[Top] [All Lists]

Re: Oops from unlock in xfs_trans_ail_delete_bulk()

To: Alex Elder <elder@xxxxxxxxxxx>
Subject: Re: Oops from unlock in xfs_trans_ail_delete_bulk()
From: Ben Myers <bpm@xxxxxxx>
Date: Thu, 6 Sep 2012 10:41:58 -0500
Cc: xfs@xxxxxxxxxxx
In-reply-to: <504891EB.2090202@xxxxxxxxxxx>
References: <504891EB.2090202@xxxxxxxxxxx>
User-agent: Mutt/1.5.20 (2009-06-14)
Hi Alex,

On Thu, Sep 06, 2012 at 07:07:07AM -0500, Alex Elder wrote:
> I'm trying to capture something useful here because I need to move
> on with what I'm doing.  I was running xfstests using a kernel with
> this as the latest commit affecting fs/xfs:
>     commit a0e881b7c189fa2bd76c024dbff91e79511c971d
>     Merge: eff0d13 dbc6e02
>     Author: Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
>     Date:   Wed Aug 1 10:26:23 2012 -0700
>     Merge branch 'for-linus' of
> git://git.kernel.org/pub/scm/linux/kernel/git/vi
> While running test 135 (it had started but its completion time had
> not yet been reported) I got an oops whose details I'll include below.
> My console window isn't properly wrapping long lines so some info
> has been truncated below.  The fault address is at
> __ticket_spin_is_locked+0x10, called via xfs_trans_ail_delete_bulk().
> I may be able to offer a little more information if someone is
> interested, and if I find this is repeatable I will certainly
> report it.

Thanks for the report.  This is one we've been struggling with for a little

Mark has a patch in this thread that seems to work ok:

And Dave is also working toward a fix in this thread:

I think that this will be resolved soon.  For now there are a few options that
seem to work around the problem, as described in the first thread.


<Prev in Thread] Current Thread [Next in Thread>