Peter Jeremy via freebsd-net
2021-04-30 04:57:29 UTC
I've had the following panic & traceback occur twice with different
recent 12-stable kernels. I hadn't see that before I updated from
r366967 to a mid-April 12-stable. Doing some searching turns up some
older similar issues but they should all be fixed well before r366967.
Looking though commits since then doesn't turn up anything suspicious.
I don't believe there's anything particularly unusual about this system:
All the network traffic (and there shouldn't have been much) would be
going via one active em(4) NIC. That NIC has one VLAN associated with
it (though there'd be no traffic on it). There are also inactive
tap and tun NICs and a second, inactive em NIC.
Has anyone seen anything like this or have any ideas?
panic: Assertion in_epoch(net_epoch_preempt) failed at /usr/src/sys/net/if.c:2228
KDB: stack backtrace:
db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame 0xfffffe020aaf0600
vpanic() at vpanic+0x17b/frame 0xfffffe020aaf0650
panic() at panic+0x43/frame 0xfffffe020aaf06b0
ifaof_ifpforaddr() at ifaof_ifpforaddr+0x142/frame 0xfffffe020aaf06f0
ip_redir_alloc() at ip_redir_alloc+0x1f6/frame 0xfffffe020aaf0780
ip_tryforward() at ip_tryforward+0x3e4/frame 0xfffffe020aaf0800
ip_input() at ip_input+0x4a0/frame 0xfffffe020aaf08b0
netisr_dispatch_src() at netisr_dispatch_src+0x9b/frame 0xfffffe020aaf0910
ether_demux() at ether_demux+0x156/frame 0xfffffe020aaf0940
ether_nh_input() at ether_nh_input+0x3f3/frame 0xfffffe020aaf09a0
netisr_dispatch_src() at netisr_dispatch_src+0x9b/frame 0xfffffe020aaf0a00
ether_input() at ether_input+0x86/frame 0xfffffe020aaf0a30
_task_fn_rx() at _task_fn_rx+0xc4a/frame 0xfffffe020aaf0b20
gtaskqueue_run_locked() at gtaskqueue_run_locked+0x135/frame 0xfffffe020aaf0b80
gtaskqueue_thread_loop() at gtaskqueue_thread_loop+0x94/frame 0xfffffe020aaf0bb0
fork_exit() at fork_exit+0x80/frame 0xfffffe020aaf0bf0
fork_trampoline() at fork_trampoline+0xe/frame 0xfffffe020aaf0bf0
--- trap 0, rip = 0, rsp = 0, rbp = 0 ---
recent 12-stable kernels. I hadn't see that before I updated from
r366967 to a mid-April 12-stable. Doing some searching turns up some
older similar issues but they should all be fixed well before r366967.
Looking though commits since then doesn't turn up anything suspicious.
I don't believe there's anything particularly unusual about this system:
All the network traffic (and there shouldn't have been much) would be
going via one active em(4) NIC. That NIC has one VLAN associated with
it (though there'd be no traffic on it). There are also inactive
tap and tun NICs and a second, inactive em NIC.
Has anyone seen anything like this or have any ideas?
panic: Assertion in_epoch(net_epoch_preempt) failed at /usr/src/sys/net/if.c:2228
KDB: stack backtrace:
db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame 0xfffffe020aaf0600
vpanic() at vpanic+0x17b/frame 0xfffffe020aaf0650
panic() at panic+0x43/frame 0xfffffe020aaf06b0
ifaof_ifpforaddr() at ifaof_ifpforaddr+0x142/frame 0xfffffe020aaf06f0
ip_redir_alloc() at ip_redir_alloc+0x1f6/frame 0xfffffe020aaf0780
ip_tryforward() at ip_tryforward+0x3e4/frame 0xfffffe020aaf0800
ip_input() at ip_input+0x4a0/frame 0xfffffe020aaf08b0
netisr_dispatch_src() at netisr_dispatch_src+0x9b/frame 0xfffffe020aaf0910
ether_demux() at ether_demux+0x156/frame 0xfffffe020aaf0940
ether_nh_input() at ether_nh_input+0x3f3/frame 0xfffffe020aaf09a0
netisr_dispatch_src() at netisr_dispatch_src+0x9b/frame 0xfffffe020aaf0a00
ether_input() at ether_input+0x86/frame 0xfffffe020aaf0a30
_task_fn_rx() at _task_fn_rx+0xc4a/frame 0xfffffe020aaf0b20
gtaskqueue_run_locked() at gtaskqueue_run_locked+0x135/frame 0xfffffe020aaf0b80
gtaskqueue_thread_loop() at gtaskqueue_thread_loop+0x94/frame 0xfffffe020aaf0bb0
fork_exit() at fork_exit+0x80/frame 0xfffffe020aaf0bf0
fork_trampoline() at fork_trampoline+0xe/frame 0xfffffe020aaf0bf0
--- trap 0, rip = 0, rsp = 0, rbp = 0 ---
--
Peter Jeremy
Peter Jeremy