On Mon, Nov 24, 2014 at 12:22:16PM -0500, Theodore Ts'o wrote:

On Mon, Nov 24, 2014 at 05:38:30PM +0100, David Sterba wrote:

quoted

It is necessary and the whole .update_time callback was added
intentionally, see commits

c3b2da314834499f34cba94f7053e55f6d6f92d8
fs: introduce inode operation ->update_time

e41f941a23115e84a8550b3d901a13a14b2edc2f
Btrfs: move over to use ->update_time

Being able to signal an error if the time update fails is still
possible even if we drop update_time(), because the new write_time()
function will return an error.

Fine, means your change does not break the current status. I was
providing the more complete list of related commits.

quoted

2bc5565286121d2a77ccd728eb3484dff2035b58
Btrfs: don't update atime on RO subvolumes

Yes, but this doesn't answer my question about other places where the
VFS is only checking MS_RDONLY and MNT_READONLY besides just
update_atime().  Maybe we should be exposing an "is_readonly(inode)"
inode operations function to address this?

Yes, if this is a lightweight check then it'd would allow to remove the
filesystem-specific checks.