>
> I don't think any TCP stack is going to behave very well at all when
> every 4th packet is dropped. This prevents TCP from gaining any
> momentum, and prevents completely any possibility to perform fast
> retransmit because there are not enough packets in flight. So you
> will eat the full retransmit timeout for recovery every time, and this
> timeout will increase more and more as the connection drags on.
>
> I'm not debugging based upon this, it cannot legitimately demonstrate
> any bug.
>
Yes, Yes. I'm not looking at the TCP/IP layer. I'm thinking that I
can demonstrate the behavior of the PPP Link. I know that I can send
data between two Linux computers via RS-232C/Modems/Phone-lines, with
an occasional small data loss, less that 1/2 percent. This, from
user-mode 'C' programs.
The intermittant data-flow pattern (beat) is most exactly what I get
from PPP.
The TCP stalls that I observe occur only from PPP links. Maybe
flow-control doesn't work in PPP?
Cheers,
Dick Johnson
***** FILE SYSTEM WAS MODIFIED *****
Penguin : Linux version 2.1.118 on an i586 machine (66.15 BogoMips).
Warning : It's hard to remain at the trailing edge of technology.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.altern.org/andrebalsa/doc/lkml-faq.html