From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jan Kara Date: Wed, 22 Oct 2008 14:02:40 +0200 Subject: [Ocfs2-devel] [PATCH] ocfs2: Assign feature bits and system inodes to quota feature and quota files In-Reply-To: <20081021204318.GG2871@mail.oracle.com> References: <1224523443220-git-send-email-jack@suse.cz> <20081021204318.GG2871@mail.oracle.com> Message-ID: <20081022120240.GG13868@duck.suse.cz> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ocfs2-devel@oss.oracle.com On Tue 21-10-08 13:43:18, Joel Becker wrote: > On Mon, Oct 20, 2008 at 07:23:55PM +0200, Jan Kara wrote: > > Signed-off-by: Jan Kara > > This patch actually looks fine except for one thing: > > > diff --git a/fs/ocfs2/ocfs2_fs.h b/fs/ocfs2/ocfs2_fs.h > > index 4f61985..bb41c6a 100644 > > --- a/fs/ocfs2/ocfs2_fs.h > > +++ b/fs/ocfs2/ocfs2_fs.h > > @@ -91,7 +91,9 @@ > > | OCFS2_FEATURE_INCOMPAT_INLINE_DATA \ > > | OCFS2_FEATURE_INCOMPAT_EXTENDED_SLOT_MAP \ > > | OCFS2_FEATURE_INCOMPAT_USERSPACE_STACK) > > -#define OCFS2_FEATURE_RO_COMPAT_SUPP OCFS2_FEATURE_RO_COMPAT_UNWRITTEN > > +#define OCFS2_FEATURE_RO_COMPAT_SUPP (OCFS2_FEATURE_RO_COMPAT_UNWRITTEN \ > > + | OCFS2_FEATURE_RO_COMPAT_USRQUOTA \ > > + | OCFS2_FEATURE_RO_COMPAT_GRPQUOTA) > > The hunk turning on support should be the very last patch in the > series. It doesn't matter right now as you are working on it, but when > we go to upstream we'll want all the support in place before we turn it > on. Yes, I understand. But the only thing this patch will actually *do* is that quota system files will be loaded and S_NOQUOTA bit set on system files. No quota support is enabled by this patch. That happens in patch ocfs2-Enable-quota-accounting-on-mount-disable-on when everything necessary is in place (I even tested that the kernel compiles with only part of the patch sequence applied :). Arguably that patch should go even after the syncing thread patch and recovery patch which I now locally have, that is right and I'll probably do that. > The rest of the patch - defining the bits and the system files, > etc, is just fine. Thanks for review. Honza -- Jan Kara SUSE Labs, CR