Re: [RFC PATCH RESEND] tcp: avoid F-RTO if SACK and timestamps are disabled

On Wed, 27 Jun 2018, Yuchung Cheng wrote:

On Fri, Jun 15, 2018 at 3:35 AM, Ilpo Järvinen
[off-list ref] wrote:

quoted

On Fri, 15 Jun 2018, Michal Kubecek wrote:

quoted

My understanding is that this means that while the first ack after new
data is correctly ignored, the following ack which preserves window size
should be recognized as a dupack and (3a) should be taken.

Linux FRTO never gets that far (without my fix) if the ACK in step 2
covers beyond the RTO rexmit because 3b is prematurely invoked, that's
why you never see what would occur if 3a is taken. TCP thinks it's not
recovering anymore and therefore can send only new data (if there's some
available).

This is what I tried to tell earlier, with new data there you see there's
something else wrong too with FRTO besides the RTO loop.

agreed. Ilpo do you mind re-submitting your fix
https://patchwork.ozlabs.org/patch/883654/ (IIRC I already acked-by)

Resent as individual patch. And, no you didn't ack it but my impression
in general is that we have converged into agreement about the sender side 
fixes including this one but I'm hesitant to interpret such vague 
impression about agreement alone as acked-by.

(Now with hindsight, maybe I should have interpreted this statement
of yours above as acked-by and added it but I already did send it out 
without adding it).

tcptest suite may have to wait due to some internal workload Neal is 
juggling. 

Ok, no problem. Thanks for keeping me updated.


-- 
 i.