From: Brian Haley <brian.haley@hp.com>
To: David Miller <davem@davemloft.net>,
YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Cc: "netdev@vger.kernel.org" <netdev@vger.kernel.org>
Subject: [PATCH] IPv6: fix IPV6_RECVERR handling of locally-generated errors
Date: Mon, 03 May 2010 21:44:27 -0400 [thread overview]
Message-ID: <4BDF7BFB.1070207@hp.com> (raw)
I noticed when I added support for IPV6_DONTFRAG that if you set
IPV6_RECVERR and tried to send a UDP packet larger than 64K to an
IPv6 destination, you'd correctly get an EMSGSIZE, but reading from
MSG_ERRQUEUE returned the incorrect address in the cmsg:
struct msghdr:
msg_name 0x7fff8f3c96d0
msg_namelen 28
struct sockaddr_in6:
sin6_family 10
sin6_port 7639
sin6_flowinfo 0
sin6_addr ::ffff:38.32.0.0
sin6_scope_id 0 ((null))
It should have returned this in my case:
struct msghdr:
msg_name 0x7fffd866b510
msg_namelen 28
struct sockaddr_in6:
sin6_family 10
sin6_port 7639
sin6_flowinfo 0
sin6_addr 2620:0:a09:e000:21f:29ff:fe57:f88b
sin6_scope_id 0 ((null))
The problem is that ipv6_recv_error() assumes that if the error
wasn't generated by ICMPv6, it's an IPv4 address sitting there,
and proceeds to create a v4-mapped address from it.
--
Change ipv6_icmp_error() and ipv6_local_error() to set skb->protocol
to htons(ETH_P_IPV6) so that ipv6_recv_error() knows the address
sitting right after the extended error is IPv6, else it will
incorrectly map the first octet into an IPv4-mapped IPv6 address
in the cmsg structure returned in a recvmsg() call to obtain
the error.
Signed-off-by: Brian Haley <brian.haley@hp.com>
diff --git a/net/ipv6/datagram.c b/net/ipv6/datagram.c
index 5959230..7126846 100644
--- a/net/ipv6/datagram.c
+++ b/net/ipv6/datagram.c
@@ -222,6 +222,8 @@ void ipv6_icmp_error(struct sock *sk, struct sk_buff *skb, int err,
if (!skb)
return;
+ skb->protocol = htons(ETH_P_IPV6);
+
serr = SKB_EXT_ERR(skb);
serr->ee.ee_errno = err;
serr->ee.ee_origin = SO_EE_ORIGIN_ICMP6;
@@ -255,6 +257,8 @@ void ipv6_local_error(struct sock *sk, int err, struct flowi *fl, u32 info)
if (!skb)
return;
+ skb->protocol = htons(ETH_P_IPV6);
+
skb_put(skb, sizeof(struct ipv6hdr));
skb_reset_network_header(skb);
iph = ipv6_hdr(skb);
@@ -358,7 +362,7 @@ int ipv6_recv_error(struct sock *sk, struct msghdr *msg, int len)
sin->sin6_flowinfo = 0;
sin->sin6_port = serr->port;
sin->sin6_scope_id = 0;
- if (serr->ee.ee_origin == SO_EE_ORIGIN_ICMP6) {
+ if (skb->protocol == htons(ETH_P_IPV6)) {
ipv6_addr_copy(&sin->sin6_addr,
(struct in6_addr *)(nh + serr->addr_offset));
if (np->sndflow)
@@ -380,7 +384,7 @@ int ipv6_recv_error(struct sock *sk, struct msghdr *msg, int len)
sin->sin6_family = AF_INET6;
sin->sin6_flowinfo = 0;
sin->sin6_scope_id = 0;
- if (serr->ee.ee_origin == SO_EE_ORIGIN_ICMP6) {
+ if (skb->protocol == htons(ETH_P_IPV6)) {
ipv6_addr_copy(&sin->sin6_addr, &ipv6_hdr(skb)->saddr);
if (np->rxopt.all)
datagram_recv_ctl(sk, msg, skb);
reply other threads:[~2010-05-04 1:44 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=4BDF7BFB.1070207@hp.com \
--to=brian.haley@hp.com \
--cc=davem@davemloft.net \
--cc=netdev@vger.kernel.org \
--cc=yoshfuji@linux-ipv6.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).