Bug#538158: BUG: soft lockup - CPU#5 stuck for 62s with 2.6.26-2-686-bigmem kernel

[Jonathan Nieder <jrnieder@gmail.com>]

Hi,

Juan Miguel Corral Cano wrote:

> Hi. I had the same problem on several machines of my network (all of them
> using nfs-mounted homes). We installed kernel 2.6.32-2 from lenny-backports
> a couple months ago, and no hang since then.
>
> Hope it is helpful. Best regards. Juan.

Thanks, that _is_ helpful.

Unfortunately none of the patches between 2.6.26 and 2.6.32 jumps out
at me as likely to have fixed this.  So I would like more help, if you
are still interested (and please let me know if you are not).  Quick
questions:

 - Arcady, are you still experiencing the soft lockups in default_idle?
   If so, can you confirm Juan's finding that 2.6.32 from
   lenny-backports avoids trouble?

 - Juan, do you still have machines that can be convinced to produce
   the soft lockups again?  Could you send a log snippet (starting at
   BUG and ending at =======================) for such a soft lockup,
   either by reproducing it again or by extracting it from 2010's
   kernel logs if you still have them?

 - Assuming the lenny kernel is still broken and the lenny-backports
   kernel fixed, a useful next step would be to try a kernel from
   between (2.6.30 or so) from snapshot.debian.org to help narrow down
   the search for the fix.

Cheers,
Jonathan