xfs
[Top] [All Lists]

Re: XFS stack corruption: BUG: unable to handle kernel paging request

To: Michael Monnerie <michael.monnerie@xxxxxxxxxxxxxxxxxxx>
Subject: Re: XFS stack corruption: BUG: unable to handle kernel paging request
From: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Date: Wed, 16 Jun 2010 05:41:12 -0400
Cc: xfs@xxxxxxxxxxx
In-reply-to: <201006160810.33045@xxxxxx>
References: <201006160810.33045@xxxxxx>
User-agent: Mutt/1.5.20 (2009-08-17)
On Wed, Jun 16, 2010 at 08:10:23AM +0200, Michael Monnerie wrote:
> This happened while I did an "rsync" to that server, and it was searching the 
> files to sync.
> Messages grabbed from the console:

This might as well be another stack overflow.  You don't happen to have
more lines of the stack trace?  I fear this might be another case of
direct reclaim hitting you.

Try backporting commit 070ecdca54dde9577d2697088e74e45568f48efb
"xfs: skip writeback from reclaim context", that should get rid of
places calling into ->writepage with already deep stack usage.

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