Hi all,
I'm building up a Linux ad-hoc network by means of two miniPCI adapters from ST. Their hardware and driver is "supposed" to support WMM, but in practice, while trying to do some tests (two contemporaneous iperf traffics, one of low priority and the other of the highest one) I don't see any change in the bandwidth, as expected. I mean, I use 802.11a, channel 36, rate 6M, and try to put on the wireless medium those two traffics of equal bandwidth which sum exceeds the theorical maximum throughput of the
network... I'm expecting to see that the highest priority traffic gains more bandwidth than the other, but this is no what happens.
Am I wrong to expect something like this?
I've also tried to use a background iper traffic (lowest priority) and over this, to send some long pings, one time with the lowest priority and the second time with the highest; once again I was expecting to see some difference in the average time... but no differences at all were visible.

Of course the driver is configured with the qos on: in the code I can see 8 "upper layer" queues where packets marked with different priorities go (0-7), and 4 access categories (those of WMM) whose parameters I can set.
I was wondering if there could be something missing in my kernel configuration... do you know about the need to add something on .config?

Thank you so much!

Elise