1 files changed, 386 insertions, 166 deletions

diff --git a/drivers/net/ethernet/ibm/ibmvnic.c b/drivers/net/ethernet/ibm/ibmvnic.c
index e19a6bb3f444..3808148c1fc7 100644
--- a/drivers/net/ethernet/ibm/ibmvnic.c
+++ b/drivers/net/ethernet/ibm/ibmvnic.c
@@ -97,6 +97,8 @@ static int pending_scrq(struct ibmvnic_adapter *,
 static union sub_crq *ibmvnic_next_scrq(struct ibmvnic_adapter *,
 					struct ibmvnic_sub_crq_queue *);
 static int ibmvnic_poll(struct napi_struct *napi, int data);
+static int reset_sub_crq_queues(struct ibmvnic_adapter *adapter);
+static inline void reinit_init_done(struct ibmvnic_adapter *adapter);
 static void send_query_map(struct ibmvnic_adapter *adapter);
 static int send_request_map(struct ibmvnic_adapter *, dma_addr_t, u32, u8);
 static int send_request_unmap(struct ibmvnic_adapter *, u8);
@@ -114,6 +116,8 @@ static void ibmvnic_tx_scrq_clean_buffer(struct ibmvnic_adapter *adapter,
 static void free_long_term_buff(struct ibmvnic_adapter *adapter,
 				struct ibmvnic_long_term_buff *ltb);
 static void ibmvnic_disable_irqs(struct ibmvnic_adapter *adapter);
+static void flush_reset_queue(struct ibmvnic_adapter *adapter);
+static void print_subcrq_error(struct device *dev, int rc, const char *func);
 
 struct ibmvnic_stat {
 	char name[ETH_GSTRING_LEN];
@@ -191,9 +195,8 @@ static void ibmvnic_clean_affinity(struct ibmvnic_adapter *adapter)
 	struct ibmvnic_sub_crq_queue **rxqs;
 	struct ibmvnic_sub_crq_queue **txqs;
 	int num_rxqs, num_txqs;
-	int rc, i;
+	int i;
 
-	rc = 0;
 	rxqs = adapter->rx_scrq;
 	txqs = adapter->tx_scrq;
 	num_txqs = adapter->num_active_tx_scrqs;
@@ -231,11 +234,17 @@ static int ibmvnic_set_queue_affinity(struct ibmvnic_sub_crq_queue *queue,
 		(*stragglers)--;
 	}
 	/* atomic write is safer than writing bit by bit directly */
-	for (i = 0; i < stride; i++) {
-		cpumask_set_cpu(*cpu, mask);
-		*cpu = cpumask_next_wrap(*cpu, cpu_online_mask,
-					 nr_cpu_ids, false);
+	for_each_online_cpu_wrap(i, *cpu) {
+		if (!stride--) {
+			/* For the next queue we start from the first
+			 * unused CPU in this queue
+			 */
+			*cpu = i;
+			break;
+		}
+		cpumask_set_cpu(i, mask);
 	}
+
 	/* set queue affinity mask */
 	cpumask_copy(queue->affinity_mask, mask);
 	rc = irq_set_affinity_and_hint(queue->irq, queue->affinity_mask);
@@ -250,10 +259,11 @@ static void ibmvnic_set_affinity(struct ibmvnic_adapter *adapter)
 	struct ibmvnic_sub_crq_queue **rxqs = adapter->rx_scrq;
 	struct ibmvnic_sub_crq_queue **txqs = adapter->tx_scrq;
 	struct ibmvnic_sub_crq_queue *queue;
-	int num_rxqs = adapter->num_active_rx_scrqs;
-	int num_txqs = adapter->num_active_tx_scrqs;
+	int num_rxqs = adapter->num_active_rx_scrqs, i_rxqs = 0;
+	int num_txqs = adapter->num_active_tx_scrqs, i_txqs = 0;
 	int total_queues, stride, stragglers, i;
-	unsigned int num_cpu, cpu;
+	unsigned int num_cpu, cpu = 0;
+	bool is_rx_queue;
 	int rc = 0;
 
 	netdev_dbg(adapter->netdev, "%s: Setting irq affinity hints", __func__);
@@ -270,33 +280,33 @@ static void ibmvnic_set_affinity(struct ibmvnic_adapter *adapter)
 	stride = max_t(int, num_cpu / total_queues, 1);
 	/* number of leftover cpu's */
 	stragglers = num_cpu >= total_queues ? num_cpu % total_queues : 0;
-	/* next available cpu to assign irq to */
-	cpu = cpumask_next(-1, cpu_online_mask);
 
-	for (i = 0; i < num_txqs; i++) {
-		queue = txqs[i];
+	for (i = 0; i < total_queues; i++) {
+		is_rx_queue = false;
+		/* balance core load by alternating rx and tx assignments
+		 * ex: TX0 -> RX0 -> TX1 -> RX1 etc.
+		 */
+		if ((i % 2 == 1 && i_rxqs < num_rxqs) || i_txqs == num_txqs) {
+			queue = rxqs[i_rxqs++];
+			is_rx_queue = true;
+		} else {
+			queue = txqs[i_txqs++];
+		}
+
 		rc = ibmvnic_set_queue_affinity(queue, &cpu, &stragglers,
 						stride);
 		if (rc)
 			goto out;
 
-		if (!queue)
+		if (!queue || is_rx_queue)
 			continue;
 
 		rc = __netif_set_xps_queue(adapter->netdev,
 					   cpumask_bits(queue->affinity_mask),
-					   i, XPS_CPUS);
+					   i_txqs - 1, XPS_CPUS);
 		if (rc)
 			netdev_warn(adapter->netdev, "%s: Set XPS on queue %d failed, rc = %d.\n",
-				    __func__, i, rc);
-	}
-
-	for (i = 0; i < num_rxqs; i++) {
-		queue = rxqs[i];
-		rc = ibmvnic_set_queue_affinity(queue, &cpu, &stragglers,
-						stride);
-		if (rc)
-			goto out;
+				    __func__, i_txqs - 1, rc);
 	}
 
 out:
@@ -746,6 +756,17 @@ static void deactivate_rx_pools(struct ibmvnic_adapter *adapter)
 		adapter->rx_pool[i].active = 0;
 }
 
+static void ibmvnic_set_safe_max_ind_descs(struct ibmvnic_adapter *adapter)
+{
+	if (adapter->cur_max_ind_descs > IBMVNIC_SAFE_IND_DESC) {
+		netdev_info(adapter->netdev,
+			    "set max ind descs from %u to safe limit %u\n",
+			    adapter->cur_max_ind_descs,
+			    IBMVNIC_SAFE_IND_DESC);
+		adapter->cur_max_ind_descs = IBMVNIC_SAFE_IND_DESC;
+	}
+}
+
 static void replenish_rx_pool(struct ibmvnic_adapter *adapter,
 			      struct ibmvnic_rx_pool *pool)
 {
@@ -833,7 +854,7 @@ static void replenish_rx_pool(struct ibmvnic_adapter *adapter,
 		sub_crq->rx_add.len = cpu_to_be32(pool->buff_size << shift);
 
 		/* if send_subcrq_indirect queue is full, flush to VIOS */
-		if (ind_bufp->index == IBMVNIC_MAX_IND_DESCS ||
+		if (ind_bufp->index == adapter->cur_max_ind_descs ||
 		    i == count - 1) {
 			lpar_rc =
 				send_subcrq_indirect(adapter, handle,
@@ -852,6 +873,14 @@ static void replenish_rx_pool(struct ibmvnic_adapter *adapter,
 failure:
 	if (lpar_rc != H_PARAMETER && lpar_rc != H_CLOSED)
 		dev_err_ratelimited(dev, "rx: replenish packet buffer failed\n");
+
+	/* Detect platform limit H_PARAMETER */
+	if (lpar_rc == H_PARAMETER)
+		ibmvnic_set_safe_max_ind_descs(adapter);
+
+	/* For all error case, temporarily drop only this batch
+	 * Rely on TCP/IP retransmissions to retry and recover
+	 */
 	for (i = ind_bufp->index - 1; i >= 0; --i) {
 		struct ibmvnic_rx_buff *rx_buff;
 
@@ -1502,8 +1531,8 @@ static const char *adapter_state_to_string(enum vnic_state state)
 
 static int ibmvnic_login(struct net_device *netdev)
 {
+	unsigned long flags, timeout = msecs_to_jiffies(20000);
 	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
-	unsigned long timeout = msecs_to_jiffies(20000);
 	int retry_count = 0;
 	int retries = 10;
 	bool retry;
@@ -1524,11 +1553,9 @@ static int ibmvnic_login(struct net_device *netdev)
 
 		if (!wait_for_completion_timeout(&adapter->init_done,
 						 timeout)) {
-			netdev_warn(netdev, "Login timed out, retrying...\n");
-			retry = true;
-			adapter->init_done_rc = 0;
-			retry_count++;
-			continue;
+			netdev_warn(netdev, "Login timed out\n");
+			adapter->login_pending = false;
+			goto partial_reset;
 		}
 
 		if (adapter->init_done_rc == ABORTED) {
@@ -1570,10 +1597,69 @@ static int ibmvnic_login(struct net_device *netdev)
 					    "SCRQ irq initialization failed\n");
 				return rc;
 			}
+		/* Default/timeout error handling, reset and start fresh */
 		} else if (adapter->init_done_rc) {
 			netdev_warn(netdev, "Adapter login failed, init_done_rc = %d\n",
 				    adapter->init_done_rc);
-			return -EIO;
+
+partial_reset:
+			/* adapter login failed, so free any CRQs or sub-CRQs
+			 * and register again before attempting to login again.
+			 * If we don't do this then the VIOS may think that
+			 * we are already logged in and reject any subsequent
+			 * attempts
+			 */
+			netdev_warn(netdev,
+				    "Freeing and re-registering CRQs before attempting to login again\n");
+			retry = true;
+			adapter->init_done_rc = 0;
+			release_sub_crqs(adapter, true);
+			/* Much of this is similar logic as ibmvnic_probe(),
+			 * we are essentially re-initializing communication
+			 * with the server. We really should not run any
+			 * resets/failovers here because this is already a form
+			 * of reset and we do not want parallel resets occurring
+			 */
+			do {
+				reinit_init_done(adapter);
+				/* Clear any failovers we got in the previous
+				 * pass since we are re-initializing the CRQ
+				 */
+				adapter->failover_pending = false;
+				release_crq_queue(adapter);
+				/* If we don't sleep here then we risk an
+				 * unnecessary failover event from the VIOS.
+				 * This is a known VIOS issue caused by a vnic
+				 * device freeing and registering a CRQ too
+				 * quickly.
+				 */
+				msleep(1500);
+				/* Avoid any resets, since we are currently
+				 * resetting.
+				 */
+				spin_lock_irqsave(&adapter->rwi_lock, flags);
+				flush_reset_queue(adapter);
+				spin_unlock_irqrestore(&adapter->rwi_lock,
+						       flags);
+
+				rc = init_crq_queue(adapter);
+				if (rc) {
+					netdev_err(netdev, "login recovery: init CRQ failed %d\n",
+						   rc);
+					return -EIO;
+				}
+
+				rc = ibmvnic_reset_init(adapter, false);
+				if (rc)
+					netdev_err(netdev, "login recovery: Reset init failed %d\n",
+						   rc);
+				/* IBMVNIC_CRQ_INIT will return EAGAIN if it
+				 * fails, since ibmvnic_reset_init will free
+				 * irq's in failure, we won't be able to receive
+				 * new CRQs so we need to keep trying. probe()
+				 * handles this similarly.
+				 */
+			} while (rc == -EAGAIN && retry_count++ < retries);
 		}
 	} while (retry);
 
@@ -1585,12 +1671,22 @@ static int ibmvnic_login(struct net_device *netdev)
 
 static void release_login_buffer(struct ibmvnic_adapter *adapter)
 {
+	if (!adapter->login_buf)
+		return;
+
+	dma_unmap_single(&adapter->vdev->dev, adapter->login_buf_token,
+			 adapter->login_buf_sz, DMA_TO_DEVICE);
 	kfree(adapter->login_buf);
 	adapter->login_buf = NULL;
 }
 
 static void release_login_rsp_buffer(struct ibmvnic_adapter *adapter)
 {
+	if (!adapter->login_rsp_buf)
+		return;
+
+	dma_unmap_single(&adapter->vdev->dev, adapter->login_rsp_buf_token,
+			 adapter->login_rsp_buf_sz, DMA_FROM_DEVICE);
 	kfree(adapter->login_rsp_buf);
 	adapter->login_rsp_buf = NULL;
 }
@@ -1813,7 +1909,14 @@ static int __ibmvnic_open(struct net_device *netdev)
 		if (prev_state == VNIC_CLOSED)
 			enable_irq(adapter->tx_scrq[i]->irq);
 		enable_scrq_irq(adapter, adapter->tx_scrq[i]);
-		netdev_tx_reset_queue(netdev_get_tx_queue(netdev, i));
+		/* netdev_tx_reset_queue will reset dql stats. During NON_FATAL
+		 * resets, don't reset the stats because there could be batched
+		 * skb's waiting to be sent. If we reset dql stats, we risk
+		 * num_completed being greater than num_queued. This will cause
+		 * a BUG_ON in dql_completed().
+		 */
+		if (adapter->reset_reason != VNIC_RESET_NON_FATAL)
+			netdev_tx_reset_queue(netdev_get_tx_queue(netdev, i));
 	}
 
 	rc = set_link_state(adapter, IBMVNIC_LOGICAL_LNK_UP);
@@ -2061,63 +2164,49 @@ static int ibmvnic_close(struct net_device *netdev)
 }
 
 /**
- * build_hdr_data - creates L2/L3/L4 header data buffer
+ * get_hdr_lens - fills list of L2/L3/L4 hdr lens
  * @hdr_field: bitfield determining needed headers
  * @skb: socket buffer
- * @hdr_len: array of header lengths
- * @hdr_data: buffer to write the header to
+ * @hdr_len: array of header lengths to be filled
  *
  * Reads hdr_field to determine which headers are needed by firmware.
  * Builds a buffer containing these headers.  Saves individual header
  * lengths and total buffer length to be used to build descriptors.
+ *
+ * Return: total len of all headers
  */
-static int build_hdr_data(u8 hdr_field, struct sk_buff *skb,
-			  int *hdr_len, u8 *hdr_data)
+static int get_hdr_lens(u8 hdr_field, struct sk_buff *skb,
+			int *hdr_len)
 {
 	int len = 0;
-	u8 *hdr;
 
-	if (skb_vlan_tagged(skb) && !skb_vlan_tag_present(skb))
-		hdr_len[0] = sizeof(struct vlan_ethhdr);
-	else
-		hdr_len[0] = sizeof(struct ethhdr);
+
+	if ((hdr_field >> 6) & 1) {
+		hdr_len[0] = skb_mac_header_len(skb);
+		len += hdr_len[0];
+	}
+
+	if ((hdr_field >> 5) & 1) {
+		hdr_len[1] = skb_network_header_len(skb);
+		len += hdr_len[1];
+	}
+
+	if (!((hdr_field >> 4) & 1))
+		return len;
 
 	if (skb->protocol == htons(ETH_P_IP)) {
-		hdr_len[1] = ip_hdr(skb)->ihl * 4;
 		if (ip_hdr(skb)->protocol == IPPROTO_TCP)
 			hdr_len[2] = tcp_hdrlen(skb);
 		else if (ip_hdr(skb)->protocol == IPPROTO_UDP)
 			hdr_len[2] = sizeof(struct udphdr);
 	} else if (skb->protocol == htons(ETH_P_IPV6)) {
-		hdr_len[1] = sizeof(struct ipv6hdr);
 		if (ipv6_hdr(skb)->nexthdr == IPPROTO_TCP)
 			hdr_len[2] = tcp_hdrlen(skb);
 		else if (ipv6_hdr(skb)->nexthdr == IPPROTO_UDP)
 			hdr_len[2] = sizeof(struct udphdr);
-	} else if (skb->protocol == htons(ETH_P_ARP)) {
-		hdr_len[1] = arp_hdr_len(skb->dev);
-		hdr_len[2] = 0;
 	}
 
-	memset(hdr_data, 0, 120);
-	if ((hdr_field >> 6) & 1) {
-		hdr = skb_mac_header(skb);
-		memcpy(hdr_data, hdr, hdr_len[0]);
-		len += hdr_len[0];
-	}
-
-	if ((hdr_field >> 5) & 1) {
-		hdr = skb_network_header(skb);
-		memcpy(hdr_data + len, hdr, hdr_len[1]);
-		len += hdr_len[1];
-	}
-
-	if ((hdr_field >> 4) & 1) {
-		hdr = skb_transport_header(skb);
-		memcpy(hdr_data + len, hdr, hdr_len[2]);
-		len += hdr_len[2];
-	}
-	return len;
+	return len + hdr_len[2];
 }
 
 /**
@@ -2130,12 +2219,14 @@ static int build_hdr_data(u8 hdr_field, struct sk_buff *skb,
  *
  * Creates header and, if needed, header extension descriptors and
  * places them in a descriptor array, scrq_arr
+ *
+ * Return: Number of header descs
  */
 
 static int create_hdr_descs(u8 hdr_field, u8 *hdr_data, int len, int *hdr_len,
 			    union sub_crq *scrq_arr)
 {
-	union sub_crq hdr_desc;
+	union sub_crq *hdr_desc;
 	int tmp_len = len;
 	int num_descs = 0;
 	u8 *data, *cur;
@@ -2144,28 +2235,26 @@ static int create_hdr_descs(u8 hdr_field, u8 *hdr_data, int len, int *hdr_len,
 	while (tmp_len > 0) {
 		cur = hdr_data + len - tmp_len;
 
-		memset(&hdr_desc, 0, sizeof(hdr_desc));
-		if (cur != hdr_data) {
-			data = hdr_desc.hdr_ext.data;
+		hdr_desc = &scrq_arr[num_descs];
+		if (num_descs) {
+			data = hdr_desc->hdr_ext.data;
 			tmp = tmp_len > 29 ? 29 : tmp_len;
-			hdr_desc.hdr_ext.first = IBMVNIC_CRQ_CMD;
-			hdr_desc.hdr_ext.type = IBMVNIC_HDR_EXT_DESC;
-			hdr_desc.hdr_ext.len = tmp;
+			hdr_desc->hdr_ext.first = IBMVNIC_CRQ_CMD;
+			hdr_desc->hdr_ext.type = IBMVNIC_HDR_EXT_DESC;
+			hdr_desc->hdr_ext.len = tmp;
 		} else {
-			data = hdr_desc.hdr.data;
+			data = hdr_desc->hdr.data;
 			tmp = tmp_len > 24 ? 24 : tmp_len;
-			hdr_desc.hdr.first = IBMVNIC_CRQ_CMD;
-			hdr_desc.hdr.type = IBMVNIC_HDR_DESC;
-			hdr_desc.hdr.len = tmp;
-			hdr_desc.hdr.l2_len = (u8)hdr_len[0];
-			hdr_desc.hdr.l3_len = cpu_to_be16((u16)hdr_len[1]);
-			hdr_desc.hdr.l4_len = (u8)hdr_len[2];
-			hdr_desc.hdr.flag = hdr_field << 1;
+			hdr_desc->hdr.first = IBMVNIC_CRQ_CMD;
+			hdr_desc->hdr.type = IBMVNIC_HDR_DESC;
+			hdr_desc->hdr.len = tmp;
+			hdr_desc->hdr.l2_len = (u8)hdr_len[0];
+			hdr_desc->hdr.l3_len = cpu_to_be16((u16)hdr_len[1]);
+			hdr_desc->hdr.l4_len = (u8)hdr_len[2];
+			hdr_desc->hdr.flag = hdr_field << 1;
 		}
 		memcpy(data, cur, tmp);
 		tmp_len -= tmp;
-		*scrq_arr = hdr_desc;
-		scrq_arr++;
 		num_descs++;
 	}
 
@@ -2188,13 +2277,11 @@ static void build_hdr_descs_arr(struct sk_buff *skb,
 				int *num_entries, u8 hdr_field)
 {
 	int hdr_len[3] = {0, 0, 0};
-	u8 hdr_data[140] = {0};
 	int tot_len;
 
-	tot_len = build_hdr_data(hdr_field, skb, hdr_len,
-				 hdr_data);
-	*num_entries += create_hdr_descs(hdr_field, hdr_data, tot_len, hdr_len,
-					 indir_arr + 1);
+	tot_len = get_hdr_lens(hdr_field, skb, hdr_len);
+	*num_entries += create_hdr_descs(hdr_field, skb_mac_header(skb),
+					 tot_len, hdr_len, indir_arr + 1);
 }
 
 static int ibmvnic_xmit_workarounds(struct sk_buff *skb,
@@ -2244,9 +2331,7 @@ static void ibmvnic_tx_scrq_clean_buffer(struct ibmvnic_adapter *adapter,
 					  tx_pool->num_buffers - 1 :
 					  tx_pool->consumer_index - 1;
 		tx_buff = &tx_pool->tx_buff[index];
-		adapter->netdev->stats.tx_packets--;
-		adapter->netdev->stats.tx_bytes -= tx_buff->skb->len;
-		adapter->tx_stats_buffers[queue_num].packets--;
+		adapter->tx_stats_buffers[queue_num].batched_packets--;
 		adapter->tx_stats_buffers[queue_num].bytes -=
 						tx_buff->skb->len;
 		dev_kfree_skb_any(tx_buff->skb);
@@ -2271,8 +2356,29 @@ static void ibmvnic_tx_scrq_clean_buffer(struct ibmvnic_adapter *adapter,
 	}
 }
 
+static int send_subcrq_direct(struct ibmvnic_adapter *adapter,
+			      u64 remote_handle, u64 *entry)
+{
+	unsigned int ua = adapter->vdev->unit_address;
+	struct device *dev = &adapter->vdev->dev;
+	int rc;
+
+	/* Make sure the hypervisor sees the complete request */
+	dma_wmb();
+	rc = plpar_hcall_norets(H_SEND_SUB_CRQ, ua,
+				cpu_to_be64(remote_handle),
+				cpu_to_be64(entry[0]), cpu_to_be64(entry[1]),
+				cpu_to_be64(entry[2]), cpu_to_be64(entry[3]));
+
+	if (rc)
+		print_subcrq_error(dev, rc, __func__);
+
+	return rc;
+}
+
 static int ibmvnic_tx_scrq_flush(struct ibmvnic_adapter *adapter,
-				 struct ibmvnic_sub_crq_queue *tx_scrq)
+				 struct ibmvnic_sub_crq_queue *tx_scrq,
+				 bool indirect)
 {
 	struct ibmvnic_ind_xmit_queue *ind_bufp;
 	u64 dma_addr;
@@ -2287,17 +2393,35 @@ static int ibmvnic_tx_scrq_flush(struct ibmvnic_adapter *adapter,
 
 	if (!entries)
 		return 0;
-	rc = send_subcrq_indirect(adapter, handle, dma_addr, entries);
-	if (rc)
-		ibmvnic_tx_scrq_clean_buffer(adapter, tx_scrq);
+
+	if (indirect)
+		rc = send_subcrq_indirect(adapter, handle, dma_addr, entries);
 	else
+		rc = send_subcrq_direct(adapter, handle,
+					(u64 *)ind_bufp->indir_arr);
+
+	if (rc) {
+		dev_err_ratelimited(&adapter->vdev->dev,
+				    "tx_flush failed, rc=%u (%llu entries dma=%pad handle=%llx)\n",
+				    rc, entries, &dma_addr, handle);
+		/* Detect platform limit H_PARAMETER */
+		if (rc == H_PARAMETER)
+			ibmvnic_set_safe_max_ind_descs(adapter);
+
+		/* For all error case, temporarily drop only this batch
+		 * Rely on TCP/IP retransmissions to retry and recover
+		 */
+		ibmvnic_tx_scrq_clean_buffer(adapter, tx_scrq);
+	} else {
 		ind_bufp->index = 0;
-	return 0;
+	}
+	return rc;
 }
 
 static netdev_tx_t ibmvnic_xmit(struct sk_buff *skb, struct net_device *netdev)
 {
 	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+	u32 cur_max_ind_descs = adapter->cur_max_ind_descs;
 	int queue_num = skb_get_queue_mapping(skb);
 	u8 *hdrs = (u8 *)&adapter->tx_rx_desc_req;
 	struct device *dev = &adapter->vdev->dev;
@@ -2311,13 +2435,16 @@ static netdev_tx_t ibmvnic_xmit(struct sk_buff *skb, struct net_device *netdev)
 	unsigned int tx_map_failed = 0;
 	union sub_crq indir_arr[16];
 	unsigned int tx_dropped = 0;
-	unsigned int tx_packets = 0;
+	unsigned int tx_dpackets = 0;
+	unsigned int tx_bpackets = 0;
 	unsigned int tx_bytes = 0;
 	dma_addr_t data_dma_addr;
 	struct netdev_queue *txq;
 	unsigned long lpar_rc;
+	unsigned int skblen;
 	union sub_crq tx_crq;
 	unsigned int offset;
+	bool use_scrq_send_direct = false;
 	int num_entries = 1;
 	unsigned char *dst;
 	int bufidx = 0;
@@ -2345,7 +2472,9 @@ static netdev_tx_t ibmvnic_xmit(struct sk_buff *skb, struct net_device *netdev)
 		tx_dropped++;
 		tx_send_failed++;
 		ret = NETDEV_TX_OK;
-		ibmvnic_tx_scrq_flush(adapter, tx_scrq);
+		lpar_rc = ibmvnic_tx_scrq_flush(adapter, tx_scrq, true);
+		if (lpar_rc != H_SUCCESS)
+			goto tx_err;
 		goto out;
 	}
 
@@ -2360,8 +2489,10 @@ static netdev_tx_t ibmvnic_xmit(struct sk_buff *skb, struct net_device *netdev)
 		dev_kfree_skb_any(skb);
 		tx_send_failed++;
 		tx_dropped++;
-		ibmvnic_tx_scrq_flush(adapter, tx_scrq);
 		ret = NETDEV_TX_OK;
+		lpar_rc = ibmvnic_tx_scrq_flush(adapter, tx_scrq, true);
+		if (lpar_rc != H_SUCCESS)
+			goto tx_err;
 		goto out;
 	}
 
@@ -2373,6 +2504,20 @@ static netdev_tx_t ibmvnic_xmit(struct sk_buff *skb, struct net_device *netdev)
 	memset(dst, 0, tx_pool->buf_size);
 	data_dma_addr = ltb->addr + offset;
 
+	/* if we are going to send_subcrq_direct this then we need to
+	 * update the checksum before copying the data into ltb. Essentially
+	 * these packets force disable CSO so that we can guarantee that
+	 * FW does not need header info and we can send direct. Also, vnic
+	 * server must be able to xmit standard packets without header data
+	 */
+	if (*hdrs == 0 && !skb_is_gso(skb) &&
+	    !ind_bufp->index && !netdev_xmit_more()) {
+		use_scrq_send_direct = true;
+		if (skb->ip_summed == CHECKSUM_PARTIAL &&
+		    skb_checksum_help(skb))
+			use_scrq_send_direct = false;
+	}
+
 	if (skb_shinfo(skb)->nr_frags) {
 		int cur, i;
 
@@ -2392,16 +2537,26 @@ static netdev_tx_t ibmvnic_xmit(struct sk_buff *skb, struct net_device *netdev)
 		skb_copy_from_linear_data(skb, dst, skb->len);
 	}
 
-	/* post changes to long_term_buff *dst before VIOS accessing it */
-	dma_wmb();
-
 	tx_pool->consumer_index =
 	    (tx_pool->consumer_index + 1) % tx_pool->num_buffers;
 
 	tx_buff = &tx_pool->tx_buff[bufidx];
+
+	/* Sanity checks on our free map to make sure it points to an index
+	 * that is not being occupied by another skb. If skb memory is
+	 * not freed then we see congestion control kick in and halt tx.
+	 */
+	if (unlikely(tx_buff->skb)) {
+		dev_warn_ratelimited(dev, "TX free map points to untracked skb (%s %d idx=%d)\n",
+				     skb_is_gso(skb) ? "tso_pool" : "tx_pool",
+				     queue_num, bufidx);
+		dev_kfree_skb_any(tx_buff->skb);
+	}
+
 	tx_buff->skb = skb;
 	tx_buff->index = bufidx;
 	tx_buff->pool_index = queue_num;
+	skblen = skb->len;
 
 	memset(&tx_crq, 0, sizeof(tx_crq));
 	tx_crq.v1.first = IBMVNIC_CRQ_CMD;
@@ -2445,6 +2600,19 @@ static netdev_tx_t ibmvnic_xmit(struct sk_buff *skb, struct net_device *netdev)
 		tx_crq.v1.flags1 |= IBMVNIC_TX_LSO;
 		tx_crq.v1.mss = cpu_to_be16(skb_shinfo(skb)->gso_size);
 		hdrs += 2;
+	} else if (use_scrq_send_direct) {
+		/* See above comment, CSO disabled with direct xmit */
+		tx_crq.v1.flags1 &= ~(IBMVNIC_TX_CHKSUM_OFFLOAD);
+		ind_bufp->index = 1;
+		tx_buff->num_entries = 1;
+		netdev_tx_sent_queue(txq, skb->len);
+		ind_bufp->indir_arr[0] = tx_crq;
+		lpar_rc = ibmvnic_tx_scrq_flush(adapter, tx_scrq, false);
+		if (lpar_rc != H_SUCCESS)
+			goto tx_err;
+
+		tx_dpackets++;
+		goto early_exit;
 	}
 
 	if ((*hdrs >> 7) & 1)
@@ -2453,8 +2621,8 @@ static netdev_tx_t ibmvnic_xmit(struct sk_buff *skb, struct net_device *netdev)
 	tx_crq.v1.n_crq_elem = num_entries;
 	tx_buff->num_entries = num_entries;
 	/* flush buffer if current entry can not fit */
-	if (num_entries + ind_bufp->index > IBMVNIC_MAX_IND_DESCS) {
-		lpar_rc = ibmvnic_tx_scrq_flush(adapter, tx_scrq);
+	if (num_entries + ind_bufp->index > cur_max_ind_descs) {
+		lpar_rc = ibmvnic_tx_scrq_flush(adapter, tx_scrq, true);
 		if (lpar_rc != H_SUCCESS)
 			goto tx_flush_err;
 	}
@@ -2462,23 +2630,26 @@ static netdev_tx_t ibmvnic_xmit(struct sk_buff *skb, struct net_device *netdev)
 	indir_arr[0] = tx_crq;
 	memcpy(&ind_bufp->indir_arr[ind_bufp->index], &indir_arr[0],
 	       num_entries * sizeof(struct ibmvnic_generic_scrq));
+
 	ind_bufp->index += num_entries;
 	if (__netdev_tx_sent_queue(txq, skb->len,
 				   netdev_xmit_more() &&
-				   ind_bufp->index < IBMVNIC_MAX_IND_DESCS)) {
-		lpar_rc = ibmvnic_tx_scrq_flush(adapter, tx_scrq);
+				   ind_bufp->index < cur_max_ind_descs)) {
+		lpar_rc = ibmvnic_tx_scrq_flush(adapter, tx_scrq, true);
 		if (lpar_rc != H_SUCCESS)
 			goto tx_err;
 	}
 
+	tx_bpackets++;
+
+early_exit:
 	if (atomic_add_return(num_entries, &tx_scrq->used)
 					>= adapter->req_tx_entries_per_subcrq) {
 		netdev_dbg(netdev, "Stopping queue %d\n", queue_num);
 		netif_stop_subqueue(netdev, queue_num);
 	}
 
-	tx_packets++;
-	tx_bytes += skb->len;
+	tx_bytes += skblen;
 	txq_trans_cond_update(txq);
 	ret = NETDEV_TX_OK;
 	goto out;
@@ -2505,12 +2676,10 @@ tx_err:
 	}
 out:
 	rcu_read_unlock();
-	netdev->stats.tx_dropped += tx_dropped;
-	netdev->stats.tx_bytes += tx_bytes;
-	netdev->stats.tx_packets += tx_packets;
 	adapter->tx_send_failed += tx_send_failed;
 	adapter->tx_map_failed += tx_map_failed;
-	adapter->tx_stats_buffers[queue_num].packets += tx_packets;
+	adapter->tx_stats_buffers[queue_num].batched_packets += tx_bpackets;
+	adapter->tx_stats_buffers[queue_num].direct_packets += tx_dpackets;
 	adapter->tx_stats_buffers[queue_num].bytes += tx_bytes;
 	adapter->tx_stats_buffers[queue_num].dropped_packets += tx_dropped;
 
@@ -3309,6 +3478,25 @@ err:
 	return -ret;
 }
 
+static void ibmvnic_get_stats64(struct net_device *netdev,
+				struct rtnl_link_stats64 *stats)
+{
+	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+	int i;
+
+	for (i = 0; i < adapter->req_rx_queues; i++) {
+		stats->rx_packets += adapter->rx_stats_buffers[i].packets;
+		stats->rx_bytes   += adapter->rx_stats_buffers[i].bytes;
+	}
+
+	for (i = 0; i < adapter->req_tx_queues; i++) {
+		stats->tx_packets += adapter->tx_stats_buffers[i].batched_packets;
+		stats->tx_packets += adapter->tx_stats_buffers[i].direct_packets;
+		stats->tx_bytes   += adapter->tx_stats_buffers[i].bytes;
+		stats->tx_dropped += adapter->tx_stats_buffers[i].dropped_packets;
+	}
+}
+
 static void ibmvnic_tx_timeout(struct net_device *dev, unsigned int txqueue)
 {
 	struct ibmvnic_adapter *adapter = netdev_priv(dev);
@@ -3424,23 +3612,20 @@ restart_poll:
 
 		length = skb->len;
 		napi_gro_receive(napi, skb); /* send it up */
-		netdev->stats.rx_packets++;
-		netdev->stats.rx_bytes += length;
 		adapter->rx_stats_buffers[scrq_num].packets++;
 		adapter->rx_stats_buffers[scrq_num].bytes += length;
 		frames_processed++;
 	}
 
 	if (adapter->state != VNIC_CLOSING &&
-	    ((atomic_read(&adapter->rx_pool[scrq_num].available) <
-	      adapter->req_rx_add_entries_per_subcrq / 2) ||
-	      frames_processed < budget))
+	    (atomic_read(&adapter->rx_pool[scrq_num].available) <
+	      adapter->req_rx_add_entries_per_subcrq / 2))
 		replenish_rx_pool(adapter, &adapter->rx_pool[scrq_num]);
 	if (frames_processed < budget) {
 		if (napi_complete_done(napi, frames_processed)) {
 			enable_scrq_irq(adapter, rx_scrq);
 			if (pending_scrq(adapter, rx_scrq)) {
-				if (napi_reschedule(napi)) {
+				if (napi_schedule(napi)) {
 					disable_scrq_irq(adapter, rx_scrq);
 					goto restart_poll;
 				}
@@ -3536,6 +3721,7 @@ static const struct net_device_ops ibmvnic_netdev_ops = {
 	.ndo_set_rx_mode	= ibmvnic_set_multi,
 	.ndo_set_mac_address	= ibmvnic_set_mac,
 	.ndo_validate_addr	= eth_validate_addr,
+	.ndo_get_stats64	= ibmvnic_get_stats64,
 	.ndo_tx_timeout		= ibmvnic_tx_timeout,
 	.ndo_change_mtu		= ibmvnic_change_mtu,
 	.ndo_features_check     = ibmvnic_features_check,
@@ -3672,29 +3858,20 @@ static void ibmvnic_get_strings(struct net_device *dev, u32 stringset, u8 *data)
 	if (stringset != ETH_SS_STATS)
 		return;
 
-	for (i = 0; i < ARRAY_SIZE(ibmvnic_stats); i++, data += ETH_GSTRING_LEN)
-		memcpy(data, ibmvnic_stats[i].name, ETH_GSTRING_LEN);
+	for (i = 0; i < ARRAY_SIZE(ibmvnic_stats); i++)
+		ethtool_puts(&data, ibmvnic_stats[i].name);
 
 	for (i = 0; i < adapter->req_tx_queues; i++) {
-		snprintf(data, ETH_GSTRING_LEN, "tx%d_packets", i);
-		data += ETH_GSTRING_LEN;
-
-		snprintf(data, ETH_GSTRING_LEN, "tx%d_bytes", i);
-		data += ETH_GSTRING_LEN;
-
-		snprintf(data, ETH_GSTRING_LEN, "tx%d_dropped_packets", i);
-		data += ETH_GSTRING_LEN;
+		ethtool_sprintf(&data, "tx%d_batched_packets", i);
+		ethtool_sprintf(&data, "tx%d_direct_packets", i);
+		ethtool_sprintf(&data, "tx%d_bytes", i);
+		ethtool_sprintf(&data, "tx%d_dropped_packets", i);
 	}
 
 	for (i = 0; i < adapter->req_rx_queues; i++) {
-		snprintf(data, ETH_GSTRING_LEN, "rx%d_packets", i);
-		data += ETH_GSTRING_LEN;
-
-		snprintf(data, ETH_GSTRING_LEN, "rx%d_bytes", i);
-		data += ETH_GSTRING_LEN;
-
-		snprintf(data, ETH_GSTRING_LEN, "rx%d_interrupts", i);
-		data += ETH_GSTRING_LEN;
+		ethtool_sprintf(&data, "rx%d_packets", i);
+		ethtool_sprintf(&data, "rx%d_bytes", i);
+		ethtool_sprintf(&data, "rx%d_interrupts", i);
 	}
 }
 
@@ -3741,7 +3918,9 @@ static void ibmvnic_get_ethtool_stats(struct net_device *dev,
 				      (adapter, ibmvnic_stats[i].offset));
 
 	for (j = 0; j < adapter->req_tx_queues; j++) {
-		data[i] = adapter->tx_stats_buffers[j].packets;
+		data[i] = adapter->tx_stats_buffers[j].batched_packets;
+		i++;
+		data[i] = adapter->tx_stats_buffers[j].direct_packets;
 		i++;
 		data[i] = adapter->tx_stats_buffers[j].bytes;
 		i++;
@@ -3858,7 +4037,7 @@ static void release_sub_crq_queue(struct ibmvnic_adapter *adapter,
 	}
 
 	dma_free_coherent(dev,
-			  IBMVNIC_IND_ARR_SZ,
+			  IBMVNIC_IND_MAX_ARR_SZ,
 			  scrq->ind_buf.indir_arr,
 			  scrq->ind_buf.indir_dma);
 
@@ -3915,7 +4094,7 @@ static struct ibmvnic_sub_crq_queue *init_sub_crq_queue(struct ibmvnic_adapter
 
 	scrq->ind_buf.indir_arr =
 		dma_alloc_coherent(dev,
-				   IBMVNIC_IND_ARR_SZ,
+				   IBMVNIC_IND_MAX_ARR_SZ,
 				   &scrq->ind_buf.indir_dma,
 				   GFP_KERNEL);
 
@@ -3978,6 +4157,12 @@ static void release_sub_crqs(struct ibmvnic_adapter *adapter, bool do_h_free)
 		adapter->num_active_tx_scrqs = 0;
 	}
 
+	/* Clean any remaining outstanding SKBs
+	 * we freed the irq so we won't be hearing
+	 * from them
+	 */
+	clean_tx_pools(adapter);
+
 	if (adapter->rx_scrq) {
 		for (i = 0; i < adapter->num_active_rx_scrqs; i++) {
 			if (!adapter->rx_scrq[i])
@@ -4068,20 +4253,17 @@ static int ibmvnic_complete_tx(struct ibmvnic_adapter *adapter,
 			       struct ibmvnic_sub_crq_queue *scrq)
 {
 	struct device *dev = &adapter->vdev->dev;
+	int num_packets = 0, total_bytes = 0;
 	struct ibmvnic_tx_pool *tx_pool;
 	struct ibmvnic_tx_buff *txbuff;
 	struct netdev_queue *txq;
 	union sub_crq *next;
-	int index;
-	int i;
+	int index, i;
 
 restart_loop:
 	while (pending_scrq(adapter, scrq)) {
 		unsigned int pool = scrq->pool_index;
 		int num_entries = 0;
-		int total_bytes = 0;
-		int num_packets = 0;
-
 		next = ibmvnic_next_scrq(adapter, scrq);
 		for (i = 0; i < next->tx_comp.num_comps; i++) {
 			index = be32_to_cpu(next->tx_comp.correlators[i]);
@@ -4117,8 +4299,6 @@ restart_loop:
 		/* remove tx_comp scrq*/
 		next->tx_comp.first = 0;
 
-		txq = netdev_get_tx_queue(adapter->netdev, scrq->pool_index);
-		netdev_tx_completed_queue(txq, num_packets, total_bytes);
 
 		if (atomic_sub_return(num_entries, &scrq->used) <=
 		    (adapter->req_tx_entries_per_subcrq / 2) &&
@@ -4143,6 +4323,9 @@ restart_loop:
 		goto restart_loop;
 	}
 
+	txq = netdev_get_tx_queue(adapter->netdev, scrq->pool_index);
+	netdev_tx_completed_queue(txq, num_packets, total_bytes);
+
 	return 0;
 }
 
@@ -4694,6 +4877,18 @@ static void vnic_add_client_data(struct ibmvnic_adapter *adapter,
 	strscpy(vlcd->name, adapter->netdev->name, len);
 }
 
+static void ibmvnic_print_hex_dump(struct net_device *dev, void *buf,
+				   size_t len)
+{
+	unsigned char hex_str[16 * 3];
+
+	for (size_t i = 0; i < len; i += 16) {
+		hex_dump_to_buffer((unsigned char *)buf + i, len - i, 16, 8,
+				   hex_str, sizeof(hex_str), false);
+		netdev_dbg(dev, "%s\n", hex_str);
+	}
+}
+
 static int send_login(struct ibmvnic_adapter *adapter)
 {
 	struct ibmvnic_login_rsp_buffer *login_rsp_buffer;
@@ -4804,10 +4999,8 @@ static int send_login(struct ibmvnic_adapter *adapter)
 	vnic_add_client_data(adapter, vlcd);
 
 	netdev_dbg(adapter->netdev, "Login Buffer:\n");
-	for (i = 0; i < (adapter->login_buf_sz - 1) / 8 + 1; i++) {
-		netdev_dbg(adapter->netdev, "%016lx\n",
-			   ((unsigned long *)(adapter->login_buf))[i]);
-	}
+	ibmvnic_print_hex_dump(adapter->netdev, adapter->login_buf,
+			       adapter->login_buf_sz);
 
 	memset(&crq, 0, sizeof(crq));
 	crq.login.first = IBMVNIC_CRQ_CMD;
@@ -4820,11 +5013,14 @@ static int send_login(struct ibmvnic_adapter *adapter)
 	if (rc) {
 		adapter->login_pending = false;
 		netdev_err(adapter->netdev, "Failed to send login, rc=%d\n", rc);
-		goto buf_rsp_map_failed;
+		goto buf_send_failed;
 	}
 
 	return 0;
 
+buf_send_failed:
+	dma_unmap_single(dev, rsp_buffer_token, rsp_buffer_size,
+			 DMA_FROM_DEVICE);
 buf_rsp_map_failed:
 	kfree(login_rsp_buffer);
 	adapter->login_rsp_buf = NULL;
@@ -5165,7 +5361,8 @@ static void handle_vpd_rsp(union ibmvnic_crq *crq,
 	/* copy firmware version string from vpd into adapter */
 	if ((substr + 3 + fw_level_len) <
 	    (adapter->vpd->buff + adapter->vpd->len)) {
-		strncpy((char *)adapter->fw_version, substr + 3, fw_level_len);
+		strscpy(adapter->fw_version, substr + 3,
+			sizeof(adapter->fw_version));
 	} else {
 		dev_info(dev, "FW substr extrapolated VPD buff\n");
 	}
@@ -5180,15 +5377,13 @@ static void handle_query_ip_offload_rsp(struct ibmvnic_adapter *adapter)
 {
 	struct device *dev = &adapter->vdev->dev;
 	struct ibmvnic_query_ip_offload_buffer *buf = &adapter->ip_offload_buf;
-	int i;
 
 	dma_unmap_single(dev, adapter->ip_offload_tok,
 			 sizeof(adapter->ip_offload_buf), DMA_FROM_DEVICE);
 
 	netdev_dbg(adapter->netdev, "Query IP Offload Buffer:\n");
-	for (i = 0; i < (sizeof(adapter->ip_offload_buf) - 1) / 8 + 1; i++)
-		netdev_dbg(adapter->netdev, "%016lx\n",
-			   ((unsigned long *)(buf))[i]);
+	ibmvnic_print_hex_dump(adapter->netdev, buf,
+			       sizeof(adapter->ip_offload_buf));
 
 	netdev_dbg(adapter->netdev, "ipv4_chksum = %d\n", buf->ipv4_chksum);
 	netdev_dbg(adapter->netdev, "ipv6_chksum = %d\n", buf->ipv6_chksum);
@@ -5386,6 +5581,7 @@ static int handle_login_rsp(union ibmvnic_crq *login_rsp_crq,
 	int num_tx_pools;
 	int num_rx_pools;
 	u64 *size_array;
+	u32 rsp_len;
 	int i;
 
 	/* CHECK: Test/set of login_pending does not need to be atomic
@@ -5397,11 +5593,6 @@ static int handle_login_rsp(union ibmvnic_crq *login_rsp_crq,
 	}
 	adapter->login_pending = false;
 
-	dma_unmap_single(dev, adapter->login_buf_token, adapter->login_buf_sz,
-			 DMA_TO_DEVICE);
-	dma_unmap_single(dev, adapter->login_rsp_buf_token,
-			 adapter->login_rsp_buf_sz, DMA_FROM_DEVICE);
-
 	/* If the number of queues requested can't be allocated by the
 	 * server, the login response will return with code 1. We will need
 	 * to resend the login buffer with fewer queues requested.
@@ -5423,10 +5614,8 @@ static int handle_login_rsp(union ibmvnic_crq *login_rsp_crq,
 	netdev->mtu = adapter->req_mtu - ETH_HLEN;
 
 	netdev_dbg(adapter->netdev, "Login Response Buffer:\n");
-	for (i = 0; i < (adapter->login_rsp_buf_sz - 1) / 8 + 1; i++) {
-		netdev_dbg(adapter->netdev, "%016lx\n",
-			   ((unsigned long *)(adapter->login_rsp_buf))[i]);
-	}
+	ibmvnic_print_hex_dump(netdev, adapter->login_rsp_buf,
+			       adapter->login_rsp_buf_sz);
 
 	/* Sanity checks */
 	if (login->num_txcomp_subcrqs != login_rsp->num_txsubm_subcrqs ||
@@ -5437,6 +5626,23 @@ static int handle_login_rsp(union ibmvnic_crq *login_rsp_crq,
 		ibmvnic_reset(adapter, VNIC_RESET_FATAL);
 		return -EIO;
 	}
+
+	rsp_len = be32_to_cpu(login_rsp->len);
+	if (be32_to_cpu(login->login_rsp_len) < rsp_len ||
+	    rsp_len <= be32_to_cpu(login_rsp->off_txsubm_subcrqs) ||
+	    rsp_len <= be32_to_cpu(login_rsp->off_rxadd_subcrqs) ||
+	    rsp_len <= be32_to_cpu(login_rsp->off_rxadd_buff_size) ||
+	    rsp_len <= be32_to_cpu(login_rsp->off_supp_tx_desc)) {
+		/* This can happen if a login request times out and there are
+		 * 2 outstanding login requests sent, the LOGIN_RSP crq
+		 * could have been for the older login request. So we are
+		 * parsing the newer response buffer which may be incomplete
+		 */
+		dev_err(dev, "FATAL: Login rsp offsets/lengths invalid\n");
+		ibmvnic_reset(adapter, VNIC_RESET_FATAL);
+		return -EIO;
+	}
+
 	size_array = (u64 *)((u8 *)(adapter->login_rsp_buf) +
 		be32_to_cpu(adapter->login_rsp_buf->off_rxadd_buff_size));
 	/* variable buffer sizes are not supported, so just read the
@@ -6194,6 +6400,19 @@ static int ibmvnic_reset_init(struct ibmvnic_adapter *adapter, bool reset)
 			rc = reset_sub_crq_queues(adapter);
 		}
 	} else {
+		if (adapter->reset_reason == VNIC_RESET_MOBILITY) {
+			/* After an LPM, reset the max number of indirect
+			 * subcrq descriptors per H_SEND_SUB_CRQ_INDIRECT
+			 * hcall to the default max (e.g POWER8 -> POWER10)
+			 *
+			 * If the new destination platform does not support
+			 * the higher limit max (e.g. POWER10-> POWER8 LPM)
+			 * H_PARAMETER will trigger automatic fallback to the
+			 * safe minimum limit.
+			 */
+			adapter->cur_max_ind_descs = IBMVNIC_MAX_IND_DESCS;
+		}
+
 		rc = init_sub_crqs(adapter);
 	}
 
@@ -6345,6 +6564,7 @@ static int ibmvnic_probe(struct vio_dev *dev, const struct vio_device_id *id)
 
 	adapter->wait_for_reset = false;
 	adapter->last_reset_time = jiffies;
+	adapter->cur_max_ind_descs = IBMVNIC_MAX_IND_DESCS;
 
 	rc = register_netdev(netdev);
 	if (rc) {