Summary: Cleanup mount path by avoiding calling btrfs_mount() twice. No functional change.
change to v2: split the patch into three parts. Long Explanation: btrfs uses mount_subtree() to mount a subvolume directly. This function needs a vfsmount* of device's root (/), which is a return value of vfs_kern_mount() (therefore root has to be mounted internally anyway). Current approach of getting root's vfsmount* in mount time is a bit tricky: 1. mount systemcall calls vfs_kern_mount() on the way 2. btrfs_mount() is called 3. btrfs_parse_early_options() parses "subvolid=" mount option and set the value to subvol_objectid. Otherwise, subvol_objectid has the initial value of 0 4. check subvol_objectid is 5 or not. This time id is not 5, and btrfs_mount() returns by calling mount_subvol() 5. In mount_subvol(), original mount options are modified to contain "subvolid=0" in setup_root_args(). Then, vfs_kern_mount() is called with this new options to get root's vfsmount* 6. btrfs_mount() is called again 7. btrfs_parse_early_options() parses "subvolid=0" and set 5 (instead of 0) to subvol_objectid 8. check subvol_objectid is 5 or not. This time id is 5 and mount_subvol() is not called. btrfs_mount() finishes mounting a root 9. (in mount_subvol()) with using a return vale of vfs_kern_mount(), it calls mount_subtree() 10 return subvolume's dentry As illustrated above, calling btrfs_mount() twice complicates the problem. Callback function of mount time (btrfs_mount()) is specified in struct file_system_type which is passed to vfs_kern_mount(). Therefore, we can avoid this by using another file_system_type for arguments of our vfs_kern_mount() call. There is no need of modifying mount options. In this approach: 1. btrfs_mount() is called 2. parse "subvolid=" opiton and set the value to subvol_objectid 3. mount device's root by calling vfs_kern_mount() with different file_system_type specified. Then, different callback function is called (mount_root()). Most of this new function is the same as the original btrfs_mount() 4. return by calling mount_subtree() I think this approach is the same as nfsv4, which is the only other filesystem using mount_subtree() currently, and easy to understand. Most of the change is done by just reorganizing the original code of btrfs_mount()/mount_subvol() into btrfs_mount()/mount_subvol()/mount_root() btrfs_parse_early_options() is split into two parts to avoid "device=" option will be handled twice (though it cause no harm). setup_root_args() is deleted as not needed anymore. Tomohiro Misono (3): change btrfs_mount() to mount_root() split parse_early_options() in two introduce new btrfs_mount() fs/btrfs/super.c | 231 ++++++++++++++++++++++++++++++------------------------- 1 file changed, 128 insertions(+), 103 deletions(-) -- 2.9.5 -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
