Odd network trace...

Odd network trace...

[Rogier Wolff]

Hi, 

In the following trace, near the end, I see the server ack-ing all
data sent sofar, but the client (Linux version 2.2.17 (root@cave) (gcc
version egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)) #4 Fri Oct
27 14:54:42 MEST 2000) does not continue to send further data,
although that IS available (the transfer should end at about 6Mb).

Am I missing something?


17:04:54.054485 client.1880 > server.http: P 283616:285024(1408) ack 1 win 32660 <nop,nop,timestamp 363440696 2799998> (DF)
17:04:54.054632 client.1880 > server.http: P 285024:286432(1408) ack 1 win 32660 <nop,nop,timestamp 363440696 2799998> (DF)
17:04:54.784416 server.http > client.1880: . 1:1(0) ack 280800 win 7112 <nop,nop,timestamp 2799999 363440587,nop,nop, sack 1 {282208:285024} > (DF)
17:04:54.784501 client.1880 > server.http: P 286432:287840(1408) ack 1 win 32660 <nop,nop,timestamp 363440769 2799999> (DF)
17:04:54.804362 server.http > client.1880: . 1:1(0) ack 280800 win 7112 <nop,nop,timestamp 2799999 363440587,nop,nop, sack 1 {282208:286432} > (DF)
17:04:54.804399 client.1880 > server.http: . 280800:282208(1408) ack 1 win 32660 <nop,nop,timestamp 363440771 2799999> (DF)
17:04:55.374394 server.http > client.1880: . 1:1(0) ack 280800 win 7112 <nop,nop,timestamp 2800000 363440587,nop,nop, sack 1 {282208:287840} > (DF)
17:04:55.394320 server.http > client.1880: . 1:1(0) ack 287840 win 72 <nop,nop,timestamp 2800000 363440771> (DF)
17:04:59.444204 client.1880 > server.http: . 287840:287912(72) ack 1 win 32660 <nop,nop,timestamp 363441235 2800000> (DF)
17:04:59.794183 server.http > client.1880: . 1:1(0) ack 287912 win 0 <nop,nop,timestamp 2800006 363441235> (DF)
17:05:07.893790 client.1880 > server.http: . 287911:287911(0) ack 1 win 32660 <nop,nop,timestamp 363442080 2800006> (DF)
17:05:25.012976 client.1880 > server.http: . 287911:287911(0) ack 1 win 32660 <nop,nop,timestamp 363443792 2800006> (DF)
17:05:59.251356 client.1880 > server.http: . 287911:287911(0) ack 1 win 32660 <nop,nop,timestamp 363447216 2800006> (DF)
17:05:59.961324 server.http > client.1880: . 1:1(0) ack 287912 win 0 <nop,nop,timestamp 2800084 363441235> (DF)
17:07:08.438185 client.1880 > server.http: . 287911:287911(0) ack 1 win 32660 <nop,nop,timestamp 363454135 2800084> (DF)
17:07:08.778122 server.http > client.1880: . 1:1(0) ack 287912 win 0 <nop,nop,timestamp 2800175 363441235> (DF)
17:08:43.793808 server.http > client.1880: R 1:1(0) ack 287912 win 0 (DF)



			Roger. 

-- 
** R.E.Wolff@BitWizard.nl ** http://www.BitWizard.nl/ ** +31-15-2137555 **
*-- BitWizard writes Linux device drivers for any device you may have! --*
* There are old pilots, and there are bold pilots. 
* There are also old, bald pilots. 
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: Odd network trace...

[David S. Miller]

Rogier Wolff writes:
 > Am I missing something?
 ...
 > 17:05:59.961324 server.http > client.1880: . 1:1(0) ack 287912 win 0 <nop,nop,timestamp 2800084 363441235> (DF)

server advertises zero window, no data may be sent.

Until server advertises a non-zero window, the data transfer
may not proceed.

Later,
David S. Miller
davem@redhat.com
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: Odd network trace...

[Rogier Wolff]

David S. Miller wrote:
> 
> Rogier Wolff writes:
>  > Am I missing something?
>  ...
>  > 17:05:59.961324 server.http > client.1880: . 1:1(0) ack 287912 win 0 <nop,nop,timestamp 2800084 363441235> (DF)
> 
> server advertises zero window, no data may be sent.

Yep, I WAS missing someting. Thanks.

				Roger. 

-- 
** R.E.Wolff@BitWizard.nl ** http://www.BitWizard.nl/ ** +31-15-2137555 **
*-- BitWizard writes Linux device drivers for any device you may have! --*
* There are old pilots, and there are bold pilots. 
* There are also old, bald pilots. 
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/