[CUWiN-Dev] IPv6 for Community Networks: Back from the Dead.

Quantum Scientific Info at Quantum-Sci.com
Sun Apr 24 08:50:58 CDT 2005


On Sunday 24 April 2005 2:35, David Young wrote:
> > Belkin's intransigence with GPL alone, may well doom our project.  The source 
> > on their GPL page is vastly incomplete.  The major problem is, I am a 
> > perfectionist and want our router to be an *improvement*, so must have the 
> > toolchain specifically tuned to the Broadcom BCM4704 processor and 
> > accessories.  Starting to look like I'm not going to get it in time.
> 
> Carl,
> 
> It's worse than that.  Broadcom docs are closed tight.  Belkin will
> release them over Broadcom's dead body, you could say.  Any luck getting
> doco on the MIMO/pre-N chips from Airgo or Video54?


I'm trying to get docs although granted, it may not happen.  I looked to 
-Linksys'- MIMO source, which actually has the right toolchain, but 
unfortunately it is corrupted 3/4 into the archive.  They claim it's all they 
have, but clearly they are not using corrupted source.  It would take another 
fight, which I don't have in me.  DLink's MIMO is pure crap, given the 
reviews.

Zero from Airgo, and Askey (who actually wrote the driver), after months of 
trying.  Airgo won't even answer antenna placement questions!  This could be 
because either:
a. Belkin assumed support responsibility when they bought Airgo's reference 
design (although Belkin does not have the technical depth to answer my 
questions); or
b. Because Airgo's MIMO proposal has now been declined by IEEE and they are 
pissed off & working on a new one.

The radio driver is binary of course, compiled to Linux kernel 2.4.20, and I 
do not know yet whether it has ad-hoc/WDS.  There are five driver files (an 
ASIC.out, and two for the Polaris radio), and I don't know how much is handled 
by ASIC and how much by the driver.  Given speed considerations, much must 
be in ASIC, although this radio is 'very adjustable' by the developer.  As 
Airgo's design was declined ('n' finalization isn't until next year), there may 
never be a BSD driver, and it's probably not worth reversing.  There's sure 
nothing wrong with their design technically, based on my field testing, so I 
suspect they're being punished for releasing early.  Doesn't matter to me 
though.
 
I've been quite busy with other things lately, but will resume work shortly.

Best,

Carl Cook


More information about the CU-Wireless-Dev mailing list