WORKAROUND APPLIED: nohz=off

The panic has been reproduced on kernels 7.0.0-28-generic and 7.0.0-29-generic
(Ubuntu 26.04). As a temporary workaround I added "nohz=off" to the kernel
command line (GRUB). The workaround is confirmed active:

  # cat /proc/cmdline
  BOOT_IMAGE=/boot/vmlinuz root=UUID=... ro quiet nohz=off 
systemd.unit=multi-user.target

With nohz=off, the failing code path (_nohz_idle_balance ->
sched_balance_update_blocked_averages -> raw_spin_rq_lock_nested ->
native_queued_spin_lock_slowpath) is no longer exercised, and the machine
has not panicked since.

Root cause hypothesis: UBSAN array-index-out-of-bounds in
kernel/locking/qspinlock.h:68 (index -1 out of range for
'long unsigned int [8192]') during NOHZ idle balance, in interrupt
context (swapper/N), leading to a page fault in
native_queued_spin_lock_slowpath and a kernel panic.

Hardware: Intel NUC7i7BNH (i7-7567U, 4 cores).
Full pstore dumps attached in kernel-bug-evidence.tar.gz.

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2163292

Title:
  The machine dies periodically (1-7 day intervals, usually at idle,
  often at night). EFI pstore dumps captured on 4 separate occasions all
  show the IDENTICAL kernel panic

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2163292/+subscriptions


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

Reply via email to