DragonFlyBSD

is this fixed?

The problem we had encounted in bridge code should have been resolved
by sys/net/if_bridge.c rev 1.25
I think it is a better solution than recursive serializer.

Since the original problem is reported by you, you can test to verify it :)

Best Regards,
sephe

Ah no, I believe you. Closing.

Does this fix the momentary (1-2 second) 'freezes' one gets when wpa_supplicant is cycling through
configs trying to associate with an AP ?

Maybe it's just me, my ath driver isn't completely standard ... ;)

--Andrew

Cycling through? Do you mean rekeying? I have tested (iperf on both
direction) ath(4) (STA mode) for two days (24hours for CCMP/24 hours
for TKIP, with WPA) rekeying every 10 minutes, before importing the
new hal; I didn't notice any slow down (iperf reported stats every 10
seconds) in my test. But all the ath(4) that I can access are 5212
typed (different mac/phy revision tho).

Best Regards,
sephe