netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH net v2] ipv6: socket SO_BINDTODEVICE lookup routing fail without IPv6 rule.
@ 2025-01-03  5:43 Shiming Cheng
  2025-01-03 10:27 ` Shiming Cheng (成诗明)
  2025-01-04  2:14 ` Jakub Kicinski
  0 siblings, 2 replies; 6+ messages in thread
From: Shiming Cheng @ 2025-01-03  5:43 UTC (permalink / raw)
  To: willemdebruijn.kernel, davem, dsahern, edumazet, kuba, pabeni,
	horms, matthias.bgg, angelogioacchino.delregno, linux-kernel,
	linux-arm-kernel, linux-mediatek
  Cc: netdev, lena.wang, Shiming Cheng

When using socket IPv6 with SO_BINDTODEVICE, if IPv6 rule is not
matched, it will return ENETUNREACH. In fact, IPv4 does not behave
this way. IPv4 prioritizes looking up IP rules for routing and
forwarding, if not matched it will use socket-bound out interface
to send packets. The modification here is to make IPv6 behave the
same as IPv4. If IP rule is not found, it will also use socket-bound
out interface to send packts.

Fixes: 6f21c96a78b8 ("ipv6: enforce flowi6_oif usage in ip6_dst_lookup_tail()")
Signed-off-by: Shiming Cheng <shiming.cheng@mediatek.com>
---
 include/net/ip6_route.h |  2 ++
 net/ipv6/ip6_output.c   |  7 ++++++-
 net/ipv6/route.c        | 34 ++++++++++++++++++++++++++++++++++
 3 files changed, 42 insertions(+), 1 deletion(-)

diff --git a/include/net/ip6_route.h b/include/net/ip6_route.h
index 6dbdf60b342f..0625597def6f 100644
--- a/include/net/ip6_route.h
+++ b/include/net/ip6_route.h
@@ -214,6 +214,8 @@ void rt6_multipath_rebalance(struct fib6_info *f6i);
 
 void rt6_uncached_list_add(struct rt6_info *rt);
 void rt6_uncached_list_del(struct rt6_info *rt);
+struct rt6_info *ip6_create_rt_oif_rcu(struct net *net, const struct sock *sk,
+		struct flowi6 *fl6, int flags);
 
 static inline const struct rt6_info *skb_rt6_info(const struct sk_buff *skb)
 {
diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
index f7b4608bb316..95728c8921cb 100644
--- a/net/ipv6/ip6_output.c
+++ b/net/ipv6/ip6_output.c
@@ -1156,8 +1156,13 @@ static int ip6_dst_lookup_tail(struct net *net, const struct sock *sk,
 		*dst = ip6_route_output_flags(net, sk, fl6, flags);
 
 	err = (*dst)->error;
-	if (err)
+	if (err && (flags & RT6_LOOKUP_F_IFACE)) {
+		*dst = (struct dst_entry *)ip6_create_rt_oif_rcu(net, sk, fl6, flags);
+		if (!*dst)
+			goto out_err_release;
+	} else if (err) {
 		goto out_err_release;
+	}
 
 #ifdef CONFIG_IPV6_OPTIMISTIC_DAD
 	/*
diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index 67ff16c04718..7d7450fab44f 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -1214,6 +1214,40 @@ static struct rt6_info *ip6_create_rt_rcu(const struct fib6_result *res)
 	return nrt;
 }
 
+struct rt6_info *ip6_create_rt_oif_rcu(struct net *net, const struct sock *sk,
+				       struct flowi6 *fl6, int flags)
+{
+	struct rt6_info *rt;
+	unsigned int prefs;
+	int err;
+	struct net_device *dev = dev_get_by_index_rcu(net, fl6->flowi6_oif);
+
+	if (!dev)
+		return NULL;
+	rt = ip6_dst_alloc(dev_net(dev), dev, flags);
+
+	if (!rt)
+		return NULL;
+	rt->dst.error = 0;
+	rt->dst.output = ip6_output;
+	rt->dst.lastuse = jiffies;
+	prefs = sk ? inet6_sk(sk)->srcprefs : 0;
+	err = ipv6_dev_get_saddr(net, dev, &fl6->daddr, prefs, &fl6->saddr);
+
+	if (err) {
+		dst_release(&rt->dst);
+		return NULL;
+	}
+	rt->rt6i_dst.addr = fl6->daddr;
+	rt->rt6i_dst.plen = 128;
+	rt->rt6i_src.addr = fl6->saddr;
+	rt->rt6i_dst.plen = 128;
+	rt->rt6i_idev = in6_dev_get(dev);
+	rt->rt6i_flags = flags;
+	return rt;
+}
+EXPORT_SYMBOL_GPL(ip6_create_rt_oif_rcu);
+
 INDIRECT_CALLABLE_SCOPE struct rt6_info *ip6_pol_route_lookup(struct net *net,
 					     struct fib6_table *table,
 					     struct flowi6 *fl6,
-- 
2.45.2


^ permalink raw reply related	[flat|nested] 6+ messages in thread
end of thread, other threads:[~2025-01-19 23:21 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-01-03  5:43 [PATCH net v2] ipv6: socket SO_BINDTODEVICE lookup routing fail without IPv6 rule Shiming Cheng
2025-01-03 10:27 ` Shiming Cheng (成诗明)
2025-01-03 15:31   ` David Ahern
2025-01-14  3:11     ` Shiming Cheng (成诗明)
2025-01-19 23:21       ` David Ahern
2025-01-04  2:14 ` Jakub Kicinski

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).