xfs
[Top] [All Lists]

Re: [PATCH, RFC] more reserved blocks fixups

To: Dave Chinner <david@xxxxxxxxxxxxx>
Subject: Re: [PATCH, RFC] more reserved blocks fixups
From: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Date: Thu, 4 Feb 2010 11:18:47 -0500
Cc: Eric Sandeen <sandeen@xxxxxxxxxxx>, xfs mailing list <xfs@xxxxxxxxxxx>
In-reply-to: <20100128015817.GH15853@xxxxxxxxxxxxxxxx>
References: <4B60C8EE.5080700@xxxxxxxxxxx> <20100128015817.GH15853@xxxxxxxxxxxxxxxx>
User-agent: Mutt/1.5.19 (2009-01-05)
On Thu, Jan 28, 2010 at 12:58:17PM +1100, Dave Chinner wrote:
> On Wed, Jan 27, 2010 at 05:14:54PM -0600, Eric Sandeen wrote:
> > This mangles the reserved blocks counts a little more.
> > 
> > 1) add a helper function for the default reserved count
> > 2) add helper functions to save/restore counts on ro/rw
> > 3) save/restore reserved blocks on freeze/thaw
> > 4) disallow changing reserved count while readonly
> > 
> > for 2) - maybe better names (save_and_clear?)
> > for 4) - maybe allow, but change the _ro field instead?
> > 
> > (TBH not tested yet but wondered if this seems sane)
> 
> I was wondering if the save/restore could be encapsualted entirely
> within xfs_quiesce_attr(). i.e. save before the superblock write,
> restore directly after. That removes the need for a variable in
> the xfs_mount structure, and catches both remount,ro and freeze.

Seems like a bit fragile.  But I think we want Eric's version
over yours for a start - we should have a clean filesystem not
only after remount,ro but also for a frozen filesystem.

<Prev in Thread] Current Thread [Next in Thread>
  • Re: [PATCH, RFC] more reserved blocks fixups, Christoph Hellwig <=