lkml.org 
[lkml]   [1998]   [Jun]   [28]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: Thread implementations...

On Sun, 28 Jun 1998, Michael O'Reilly wrote:

> Umm. Squid does a LOT of different things. One of the things squid
> does a lot of is uncacheable data (i.e. it has some authentication
> method, or cookiefied or whatever). Such data never goes to disk, and
> (after the headers) isn't looked at by squid. Such transfers make up
> nearly 40% of all bytes passed by squid.

(oops, cool. The remaining 60% are still there ... :) But it's cool that
an squid could support the socket->socket copy scheme out of box, but it
does not seem to be trivial to support 'short-switched' sockets. Although
the Linux TCP layer knows about iovecs, so it could theoretically almost
transfer to the outside what an incoming socket produces (linked lists of
skbuffs), but the TCP layer has to know _in advance_ what buffers it has
to send. So we cannot just shuffle iovecs in and out of the socket layer
... this needs support deep in the TCP code which does not seem to be a
2.2 item :(

-- mingo


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu

\
 
 \ /
  Last update: 2005-03-22 13:43    [W:0.092 / U:0.128 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site