xfs
[Top] [All Lists]

Re: [PATCH 13/17] xfs: get rid of indirections in the quotaops implement

To: Christoph Hellwig <hch@xxxxxxxxxxxxx>, xfs@xxxxxxxxxxx
Subject: Re: [PATCH 13/17] xfs: get rid of indirections in the quotaops implementation
From: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Date: Mon, 9 Feb 2009 02:46:32 -0500
In-reply-to: <20090209024115.GH8830@disturbed>
References: <20090126073136.384490000@xxxxxxxxxxxxxxxxxxxxxx> <20090126073202.961302000@xxxxxxxxxxxxxxxxxxxxxx> <20090209024115.GH8830@disturbed>
User-agent: Mutt/1.5.18 (2008-05-17)
On Mon, Feb 09, 2009 at 01:41:15PM +1100, Dave Chinner wrote:
> > +STATIC int
> > +xfs_fs_get_xstate(
> > +   struct super_block      *sb,
> > +   struct fs_quota_stat    *fqs)
> > +{
> > +   struct xfs_mount        *mp = XFS_M(sb);
> > +
> > +   if (!XFS_IS_QUOTA_RUNNING(mp))
> > +           return -ENOSYS;
> > +   return xfs_qm_scall_getqstat(mp, fqs);
> 
> The return of xfs_qm_scall_getqstat() needs to be negated.

Currently it only ever returns 0.  But I agree, if this ever returns
an error it would be a positive one, so I added the inversion.

> > +   return xfs_qm_scall_setqlim(mp, id, xfs_quota_type(type), fdq);
> 
> That should be negated as well.

But this one can already return a positive error, so it's definitively
needed.

Thanks, I've commited the updated version.

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