[Bug 268246] crash and panic using pfsync on 13.1-RELEASE

From: <bugzilla-noreply_at_freebsd.org>
Date: Mon, 20 Mar 2023 08:33:45 UTC
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=268246

--- Comment #93 from Kristof Provost <kp@freebsd.org> ---
(In reply to mikael.le-bohec from comment #92)
You're tripping over a different panic. (And it would have been helpful to
actually include the backtrace in your report.)

> panic: _mtx_lock_sleep: recursed on non-recursive mutex pfsync bucket @ /usr/src/sys/netpfil/pf/if_pfsync.c:2295
> 
> cpuid = 0
> time = 1679262157
> KDB: stack backtrace:
> db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame 0xfffffe00c72cbc50
> vpanic() at vpanic+0x152/frame 0xfffffe00c72cbca0
> panic() at panic+0x43/frame 0xfffffe00c72cbd00
> __mtx_lock_sleep() at __mtx_lock_sleep+0x43d/frame 0xfffffe00c72cbd90
> __mtx_lock_flags() at __mtx_lock_flags+0xe5/frame 0xfffffe00c72cbde0
> pfsync_timeout() at pfsync_timeout+0x64/frame 0xfffffe00c72cbe10
> softclock_call_cc() at softclock_call_cc+0x14b/frame 0xfffffe00c72cbec0
> softclock_thread() at softclock_thread+0xc6/frame 0xfffffe00c72cbef0
> fork_exit() at fork_exit+0x80/frame 0xfffffe00c72cbf30
> fork_trampoline() at fork_trampoline+0xe/frame 0xfffffe00c72cbf30
> --- trap 0, rip = 0, rsp = 0, rbp = 0 ---
> KDB: enter: panic

Which I can reproduce but do not understand. There's no obvious way for us to
end up recursing on the bucket lock there, nor is there a clear path where we
could fail to unlock a bucket.

-- 
You are receiving this mail because:
You are the assignee for the bug.