Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Tue, 14 May 2019 18:10:41 +0200
From:   Stefano Garzarella <sgarzare@redhat.com>
To:     Jason Wang <jasowang@redhat.com>
Cc:     netdev@vger.kernel.org, "David S. Miller" <davem@davemloft.net>,
        "Michael S. Tsirkin" <mst@redhat.com>,
        virtualization@lists.linux-foundation.org,
        linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
        Stefan Hajnoczi <stefanha@redhat.com>
Subject: Re: [PATCH v2 8/8] vsock/virtio: make the RX buffer size tunable
Message-ID: <20190514161041.y4exigcwwys34naf@steredhat>
References: <20190510125843.95587-1-sgarzare@redhat.com>
 <20190510125843.95587-9-sgarzare@redhat.com>
 <eddb5a89-ed44-3a65-0181-84f7f27dd2cb@redhat.com>
 <8e72ef5e-cf6a-a635-3f76-bdeac95761b8@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <8e72ef5e-cf6a-a635-3f76-bdeac95761b8@redhat.com>
User-Agent: NeoMutt/20180716
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Mon, May 13, 2019 at 08:46:19PM +0800, Jason Wang wrote:
> 
> On 2019/5/13 下午6:05, Jason Wang wrote:
> > 
> > On 2019/5/10 下午8:58, Stefano Garzarella wrote:
> > > The RX buffer size determines the memory consumption of the
> > > vsock/virtio guest driver, so we make it tunable through
> > > a module parameter.
> > > 
> > > The size allowed are between 4 KB and 64 KB in order to be
> > > compatible with old host drivers.
> > > 
> > > Suggested-by: Stefan Hajnoczi <stefanha@redhat.com>
> > > Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
> > 
> > 
> > I don't see much value of doing this through kernel command line. We
> > should deal with them automatically like what virtio-net did. Or even a
> > module parameter is better.
> > 
> > Thanks
> 
> 
> Sorry, I misread the patch. But even module parameter is something not
> flexible enough. We should deal with them transparently.
> 

Okay, I'll try to understand how we can automatically adapt the RX
buffer size. Since the flow is stream based, the receiver doesn't know the
original packet size.

Maybe I can reuse the EWMA approach to understand if the buffers are
entirely filled or not.
In that case I can increase (e.g. double) or decrease the size.

I'll try to do it!

Thanks,
Stefano