Re: -mm: xfs lockdep warning

[Torsten Kaiser <just.for.lkml@xxxxxxxxxxxxxx>]

On Tue, Oct 5, 2010 at 6:51 PM, Tejun Heo <tj@xxxxxxxxxx> wrote:
> Hello,
>
> This is getting confusing.
>
> On 10/05/2010 12:09 PM, Torsten Kaiser wrote:
>> 2) the hang with 2.6.35-rc5 I have seen twice (triggered probably by
>> high load while building KOffice on an tmpfs)
>
> The trace doesn't seem to firmly point at xfs locking up.  There are
> md and dm crypt involved.  Are you sure this is locking up inside xfs?

Not 100%, but apart from the md/dm-worker the other prcesses got stuck
in sync_page() and do_lockup() and these function look filesystem
related. And the only filesystem I'm using on that system is XFS.

And having just read about some problems (that were fixed in
2.6.36-rc5) regarding XFS vs the new workqueues, I "feared" that I
somehow triggered some funny interaction between the workqueues of
XFS, md-raid1 and dm-crypt.

> Also, can you please enable frame pointers so that we can get more
> reliable backtrace?

I will do that and see if -rc6 gets stuck the same way.
But it might take until tomorrow to test this.

>> 3) the lockdep issue that started this thread and that I am seeing in
>> mainline 2.6.36-rc5 and -rc6 and that seems to be rather easy to
>> trigger for me.
>>
>> Because 3) is regarded as a false positive, it should not be the cause
>> of 2)
>
> For now, let's not draw any conclusion or mix the issues.  I'll look
> up the original thread and look into what the lockdep warning means.
>
> Thanks.
>
> --
> tejun
>