summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorTommy S. Christensen <tommy.christensen@tpack.net>2005-05-19 22:07:32 +0200
committerDavid S. Miller <davem@davemloft.net>2005-05-19 22:07:32 +0200
commitaa1c6a6f7f0518b42994d02756a41cbfdcac1916 (patch)
treef56cbf31594d904ecc711269c14c206cf08c0126
parent[NETLINK]: Move broadcast skb_orphan to the skb_get path. (diff)
downloadlinux-aa1c6a6f7f0518b42994d02756a41cbfdcac1916.tar.xz
linux-aa1c6a6f7f0518b42994d02756a41cbfdcac1916.zip
[NETLINK]: Defer socket destruction a bit
In netlink_broadcast() we're sending shared skb's to netlink listeners when possible (saves some copying). This is OK, since we hold the only other reference to the skb. However, this implies that we must drop our reference on the skb, before allowing a receiving socket to disappear. Otherwise, the socket buffer accounting is disrupted. Signed-off-by: Tommy S. Christensen <tommy.christensen@tpack.net> Acked-by: Herbert Xu <herbert@gondor.apana.org.au> Signed-off-by: David S. Miller <davem@davemloft.net>
-rw-r--r--net/netlink/af_netlink.c3
1 files changed, 2 insertions, 1 deletions
diff --git a/net/netlink/af_netlink.c b/net/netlink/af_netlink.c
index 4b91f4b84cb7..e41ce458c2a9 100644
--- a/net/netlink/af_netlink.c
+++ b/net/netlink/af_netlink.c
@@ -789,11 +789,12 @@ int netlink_broadcast(struct sock *ssk, struct sk_buff *skb, u32 pid,
sk_for_each_bound(sk, node, &nl_table[ssk->sk_protocol].mc_list)
do_one_broadcast(sk, &info);
+ kfree_skb(skb);
+
netlink_unlock_table();
if (info.skb2)
kfree_skb(info.skb2);
- kfree_skb(skb);
if (info.delivered) {
if (info.congested && (allocation & __GFP_WAIT))