Useless networking code in 2.4.x ?

Useless networking code in 2.4.x ?

[an7]

Hi,

If we have a look at tcp_recv_skb, and
tcp_read_sock(),

we notice that there is a SYN check, and if the flag
is on, we do offset-- (sequence number not
corresponding to real data byte). 

This Syn check is useless, as the function cannot be
called at the beginning of a connection (since we have
not copied_seq filled with the last sequence number of
the last packet passed to the upper layer)

What do you think of that ?


___________________________________________________________
Do You Yahoo!? -- Une adresse @yahoo.fr gratuite et en français !
Yahoo! Mail : http://fr.mail.yahoo.com

Re: Useless networking code in 2.4.x ?

[David S. Miller]

On Thu, 23 Oct 2003 10:58:01 +0200 (CEST)
an7 <an3h0ny@yahoo.fr> wrote:

> If we have a look at tcp_recv_skb, and
> tcp_read_sock(),
>
> we notice that there is a SYN check, and if the flag
> is on, we do offset-- (sequence number not
> corresponding to real data byte). 
>
> This Syn check is useless, as the function cannot be
> called at the beginning of a connection (since we have
> not copied_seq filled with the last sequence number of
> the last packet passed to the upper layer)
> 
> What do you think of that ?

Please next time, take this kind of question to netdev@oss.sgi.com
where the networking developers are, most of them are not subscribed
to linux-kernel.

As to your question, if we ever support accepting data in the
initial final SYN-ACK packet, this code would be needed, so it's
better to keep this code around.