p2p start listen

Jouni Malinen j at w1.fi
Sun Dec 18 09:24:17 EST 2011


On Sun, Dec 18, 2011 at 08:55:41PM +0800, Huijun Gan wrote:
> Thanks. My fault that expressed the opposite meaning. I wonder why
> wpa_drv_probe_req_report() is called first.
> 
> My understanding on  remain-on-channel is to ensure that the device is
> on the indicated channel to receive the probe req frames.

Well, these should really be done in an atomic operation, so I don't
really consider either sequence to be better than the reverse order.

> And besides, for OFFCHANNEL_TX driver, after sending action frame
> (like GO neg req) out on peer's listen or operating channel, how does
> our own driver/device know how long it should stay on this channel, to
> guerantee peer's resp? There is a wait_time parameter when sending
> action frame, I think this is to ensure TX success (the max time
> waiting for TX_STATUS, right?).

This depends on the driver capabilities. You can either depend on the
remain-on-channel operations or use the wait_time to offload this to the
driver/firmware. In the offload case, the wait_time parameter is the
time that the device needs to wait for the response.

-- 
Jouni Malinen                                            PGP id EFC895FA


More information about the HostAP mailing list