[Top] [All Lists]

Re: [PATCH] Prevent log tail pushing from blocking on buffer locks

To: Lachlan McIlroy <lachlan@xxxxxxx>
Subject: Re: [PATCH] Prevent log tail pushing from blocking on buffer locks
From: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Date: Thu, 24 Jul 2008 01:45:31 -0400
Cc: Christoph Hellwig <hch@xxxxxxxxxxxxx>, xfs-dev <xfs-dev@xxxxxxx>, xfs-oss <xfs@xxxxxxxxxxx>
In-reply-to: <48881788.7010709@xxxxxxx>
References: <48857EFB.3030301@xxxxxxx> <20080723112154.GA17338@xxxxxxxxxxxxx> <48881788.7010709@xxxxxxx>
Sender: xfs-bounce@xxxxxxxxxxx
User-agent: Mutt/1.5.18 (2008-05-17)
On Thu, Jul 24, 2008 at 03:47:52PM +1000, Lachlan McIlroy wrote:
>> The stuck propagation looks good, but I don't think this should be
>> blindly done for all errors.  The only error where it makes sense is
>> the EAGAIN from xfs_iflush.  All other returns inside the item_push
>> handlers basically indicate filesystem corruption.
> Good point.  Regardless of the error it's still an item that could not
> be pushed and is effectively 'stuck'.  What do you recommend I do for
> other errors?  Shutdown the filesystem?

Good questions.  When I did a quick audit many error do in fact come
from tests for shutdown filesystems, so they shouldn't ever happens.
So maybe add an assert that we're never getting here for shut down
filesystems and otherwise shut it down.

But all this is getting a little complicated and riksy, so I'd say push
your original patch in first and make the error handling a second one.

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