VFS: Fix the remaining automounter semantics regressions
authorTrond Myklebust <Trond.Myklebust@netapp.com>
Tue, 27 Sep 2011 00:36:09 +0000 (20:36 -0400)
committerLinus Torvalds <torvalds@linux-foundation.org>
Tue, 27 Sep 2011 02:16:46 +0000 (19:16 -0700)
The concensus seems to be that system calls such as stat() etc should
not trigger an automount.  Neither should the l* versions.

This patch therefore adds a LOOKUP_AUTOMOUNT flag to tag those lookups
that _should_ trigger an automount on the last path element.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
[ Edited to leave out the cases that are already covered by LOOKUP_OPEN,
  LOOKUP_DIRECTORY and LOOKUP_CREATE - all of which also fundamentally
  force automounting for their own reasons   - Linus ]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
fs/namespace.c
fs/nfs/super.c
fs/quota/quota.c

index 22bfe8273c680b441f43a0191c7bd0e805a51ec2..b4febb29d3bb3d855793daa1b3f07cb303e1f6ea 100644 (file)
@@ -1757,7 +1757,7 @@ static int do_loopback(struct path *path, char *old_name,
                return err;
        if (!old_name || !*old_name)
                return -EINVAL;
-       err = kern_path(old_name, LOOKUP_FOLLOW, &old_path);
+       err = kern_path(old_name, LOOKUP_FOLLOW|LOOKUP_AUTOMOUNT, &old_path);
        if (err)
                return err;
 
index 9b7dd7013b155e2f0e38a783b0aebe44446b9655..5b19b6aabe18d6e7f50b5d9193304ea70fa19487 100644 (file)
@@ -2798,7 +2798,7 @@ static struct dentry *nfs_follow_remote_path(struct vfsmount *root_mnt,
                goto out_put_mnt_ns;
 
        ret = vfs_path_lookup(root_mnt->mnt_root, root_mnt,
-                       export_path, LOOKUP_FOLLOW, &path);
+                       export_path, LOOKUP_FOLLOW|LOOKUP_AUTOMOUNT, &path);
 
        nfs_referral_loop_unprotect();
        put_mnt_ns(ns_private);
index b34bdb25490c8956b4733c29771bea3109ec2fcf..10b6be3ca280963e90e5de2b11f5556e576cea4a 100644 (file)
@@ -355,7 +355,7 @@ SYSCALL_DEFINE4(quotactl, unsigned int, cmd, const char __user *, special,
         * resolution (think about autofs) and thus deadlocks could arise.
         */
        if (cmds == Q_QUOTAON) {
-               ret = user_path_at(AT_FDCWD, addr, LOOKUP_FOLLOW, &path);
+               ret = user_path_at(AT_FDCWD, addr, LOOKUP_FOLLOW|LOOKUP_AUTOMOUNT, &path);
                if (ret)
                        pathp = ERR_PTR(ret);
                else