From owner-freebsd-hackers  Sun Mar  8 21:22:10 1998
Return-Path: <owner-freebsd-hackers@FreeBSD.ORG>
Received: (from majordom@localhost)
          by hub.freebsd.org (8.8.8/8.8.8) id VAA01272
          for freebsd-hackers-outgoing; Sun, 8 Mar 1998 21:22:10 -0800 (PST)
          (envelope-from owner-freebsd-hackers@FreeBSD.ORG)
Received: from rover.village.org (rover.village.org [204.144.255.49])
          by hub.freebsd.org (8.8.8/8.8.8) with SMTP id VAA01230
          for <hackers@freebsd.org>; Sun, 8 Mar 1998 21:22:02 -0800 (PST)
          (envelope-from imp@village.org)
Received: from pencil-box.village.org [10.0.0.22] 
	by rover.village.org with esmtp (Exim 1.71 #1)
	id 0yBv0c-0005O4-00; Sun, 8 Mar 1998 22:21:54 -0700
Received: from pencil-box.village.org (localhost [127.0.0.1]) by pencil-box.village.org (8.8.8/8.8.3) with ESMTP id WAA04154; Sun, 8 Mar 1998 22:21:09 -0700 (MST)
Message-Id: <199803090521.WAA04154@pencil-box.village.org>
To: Marc Slemko <marcs@znep.com>
Subject: Re: kernel wishlist for web server performance 
Cc: Julian Elischer <julian@whistle.com>, hackers@FreeBSD.ORG
In-reply-to: Your message of "Sun, 08 Mar 1998 08:38:38 MST."
		<Pine.BSF.3.95.980308083137.2799W-100000@alive.znep.com> 
References: <Pine.BSF.3.95.980308083137.2799W-100000@alive.znep.com>  
Date: Sun, 08 Mar 1998 22:21:09 -0700
From: "M. Warner Losh" <imp@village.org>
Sender: owner-freebsd-hackers@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.ORG

In message <Pine.BSF.3.95.980308083137.2799W-100000@alive.znep.com> Marc Slemko writes:
: But the point is that you still have to copy it into the mbuf, you still
: have to use the memory for the mbuf, etc.  This uses more CPU and memory
: bandwidth, increases memory use, and means you may have to chop things up
: smaller to avoid using too much memory for mbufs.

And how does sendfile() differ from this?  It still has to insert
protocol headers somehow.

: Or, you can use sendfile() and code it so it doesn't have to copy into
: mbufs.  This is easier because of where the data is coming from.

Then don't you lose performance setting up two DMA transfers for the
packet?  Or do most cards have scatter/gather operations for
transmitting packets?

Warner

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message