xfs
[Top] [All Lists]

Re: [BUG] assert in xfs_imap_to_bp, line 185

To: Ben Myers <bpm@xxxxxxx>
Subject: Re: [BUG] assert in xfs_imap_to_bp, line 185
From: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Date: Tue, 8 Nov 2011 12:06:07 -0500
Cc: xfs@xxxxxxxxxxx
In-reply-to: <20111108163932.GA22095@xxxxxxx>
References: <20111108163932.GA22095@xxxxxxx>
User-agent: Mutt/1.5.21 (2010-09-15)
On Tue, Nov 08, 2011 at 10:39:32AM -0600, Ben Myers wrote:
> Hi XFS Folks,
> 
> I am able to hit this assert consistently when running test 111 on recent
> (commit 9e4c109a) oss xfs bits:

That's expected - the test must only be run on non-debug kernels.  It
corrupt the inode, and then we read it.  A normal kernel shouldn't
panic, but for the debug kernel we settled on an assert for it.

I'm not sure it's a smart idea, and WANT_CORRUPTED_GOTO might be
better in the end.

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