Re: [PATCH 1/4] fs: split update_time() into update_time() and write_time()

[Theodore Ts'o <tytso@xxxxxxx>]

On Mon, Nov 24, 2014 at 07:21:01AM -0800, Christoph Hellwig wrote:
> On Fri, Nov 21, 2014 at 02:59:21PM -0500, Theodore Ts'o wrote:
> > We needed to preserve update_time() because btrfs wants to have a
> > special btrfs_root_readonly() check; otherwise we could drop the
> > update_time() inode operation entirely.
> 
> Can't btrfs just set the immutable flag on every inode that is read
> when the root has the BTRFS_ROOT_SUBVOL_RDONLY flag?  That would
> cut down the places that need this check to the ioctl path so that
> we prevent users from clearling the immutable flag.

Sounds like a good plan to me, although I'm not sure I understand how
BTRFS_ROOT_SUBVOL_RDONLY flag works, since I would have thought there
are all sorts of places in the VFS layer where it is currently
checking MS_RDONLY and MNT_READONLY and _not_ checking
BTRFS_ROOT_SUBVOL_RDONLY isn't causing other problems.

But unless there's something more subtle going on, it would seem to me
that setting the immutable flag on each inode would be a better way to
go in any case.

                                        - Ted