From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <netdev-owner@vger.kernel.org>
Received: from mx3-rdu2.redhat.com ([66.187.233.73]:40880 "EHLO mx1.redhat.com"
        rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP
        id S1751726AbeB0C3K (ORCPT <rfc822;netdev@vger.kernel.org>);
        Mon, 26 Feb 2018 21:29:10 -0500
Subject: Re: [net PATCH 1/4] virtio_net: disable XDP_REDIRECT in
 receive_mergeable() case
From: Jason Wang <jasowang@redhat.com>
To: "Michael S. Tsirkin" <mst@redhat.com>,
        Jesper Dangaard Brouer <brouer@redhat.com>
Cc: netdev@vger.kernel.org, John Fastabend <john.fastabend@gmail.com>,
        Alexei Starovoitov <alexei.starovoitov@gmail.com>,
        Saeed Mahameed <saeedm@mellanox.com>,
        Daniel Borkmann <borkmann@iogearbox.net>,
        "David S. Miller" <davem@davemloft.net>,
        Tariq Toukan <tariqt@mellanox.com>
References: <151913348634.28247.17519468037896960567.stgit@firesoul>
 <151913352486.28247.12339812865877070160.stgit@firesoul>
 <20180227023529-mutt-send-email-mst@kernel.org>
 <86ea9887-0ea2-1f58-7710-26d4d126a76d@redhat.com>
Message-ID: <71fc11a2-ea95-766c-0d7a-2433abd9dcbc@redhat.com>
Date: Tue, 27 Feb 2018 10:28:58 +0800
MIME-Version: 1.0
In-Reply-To: <86ea9887-0ea2-1f58-7710-26d4d126a76d@redhat.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 8bit
Content-Language: en-US
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>



On 2018年02月27日 10:25, Jason Wang wrote:
>
>
> On 2018年02月27日 08:40, Michael S. Tsirkin wrote:
>>> IMHO we should consider NOT supporting XDP in receive_mergeable() at
>>> all, because the principles behind XDP are to gain speed by (1) code
>>> simplicity, (2) sacrificing memory and (3) where possible moving
>>> runtime checks to setup time.  These principles are clearly being
>>> violated in receive_mergeable(), that e.g. runtime track average
>>> buffer size to save memory consumption.
>>>
>>> In the longer run, we should consider introducing a separate receive
>>> function when attaching an XDP program, and also change the memory
>>> model to be compatible with XDP when attaching an XDP prog.
>> I agree with a separate function approach.
>>
>> So each buffer is tagged as xdp/non xdp, we check that
>> and handle appropriately - where non xdp could be handled
>> by the generic path.
>
> If we want to have separated function, should we do it for all XDP 
> capable drivers instead of virtio-net only?
>
> Thanks
>

What's more, since we don't stop device during XDP set. Even if a buffer 
is tagged as xdp/non xdp during refill, we still can't make sure whether 
or not XDP was used at receive function. So we can not handle them 
separately.

Thanks