| Message ID | 1478711246-17440-1-git-send-email-dsa@cumulusnetworks.com |
|---|---|
| State | Accepted, archived |
| Delegated to: | David Miller |
| Headers | show |
On Thu, Nov 10, 2016 at 2:07 AM, David Ahern <dsa@cumulusnetworks.com> wrote: > Revert the change to ip_send_unicast_reply and tcp_v6_send_response such > that the oif in the flow is set to the skb_iif only if skb_iif is an L3 > master. This fixes the IPv4 and IPv6 tests, thanks! Tested-by: Lorenzo Colitti <lorenzo@google.com> Acked-by: Lorenzo Colitti <lorenzo@google.com>
From: David Ahern <dsa@cumulusnetworks.com> Date: Wed, 9 Nov 2016 09:07:26 -0800 > Lorenzo noted an Android unit test failed due to e0d56fdd7342: > "The expectation in the test was that the RST replying to a SYN sent to a > closed port should be generated with oif=0. In other words it should not > prefer the interface where the SYN came in on, but instead should follow > whatever the routing table says it should do." > > Revert the change to ip_send_unicast_reply and tcp_v6_send_response such > that the oif in the flow is set to the skb_iif only if skb_iif is an L3 > master. > > Fixes: e0d56fdd7342 ("net: l3mdev: remove redundant calls") > Reported-by: Lorenzo Colitti <lorenzo@google.com> > Signed-off-by: David Ahern <dsa@cumulusnetworks.com> Applied, thanks David.
diff --git a/net/ipv4/ip_output.c b/net/ipv4/ip_output.c index 03e7f7310423..7cf627b0b248 100644 --- a/net/ipv4/ip_output.c +++ b/net/ipv4/ip_output.c @@ -1579,7 +1579,8 @@ void ip_send_unicast_reply(struct sock *sk, struct sk_buff *skb, } oif = arg->bound_dev_if; - oif = oif ? : skb->skb_iif; + if (!oif && netif_index_is_l3_master(net, skb->skb_iif)) + oif = skb->skb_iif; flowi4_init_output(&fl4, oif, IP4_REPLY_MARK(net, skb->mark), diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 5a27ab4eab39..6ca23c2e76f7 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -818,8 +818,12 @@ static void tcp_v6_send_response(const struct sock *sk, struct sk_buff *skb, u32 fl6.flowi6_proto = IPPROTO_TCP; if (rt6_need_strict(&fl6.daddr) && !oif) fl6.flowi6_oif = tcp_v6_iif(skb); - else - fl6.flowi6_oif = oif ? : skb->skb_iif; + else { + if (!oif && netif_index_is_l3_master(net, skb->skb_iif)) + oif = skb->skb_iif; + + fl6.flowi6_oif = oif; + } fl6.flowi6_mark = IP6_REPLY_MARK(net, skb->mark); fl6.fl6_dport = t1->dest;
Lorenzo noted an Android unit test failed due to e0d56fdd7342: "The expectation in the test was that the RST replying to a SYN sent to a closed port should be generated with oif=0. In other words it should not prefer the interface where the SYN came in on, but instead should follow whatever the routing table says it should do." Revert the change to ip_send_unicast_reply and tcp_v6_send_response such that the oif in the flow is set to the skb_iif only if skb_iif is an L3 master. Fixes: e0d56fdd7342 ("net: l3mdev: remove redundant calls") Reported-by: Lorenzo Colitti <lorenzo@google.com> Signed-off-by: David Ahern <dsa@cumulusnetworks.com> --- net/ipv4/ip_output.c | 3 ++- net/ipv6/tcp_ipv6.c | 8 ++++++-- 2 files changed, 8 insertions(+), 3 deletions(-)