[PATCH] VFS: Fix automount for negative autofs dentries

[PATCH] VFS: Fix automount for negative autofs dentries

[David Howells]

Autofs may set the DCACHE_NEED_AUTOMOUNT flag on negative dentries.  These
need attention from the automounter daemon regardless of the LOOKUP_FOLLOW flag.

Signed-off-by: David Howells <dhowells@redhat.com>
Acked-by: Ian Kent <raven@themaw.net>
---

 fs/namei.c |   27 +++++++++++++++++++--------
 1 files changed, 19 insertions(+), 8 deletions(-)

diff --git a/fs/namei.c b/fs/namei.c
index 0223c41..93e221e 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -777,18 +777,29 @@ static int follow_automount(struct path *path, unsigned flags,
 	if ((flags & LOOKUP_NO_AUTOMOUNT) && !(flags & LOOKUP_CONTINUE))
 		return -EISDIR; /* we actually want to stop here */
 
-	/* We want to mount if someone is trying to open/create a file of any
-	 * type under the mountpoint, wants to traverse through the mountpoint
-	 * or wants to open the mounted directory.
-	 *
-	 * We don't want to mount if someone's just doing a stat and they've
+	/* We don't want to mount if someone's just doing a stat and they've
 	 * set AT_SYMLINK_NOFOLLOW - unless they're stat'ing a directory and
 	 * appended a '/' to the name.
 	 */
-	if (!(flags & LOOKUP_FOLLOW) &&
-	    !(flags & (LOOKUP_CONTINUE | LOOKUP_DIRECTORY |
-		       LOOKUP_OPEN | LOOKUP_CREATE)))
+	if (!(flags & LOOKUP_FOLLOW)) {
+		/* We do, however, want to mount if someone wants to open or
+		 * create a file of any type under the mountpoint, wants to
+		 * traverse through the mountpoint or wants to open the mounted
+		 * directory.
+		 */
+		if (flags & (LOOKUP_CONTINUE | LOOKUP_DIRECTORY |
+			     LOOKUP_OPEN | LOOKUP_CREATE))
+			goto need_automount;
+
+		/* Also, autofs may mark negative dentries as being automount
+		 * points.  These will need the attentions of the daemon to
+		 * instantiate them before they can be used.
+		 */
+		if (!path->dentry->d_inode)
+			goto need_automount;
 		return -EISDIR;
+	}
+need_automount:
 
 	current->total_link_count++;
 	if (current->total_link_count >= 40)

Re: [PATCH] VFS: Fix automount for negative autofs dentries

[Christoph Hellwig]

> +	if (!(flags & LOOKUP_FOLLOW)) {
> +		/* We do, however, want to mount if someone wants to open or
> +		 * create a file of any type under the mountpoint, wants to
> +		 * traverse through the mountpoint or wants to open the mounted
> +		 * directory.
> +		 */
> +		if (flags & (LOOKUP_CONTINUE | LOOKUP_DIRECTORY |
> +			     LOOKUP_OPEN | LOOKUP_CREATE))
> +			goto need_automount;
> +
> +		/* Also, autofs may mark negative dentries as being automount
> +		 * points.  These will need the attentions of the daemon to
> +		 * instantiate them before they can be used.
> +		 */
> +		if (!path->dentry->d_inode)
> +			goto need_automount;
>  		return -EISDIR;
> +	}
> +need_automount:

That's some really odd code structure.

	if (!(flags & (LOOKUP_FOLLOW | LOOKUP_CONTINUE | LOOKUP_DIRECTORY |
		        LOOKUP_OPEN | LOOKUP_CREATE)) &&
	    path->dentry->d_inode)
		return -EISDIR;

would do the same.

Re: [PATCH] VFS: Fix automount for negative autofs dentries

[David Howells]

Christoph Hellwig <hch@infradead.org> wrote:

> would do the same.

But is much less obvious.  The LOOKUP_FOLLOW flag is the primary reason for
this statement.  The rest are subordinate and would be wholly irrelevant if
LOOKUP_FOLLOW was to be removed from the list.

David

Re: [PATCH] VFS: Fix automount for negative autofs dentries

[Christoph Hellwig]

On Mon, Jul 11, 2011 at 03:13:41PM +0100, David Howells wrote:
> Christoph Hellwig <hch@infradead.org> wrote:
> 
> > would do the same.
> 
> But is much less obvious.  The LOOKUP_FOLLOW flag is the primary reason for
> this statement.  The rest are subordinate and would be wholly irrelevant if
> LOOKUP_FOLLOW was to be removed from the list.

Then keep the LOOKUP_FOLLOW semi-separate as in the original code.
But there's absolutely no reason for the goto spaghetti.