Re: tuntap: Overload handling

["Michael S. Tsirkin" <mst@redhat.com>]

On Thu, Feb 14, 2013 at 08:32:27AM -0800, Eric Dumazet wrote:
> On Thu, 2013-02-14 at 12:50 +0100, Sebastian Pöhn wrote:
> > I am having a look on the tun driver to realize an userspace network
> > driver ( TAP + UIO ). Maybe that's not the use-case tun is intended
> > for.
> > 
> > What I've noticed is that in tun.c Line 741
> > 
> > static netdev_tx_t tun_net_xmit(struct sk_buff *skb, struct net_device *dev)
> > 
> > /* Limit the number of packets queued by dividing txq length with the
> >  * number of queues.
> >  */
> > if (skb_queue_len(&tfile->socket.sk->sk_receive_queue)
> >  >= dev->tx_queue_len / tun->numqueues)
> > goto drop;
> > 
> > If a frame can not be tx it is dropped by the driver.
> > Wouldn't it be more correct to netif_tx_stop_queue() so that packet
> > drops are performed by the overlying traffic control code?
> > 
> > Of course this is not very likely in virtual environments but as soon
> > as any real network hop is involved it could be important.
> > 
> > (I also had a look on some two year old version of tun.c. There
> > queue/tx stopping was done correctly.)

Hmm so ~1000 packets in the tun queue is not enough?
You always have the option to increase it some more ...

> You should ask Michael S. Tsirkin, as he removed the flow control
> in commit 5d097109257c03a71845729f8db6b5770c4bbedc
> (tun: only queue packets on device)
> 

Eric in the past you said the following things
(http://lkml.indiana.edu/hypermail/linux/kernel/1204.1/00784.html)
> > In your case I would just not use qdisc at all, like other virtual
> > devices.
...
> > Anyway, with a 500 packet limit in TUN queue itself, qdisc layer should
> > be always empty. Whats the point storing more than 500 packets for a
> > device ? Thats a latency killer.
you don't think this applies, anymore?

-- 
MST