3 files changed, 101 insertions, 82 deletions

diff --git a/drivers/net/ethernet/renesas/ravb_main.c b/drivers/net/ethernet/renesas/ravb_main.c
index ac0f093f647a..bc395294a32d 100644
--- a/drivers/net/ethernet/renesas/ravb_main.c
+++ b/drivers/net/ethernet/renesas/ravb_main.c
@@ -2763,6 +2763,7 @@ static const struct ravb_hw_info ravb_rzv2m_hw_info = {
 	.net_features = NETIF_F_RXCSUM,
 	.stats_len = ARRAY_SIZE(ravb_gstrings_stats),
 	.tccr_mask = TCCR_TSRQ0 | TCCR_TSRQ1 | TCCR_TSRQ2 | TCCR_TSRQ3,
+	.tx_max_frame_size = SZ_2K,
 	.rx_max_frame_size = SZ_2K,
 	.rx_buffer_size = SZ_2K +
 			  SKB_DATA_ALIGN(sizeof(struct skb_shared_info)),
diff --git a/drivers/net/ethernet/renesas/rswitch.c b/drivers/net/ethernet/renesas/rswitch.c
index 8d18dae4d8fb..9ac6e2aad18f 100644
--- a/drivers/net/ethernet/renesas/rswitch.c
+++ b/drivers/net/ethernet/renesas/rswitch.c
@@ -547,7 +547,6 @@ static int rswitch_gwca_ts_queue_alloc(struct rswitch_private *priv)
 	desc = &gq->ts_ring[gq->ring_size];
 	desc->desc.die_dt = DT_LINKFIX;
 	rswitch_desc_set_dptr(&desc->desc, gq->ring_dma);
-	INIT_LIST_HEAD(&priv->gwca.ts_info_list);
 
 	return 0;
 }
@@ -862,13 +861,10 @@ static void rswitch_tx_free(struct net_device *ndev)
 	struct rswitch_ext_desc *desc;
 	struct sk_buff *skb;
 
-	for (; rswitch_get_num_cur_queues(gq) > 0;
-	     gq->dirty = rswitch_next_queue_index(gq, false, 1)) {
-		desc = &gq->tx_ring[gq->dirty];
-		if ((desc->desc.die_dt & DT_MASK) != DT_FEMPTY)
-			break;
-
+	desc = &gq->tx_ring[gq->dirty];
+	while ((desc->desc.die_dt & DT_MASK) == DT_FEMPTY) {
 		dma_rmb();
+
 		skb = gq->skbs[gq->dirty];
 		if (skb) {
 			rdev->ndev->stats.tx_packets++;
@@ -879,7 +875,10 @@ static void rswitch_tx_free(struct net_device *ndev)
 			dev_kfree_skb_any(gq->skbs[gq->dirty]);
 			gq->skbs[gq->dirty] = NULL;
 		}
+
 		desc->desc.die_dt = DT_EEMPTY;
+		gq->dirty = rswitch_next_queue_index(gq, false, 1);
+		desc = &gq->tx_ring[gq->dirty];
 	}
 }
 
@@ -908,8 +907,10 @@ retry:
 
 	if (napi_complete_done(napi, budget - quota)) {
 		spin_lock_irqsave(&priv->lock, flags);
-		rswitch_enadis_data_irq(priv, rdev->tx_queue->index, true);
-		rswitch_enadis_data_irq(priv, rdev->rx_queue->index, true);
+		if (test_bit(rdev->port, priv->opened_ports)) {
+			rswitch_enadis_data_irq(priv, rdev->tx_queue->index, true);
+			rswitch_enadis_data_irq(priv, rdev->rx_queue->index, true);
+		}
 		spin_unlock_irqrestore(&priv->lock, flags);
 	}
 
@@ -1001,9 +1002,10 @@ static int rswitch_gwca_request_irqs(struct rswitch_private *priv)
 static void rswitch_ts(struct rswitch_private *priv)
 {
 	struct rswitch_gwca_queue *gq = &priv->gwca.ts_queue;
-	struct rswitch_gwca_ts_info *ts_info, *ts_info2;
 	struct skb_shared_hwtstamps shhwtstamps;
 	struct rswitch_ts_desc *desc;
+	struct rswitch_device *rdev;
+	struct sk_buff *ts_skb;
 	struct timespec64 ts;
 	unsigned int num;
 	u32 tag, port;
@@ -1013,23 +1015,28 @@ static void rswitch_ts(struct rswitch_private *priv)
 		dma_rmb();
 
 		port = TS_DESC_DPN(__le32_to_cpu(desc->desc.dptrl));
-		tag = TS_DESC_TSUN(__le32_to_cpu(desc->desc.dptrl));
-
-		list_for_each_entry_safe(ts_info, ts_info2, &priv->gwca.ts_info_list, list) {
-			if (!(ts_info->port == port && ts_info->tag == tag))
-				continue;
-
-			memset(&shhwtstamps, 0, sizeof(shhwtstamps));
-			ts.tv_sec = __le32_to_cpu(desc->ts_sec);
-			ts.tv_nsec = __le32_to_cpu(desc->ts_nsec & cpu_to_le32(0x3fffffff));
-			shhwtstamps.hwtstamp = timespec64_to_ktime(ts);
-			skb_tstamp_tx(ts_info->skb, &shhwtstamps);
-			dev_consume_skb_irq(ts_info->skb);
-			list_del(&ts_info->list);
-			kfree(ts_info);
-			break;
-		}
+		if (unlikely(port >= RSWITCH_NUM_PORTS))
+			goto next;
+		rdev = priv->rdev[port];
 
+		tag = TS_DESC_TSUN(__le32_to_cpu(desc->desc.dptrl));
+		if (unlikely(tag >= TS_TAGS_PER_PORT))
+			goto next;
+		ts_skb = xchg(&rdev->ts_skb[tag], NULL);
+		smp_mb(); /* order rdev->ts_skb[] read before bitmap update */
+		clear_bit(tag, rdev->ts_skb_used);
+
+		if (unlikely(!ts_skb))
+			goto next;
+
+		memset(&shhwtstamps, 0, sizeof(shhwtstamps));
+		ts.tv_sec = __le32_to_cpu(desc->ts_sec);
+		ts.tv_nsec = __le32_to_cpu(desc->ts_nsec & cpu_to_le32(0x3fffffff));
+		shhwtstamps.hwtstamp = timespec64_to_ktime(ts);
+		skb_tstamp_tx(ts_skb, &shhwtstamps);
+		dev_consume_skb_irq(ts_skb);
+
+next:
 		gq->cur = rswitch_next_queue_index(gq, true, 1);
 		desc = &gq->ts_ring[gq->cur];
 	}
@@ -1114,25 +1121,40 @@ static int rswitch_etha_wait_link_verification(struct rswitch_etha *etha)
 
 static void rswitch_rmac_setting(struct rswitch_etha *etha, const u8 *mac)
 {
-	u32 val;
+	u32 pis, lsc;
 
 	rswitch_etha_write_mac_address(etha, mac);
 
+	switch (etha->phy_interface) {
+	case PHY_INTERFACE_MODE_SGMII:
+		pis = MPIC_PIS_GMII;
+		break;
+	case PHY_INTERFACE_MODE_USXGMII:
+	case PHY_INTERFACE_MODE_5GBASER:
+		pis = MPIC_PIS_XGMII;
+		break;
+	default:
+		pis = FIELD_GET(MPIC_PIS, ioread32(etha->addr + MPIC));
+		break;
+	}
+
 	switch (etha->speed) {
 	case 100:
-		val = MPIC_LSC_100M;
+		lsc = MPIC_LSC_100M;
 		break;
 	case 1000:
-		val = MPIC_LSC_1G;
+		lsc = MPIC_LSC_1G;
 		break;
 	case 2500:
-		val = MPIC_LSC_2_5G;
+		lsc = MPIC_LSC_2_5G;
 		break;
 	default:
-		return;
+		lsc = FIELD_GET(MPIC_LSC, ioread32(etha->addr + MPIC));
+		break;
 	}
 
-	iowrite32(MPIC_PIS_GMII | val, etha->addr + MPIC);
+	rswitch_modify(etha->addr, MPIC, MPIC_PIS | MPIC_LSC,
+		       FIELD_PREP(MPIC_PIS, pis) | FIELD_PREP(MPIC_LSC, lsc));
 }
 
 static void rswitch_etha_enable_mii(struct rswitch_etha *etha)
@@ -1538,20 +1560,20 @@ static int rswitch_open(struct net_device *ndev)
 	struct rswitch_device *rdev = netdev_priv(ndev);
 	unsigned long flags;
 
-	phy_start(ndev->phydev);
+	if (bitmap_empty(rdev->priv->opened_ports, RSWITCH_NUM_PORTS))
+		iowrite32(GWCA_TS_IRQ_BIT, rdev->priv->addr + GWTSDIE);
 
 	napi_enable(&rdev->napi);
-	netif_start_queue(ndev);
 
 	spin_lock_irqsave(&rdev->priv->lock, flags);
+	bitmap_set(rdev->priv->opened_ports, rdev->port, 1);
 	rswitch_enadis_data_irq(rdev->priv, rdev->tx_queue->index, true);
 	rswitch_enadis_data_irq(rdev->priv, rdev->rx_queue->index, true);
 	spin_unlock_irqrestore(&rdev->priv->lock, flags);
 
-	if (bitmap_empty(rdev->priv->opened_ports, RSWITCH_NUM_PORTS))
-		iowrite32(GWCA_TS_IRQ_BIT, rdev->priv->addr + GWTSDIE);
+	phy_start(ndev->phydev);
 
-	bitmap_set(rdev->priv->opened_ports, rdev->port, 1);
+	netif_start_queue(ndev);
 
 	return 0;
 };
@@ -1559,31 +1581,34 @@ static int rswitch_open(struct net_device *ndev)
 static int rswitch_stop(struct net_device *ndev)
 {
 	struct rswitch_device *rdev = netdev_priv(ndev);
-	struct rswitch_gwca_ts_info *ts_info, *ts_info2;
+	struct sk_buff *ts_skb;
 	unsigned long flags;
+	unsigned int tag;
 
 	netif_tx_stop_all_queues(ndev);
-	bitmap_clear(rdev->priv->opened_ports, rdev->port, 1);
-
-	if (bitmap_empty(rdev->priv->opened_ports, RSWITCH_NUM_PORTS))
-		iowrite32(GWCA_TS_IRQ_BIT, rdev->priv->addr + GWTSDID);
 
-	list_for_each_entry_safe(ts_info, ts_info2, &rdev->priv->gwca.ts_info_list, list) {
-		if (ts_info->port != rdev->port)
-			continue;
-		dev_kfree_skb_irq(ts_info->skb);
-		list_del(&ts_info->list);
-		kfree(ts_info);
-	}
+	phy_stop(ndev->phydev);
 
 	spin_lock_irqsave(&rdev->priv->lock, flags);
 	rswitch_enadis_data_irq(rdev->priv, rdev->tx_queue->index, false);
 	rswitch_enadis_data_irq(rdev->priv, rdev->rx_queue->index, false);
+	bitmap_clear(rdev->priv->opened_ports, rdev->port, 1);
 	spin_unlock_irqrestore(&rdev->priv->lock, flags);
 
-	phy_stop(ndev->phydev);
 	napi_disable(&rdev->napi);
 
+	if (bitmap_empty(rdev->priv->opened_ports, RSWITCH_NUM_PORTS))
+		iowrite32(GWCA_TS_IRQ_BIT, rdev->priv->addr + GWTSDID);
+
+	for (tag = find_first_bit(rdev->ts_skb_used, TS_TAGS_PER_PORT);
+	     tag < TS_TAGS_PER_PORT;
+	     tag = find_next_bit(rdev->ts_skb_used, TS_TAGS_PER_PORT, tag + 1)) {
+		ts_skb = xchg(&rdev->ts_skb[tag], NULL);
+		clear_bit(tag, rdev->ts_skb_used);
+		if (ts_skb)
+			dev_kfree_skb(ts_skb);
+	}
+
 	return 0;
 };
 
@@ -1594,20 +1619,17 @@ static bool rswitch_ext_desc_set_info1(struct rswitch_device *rdev,
 	desc->info1 = cpu_to_le64(INFO1_DV(BIT(rdev->etha->index)) |
 				  INFO1_IPV(GWCA_IPV_NUM) | INFO1_FMT);
 	if (skb_shinfo(skb)->tx_flags & SKBTX_HW_TSTAMP) {
-		struct rswitch_gwca_ts_info *ts_info;
+		unsigned int tag;
 
-		ts_info = kzalloc(sizeof(*ts_info), GFP_ATOMIC);
-		if (!ts_info)
+		tag = find_first_zero_bit(rdev->ts_skb_used, TS_TAGS_PER_PORT);
+		if (tag == TS_TAGS_PER_PORT)
 			return false;
+		smp_mb(); /* order bitmap read before rdev->ts_skb[] write */
+		rdev->ts_skb[tag] = skb_get(skb);
+		set_bit(tag, rdev->ts_skb_used);
 
 		skb_shinfo(skb)->tx_flags |= SKBTX_IN_PROGRESS;
-		rdev->ts_tag++;
-		desc->info1 |= cpu_to_le64(INFO1_TSUN(rdev->ts_tag) | INFO1_TXC);
-
-		ts_info->skb = skb_get(skb);
-		ts_info->port = rdev->port;
-		ts_info->tag = rdev->ts_tag;
-		list_add_tail(&ts_info->list, &rdev->priv->gwca.ts_info_list);
+		desc->info1 |= cpu_to_le64(INFO1_TSUN(tag) | INFO1_TXC);
 
 		skb_tx_timestamp(skb);
 	}
@@ -1681,8 +1703,11 @@ static netdev_tx_t rswitch_start_xmit(struct sk_buff *skb, struct net_device *nd
 	if (dma_mapping_error(ndev->dev.parent, dma_addr_orig))
 		goto err_kfree;
 
-	gq->skbs[gq->cur] = skb;
-	gq->unmap_addrs[gq->cur] = dma_addr_orig;
+	/* Stored the skb at the last descriptor to avoid skb free before hardware completes send */
+	gq->skbs[(gq->cur + nr_desc - 1) % gq->ring_size] = skb;
+	gq->unmap_addrs[(gq->cur + nr_desc - 1) % gq->ring_size] = dma_addr_orig;
+
+	dma_wmb();
 
 	/* DT_FSTART should be set at last. So, this is reverse order. */
 	for (i = nr_desc; i-- > 0; ) {
@@ -1694,14 +1719,13 @@ static netdev_tx_t rswitch_start_xmit(struct sk_buff *skb, struct net_device *nd
 			goto err_unmap;
 	}
 
-	wmb();	/* gq->cur must be incremented after die_dt was set */
-
 	gq->cur = rswitch_next_queue_index(gq, true, nr_desc);
 	rswitch_modify(rdev->addr, GWTRC(gq->index), 0, BIT(gq->index % 32));
 
 	return ret;
 
 err_unmap:
+	gq->skbs[(gq->cur + nr_desc - 1) % gq->ring_size] = NULL;
 	dma_unmap_single(ndev->dev.parent, dma_addr_orig, skb->len, DMA_TO_DEVICE);
 
 err_kfree:
@@ -1889,7 +1913,6 @@ static int rswitch_device_alloc(struct rswitch_private *priv, unsigned int index
 	rdev->np_port = rswitch_get_port_node(rdev);
 	rdev->disabled = !rdev->np_port;
 	err = of_get_ethdev_address(rdev->np_port, ndev);
-	of_node_put(rdev->np_port);
 	if (err) {
 		if (is_valid_ether_addr(rdev->etha->mac_addr))
 			eth_hw_addr_set(ndev, rdev->etha->mac_addr);
@@ -1919,6 +1942,7 @@ out_txdmac:
 
 out_rxdmac:
 out_get_params:
+	of_node_put(rdev->np_port);
 	netif_napi_del(&rdev->napi);
 	free_netdev(ndev);
 
@@ -1932,6 +1956,7 @@ static void rswitch_device_free(struct rswitch_private *priv, unsigned int index
 
 	rswitch_txdmac_free(ndev);
 	rswitch_rxdmac_free(ndev);
+	of_node_put(rdev->np_port);
 	netif_napi_del(&rdev->napi);
 	free_netdev(ndev);
 }
diff --git a/drivers/net/ethernet/renesas/rswitch.h b/drivers/net/ethernet/renesas/rswitch.h
index 72e3ff596d31..d8d4ed7d7f8b 100644
--- a/drivers/net/ethernet/renesas/rswitch.h
+++ b/drivers/net/ethernet/renesas/rswitch.h
@@ -724,13 +724,13 @@ enum rswitch_etha_mode {
 
 #define EAVCC_VEM_SC_TAG	(0x3 << 16)
 
-#define MPIC_PIS_MII		0x00
-#define MPIC_PIS_GMII		0x02
-#define MPIC_PIS_XGMII		0x04
-#define MPIC_LSC_SHIFT		3
-#define MPIC_LSC_100M		(1 << MPIC_LSC_SHIFT)
-#define MPIC_LSC_1G		(2 << MPIC_LSC_SHIFT)
-#define MPIC_LSC_2_5G		(3 << MPIC_LSC_SHIFT)
+#define MPIC_PIS		GENMASK(2, 0)
+#define MPIC_PIS_GMII		2
+#define MPIC_PIS_XGMII		4
+#define MPIC_LSC		GENMASK(5, 3)
+#define MPIC_LSC_100M		1
+#define MPIC_LSC_1G		2
+#define MPIC_LSC_2_5G		3
 
 #define MDIO_READ_C45		0x03
 #define MDIO_WRITE_C45		0x01
@@ -972,14 +972,6 @@ struct rswitch_gwca_queue {
 	};
 };
 
-struct rswitch_gwca_ts_info {
-	struct sk_buff *skb;
-	struct list_head list;
-
-	int port;
-	u8 tag;
-};
-
 #define RSWITCH_NUM_IRQ_REGS	(RSWITCH_MAX_NUM_QUEUES / BITS_PER_TYPE(u32))
 struct rswitch_gwca {
 	unsigned int index;
@@ -989,7 +981,6 @@ struct rswitch_gwca {
 	struct rswitch_gwca_queue *queues;
 	int num_queues;
 	struct rswitch_gwca_queue ts_queue;
-	struct list_head ts_info_list;
 	DECLARE_BITMAP(used, RSWITCH_MAX_NUM_QUEUES);
 	u32 tx_irq_bits[RSWITCH_NUM_IRQ_REGS];
 	u32 rx_irq_bits[RSWITCH_NUM_IRQ_REGS];
@@ -997,6 +988,7 @@ struct rswitch_gwca {
 };
 
 #define NUM_QUEUES_PER_NDEV	2
+#define TS_TAGS_PER_PORT	256
 struct rswitch_device {
 	struct rswitch_private *priv;
 	struct net_device *ndev;
@@ -1004,7 +996,8 @@ struct rswitch_device {
 	void __iomem *addr;
 	struct rswitch_gwca_queue *tx_queue;
 	struct rswitch_gwca_queue *rx_queue;
-	u8 ts_tag;
+	struct sk_buff *ts_skb[TS_TAGS_PER_PORT];
+	DECLARE_BITMAP(ts_skb_used, TS_TAGS_PER_PORT);
 	bool disabled;
 
 	int port;