Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net

Pull networking fixes from David Miller:
 "It looks like a decent sized set of fixes, but a lot of these are one
  liner off-by-one and similar type changes:

   1) Fix netlink header pointer to calcular bad attribute offset
      reported to user. From Pablo Neira Ayuso.

   2) Don't double clear PHY interrupts when ->did_interrupt is set,
      from Heiner Kallweit.

   3) Add missing validation of various (devlink, nl802154, fib, etc.)
      attributes, from Jakub Kicinski.

   4) Missing *pos increments in various netfilter seq_next ops, from
      Vasily Averin.

   5) Missing break in of_mdiobus_register() loop, from Dajun Jin.

   6) Don't double bump tx_dropped in veth driver, from Jiang Lidong.

   7) Work around FMAN erratum A050385, from Madalin Bucur.

   8) Make sure ARP header is pulled early enough in bonding driver,
      from Eric Dumazet.

   9) Do a cond_resched() during multicast processing of ipvlan and
      macvlan, from Mahesh Bandewar.

  10) Don't attach cgroups to unrelated sockets when in interrupt
      context, from Shakeel Butt.

  11) Fix tpacket ring state management when encountering unknown GSO
      types. From Willem de Bruijn.

  12) Fix MDIO bus PHY resume by checking mdio_bus_phy_may_suspend()
      only in the suspend context. From Heiner Kallweit"

* git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (112 commits)
  net: systemport: fix index check to avoid an array out of bounds access
  tc-testing: add ETS scheduler to tdc build configuration
  net: phy: fix MDIO bus PM PHY resuming
  net: hns3: clear port base VLAN when unload PF
  net: hns3: fix RMW issue for VLAN filter switch
  net: hns3: fix VF VLAN table entries inconsistent issue
  net: hns3: fix "tc qdisc del" failed issue
  taprio: Fix sending packets without dequeueing them
  net: mvmdio: avoid error message for optional IRQ
  net: dsa: mv88e6xxx: Add missing mask of ATU occupancy register
  net: memcg: fix lockdep splat in inet_csk_accept()
  s390/qeth: implement smarter resizing of the RX buffer pool
  s390/qeth: refactor buffer pool code
  s390/qeth: use page pointers to manage RX buffer pool
  seg6: fix SRv6 L2 tunnels to use IANA-assigned protocol number
  net: dsa: Don't instantiate phylink for CPU/DSA ports unless needed
  net/packet: tpacket_rcv: do not increment ring index on drop
  sxgbe: Fix off by one in samsung driver strncpy size arg
  net: caif: Add lockdep expression to RCU traversal primitive
  MAINTAINERS: remove Sathya Perla as Emulex NIC maintainer
  ...

47 files changed, 484 insertions, 148 deletions

diff --git a/drivers/net/bonding/bond_alb.c b/drivers/net/bonding/bond_alb.c
index 1cc2cd894f87..c81698550e5a 100644
--- a/drivers/net/bonding/bond_alb.c
+++ b/drivers/net/bonding/bond_alb.c
@@ -50,11 +50,6 @@ struct arp_pkt {
 };
 #pragma pack()
 
-static inline struct arp_pkt *arp_pkt(const struct sk_buff *skb)
-{
-	return (struct arp_pkt *)skb_network_header(skb);
-}
-
 /* Forward declaration */
 static void alb_send_learning_packets(struct slave *slave, u8 mac_addr[],
 				      bool strict_match);
@@ -553,10 +548,11 @@ static void rlb_req_update_subnet_clients(struct bonding *bond, __be32 src_ip)
 	spin_unlock(&bond->mode_lock);
 }
 
-static struct slave *rlb_choose_channel(struct sk_buff *skb, struct bonding *bond)
+static struct slave *rlb_choose_channel(struct sk_buff *skb,
+					struct bonding *bond,
+					const struct arp_pkt *arp)
 {
 	struct alb_bond_info *bond_info = &(BOND_ALB_INFO(bond));
-	struct arp_pkt *arp = arp_pkt(skb);
 	struct slave *assigned_slave, *curr_active_slave;
 	struct rlb_client_info *client_info;
 	u32 hash_index = 0;
@@ -653,8 +649,12 @@ static struct slave *rlb_choose_channel(struct sk_buff *skb, struct bonding *bon
  */
 static struct slave *rlb_arp_xmit(struct sk_buff *skb, struct bonding *bond)
 {
-	struct arp_pkt *arp = arp_pkt(skb);
 	struct slave *tx_slave = NULL;
+	struct arp_pkt *arp;
+
+	if (!pskb_network_may_pull(skb, sizeof(*arp)))
+		return NULL;
+	arp = (struct arp_pkt *)skb_network_header(skb);
 
 	/* Don't modify or load balance ARPs that do not originate locally
 	 * (e.g.,arrive via a bridge).
@@ -664,7 +664,7 @@ static struct slave *rlb_arp_xmit(struct sk_buff *skb, struct bonding *bond)
 
 	if (arp->op_code == htons(ARPOP_REPLY)) {
 		/* the arp must be sent on the selected rx channel */
-		tx_slave = rlb_choose_channel(skb, bond);
+		tx_slave = rlb_choose_channel(skb, bond, arp);
 		if (tx_slave)
 			bond_hw_addr_copy(arp->mac_src, tx_slave->dev->dev_addr,
 					  tx_slave->dev->addr_len);
@@ -676,7 +676,7 @@ static struct slave *rlb_arp_xmit(struct sk_buff *skb, struct bonding *bond)
 		 * When the arp reply is received the entry will be updated
 		 * with the correct unicast address of the client.
 		 */
-		tx_slave = rlb_choose_channel(skb, bond);
+		tx_slave = rlb_choose_channel(skb, bond, arp);
 
 		/* The ARP reply packets must be delayed so that
 		 * they can cancel out the influence of the ARP request.
diff --git a/drivers/net/can/dev.c b/drivers/net/can/dev.c
index 6ee06a49fb4c..68834a2853c9 100644
--- a/drivers/net/can/dev.c
+++ b/drivers/net/can/dev.c
@@ -883,6 +883,7 @@ static const struct nla_policy can_policy[IFLA_CAN_MAX + 1] = {
 				= { .len = sizeof(struct can_bittiming) },
 	[IFLA_CAN_DATA_BITTIMING_CONST]
 				= { .len = sizeof(struct can_bittiming_const) },
+	[IFLA_CAN_TERMINATION]	= { .type = NLA_U16 },
 };
 
 static int can_validate(struct nlattr *tb[], struct nlattr *data[],
diff --git a/drivers/net/dsa/mv88e6xxx/chip.c b/drivers/net/dsa/mv88e6xxx/chip.c
index 8c9289549688..2f993e673ec7 100644
--- a/drivers/net/dsa/mv88e6xxx/chip.c
+++ b/drivers/net/dsa/mv88e6xxx/chip.c
@@ -2769,6 +2769,8 @@ static u64 mv88e6xxx_devlink_atu_bin_get(struct mv88e6xxx_chip *chip,
 		goto unlock;
 	}
 
+	occupancy &= MV88E6XXX_G2_ATU_STATS_MASK;
+
 unlock:
 	mv88e6xxx_reg_unlock(chip);
 
diff --git a/drivers/net/dsa/mv88e6xxx/global2.c b/drivers/net/dsa/mv88e6xxx/global2.c
index 01503014b1ee..8fd483020c5b 100644
--- a/drivers/net/dsa/mv88e6xxx/global2.c
+++ b/drivers/net/dsa/mv88e6xxx/global2.c
@@ -1099,6 +1099,13 @@ int mv88e6xxx_g2_irq_setup(struct mv88e6xxx_chip *chip)
 {
 	int err, irq, virq;
 
+	chip->g2_irq.masked = ~0;
+	mv88e6xxx_reg_lock(chip);
+	err = mv88e6xxx_g2_int_mask(chip, ~chip->g2_irq.masked);
+	mv88e6xxx_reg_unlock(chip);
+	if (err)
+		return err;
+
 	chip->g2_irq.domain = irq_domain_add_simple(
 		chip->dev->of_node, 16, 0, &mv88e6xxx_g2_irq_domain_ops, chip);
 	if (!chip->g2_irq.domain)
@@ -1108,7 +1115,6 @@ int mv88e6xxx_g2_irq_setup(struct mv88e6xxx_chip *chip)
 		irq_create_mapping(chip->g2_irq.domain, irq);
 
 	chip->g2_irq.chip = mv88e6xxx_g2_irq_chip;
-	chip->g2_irq.masked = ~0;
 
 	chip->device_irq = irq_find_mapping(chip->g1_irq.domain,
 					    MV88E6XXX_G1_STS_IRQ_DEVICE);
diff --git a/drivers/net/dsa/sja1105/sja1105_main.c b/drivers/net/dsa/sja1105/sja1105_main.c
index 03ba6d25f7fe..7edea5741a5f 100644
--- a/drivers/net/dsa/sja1105/sja1105_main.c
+++ b/drivers/net/dsa/sja1105/sja1105_main.c
@@ -1741,7 +1741,8 @@ static void sja1105_teardown(struct dsa_switch *ds)
 		if (!dsa_is_user_port(ds, port))
 			continue;
 
-		kthread_destroy_worker(sp->xmit_worker);
+		if (sp->xmit_worker)
+			kthread_destroy_worker(sp->xmit_worker);
 	}
 
 	sja1105_tas_teardown(ds);
diff --git a/drivers/net/ethernet/broadcom/bcmsysport.c b/drivers/net/ethernet/broadcom/bcmsysport.c
index e0611cba87f9..15b31cddc054 100644
--- a/drivers/net/ethernet/broadcom/bcmsysport.c
+++ b/drivers/net/ethernet/broadcom/bcmsysport.c
@@ -2135,7 +2135,7 @@ static int bcm_sysport_rule_set(struct bcm_sysport_priv *priv,
 		return -ENOSPC;
 
 	index = find_first_zero_bit(priv->filters, RXCHK_BRCM_TAG_MAX);
-	if (index > RXCHK_BRCM_TAG_MAX)
+	if (index >= RXCHK_BRCM_TAG_MAX)
 		return -ENOSPC;
 
 	/* Location is the classification ID, and index is the position
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index f9a8151f092c..c5c8effc0139 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -10982,13 +10982,13 @@ static int bnxt_change_mtu(struct net_device *dev, int new_mtu)
 	struct bnxt *bp = netdev_priv(dev);
 
 	if (netif_running(dev))
-		bnxt_close_nic(bp, false, false);
+		bnxt_close_nic(bp, true, false);
 
 	dev->mtu = new_mtu;
 	bnxt_set_ring_params(bp);
 
 	if (netif_running(dev))
-		return bnxt_open_nic(bp, false, false);
+		return bnxt_open_nic(bp, true, false);
 
 	return 0;
 }
diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
index e8fc1671c581..1f67e6729a2c 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
@@ -2007,8 +2007,8 @@ int bnxt_flash_package_from_file(struct net_device *dev, const char *filename,
 	struct hwrm_nvm_install_update_output *resp = bp->hwrm_cmd_resp_addr;
 	struct hwrm_nvm_install_update_input install = {0};
 	const struct firmware *fw;
-	int rc, hwrm_err = 0;
 	u32 item_len;
+	int rc = 0;
 	u16 index;
 
 	bnxt_hwrm_fw_set_time(bp);
@@ -2052,15 +2052,14 @@ int bnxt_flash_package_from_file(struct net_device *dev, const char *filename,
 			memcpy(kmem, fw->data, fw->size);
 			modify.host_src_addr = cpu_to_le64(dma_handle);
 
-			hwrm_err = hwrm_send_message(bp, &modify,
-						     sizeof(modify),
-						     FLASH_PACKAGE_TIMEOUT);
+			rc = hwrm_send_message(bp, &modify, sizeof(modify),
+					       FLASH_PACKAGE_TIMEOUT);
 			dma_free_coherent(&bp->pdev->dev, fw->size, kmem,
 					  dma_handle);
 		}
 	}
 	release_firmware(fw);
-	if (rc || hwrm_err)
+	if (rc)
 		goto err_exit;
 
 	if ((install_type & 0xffff) == 0)
@@ -2069,20 +2068,19 @@ int bnxt_flash_package_from_file(struct net_device *dev, const char *filename,
 	install.install_type = cpu_to_le32(install_type);
 
 	mutex_lock(&bp->hwrm_cmd_lock);
-	hwrm_err = _hwrm_send_message(bp, &install, sizeof(install),
-				      INSTALL_PACKAGE_TIMEOUT);
-	if (hwrm_err) {
+	rc = _hwrm_send_message(bp, &install, sizeof(install),
+				INSTALL_PACKAGE_TIMEOUT);
+	if (rc) {
 		u8 error_code = ((struct hwrm_err_output *)resp)->cmd_err;
 
 		if (resp->error_code && error_code ==
 		    NVM_INSTALL_UPDATE_CMD_ERR_CODE_FRAG_ERR) {
 			install.flags |= cpu_to_le16(
 			       NVM_INSTALL_UPDATE_REQ_FLAGS_ALLOWED_TO_DEFRAG);
-			hwrm_err = _hwrm_send_message(bp, &install,
-						      sizeof(install),
-						      INSTALL_PACKAGE_TIMEOUT);
+			rc = _hwrm_send_message(bp, &install, sizeof(install),
+						INSTALL_PACKAGE_TIMEOUT);
 		}
-		if (hwrm_err)
+		if (rc)
 			goto flash_pkg_exit;
 	}
 
@@ -2094,7 +2092,7 @@ int bnxt_flash_package_from_file(struct net_device *dev, const char *filename,
 flash_pkg_exit:
 	mutex_unlock(&bp->hwrm_cmd_lock);
 err_exit:
-	if (hwrm_err == -EACCES)
+	if (rc == -EACCES)
 		bnxt_print_admin_err(bp);
 	return rc;
 }
diff --git a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
index 649842a8aa28..97f90edbc068 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
+++ b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
@@ -5381,12 +5381,11 @@ static inline bool is_x_10g_port(const struct link_config *lc)
 static int cfg_queues(struct adapter *adap)
 {
 	u32 avail_qsets, avail_eth_qsets, avail_uld_qsets;
+	u32 i, n10g = 0, qidx = 0, n1g = 0;
+	u32 ncpus = num_online_cpus();
 	u32 niqflint, neq, num_ulds;
 	struct sge *s = &adap->sge;
-	u32 i, n10g = 0, qidx = 0;
-#ifndef CONFIG_CHELSIO_T4_DCB
-	int q10g = 0;
-#endif
+	u32 q10g = 0, q1g;
 
 	/* Reduce memory usage in kdump environment, disable all offload. */
 	if (is_kdump_kernel() || (is_uld(adap) && t4_uld_mem_alloc(adap))) {
@@ -5424,44 +5423,50 @@ static int cfg_queues(struct adapter *adap)
 		n10g += is_x_10g_port(&adap2pinfo(adap, i)->link_cfg);
 
 	avail_eth_qsets = min_t(u32, avail_qsets, MAX_ETH_QSETS);
+
+	/* We default to 1 queue per non-10G port and up to # of cores queues
+	 * per 10G port.
+	 */
+	if (n10g)
+		q10g = (avail_eth_qsets - (adap->params.nports - n10g)) / n10g;
+
+	n1g = adap->params.nports - n10g;
 #ifdef CONFIG_CHELSIO_T4_DCB
 	/* For Data Center Bridging support we need to be able to support up
 	 * to 8 Traffic Priorities; each of which will be assigned to its
 	 * own TX Queue in order to prevent Head-Of-Line Blocking.
 	 */
+	q1g = 8;
 	if (adap->params.nports * 8 > avail_eth_qsets) {
 		dev_err(adap->pdev_dev, "DCB avail_eth_qsets=%d < %d!\n",
 			avail_eth_qsets, adap->params.nports * 8);
 		return -ENOMEM;
 	}
 
-	for_each_port(adap, i) {
-		struct port_info *pi = adap2pinfo(adap, i);
+	if (adap->params.nports * ncpus < avail_eth_qsets)
+		q10g = max(8U, ncpus);
+	else
+		q10g = max(8U, q10g);
 
-		pi->first_qset = qidx;
-		pi->nqsets = is_kdump_kernel() ? 1 : 8;
-		qidx += pi->nqsets;
-	}
-#else /* !CONFIG_CHELSIO_T4_DCB */
-	/* We default to 1 queue per non-10G port and up to # of cores queues
-	 * per 10G port.
-	 */
-	if (n10g)
-		q10g = (avail_eth_qsets - (adap->params.nports - n10g)) / n10g;
-	if (q10g > netif_get_num_default_rss_queues())
-		q10g = netif_get_num_default_rss_queues();
+	while ((q10g * n10g) > (avail_eth_qsets - n1g * q1g))
+		q10g--;
 
-	if (is_kdump_kernel())
+#else /* !CONFIG_CHELSIO_T4_DCB */
+	q1g = 1;
+	q10g = min(q10g, ncpus);
+#endif /* !CONFIG_CHELSIO_T4_DCB */
+	if (is_kdump_kernel()) {
 		q10g = 1;
+		q1g = 1;
+	}
 
 	for_each_port(adap, i) {
 		struct port_info *pi = adap2pinfo(adap, i);
 
 		pi->first_qset = qidx;
-		pi->nqsets = is_x_10g_port(&pi->link_cfg) ? q10g : 1;
+		pi->nqsets = is_x_10g_port(&pi->link_cfg) ? q10g : q1g;
 		qidx += pi->nqsets;
 	}
-#endif /* !CONFIG_CHELSIO_T4_DCB */
 
 	s->ethqsets = qidx;
 	s->max_ethqsets = qidx;   /* MSI-X may lower it later */
@@ -5473,7 +5478,7 @@ static int cfg_queues(struct adapter *adap)
 		 * capped by the number of available cores.
 		 */
 		num_ulds = adap->num_uld + adap->num_ofld_uld;
-		i = min_t(u32, MAX_OFLD_QSETS, num_online_cpus());
+		i = min_t(u32, MAX_OFLD_QSETS, ncpus);
 		avail_uld_qsets = roundup(i, adap->params.nports);
 		if (avail_qsets < num_ulds * adap->params.nports) {
 			adap->params.offload = 0;
diff --git a/drivers/net/ethernet/freescale/dpaa/dpaa_eth.c b/drivers/net/ethernet/freescale/dpaa/dpaa_eth.c
index fd93d542f497..ca74a684a904 100644
--- a/drivers/net/ethernet/freescale/dpaa/dpaa_eth.c
+++ b/drivers/net/ethernet/freescale/dpaa/dpaa_eth.c
@@ -1,4 +1,5 @@
 /* Copyright 2008 - 2016 Freescale Semiconductor Inc.
+ * Copyright 2020 NXP
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are met:
@@ -123,7 +124,22 @@ MODULE_PARM_DESC(tx_timeout, "The Tx timeout in ms");
 #define FSL_QMAN_MAX_OAL	127
 
 /* Default alignment for start of data in an Rx FD */
+#ifdef CONFIG_DPAA_ERRATUM_A050385
+/* aligning data start to 64 avoids DMA transaction splits, unless the buffer
+ * is crossing a 4k page boundary
+ */
+#define DPAA_FD_DATA_ALIGNMENT  (fman_has_errata_a050385() ? 64 : 16)
+/* aligning to 256 avoids DMA transaction splits caused by 4k page boundary
+ * crossings; also, all SG fragments except the last must have a size multiple
+ * of 256 to avoid DMA transaction splits
+ */
+#define DPAA_A050385_ALIGN 256
+#define DPAA_FD_RX_DATA_ALIGNMENT (fman_has_errata_a050385() ? \
+				   DPAA_A050385_ALIGN : 16)
+#else
 #define DPAA_FD_DATA_ALIGNMENT  16
+#define DPAA_FD_RX_DATA_ALIGNMENT DPAA_FD_DATA_ALIGNMENT
+#endif
 
 /* The DPAA requires 256 bytes reserved and mapped for the SGT */
 #define DPAA_SGT_SIZE 256
@@ -158,8 +174,13 @@ MODULE_PARM_DESC(tx_timeout, "The Tx timeout in ms");
 #define DPAA_PARSE_RESULTS_SIZE sizeof(struct fman_prs_result)
 #define DPAA_TIME_STAMP_SIZE 8
 #define DPAA_HASH_RESULTS_SIZE 8
+#ifdef CONFIG_DPAA_ERRATUM_A050385
+#define DPAA_RX_PRIV_DATA_SIZE (DPAA_A050385_ALIGN - (DPAA_PARSE_RESULTS_SIZE\
+	 + DPAA_TIME_STAMP_SIZE + DPAA_HASH_RESULTS_SIZE))
+#else
 #define DPAA_RX_PRIV_DATA_SIZE	(u16)(DPAA_TX_PRIV_DATA_SIZE + \
 					dpaa_rx_extra_headroom)
+#endif
 
 #define DPAA_ETH_PCD_RXQ_NUM	128
 
@@ -180,7 +201,12 @@ static struct dpaa_bp *dpaa_bp_array[BM_MAX_NUM_OF_POOLS];
 
 #define DPAA_BP_RAW_SIZE 4096
 
+#ifdef CONFIG_DPAA_ERRATUM_A050385
+#define dpaa_bp_size(raw_size) (SKB_WITH_OVERHEAD(raw_size) & \
+				~(DPAA_A050385_ALIGN - 1))
+#else
 #define dpaa_bp_size(raw_size) SKB_WITH_OVERHEAD(raw_size)
+#endif
 
 static int dpaa_max_frm;
 
@@ -1192,7 +1218,7 @@ static int dpaa_eth_init_rx_port(struct fman_port *port, struct dpaa_bp *bp,
 	buf_prefix_content.pass_prs_result = true;
 	buf_prefix_content.pass_hash_result = true;
 	buf_prefix_content.pass_time_stamp = true;
-	buf_prefix_content.data_align = DPAA_FD_DATA_ALIGNMENT;
+	buf_prefix_content.data_align = DPAA_FD_RX_DATA_ALIGNMENT;
 
 	rx_p = &params.specific_params.rx_params;
 	rx_p->err_fqid = errq->fqid;
@@ -1662,6 +1688,8 @@ static u8 rx_csum_offload(const struct dpaa_priv *priv, const struct qm_fd *fd)
 	return CHECKSUM_NONE;
 }
 
+#define PTR_IS_ALIGNED(x, a) (IS_ALIGNED((unsigned long)(x), (a)))
+
 /* Build a linear skb around the received buffer.
  * We are guaranteed there is enough room at the end of the data buffer to
  * accommodate the shared info area of the skb.
@@ -1733,8 +1761,7 @@ static struct sk_buff *sg_fd_to_skb(const struct dpaa_priv *priv,
 
 		sg_addr = qm_sg_addr(&sgt[i]);
 		sg_vaddr = phys_to_virt(sg_addr);
-		WARN_ON(!IS_ALIGNED((unsigned long)sg_vaddr,
-				    SMP_CACHE_BYTES));
+		WARN_ON(!PTR_IS_ALIGNED(sg_vaddr, SMP_CACHE_BYTES));
 
 		dma_unmap_page(priv->rx_dma_dev, sg_addr,
 			       DPAA_BP_RAW_SIZE, DMA_FROM_DEVICE);
@@ -2022,6 +2049,75 @@ static inline int dpaa_xmit(struct dpaa_priv *priv,
 	return 0;
 }
 
+#ifdef CONFIG_DPAA_ERRATUM_A050385
+int dpaa_a050385_wa(struct net_device *net_dev, struct sk_buff **s)
+{
+	struct dpaa_priv *priv = netdev_priv(net_dev);
+	struct sk_buff *new_skb, *skb = *s;
+	unsigned char *start, i;
+
+	/* check linear buffer alignment */
+	if (!PTR_IS_ALIGNED(skb->data, DPAA_A050385_ALIGN))
+		goto workaround;
+
+	/* linear buffers just need to have an aligned start */
+	if (!skb_is_nonlinear(skb))
+		return 0;
+
+	/* linear data size for nonlinear skbs needs to be aligned */
+	if (!IS_ALIGNED(skb_headlen(skb), DPAA_A050385_ALIGN))
+		goto workaround;
+
+	for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
+		skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
+
+		/* all fragments need to have aligned start addresses */
+		if (!IS_ALIGNED(skb_frag_off(frag), DPAA_A050385_ALIGN))
+			goto workaround;
+
+		/* all but last fragment need to have aligned sizes */
+		if (!IS_ALIGNED(skb_frag_size(frag), DPAA_A050385_ALIGN) &&
+		    (i < skb_shinfo(skb)->nr_frags - 1))
+			goto workaround;
+	}
+
+	return 0;
+
+workaround:
+	/* copy all the skb content into a new linear buffer */
+	new_skb = netdev_alloc_skb(net_dev, skb->len + DPAA_A050385_ALIGN - 1 +
+						priv->tx_headroom);
+	if (!new_skb)
+		return -ENOMEM;
+
+	/* NET_SKB_PAD bytes already reserved, adding up to tx_headroom */
+	skb_reserve(new_skb, priv->tx_headroom - NET_SKB_PAD);
+
+	/* Workaround for DPAA_A050385 requires data start to be aligned */
+	start = PTR_ALIGN(new_skb->data, DPAA_A050385_ALIGN);
+	if (start - new_skb->data != 0)
+		skb_reserve(new_skb, start - new_skb->data);
+
+	skb_put(new_skb, skb->len);
+	skb_copy_bits(skb, 0, new_skb->data, skb->len);
+	skb_copy_header(new_skb, skb);
+	new_skb->dev = skb->dev;
+
+	/* We move the headroom when we align it so we have to reset the
+	 * network and transport header offsets relative to the new data
+	 * pointer. The checksum offload relies on these offsets.
+	 */
+	skb_set_network_header(new_skb, skb_network_offset(skb));
+	skb_set_transport_header(new_skb, skb_transport_offset(skb));
+
+	/* TODO: does timestamping need the result in the old skb? */
+	dev_kfree_skb(skb);
+	*s = new_skb;
+
+	return 0;
+}
+#endif
+
 static netdev_tx_t
 dpaa_start_xmit(struct sk_buff *skb, struct net_device *net_dev)
 {
@@ -2068,6 +2164,14 @@ dpaa_start_xmit(struct sk_buff *skb, struct net_device *net_dev)
 		nonlinear = skb_is_nonlinear(skb);
 	}
 
+#ifdef CONFIG_DPAA_ERRATUM_A050385
+	if (unlikely(fman_has_errata_a050385())) {
+		if (dpaa_a050385_wa(net_dev, &skb))
+			goto enomem;
+		nonlinear = skb_is_nonlinear(skb);
+	}
+#endif
+
 	if (nonlinear) {
 		/* Just create a S/G fd based on the skb */
 		err = skb_to_sg_fd(priv, skb, &fd);
@@ -2741,9 +2845,7 @@ static inline u16 dpaa_get_headroom(struct dpaa_buffer_layout *bl)
 	headroom = (u16)(bl->priv_data_size + DPAA_PARSE_RESULTS_SIZE +
 		DPAA_TIME_STAMP_SIZE + DPAA_HASH_RESULTS_SIZE);
 
-	return DPAA_FD_DATA_ALIGNMENT ? ALIGN(headroom,
-					      DPAA_FD_DATA_ALIGNMENT) :
-					headroom;
+	return ALIGN(headroom, DPAA_FD_DATA_ALIGNMENT);
 }
 
 static int dpaa_eth_probe(struct platform_device *pdev)
diff --git a/drivers/net/ethernet/freescale/fec_main.c b/drivers/net/ethernet/freescale/fec_main.c
index 4432a59904c7..23c5fef2f1ad 100644
--- a/drivers/net/ethernet/freescale/fec_main.c
+++ b/drivers/net/ethernet/freescale/fec_main.c
@@ -2529,15 +2529,15 @@ fec_enet_set_coalesce(struct net_device *ndev, struct ethtool_coalesce *ec)
 		return -EINVAL;
 	}
 
-	cycle = fec_enet_us_to_itr_clock(ndev, fep->rx_time_itr);
+	cycle = fec_enet_us_to_itr_clock(ndev, ec->rx_coalesce_usecs);
 	if (cycle > 0xFFFF) {
 		dev_err(dev, "Rx coalesced usec exceed hardware limitation\n");
 		return -EINVAL;
 	}
 
-	cycle = fec_enet_us_to_itr_clock(ndev, fep->tx_time_itr);
+	cycle = fec_enet_us_to_itr_clock(ndev, ec->tx_coalesce_usecs);
 	if (cycle > 0xFFFF) {
-		dev_err(dev, "Rx coalesced usec exceed hardware limitation\n");
+		dev_err(dev, "Tx coalesced usec exceed hardware limitation\n");
 		return -EINVAL;
 	}
 
diff --git a/drivers/net/ethernet/freescale/fman/Kconfig b/drivers/net/ethernet/freescale/fman/Kconfig
index 0139cb9042ec..34150182cc35 100644
--- a/drivers/net/ethernet/freescale/fman/Kconfig
+++ b/drivers/net/ethernet/freescale/fman/Kconfig
@@ -8,3 +8,31 @@ config FSL_FMAN
 	help
 		Freescale Data-Path Acceleration Architecture Frame Manager
 		(FMan) support
+
+config DPAA_ERRATUM_A050385
+	bool
+	depends on ARM64 && FSL_DPAA
+	default y
+	help
+		DPAA FMan erratum A050385 software workaround implementation:
+		align buffers, data start, SG fragment length to avoid FMan DMA
+		splits.
+		FMAN DMA read or writes under heavy traffic load may cause FMAN
+		internal resource leak thus stopping further packet processing.
+		The FMAN internal queue can overflow when FMAN splits single
+		read or write transactions into multiple smaller transactions
+		such that more than 17 AXI transactions are in flight from FMAN
+		to interconnect. When the FMAN internal queue overflows, it can
+		stall further packet processing. The issue can occur with any
+		one of the following three conditions:
+		1. FMAN AXI transaction crosses 4K address boundary (Errata
+		A010022)
+		2. FMAN DMA address for an AXI transaction is not 16 byte
+		aligned, i.e. the last 4 bits of an address are non-zero
+		3. Scatter Gather (SG) frames have more than one SG buffer in
+		the SG list and any one of the buffers, except the last
+		buffer in the SG list has data size that is not a multiple
+		of 16 bytes, i.e., other than 16, 32, 48, 64, etc.
+		With any one of the above three conditions present, there is
+		likelihood of stalled FMAN packet processing, especially under
+		stress with multiple ports injecting line-rate traffic.
diff --git a/drivers/net/ethernet/freescale/fman/fman.c b/drivers/net/ethernet/freescale/fman/fman.c
index 934111def0be..f151d6e111dd 100644
--- a/drivers/net/ethernet/freescale/fman/fman.c
+++ b/