From: David S. Miller <davem@davemloft.net>
Date: Fri, 30 Aug 2013 21:39:33 +0000 (-0400)
Subject: Revert "ipv6: Don't depend on per socket memory for neighbour discovery messages"
X-Git-Url: http://git.maquefel.me/?a=commitdiff_plain;h=25ad6117e73656071b38fd19fa67ae325471c758;p=linux.git

Revert "ipv6: Don't depend on per socket memory for neighbour discovery messages"

This reverts commit 1f324e38870cc09659cf23bc626f1b8869e201f2.

It seems to cause regressions, and in particular the output path
really depends upon there being a socket attached to skb->sk for
checks such as sk_mc_loop(skb->sk) for example.  See ip6_output_finish2().

Reported-by: Stephen Warren <swarren@wwwdotorg.org>
Reported-by: Fabio Estevam <festevam@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---

diff --git a/net/ipv6/ndisc.c b/net/ipv6/ndisc.c
index 5cb98df966c29..04d31c2fbef1e 100644
--- a/net/ipv6/ndisc.c
+++ b/net/ipv6/ndisc.c
@@ -370,12 +370,16 @@ static struct sk_buff *ndisc_alloc_skb(struct net_device *dev,
 {
 	int hlen = LL_RESERVED_SPACE(dev);
 	int tlen = dev->needed_tailroom;
+	struct sock *sk = dev_net(dev)->ipv6.ndisc_sk;
 	struct sk_buff *skb;
+	int err;
 
-	skb = alloc_skb(hlen + sizeof(struct ipv6hdr) + len + tlen, GFP_ATOMIC);
+	skb = sock_alloc_send_skb(sk,
+				  hlen + sizeof(struct ipv6hdr) + len + tlen,
+				  1, &err);
 	if (!skb) {
-		ND_PRINTK(0, err, "ndisc: %s failed to allocate an skb\n",
-			  __func__);
+		ND_PRINTK(0, err, "ndisc: %s failed to allocate an skb, err=%d\n",
+			  __func__, err);
 		return NULL;
 	}