mac80211: reset probe_send_count also in HW_CONNECTION_MONITOR case
authorEliad Peller <eliad@wizery.com>
Tue, 11 Feb 2014 10:30:18 +0000 (12:30 +0200)
committerJohannes Berg <johannes.berg@intel.com>
Tue, 11 Feb 2014 11:58:32 +0000 (12:58 +0100)
In case of beacon_loss with IEEE80211_HW_CONNECTION_MONITOR
device, mac80211 probes the ap (and disconnects on timeout)
but ignores the ack.

If we already got an ack, there's no reason to continue
disconnecting. this can help devices that supports
IEEE80211_HW_CONNECTION_MONITOR only partially (e.g. take
care of keep alives, but does not probe the ap.

In case the device wants to disconnect without probing,
it can just call ieee80211_connection_loss.

Signed-off-by: Eliad Peller <eliadx.peller@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
include/net/mac80211.h
net/mac80211/mlme.c

index 4f0f29d..4005c5b 100644 (file)
@@ -1507,8 +1507,6 @@ struct ieee80211_tx_control {
  * @IEEE80211_HW_CONNECTION_MONITOR:
  *     The hardware performs its own connection monitoring, including
  *     periodic keep-alives to the AP and probing the AP on beacon loss.
- *     When this flag is set, signaling beacon-loss will cause an immediate
- *     change to disassociated state.
  *
  * @IEEE80211_HW_NEED_DTIM_BEFORE_ASSOC:
  *     This device needs to get data from beacon before association (i.e.
index 6160483..b9432b5 100644 (file)
@@ -131,13 +131,13 @@ void ieee80211_sta_reset_conn_monitor(struct ieee80211_sub_if_data *sdata)
        if (unlikely(!sdata->u.mgd.associated))
                return;
 
+       ifmgd->probe_send_count = 0;
+
        if (sdata->local->hw.flags & IEEE80211_HW_CONNECTION_MONITOR)
                return;
 
        mod_timer(&sdata->u.mgd.conn_mon_timer,
                  round_jiffies_up(jiffies + IEEE80211_CONNECTION_IDLE_TIME));
-
-       ifmgd->probe_send_count = 0;
 }
 
 static int ecw2cw(int ecw)