diff --git a/drivers/net/wireless/mediatek/mt76/mt76.h b/drivers/net/wireless/mediatek/mt76/mt76.h
index 640061276d76..dc160d9ae537 100644
--- a/drivers/net/wireless/mediatek/mt76/mt76.h
+++ b/drivers/net/wireless/mediatek/mt76/mt76.h
@@ -539,6 +539,7 @@ struct mt76_hw_cap {
 #define MT_DRV_HW_MGMT_TXQ		BIT(4)
 #define MT_DRV_AMSDU_OFFLOAD		BIT(5)
 #define MT_DRV_IGNORE_TXS_FAILED	BIT(6)
+#define MT_DRV_HW_PS_BUFFERING		BIT(7)
 
 struct mt76_driver_ops {
 	u32 drv_flags;
@@ -1541,6 +1542,8 @@ void mt76_release_buffered_frames(struct ieee80211_hw *hw,
 				  u16 tids, int nframes,
 				  enum ieee80211_frame_release_type reason,
 				  bool more_data);
+void mt76_sta_ps_transition(struct mt76_dev *dev, struct mt76_wcid *wcid,
+			    bool ps);
 bool mt76_has_tx_pending(struct mt76_phy *phy);
 int mt76_update_channel(struct mt76_phy *phy);
 void mt76_update_survey(struct mt76_phy *phy);
diff --git a/drivers/net/wireless/mediatek/mt76/mt76_connac_mcu.h b/drivers/net/wireless/mediatek/mt76/mt76_connac_mcu.h
index 51380849d24e..0ededa569e9a 100644
--- a/drivers/net/wireless/mediatek/mt76/mt76_connac_mcu.h
+++ b/drivers/net/wireless/mediatek/mt76/mt76_connac_mcu.h
@@ -1094,6 +1094,7 @@ enum {
 	MCU_UNI_EVENT_IE_COUNTDOWN = 0x09,
 	MCU_UNI_EVENT_COREDUMP = 0x0a,
 	MCU_UNI_EVENT_BSS_BEACON_LOSS = 0x0c,
+	MCU_UNI_EVENT_PS_SYNC = 0x0d,
 	MCU_UNI_EVENT_SCAN_DONE = 0x0e,
 	MCU_UNI_EVENT_RDD_REPORT = 0x11,
 	MCU_UNI_EVENT_ROC = 0x27,
diff --git a/drivers/net/wireless/mediatek/mt76/tx.c b/drivers/net/wireless/mediatek/mt76/tx.c
index b03be0eb4712..12c615b1294b 100644
--- a/drivers/net/wireless/mediatek/mt76/tx.c
+++ b/drivers/net/wireless/mediatek/mt76/tx.c
@@ -267,6 +267,21 @@ void __mt76_tx_complete_skb(struct mt76_dev *dev, u16 wcid_idx, struct sk_buff *
 	wcid = __mt76_wcid_ptr(dev, wcid_idx);
 	mt76_tx_check_non_aql(dev, wcid, skb);
 
+	if (wcid && (dev->drv->drv_flags & MT_DRV_HW_PS_BUFFERING) &&
+	    test_bit(MT_WCID_FLAG_PS, &wcid->flags)) {
+		struct ieee80211_sta *sta = wcid_to_sta(wcid);
+
+		if (sta) {
+			struct ieee80211_hw *hw = mt76_phy_hw(dev, wcid->phy_idx);
+			int i;
+
+			for (i = 0; i < ARRAY_SIZE(sta->txq); i++)
+				if (sta->txq[i])
+					ieee80211_schedule_txq(hw, sta->txq[i]);
+			mt76_worker_schedule(&dev->tx_worker);
+		}
+	}
+
 #ifdef CONFIG_NL80211_TESTMODE
 	if (mt76_is_testmode_skb(dev, skb, &hw)) {
 		struct mt76_phy *phy = hw->priv;
@@ -476,8 +491,12 @@ mt76_txq_send_burst(struct mt76_phy *phy, struct mt76_queue *q,
 	bool stop = false;
 	int idx;
 
-	if (test_bit(MT_WCID_FLAG_PS, &wcid->flags))
-		return 0;
+	if (test_bit(MT_WCID_FLAG_PS, &wcid->flags)) {
+		if (!(dev->drv->drv_flags & MT_DRV_HW_PS_BUFFERING))
+			return 0;
+		if (ieee80211_txq_aql_pending(phy->hw, txq))
+			return 0;
+	}
 
 	if (atomic_read(&wcid->non_aql_packets) >= MT_MAX_NON_AQL_PKT)
 		return 0;
@@ -497,6 +516,9 @@ mt76_txq_send_burst(struct mt76_phy *phy, struct mt76_queue *q,
 	if (idx < 0)
 		return idx;
 
+	if (test_bit(MT_WCID_FLAG_PS, &wcid->flags))
+		goto out;
+
 	do {
 		if (test_bit(MT76_RESET, &phy->state) || phy->offchannel)
 			break;
@@ -522,6 +544,7 @@ mt76_txq_send_burst(struct mt76_phy *phy, struct mt76_queue *q,
 		n_frames++;
 	} while (1);
 
+out:
 	spin_lock(&q->lock);
 	dev->queue_ops->kick(dev, q);
 	spin_unlock(&q->lock);
@@ -548,10 +571,7 @@ mt76_txq_schedule_list(struct mt76_phy *phy, enum mt76_txq_id qid)
 
 		mtxq = (struct mt76_txq *)txq->drv_priv;
 		wcid = __mt76_wcid_ptr(dev, mtxq->wcid);
-		if (!wcid || test_bit(MT_WCID_FLAG_PS, &wcid->flags))
-			continue;
-
-		if (atomic_read(&wcid->non_aql_packets) >= MT_MAX_NON_AQL_PKT)
+		if (!wcid)
 			continue;
 
 		phy = mt76_dev_phy(dev, wcid->phy_idx);
@@ -559,6 +579,24 @@ mt76_txq_schedule_list(struct mt76_phy *phy, enum mt76_txq_id qid)
 			continue;
 
 		q = phy->q_tx[qid];
+
+		if (test_bit(MT_WCID_FLAG_PS, &wcid->flags)) {
+			if (!(dev->drv->drv_flags & MT_DRV_HW_PS_BUFFERING))
+				continue;
+
+			if (!mt76_txq_stopped(q))
+				n_frames = mt76_txq_send_burst(phy, q, mtxq, wcid);
+
+			ieee80211_return_txq(phy->hw, txq, false);
+
+			if (unlikely(n_frames < 0))
+				return n_frames;
+			ret += n_frames;
+			continue;
+		}
+
+		if (atomic_read(&wcid->non_aql_packets) >= MT_MAX_NON_AQL_PKT)
+			continue;
 		if (dev->queue_ops->tx_cleanup &&
 		    q->queued + 2 * MT_TXQ_FREE_THR >= q->ndesc) {
 			dev->queue_ops->tx_cleanup(dev, q, false);
@@ -765,6 +803,33 @@ void mt76_stop_tx_queues(struct mt76_phy *phy, struct ieee80211_sta *sta,
 }
 EXPORT_SYMBOL_GPL(mt76_stop_tx_queues);
 
+void mt76_sta_ps_transition(struct mt76_dev *dev, struct mt76_wcid *wcid,
+			    bool ps)
+{
+	struct ieee80211_sta *sta;
+	struct ieee80211_hw *hw;
+	int i;
+
+	if (ps) {
+		set_bit(MT_WCID_FLAG_PS, &wcid->flags);
+		mt76_worker_schedule(&dev->tx_worker);
+		return;
+	}
+
+	clear_bit(MT_WCID_FLAG_PS, &wcid->flags);
+
+	sta = wcid_to_sta(wcid);
+	if (!sta)
+		return;
+
+	hw = mt76_phy_hw(dev, wcid->phy_idx);
+	for (i = 0; i < ARRAY_SIZE(sta->txq); i++)
+		if (sta->txq[i])
+			ieee80211_schedule_txq(hw, sta->txq[i]);
+	mt76_worker_schedule(&dev->tx_worker);
+}
+EXPORT_SYMBOL_GPL(mt76_sta_ps_transition);
+
 void mt76_wake_tx_queue(struct ieee80211_hw *hw, struct ieee80211_txq *txq)
 {
 	struct mt76_phy *phy = hw->priv;