[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <faba5cf5-8659-e953-d828-ad1f472ac0c9@lab.ntt.co.jp>
Date: Thu, 26 Apr 2018 19:46:26 +0900
From: Toshiaki Makita <makita.toshiaki@....ntt.co.jp>
To: Jesper Dangaard Brouer <brouer@...hat.com>
Cc: Toshiaki Makita <toshiaki.makita1@...il.com>,
netdev@...r.kernel.org
Subject: Re: [PATCH RFC 2/9] veth: Add driver XDP
Hi Jesper,
Thanks for taking a look!
On 2018/04/26 5:39, Jesper Dangaard Brouer wrote:
> On Tue, 24 Apr 2018 23:39:16 +0900
> Toshiaki Makita <toshiaki.makita1@...il.com> wrote:
>
>> This is basic implementation of veth driver XDP.
>>
>> Incoming packets are sent from the peer veth device in the form of skb,
>> so this is generally doing the same thing as generic XDP.
>
> I'm unsure that context you are calling veth_xdp_rcv_skb() from. The
> XDP (RX side) depend heavily on the protection provided by NAPI context.
> It looks like you are adding NAPI handler later.
This is called from softirq or bh disabled context.
I can see XDP REDIRECT depends on NAPI since it uses per-cpu temporary
storage which is used in ndo_xdp_flush. I thought DROP and PASS is safe
here. Also this is basically the same context as generic XDP, which is
called from netif_rx_internal.
Anyway this is a temporary state and not needed. It looks like this does
not help review so I'll squash this and patch 4 (napi patch).
--
Toshiaki Makita
Powered by blists - more mailing lists