Possible duplicate / same regression: I filed bug #2169477 (https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2169477) on 2026-10-04 for what looks like the same defect -- same boundary (6.8.0-142 good, 6.8.0-146 bad, noble), same refcount_t "addition on 0; use-after-free" on an L2TP session object, same L2TP-over-IPsec trigger.
Your trace surfaces in l2tp_tunnel_get_session() via xfrm_trans_reinject(); on my host the only changed l2tp object between -142 and -146 is l2tp_ppp.ko (l2tp_core.ko is identical), and I suspect the session refcount handling in pppol2tp_ioctl() touched by the CVE-2026-53262 backport. My report has 11 EFI pstore dumps attached if they are useful for comparison. One data point that may narrow the trigger: the hang only occurs once an L2TP session is actually established. Behind NAT, where strongSwan never completes and no session is created, the same box on the same -146 kernel stayed up 12-17 minutes with no splat; on a public uplink with ppp0 up at ~19 s it died 46 s and 108 s into boot. ** CVE added: https://cve.org/CVERecord?id=CVE-2026-53262 -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/2169269 Title: l2tp_core: refcount_t "addition on 0; use-after-free" in l2tp_tunnel_get_session on every L2TP/IPsec connect, system freezes (regression 6.8.0-142 -> 6.8.0-146, noble) To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2169269/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
