xfs
[Top] [All Lists]

Re: Assertion failed: RB_EMPTY_NODE(&bp->b_rbnode)

To: Ben Myers <bpm@xxxxxxx>
Subject: Re: Assertion failed: RB_EMPTY_NODE(&bp->b_rbnode)
From: Dave Chinner <david@xxxxxxxxxxxxx>
Date: Tue, 1 May 2012 09:12:23 +1000
Cc: xfs@xxxxxxxxxxx
In-reply-to: <20120430143200.GL16881@xxxxxxx>
References: <1335160747-17254-1-git-send-email-david@xxxxxxxxxxxxx> <20120430143200.GL16881@xxxxxxx>
User-agent: Mutt/1.5.21 (2010-09-15)
On Mon, Apr 30, 2012 at 09:32:00AM -0500, Ben Myers wrote:
> While testing this patchset without Jan's freeze work, I hit this assert.  I
> believe I have not seen this one before.
> 
> -Ben
> 
> [36850.416217] XFS: Assertion failed: RB_EMPTY_NODE(&bp->b_rbnode), file: 
> /work/xfs/fs/xfs/xfs_buf.c, line: 838
> [36850.445669] BUG: unable to handle kernel NULL pointer dereference at 
> 0000000000000038

Two problems - the BUG seems to be one that you have reported
before, while the assert failure tends to indicate a use after free
situation on a buffer (i.e.  a log buffer has been torn down while
IO is still in progress). That smells like different symptoms of the
same problem....

I think you are going to have to debug this one, Ben, because I have
not seen anything at all like these problems on any of my test or
stress workloads. I don't even know what test you are running at the
time these failures are occurring, so I can't even start looking...

Cheers,

Dave.
-- 
Dave Chinner
david@xxxxxxxxxxxxx

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