mt76: unify queue tx cleanup code

Cleanup and preparation for changing tx scheduling behavior
Signed-off-by: Felix Fietkau <nbd@nbd.name>

drivers/net/wireless/mediatek/mt76/dma.c

@@ -165,16 +165,8 @@ mt76_dma_tx_cleanup(struct mt76_dev *dev, enum mt76_txq_id qid, bool flush)
 		last = readl(&q->regs->dma_idx);
 
 	while (q->queued > 0 && q->tail != last) {
-		int swq_qid = -1;
-
 		mt76_dma_tx_cleanup_idx(dev, q, q->tail, &entry);
-		if (entry.schedule)
-			swq_qid = entry.qid;
-
-		q->tail = (q->tail + 1) % q->ndesc;
-
-		if (entry.skb)
-			dev->drv->tx_complete_skb(dev, qid, &entry);
+		mt76_queue_tx_complete(dev, q, &entry);
 
 		if (entry.txwi) {
 			if (!(dev->drv->drv_flags & MT_DRV_TXWI_NO_FREE))
@@ -185,13 +177,6 @@ mt76_dma_tx_cleanup(struct mt76_dev *dev, enum mt76_txq_id qid, bool flush)
 		if (!flush && q->tail == last)
 			last = readl(&q->regs->dma_idx);
 
-		spin_lock_bh(&q->lock);
-		if (swq_qid >= 4)
-			dev->q_tx[__MT_TXQ_MAX + swq_qid - 4].swq_queued--;
-		else if (swq_qid >= 0)
-			dev->q_tx[swq_qid].swq_queued--;
-		q->queued--;
-		spin_unlock_bh(&q->lock);
 	}
 
 	if (flush) {

drivers/net/wireless/mediatek/mt76/mt76.h

@@ -1017,6 +1017,8 @@ void mt76_rx_poll_complete(struct mt76_dev *dev, enum mt76_rxq_id q,
 			   struct napi_struct *napi);
 void mt76_rx_aggr_reorder(struct sk_buff *skb, struct sk_buff_head *frames);
 void mt76_testmode_tx_pending(struct mt76_dev *dev);
+void mt76_queue_tx_complete(struct mt76_dev *dev, struct mt76_queue *q,
+			    struct mt76_queue_entry *e);
 
 /* usb */
 static inline bool mt76u_urb_error(struct urb *urb)

drivers/net/wireless/mediatek/mt76/sdio.c

@@ -133,38 +133,28 @@ mt76s_process_rx_queue(struct mt76_dev *dev, struct mt76_queue *q)
 	return nframes;
 }
 
-static int mt76s_process_tx_queue(struct mt76_dev *dev, enum mt76_txq_id qid)
+static void mt76s_process_tx_queue(struct mt76_dev *dev, enum mt76_txq_id qid)
 {
 	struct mt76_sw_queue *sq = &dev->q_tx[qid];
-	u32 n_dequeued = 0, n_sw_dequeued = 0;
 	struct mt76_queue_entry entry;
 	struct mt76_queue *q = sq->q;
 	bool wake;
 
-	while (q->queued > n_dequeued) {
+	while (q->queued > 0) {
 		if (!q->entry[q->tail].done)
 			break;
 
-		if (q->entry[q->tail].schedule) {
-			q->entry[q->tail].schedule = false;
-			n_sw_dequeued++;
-		}
-
 		entry = q->entry[q->tail];
 		q->entry[q->tail].done = false;
-		q->tail = (q->tail + 1) % q->ndesc;
-		n_dequeued++;
+		q->entry[q->tail].schedule = false;
 
-		if (qid == MT_TXQ_MCU)
+		if (qid == MT_TXQ_MCU) {
 			dev_kfree_skb(entry.skb);
-		else
-			dev->drv->tx_complete_skb(dev, qid, &entry);
+			entry.skb = NULL;
 		}
 
-	spin_lock_bh(&q->lock);
-
-	sq->swq_queued -= n_sw_dequeued;
-	q->queued -= n_dequeued;
+		mt76_queue_tx_complete(dev, q, &entry);
+	}
 
 	wake = q->stopped && q->queued < q->ndesc - 8;
 	if (wake)
@@ -173,18 +163,13 @@ static int mt76s_process_tx_queue(struct mt76_dev *dev, enum mt76_txq_id qid)
 	if (!q->queued)
 		wake_up(&dev->tx_wait);
 
-	spin_unlock_bh(&q->lock);
-
 	if (qid == MT_TXQ_MCU)
-		goto out;
+		return;
 
 	mt76_txq_schedule(&dev->phy, qid);
 
 	if (wake)
 		ieee80211_wake_queue(dev->hw, qid);
-
-out:
-	return n_dequeued;
 }
 
 static void mt76s_tx_status_data(struct work_struct *work)

drivers/net/wireless/mediatek/mt76/tx.c

@@ -696,3 +696,25 @@ int mt76_skb_adjust_pad(struct sk_buff *skb)
 	return 0;
 }
 EXPORT_SYMBOL_GPL(mt76_skb_adjust_pad);
+
+void mt76_queue_tx_complete(struct mt76_dev *dev, struct mt76_queue *q,
+			    struct mt76_queue_entry *e)
+{
+	enum mt76_txq_id qid = e->qid % 4;
+	bool ext_phy = e->qid >= 4;
+
+	if (e->skb)
+		dev->drv->tx_complete_skb(dev, qid, e);
+
+	spin_lock_bh(&q->lock);
+	q->tail = (q->tail + 1) % q->ndesc;
+	q->queued--;
+
+	if (ext_phy)
+		qid += __MT_TXQ_MAX;
+
+	if (e->schedule)
+		dev->q_tx[qid].swq_queued--;
+	spin_unlock_bh(&q->lock);
+}
+EXPORT_SYMBOL_GPL(mt76_queue_tx_complete);

drivers/net/wireless/mediatek/mt76/usb.c

@@ -802,33 +802,20 @@ static void mt76u_tx_tasklet(unsigned long data)
 	int i;
 
 	for (i = 0; i < IEEE80211_NUM_ACS; i++) {
-		u32 n_dequeued = 0, n_sw_dequeued = 0;
-
 		sq = &dev->q_tx[i];
 		q = sq->q;
 
-		while (q->queued > n_dequeued) {
+		while (q->queued > 0) {
 			if (!q->entry[q->tail].done)
 				break;
 
-			if (q->entry[q->tail].schedule) {
-				q->entry[q->tail].schedule = false;
-				n_sw_dequeued++;
-			}
-
 			entry = q->entry[q->tail];
 			q->entry[q->tail].done = false;
-			q->tail = (q->tail + 1) % q->ndesc;
-			n_dequeued++;
+			q->entry[q->tail].schedule = false;
 
-			dev->drv->tx_complete_skb(dev, i, &entry);
+			mt76_queue_tx_complete(dev, q, &entry);
 		}
 
-		spin_lock_bh(&q->lock);
-
-		sq->swq_queued -= n_sw_dequeued;
-		q->queued -= n_dequeued;
-
 		wake = q->stopped && q->queued < q->ndesc - 8;
 		if (wake)
 			q->stopped = false;
@@ -836,8 +823,6 @@ static void mt76u_tx_tasklet(unsigned long data)
 		if (!q->queued)
 			wake_up(&dev->tx_wait);
 
-		spin_unlock_bh(&q->lock);
-
 		mt76_txq_schedule(&dev->phy, i);
 
 		if (dev->drv->tx_status_data &&
@@ -1068,16 +1053,11 @@ void mt76u_stop_tx(struct mt76_dev *dev)
 			if (!q)
 				continue;
 
-			/* Assure we are in sync with killed tasklet. */
-			spin_lock_bh(&q->lock);
-			while (q->queued) {
 			entry = q->entry[q->tail];
-				q->tail = (q->tail + 1) % q->ndesc;
-				q->queued--;
+			q->entry[q->tail].done = false;
+			q->entry[q->tail].schedule = false;
 
-				dev->drv->tx_complete_skb(dev, i, &entry);
-			}
-			spin_unlock_bh(&q->lock);
+			mt76_queue_tx_complete(dev, q, &entry);
 		}
 	}