Panic in tcp_input

[Andre Oppermann]

On 19.06.2013 11:34, Ian FREISLICH wrote:
> Andre Oppermann wrote:
>> On 19.06.2013 11:10, Ian FREISLICH wrote:
>>> Hi
>>>
>>> I'm seeing this panic quite regularly now.  Most recent sighting on r251858.
>>
>> This panic message is not very informative and very hard to extract any
>> meaningful hints.  Do you have a core dump and matching debug kernel?
>
> I do.  If you can send me your public ssh key in private email, I
> can give you access to the server in question.

It bombs while accessing a per-cpu counter:

(kgdb) frame 11
#11 0xffffffff8054518e in tcp_input (m=<value optimized out>, off0=Cannot access
memory at address 0x14)
at counter.h:45
45              __asm __volatile("addq\t%1,%%gs:(%0)"

This seems to be a fallout of the recent UMA changes and its interaction
with the per-cpu counter(9) system.

Adding in and handing over to Gleb and Jeff as they touched this area last.

-- 
Andre