[PATCH] appletalk: Fix skb allocation size in loopback case

[PATCH] appletalk: Fix skb allocation size in loopback case

[Doug Brown]

If a DDP broadcast packet is sent out to a non-gateway target, it is
also looped back. There is a potential for the loopback device to have a
longer hardware header length than the original target route's device,
which can result in the skb not being created with enough room for the
loopback device's hardware header. This patch fixes the issue by
determining that a loopback will be necessary prior to allocating the
skb, and if so, ensuring the skb has enough room.

This was discovered while testing a new driver that creates a LocalTalk
network interface (LTALK_HLEN = 1). It caused an skb_under_panic.

Signed-off-by: Doug Brown <doug@schmorgal.com>
---
 net/appletalk/ddp.c | 33 +++++++++++++++++++++------------
 1 file changed, 21 insertions(+), 12 deletions(-)

diff --git a/net/appletalk/ddp.c b/net/appletalk/ddp.c
index ca1a0d07a087..ebda397fa95a 100644
--- a/net/appletalk/ddp.c
+++ b/net/appletalk/ddp.c
@@ -1577,8 +1577,8 @@ static int atalk_sendmsg(struct socket *sock, struct msghdr *msg, size_t len)
 	struct sk_buff *skb;
 	struct net_device *dev;
 	struct ddpehdr *ddp;
-	int size;
-	struct atalk_route *rt;
+	int size, hard_header_len;
+	struct atalk_route *rt, *rt_lo = NULL;
 	int err;
 
 	if (flags & ~(MSG_DONTWAIT|MSG_CMSG_COMPAT))
@@ -1641,7 +1641,22 @@ static int atalk_sendmsg(struct socket *sock, struct msghdr *msg, size_t len)
 	SOCK_DEBUG(sk, "SK %p: Size needed %d, device %s\n",
 			sk, size, dev->name);
 
-	size += dev->hard_header_len;
+	hard_header_len = dev->hard_header_len;
+	/* Leave room for loopback hardware header if necessary */
+	if (usat->sat_addr.s_node == ATADDR_BCAST &&
+	    (dev->flags & IFF_LOOPBACK || !(rt->flags & RTF_GATEWAY))) {
+		struct atalk_addr at_lo;
+
+		at_lo.s_node = 0;
+		at_lo.s_net  = 0;
+
+		rt_lo = atrtr_find(&at_lo);
+
+		if (rt_lo && rt_lo->dev->hard_header_len > hard_header_len)
+			hard_header_len = rt_lo->dev->hard_header_len;
+	}
+
+	size += hard_header_len;
 	release_sock(sk);
 	skb = sock_alloc_send_skb(sk, size, (flags & MSG_DONTWAIT), &err);
 	lock_sock(sk);
@@ -1649,7 +1664,7 @@ static int atalk_sendmsg(struct socket *sock, struct msghdr *msg, size_t len)
 		goto out;
 
 	skb_reserve(skb, ddp_dl->header_length);
-	skb_reserve(skb, dev->hard_header_len);
+	skb_reserve(skb, hard_header_len);
 	skb->dev = dev;
 
 	SOCK_DEBUG(sk, "SK %p: Begin build.\n", sk);
@@ -1700,18 +1715,12 @@ static int atalk_sendmsg(struct socket *sock, struct msghdr *msg, size_t len)
 		/* loop back */
 		skb_orphan(skb);
 		if (ddp->deh_dnode == ATADDR_BCAST) {
-			struct atalk_addr at_lo;
-
-			at_lo.s_node = 0;
-			at_lo.s_net  = 0;
-
-			rt = atrtr_find(&at_lo);
-			if (!rt) {
+			if (!rt_lo) {
 				kfree_skb(skb);
 				err = -ENETUNREACH;
 				goto out;
 			}
-			dev = rt->dev;
+			dev = rt_lo->dev;
 			skb->dev = dev;
 		}
 		ddp_dl->request(ddp_dl, skb, dev->dev_addr);
-- 
2.25.1

Re: [PATCH] appletalk: Fix skb allocation size in loopback case

[patchwork-bot+netdevbpf]

Hello:

This patch was applied to netdev/net.git (refs/heads/master):

On Thu, 11 Feb 2021 21:27:54 -0800 you wrote:
> If a DDP broadcast packet is sent out to a non-gateway target, it is
> also looped back. There is a potential for the loopback device to have a
> longer hardware header length than the original target route's device,
> which can result in the skb not being created with enough room for the
> loopback device's hardware header. This patch fixes the issue by
> determining that a loopback will be necessary prior to allocating the
> skb, and if so, ensuring the skb has enough room.
> 
> [...]

Here is the summary with links:
  - appletalk: Fix skb allocation size in loopback case
    https://git.kernel.org/netdev/net/c/39935dccb21c

You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html