pfSense 2.6.0 crashes Panic String: sleeping thread

Yves_

Hi there,

I am not 100% sure if this correlates but it seams to happen if the ISP gives me a new IP. Like I said not 100% sure...

What happens is I get tons and tons of:

<7>arpresolve: can't allocate llinfo for 46.126.232.1 on igb0
<7>arpresolve: can't allocate llinfo for 46.126.232.1 on igb0
<7>arpresolve: can't allocate llinfo for 46.126.232.1 on igb0
<7>arpresolve: can't allocate llinfo for 46.126.232.1 on igb0
<7>arpresolve: can't allocate llinfo for 46.126.232.1 on igb0
<7>arpresolve: can't allocate llinfo for 46.126.232.1 on igb0
<7>arpresolve: can't allocate llinfo for 46.126.232.1 on igb0
<7>arpresolve: can't allocate llinfo for 46.126.232.1 on igb0

later some:

<6>igb0: link state changed to UP
<6>igb0: link state changed to DOWN
<7>arpresolve: can't allocate llinfo for 46.126.232.1 on igb0
<7>arpresolve: can't allocate llinfo for 46.126.232.1 on igb0
<7>arpresolve: can't allocate llinfo for 46.126.232.1 on igb0
<7>arpresolve: can't allocate llinfo for 46.126.232.1 on igb0
<7>arpresolve: can't allocate llinfo for 46.126.232.1 on igb0
<7>arpresolve: can't allocate llinfo for 46.126.232.1 on igb0

and later we crash:

<7>arpresolve: can't allocate llinfo for 46.126.232.1 on igb0
<6>ovpns1: link state changed to DOWN
<6>ovpns1: link state changed to UP
Sleeping thread (tid 100277, pid 87983) owns a non-sleepable lock
KDB: stack backtrace of thread 100277:
sched_switch() at sched_switch+0x630/frame 0xfffffe008c36b710
mi_switch() at mi_switch+0xd4/frame 0xfffffe008c36b740
sleepq_wait() at sleepq_wait+0x2c/frame 0xfffffe008c36b770
_sx_xlock_hard() at _sx_xlock_hard+0x3ee/frame 0xfffffe008c36b820
iflib_if_ioctl() at iflib_if_ioctl+0x9cc/frame 0xfffffe008c36b890
if_allmulti() at if_allmulti+0x85/frame 0xfffffe008c36b8e0
X_ip_mrouter_set() at X_ip_mrouter_set+0x17b4/frame 0xfffffe008c36b9b0
rip_ctloutput() at rip_ctloutput+0xf3/frame 0xfffffe008c36b9e0

which would be okayish... but since I always need to manually add my 25GB Broadcom patch

ifconfig bnxt0 promisc

firewall goes down hard and I can not do anything until I have physical access to it...

I had this twice within the last 7 days. Which is a bit painfull...

Any ideas to workaround? I read that it can happen because I monitor the gateway of the igb0 but does this not work automagicly ;-) that if I get an IP in another range it starts monitoring the new gateway?

Thanks for your help guys

Cheers,
Yves

stephenw10

Do you have a full crash report?

Are you applying that setting to the bntx NIC manually at every boot? Have you tried using a shellcmd instead?
https://docs.netgate.com/pfsense/en/latest/development/boot-commands.html#shellcmd-option

Steve

Yves_

@stephenw10 Hi Steve, yes I do. But I did not want to publicly post it. Is there a way that I can get it to you directly?

stephenw10

Sure, you can upload it here: https://nc.netgate.com/nextcloud/s/2tqXwS5nHw6M9ww

Steve

Yves_

@stephenw10: Done

stephenw10

Great, I have those. Checking them now.

Yves_

@stephenw10 Hi Steve, did you figure something out already? Cheers Yves

stephenw10

It's not anything familiar to me. I have asked internally.

stephenw10

It looks like it's something in multicast routing though which is probably why more people are not hitting it.
I don't see either IGMPproxy or pimd shown in your console output but do you have either configured?
It could be same root cause as this: https://redmine.pfsense.org/issues/12079
It's very difficult to reproduce that issue unfortunately.

Steve