lkml.org 
[lkml]   [1996]   [Jun]   [9]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: Speed of memcpy, csum_partial and csum_partial_copy
Date
> On the TCP i've been seting up for 2.1 there is no need for copy
> number 2 since the skb is cloned (the copy is restricted to the skb
> header).

Providing tcp.c doesnt scribble into the packet itself then yes this
is true. Consider however a packet that is source routed via loopback
then out. You will then need to rejig the header.

> Well, gonna try removing that copy you mention to see if it increases
> performance. Is it done on the loopback device itself ?

Yes. I think we need to clone things more and drop code in the forwarding
routines along the line of


newskb=skb_unshare(skb); /* in case its a clone */
if(!skb)
dropit(skb);
else
{
scribble_on_packet();
sendit();
}

which solves the issues with network taps as well.

Alan



\
 
 \ /
  Last update: 2005-03-22 13:37    [W:0.070 / U:0.956 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site