From: Eric Dumazet <eric.dumazet@gmail.com>
To: Herbert Xu <herbert@gondor.apana.org.au>
Cc: davem@davemloft.net, netdev@vger.kernel.org,
Patrick McHardy <kaber@trash.net>
Subject: Re: [PATCH] ip_gre: CONFIG_IPV6_MODULE support
Date: Wed, 29 Sep 2010 06:18:30 +0200 [thread overview]
Message-ID: <1285733911.22570.20.camel@edumazet-laptop> (raw)
In-Reply-To: <1285663738.2607.30.camel@edumazet-laptop>
Le mardi 28 septembre 2010 à 10:48 +0200, Eric Dumazet a écrit :
> Le mardi 28 septembre 2010 à 17:41 +0900, Herbert Xu a écrit :
> > Eric Dumazet <eric.dumazet@gmail.com> wrote:
> > > ipv6 can be a module, we should test CONFIG_IPV6 and CONFIG_IPV6_MODULE
> > > to enable ipv6 bits in ip_gre.
> > >
> > > Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
> >
> > To do this safely, we also need to disallow IPV6=m and GRE=y.
> >
> > Cheers,
>
I even remember Patrick warned about this 30 months ago.... Doh...
http://www.kerneltrap.com/mailarchive/linux-netdev/2008/1/22/590393
Given its now rare ipv6 is not loaded, what about following patch ?
Thanks !
Tested here, with a non loaded ipv6 module.
[PATCH] ipv6: introduce call_icmpv6_send()
ip_gre module has a dependency against IPv6:
It needs to call icmpv6_send(), while this symbol is available only if
IPv6 module is loaded (or static)
Introduce call_icmpv6_send(), provided in core network, to indirectly
call icmpv6_send() function if available.
Reported-by: Patrick McHardy <kaber@trash.net>
Reported-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
---
include/net/icmp.h | 6 ++++++
net/ipv4/icmp.c | 33 +++++++++++++++++++++++++++++++++
net/ipv4/ip_gre.c | 2 +-
net/ipv6/icmp.c | 2 ++
4 files changed, 42 insertions(+), 1 deletion(-)
diff --git a/include/net/icmp.h b/include/net/icmp.h
index 6e991e0..ec45a4f 100644
--- a/include/net/icmp.h
+++ b/include/net/icmp.h
@@ -48,4 +48,10 @@ extern void icmp_out_count(struct net *net, unsigned char type);
/* Move into dst.h ? */
extern int xrlim_allow(struct dst_entry *dst, int timeout);
+extern int register_icmpv6_send(void (*func)(struct sk_buff *skb, u8 type,
+ u8 code, __u32 info));
+extern int unregister_icmpv6_send(void);
+extern void call_icmpv6_send(struct sk_buff *skb, u8 type,
+ u8 code, __u32 info);
+
#endif /* _ICMP_H */
diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c
index a0d847c..4c5b817 100644
--- a/net/ipv4/icmp.c
+++ b/net/ipv4/icmp.c
@@ -1217,3 +1217,36 @@ int __init icmp_init(void)
{
return register_pernet_subsys(&icmp_sk_ops);
}
+
+
+/* wrappers to call icmpv6_send() if ipv6 module is loaded */
+
+static void __rcu (*icmpv6_send_ptr)(struct sk_buff *skb, u8 type, u8 code,
+ __u32 info);
+
+int register_icmpv6_send(void (*func)(struct sk_buff *skb, u8 type, u8 code,
+ __u32 info))
+{
+ rcu_assign_pointer(icmpv6_send_ptr, func);
+ return 0;
+}
+EXPORT_SYMBOL(register_icmpv6_send);
+
+int unregister_icmpv6_send(void)
+{
+ rcu_assign_pointer(icmpv6_send_ptr, NULL);
+ synchronize_rcu();
+ return 0;
+}
+EXPORT_SYMBOL(unregister_icmpv6_send);
+
+/* called with rcu_read_lock */
+void call_icmpv6_send(struct sk_buff *skb, u8 type, u8 code, __u32 info)
+{
+ void (*func)(struct sk_buff *skb, u8 type, u8 code, __u32 info);
+
+ func = rcu_dereference(icmpv6_send_ptr);
+ if (func)
+ (*func)(skb, type, code, info);
+}
+EXPORT_SYMBOL(call_icmpv6_send);
diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c
index 35c93e8..6cbd37c 100644
--- a/net/ipv4/ip_gre.c
+++ b/net/ipv4/ip_gre.c
@@ -788,7 +788,7 @@ static netdev_tx_t ipgre_tunnel_xmit(struct sk_buff *skb, struct net_device *dev
}
if (mtu >= IPV6_MIN_MTU && mtu < skb->len - tunnel->hlen + gre_hlen) {
- icmpv6_send(skb, ICMPV6_PKT_TOOBIG, 0, mtu);
+ call_icmpv6_send(skb, ICMPV6_PKT_TOOBIG, 0, mtu);
ip_rt_put(rt);
goto tx_error;
}
diff --git a/net/ipv6/icmp.c b/net/ipv6/icmp.c
index 03e62f9..25e9ad6 100644
--- a/net/ipv6/icmp.c
+++ b/net/ipv6/icmp.c
@@ -862,6 +862,7 @@ int __init icmpv6_init(void)
err = -EAGAIN;
if (inet6_add_protocol(&icmpv6_protocol, IPPROTO_ICMPV6) < 0)
goto fail;
+ register_icmpv6_send(icmpv6_send);
return 0;
fail:
@@ -874,6 +875,7 @@ void icmpv6_cleanup(void)
{
unregister_pernet_subsys(&icmpv6_sk_ops);
inet6_del_protocol(&icmpv6_protocol, IPPROTO_ICMPV6);
+ unregister_icmpv6_send();
}
next prev parent reply other threads:[~2010-09-29 4:18 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-09-20 13:01 [PATCH] ip_gre: CONFIG_IPV6_MODULE support Eric Dumazet
2010-09-20 17:06 ` David Miller
2010-09-28 8:41 ` Herbert Xu
2010-09-28 8:48 ` Eric Dumazet
2010-09-29 4:18 ` Eric Dumazet [this message]
2010-09-29 4:24 ` Eric Dumazet
2010-09-29 4:36 ` Herbert Xu
2010-09-29 4:41 ` David Miller
2010-09-29 5:38 ` David Miller
2010-09-29 6:47 ` Eric Dumazet
2010-09-29 7:25 ` David Miller
2010-09-29 7:45 ` Eric Dumazet
2010-09-29 7:57 ` Herbert Xu
2010-09-29 8:00 ` David Miller
2010-09-29 8:03 ` Herbert Xu
2010-09-29 8:22 ` Eric Dumazet
2010-09-29 7:59 ` David Miller
2010-09-29 6:24 ` Eric Dumazet
2010-09-29 7:11 ` Herbert Xu
2010-09-29 7:26 ` David Miller
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=1285733911.22570.20.camel@edumazet-laptop \
--to=eric.dumazet@gmail.com \
--cc=davem@davemloft.net \
--cc=herbert@gondor.apana.org.au \
--cc=kaber@trash.net \
--cc=netdev@vger.kernel.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