Re: [PATCH 1/2] mac80211: add offload channel switch support

[Johannes Berg <johannes@sipsolutions.net>]

Sorry, just thought of something.

On Wed, 2010-05-05 at 20:34 -0700, wey-yi.w.guy@intel.com wrote:

> + * @timestamp: value in microseconds of the 64-bit Time Synchronization
> + *	Function (TSF) timer when the channel switch ie received.

This is not accurate now but see below.

> +		if (beacon)
> +			timestamp = le64_to_cpu(mgmt->u.beacon.timestamp);
> +		else
> +			timestamp = le64_to_cpu(mgmt->u.probe_resp.timestamp);
> +		ieee80211_sta_process_chanswitch(sdata, sw_elem,
> +						 bss, timestamp);
>  	}
>  }
>  
> @@ -1647,7 +1691,8 @@ static void ieee80211_sta_rx_queued_mgmt(struct ieee80211_sub_if_data *sdata,
>  
>  			ieee80211_sta_process_chanswitch(sdata,
>  					&mgmt->u.action.u.chan_switch.sw_elem,
> -					(void *)ifmgd->associated->priv);
> +					(void *)ifmgd->associated->priv,
> +					rx_status->mactime);

For some reason I didn't notice the action frame handling before. Maybe
it would be more consistent to keep using rx_status->mactime even for
beacons? I apologise, this is totally my fault since I suggested using
the beacon time before, but that has a slightly different value relative
to the frame... Hmm. Maybe it would be better after all to require
drivers to provide mactime at least for mgmt frames (beacons + actions)
when they want to implement this callback.

Anyone else have ideas?

johannes