Hi, Ian,
I'm not sure what purpose the following check serves in the cache_clean
function:
if (is_mounted(_PATH_MOUNTED, path)) {
free(path);
continue;
}
In the event that the map was updated, we signal the daemon to do a re-read
of the entire map. If the mount finishes before the daemon is able to
process this signal, we end up not pruning the old entry for the given
path. Now, because we append entries to the cache, we end up with the
stale entry first, followed by the updated entry. This means that
subsequent lookups will give the old entry.
As I don't see the cache information being used aside from during mount, I
think we can safely get rid of this check. Do you agree?
(Note that if you don't remove this check, there is a memory leak in the
second check for is_mounted; we don't free path in that case)
-Jeff
Index: lib/cache.c
===================================================================
RCS file: /root/autofs/lib/cache.c,v
retrieving revision 1.14
diff -u -p -r1.14 cache.c
--- a/lib/cache.c 25 Apr 2005 03:42:08 -0000 1.14
+++ b/lib/cache.c 25 Apr 2005 19:17:28 -0000
@@ -312,11 +312,6 @@ void cache_clean(const char *root, time_
if (!path)
return;
- if (is_mounted(_PATH_MOUNTED, path)) {
- free(path);
- continue;
- }
-
if (me->age < age) {
pred->next = me->next;
free(me->key);
@@ -336,9 +331,6 @@ void cache_clean(const char *root, time_
path = cache_fullpath(root, me->key);
if (!path)
return;
-
- if (is_mounted(_PATH_MOUNTED, path))
- continue;
if (me->age < age) {
mapent_hash[i] = me->next;
_______________________________________________
autofs mailing list
autofs@linux.kernel.org
http://linux.kernel.org/mailman/listinfo/autofs
