Re: 答复: 答复: about Mac80211 scan issue

[Stanislaw Gruszka <sgruszka@redhat.com>]

On Thu, Feb 10, 2011 at 03:48:57PM +0800, 李朝明 wrote:
> Yes, that flush can work well, but before drv_flush in
> ieee80211_start_sw_scan have no queue stop actions.
> Below are corresponding code:
> static int ieee80211_start_sw_scan(struct ieee80211_local *local)
> {
> 	/*
> 	 * Hardware/driver doesn't support hw_scan, so use software
> 	 * scanning instead. First send a nullfunc frame with power save
> 	 * bit on so that AP will buffer the frames for us while we are not
> 	 * listening, then send probe requests to each channel and wait for
> 	 * the responses. After all channels are scanned, tune back to the
> 	 * original channel and send a nullfunc frame with power save bit
> 	 * off to trigger the AP to send us all the buffered frames.
> 	 *
> 	 * Note that while local->sw_scanning is true everything else but
> 	 * nullfunc frames and probe requests will be dropped in
> 	 * ieee80211_tx_h_check_assoc().
> 	 */
> 	drv_sw_scan_start(local);
> 
> 	ieee80211_offchannel_stop_beaconing(local);
> 
> 	local->leave_oper_channel_time = 0;
> 	local->next_scan_state = SCAN_DECISION;
> 	local->scan_channel_idx = 0;
> 
> 	drv_flush(local, false);

I looked at wireless-testing tree where code was currently
changed by b23b025fe246f3acc2988eb6d400df34c27cb8ae
"mac80211: Optimize scans on current operating channel."
We stop queues and flush only when leaving operating channel in
ieee80211_scan_state_leave_oper_channel(), what seems to be correct.

Regarding linux-2.6 tree. I think you have right and we have bug
in mac80211, because we do not stop queues before flush in
ieee80211_offchannel_stop_beaconing() when working in STA mode.
Also comment in above function seems to be wrong. 

I'm not sure what to do with that, maybe Johannes will
have some insight?

Anyway, chaoming_li, could you clone current wireless-testing
tree and see how things work there?