2013-02-14 11:58:51

by Magnus Cederlöf

[permalink] [raw]
Subject: Possible bug with probe_send_count set to 0 and then checked if greater than 0

In net/mac80211/mlme.c function ieee80211_sta_tx_notify:
If a null frame was acked, ieee80211_sta_reset_conn_monitor is called.
This function sets probe_send_count to 0.
After the call to ieee80211_sta_reset_conn_monitor, it is checked if
probe_set_count is > 0.
This means that ieee80211_queue_work will not be called even though
the nullfunc was successful.

The symptom of this error is that when using power save, it will take
500ms to enter power save again after a successfully acked null frame.

What case will the check for 'probe_send_count > 0' cover?

Regards,
Magnus Cederl?f


2013-02-14 12:14:44

by Wojciech Dubowik

[permalink] [raw]
Subject: Re: Possible bug with probe_send_count set to 0 and then checked if greater than 0

On 02/14/2013 12:58 PM, Magnus Cederl?f wrote:
> In net/mac80211/mlme.c function ieee80211_sta_tx_notify:
> If a null frame was acked, ieee80211_sta_reset_conn_monitor is called.
> This function sets probe_send_count to 0.
> After the call to ieee80211_sta_reset_conn_monitor, it is checked if
> probe_set_count is > 0.
> This means that ieee80211_queue_work will not be called even though
> the nullfunc was successful.
>
> The symptom of this error is that when using power save, it will take
> 500ms to enter power save again after a successfully acked null frame.
I am looking at it now and I see that it happens without power save
as well. I have just started putting printk around so I am not so far.

Wojtek
> What case will the check for 'probe_send_count > 0' cover?
>
> Regards,
> Magnus Cederl?f
> --
> To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html