Slashdot Mirror

← Back to Stories (view on slashdot.org)

Explaining WLAN Chips' Poor Linux Support

Posted by timothy on from the holding-the-reins dept.

morcheeba writes "Kernel Traffic is reporting (mirror mirror list) that 'Some WLAN Chip Specs Secret To Protect Military Communications.' While this is stretching it a bit -- these radios are generally limited to a narrow frequency range and few modulation types -- software can cause illegal radio operation, especially when the laws vary by country. Is Linux support for 802.11g and Centrino chipsets going to be delayed by manufacturers afraid of FCC harassment? An interesting discussion on the future of Openness in radio chipsets." Interesting comments from Alan Cox in here about just how flexible some of these chips are.

6 of 218 comments (clear)

  1. drivers take a little while by g4dget · · Score: 5, Informative

    Linux drivers don't come from the Driver Fairy, they usually get written by volunteers. That takes a while: getting the specs, implementing the drivers, testing them, etc. And it usually only happens after the hardware is starting to sell. So, it may well take a year or two for Linux drivers to appear for a piece of hardware. If you want it to happen faster, volunteer yourself.

    Of course, a few manufacturers do ship their own Linux drivers. That's nice, but it isn't all that common yet. And many of the drivers that do ship from manufacturers are based on proprietary, commercial driver toolkits and have to be closed source.

    Centrino is a special case. Centrino is largely a marketing construct, not a technology, and the marketing group that pushed Centrino inside Intel apparently wants to make Microsoft happy and doesn't like Linux. I doubt this is going to last: Linux is too important for Intel to maintain this position.

    1. Re:drivers take a little while by PurpleFloyd · · Score: 5, Informative
      That takes a while: getting the specs, implementing the drivers, testing them, etc. And it usually only happens after the hardware is starting to sell. So, it may well take a year or two for Linux drivers to appear for a piece of hardware. If you want it to happen faster, volunteer yourself.
      The problem with 802.11x support under Linux is that the specs are being kept under wraps. The card manufacturers say the reason is that their cards could be reprogrammed to transmit on reserved frequencies (military, air traffic control, etc.). Apparently, many of these manufacturers would like to make Linux drivers available, but they can't be free-as-in-speech without allowing any decent hacker to change the operating frequency and spy on or disrupt sensitive transmissions.

      As for Linux Centrino support, Linux is not particularly important in the mobile market (excepting Linux-based embedded systems). Servers remain the primary Linux market; Linux laptops are more the domain of hobbyists. While I have no doubt that Centrino drivers will be available for Linux, if what you said about the marketing team in charge of Centrino is true, they will probably come from the hacker community rather than from Intel.

      --

      That's it. I'm no longer part of Team Sanity.
  2. the _REAL_ reason for no drivers... by kwj8fty1 · · Score: 5, Informative

    Many of the chipset makers feel that their 'drivers' are also their IP. In the wireless space, the first to market folks get to make the rules. In the case of 802.11a, Atheros was the first to market. There existed a 'binary only' driver that was built on a mandrake linux box. The bad news is the way it was built made it completely useless. I've not heard of anyone having sucuess using it. Rumor has it that Atheros built this driver & they would release 'formal' drivers for their chipsets. This hasn't happened yet. I doubt it will until they have somebody else providing 802.11a chipsets. To them, it's about getting market share & protecting IP.

    Recent developments:

    Reyk Floeter has started building a GPL driver. It's amusing based on the context of this article, because all this driver can do is SNIFF. That's right, RX Only. Progress has been very slow, and there have been several questions to the list as to how this driver exists, and how it's being built. It would seem that Reyk doesn't have any of the specs & hasn't signed an NDA. I assume he's reverse engineering the windows drivers, but he hasn't stated as much. The development progress has been _VERY_ slow, and this project needs help from OSS devs. Anyone up for a challenge?

    Intellegraphics signed the NDA, and has a driver 'for sale'.

    While the government has it's paws in everything, I doubt this is the case at this point. This whole article is based on FUD.

    That's all. -Eric Johanson, SeattleWireless

  3. Read the linked article, please. by Anonymous Coward · · Score: 5, Informative

    People, come on, RTFA, ok?

    This is not about 802.11b, it's about 802.11g, the newer standard. The one that can do 54 Mbps? Look arround for drivers for those chipsets (mostly Broadcom or Intel's). You'll find none. Why? Because these things can be programmed to receive *and* transmit on any frequency. Any. That inclues military frequencies. Building a receiver for any frequency is not rocket science. In fact it's boring since it's well known. The problem is that the FCC has to approve this things. The vendor builds it and they have to get approval from the FCC to market it. If the FCC catches word that the vendor is giving the specs to a bunch of hippies, the hardware might not get approved (nothing to do with the FCC per se, it's just politics). The problem is some people out there are _very_ willing to go on a disrupt police communications. And what could be easier that just taking a laptop on a car and just war drive. It gives a whole new meaning to the term, doesn't it?

    Some people (RTFA) have proposed solutions, basically signed frequency tables, but since the hardware is out, it's too fscking late. That's going to be in the next gen hardware, but not in the current round. By the look of it, 802.11g on Linux is screwed for the time being.

    1. Re:Read the linked article, please. by iabervon · · Score: 3, Informative

      The FCC doesn't care if you give out the specs; the FCC just cares that there not be any way to send on restricted frequencies (and they'd like it if you can't receive on other restricted frequencies, either, at least not without knowing you're on a restricted frequency, so you can obey the law on disclosure of what you get).

      Basically, the card makers may some stuff configurable which shouldn't be configurable, and they need to deal with getting it configured in the legal way by default. The tricky thing is that there isn't a universal legal default.

      The situation is that it's illegal to disrupt police communications, and it's pretty easy, but it's also pretty obvious. If you're doing it, they can trivially track you down and arrest you. The manufacturer gets in trouble if you're doing it with an unmodified transmitter, because you might not realize you're breaking the law. If anyone who uses the SuSE driver for a Broadcom card (configured, of course, for Germany) in the US is breaking the law and causing problems for the police, someone will get in trouble, and it's likely to be Broadcom.

  4. Re:Time by epsworth · · Score: 3, Informative
    I think you may be confused between the 650 and 650+. Despite the similar names, they use entirely different chipsets. I think the 650 has very good Linux support, but I know for sure that the 650+ has major problems under Linux, since I bought two of them only to find they don't work!

    The 650+ uses the Texas Instruments ACX100 chipset, and they are not willing to release the necessary specs to write drivers for it. Check the ACX100 project on sourceforge to get the details. The only known driver is a binary-only module that will only work for an obscure Mandrake kernel version (PCI version only), and I have only heard of one person who got it to do anything at all.

    You might be right about the (lack of) speed difference between the 11Mbps and 22Mbps products, but I haven't tested them. As always, take marketing-speak with a pinch of salt... It can't be entirely bullshit though, or I would have expected to hear a lot more complaints. Does anyone know more about the difference between the two standards that could explain?