On Fri, Nov 21, 2014 at 04:42:45PM -0500, Theodore Ts'o wrote:

Out of curiosity, why does btrfs_update_time() need to call
btrfs_root_readonly()?  Why can't it just depend on the
__mnt_want_write() call in touch_atime()?

mnt_want_write looks only at the mountpoint flags, the readonly
subvolume status is external to that.

Surely if there are times when it's not OK to write into a btrfs file
system and mnt_is_readonly() returns false, the VFS is going to get
very confused abyway.

If the btrfs_update_time() is not necessary, then we could drop
btrfs_update_time() and update_time() from the inode operations
entirely, and depend on the VFS-level code in update_time().

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

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