From: Renato Westphal <renato@opensourcerouting.org>
Date: Mon, 31 May 2021 13:27:51 +0000 (-0300)
Subject: ospfd: fix cleanup of MaxAge LSAs on exit
X-Git-Tag: base_8.1~445^2~5
X-Git-Url: https://git.puffer.fish/?a=commitdiff_plain;h=69843e810acc27b152c19fe91bb2e9b67f808108;p=matthieu%2Ffrr.git

ospfd: fix cleanup of MaxAge LSAs on exit

During shutdown, the ospf->maxage_lsa table is iterated over to
clean up all existing entries. While doing that, route_unlock_node()
should be called only for the nodes that have an associated entry,
otherwise the table will get corrupted and ospfd will crash.

As a side note, using a routing table to store MaxAge LSAs was a
very poor choice of a data structure, considering that a simple
rb-tree or hash table would get the job done with a much simpler
(and less error-prone) API. Something to cleanup in the future...

Signed-off-by: Renato Westphal <renato@opensourcerouting.org>
---

diff --git a/ospfd/ospfd.c b/ospfd/ospfd.c
index 106ce3bd7d..d053bdabbd 100644
--- a/ospfd/ospfd.c
+++ b/ospfd/ospfd.c
@@ -828,8 +828,8 @@ static void ospf_finish_final(struct ospf *ospf)
 		if ((lsa = rn->info) != NULL) {
 			ospf_lsa_unlock(&lsa);
 			rn->info = NULL;
+			route_unlock_node(rn);
 		}
-		route_unlock_node(rn);
 	}
 	route_table_finish(ospf->maxage_lsa);