On Tue, Mar 28, 2023 at 06:13:09PM +0200, Christian Brauner wrote:
> Currently, lock_mount() uses a goto to retry the lookup until it
> succeeded in acquiring the namespace_lock() preventing the top mount
> from being overmounted. While that's perfectly fine we want to lookup
> the mountpoint on the parent of the top mount in later patches. So adapt
> the code to make this easier to implement. Also, the for loop is
> arguably a little cleaner and makes the code easier to follow. No
> functional changes intended.
>
> Signed-off-by: Christian Brauner <brauner@xxxxxxxxxx>
> ---
> fs/namespace.c | 50 +++++++++++++++++++++++++++++---------------------
> 1 file changed, 29 insertions(+), 21 deletions(-)
>
> diff --git a/fs/namespace.c b/fs/namespace.c
> index 42dc87f86f34..7f22fcfd8eab 100644
> --- a/fs/namespace.c
> +++ b/fs/namespace.c
> @@ -2308,31 +2308,39 @@ static int attach_recursive_mnt(struct mount *source_mnt,
>
> static struct mountpoint *lock_mount(struct path *path)
> {
> - struct vfsmount *mnt;
> - struct dentry *dentry = path->dentry;
> -retry:
> - inode_lock(dentry->d_inode);
> - if (unlikely(cant_mount(dentry))) {
> - inode_unlock(dentry->d_inode);
> - return ERR_PTR(-ENOENT);
> - }
> - namespace_lock();
> - mnt = lookup_mnt(path);
> - if (likely(!mnt)) {
> - struct mountpoint *mp = get_mountpoint(dentry);
> - if (IS_ERR(mp)) {
> - namespace_unlock();
> + struct vfsmount *mnt = path->mnt;
One small complaint, mnt is reassigned before it is used, so this
assignment is unnecessary. Otherwise looks good.
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@xxxxxxxxxx>
