> Date: Wed, 24 Jul 2013 02:50:13 -0600
> From: Nathan Goings <[email protected]>
> 
> Well, I'm a moron.  I spent several weeks working on the Ralink driver, 
> profiling linux vendor provided code; looking at changes between 2860, 
> 35xx, and 3090; and nagging various people.  I made a single change and 
> the ral(4) driver is working on the Ralink 35xx series in both client 
> and hostap mode (currently, 11b mode).

Welcome to the world of hardware hacking.

> What is the protocol for testing?  Here's my agenda for testing:
>   1. Try different (faster) wireless modes.
>   2. Stress-test bandwidth with file transfers
>   3. Long term monitor for packet-loss, reconnects, and ping spikes.
>   4. Throw away that Fedora ISO. (supports 35xx out-of-the-box)
>   5. Bathe in the glory of OpenBSD.

The protocol for testing is simple: make sure it doesn't break any of
the devices that are already working.

> The change I made was simple:
> /usr/src/sys/dev/ic/rt2860.c : Line 202
> 
> Change "RT3070_DEF_RF" to "RT3572_DEF_RF"
> 
> That causes the rt3090_def_rf[] array to fill with defaults for the 
> 35xx.  I'm surprised this change works.

Well, that will certainly break existing support for the RT3090 chips.
So a bit more work will be needed.  First step would be to provide a
proper diff, and show us your dmesg!

Reply via email to