OpenBSD Journal

OpenBSD goes 10Gbps

Contributed by mk/reverse on from the the-wow-a-28k8-bps-modem-days-are-over dept.

brad@ just committed support for Intel's 10Gbps network interfaces:

Modified files:
        sys/dev/pci    : files.pci
Added files:
        sys/dev/pci    : if_ixgb.c if_ixgb.h if_ixgb_osdep.h ixgb_ee.c
                         ixgb_ee.h ixgb_hw.c ixgb_hw.h ixgb_ids.h
 
Log message:
add Intel 10Gb Ethernet driver with support for LR/SR and CX4 cards.

brad@ also has a driver ready for Neterion's Xframe-I 10Gbps interface, but he has no hardware to test it with. He's also in need of a 10Gbps switch, so if you can donate either please contact him.

(Comments are closed)


Comments
  1. By Anonymous Coward (134.58.253.131) on

    Did Intel provide the necessary documentation, or was there reverse-engineering and such involved?

    Comments
    1. By Michael Knudsen (130.225.192.105) on

      I left out a bit if the commit message, namely the ok's and "From: Intel's web-site" which is why I linked to the complete commit message. As far as I know, this is a port to OpenBSD of the Intel FreeBSD driver.

    2. By Anonymous Coward (193.63.217.208) on

      Intel are offering driver source code for FreeBSD and Linux. The FreeBSD driver requires you to click-through a copy of the GPL but the source code itself has a BSD license in it.

      I hope this is a sign of a more co-operative Intel, maybe we'll even see an end to closed-source firmware?

      Comments
      1. By baldusi (209.13.233.223) on

        It's just that Linux+FreeBSD are quite a bit of the target market of 10Gbps Ethernet cards. They simply can't be in that market without a BSD driver. That's why you see drivers for Broadcom and Intel in the Ethernet area. But since the bulk of the wifi card are notebooks, you won't see much cooperation there. And the accelleration and encryption stuff I don't think we'll ever see any of that.

      2. By Brad (216.138.195.228) brad at comstyle dot com on

        The click-through also has a BSD license. You are mistaken.

  2. By Anonymous Coward (82.197.192.49) on

    The complete commit message includes the origin. The driver is coming from Intel itself.

  3. By Anonymous Coward (62.179.8.122) on

    30 Sep 2004 - NetBSD again sets Internet2 Land Speed World Record (top)

    NetBSD does it again: after the original Internet2 Land Speed Record set by NetBSD in 2004 May 3 was broken, NetBSD shines again: Once more researchers at the Swedish University Network (SUNET) have broken the Internet2 Land Speed Record, using the upcoming version of NetBSD 2.0.

    The new records are 124.935 Pbmps in a single stream (was 69.073 Pbmps), and 122.367 Pbmps in multiple streams. NetBSD was used once more due to the “scalability of it's TCP code”.

    More information about this record including the NetBSD configuration can be found at: http://proj.sunet.se/LSR3-s/ for single stream and http://proj.sunet.se/LSR3-m/ for multiple streams. And the website of the Internet2 Land Speed Record (I2-LSR) competition is located at: http://lsr.internet2.edu/.

    Comments
    1. By Anonymous Coward (66.11.66.41) on

      What's a Pbmps?

      Comments
      1. By corentin (81.56.152.193) on

        Petabit-meters per second.

    2. By m0rf (68.104.17.51) on

      don't you have a website/mailing-list you can post your exciting (+1 year old) netbsd news for netbsd users to see (netbsd-advocacy@ maybe) instead of trolling for netbsd on the openbsd journal?

      "ohnoes openbsd is teh slow, netbsd is teh fast, use netbsd it am teh best."

    3. By Anonymous Coward (204.209.209.129) on

      and if you could explain to me how the tcp code is so different that would be great.

    4. By Chris Wareham (144.178.103.79) chriswareham@chriswareham.demon.co.uk on http://www.chriswareham.demon.co.uk/

      Hubert is that you??? If so, then please don't bait the OpenBSD people.

      Comments
      1. By Anonymous Coward (64.74.207.33) on

        I doubt it; the poster's IP address is Polish. Hubert is German.

        Comments
        1. By Anonymous Coward (84.177.74.15) on

          Actually Hubert is hiding in the US, that's why it's been a bit quiet ;-)

          Hubert, we miss you. I need a laugh once a while ;)

    5. By m0rf (68.104.17.51) on

      exactly why is /anyone/ modding up netbsd news from 2004?

    6. By m0rf (68.104.17.51) on

      http://lsr.internet2.edu/history.html:

      19 January 2005, IPv6 Single and Multiple Stream, operating system: Linux kernel 2.6.8

      24 December 2004, IPv4 Single and Multiple Stream, operating system: Linux kernel 2.6.6

      not only old, outdated.

  4. By Anonymous Coward (67.167.250.95) on

    Does anyone make a 10Gbps switch? All I've seen are switches that have 1 or 2 10Gbps ports, or modules for cisco switches.

    Comments
    1. By Brad (216.138.195.228) brad at comstyle dot com on

      Yes, Foundry Networks makes a L2 8 port 10G EdgeIron switch.

      Comments
      1. By Brad (216.138.195.228) brad at comstyle dot com on

        or there is the HP ProCurve 6410cl .

        Comments
        1. By Anonymous Coward (67.167.250.95) on

          The Foundry Networks switch is about $20,000 while the HP one is around $4,000. Why the big difference?

          Comments
          1. By brad (216.138.195.228) brad at comstyle dot com on

            The 6410cl is actually around $7,000, the 6400cl is $4,000, but is CX4 capable only. You're paying for the Foundry name, no different than you would pay for Cisco.

            Comments
            1. By Philip Olsson (85.30.133.54) olsson@puffy.nu on http://www.puffy.nu

              The HP and Foundry switch is the same switch. HP buys alot of switches from foundry, strips some features, add others and relabe. There are a couple of switches in the hp/foundryline where you can exchange chassi,linecards,powersupplys in between. You can even reflash your hp to a foundry.

    2. By Blake (62.4.77.94) on two one one two dot net

      The MRV 2408 is pretty good and very inexpensive (as 10g switches go). We've tested a 24 port 10/100 + 4 port GigE switch in this same series of produts and found it to be quite good - it uses an IOS-like cli interface and apart from a few niggles (only supports one TACACS+ server, 8 character usernames, and vlan trunking config is a little wierd) we found it perfectly adequate for 99% of the customer installations we do. Digital diagnostic support (a.k.a. optical power meter) on all pluggale ports is a nice bonus.

      -Blake

Latest Articles

Credits

Copyright © - Daniel Hartmeier. All rights reserved. Articles and comments are copyright their respective authors, submission implies license to publish on this web site. Contents of the archive prior to as well as images and HTML templates were copied from the fabulous original deadly.org with Jose's and Jim's kind permission. This journal runs as CGI with httpd(8) on OpenBSD, the source code is BSD licensed. undeadly \Un*dead"ly\, a. Not subject to death; immortal. [Obs.]