Re: Panic in tcp_input

On Wed, Jun 19, 2013 at 02:20:14PM +0200, Andre Oppermann wrote:
A> On 19.06.2013 11:34, Ian FREISLICH wrote:
A> > Andre Oppermann wrote:
A> >> On 19.06.2013 11:10, Ian FREISLICH wrote:
A> >>> Hi
A> >>>
A> >>> I'm seeing this panic quite regularly now.  Most recent sighting on r251858.
A> >>
A> >> This panic message is not very informative and very hard to extract any
A> >> meaningful hints.  Do you have a core dump and matching debug kernel?
A> >
A> > I do.  If you can send me your public ssh key in private email, I
A> > can give you access to the server in question.
A> 
A> It bombs while accessing a per-cpu counter:
A> 
A> (kgdb) frame 11
A> #11 0xffffffff8054518e in tcp_input (m=<value optimized out>, off0=Cannot access
A> memory at address 0x14)
A> at counter.h:45
A> 45              __asm __volatile("addq\t%1,%%gs:(%0)"
A> 
A> This seems to be a fallout of the recent UMA changes and its interaction
A> with the per-cpu counter(9) system.
A> 
A> Adding in and handing over to Gleb and Jeff as they touched this area last.

Ian has reported this panic several days before Jeff's changes :(

-- 
Totus tuus, Glebius.