All of lore.kernel.org
 help / color / mirror / Atom feed
From: Subba Rao <subba9@home.com>
To: Andrew Morton <akpm@zip.com.au>
Cc: Linux Kernel <linux-kernel@vger.kernel.org>
Subject: Re: Half Duplex and Zero Copy IP
Date: Sat, 11 Aug 2001 07:59:06 +0000	[thread overview]
Message-ID: <20010811075905.A32210@home.com> (raw)
In-Reply-To: <20010810095313.A6219@home.com> <3B742996.3F2C8DEC@zip.com.au>
In-Reply-To: <3B742996.3F2C8DEC@zip.com.au>; from akpm@zip.com.au on Fri, Aug 10, 2001 at 11:36:06AM -0700

On  0, Andrew Morton <akpm@zip.com.au> wrote:
> Subba Rao wrote:
> > 
> > Hello,
> > 
> > I have 2 3Com NICs on my system. They are 3c905C Tornado PCI cards.
> > The drivers are compiled into the kernel (Slackware 8.0 with kernel 2.4.7).
> > 
> > One of the interfaces will be used as a sniffer interface (without IP address)
> > and a very high traffic pipes. I do not wish to loose any packets coming to this
> > interface. Is it better if I initialize the interface in HALF DUPLEX mode? If yes,
> > how do I set the card to HALF DUPLEX mode? How can I find out the HW (NIC) settings
> > on the system?
> 
> No, this will provide no benefit.

So, is the card set in half-duplex mode by default and full-duplex is a forced
option?

> 
> > Another question about 3Com NICs, do they perform zero-copy IP?
> 
> Linux's zerocopy infrastructure allows the sendfile() system call
> to save a copy with NICs which have hardware checksumming and
> scatter/gather.  3c905C is one such NIC.  Kernel is not generally
> "zero copy", but large savings are available in certain situations.
> NFS packet reassembly benefits from 905C's as well.
> 
> > I read that the performance improves a lot WITHOUT zero-copy IP.
> 
> Not right.  Where did you read that?
> 
	http://www.fefe.de/linuxeth/

"To achieve gigabit throughput, it is important that the operating system does 
not copy the data in the packets before sending them (this is called zero-copy 
IP). Unfortunately, the kernel needs to put a header before the data in the
packet, so not copying the data to a buffer in kernel space means that the NIC
needs to be able to fetch the header from a different place in memory than the
user data in the packet. This is called scatter/gather and is necessary for
zero-copy IP." 

-- 

Subba Rao
subba9@home.com
http://members.home.net/subba9/

GPG public key ID 27FC9217
Key fingerprint = 2B4C 498E 1860 5A2B 6570  5852 7527 882A 27FC 9217

  reply	other threads:[~2001-08-11 11:58 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2001-08-10  9:53 Half Duplex and Zero Copy IP Subba Rao
2001-08-10 18:36 ` Andrew Morton
2001-08-11  7:59   ` Subba Rao [this message]
2001-08-10 18:46 ` Andi Kleen
2001-08-10 23:50   ` Jamie Lokier
2001-08-11  0:00     ` Andi Kleen

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20010811075905.A32210@home.com \
    --to=subba9@home.com \
    --cc=akpm@zip.com.au \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.