Built a local test kernel (7.0.0-30.999, based on 7.0.0-30.30) with
Thomas Hellström's "drm/ttm: Represent LRU bulk moves as nested
sublists" patch
(lore.kernel.org/all/[email protected])
applied on top.

Installed and running for ~24 hours now, covering:
- Multiple plain suspend/resume cycles
- Multiple suspend-then-hibernate cycles (including full hibernation, not just 
the delay-then-suspend path)
- Normal desktop use, including GPU-heavy workloads right after resume (the 
second crash I originally reported was triggered this way, without hibernation 
involved)

No recurrence of the ttm_lru_bulk_move_tail NULL pointer dereference so
far — previously this was crashing multiple times a day. journalctl -k
shows no TTM/amdgpu BUG/Oops entries across the whole uptime.

Will keep running it and report back if anything turns up, but this is
looking like a strong fix. Happy to keep testing or try anything else
that would help get this verified for backport.

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2163363

Title:
  [amdgpu/ttm] TTM list corruption and NULL dereference under GPU memory
  pressure on ASUS ProArt PX13 (7.0.0-29)

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2163363/+subscriptions


-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to