Kernel 2.6.30.4 XFS(..?) regression (& with/2.6.31-rc6)

[Justin Piszcz]

On Sat, 22 Aug 2009, Christoph Hellwig wrote:

> On Sat, Aug 22, 2009 at 10:45:13AM -0400, Justin Piszcz wrote:
>>> Yes, entirely possible.   All the crypto loop / dm implementation used
>>> to have a tendency to never complete I/O once in a while, although that
>>> didn't happen anymore recently.
>>
>> Ok,
>>
>> What is the best way to debug the issue? git bi-sects between 2.6.29.x
>> and 2.6.31-rc6? or?
>
> Well, that would be a good way to find it.  But I'm personally not
> interested in spending much time on debugging a problem if it only
> happens with fishy out of tree modules.
>

.. No out of tree modules are used ..

Here is the kernel config used:
http://home.comcast.net/~jpiszcz/20090822/config.txt

After 24hrs of various tar-like read/write/verify I/O, the error usually 
recurs.

Justin.