On 2/12/24 13:44, Michael Butler wrote:
On 2/12/24 12:36, John Baldwin wrote:
[ .. trimmed .. ]
Short of a stack trace, you can at least use lldb or gdb to lookup
the source line associated with the faulting instruction pointer (as
long as it isn't in a kernel module), e.g. for gdb you would use 'gdb
/boot/kernel/kernel' and then 'l *<instruction pointer address>',
e.g. from above: 'l *0xffffffff80acb962'
I still didn't manage to get a core but .. does this make any sense in
htis context?
I apologize .. too many crashes and I grabbed the wrong instruction
pointer; this was the most recent attempt. I have no idea why this
networking code and PCI configurations are seemingly related :-(
(kgdb) l *0xffffffff80acbc02
0xffffffff80acbc02 is in cc_cong_signal
(/usr/src/sys/netinet/tcp_input.c:465).
460 tp->t_flags &= ~TF_PREVVALID;
461 tp->t_badrxtwin = 0;
462 break;
463 }
464
465 if (CC_ALGO(tp)->cong_signal != NULL) {
466 if (th != NULL)
467 tp->t_ccv.curack = th->th_ack;
468 CC_ALGO(tp)->cong_signal(&tp->t_ccv, type);
469 }
