Skip to contentSkip to conversation

ath9k patches

RFC In review Sujith Manoharan 1 participant 6 messages Sep 24, 2014 Qualcomm Atheros ATH9K wireless driver

lore
Share & tools
lore
#0 lore raw

cover ath9k patches

From: Sujith Manoharan <[email protected]>

MCC fixes. Please review !

Sujith Manoharan (4):
ath9k: Fix queue management
ath9k: Use normal queues for offchannel frames
ath9k: Fix force_channel usage for offchannel frames
ath9k: Fix offchannel queuing

 drivers/net/wireless/ath/ath9k/xmit.c | 20 ++++++--------------
 1 file changed, 6 insertions(+), 14 deletions(-)
Signature · 1 line
2.1.0
#1 lore raw

RFC 1/4 ath9k: Fix queue management

From: Sujith Manoharan <[email protected]>

Since we use IEEE80211_HW_QUEUE_CONTROL now, the CAB/Offchannel queues are registered as the last two queues. There is no need to check and reassign the queues in the TX start()/done() routines.

CAB frames will not reach the tx() callback since we set IEEE80211_HW_HOST_BROADCAST_PS_BUFFERING and pull the buffered frames during beacon transmission. We also don't have a special HW queue for handling off-channel frames.

---
 drivers/net/wireless/ath/ath9k/xmit.c | 10 +++-------
 1 file changed, 3 insertions(+), 7 deletions(-)
drivers/net/wireless/ath/ath9k/xmit.c +3 -7
index 93ad31b..2c8327f 100644
@@ -158,7 +158,6 @@ static void ath_txq_skb_done(struct ath_softc *sc, struct ath_txq *txq,
 {
 	struct ieee80211_tx_info *info = IEEE80211_SKB_CB(skb);
 	struct ath_frame_info *fi = get_frame_info(skb);
-	int hw_queue;
 	int q = fi->txq;
 
 	if (q < 0)
@@ -168,10 +167,9 @@ static void ath_txq_skb_done(struct ath_softc *sc, struct ath_txq *txq,
 	if (WARN_ON(--txq->pending_frames < 0))
 		txq->pending_frames = 0;
 
-	hw_queue = (info->hw_queue >= sc->hw->queues - 2) ? q : info->hw_queue;
 	if (txq->stopped &&
 	    txq->pending_frames < sc->tx.txq_max_pending[q]) {
-		ieee80211_wake_queue(sc->hw, hw_queue);
+		ieee80211_wake_queue(sc->hw, info->hw_queue);
 		txq->stopped = false;
 	}
 }
@@ -2208,8 +2206,7 @@ int ath_tx_start(struct ieee80211_hw *hw, struct sk_buff *skb,
 	struct ath_atx_tid *tid = NULL;
 	struct ath_buf *bf;
 	bool queue;
-	int q, hw_queue;
-	int ret;
+	int q, ret;
 
 	if (vif)
 		avp = (void *)vif->drv_priv;
@@ -2228,14 +2225,13 @@ int ath_tx_start(struct ieee80211_hw *hw, struct sk_buff *skb,
 	 */
 
 	q = skb_get_queue_mapping(skb);
-	hw_queue = (info->hw_queue >= sc->hw->queues - 2) ? q : info->hw_queue;
 
 	ath_txq_lock(sc, txq);
 	if (txq == sc->tx.txq_map[q]) {
 		fi->txq = q;
 		if (++txq->pending_frames > sc->tx.txq_max_pending[q] &&
 		    !txq->stopped) {
-			ieee80211_stop_queue(sc->hw, hw_queue);
+			ieee80211_stop_queue(sc->hw, info->hw_queue);
 			txq->stopped = true;
 		}
 	}
Signature · 1 line
2.1.0
#2 lore raw

RFC 2/4 ath9k: Use normal queues for offchannel frames

In reply to Sujith Manoharan

From: Sujith Manoharan <[email protected]>

There is no reason why frames marked with IEEE80211_TX_CTL_TX_OFFCHAN have to be sent using the UAPSD queue. Since mac80211 makes sure that RoC is done before pushing an offchannel frame to the driver, we can use the normal TX queues for transmission.

---
 drivers/net/wireless/ath/ath9k/xmit.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
drivers/net/wireless/ath/ath9k/xmit.c +1 -2
index 2c8327f..00931b6 100644
@@ -2253,8 +2253,7 @@ int ath_tx_start(struct ieee80211_hw *hw, struct sk_buff *skb,
 	if (txctl->an && queue)
 		tid = ath_get_skb_tid(sc, txctl->an, skb);
 
-	if (info->flags & (IEEE80211_TX_CTL_PS_RESPONSE |
-			   IEEE80211_TX_CTL_TX_OFFCHAN)) {
+	if (info->flags & IEEE80211_TX_CTL_PS_RESPONSE) {
 		ath_txq_unlock(sc, txq);
 		txq = sc->tx.uapsdq;
 		ath_txq_lock(sc, txq);
Signature · 1 line
2.1.0
#3 lore raw

RFC 3/4 ath9k: Fix force_channel usage for offchannel frames

In reply to Sujith Manoharan

From: Sujith Manoharan <[email protected]>

Since RoC is done before frames marked with IEEE80211_TX_CTL_TX_OFFCHAN are received by the driver, setting force_channel is useless. We will be in the required offchannel, so incoming frames can be transmitted immediately.

---
 drivers/net/wireless/ath/ath9k/xmit.c | 3 ---
 1 file changed, 3 deletions(-)
drivers/net/wireless/ath/ath9k/xmit.c +0 -3
index 00931b6..4d00f59 100644
@@ -2211,9 +2211,6 @@ int ath_tx_start(struct ieee80211_hw *hw, struct sk_buff *skb,
 	if (vif)
 		avp = (void *)vif->drv_priv;
 
-	if (info->flags & IEEE80211_TX_CTL_TX_OFFCHAN)
-		txctl->force_channel = true;
-
 	ret = ath_tx_prepare(hw, skb, txctl);
 	if (ret)
 	    return ret;
Signature · 1 line
2.1.0
#4 lore raw

Re: [RFC 3/4] ath9k: Fix force_channel usage for offchannel frames

Quoted 18 lines · Sujith Manoharan

From: Sujith Manoharan <[email protected]>

Since RoC is done before frames marked with IEEE80211_TX_CTL_TX_OFFCHAN are received by the driver, setting force_channel is useless. We will be in the required offchannel, so incoming frames can be transmitted immediately.

---
 drivers/net/wireless/ath/ath9k/xmit.c | 3 ---
 1 file changed, 3 deletions(-)
drivers/net/wireless/ath/ath9k/xmit.c +0 -0
index 00931b6..4d00f59 100644

drivers/net/wireless/ath/ath9k/xmit.c int ath_tx_start(struct ieee80211_hw *hw, struct sk_buff *skb, · line 2216

Quoted 9 lines
drivers/net/wireless/ath/ath9k/xmit.c +0 -3
@@ -2211,9 +2211,6 @@ int ath_tx_start(struct ieee80211_hw *hw, struct sk_buff *skb,
 	if (vif)
 		avp = (void *)vif->drv_priv;
 
-	if (info->flags & IEEE80211_TX_CTL_TX_OFFCHAN)
-		txctl->force_channel = true;
-
 	ret = ath_tx_prepare(hw, skb, txctl);
 	if (ret)
 	    return ret;

This breaks things, so this patch can be dropped.

Sujith

#5 lore raw

RFC 4/4 ath9k: Fix offchannel queuing

In reply to Sujith Manoharan

From: Sujith Manoharan <[email protected]>

Clearing IEEE80211_TX_CTL_PS_RESPONSE in a frame that is not in the current context doesn't seem right. Instead make sure that we don't add such frames to the UAPSD queue by using a local variable.

---
 drivers/net/wireless/ath/ath9k/xmit.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/net/wireless/ath/ath9k/xmit.c +3 -3
index 4d00f59..3e3a41c 100644
@@ -2205,7 +2205,7 @@ int ath_tx_start(struct ieee80211_hw *hw, struct sk_buff *skb,
 	struct ath_txq *txq = txctl->txq;
 	struct ath_atx_tid *tid = NULL;
 	struct ath_buf *bf;
-	bool queue;
+	bool queue, skip_uapsd;
 	int q, ret;
 
 	if (vif)
@@ -2243,14 +2243,14 @@ int ath_tx_start(struct ieee80211_hw *hw, struct sk_buff *skb,
 	     sc->cur_chan->stopped) && !txctl->force_channel) {
 		if (!txctl->an)
 			txctl->an = &avp->mcast_node;
-		info->flags &= ~IEEE80211_TX_CTL_PS_RESPONSE;
 		queue = true;
+		skip_uapsd = true;
 	}
 
 	if (txctl->an && queue)
 		tid = ath_get_skb_tid(sc, txctl->an, skb);
 
-	if (info->flags & IEEE80211_TX_CTL_PS_RESPONSE) {
+	if (!skip_uapsd && (info->flags & IEEE80211_TX_CTL_PS_RESPONSE)) {
 		ath_txq_unlock(sc, txq);
 		txq = sc->tx.uapsdq;
 		ath_txq_lock(sc, txq);
Signature · 1 line
2.1.0