Re: [PATCH] tcp: don't use timestamp from repaired skb-s to calculate RTT

[Andrew Vagin <avagin@parallels.com>]

On Tue, Aug 12, 2014 at 05:15:01AM -0700, Eric Dumazet wrote:
> On Tue, 2014-08-12 at 13:45 +0400, Andrey Vagin wrote:
> > We don't know right timestamp for repaired skb-s. Wrong RTT estimations
> > isn't good, because some congestion modules heavily depends on it.
> > 
> > This patch adds the TCPCB_REPAIRED flag, which is included in
> > TCPCB_RETRANS.
> 
> ...
> 
> > +
> > +				/* All packets are restored as if they have
> > +				 * already been sent. skb_mstamp isn't set to
> > +				 * avoid wrong rtt estimation.
> > +				 */
> > +				if (tp->repair) {
> > +					TCP_SKB_CB(skb)->sacked |= TCPCB_REPAIRED;
> > +					TCP_SKB_CB(skb)->when = tcp_time_stamp;
> > +				}
> >  			}
> >  
> >  			/* Try to append data to the end of skb. */
> 
> 
> Are you sure TCP_SKB_CB(skb)->when needs to be set ?

It's used in tcp_rearm_rto() for calculating a retransmit timeout.
...
	const u32 rto_time_stamp = TCP_SKB_CB(skb)->when + rto;
	s32 delta = (s32)(rto_time_stamp - tcp_time_stamp);
...

"when" is used as a start point, so I think it's acceptable here.

I will add a comment. Thanks.

> 
> It should not anymore.
> 
> If yes, I believe a comment would help a lot here.
> 
> Also, please include this tag to ease stable backports (3.15+) :
> 
> Fixes: 431a91242d8d ("tcp: timestamp SYN+DATA messages")
> Fixes: 740b0f1841f6 ("tcp: switch rtt estimations to usec resolution")

Thanks,
Andrew