sfc: per-queue stats

index 6ded44b86052..a8baeacd83c0 100644
@@ -75,7 +75,6 @@ static const struct efx_sw_stat_desc efx_sw_stat_desc[] = {
 	EFX_ETHTOOL_UINT_TXQ_STAT(pio_packets),
 	EFX_ETHTOOL_UINT_TXQ_STAT(cb_packets),
 	EFX_ETHTOOL_ATOMIC_NIC_ERROR_STAT(rx_reset),
-	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_tobe_disc),
 	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_ip_hdr_chksum_err),
 	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_tcp_udp_chksum_err),
 	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_inner_ip_hdr_chksum_err),

@@ -83,7 +82,6 @@ static const struct efx_sw_stat_desc efx_sw_stat_desc[] = {
 	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_outer_ip_hdr_chksum_err),
 	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_outer_tcp_udp_chksum_err),
 	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_eth_crc_err),
-	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_mcast_mismatch),
 	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_frm_trunc),
 	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_merge_events),
 	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_merge_packets),

index b85c51cbe7f9..4d904e1404d4 100644
@@ -451,10 +451,8 @@ enum efx_sync_events_state {
  * @filter_work: Work item for efx_filter_rfs_expire()
  * @rps_flow_id: Flow IDs of filters allocated for accelerated RFS,
  *      indexed by filter ID
- * @n_rx_tobe_disc: Count of RX_TOBE_DISC errors
  * @n_rx_ip_hdr_chksum_err: Count of RX IP header checksum errors
  * @n_rx_tcp_udp_chksum_err: Count of RX TCP and UDP checksum errors
- * @n_rx_mcast_mismatch: Count of unmatched multicast frames
  * @n_rx_frm_trunc: Count of RX_FRM_TRUNC errors
  * @n_rx_overlength: Count of RX_OVERLENGTH errors
  * @n_skbuff_leaks: Count of skbuffs leaked due to RX overrun

@@ -511,7 +509,6 @@ struct efx_channel {
 	u32 *rps_flow_id;
 #endif
 
-	unsigned int n_rx_tobe_disc;
 	unsigned int n_rx_ip_hdr_chksum_err;
 	unsigned int n_rx_tcp_udp_chksum_err;
 	unsigned int n_rx_outer_ip_hdr_chksum_err;

@@ -519,7 +516,6 @@ struct efx_channel {
 	unsigned int n_rx_inner_ip_hdr_chksum_err;
 	unsigned int n_rx_inner_tcp_udp_chksum_err;
 	unsigned int n_rx_eth_crc_err;
-	unsigned int n_rx_mcast_mismatch;
 	unsigned int n_rx_frm_trunc;
 	unsigned int n_rx_overlength;
 	unsigned int n_skbuff_leaks;

index 83d9db71d7d7..992151775cb8 100644
@@ -134,6 +134,8 @@ void __ef100_rx_packet(struct efx_channel *channel)
 		goto free_rx_buffer;
 	}
 
+	++rx_queue->rx_packets;
+
 	efx_rx_packet_gro(channel, rx_buf, channel->rx_pkt_n_frags, eh, csum);
 	goto out;
 

@@ -149,8 +151,6 @@ static void ef100_rx_packet(struct efx_rx_queue *rx_queue, unsigned int index)
 	struct efx_channel *channel = efx_rx_queue_channel(rx_queue);
 	struct efx_nic *efx = rx_queue->efx;
 
-	++rx_queue->rx_packets;
-
 	netif_vdbg(efx, rx_status, efx->net_dev,
 		   "RX queue %d received id %x\n",
 		   efx_rx_queue_index(rx_queue), index);

93 lines not on this page · raw message

index c9e17a8208a9..834d51812e2b 100644
@@ -1209,6 +1209,8 @@ static int efx_process_channel(struct efx_channel *channel, int budget)
 						  tx_queue->pkts_compl,
 						  tx_queue->bytes_compl);
 		}
+		tx_queue->complete_packets += tx_queue->pkts_compl;
+		tx_queue->complete_bytes += tx_queue->bytes_compl;
 	}
 
 	/* Receive any packets we queued up */

index 4d904e1404d4..83c33c1ca120 100644
@@ -193,6 +193,10 @@ struct efx_tx_buffer {
  * @initialised: Has hardware queue been initialised?
  * @timestamping: Is timestamping enabled for this channel?
  * @xdp_tx: Is this an XDP tx queue?
+ * @old_complete_packets: Value of @complete_packets as of last
+ *	efx_init_tx_queue()
+ * @old_complete_bytes: Value of @complete_bytes as of last
+ *	efx_init_tx_queue()
  * @read_count: Current read pointer.
  *	This is the number of buffers that have been removed from both rings.
  * @old_write_count: The value of @write_count when last checked.

@@ -202,6 +206,16 @@ struct efx_tx_buffer {
  *	avoid cache-line ping-pong between the xmit path and the
  *	completion path.
  * @merge_events: Number of TX merged completion events
+ * @bytes_compl: Number of bytes completed during this NAPI poll
+ *	(efx_process_channel()).  For BQL.
+ * @pkts_compl: Number of packets completed during this NAPI poll.
+ * @complete_packets: Number of packets completed since this struct was
+ *	created.  Only counts SKB packets, not XDP TX (it accumulates
+ *	the same values that are reported to BQL).
+ * @complete_bytes: Number of bytes completed since this struct was
+ *	created.  For TSO, counts the superframe size, not the sizes of
+ *	generated frames on the wire (i.e. the headers are only counted
+ *	once)
  * @completed_timestamp_major: Top part of the most recent tx timestamp.
  * @completed_timestamp_minor: Low part of the most recent tx timestamp.
  * @insert_count: Current insert pointer

@@ -232,6 +246,7 @@ struct efx_tx_buffer {
  * @xmit_pending: Are any packets waiting to be pushed to the NIC
  * @cb_packets: Number of times the TX copybreak feature has been used
  * @notify_count: Count of notified descriptors to the NIC
+ * @tx_packets: Number of packets sent since this struct was created
  * @empty_read_count: If the completion path has seen the queue as empty
  *	and the transmission path has not yet checked this, the value of
  *	@read_count bitwise-added to %EFX_EMPTY_COUNT_VALID; otherwise 0.

@@ -255,6 +270,8 @@ struct efx_tx_queue {
 	bool initialised;
 	bool timestamping;
 	bool xdp_tx;
+	unsigned long old_complete_packets;
+	unsigned long old_complete_bytes;
 
 	/* Members used mainly on the completion path */
 	unsigned int read_count ____cacheline_aligned_in_smp;

@@ -262,6 +279,8 @@ struct efx_tx_queue {
 	unsigned int merge_events;
 	unsigned int bytes_compl;
 	unsigned int pkts_compl;
+	unsigned long complete_packets;
+	unsigned long complete_bytes;
 	u32 completed_timestamp_major;
 	u32 completed_timestamp_minor;
 

@@ -370,6 +389,8 @@ struct efx_rx_page_state {
  * @recycle_count: RX buffer recycle counter.
  * @slow_fill: Timer used to defer efx_nic_generate_fill_event().
  * @grant_work: workitem used to grant credits to the MAE if @grant_credits
+ * @rx_packets: Number of packets received since this struct was created
+ * @old_rx_packets: Value of @rx_packets as of last efx_init_rx_queue()
  * @xdp_rxq_info: XDP specific RX queue information.
  * @xdp_rxq_info_valid: Is xdp_rxq_info valid data?.
  */

@@ -406,6 +427,7 @@ struct efx_rx_queue {
 	struct work_struct grant_work;
 	/* Statistics to supplement MAC stats */
 	unsigned long rx_packets;
+	unsigned long old_rx_packets;
 	struct xdp_rxq_info xdp_rxq_info;
 	bool xdp_rxq_info_valid;
 };

index f77a2d3ef37e..f07495582125 100644
@@ -125,8 +125,6 @@ void efx_rx_packet(struct efx_rx_queue *rx_queue, unsigned int index,
 	struct efx_channel *channel = efx_rx_queue_channel(rx_queue);
 	struct efx_rx_buffer *rx_buf;
 
-	rx_queue->rx_packets++;
-
 	rx_buf = efx_rx_buffer(rx_queue, index);
 	rx_buf->flags |= flags;
 

@@ -394,6 +392,8 @@ void __efx_rx_packet(struct efx_channel *channel)
 		goto out;
 	}
 
+	rx_queue->rx_packets++;
+
 	if (!efx_do_xdp(efx, channel, rx_buf, &eh))
 		goto out;
 

index 0b7dc75c40f9..bdb4325a7c2c 100644
@@ -241,6 +241,8 @@ void efx_init_rx_queue(struct efx_rx_queue *rx_queue)
 	rx_queue->page_recycle_failed = 0;
 	rx_queue->page_recycle_full = 0;
 
+	rx_queue->old_rx_packets = rx_queue->rx_packets;
+
 	/* Initialise limit fields */
 	max_fill = efx->rxq_entries - EFX_RXD_HEAD_ROOM;
 	max_trigger =

index 2adb132b2f7e..6d47927e1c2c 100644
@@ -86,6 +86,9 @@ void efx_init_tx_queue(struct efx_tx_queue *tx_queue)
 	tx_queue->completed_timestamp_major = 0;
 	tx_queue->completed_timestamp_minor = 0;
 
+	tx_queue->old_complete_packets = tx_queue->complete_packets;
+	tx_queue->old_complete_bytes = tx_queue->complete_bytes;
+
 	tx_queue->xdp_tx = efx_channel_is_xdp_tx(tx_queue->channel);
 	tx_queue->tso_version = 0;
 

index a8baeacd83c0..ae32e08540fa 100644
@@ -83,6 +83,7 @@ static const struct efx_sw_stat_desc efx_sw_stat_desc[] = {
 	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_outer_tcp_udp_chksum_err),
 	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_eth_crc_err),
 	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_frm_trunc),
+	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_overlength),
 	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_merge_events),
 	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_merge_packets),
 	EFX_ETHTOOL_UINT_CHANNEL_STAT(rx_xdp_drops),

index 21dc4b885542..ea1e0e8ecbdd 100644
@@ -690,6 +690,9 @@ static void efx_get_base_stats(struct net_device *net_dev,
 				tx->packets += tx_queue->old_complete_packets;
 				tx->bytes += tx_queue->old_complete_bytes;
 			}
+			/* Include XDP TX in device-wide stats */
+			tx->packets += tx_queue->complete_xdp_packets;
+			tx->bytes += tx_queue->complete_xdp_bytes;
 		}
 	}
 }

index 83c33c1ca120..aba106d03d41 100644
@@ -216,6 +216,10 @@ struct efx_tx_buffer {
  *	created.  For TSO, counts the superframe size, not the sizes of
  *	generated frames on the wire (i.e. the headers are only counted
  *	once)
+ * @complete_xdp_packets: Number of XDP TX packets completed since this
+ *	struct was created.
+ * @complete_xdp_bytes: Number of XDP TX bytes completed since this
+ *	struct was created.
  * @completed_timestamp_major: Top part of the most recent tx timestamp.
  * @completed_timestamp_minor: Low part of the most recent tx timestamp.
  * @insert_count: Current insert pointer

@@ -281,6 +285,8 @@ struct efx_tx_queue {
 	unsigned int pkts_compl;
 	unsigned long complete_packets;
 	unsigned long complete_bytes;
+	unsigned long complete_xdp_packets;
+	unsigned long complete_xdp_bytes;
 	u32 completed_timestamp_major;
 	u32 completed_timestamp_minor;
 

index fe2d476028e7..822ec6564b2d 100644
@@ -553,6 +553,7 @@ netdev_tx_t efx_hard_start_xmit(struct sk_buff *skb,
 
 void efx_xmit_done_single(struct efx_tx_queue *tx_queue)
 {
+	unsigned int xdp_pkts_compl = 0, xdp_bytes_compl = 0;
 	unsigned int pkts_compl = 0, bytes_compl = 0;
 	unsigned int efv_pkts_compl = 0;
 	unsigned int read_ptr;

@@ -577,7 +578,8 @@ void efx_xmit_done_single(struct efx_tx_queue *tx_queue)
 		if (buffer->flags & EFX_TX_BUF_SKB)
 			finished = true;
 		efx_dequeue_buffer(tx_queue, buffer, &pkts_compl, &bytes_compl,
-				   &efv_pkts_compl);
+				   &efv_pkts_compl, &xdp_pkts_compl,
+				   &xdp_bytes_compl);
 
 		++tx_queue->read_count;
 		read_ptr = tx_queue->read_count & tx_queue->ptr_mask;

@@ -585,6 +587,8 @@ void efx_xmit_done_single(struct efx_tx_queue *tx_queue)
 
 	tx_queue->pkts_compl += pkts_compl;
 	tx_queue->bytes_compl += bytes_compl;
+	tx_queue->complete_xdp_packets += xdp_pkts_compl;
+	tx_queue->complete_xdp_bytes += xdp_bytes_compl;
 
 	EFX_WARN_ON_PARANOID(pkts_compl + efv_pkts_compl != 1);
 

index 6d47927e1c2c..2013a609f9be 100644
@@ -112,12 +112,14 @@ void efx_fini_tx_queue(struct efx_tx_queue *tx_queue)
 
 	/* Free any buffers left in the ring */
 	while (tx_queue->read_count != tx_queue->write_count) {
+		unsigned int xdp_pkts_compl = 0, xdp_bytes_compl = 0;
 		unsigned int pkts_compl = 0, bytes_compl = 0;
 		unsigned int efv_pkts_compl = 0;
 
 		buffer = &tx_queue->buffer[tx_queue->read_count & tx_queue->ptr_mask];
 		efx_dequeue_buffer(tx_queue, buffer, &pkts_compl, &bytes_compl,
-				   &efv_pkts_compl);
+				   &efv_pkts_compl, &xdp_pkts_compl,
+				   &xdp_bytes_compl);
 
 		++tx_queue->read_count;
 	}

@@ -153,7 +155,9 @@ void efx_dequeue_buffer(struct efx_tx_queue *tx_queue,
 			struct efx_tx_buffer *buffer,
 			unsigned int *pkts_compl,
 			unsigned int *bytes_compl,
-			unsigned int *efv_pkts_compl)
+			unsigned int *efv_pkts_compl,
+			unsigned int *xdp_pkts,
+			unsigned int *xdp_bytes)
 {
 	if (buffer->unmap_len) {
 		struct device *dma_dev = &tx_queue->efx->pci_dev->dev;

@@ -198,6 +202,10 @@ void efx_dequeue_buffer(struct efx_tx_queue *tx_queue,
 			   tx_queue->queue, tx_queue->read_count);
 	} else if (buffer->flags & EFX_TX_BUF_XDP) {
 		xdp_return_frame_rx_napi(buffer->xdpf);
+		if (xdp_pkts)
+			(*xdp_pkts)++;
+		if (xdp_bytes)
+			(*xdp_bytes) += buffer->xdpf->len;
 	}
 
 	buffer->len = 0;

@@ -213,7 +221,9 @@ static void efx_dequeue_buffers(struct efx_tx_queue *tx_queue,
 				unsigned int index,
 				unsigned int *pkts_compl,
 				unsigned int *bytes_compl,
-				unsigned int *efv_pkts_compl)
+				unsigned int *efv_pkts_compl,
+				unsigned int *xdp_pkts,
+				unsigned int *xdp_bytes)
 {
 	struct efx_nic *efx = tx_queue->efx;
 	unsigned int stop_index, read_ptr;

@@ -233,7 +243,7 @@ static void efx_dequeue_buffers(struct efx_tx_queue *tx_queue,
 		}
 
 		efx_dequeue_buffer(tx_queue, buffer, pkts_compl, bytes_compl,
-				   efv_pkts_compl);
+				   efv_pkts_compl, xdp_pkts, xdp_bytes);
 
 		++tx_queue->read_count;
 		read_ptr = tx_queue->read_count & tx_queue->ptr_mask;

@@ -256,15 +266,18 @@ void efx_xmit_done_check_empty(struct efx_tx_queue *tx_queue)
 int efx_xmit_done(struct efx_tx_queue *tx_queue, unsigned int index)
 {
 	unsigned int fill_level, pkts_compl = 0, bytes_compl = 0;
+	unsigned int xdp_pkts_compl = 0, xdp_bytes_compl = 0;
 	unsigned int efv_pkts_compl = 0;
 	struct efx_nic *efx = tx_queue->efx;
 
 	EFX_WARN_ON_ONCE_PARANOID(index > tx_queue->ptr_mask);
 
 	efx_dequeue_buffers(tx_queue, index, &pkts_compl, &bytes_compl,
-			    &efv_pkts_compl);
+			    &efv_pkts_compl, &xdp_pkts_compl, &xdp_bytes_compl);
 	tx_queue->pkts_compl += pkts_compl;
 	tx_queue->bytes_compl += bytes_compl;
+	tx_queue->complete_xdp_packets += xdp_pkts_compl;
+	tx_queue->complete_xdp_bytes += xdp_bytes_compl;
 
 	if (pkts_compl + efv_pkts_compl > 1)
 		++tx_queue->merge_events;

@@ -293,6 +306,8 @@ int efx_xmit_done(struct efx_tx_queue *tx_queue, unsigned int index)
 void efx_enqueue_unwind(struct efx_tx_queue *tx_queue,
 			unsigned int insert_count)
 {
+	unsigned int xdp_bytes_compl = 0;
+	unsigned int xdp_pkts_compl = 0;
 	unsigned int efv_pkts_compl = 0;
 	struct efx_tx_buffer *buffer;
 	unsigned int bytes_compl = 0;

@@ -303,7 +318,8 @@ void efx_enqueue_unwind(struct efx_tx_queue *tx_queue,
 		--tx_queue->insert_count;
 		buffer = __efx_tx_queue_get_insert_buffer(tx_queue);
 		efx_dequeue_buffer(tx_queue, buffer, &pkts_compl, &bytes_compl,
-				   &efv_pkts_compl);
+				   &efv_pkts_compl, &xdp_pkts_compl,
+				   &xdp_bytes_compl);
 	}
 }
 

index 1e9f42938aac..039eefafba23 100644
@@ -20,7 +20,9 @@ void efx_dequeue_buffer(struct efx_tx_queue *tx_queue,
 			struct efx_tx_buffer *buffer,
 			unsigned int *pkts_compl,
 			unsigned int *bytes_compl,
-			unsigned int *efv_pkts_compl);
+			unsigned int *efv_pkts_compl,
+			unsigned int *xdp_pkts,
+			unsigned int *xdp_bytes);
 
 static inline bool efx_tx_buffer_in_use(struct efx_tx_buffer *buffer)
 {

index ea1e0e8ecbdd..b0ea4ca82cd8 100644
@@ -638,6 +638,10 @@ static void efx_get_queue_stats_rx(struct net_device *net_dev, int idx,
 	rx_queue = efx_channel_get_rx_queue(channel);
 	/* Count only packets since last time datapath was started */
 	stats->packets = rx_queue->rx_packets - rx_queue->old_rx_packets;
+	stats->hw_drops = efx_get_queue_stat_rx_hw_drops(channel) -
+			  channel->old_n_rx_hw_drops;
+	stats->hw_drop_overruns = channel->n_rx_nodesc_trunc -
+				  channel->old_n_rx_hw_drop_overruns;
 }
 
 static void efx_get_queue_stats_tx(struct net_device *net_dev, int idx,

@@ -668,6 +672,8 @@ static void efx_get_base_stats(struct net_device *net_dev,
 	struct efx_channel *channel;
 
 	rx->packets = 0;
+	rx->hw_drops = 0;
+	rx->hw_drop_overruns = 0;
 	tx->packets = 0;
 	tx->bytes = 0;
 

@@ -676,10 +682,15 @@ static void efx_get_base_stats(struct net_device *net_dev,
 	 */
 	efx_for_each_channel(channel, efx) {
 		rx_queue = efx_channel_get_rx_queue(channel);
-		if (channel->channel >= net_dev->real_num_rx_queues)
+		if (channel->channel >= net_dev->real_num_rx_queues) {
 			rx->packets += rx_queue->rx_packets;
-		else
+			rx->hw_drops += efx_get_queue_stat_rx_hw_drops(channel);
+			rx->hw_drop_overruns += channel->n_rx_nodesc_trunc;
+		} else {
 			rx->packets += rx_queue->old_rx_packets;
+			rx->hw_drops += channel->old_n_rx_hw_drops;
+			rx->hw_drop_overruns += channel->old_n_rx_hw_drop_overruns;
+		}
 		efx_for_each_channel_tx_queue(tx_queue, channel) {
 			if (channel->channel < efx->tx_channel_offset ||
 			    channel->channel >= efx->tx_channel_offset +

index 834d51812e2b..44d92c0e1b63 100644
@@ -1100,6 +1100,10 @@ void efx_start_channels(struct efx_nic *efx)
 			atomic_inc(&efx->active_queues);
 		}
 
+		/* reset per-queue stats */
+		channel->old_n_rx_hw_drops = efx_get_queue_stat_rx_hw_drops(channel);
+		channel->old_n_rx_hw_drop_overruns = channel->n_rx_nodesc_trunc;
+
 		efx_for_each_channel_rx_queue(rx_queue, channel) {
 			efx_init_rx_queue(rx_queue);
 			atomic_inc(&efx->active_queues);

index 46b702648721..547cf94014a3 100644
@@ -43,6 +43,13 @@ struct efx_channel *efx_copy_channel(const struct efx_channel *old_channel);
 void efx_start_channels(struct efx_nic *efx);
 void efx_stop_channels(struct efx_nic *efx);
 
+static inline u64 efx_get_queue_stat_rx_hw_drops(struct efx_channel *channel)
+{
+	return channel->n_rx_eth_crc_err + channel->n_rx_frm_trunc +
+	       channel->n_rx_overlength + channel->n_rx_nodesc_trunc +
+	       channel->n_rx_mport_bad;
+}
+
 void efx_init_napi_channel(struct efx_channel *channel);
 void efx_init_napi(struct efx_nic *efx);
 void efx_fini_napi_channel(struct efx_channel *channel);

index aba106d03d41..f6632f8185b2 100644
@@ -494,6 +494,10 @@ enum efx_sync_events_state {
  * @n_rx_xdp_redirect: Count of RX packets redirected to a different NIC by XDP
  * @n_rx_mport_bad: Count of RX packets dropped because their ingress mport was
  *	not recognised
+ * @old_n_rx_hw_drops: Count of all RX packets dropped for any reason as of last
+ *	efx_start_channels()
+ * @old_n_rx_hw_drop_overruns: Value of @n_rx_nodesc_trunc as of last
+ *	efx_start_channels()
  * @rx_pkt_n_frags: Number of fragments in next packet to be delivered by
  *	__efx_rx_packet(), or zero if there is none
  * @rx_pkt_index: Ring index of first buffer for next packet to be delivered

@@ -556,6 +560,9 @@ struct efx_channel {
 	unsigned int n_rx_xdp_redirect;
 	unsigned int n_rx_mport_bad;
 
+	unsigned int old_n_rx_hw_drops;
+	unsigned int old_n_rx_hw_drop_overruns;
+
 	unsigned int rx_pkt_n_frags;
 	unsigned int rx_pkt_index;
 

index b0ea4ca82cd8..68ddb28d3141 100644
@@ -654,11 +654,21 @@ static void efx_get_queue_stats_tx(struct net_device *net_dev, int idx,
 	channel = efx_get_tx_channel(efx, idx);
 	stats->packets = 0;
 	stats->bytes = 0;
+	stats->hw_gso_packets = 0;
+	stats->hw_gso_wire_packets = 0;
 	efx_for_each_channel_tx_queue(tx_queue, channel) {
 		stats->packets += tx_queue->complete_packets -
 				  tx_queue->old_complete_packets;
 		stats->bytes += tx_queue->complete_bytes -
 				tx_queue->old_complete_bytes;
+		/* Note that, unlike stats->packets and stats->bytes,
+		 * these count TXes enqueued, rather than completed,
+		 * which may not be what users expect.
+		 */
+		stats->hw_gso_packets += tx_queue->tso_bursts -
+					 tx_queue->old_tso_bursts;
+		stats->hw_gso_wire_packets += tx_queue->tso_packets -
+					      tx_queue->old_tso_packets;
 	}
 }
 

@@ -676,6 +686,8 @@ static void efx_get_base_stats(struct net_device *net_dev,
 	rx->hw_drop_overruns = 0;
 	tx->packets = 0;
 	tx->bytes = 0;
+	tx->hw_gso_packets = 0;
+	tx->hw_gso_wire_packets = 0;
 
 	/* Count all packets on non-core queues, and packets before last
 	 * datapath start on core queues.

@@ -697,9 +709,13 @@ static void efx_get_base_stats(struct net_device *net_dev,
 						net_dev->real_num_tx_queues) {
 				tx->packets += tx_queue->complete_packets;
 				tx->bytes += tx_queue->complete_bytes;
+				tx->hw_gso_packets += tx_queue->tso_bursts;
+				tx->hw_gso_wire_packets += tx_queue->tso_packets;
 			} else {
 				tx->packets += tx_queue->old_complete_packets;
 				tx->bytes += tx_queue->old_complete_bytes;
+				tx->hw_gso_packets += tx_queue->old_tso_bursts;
+				tx->hw_gso_wire_packets += tx_queue->old_tso_packets;
 			}
 			/* Include XDP TX in device-wide stats */
 			tx->packets += tx_queue->complete_xdp_packets;

index f6632f8185b2..4ca48db3e168 100644
@@ -197,6 +197,8 @@ struct efx_tx_buffer {
  *	efx_init_tx_queue()
  * @old_complete_bytes: Value of @complete_bytes as of last
  *	efx_init_tx_queue()
+ * @old_tso_bursts: Value of @tso_bursts as of last efx_init_tx_queue()
+ * @old_tso_packets: Value of @tso_packets as of last efx_init_tx_queue()
  * @read_count: Current read pointer.
  *	This is the number of buffers that have been removed from both rings.
  * @old_write_count: The value of @write_count when last checked.

@@ -276,6 +278,8 @@ struct efx_tx_queue {
 	bool xdp_tx;
 	unsigned long old_complete_packets;
 	unsigned long old_complete_bytes;
+	unsigned int old_tso_bursts;
+	unsigned int old_tso_packets;
 
 	/* Members used mainly on the completion path */
 	unsigned int read_count ____cacheline_aligned_in_smp;

index 2013a609f9be..a22a0d634ffc 100644
@@ -88,6 +88,8 @@ void efx_init_tx_queue(struct efx_tx_queue *tx_queue)
 
 	tx_queue->old_complete_packets = tx_queue->complete_packets;
 	tx_queue->old_complete_bytes = tx_queue->complete_bytes;
+	tx_queue->old_tso_bursts = tx_queue->tso_bursts;
+	tx_queue->old_tso_packets = tx_queue->tso_packets;
 
 	tx_queue->xdp_tx = efx_channel_is_xdp_tx(tx_queue->channel);
 	tx_queue->tso_version = 0;

index 992151775cb8..44dc75feb162 100644
@@ -135,6 +135,7 @@ void __ef100_rx_packet(struct efx_channel *channel)
 	}
 
 	++rx_queue->rx_packets;
+	rx_queue->rx_bytes += rx_buf->len;
 
 	efx_rx_packet_gro(channel, rx_buf, channel->rx_pkt_n_frags, eh, csum);
 	goto out;

index 68ddb28d3141..90bb7db15519 100644
@@ -638,6 +638,7 @@ static void efx_get_queue_stats_rx(struct net_device *net_dev, int idx,
 	rx_queue = efx_channel_get_rx_queue(channel);
 	/* Count only packets since last time datapath was started */
 	stats->packets = rx_queue->rx_packets - rx_queue->old_rx_packets;
+	stats->bytes = rx_queue->rx_bytes - rx_queue->old_rx_bytes;
 	stats->hw_drops = efx_get_queue_stat_rx_hw_drops(channel) -
 			  channel->old_n_rx_hw_drops;
 	stats->hw_drop_overruns = channel->n_rx_nodesc_trunc -

@@ -682,6 +683,7 @@ static void efx_get_base_stats(struct net_device *net_dev,
 	struct efx_channel *channel;
 
 	rx->packets = 0;
+	rx->bytes = 0;
 	rx->hw_drops = 0;
 	rx->hw_drop_overruns = 0;
 	tx->packets = 0;

@@ -696,10 +698,12 @@ static void efx_get_base_stats(struct net_device *net_dev,
 		rx_queue = efx_channel_get_rx_queue(channel);
 		if (channel->channel >= net_dev->real_num_rx_queues) {
 			rx->packets += rx_queue->rx_packets;
+			rx->bytes += rx_queue->rx_bytes;
 			rx->hw_drops += efx_get_queue_stat_rx_hw_drops(channel);
 			rx->hw_drop_overruns += channel->n_rx_nodesc_trunc;
 		} else {
 			rx->packets += rx_queue->old_rx_packets;
+			rx->bytes += rx_queue->old_rx_bytes;
 			rx->hw_drops += channel->old_n_rx_hw_drops;
 			rx->hw_drop_overruns += channel->old_n_rx_hw_drop_overruns;
 		}

index 4ca48db3e168..b54662d32f55 100644
@@ -400,7 +400,9 @@ struct efx_rx_page_state {
  * @slow_fill: Timer used to defer efx_nic_generate_fill_event().
  * @grant_work: workitem used to grant credits to the MAE if @grant_credits
  * @rx_packets: Number of packets received since this struct was created
+ * @rx_bytes: Number of bytes received since this struct was created
  * @old_rx_packets: Value of @rx_packets as of last efx_init_rx_queue()
+ * @old_rx_bytes: Value of @rx_bytes as of last efx_init_rx_queue()
  * @xdp_rxq_info: XDP specific RX queue information.
  * @xdp_rxq_info_valid: Is xdp_rxq_info valid data?.
  */

@@ -437,7 +439,9 @@ struct efx_rx_queue {
 	struct work_struct grant_work;
 	/* Statistics to supplement MAC stats */
 	unsigned long rx_packets;
+	unsigned long rx_bytes;
 	unsigned long old_rx_packets;
+	unsigned long old_rx_bytes;
 	struct xdp_rxq_info xdp_rxq_info;
 	bool xdp_rxq_info_valid;
 };

index f07495582125..ffca82207e47 100644
@@ -393,6 +393,7 @@ void __efx_rx_packet(struct efx_channel *channel)
 	}
 
 	rx_queue->rx_packets++;
+	rx_queue->rx_bytes += rx_buf->len;
 
 	if (!efx_do_xdp(efx, channel, rx_buf, &eh))
 		goto out;

index bdb4325a7c2c..ab358fe13e1d 100644
@@ -242,6 +242,7 @@ void efx_init_rx_queue(struct efx_rx_queue *rx_queue)
 	rx_queue->page_recycle_full = 0;
 
 	rx_queue->old_rx_packets = rx_queue->rx_packets;
+	rx_queue->old_rx_bytes = rx_queue->rx_bytes;
 
 	/* Initialise limit fields */
 	max_fill = efx->rxq_entries - EFX_RXD_HEAD_ROOM;