[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-Id: <20181117.200618.1664143348881472089.davem@davemloft.net>
Date: Sat, 17 Nov 2018 20:06:18 -0800 (PST)
From: David Miller <davem@...emloft.net>
To: eric.dumazet@...il.com
Cc: lirongqing@...du.com, netdev@...r.kernel.org
Subject: Re: [PATCH][net-next] net: slightly optimize eth_type_trans
From: Eric Dumazet <eric.dumazet@...il.com>
Date: Sat, 17 Nov 2018 17:19:34 -0800
> I might have been not very clear.
>
> Issue is that with macvlan and GRO-friend traffic we can receive the following packets :
>
> P1-P2 with PACKET_OTHERHOST because eth_type_trans() detected the dst MAC is not the eth0 device mac address.
>
> P2 has been aggregated to P1, so the sk_buff has been put to napi->skb via napi_reuse_skb()
>
> Then we receive on same NAPI packet P3, for this host, reusing napi->skb that was saved (old P2 sk_buff).
> skb->pkt_type is PACKET_OTHERHOST.
>
> eth_type_trans() does not change skb->pkt_type because ethernet dst mac address is our ethernet mac address.
>
> -> We feed the upper stack with P3, with incorrect pkt_type.
>
> -> packet is dropped because pkt_type != PACKET_HOST, for example in tcp_v4_rcv()
Oh I see, it is about defaults when using skb freshly from cache for another new packet.
Yes, I completely agree with your suggested patch.
Thanks for explaining.
Powered by blists - more mailing lists