net: stmmac: use per-queue 64 bit statistics where necessary

Currently, there are two major issues with stmmac driver statistics
First of all, statistics in stmmac_extra_stats, stmmac_rxq_stats
and stmmac_txq_stats are 32 bit variables on 32 bit platforms. This
can cause some stats to overflow after several minutes of
high traffic, for example rx_pkt_n, tx_pkt_n and so on.

Secondly, if HW supports multiqueues, there are frequent cacheline
ping pongs on some driver statistic vars, for example, normal_irq_n,
tx_pkt_n and so on. What's more, frequent cacheline ping pongs on
normal_irq_n happens in ISR, this makes the situation worse.

To improve the driver, we convert those statistics to 64 bit, implement
ndo_get_stats64 and update .get_ethtool_stats implementation
accordingly. We also use per-queue statistics where necessary to remove
the cacheline ping pongs as much as possible to make multiqueue
operations faster. Those statistics which are not possible to overflow
and not frequently updated are kept as is.

Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Link: https://lore.kernel.org/r/20230717160630.1892-3-jszhang@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

1 files changed, 156 insertions, 44 deletions

diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
index a29a9619b48e..e7ca52f0d2f2 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_main.c
@@ -2434,6 +2434,8 @@ static bool stmmac_xdp_xmit_zc(struct stmmac_priv *priv, u32 queue, u32 budget)
 	struct dma_desc *tx_desc = NULL;
 	struct xdp_desc xdp_desc;
 	bool work_done = true;
+	u32 tx_set_ic_bit = 0;
+	unsigned long flags;
 
 	/* Avoids TX time-out as we are sharing with slow path */
 	txq_trans_cond_update(nq);
@@ -2494,7 +2496,7 @@ static bool stmmac_xdp_xmit_zc(struct stmmac_priv *priv, u32 queue, u32 budget)
 		if (set_ic) {
 			tx_q->tx_count_frames = 0;
 			stmmac_set_tx_ic(priv, tx_desc);
-			priv->xstats.tx_set_ic_bit++;
+			tx_set_ic_bit++;
 		}
 
 		stmmac_prepare_tx_desc(priv, tx_desc, 1, xdp_desc.len,
@@ -2506,6 +2508,9 @@ static bool stmmac_xdp_xmit_zc(struct stmmac_priv *priv, u32 queue, u32 budget)
 		tx_q->cur_tx = STMMAC_GET_ENTRY(tx_q->cur_tx, priv->dma_conf.dma_tx_size);
 		entry = tx_q->cur_tx;
 	}
+	flags = u64_stats_update_begin_irqsave(&tx_q->txq_stats.syncp);
+	tx_q->txq_stats.tx_set_ic_bit += tx_set_ic_bit;
+	u64_stats_update_end_irqrestore(&tx_q->txq_stats.syncp, flags);
 
 	if (tx_desc) {
 		stmmac_flush_tx_descriptors(priv, queue);
@@ -2547,11 +2552,11 @@ static int stmmac_tx_clean(struct stmmac_priv *priv, int budget, u32 queue)
 	struct stmmac_tx_queue *tx_q = &priv->dma_conf.tx_queue[queue];
 	unsigned int bytes_compl = 0, pkts_compl = 0;
 	unsigned int entry, xmits = 0, count = 0;
+	u32 tx_packets = 0, tx_errors = 0;
+	unsigned long flags;
 
 	__netif_tx_lock_bh(netdev_get_tx_queue(priv->dev, queue));
 
-	priv->xstats.tx_clean++;
-
 	tx_q->xsk_frames_done = 0;
 
 	entry = tx_q->dirty_tx;
@@ -2582,8 +2587,7 @@ static int stmmac_tx_clean(struct stmmac_priv *priv, int budget, u32 queue)
 		else
 			p = tx_q->dma_tx + entry;
 
-		status = stmmac_tx_status(priv, &priv->dev->stats,
-				&priv->xstats, p, priv->ioaddr);
+		status = stmmac_tx_status(priv,	&priv->xstats, p, priv->ioaddr);
 		/* Check if the descriptor is owned by the DMA */
 		if (unlikely(status & tx_dma_own))
 			break;
@@ -2599,13 +2603,11 @@ static int stmmac_tx_clean(struct stmmac_priv *priv, int budget, u32 queue)
 		if (likely(!(status & tx_not_ls))) {
 			/* ... verify the status error condition */
 			if (unlikely(status & tx_err)) {
-				priv->dev->stats.tx_errors++;
+				tx_errors++;
 				if (unlikely(status & tx_err_bump_tc))
 					stmmac_bump_dma_threshold(priv, queue);
 			} else {
-				priv->dev->stats.tx_packets++;
-				priv->xstats.tx_pkt_n++;
-				priv->xstats.txq_stats[queue].tx_pkt_n++;
+				tx_packets++;
 			}
 			if (skb)
 				stmmac_get_tx_hwtstamp(priv, p, skb);
@@ -2709,6 +2711,14 @@ static int stmmac_tx_clean(struct stmmac_priv *priv, int budget, u32 queue)
 			      STMMAC_COAL_TIMER(priv->tx_coal_timer[queue]),
 			      HRTIMER_MODE_REL);
 
+	flags = u64_stats_update_begin_irqsave(&tx_q->txq_stats.syncp);
+	tx_q->txq_stats.tx_packets += tx_packets;
+	tx_q->txq_stats.tx_pkt_n += tx_packets;
+	tx_q->txq_stats.tx_clean++;
+	u64_stats_update_end_irqrestore(&tx_q->txq_stats.syncp, flags);
+
+	priv->xstats.tx_errors += tx_errors;
+
 	__netif_tx_unlock_bh(netdev_get_tx_queue(priv->dev, queue));
 
 	/* Combine decisions from TX clean and XSK TX */
@@ -2736,7 +2746,7 @@ static void stmmac_tx_err(struct stmmac_priv *priv, u32 chan)
 			    tx_q->dma_tx_phy, chan);
 	stmmac_start_tx_dma(priv, chan);
 
-	priv->dev->stats.tx_errors++;
+	priv->xstats.tx_errors++;
 	netif_tx_wake_queue(netdev_get_tx_queue(priv->dev, chan));
 }
 
@@ -4109,6 +4119,7 @@ static netdev_tx_t stmmac_tso_xmit(struct sk_buff *skb, struct net_device *dev)
 	struct stmmac_tx_queue *tx_q;
 	bool has_vlan, set_ic;
 	u8 proto_hdr_len, hdr;
+	unsigned long flags;
 	u32 pay_len, mss;
 	dma_addr_t des;
 	int i;
@@ -4257,7 +4268,6 @@ static netdev_tx_t stmmac_tso_xmit(struct sk_buff *skb, struct net_device *dev)
 
 		tx_q->tx_count_frames = 0;
 		stmmac_set_tx_ic(priv, desc);
-		priv->xstats.tx_set_ic_bit++;
 	}
 
 	/* We've used all descriptors we need for this skb, however,
@@ -4273,9 +4283,13 @@ static netdev_tx_t stmmac_tso_xmit(struct sk_buff *skb, struct net_device *dev)
 		netif_tx_stop_queue(netdev_get_tx_queue(priv->dev, queue));
 	}
 
-	dev->stats.tx_bytes += skb->len;
-	priv->xstats.tx_tso_frames++;
-	priv->xstats.tx_tso_nfrags += nfrags;
+	flags = u64_stats_update_begin_irqsave(&tx_q->txq_stats.syncp);
+	tx_q->txq_stats.tx_bytes += skb->len;
+	tx_q->txq_stats.tx_tso_frames++;
+	tx_q->txq_stats.tx_tso_nfrags += nfrags;
+	if (set_ic)
+		tx_q->txq_stats.tx_set_ic_bit++;
+	u64_stats_update_end_irqrestore(&tx_q->txq_stats.syncp, flags);
 
 	if (priv->sarc_type)
 		stmmac_set_desc_sarc(priv, first, priv->sarc_type);
@@ -4325,7 +4339,7 @@ static netdev_tx_t stmmac_tso_xmit(struct sk_buff *skb, struct net_device *dev)
 dma_map_err:
 	dev_err(priv->device, "Tx dma map failed\n");
 	dev_kfree_skb(skb);
-	priv->dev->stats.tx_dropped++;
+	priv->xstats.tx_dropped++;
 	return NETDEV_TX_OK;
 }
 
@@ -4351,6 +4365,7 @@ static netdev_tx_t stmmac_xmit(struct sk_buff *skb, struct net_device *dev)
 	struct stmmac_tx_queue *tx_q;
 	bool has_vlan, set_ic;
 	int entry, first_tx;
+	unsigned long flags;
 	dma_addr_t des;
 
 	tx_q = &priv->dma_conf.tx_queue[queue];
@@ -4479,7 +4494,6 @@ static netdev_tx_t stmmac_xmit(struct sk_buff *skb, struct net_device *dev)
 
 		tx_q->tx_count_frames = 0;
 		stmmac_set_tx_ic(priv, desc);
-		priv->xstats.tx_set_ic_bit++;
 	}
 
 	/* We've used all descriptors we need for this skb, however,
@@ -4506,7 +4520,11 @@ static netdev_tx_t stmmac_xmit(struct sk_buff *skb, struct net_device *dev)
 		netif_tx_stop_queue(netdev_get_tx_queue(priv->dev, queue));
 	}
 
-	dev->stats.tx_bytes += skb->len;
+	flags = u64_stats_update_begin_irqsave(&tx_q->txq_stats.syncp);
+	tx_q->txq_stats.tx_bytes += skb->len;
+	if (set_ic)
+		tx_q->txq_stats.tx_set_ic_bit++;
+	u64_stats_update_end_irqrestore(&tx_q->txq_stats.syncp, flags);
 
 	if (priv->sarc_type)
 		stmmac_set_desc_sarc(priv, first, priv->sarc_type);
@@ -4568,7 +4586,7 @@ static netdev_tx_t stmmac_xmit(struct sk_buff *skb, struct net_device *dev)
 dma_map_err:
 	netdev_err(priv->dev, "Tx DMA map failed\n");
 	dev_kfree_skb(skb);
-	priv->dev->stats.tx_dropped++;
+	priv->xstats.tx_dropped++;
 	return NETDEV_TX_OK;
 }
 
@@ -4769,9 +4787,12 @@ static int stmmac_xdp_xmit_xdpf(struct stmmac_priv *priv, int queue,
 		set_ic = false;
 
 	if (set_ic) {
+		unsigned long flags;
 		tx_q->tx_count_frames = 0;
 		stmmac_set_tx_ic(priv, tx_desc);
-		priv->xstats.tx_set_ic_bit++;
+		flags = u64_stats_update_begin_irqsave(&tx_q->txq_stats.syncp);
+		tx_q->txq_stats.tx_set_ic_bit++;
+		u64_stats_update_end_irqrestore(&tx_q->txq_stats.syncp, flags);
 	}
 
 	stmmac_enable_dma_transmission(priv, priv->ioaddr);
@@ -4916,16 +4937,18 @@ static void stmmac_dispatch_skb_zc(struct stmmac_priv *priv, u32 queue,
 				   struct dma_desc *p, struct dma_desc *np,
 				   struct xdp_buff *xdp)
 {
+	struct stmmac_rx_queue *rx_q = &priv->dma_conf.rx_queue[queue];
 	struct stmmac_channel *ch = &priv->channel[queue];
 	unsigned int len = xdp->data_end - xdp->data;
 	enum pkt_hash_types hash_type;
 	int coe = priv->hw->rx_csum;
+	unsigned long flags;
 	struct sk_buff *skb;
 	u32 hash;
 
 	skb = stmmac_construct_skb_zc(ch, xdp);
 	if (!skb) {
-		priv->dev->stats.rx_dropped++;
+		priv->xstats.rx_dropped++;
 		return;
 	}
 
@@ -4944,8 +4967,10 @@ static void stmmac_dispatch_skb_zc(struct stmmac_priv *priv, u32 queue,
 	skb_record_rx_queue(skb, queue);
 	napi_gro_receive(&ch->rxtx_napi, skb);
 
-	priv->dev->stats.rx_packets++;
-	priv->dev->stats.rx_bytes += len;
+	flags = u64_stats_update_begin_irqsave(&rx_q->rxq_stats.syncp);
+	rx_q->rxq_stats.rx_pkt_n++;
+	rx_q->rxq_stats.rx_bytes += len;
+	u64_stats_update_end_irqrestore(&rx_q->rxq_stats.syncp, flags);
 }
 
 static bool stmmac_rx_refill_zc(struct stmmac_priv *priv, u32 queue, u32 budget)
@@ -5022,9 +5047,11 @@ static int stmmac_rx_zc(struct stmmac_priv *priv, int limit, u32 queue)
 	unsigned int count = 0, error = 0, len = 0;
 	int dirty = stmmac_rx_dirty(priv, queue);
 	unsigned int next_entry = rx_q->cur_rx;
+	u32 rx_errors = 0, rx_dropped = 0;
 	unsigned int desc_size;
 	struct bpf_prog *prog;
 	bool failure = false;
+	unsigned long flags;
 	int xdp_status = 0;
 	int status = 0;
 
@@ -5080,8 +5107,7 @@ read_again:
 			p = rx_q->dma_rx + entry;
 
 		/* read the status of the incoming frame */
-		status = stmmac_rx_status(priv, &priv->dev->stats,
-					  &priv->xstats, p);
+		status = stmmac_rx_status(priv, &priv->xstats, p);
 		/* check if managed by the DMA otherwise go ahead */
 		if (unlikely(status & dma_own))
 			break;
@@ -5103,8 +5129,7 @@ read_again:
 			break;
 
 		if (priv->extend_desc)
-			stmmac_rx_extended_status(priv, &priv->dev->stats,
-						  &priv->xstats,
+			stmmac_rx_extended_status(priv, &priv->xstats,
 						  rx_q->dma_erx + entry);
 		if (unlikely(status == discard_frame)) {
 			xsk_buff_free(buf->xdp);
@@ -5112,7 +5137,7 @@ read_again:
 			dirty++;
 			error = 1;
 			if (!priv->hwts_rx_en)
-				priv->dev->stats.rx_errors++;
+				rx_errors++;
 		}
 
 		if (unlikely(error && (status & rx_not_ls)))
@@ -5160,7 +5185,7 @@ read_again:
 			break;
 		case STMMAC_XDP_CONSUMED:
 			xsk_buff_free(buf->xdp);
-			priv->dev->stats.rx_dropped++;
+			rx_dropped++;
 			break;
 		case STMMAC_XDP_TX:
 		case STMMAC_XDP_REDIRECT:
@@ -5181,8 +5206,12 @@ read_again:
 
 	stmmac_finalize_xdp_rx(priv, xdp_status);
 
-	priv->xstats.rx_pkt_n += count;
-	priv->xstats.rxq_stats[queue].rx_pkt_n += count;
+	flags = u64_stats_update_begin_irqsave(&rx_q->rxq_stats.syncp);
+	rx_q->rxq_stats.rx_pkt_n += count;
+	u64_stats_update_end_irqrestore(&rx_q->rxq_stats.syncp, flags);
+
+	priv->xstats.rx_dropped += rx_dropped;
+	priv->xstats.rx_errors += rx_errors;
 
 	if (xsk_uses_need_wakeup(rx_q->xsk_pool)) {
 		if (failure || stmmac_rx_dirty(priv, queue) > 0)
@@ -5206,6 +5235,7 @@ read_again:
  */
 static int stmmac_rx(struct stmmac_priv *priv, int limit, u32 queue)
 {
+	u32 rx_errors = 0, rx_dropped = 0, rx_bytes = 0, rx_packets = 0;
 	struct stmmac_rx_queue *rx_q = &priv->dma_conf.rx_queue[queue];
 	struct stmmac_channel *ch = &priv->channel[queue];
 	unsigned int count = 0, error = 0, len = 0;
@@ -5215,6 +5245,7 @@ static int stmmac_rx(struct stmmac_priv *priv, int limit, u32 queue)
 	unsigned int desc_size;
 	struct sk_buff *skb = NULL;
 	struct stmmac_xdp_buff ctx;
+	unsigned long flags;
 	int xdp_status = 0;
 	int buf_sz;
 
@@ -5270,8 +5301,7 @@ read_again:
 			p = rx_q->dma_rx + entry;
 
 		/* read the status of the incoming frame */
-		status = stmmac_rx_status(priv, &priv->dev->stats,
-				&priv->xstats, p);
+		status = stmmac_rx_status(priv, &priv->xstats, p);
 		/* check if managed by the DMA otherwise go ahead */
 		if (unlikely(status & dma_own))
 			break;
@@ -5288,14 +5318,13 @@ read_again:
 		prefetch(np);
 
 		if (priv->extend_desc)
-			stmmac_rx_extended_status(priv, &priv->dev->stats,
-					&priv->xstats, rx_q->dma_erx + entry);
+			stmmac_rx_extended_status(priv, &priv->xstats, rx_q->dma_erx + entry);
 		if (unlikely(status == discard_frame)) {
 			page_pool_recycle_direct(rx_q->page_pool, buf->page);
 			buf->page = NULL;
 			error = 1;
 			if (!priv->hwts_rx_en)
-				priv->dev->stats.rx_errors++;
+				rx_errors++;
 		}
 
 		if (unlikely(error && (status & rx_not_ls)))
@@ -5363,7 +5392,7 @@ read_again:
 							   virt_to_head_page(ctx.xdp.data),
 							   sync_len, true);
 					buf->page = NULL;
-					priv->dev->stats.rx_dropped++;
+					rx_dropped++;
 
 					/* Clear skb as it was set as
 					 * status by XDP program.
@@ -5392,7 +5421,7 @@ read_again:
 
 			skb = napi_alloc_skb(&ch->rx_napi, buf1_len);
 			if (!skb) {
-				priv->dev->stats.rx_dropped++;
+				rx_dropped++;
 				count++;
 				goto drain_data;
 			}
@@ -5452,8 +5481,8 @@ drain_data:
 		napi_gro_receive(&ch->rx_napi, skb);
 		skb = NULL;
 
-		priv->dev->stats.rx_packets++;
-		priv->dev->stats.rx_bytes += len;
+		rx_packets++;
+		rx_bytes += len;
 		count++;
 	}
 
@@ -5468,8 +5497,14 @@ drain_data:
 
 	stmmac_rx_refill(priv, queue);
 
-	priv->xstats.rx_pkt_n += count;
-	priv->xstats.rxq_stats[queue].rx_pkt_n += count;
+	flags = u64_stats_update_begin_irqsave(&rx_q->rxq_stats.syncp);
+	rx_q->rxq_stats.rx_packets += rx_packets;
+	rx_q->rxq_stats.rx_bytes += rx_bytes;
+	rx_q->rxq_stats.rx_pkt_n += count;
+	u64_stats_update_end_irqrestore(&rx_q->rxq_stats.syncp, flags);
+
+	priv->xstats.rx_dropped += rx_dropped;
+	priv->xstats.rx_errors += rx_errors;
 
 	return count;
 }
@@ -5479,10 +5514,15 @@ static int stmmac_napi_poll_rx(struct napi_struct *napi, int budget)
 	struct stmmac_channel *ch =
 		container_of(napi, struct stmmac_channel, rx_napi);
 	struct stmmac_priv *priv = ch->priv_data;
+	struct stmmac_rx_queue *rx_q;
 	u32 chan = ch->index;
+	unsigned long flags;
 	int work_done;
 
-	priv->xstats.napi_poll++;
+	rx_q = &priv->dma_conf.rx_queue[chan];
+	flags = u64_stats_update_begin_irqsave(&rx_q->rxq_stats.syncp);
+	rx_q->rxq_stats.napi_poll++;
+	u64_stats_update_end_irqrestore(&rx_q->rxq_stats.syncp, flags);
 
 	work_done = stmmac_rx(priv, budget, chan);
 	if (work_done < budget && napi_complete_done(napi, work_done)) {
@@ -5501,10 +5541,15 @@ static int stmmac_napi_poll_tx(struct napi_struct *napi, int budget)
 	struct stmmac_channel *ch =
 		container_of(napi, struct stmmac_channel, tx_napi);
 	struct stmmac_priv *priv = ch->priv_data;
+	struct stmmac_tx_queue *tx_q;
 	u32 chan = ch->index;
+	unsigned long flags;
 	int work_done;
 
-	priv->xstats.napi_poll++;
+	tx_q = &priv->dma_conf.tx_queue[chan];
+	flags = u64_stats_update_begin_irqsave(&tx_q->txq_stats.syncp);
+	tx_q->txq_stats.napi_poll++;
+	u64_stats_update_end_irqrestore(&tx_q->txq_stats.syncp, flags);
 
 	work_done = stmmac_tx_clean(priv, budget, chan);
 	work_done = min(work_done, budget);
@@ -5526,9 +5571,20 @@ static int stmmac_napi_poll_rxtx(struct napi_struct *napi, int budget)
 		container_of(napi, struct stmmac_channel, rxtx_napi);
 	struct stmmac_priv *priv = ch->priv_data;
 	int rx_done, tx_done, rxtx_done;
+	struct stmmac_rx_queue *rx_q;
+	struct stmmac_tx_queue *tx_q;
 	u32 chan = ch->index;
+	unsigned long flags;
+
+	rx_q = &priv->dma_conf.rx_queue[chan];
+	flags = u64_stats_update_begin_irqsave(&rx_q->rxq_stats.syncp);
+	rx_q->rxq_stats.napi_poll++;
+	u64_stats_update_end_irqrestore(&rx_q->rxq_stats.syncp, flags);
 
-	priv->xstats.napi_poll++;
+	tx_q = &priv->dma_conf.tx_queue[chan];
+	flags = u64_stats_update_begin_irqsave(&tx_q->txq_stats.syncp);
+	tx_q->txq_stats.napi_poll++;
+	u64_stats_update_end_irqrestore(&tx_q->txq_stats.syncp, flags);
 
 	tx_done = stmmac_tx_clean(priv, budget, chan);
 	tx_done = min(tx_done, budget);
@@ -6788,6 +6844,56 @@ int stmmac_xsk_wakeup(struct net_device *dev, u32 queue, u32 flags)
 	return 0;
 }
 
+static void stmmac_get_stats64(struct net_device *dev, struct rtnl_link_stats64 *stats)
+{
+	struct stmmac_priv *priv = netdev_priv(dev);
+	u32 tx_cnt = priv->plat->tx_queues_to_use;
+	u32 rx_cnt = priv->plat->rx_queues_to_use;
+	unsigned int start;
+	int q;
+
+	for (q = 0; q < tx_cnt; q++) {
+		struct stmmac_txq_stats *txq_stats = &priv->dma_conf.tx_queue[q].txq_stats;
+		u64 tx_packets;
+		u64 tx_bytes;
+
+		do {
+			start = u64_stats_fetch_begin(&txq_stats->syncp);
+			tx_packets = txq_stats->tx_packets;
+			tx_bytes   = txq_stats->tx_bytes;
+		} while (u64_stats_fetch_retry(&txq_stats->syncp, start));
+
+		stats->tx_packets += tx_packets;
+		stats->tx_bytes += tx_bytes;
+	}
+
+	for (q = 0; q < rx_cnt; q++) {
+		struct stmmac_rxq_stats *rxq_stats = &priv->dma_conf.rx_queue[q].rxq_stats;
+		u64 rx_packets;
+		u64 rx_bytes;
+
+		do {
+			start = u64_stats_fetch_begin(&rxq_stats->syncp);
+			rx_packets = rxq_stats->rx_packets;
+			rx_bytes   = rxq_stats->rx_bytes;
+		} while (u64_stats_fetch_retry(&rxq_stats->syncp, start));
+
+		stats->rx_packets += rx_packets;
+		stats->rx_bytes += rx_bytes;
+	}
+
+	stats->rx_dropped = priv->xstats.rx_dropped;
+	stats->rx_errors = priv->xstats.rx_errors;
+	stats->tx_dropped = priv->xstats.tx_dropped;
+	stats->tx_errors = priv->xstats.tx_errors;
+	stats->tx_carrier_errors = priv->xstats.tx_losscarrier + priv->xstats.tx_carrier;
+	stats->collisions = priv->xstats.tx_collision + priv->xstats.rx_collision;
+	stats->rx_length_errors = priv->xstats.rx_length;
+	stats->rx_crc_errors = priv->xstats.rx_crc_errors;
+	stats->rx_over_errors = priv->xstats.rx_overflow_cntr;
+	stats->rx_missed_errors = priv->xstats.rx_missed_cntr;
+}
+
 static const struct net_device_ops stmmac_netdev_ops = {
 	.ndo_open = stmmac_open,
 	.ndo_start_xmit = stmmac_xmit,
@@ -6798,6 +6904,7 @@ static const struct net_device_ops stmmac_netdev_ops = {
 	.ndo_set_rx_mode = stmmac_set_rx_mode,
 	.ndo_tx_timeout = stmmac_tx_timeout,
 	.ndo_eth_ioctl = stmmac_ioctl,
+	.ndo_get_stats64 = stmmac_get_stats64,
 	.ndo_setup_tc = stmmac_setup_tc,
 	.ndo_select_queue = stmmac_select_queue,
 #ifdef CONFIG_NET_POLL_CONTROLLER
@@ -7161,6 +7268,11 @@ int stmmac_dvr_probe(struct device *device,
 	priv->device = device;
 	priv->dev = ndev;
 
+	for (i = 0; i < MTL_MAX_RX_QUEUES; i++)
+		u64_stats_init(&priv->dma_conf.rx_queue[i].rxq_stats.syncp);
+	for (i = 0; i < MTL_MAX_TX_QUEUES; i++)
+		u64_stats_init(&priv->dma_conf.tx_queue[i].txq_stats.syncp);
+
 	stmmac_set_ethtool_ops(ndev);
 	priv->pause = pause;
 	priv->plat = plat_dat;