autofs4-2.6.34-rc1 - fix link_count usage
authorIan Kent <raven@themaw.net>
Mon, 10 May 2010 08:46:08 +0000 (16:46 +0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Mon, 10 May 2010 16:48:10 +0000 (09:48 -0700)
After commit 1f36f774b2 ("Switch !O_CREAT case to use of do_last()") in
2.6.34-rc1 autofs direct mounts stopped working.  This is caused by
current->link_count being 0 when ->follow_link() is called from
do_filp_open().

I can't work out why this hasn't been seen before Als patch series.

This patch removes the autofs dependence on current->link_count.

Signed-off-by: Ian Kent <raven@themaw.net>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
fs/autofs4/root.c

index 109a6c606d9292d3438795291d22c7a3c9a94b9e..e8e5e63ac9503b6f42d20681a26fc6565e43641b 100644 (file)
@@ -177,8 +177,7 @@ static int try_to_fill_dentry(struct dentry *dentry, int flags)
                }
        /* Trigger mount for path component or follow link */
        } else if (ino->flags & AUTOFS_INF_PENDING ||
-                       autofs4_need_mount(flags) ||
-                       current->link_count) {
+                       autofs4_need_mount(flags)) {
                DPRINTK("waiting for mount name=%.*s",
                        dentry->d_name.len, dentry->d_name.name);
 
@@ -262,7 +261,7 @@ static void *autofs4_follow_link(struct dentry *dentry, struct nameidata *nd)
                spin_unlock(&dcache_lock);
                spin_unlock(&sbi->fs_lock);
 
-               status = try_to_fill_dentry(dentry, 0);
+               status = try_to_fill_dentry(dentry, nd->flags);
                if (status)
                        goto out_error;