public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* TCP ACKs carrying data
@ 2005-03-04 12:49 se go
  2005-03-04 13:17 ` Michael Tokarev
  0 siblings, 1 reply; 2+ messages in thread
From: se go @ 2005-03-04 12:49 UTC (permalink / raw)
  To: linux-kernel

Hello,

i got some problems concerning tcp realization in kernel (v 2.4).

i need to make some modifications to tcp to be able to transfer data within ACK segments...(initially their data field is empty)

i suppose i should modify the following functions:

/net/ipv4/tcp_input.c > tcp_ack()  - for processing incoming acks, to read data from their data field

and

/net/ipv4/tcp_output.c > tcp_send_ack() - for processing outgoing acks, to put data in them

my questions are: 
 - am i right in my decision?
 - what problems these modifications may cause (i tried to modify and they do cause!... i just cannot understand what they are)
 - what functions else i may need to modify?
 - how must i modify ACKs to make them do what i want and to make linux kernel correctly treating them.

if you need i can reply the list of modifications i've done.

i'd like to be personally CC'd.


thanks in advance,

Serge Goodenko
Moscow Institute of Physics and Technology
Moscow, Russia


^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: TCP ACKs carrying data
  2005-03-04 12:49 TCP ACKs carrying data se go
@ 2005-03-04 13:17 ` Michael Tokarev
  0 siblings, 0 replies; 2+ messages in thread
From: Michael Tokarev @ 2005-03-04 13:17 UTC (permalink / raw)
  To: se go; +Cc: linux-kernel

se go wrote:
> Hello,
> 
> i got some problems concerning tcp realization in kernel (v 2.4).
> 
> i need to make some modifications to tcp to be able to transfer data within ACK segments...(initially their data field is empty)

I thought it is already here.  That is, if the kernel have something
to send at the moment when incoming packet arrives, it sends one
packet out, containing both the ACK for previously received packet
and the data.

Ie, if the app is sending data in both directions, there will be no
(well, almost: initial handshake and final sequence are still unique)
"bare" ACKs, or packets without data.

Or Am I just dreaming? ;)

/mjt

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2005-03-04 13:24 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2005-03-04 12:49 TCP ACKs carrying data se go
2005-03-04 13:17 ` Michael Tokarev

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox