On Wed, Feb 19, 2014 at 03:50:14PM +0100, Johannes Berg wrote:
> On Wed, 2014-02-19 at 15:48 +0100, Stanislaw Gruszka wrote:
>
> > Ok, I think I see this, it seems to be race condition in
> > ieee80211_sta_ps_deliver_wakeup().
> >
> > Perhaps it could be solved by modifying
> > ieee80211_add_pending_skbs_fn() to take list of queues as argument,
> > that function seems properly stop queues, add buffered frames to
> > pending queue, clear WLAN_STA_PS_STA and then wake up queues. Or
> > just stop using ieee80211_add_pending_skbs_fn() and do the same
> > sequence directly on ieee80211_sta_ps_deliver_wakeup() .
>
> That might work, but we were considering just locking the sta->lock
> around the whole "pending" thing inside deliver_wakeup() and in the "is
> station sleeping" part. Then you also need to re-check ("is it still
> sleeping?") inside the lock.
Looks good to me, feel free to post that.
Stanislaw
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
