-jammy. In either case, without details of your testing we will not be able to proceed.

Further information regarding the verification process can be found at https://wiki.ubuntu.com/QATeam/PerformingSRUVerification . Thank you in advance for helping!

N.B. The updated package will be released to -updates after the bug(s) fixed by this package have been verified and the package has been in -proposed for a minimum of 7 days.

I accepted 22.0.5-0ubuntu0.3 which does NOT fix this bug, so I'm reopening it.

This bug was fixed in the package mesa - 22.0.5-0ubuntu0.3

---------------
mesa (22.0.5-0ubuntu0.3) jammy; urgency=medium

  * crocus-fix-leak-in-query-code.patch: Drop the patch so the other fix
    can get through to updates.

mesa (22.0.5-0ubuntu0.2) jammy; urgency=medium

  * crocus-fix-leak-in-query-code.patch: Fix gnome-shell crashing on
    older Intel hw. (LP: #1972977)
  * patches: Update pci-id's for ADL-S/RPL-S. (LP: #1998893)

 -- Timo Aaltonen <email address hidden> Thu, 05 Jan 2023 10:45:40 +0200

nope, reopening so that this will be fixed via the kinetic backport

While 22.04 isn't fixed yet, please try editing /etc/environment and add:

  MESA_LOADER_DRIVER_OVERRIDE=i965

then reboot. It might be a suitable workaround.

is fixed for real now