Re: [PATCH][2.5.22] OOPS in tcp_v6_get_port

Em Mon, Jun 17, 2002 at 02:33:19PM -0700, David S. Miller escreveu:

This is a known bug introduced by the struct sock splitup into
external per-protocol pieces done by Arnaldo de Melo.  He is working
on the proper fix, your proposed change will just paper over the real
bug.

Carl,

	Can you try this patch?

- Arnaldo

--- orig/net/ipv6/tcp_ipv6.c	Sat May 25 23:13:56 2002
+++ linux/net/ipv6/tcp_ipv6.c	Fri Jun 14 23:23:07 2002

@@ -1240,6 +1240,7 @@
 					  struct dst_entry *dst)
 {
 	struct ipv6_pinfo *newnp, *np = inet6_sk(sk);
+	struct tcp6_sock *newtcp6sk;
 	struct flowi fl;
 	struct inet_opt *newinet;
 	struct tcp_opt *newtp;

@@ -1256,10 +1257,15 @@
 		if (newsk == NULL) 
 			return NULL;
 
+		newtcp6sk = (struct tcp6_sock *)newsk;
+		newtcp6sk->pinet6 = &newtcp6sk->inet6;
+
 		newinet = inet_sk(newsk);
 		newnp = inet6_sk(newsk);
 		newtp = tcp_sk(newsk);
 
+		memcpy(newnp, np, sizeof(struct ipv6_pinfo));
+
 		ipv6_addr_set(&newnp->daddr, 0, 0, htonl(0x0000FFFF),
 			      newinet->daddr);
 

@@ -1336,9 +1342,15 @@
 	ip6_dst_store(newsk, dst, NULL);
 	sk->route_caps = dst->dev->features&~NETIF_F_IP_CSUM;
 
+	newtcp6sk = (struct tcp6_sock *)newsk;
+	newtcp6sk->pinet6 = &newtcp6sk->inet6;
+
 	newtp = tcp_sk(newsk);
 	newinet = inet_sk(newsk);
 	newnp = inet6_sk(newsk);
+
+	memcpy(newnp, np, sizeof(struct ipv6_pinfo));
+
 	ipv6_addr_copy(&newnp->daddr, &req->af.v6_req.rmt_addr);
 	ipv6_addr_copy(&newnp->saddr, &req->af.v6_req.loc_addr);
 	ipv6_addr_copy(&newnp->rcv_saddr, &req->af.v6_req.loc_addr);

-- 
Russell King (rmk@arm.linux.org.uk)                The developer of ARM Linux
             http://www.arm.linux.org.uk/personal/aboutme.html

   From: Arnaldo Carvalho de Melo [off-list ref]
   Date: Mon, 17 Jun 2002 21:57:35 -0300

   --- orig/net/ipv6/tcp_ipv6.c	Sat May 25 23:13:56 2002
   +++ linux/net/ipv6/tcp_ipv6.c	Fri Jun 14 23:23:07 2002

I've installed this change into my tree in the meantime.
If we find a better fix, we can just revert this.

Thanks.

On Mon, 17 Jun 2002, Arnaldo Carvalho de Melo wrote:

Em Mon, Jun 17, 2002 at 02:33:19PM -0700, David S. Miller escreveu:

quoted

This is a known bug introduced by the struct sock splitup into
external per-protocol pieces done by Arnaldo de Melo.  He is working
on the proper fix, your proposed change will just paper over the real
bug.

I suspected something like that.

This patch doesn't help, it produces exactly the same oops when decoded, I 
assume the newer patch you mention to Dave is the correct fix?

Thanks,
Carl