Re: TCP and reordering

[Eric Dumazet <eric.dumazet@gmail.com>]

On Wed, 2012-11-28 at 10:24 -0800, Rick Jones wrote:
> On 11/28/2012 12:54 AM, Saku Ytti wrote:
> > On (2012-11-28 00:35 -0800), Vijay Subramanian wrote:
> >
> >> Also note that reordering is tracked on the sender side using the
> >> per flow variable tp->reordering . This measures the amount of
> >> reordering on the connection so that fast retransmit and other loss
> >> recovery mechanisms are not entered prematurely. Doesn't this
> >> behavior at the  sender already provide the behavior you seek?
> >
> > Sorry I don't seem to understand what you mean. Do you mind explaining how
> > the sender can help to restore performance on reordering network?
> 
> tp->reordering is initialized via the sysctl net.ipv4.tcp_reordering 
> which controls how anxious TCP will be to fast retransmit.
> 
> By increasing net.ipv4.tcp_reordering you make the sending TCP less 
> "sensitive" to duplicate ACKs and so less sensitive to reordering 
> detected by the receiver.  The receiver is generating as many 
> (duplicate) ACKs as before, it is just that the sender is ignoring them 
> a bit more.

Note that this sysctl controls the initial value of the per socket
reordering value. It _does_ increase automatically (assuming SACK is
enabled of course)