Re: [PATCH v3] btrfs: Introduce new mount option to disable tree log replay

[David Sterba <dsterba@suse.cz>]

On Mon, Dec 14, 2015 at 12:50:37PM -0500, Austin S. Hemmelgarn wrote:
> On 2015-12-14 12:32, David Sterba wrote:
> > On Thu, Dec 10, 2015 at 10:34:06AM +0800, Qu Wenruo wrote:
> >> Introduce a new mount option "nologreplay" to co-operate with "ro" mount
> >> option to get real readonly mount, like "norecovery" in ext* and xfs.
> >>
> >> Since the new parse_options() need to check new flags at remount time,
> >> so add a new parameter for parse_options().
> >>
> >> Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
> >> Reviewed-by: Chandan Rajendra <chandan@linux.vnet.ibm.com>
> >> Tested-by: Austin S. Hemmelgarn <ahferroin7@gmail.com>
> >
> > I've read the discussions around the change and from the user's POV I'd
> > suggest to add another mount option that would be just an alias for any
> > mount options that would implement the 'hard-ro' semantics.
> >
> > Say it's called 'nowr'. Now it would imply 'nologreplay', but may cover
> > more options in the future.
> It should also imply noatime.  I'm not sure how BTRFS handles atime when 
> mounted RO, but I know a lot of old UNIX systems updated atime even on 
> filesystems mounted RO, and I know that at least at one point Linux did too.

A mount with -o ro will not touch atimes. At one point the read-only
snapshots changed atimes, but this has been fixed since.

http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/fs/inode.c#n1602
http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/fs/btrfs/inode.c#n5973

> >   mount -o ro,nowr /dev/sdx /mnt
> >
> > would work when switching kernels.
> 
> I like this idea, but I think that having a name like true-ro or hard-ro 
> and making it imply ro (and noatime) would probably be better (or at 
> least, simpler to use from a user perspective).

Ok, a single option to do the real-ro sounds better than ro,something.