diff mbox

[net-next] ip6_tunnel: Fix missing tunnel encapsulation limit option

Message ID 20170426170707.165201-1-kraigatgoog@gmail.com
State Superseded, archived
Delegated to: David Miller
Headers show

Commit Message

Craig Gallek April 26, 2017, 5:07 p.m. UTC 
From: Craig Gallek <kraig@google.com>

The IPv6 tunneling code tries to insert IPV6_TLV_TNL_ENCAP_LIMIT and
IPV6_TLV_PADN options when an encapsulation limit is defined (the
default is a limit of 4).  An MTU adjustment is done to account for
these options as well.  However, the options are never present in the
generated packets.

ipv6_push_nfrag_opts requires that IPV6_RTHDR be present in order to
include any IPV6_DSTOPTS options.  The v6 tunnel code does not
use routing options, so the encap limit options are not included.

A brief reading of RFC 3542 section 9.2 (specifically the 4th paragraph)
makes me believe that this requirement in the kernel is incorrect.

Fixes: 333fad5364d6: ("[IPV6]: Support several new sockopt / ancillary data in Advanced API (RFC3542)")
Signed-off-by: Craig Gallek <kraig@google.com>
---
 net/ipv6/exthdrs.c | 13 +++++--------
 1 file changed, 5 insertions(+), 8 deletions(-)

Comments

Craig Gallek April 26, 2017, 5:59 p.m. UTC | #1 
On Wed, Apr 26, 2017 at 1:07 PM, Craig Gallek <kraigatgoog@gmail.com> wrote:
> From: Craig Gallek <kraig@google.com>
>
> The IPv6 tunneling code tries to insert IPV6_TLV_TNL_ENCAP_LIMIT and
> IPV6_TLV_PADN options when an encapsulation limit is defined (the
> default is a limit of 4).  An MTU adjustment is done to account for
> these options as well.  However, the options are never present in the
> generated packets.
>
> ipv6_push_nfrag_opts requires that IPV6_RTHDR be present in order to
> include any IPV6_DSTOPTS options.  The v6 tunnel code does not
> use routing options, so the encap limit options are not included.
>
> A brief reading of RFC 3542 section 9.2 (specifically the 4th paragraph)
> makes me believe that this requirement in the kernel is incorrect.
Looking more closely, I think I'm wrong here.  Specifically, the cmsg
parser puts IPV6_RTHDRDSTOPTS in dst0opt and IPV6_DSTOPTS in dst1opt.
The tunnel code is currently building dst0opt and using
ipv6_push_nfrag_opts.  Perhaps it should be building dst1opt and
calling ipv6_push_frag_opts?
diff mbox

Patch

diff --git a/net/ipv6/exthdrs.c b/net/ipv6/exthdrs.c
index 25192a3b0cd7..224a89e68a42 100644
--- a/net/ipv6/exthdrs.c
+++ b/net/ipv6/exthdrs.c
@@ -932,15 +932,12 @@  void ipv6_push_nfrag_opts(struct sk_buff *skb, struct ipv6_txoptions *opt,
 			  u8 *proto,
 			  struct in6_addr **daddr, struct in6_addr *saddr)
 {
-	if (opt->srcrt) {
+	if (opt->srcrt)
 		ipv6_push_rthdr(skb, proto, opt->srcrt, daddr, saddr);
-		/*
-		 * IPV6_RTHDRDSTOPTS is ignored
-		 * unless IPV6_RTHDR is set (RFC3542).
-		 */
-		if (opt->dst0opt)
-			ipv6_push_exthdr(skb, proto, NEXTHDR_DEST, opt->dst0opt);
-	}
+
+	if (opt->dst0opt)
+		ipv6_push_exthdr(skb, proto, NEXTHDR_DEST, opt->dst0opt);
+
 	if (opt->hopopt)
 		ipv6_push_exthdr(skb, proto, NEXTHDR_HOP, opt->hopopt);
 }