Johann Beretta <invalid@invalid.org> hath wroth:

>It's not even that, it's that they won't release ANY of the specs. Linux hackers
>would write the drivers themselves if they could, but the manufacturers won't
>release any specs, and so the coders are relegated to trying to reverse engineer
>the hardware which is MUCH harder than just writing a driver.

Actually, reverse engineering is often *MORE* accurate than using the
published specs. I've been ensnarled in at least 2 debates over
whether to build wireless devices in accordance to the specs, or to
built them to match the goofy hardware de jure which doesn't quite
follow the specs. Lots of fun. It's interesting to note that the
best drivers seem to be the ones that either include company support
(Prism, Atheros), or the limited number of devices that have been
revese engineered (Broadcom).

For what little I know about (which is not every chipset on the
planet). I have clues on the others (Atmel, Intel, Prism, Marvell,
RealTek, Zydas) but prefer to let the Linux experts doing the
complaining. Nice article on the current open source wireless mess:
<http://www.thejemreport.com/mambo/content/view/293/>
Well, maybe not so current. The ar5k Atheros driver is obselete under
Linux.

Broadcom: Doesn't supply any info except under NDA. Does not support
open source. Ships drivers in the form of binary APL's. However, it
has been reverse engineered quite successfully:
http://bcm-specs.sipsolutions.net
http://bcm-v4.sipsolutions.net
http://bcm43xx.berlios.de

Atheros: Generally supports open source drivers for Linux and
FreeBSD.
http://www.atheros.com/news/linux.html



--
Jeff Liebermann jeffl@comix.santa-cruz.ca.us
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558