Troubleshooting ral0 device timeouts

[Michaël Grünewald]

I have set up a FreeBSD access point, it is equipped with a ralink based 
card and works most of the time. I would appreciate some help for 
troubleshooting ``the rest of the time''. Thanks!

First there is nothing fancy about my wirelesse setup, I merely use the 
ralink card as an ethernet switch to create a LAN. I therefore 
configured my ralink R2600 ral0 interface, an ethernet card and bridged 
them together as described in if_bridge(4) or ifconfig(8). Note that 
only the bridge has got an IP address. Also, hostapd is up and running, 
managing the ral0 interface. This setup somehow works: I can succefully 
connect stations to the AP.

However, from times to times, the ral(4) driver emits a `device timeout' 
message and the interface is hen stuck. All connections to the AP are 
lost and no station would notice the AP anymore.

I could not successfully reset the interface with the sequence:

# /etc/rc.d/netif stop ral0
# /etc/rc.d/hostapd stop
# /etc/rc.d/netif start ral0
# /etc/rc.d/hostapd start

Although it does not emit any error message, the AP remains 
undetectable. Rebooting the machine brings back the AP to a working 
state, but that is annoying!

I am running amd64/7.2, the ralink chip is reported to be 2600 or 2610 
by the kernel.
-- 
Best regards,
Michaël