After an expire if the current map entry is stale because it could not
be cleaned up during the map entry prune (due to the presence of a real
mount) it needs to be cleaned up after a successful expire.
Currently this is done by update_map_cache() if the entry is contained
in an invalid map but it should be doing it if the entry is no longer
valid. In addition, if update_map_cache() gets called the umount has
been successful so the ioctlfd does not need to be checked (and is in
fact updated later in the expire process).
Signed-off-by: Ian Kent <raven@xxxxxxxxxx>
---
CHANGELOG | 1 +
daemon/automount.c | 9 ++-------
2 files changed, 3 insertions(+), 7 deletions(-)
diff --git a/CHANGELOG b/CHANGELOG
index d9daf1e2e..ebdfe32b7 100644
--- a/CHANGELOG
+++ b/CHANGELOG
@@ -35,6 +35,7 @@
- quiet possibly noisy log message.
- fix devid update on reload.
- fix cache writelock must be taken in update_map_cache().
+- fix skip valid map entries on expire cleanup.
02/11/2023 autofs-5.1.9
- fix kernel mount status notification.
diff --git a/daemon/automount.c b/daemon/automount.c
index 5476e84ba..5578a965b 100644
--- a/daemon/automount.c
+++ b/daemon/automount.c
@@ -519,16 +519,11 @@ static void update_map_cache(struct autofs_point *ap, const char *path)
while (map) {
struct mapent *me = NULL;
- /* Skip current, in-use cache */
- if (ap->entry->age <= map->age) {
- map = map->next;
- continue;
- }
-
mc = map->mc;
cache_writelock(mc);
me = cache_lookup_distinct(mc, key);
- if (me && me->ioctlfd == -1)
+ /* Only for invalid map entries */
+ if (me && map->age > me->age)
cache_delete(mc, key);
cache_unlock(mc);
--
2.50.1
