OpenBSD Journal

OpenBSD to support more wireless chipsets

Contributed by dhartmei on from the press-from-down-under dept.

Sam Varghese writes in The Sydney Morning Herald about the support for more wireless chipsets in the upcoming OpenBSD release 3.7 (ath, atu, atw, ral, rtw, ipw, and iwi):

    Damien Miller, a developer with the OpenSSH project, a companion project to OpenBSD, said what was puzzling about the wireless situation was that there were developers who are willing to work free and write drivers for these chipsets, but some manufacturers, like Intel and Atheros, were unwilling to release device firmware (Intel) or card specifications (Atheros). [...]

(Comments are closed)

Comments

  1. By Eduardo Alvarenga (66.110.114.5) on

    Having {iwi,ipw}-firmware ports wouldn't be much easier for a beginner to use?
    Any license restrictions about creating a port for it?

    Comments

    1. By SH (82.182.103.172) on

      As I understand it, you'll have to agree to a license before downloading the firmware. Not much point in making a port where you'll have to manually put files in /usr/ports/distfiles when you can just as easily put them in /etc/firmware yourself.

  2. By bert (216.175.250.42) thrashbluegrass at antisocial dot com on

    It absolutely stumps me why hardware vendors aren't more supportive of driver developers, especially given that it takes little or no effort on their part. One would think that larger potential user base == greater amount of hardware sold. But, then again, I'm not an MBA, so maybe I'm not in the loop on the logic.

    Comments

    1. By vesa (130.233.172.73) on

      I think it surely costs money to make public documentation. Companys internal documents may very well contain things they don't want to make public such as employees' names and secrets about hardware design and future plans. Someone would have to go through all that material, which of course costs money. Therefore, companies don't release the docs.

      Comments

      1. By Levi (24.173.243.99) on

        That, and according to some accounts that I've read, the fact that Intel, for example, doesn't own most/any/all of the IP found in it's wireless products. Since it's likely licensing IP from a third party (Symbol?) that doesn't play friendly with opensource, it can't/won't do anything to help the cause.

        Comments

        1. By jsg (220.253.7.158) on

          Oh come on, in the case of intel we just want to be able to freely distribute the firmware. Intel clearly has the right to distribute the firmware itself so it is not out of their hands. The firmware should not be a port, you should read the man page and contact intel before downloading it.

          Comments

          1. By Anonymous Coward (67.121.189.50) on

            That Intel has the legal right to freely distribute someone else's IP does not mean that they have the legal right to GRANT others the same.

            Comments

            1. By henning (213.128.133.133) on

              That Intel has the legal right to freely distribute someone else's IP does not mean that they have the legal right to GRANT others the same.

              and intel sure has the power to get that right.

      2. By Anonymous Coward (65.198.20.164) on

        Not only that, but it would make public all if any flaws in thier hardware. Flaws which they may or may not know about. I doubt they'd want to do that.

        Comments

        1. By Anonymous Coward (69.197.92.181) on

          What kind of retarded statement is that? If releasing documentation would reveal all their hardware flaws, that means that they documented those flaws. Which means they wouldn't be unknown.

          Comments

          1. By Bert (68.50.4.145) thrashbluegrass at antisocial dot com on

            Yes!

            Flamewar!

            Exactly what I set out to do when I posted my comment!

      3. By Peter Hessler (64.173.147.27) on

        Lets say for the sake of example, it costs $5000 to clean up the documents so it can be publicly avaliable. If the product costs $100, and the company makes $50 profit on it, they only have to sell 101 of them to make money. You can make that up in OpenBSD users alone, not to mention FreeBSD/NetBSD/Linux/Other OS users. And you save on the support costs, since you can just tell the users "we didn't write the driver, contact them".

        Yes, I know my example has bogus numbers, the the point is still the same.

        Comments

        1. By Anonymous Coward (80.219.121.189) on

          evidently, the economic cost is close to nil, probably less than a random management dinner party.

          someone else mentioned docs exposing bugs, well, I seem to remember an article on ross anderson's website about trying to obtain the latest IBM crypto processors, used in higher-level financial transaction checking etc.

          IBM refused to sell them one, so he sent a post-grad student to check out *available* docs on IBM's site, within 2-3 three months they devised a hardware attack without ever having seen one.

          moral? to prevent bug exposure, do not document and do not sell.

          seriously, if someone is motivated enough they will expose your flaky SW/HW and you''ll just look silly, or worse dishonest.

  3. By ciph3r (213.9.211.12) on

    Do you know if any of these new supported chipsets could be used to build access points, like hostap in wi?

    Comments

    1. By SH (82.182.103.172) on

      The Atheros chipset (or the driver) supports it, which is why I've bought a PCI card using AR5212 chipset in the first place.

      Comments

      1. By Anonymous Coward (69.158.152.86) on

        Could you please post some more details about the PCI card that you bought? Exactly which vendor and model number or name? Also which release of OpenBSD you're using with any issues. I would like to use 802.11g cards in OpenBSD routers but from what I had read so far I was under the impression that those chipsets wouldn't be supported until vendors started cooperating. I'm hoping to use OpenBSD firewall with 802.11g wireless card and several Linksys based access points as a mesh network. All with IPSEC traffic with OpenBSD as the gateway to internet. Am I wrong here to assume that the hardware mentioned will work in this setup?

        Comments

        1. By SH (82.182.103.172) on

          Sure, a dmesg with current as of 3rd of march : 

          ath0 at pci2 dev 7 function 0 "Atheros AR5212" rev 0x01: irq 9
ath0: mac 80.6 phy 4.1 radio 1.7 2.3, 802.11a/b/g, APL1, address xx:xx:xx:xx:xx:xx
gpio at ath0 not configured

          It's a D-Link DWL-AG520/EU H/W: A1 using AR5212 chipset. Support not completed yet (a "ifconfig ath0 up" gives a panic, reported to bugs@). A "wicontrol -L ath0" seems to work, though. The card is intended to complement my Prism 2.5 based wireless access point.

          Comments

          1. By Anonymous Coward (69.158.152.86) on

            "Support not completed yet (a "ifconfig ath0 up" gives a panic, reported to bugs@)."

            =(

            For the longest time I was hoping that there was atleast one or two 802.11g cards that I could use with OpenBSD with no worries. What about any other reliability or performance issues? I guess I might have to wait till the 3.7 release or just stick with 802.11b. Thanks for the details.

    2. By Bob Beck (129.128.11.43) beck@openbsd.org on



      Personally I detest G based access points. Maybe if all you're
      doing is making a private one that nobody buy you and a few people
      with G cards ever see, great, but when used in a public setting
      it's worse than useless. It drops to b whenever someone is around
      with a B beacon, and half the G cards/drivers don't take that
      very well.

      As a toy, fine. If you're building anythig real? use B - disable
      the G.

      -Bob

      Comments

      1. By SH (82.182.103.172) on

        With dual band cards (like the D-Link DWL-AG520 I bought) one may use 802.11a as well. In areas with many 802.11b/g devices/access points, the 802.11a will most likely performs better.

      2. By Anonymous Coward (69.197.92.181) on

        I don't know about the openbsd drivers and hostap mode, but with most actual APs you can set it to g only, and b clients will not be able to connect, and so the AP obviously will not go to mixed mode.

  4. By Anonymous Coward (213.118.35.44) on

    Will they be released for 3.6 as well, by any chance?

    Comments

    1. By SH (82.182.103.172) on

      No, they will not. If you want to start using new supported hardware you've to install a snapshot.

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.]