Re: [PATCH] tcp: detect loss above high_seq in recovery

["Ilpo Järvinen" <ilpo.jarvinen@xxxxxxxxxxx>]

On Tue, 3 Jan 2012, Yuchung Cheng wrote:

> On Fri, Dec 30, 2011 at 3:19 AM, Ilpo Järvinen
> <ilpo.jarvinen@xxxxxxxxxxx> wrote:
> > On Thu, 29 Dec 2011, Yuchung Cheng wrote:
> >
> >> To summarize:
> >> 1) this change does NOT forbid TCP performing further CWR on new losses
> >> 2) this change (merely) marks some forward-retransmitted packets LOST.
> >> 3) the commit message is to be explicit that CWR may not always be performed,
> >>     in both my change and current kernel.
> >
> > I agree with these points. I wonder if you intend to look to fixing the
> > lacking CWR problem or should I look to that once I get some spare time?
> >
> >> HTH. I will update commit message if people can make sense of it :)
> >
> > Please do.
> Great, I'll update the commit message first, and work on another
> feature patch that does CWR if
> a) new sequence is lost
> b) a retransmission is lost

Ok, thanks. I think that it's better to do the way you suggest... ie., to 
do the cwr on detection rather than (possibly much) later when 
rexmitting.

-- 
 i.