Date: Tue, 27 Dec 2016 16:29:05 +0800 (CST)
From: "wei zhang" <asuka.com@163.com>
To: kuznet@ms2.inr.ac.ru, davem@davemloft.net, jmorris@namei.org,
        yoshfuji@linux-ipv6.org, kaber@trash.net
Cc: netdev@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re:[PATCH] net: fix incorrect original ingress device index in
 PKTINFO
In-Reply-To: <1482825138-2125-1-git-send-email-asuka.com@163.com>
References: <1482825138-2125-1-git-send-email-asuka.com@163.com>
Content-Type: text/plain; charset=GBK
MIME-Version: 1.0
Message-ID: <64f198a9.a52b.1593f65ad7e.Coremail.asuka.com@163.com>
Sender: linux-kernel-owner@vger.kernel.org
Content-Transfer-Encoding: 8bit
Content-Length: 2073
Lines: 48

At 2016-12-27 15:52:18, "Wei Zhang" <asuka.com@163.com> wrote:
>When we send a packet for our own local address on a non-loopback interface
>(e.g. eth0), due to the change had been introduced from commit 0b922b7a829c
>("net: original ingress device index in PKTINFO"), the original ingress
>device index would be set as the loopback interface. However, the packet
>should be considered as if it is being arrived via the sending interface
>(eth0), otherwise it would break the expectation of the userspace
>application (e.g. the DHCPRELEASE message from dhcp_release binary would
>be ignored by the dnsmasq daemon)
>
>Signed-off-by: Wei Zhang <asuka.com@163.com>
>---
> net/ipv4/ip_sockglue.c | 7 ++++++-
> 1 file changed, 6 insertions(+), 1 deletion(-)
>
>diff --git a/net/ipv4/ip_sockglue.c b/net/ipv4/ip_sockglue.c
>index b8a2d63..b6a6d35 100644
>--- a/net/ipv4/ip_sockglue.c
>+++ b/net/ipv4/ip_sockglue.c
>@@ -1202,8 +1202,13 @@ void ipv4_pktinfo_prepare(const struct sock *sk, struct sk_buff *skb)
> 		 * which has interface index (iif) as the first member of the
> 		 * underlying inet{6}_skb_parm struct. This code then overlays
> 		 * PKTINFO_SKB_CB and in_pktinfo also has iif as the first
>-		 * element so the iif is picked up from the prior IPCB
>+		 * element so the iif is picked up from the prior IPCB except
>+		 * iif is loopback interface which the packet should be 
>+		 * considered as if it is being arrived via the sending interface
> 		 */
>+		if (pktinfo->ipi_ifindex == LOOPBACK_IFINDEX) {
>+			pktinfo->ipi_ifindex = inet_iif(skb);
>+		}
> 		pktinfo->ipi_spec_dst.s_addr = fib_compute_spec_dst(skb);
> 	} else {
> 		pktinfo->ipi_ifindex = 0;
>-- 
>1.8.3.1
>

  When I upgrade to the 4.9, the dhcp_release could not release the dhcp
lease, the dnsmasq ignored all the DHCPRELEASE message which it think come
from lo. I think this is due to the commit 0b922b7a829c, which set the
IPCB(skb)->iif = skb->skb_iif in the ip_rcv()!

  And I'm very sorry about forgetting checkpatch, I will resend the patch,
hope I'm not bothering you!

Thanks,
Wei Zhang