ipv4: Fix a data-race around sysctl_fib_sync_mem.

[ Upstream commit 73318c4b7dbd0e781aaababff17376b2894745c0 ]

While reading sysctl_fib_sync_mem, it can be changed concurrently.
So, we need to add READ_ONCE() to avoid a data-race.

Fixes: 9ab948a91b2c ("ipv4: Allow amount of dirty memory from fib resizing to be controllable")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>

diff --git a/net/ipv4/fib_trie.c b/net/ipv4/fib_trie.c
index 51673d00bbeac215f810552d3e5ff5f2cc3df7f8..a1f830da4ad308f1ec814ddbfa0acd0c51eb36d5 100644 (file)
--- a/net/ipv4/fib_trie.c
+++ b/net/ipv4/fib_trie.c
@@ -500,7 +500,7 @@ static void tnode_free(struct key_vector *tn)
                tn = container_of(head, struct tnode, rcu)->kv;
        }
 
-       if (tnode_free_size >= sysctl_fib_sync_mem) {
+       if (tnode_free_size >= READ_ONCE(sysctl_fib_sync_mem)) {
                tnode_free_size = 0;
                synchronize_rcu();
        }