diff --git a/sys/dev/rtwn/if_rtwnvar.h b/sys/dev/rtwn/if_rtwnvar.h
--- a/sys/dev/rtwn/if_rtwnvar.h
+++ b/sys/dev/rtwn/if_rtwnvar.h
@@ -133,7 +133,8 @@
  */
 enum {
 	RTWN_RX_DATA,
-	RTWN_RX_TX_REPORT,
+	RTWN_RX_TX_REPORT,	/* Per-packet */
+	RTWN_RX_TX_REPORT2,	/* Per-MACID summary */
 	RTWN_RX_OTHER
 };
 
@@ -348,6 +349,8 @@
 	int		(*sc_classify_intr)(struct rtwn_softc *, void *, int);
 	void		(*sc_handle_tx_report)(struct rtwn_softc *, uint8_t *,
 			    int);
+	void		(*sc_handle_tx_report2)(struct rtwn_softc *, uint8_t *,
+			    int);
 	void		(*sc_handle_c2h_report)(struct rtwn_softc *,
 			    uint8_t *, int);
 	int		(*sc_check_frame)(struct rtwn_softc *, struct mbuf *);
@@ -549,6 +552,8 @@
 	(((_sc)->sc_classify_intr)((_sc), (_buf), (_len)))
 #define rtwn_handle_tx_report(_sc, _buf, _len) \
 	(((_sc)->sc_handle_tx_report)((_sc), (_buf), (_len)))
+#define rtwn_handle_tx_report2(_sc, _buf, _len) \
+	(((_sc)->sc_handle_tx_report2)((_sc), (_buf), (_len)))
 #define rtwn_handle_c2h_report(_sc, _buf, _len) \
 	(((_sc)->sc_handle_c2h_report)((_sc), (_buf), (_len)))
 #define rtwn_check_frame(_sc, _m) \
diff --git a/sys/dev/rtwn/pci/rtwn_pci_rx.c b/sys/dev/rtwn/pci/rtwn_pci_rx.c
--- a/sys/dev/rtwn/pci/rtwn_pci_rx.c
+++ b/sys/dev/rtwn/pci/rtwn_pci_rx.c
@@ -236,6 +236,33 @@
 #endif
 }
 
+static void
+rtwn_pci_tx_report2(struct rtwn_pci_softc *pc, int len)
+{
+	struct rtwn_softc *sc = &pc->pc_sc;
+
+	if (sc->sc_ratectl != RTWN_RATECTL_NET80211) {
+		/* shouldn't happen */
+		device_printf(sc->sc_dev,
+		    "%s called while ratectl = %d!\n",
+		     __func__, sc->sc_ratectl);
+		return;
+	}
+
+	RTWN_NT_LOCK(sc);
+	rtwn_handle_tx_report2(sc, pc->pc_rx_buf, len);
+	RTWN_NT_UNLOCK(sc);
+
+#ifdef IEEE80211_SUPPORT_SUPERG
+	/*
+	 * NB: this will executed only when 'report' bit is set.
+	 */
+	if (sc->sc_tx_n_active > 0 && --sc->sc_tx_n_active <= 1)
+		rtwn_cmd_sleepable(sc, NULL, 0, rtwn_ff_flush_all);
+#endif
+}
+
+
 static void
 rtwn_pci_c2h_report(struct rtwn_pci_softc *pc, int len)
 {
@@ -341,6 +368,9 @@
 		case RTWN_RX_TX_REPORT:
 			rtwn_pci_tx_report(pc, len);
 			break;
+		case RTWN_RX_TX_REPORT2:
+			rtwn_pci_tx_report2(pc, len);
+			break;
 		case RTWN_RX_OTHER:
 			rtwn_pci_c2h_report(pc, len);
 			break;
diff --git a/sys/dev/rtwn/rtl8188e/pci/r88ee_attach.c b/sys/dev/rtwn/rtl8188e/pci/r88ee_attach.c
--- a/sys/dev/rtwn/rtl8188e/pci/r88ee_attach.c
+++ b/sys/dev/rtwn/rtl8188e/pci/r88ee_attach.c
@@ -145,6 +145,7 @@
 	sc->sc_get_rssi_ofdm		= r88e_get_rssi_ofdm;
 	sc->sc_classify_intr		= r88e_classify_intr;
 	sc->sc_handle_tx_report		= r88e_ratectl_tx_complete;
+	sc->sc_handle_tx_report2	= rtwn_nop_softc_uint8_int;
 	sc->sc_handle_c2h_report	= r88e_handle_c2h_report;
 	sc->sc_check_frame		= rtwn_nop_int_softc_mbuf;
 	sc->sc_rf_read			= r92c_rf_read;
diff --git a/sys/dev/rtwn/rtl8188e/r88e_rx.c b/sys/dev/rtwn/rtl8188e/r88e_rx.c
--- a/sys/dev/rtwn/rtl8188e/r88e_rx.c
+++ b/sys/dev/rtwn/rtl8188e/r88e_rx.c
@@ -63,8 +63,9 @@
 	case R88E_RXDW3_RPT_RX:
 		return (RTWN_RX_DATA);
 	case R88E_RXDW3_RPT_TX1:	/* per-packet Tx report */
-	case R88E_RXDW3_RPT_TX2:	/* periodical Tx report */
 		return (RTWN_RX_TX_REPORT);
+	case R88E_RXDW3_RPT_TX2:	/* periodical Tx report */
+		return (RTWN_RX_TX_REPORT2);
 	case R88E_RXDW3_RPT_HIS:
 		return (RTWN_RX_OTHER);
 	default:			/* shut up the compiler */
diff --git a/sys/dev/rtwn/rtl8188e/usb/r88eu_attach.c b/sys/dev/rtwn/rtl8188e/usb/r88eu_attach.c
--- a/sys/dev/rtwn/rtl8188e/usb/r88eu_attach.c
+++ b/sys/dev/rtwn/rtl8188e/usb/r88eu_attach.c
@@ -138,6 +138,7 @@
 	sc->sc_get_rssi_ofdm		= r88e_get_rssi_ofdm;
 	sc->sc_classify_intr		= r88e_classify_intr;
 	sc->sc_handle_tx_report		= r88e_ratectl_tx_complete;
+	sc->sc_handle_tx_report2	= rtwn_nop_softc_uint8_int;
 	sc->sc_handle_c2h_report	= r88e_handle_c2h_report;
 	sc->sc_check_frame		= rtwn_nop_int_softc_mbuf;
 	sc->sc_rf_read			= r92c_rf_read;
diff --git a/sys/dev/rtwn/rtl8192c/pci/r92ce_attach.c b/sys/dev/rtwn/rtl8192c/pci/r92ce_attach.c
--- a/sys/dev/rtwn/rtl8192c/pci/r92ce_attach.c
+++ b/sys/dev/rtwn/rtl8192c/pci/r92ce_attach.c
@@ -175,6 +175,7 @@
 	sc->sc_get_rssi_ofdm		= r92c_get_rssi_ofdm;
 	sc->sc_classify_intr		= r92c_classify_intr;
 	sc->sc_handle_tx_report		= rtwn_nop_softc_uint8_int;
+	sc->sc_handle_tx_report2	= rtwn_nop_softc_uint8_int;
 	sc->sc_handle_c2h_report	= rtwn_nop_softc_uint8_int;
 	sc->sc_check_frame		= rtwn_nop_int_softc_mbuf;
 	sc->sc_rf_read			= r92c_rf_read;
diff --git a/sys/dev/rtwn/rtl8192c/usb/r92cu_attach.c b/sys/dev/rtwn/rtl8192c/usb/r92cu_attach.c
--- a/sys/dev/rtwn/rtl8192c/usb/r92cu_attach.c
+++ b/sys/dev/rtwn/rtl8192c/usb/r92cu_attach.c
@@ -167,6 +167,7 @@
 	sc->sc_get_rssi_ofdm		= r92c_get_rssi_ofdm;
 	sc->sc_classify_intr		= r92c_classify_intr;
 	sc->sc_handle_tx_report		= rtwn_nop_softc_uint8_int;
+	sc->sc_handle_tx_report2	= rtwn_nop_softc_uint8_int;
 	sc->sc_handle_c2h_report	= rtwn_nop_softc_uint8_int;
 	sc->sc_check_frame		= rtwn_nop_int_softc_mbuf;
 	sc->sc_rf_read			= r92c_rf_read;
diff --git a/sys/dev/rtwn/rtl8192e/usb/r92eu_attach.c b/sys/dev/rtwn/rtl8192e/usb/r92eu_attach.c
--- a/sys/dev/rtwn/rtl8192e/usb/r92eu_attach.c
+++ b/sys/dev/rtwn/rtl8192e/usb/r92eu_attach.c
@@ -116,6 +116,7 @@
 	sc->sc_get_rssi_ofdm		= r88e_get_rssi_ofdm;
 	sc->sc_classify_intr		= r12au_classify_intr;
 	sc->sc_handle_tx_report		= r12a_ratectl_tx_complete;
+	sc->sc_handle_tx_report2	= rtwn_nop_softc_uint8_int;
 	sc->sc_handle_c2h_report	= r92e_handle_c2h_report;
 	sc->sc_check_frame		= rtwn_nop_int_softc_mbuf;
 	sc->sc_rf_read			= r92e_rf_read;
diff --git a/sys/dev/rtwn/rtl8812a/usb/r12au_attach.c b/sys/dev/rtwn/rtl8812a/usb/r12au_attach.c
--- a/sys/dev/rtwn/rtl8812a/usb/r12au_attach.c
+++ b/sys/dev/rtwn/rtl8812a/usb/r12au_attach.c
@@ -205,6 +205,7 @@
 	sc->sc_get_rssi_ofdm		= r88e_get_rssi_ofdm;
 	sc->sc_classify_intr		= r12au_classify_intr;
 	sc->sc_handle_tx_report		= r12a_ratectl_tx_complete;
+	sc->sc_handle_tx_report2	= rtwn_nop_softc_uint8_int;
 	sc->sc_handle_c2h_report	= r12a_handle_c2h_report;
 	sc->sc_check_frame		= r12a_check_frame_checksum;
 	sc->sc_rf_write			= r12a_rf_write;
diff --git a/sys/dev/rtwn/rtl8821a/usb/r21au_attach.c b/sys/dev/rtwn/rtl8821a/usb/r21au_attach.c
--- a/sys/dev/rtwn/rtl8821a/usb/r21au_attach.c
+++ b/sys/dev/rtwn/rtl8821a/usb/r21au_attach.c
@@ -190,6 +190,7 @@
 	sc->sc_get_rssi_ofdm		= r88e_get_rssi_ofdm;
 	sc->sc_classify_intr		= r12au_classify_intr;
 	sc->sc_handle_tx_report		= r12a_ratectl_tx_complete;
+	sc->sc_handle_tx_report2	= rtwn_nop_softc_uint8_int;
 	sc->sc_handle_c2h_report	= r12a_handle_c2h_report;
 	sc->sc_check_frame		= r12a_check_frame_checksum;
 	sc->sc_rf_read			= r12a_c_cut_rf_read;
diff --git a/sys/dev/rtwn/usb/rtwn_usb_rx.c b/sys/dev/rtwn/usb/rtwn_usb_rx.c
--- a/sys/dev/rtwn/usb/rtwn_usb_rx.c
+++ b/sys/dev/rtwn/usb/rtwn_usb_rx.c
@@ -326,6 +326,27 @@
 		rtwn_handle_tx_report(sc, buf, len);
 		RTWN_NT_UNLOCK(sc);
 
+#ifdef IEEE80211_SUPPORT_SUPERG
+		/*
+		 * NB: this will executed only when 'report' bit is set.
+		 */
+		if (sc->sc_tx_n_active > 0 && --sc->sc_tx_n_active <= 1)
+			rtwn_cmd_sleepable(sc, NULL, 0, rtwn_ff_flush_all);
+#endif
+		break;
+	case RTWN_RX_TX_REPORT2:
+		if (sc->sc_ratectl != RTWN_RATECTL_NET80211) {
+			/* shouldn't happen */
+			device_printf(sc->sc_dev,
+			    "%s called while ratectl = %d!\n",
+			    __func__, sc->sc_ratectl);
+			break;
+		}
+
+		RTWN_NT_LOCK(sc);
+		rtwn_handle_tx_report2(sc, buf, len);
+		RTWN_NT_UNLOCK(sc);
+
 #ifdef IEEE80211_SUPPORT_SUPERG
 		/*
 		 * NB: this will executed only when 'report' bit is set.