Merge branch 'for-davem' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-next

John W. Linville says:

====================
Included is a Bluetooth pull -- Gustavo says:

"These are the Bluetooth bits for inclusion in 3.8, there is basically one big
thing here which is the High Speed patches from Andrei, he did a lot of work on
A2MP and management of AMP devices. The rest are mostly clean up and bug
fixes."

Also included is an NFC pull -- Samuel says:

"With this one we have:

- pn544 p2p support.
- pn544 physical and HCI layers separation. We are getting the pn544 driver
  ready to support non i2c physical layers.
- LLCP SNL (Service Name Lookup). This is the NFC p2p service discovery
  protocol.
- LLCP datagram sockets (connection less) support.
- IDR library usage for NFC devices indexes assignement.
- NFC netlink extension for setting and getting LLCP link characteristics.
- Various code style fixes and cleanups spread over the pn533, LLCP, HCI and
  pn544 code."

There are a couple of mac80211 pulls as well -- Johannes says:

"Please pull my mac80211-next tree to get the first round of new features
for 3.8. We have:
 * finally, the mac80211 multi-channel work
 * scan improvements:
   - bg scan
   - scan flush
   - forced AP scan
 * cfg80211 tracing
 * a bit of new code to allow implementing SAE (secure authentication of
   equals) in managed mode

Along with a few random improvements, features and fixes."

and...

"Please pull from mac80211-next (per below pull request) to get a few
updates. Most important is probably the fix for the WDS regression that
my previous pull request introduced. Other than that, I have some
tracing code, two mesh updates and a change to allow drivers to
calculate the AES CMAC subkeys without having to implement the GF_mulx
operation themselves."

On top of that are the usual updates to iwlwifi, ath9k, rt2x00,
brcmfmac, mwifiex, and a few others here and there.  Of note is the
addition of the ar5523 driver, ported from an original FreeBSD driver.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 8 insertions, 7 deletions

diff --git a/net/mac80211/status.c b/net/mac80211/status.c
index 3af0cc4..21fa5c7 100644
--- a/net/mac80211/status.c
+++ b/net/mac80211/status.c
@@ -189,30 +189,31 @@ static void ieee80211_frame_acked(struct sta_info *sta, struct sk_buff *skb)
 	}
 
 	if (ieee80211_is_action(mgmt->frame_control) &&
-	    sdata->vif.type == NL80211_IFTYPE_STATION &&
 	    mgmt->u.action.category == WLAN_CATEGORY_HT &&
-	    mgmt->u.action.u.ht_smps.action == WLAN_HT_ACTION_SMPS) {
+	    mgmt->u.action.u.ht_smps.action == WLAN_HT_ACTION_SMPS &&
+	    sdata->vif.type == NL80211_IFTYPE_STATION &&
+	    ieee80211_sdata_running(sdata)) {
 		/*
 		 * This update looks racy, but isn't -- if we come
 		 * here we've definitely got a station that we're
 		 * talking to, and on a managed interface that can
 		 * only be the AP. And the only other place updating
-		 * this variable is before we're associated.
+		 * this variable in managed mode is before association.
 		 */
 		switch (mgmt->u.action.u.ht_smps.smps_control) {
 		case WLAN_HT_SMPS_CONTROL_DYNAMIC:
-			sta->sdata->u.mgd.ap_smps = IEEE80211_SMPS_DYNAMIC;
+			sdata->smps_mode = IEEE80211_SMPS_DYNAMIC;
 			break;
 		case WLAN_HT_SMPS_CONTROL_STATIC:
-			sta->sdata->u.mgd.ap_smps = IEEE80211_SMPS_STATIC;
+			sdata->smps_mode = IEEE80211_SMPS_STATIC;
 			break;
 		case WLAN_HT_SMPS_CONTROL_DISABLED:
 		default: /* shouldn't happen since we don't send that */
-			sta->sdata->u.mgd.ap_smps = IEEE80211_SMPS_OFF;
+			sdata->smps_mode = IEEE80211_SMPS_OFF;
 			break;
 		}
 
-		ieee80211_queue_work(&local->hw, &local->recalc_smps);
+		ieee80211_queue_work(&local->hw, &sdata->recalc_smps);
 	}
 }