I've found a case where invalid dentrys in a mount tree, waiting to be
cleaned up by d_invalidate, prevent the expected expire.
In this case dentrys created during a lookup for which a mount fails or has
no entry in the mount map contribute to the d_count of the parent dentry.
These dentrys may not be invalidated prior to comparing the interanl usage
count of valid autofs dentrys against the dentry d_count which makes a
mount tree appear busy so it doesn't expire.
Signed-off-by: Ian Kent <raven@themaw.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
struct autofs_info *ino = autofs4_dentry_ino(p);
unsigned int ino_count = atomic_read(&ino->count);
+ /*
+ * Clean stale dentries below that have not been
+ * invalidated after a mount fail during lookup
+ */
+ d_invalidate(p);
+
/* allow for dget above and top is already dgot */
if (p == top)
ino_count += 2;