Re: r194546 amd64: kernel panic in tcp_sack.c

Kamigishi Rei wrote:
> Lawrence Stewart wrote:
>> hmmm, handy. What is generating the TCP load on this machine? Serving 
>> to clients on the Internet? Or machines on a local lan? Is there lots 
>> of load? Little bit?
> I've forgot to mention that in both cases with this tcp_sack conditional 
> panic the uptime was approximately 40 hours.

You didn't mention whether the clients being served were local lan or 
remote. I ask because tcp_timer_rexmt() is in the backtrace indicating 
you're getting TCP RTOs firing on the connection that's triggering the 
panic.

Cheers,
Lawrence