WLAN: known issues and how to help - uptime test

Werner Almesberger werner at openmoko.org
Thu Feb 5 23:11:32 CET 2009


Philip Rhoades wrote:
> So I should upgrade to the latest (7.00.6) firmware I guess to see what  
> effect that has on OM WLAN uptime?

I think it would be good if you could upgrade, yes. I'm not sure
if it will make a difference for this problem, but they fixed a
LOT of issues, so it seems like a good move in any case.

> What about the alternative OS firmwares?

The VxWorks-based WRT54G are quite limited when it comes to
alternatives. Starting with hardware version 5, the RAM was
reduced, making it hard to run Linux, but DD-WRT managed
regardless. With version 7, Linksys/Cisco also changed the WLAN
chipset, which seems to have created a so far unsurmountable
obstacle.

See also
http://www.linuxquestions.org/questions/linux-wireless-networking-41/wrt54g-v7-firmware-504449/

Linksys/Cisco also make the specifically Linux-friendly WRT54GL.

For experimenting, you could also use a WLAN card with a PC and
run hostapd. This has the added advantage that it's easier to
patch hostapd and such than for a small embedded system.

Finding a suitable WLAN card can be a challenge, though. I did
a bit of field research, and of the readily available 802.11b/g
PCI cards, the ones based on the RaLink RT2561 seem to work best:

http://svn.openmoko.org/developers/werner/wlan/freeze/4if/HW

Reluctant to add yet another layer of trouble, I didn't venture
into USB land.

I would really love to see master mode support for the RTL-8185L,
though. These are the cheapest cards on the market, you can find
them everywhere, and they have great reception. Ironically, master
mode support used to exist in an ancient and long-forgotten driver
http://rtl8180-sa2400.sourceforge.net/ and there's another fork
that seems to have gone dormant after the new driver's been
included in mainline: http://rtl-wifi.sourceforge.net/wiki/Main_Page

- Werner



More information about the devel mailing list