[fedora-arm] BUG: scheduling while atomic: swapper/0/0/0x40000100

[Jon Masters]

On 10/25/2012 06:23 PM, Jon Masters wrote:
> On 10/24/2012 01:04 PM, Jon wrote:
> 
>> I've been seeing lots and lots of these OOPS on Pandaboard ES with
>> recent'ish f18 image.
>> By lots and lots I mean hundreds per second, causes log spam, and
>> eventually eats all the inodes in /var/spool/abrt making reports.
> 
> Should be fixed by the patch I posted for David to test. David, Peter,
> can we get the official kernel spun with this. It's going to be upstream
> shortly, has been ACKed by folks. Just swap the revert patch for the
> atomic fixup patch I posted if possible.

Here's the upstream ACK:

http://lists.infradead.org/pipermail/linux-arm-kernel/2012-October/127317.html

This should be in a kernel near you soonish, so I really see no problem
pulling into our 3.6. After all, it's more official than the revert
patch we already have as the real solution to this problem.

Separately, there are some unrelated highbank network performance fixes,
but as you can see from this mail from Jon, it's not highbank specific
the alignment fault handler triggering these BUGs.

Jon.