Panic on kern_event.c

[Mark Johnston]

On Tue, Nov 06, 2018 at 10:50:06AM +0100, Sylvain GALLIANO wrote:
> Hi,
> 
> I got random panic on Current & 11.2-STABLE on kern_event.c
> 
> Panic occur in syslog-ng (logging at high rate) with the folloging lines:
> 
>   Panic String: Bad tailq NEXT(0xfffff80039ae7a38->tqh_last) != NULL
>   Panic String: Bad tailq head 0xfffff80039f1a238 first->prev != head
> 
> It's look like knote_enqueue try to add and existings knote on TAILQ
> (confirmed by following patch).
> 
> logs after apply patch:
> XXX knote 0xfffff8012e3d33c0 already in tailq  status:1 kq_count:1  [0
> 0xfffff800327d3538]  2671
> XXX knote 0xfffff80032861780 already in tailq  status:1 kq_count:1  [0
> 0xfffff80032457938]  2671

Can you grab the stack when this happens as well, with kdb_backtrace()?
Or better, convert the print into a panic so that we can examine the
kernel dump.