https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=279899
--- Comment #14 from Zhenlei Huang <z...@freebsd.org> --- (In reply to Franco Fichtner from comment #13) > Bisecting this is a little tricky due to the random nature but I think it's > the > backport of > > https://cgit.freebsd.org/src/commit/?id=2671bde99295d9 > > which now runs pfsync_drop() which is what happens when you don't use pfsync > which > didn't happen before as the commit suggests. Yes, your analysis is right. > INVARIANTS appears to trip over: > > panic: pfsync_drop: st->sync_state == q > > without INVARIANTS it just continues and let's this crash at > pf_unlink_state() later on. Yes. Exactly. > Maybe it's wrong but I also don't appreciate the lack of enthusiasm looking > into > fixing regressions on stable branches. At least there is enough information > on the > table already to do something but I don't see any engagement. > > Cheers, > Franco Sorry for that. I have busy days. I tried to repeat but failed. Maybe I have wrong configuration of if_pfsync. May Daniel, Gyver, or you share a minimal setup? This should be repeated also on main IMO. -- You are receiving this mail because: You are the assignee for the bug.
