autofs-5.0.5 - always read file maps mount lookup map read fix

From: Ian Kent <raven@themaw.net>

If, during a mount lookup, a file map is found to have changed and the

browse option is not being used, the file map won't be refreshed, even

though file maps should always be read into the cache.

A previous change for the "always read file maps" optimization moved

the check for the browse option into the map module read function so

checking for this in send_map_update_request() is not needed and

causes nobrowse file maps to not be read if they have changed. This

isn't quite optimal as there will be a partial read of the file map

to satisfy the lookup and a full read of the map done by the queued

update request. I don't think anything can be done about that though.

---

 CHANGELOG    |    1 +

 lib/master.c |    3 ---

 2 files changed, 1 insertion(+), 3 deletions(-)

--- autofs-5.0.5.orig/CHANGELOG

+++ autofs-5.0.5/CHANGELOG

@@ -49,6 +49,7 @@

 - remove ERR_remove_state() openssl call.

 - fix init script restart option.

 - fix init script status privilege error.

+- always read file maps mount lookup map read fix.

 03/09/2009 autofs-5.0.5

 -----------------------

--- autofs-5.0.5.orig/lib/master.c

+++ autofs-5.0.5/lib/master.c

@@ -489,9 +489,6 @@ void send_map_update_request(struct auto

 	struct map_source *map;

 	int status, need_update = 0;

-	if (!(ap->flags & MOUNT_FLAG_GHOST))

-		return;

-

 	status = pthread_mutex_lock(&instance_mutex);

 	if (status)

 		fatal(status);