[conspire] "madwifi" is proprietary sludge (was: driver)

Daniel Gimpelevich daniel at gimpelevich.san-francisco.ca.us
Tue Jun 27 18:31:20 PDT 2006

Other Linuxers would gladly exchange Intel PRO/Wireless 2200 for Atheros
because the Intel chipset cannot act as an access point. Madwifi is not
just a wrapper; it is the bulk of the driver. AIUI, the proprietary HAL
primarily serves to load a proprietary firmware onto the card. There is a
project well under way to fabricate an open-source HAL compatible with the
rest of madwifi, and another project well under way to fabricate an
open-source firmware for the HAL to load. Don't blame the hardware for the
license under which software to drive it is released. Until very recently,
the Intel chipset you're so hot on worked only with the aid of ndiswrapper
and the like.

BTW, I have never heard of a miniPCI Atheros card. I was assuming the
"SuperG" in question was either CardBus or PCI. Not that it matters;
they're all equally replaceable...

On Tue, 27 Jun 2006 18:41:33 -0700, Rick Moen wrote:

> Passing this along, since Mike Williams is not on "conspire".
> "madwifi" is a open source wrapper around a proprietary, binary-only 
> proprietary "HAL" core library.  Thus, it's a proprietary driver often
> mistakenly claimed to be open source (not by you, Daniel).  Atheros
> itself has made that claim, and I think it fair to say they don't have
> the excuse of ignorance, and are, to be blunt, _lying_.
>    http://www.atheros.com/news/linux.html
>    SUNNYVALE, Calif., July 23, 2003Atheros Communications, the leading
>    developer and market share leader in advanced wireless LAN (WLAN)
>    chipsets, today announced that the first open source Linux and FreeBSD
>    software drivers for 802.11b/g and universal 802.11a/b/g products are
>    now available as a free download from the Internet.
>    Created by Open Source developer Sam Leffler, the 802.11a/b/g Linux
>    driver is intended to be used as a building block for creating fully
>    featured Linux-based 802.11b/g and 802.11a/b/g products such as network
>    adapters, access points, and home gateways. 
>    [...]
>    To download a copy of the new Linux device driver please go to:
>    https://sourceforge.net/projects/madwifi/.
> The named SourceForge project page is equally dishonest, and VA
> $WHATEVER should be ashamed of themselves.  (I actually brought to the
> SourceForge.net managers' attention in 2001 the fact that many existing 
> projects falsely claim to be open source.  They replied they didn't have
> time and staff to police the matter.)  Quoting the project page:
>    License: BSD License, GNU General Public License (GPL)
> It is in fact proprietary.  I object even more to the dishonesty than to
> the licensing, actually.
> Jean Tourrilhes's page
> (http://www.hpl.hp.com/personal/Jean_Tourrilhes/Linux/Linux.Wireless.drivers.802.11ag.html)
> repeats an often-repeated assertion from madwifi programmer Sam Leffler:
>    Sam, with the help of Atheros, had created a BSD driver for those
>    cards some time ago. Unfortunately, he was unable to release it because
>    of the FCC regulations. The Atheros hardware is basic and doesn't
>    enforce that valid operating parameters are set (such as frequency and
>    power level), however the FCC mandate that end user should not be able
>    to set invalid operating parameters. Eventually, Atheros managed to find
>    a solution that was acceptable : they create a HAL, a binary layer that
>    would sit between the hardware and the driver and enforce that FCC
>    regulations are respected. The downside is that the HAL is available
>    only for selected architectures (i386, PPC, Arm, Mips, SH4, Alpha and
>    Sparc64).
> I have no doubt that Leffler believes that -- and probably that Atheros
> Communications does, too.  However, I spent a fair amount of time
> attempting to find any "FCC regulations" to that effect.  I have found
> absolutely none.  Accordingly, I have tentatively concluded that this
> assertion is a corporate-nonsense justification for perpetuating an
> existing secrecy policy.  ("FCC _might_ have a regulation against it, so
> we'll keep our code proprietary just to be safe, and merely claim to 
> the open source people that we have to.  They'll never bother to
> check.")
> I personally would go far out of my way to avoid the need for
> proprietary drivers, including buying replacement hardware.  And so, I
> would not buy _any_ Atheros WiFi hardware; if it came with my laptop,
> I'd regard it as effectively non-existent and use something else (same
> as with winmodems).  
> If I were in specifically Mike's shoes, I'd shitcan the Atheros miniPCI
> card, and replace it with one based on an Intel PRO/Wireless 2200 chip.
> ----- Forwarded message from Daniel Gimpelevich <daniel at gimpelevich.san-francisco.ca.us> -----
> From: Daniel Gimpelevich <daniel at gimpelevich.san-francisco.ca.us>
> Date: Tue, 27 Jun 2006 16:53:43 -0700
> To: conspire at linuxmafia.com
> Subject: Re: [conspire] (forw) Re: driver
> On Mon, 26 Jun 2006 13:05:06 -0700, Rick Moen wrote:
>> Date: Mon, 26 Jun 2006 11:15:49 -0700
>> From: "Williams, Mike" <MWILLIAMS at probusiness.com>
>> To: Rick Moen <rick at linuxmafia.com>
>> Subject: driver
>> Hey Rick, 
>> The driver I was referring to was a cheaply built 802.11 superG. The disk didnt have a Lindriver I'll look for the name as I dont have my lap w / me
> "SuperG" is a trademark of Atheros and is universally supported using the
> madwifi driver. Newer cards require newer versions of the driver. The
> older the distro, the less likely it is to work. Forget the ancient ones
> you had your heart set on, and install something recently released. Unlike
> Windows, it'll "just work."
> _______________________________________________
> conspire mailing list
> conspire at linuxmafia.com
> http://linuxmafia.com/mailman/listinfo/conspire
> ----- End forwarded message -----

More information about the conspire mailing list