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

Pull networking updates from David Miller:

 1) Support 6Ghz band in ath11k driver, from Rajkumar Manoharan.

 2) Support UDP segmentation in code TSO code, from Eric Dumazet.

 3) Allow flashing different flash images in cxgb4 driver, from Vishal
    Kulkarni.

 4) Add drop frames counter and flow status to tc flower offloading,
    from Po Liu.

 5) Support n-tuple filters in cxgb4, from Vishal Kulkarni.

 6) Various new indirect call avoidance, from Eric Dumazet and Brian
    Vazquez.

 7) Fix BPF verifier failures on 32-bit pointer arithmetic, from
    Yonghong Song.

 8) Support querying and setting hardware address of a port function via
    devlink, use this in mlx5, from Parav Pandit.

 9) Support hw ipsec offload on bonding slaves, from Jarod Wilson.

10) Switch qca8k driver over to phylink, from Jonathan McDowell.

11) In bpftool, show list of processes holding BPF FD references to
    maps, programs, links, and btf objects. From Andrii Nakryiko.

12) Several conversions over to generic power management, from Vaibhav
    Gupta.

13) Add support for SO_KEEPALIVE et al. to bpf_setsockopt(), from Dmitry
    Yakunin.

14) Various https url conversions, from Alexander A. Klimov.

15) Timestamping and PHC support for mscc PHY driver, from Antoine
    Tenart.

16) Support bpf iterating over tcp and udp sockets, from Yonghong Song.

17) Support 5GBASE-T i40e NICs, from Aleksandr Loktionov.

18) Add kTLS RX HW offload support to mlx5e, from Tariq Toukan.

19) Fix the ->ndo_start_xmit() return type to be netdev_tx_t in several
    drivers. From Luc Van Oostenryck.

20) XDP support for xen-netfront, from Denis Kirjanov.

21) Support receive buffer autotuning in MPTCP, from Florian Westphal.

22) Support EF100 chip in sfc driver, from Edward Cree.

23) Add XDP support to mvpp2 driver, from Matteo Croce.

24) Support MPTCP in sock_diag, from Paolo Abeni.

25) Commonize UDP tunnel offloading code by creating udp_tunnel_nic
    infrastructure, from Jakub Kicinski.

26) Several pci_ --> dma_ API conversions, from Christophe JAILLET.

27) Add FLOW_ACTION_POLICE support to mlxsw, from Ido Schimmel.

28) Add SK_LOOKUP bpf program type, from Jakub Sitnicki.

29) Refactor a lot of networking socket option handling code in order to
    avoid set_fs() calls, from Christoph Hellwig.

30) Add rfc4884 support to icmp code, from Willem de Bruijn.

31) Support TBF offload in dpaa2-eth driver, from Ioana Ciornei.

32) Support XDP_REDIRECT in qede driver, from Alexander Lobakin.

33) Support PCI relaxed ordering in mlx5 driver, from Aya Levin.

34) Support TCP syncookies in MPTCP, from Flowian Westphal.

35) Fix several tricky cases of PMTU handling wrt. briding, from Stefano
    Brivio.

* git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next: (2056 commits)
  net: thunderx: initialize VF's mailbox mutex before first usage
  usb: hso: remove bogus check for EINPROGRESS
  usb: hso: no complaint about kmalloc failure
  hso: fix bailout in error case of probe
  ip_tunnel_core: Fix build for archs without _HAVE_ARCH_IPV6_CSUM
  selftests/net: relax cpu affinity requirement in msg_zerocopy test
  mptcp: be careful on subflow creation
  selftests: rtnetlink: make kci_test_encap() return sub-test result
  selftests: rtnetlink: correct the final return value for the test
  net: dsa: sja1105: use detected device id instead of DT one on mismatch
  tipc: set ub->ifindex for local ipv6 address
  ipv6: add ipv6_dev_find()
  net: openvswitch: silence suspicious RCU usage warning
  Revert "vxlan: fix tos value before xmit"
  ptp: only allow phase values lower than 1 period
  farsync: switch from 'pci_' to 'dma_' API
  wan: wanxl: switch from 'pci_' to 'dma_' API
  hv_netvsc: do not use VF device if link is down
  dpaa2-eth: Fix passing zero to 'PTR_ERR' warning
  net: macb: Properly handle phylink on at91sam9x
  ...

8 files changed, 309 insertions, 54 deletions

diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-debugfs.c b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-debugfs.c
index c453a23045c1..56d9927fbfda 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-debugfs.c
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-debugfs.c
@@ -21,7 +21,7 @@ static int dpaa2_dbg_cpu_show(struct seq_file *file, void *offset)
 	seq_printf(file, "Per-CPU stats for %s\n", priv->net_dev->name);
 	seq_printf(file, "%s%16s%16s%16s%16s%16s%16s%16s%16s%16s\n",
 		   "CPU", "Rx", "Rx Err", "Rx SG", "Tx", "Tx Err", "Tx conf",
-		   "Tx SG", "Tx realloc", "Enq busy");
+		   "Tx SG", "Tx converted to SG", "Enq busy");
 
 	for_each_online_cpu(i) {
 		stats = per_cpu_ptr(priv->percpu_stats, i);
@@ -35,7 +35,7 @@ static int dpaa2_dbg_cpu_show(struct seq_file *file, void *offset)
 			   stats->tx_errors,
 			   extras->tx_conf_frames,
 			   extras->tx_sg_frames,
-			   extras->tx_reallocs,
+			   extras->tx_converted_sg_frames,
 			   extras->tx_portal_busy);
 	}
 
@@ -90,6 +90,10 @@ static int dpaa2_dbg_fqs_show(struct seq_file *file, void *offset)
 		if (err)
 			fcnt = 0;
 
+		/* Skip FQs with no traffic */
+		if (!fq->stats.frames && !fcnt)
+			continue;
+
 		seq_printf(file, "%5d%16d%16d%16s%16llu%16u\n",
 			   fq->fqid,
 			   fq->target_cpu,
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-trace.h b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-trace.h
index 9801528db2a5..5fb5f14e01ec 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-trace.h
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth-trace.h
@@ -10,7 +10,6 @@
 
 #include <linux/skbuff.h>
 #include <linux/netdevice.h>
-#include "dpaa2-eth.h"
 #include <linux/tracepoint.h>
 
 #define TR_FMT "[%s] fd: addr=0x%llx, len=%u, off=%u"
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
index 0998ceb1a26e..457106e761be 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c
@@ -15,6 +15,7 @@
 #include <linux/fsl/mc.h>
 #include <linux/bpf.h>
 #include <linux/bpf_trace.h>
+#include <net/pkt_cls.h>
 #include <net/sock.h>
 
 #include "dpaa2-eth.h"
@@ -611,6 +612,10 @@ static int build_sg_fd(struct dpaa2_eth_priv *priv,
 
 	sg_init_table(scl, nr_frags + 1);
 	num_sg = skb_to_sgvec(skb, scl, 0, skb->len);
+	if (unlikely(num_sg < 0)) {
+		err = -ENOMEM;
+		goto dma_map_sg_failed;
+	}
 	num_dma_bufs = dma_map_sg(dev, scl, num_sg, DMA_BIDIRECTIONAL);
 	if (unlikely(!num_dma_bufs)) {
 		err = -ENOMEM;
@@ -681,6 +686,86 @@ dma_map_sg_failed:
 	return err;
 }
 
+/* Create a SG frame descriptor based on a linear skb.
+ *
+ * This function is used on the Tx path when the skb headroom is not large
+ * enough for the HW requirements, thus instead of realloc-ing the skb we
+ * create a SG frame descriptor with only one entry.
+ */
+static int build_sg_fd_single_buf(struct dpaa2_eth_priv *priv,
+				  struct sk_buff *skb,
+				  struct dpaa2_fd *fd)
+{
+	struct device *dev = priv->net_dev->dev.parent;
+	struct dpaa2_eth_sgt_cache *sgt_cache;
+	struct dpaa2_sg_entry *sgt;
+	struct dpaa2_eth_swa *swa;
+	dma_addr_t addr, sgt_addr;
+	void *sgt_buf = NULL;
+	int sgt_buf_size;
+	int err;
+
+	/* Prepare the HW SGT structure */
+	sgt_cache = this_cpu_ptr(priv->sgt_cache);
+	sgt_buf_size = priv->tx_data_offset + sizeof(struct dpaa2_sg_entry);
+
+	if (sgt_cache->count == 0)
+		sgt_buf = kzalloc(sgt_buf_size + DPAA2_ETH_TX_BUF_ALIGN,
+				  GFP_ATOMIC);
+	else
+		sgt_buf = sgt_cache->buf[--sgt_cache->count];
+	if (unlikely(!sgt_buf))
+		return -ENOMEM;
+
+	sgt_buf = PTR_ALIGN(sgt_buf, DPAA2_ETH_TX_BUF_ALIGN);
+	sgt = (struct dpaa2_sg_entry *)(sgt_buf + priv->tx_data_offset);
+
+	addr = dma_map_single(dev, skb->data, skb->len, DMA_BIDIRECTIONAL);
+	if (unlikely(dma_mapping_error(dev, addr))) {
+		err = -ENOMEM;
+		goto data_map_failed;
+	}
+
+	/* Fill in the HW SGT structure */
+	dpaa2_sg_set_addr(sgt, addr);
+	dpaa2_sg_set_len(sgt, skb->len);
+	dpaa2_sg_set_final(sgt, true);
+
+	/* Store the skb backpointer in the SGT buffer */
+	swa = (struct dpaa2_eth_swa *)sgt_buf;
+	swa->type = DPAA2_ETH_SWA_SINGLE;
+	swa->single.skb = skb;
+	swa->sg.sgt_size = sgt_buf_size;
+
+	/* Separately map the SGT buffer */
+	sgt_addr = dma_map_single(dev, sgt_buf, sgt_buf_size, DMA_BIDIRECTIONAL);
+	if (unlikely(dma_mapping_error(dev, sgt_addr))) {
+		err = -ENOMEM;
+		goto sgt_map_failed;
+	}
+
+	dpaa2_fd_set_offset(fd, priv->tx_data_offset);
+	dpaa2_fd_set_format(fd, dpaa2_fd_sg);
+	dpaa2_fd_set_addr(fd, sgt_addr);
+	dpaa2_fd_set_len(fd, skb->len);
+	dpaa2_fd_set_ctrl(fd, FD_CTRL_PTA);
+
+	if (priv->tx_tstamp && skb_shinfo(skb)->tx_flags & SKBTX_HW_TSTAMP)
+		enable_tx_tstamp(fd, sgt_buf);
+
+	return 0;
+
+sgt_map_failed:
+	dma_unmap_single(dev, addr, skb->len, DMA_BIDIRECTIONAL);
+data_map_failed:
+	if (sgt_cache->count >= DPAA2_ETH_SGT_CACHE_SIZE)
+		kfree(sgt_buf);
+	else
+		sgt_cache->buf[sgt_cache->count++] = sgt_buf;
+
+	return err;
+}
+
 /* Create a frame descriptor based on a linear skb */
 static int build_single_fd(struct dpaa2_eth_priv *priv,
 			   struct sk_buff *skb,
@@ -739,13 +824,16 @@ static void free_tx_fd(const struct dpaa2_eth_priv *priv,
 		       const struct dpaa2_fd *fd, bool in_napi)
 {
 	struct device *dev = priv->net_dev->dev.parent;
-	dma_addr_t fd_addr;
+	dma_addr_t fd_addr, sg_addr;
 	struct sk_buff *skb = NULL;
 	unsigned char *buffer_start;
 	struct dpaa2_eth_swa *swa;
 	u8 fd_format = dpaa2_fd_get_format(fd);
 	u32 fd_len = dpaa2_fd_get_len(fd);
 
+	struct dpaa2_eth_sgt_cache *sgt_cache;
+	struct dpaa2_sg_entry *sgt;
+
 	fd_addr = dpaa2_fd_get_addr(fd);
 	buffer_start = dpaa2_iova_to_virt(priv->iommu_domain, fd_addr);
 	swa = (struct dpaa2_eth_swa *)buffer_start;
@@ -765,16 +853,29 @@ static void free_tx_fd(const struct dpaa2_eth_priv *priv,
 					 DMA_BIDIRECTIONAL);
 		}
 	} else if (fd_format == dpaa2_fd_sg) {
-		skb = swa->sg.skb;
+		if (swa->type == DPAA2_ETH_SWA_SG) {
+			skb = swa->sg.skb;
 
-		/* Unmap the scatterlist */
-		dma_unmap_sg(dev, swa->sg.scl, swa->sg.num_sg,
-			     DMA_BIDIRECTIONAL);
-		kfree(swa->sg.scl);
+			/* Unmap the scatterlist */
+			dma_unmap_sg(dev, swa->sg.scl, swa->sg.num_sg,
+				     DMA_BIDIRECTIONAL);
+			kfree(swa->sg.scl);
 
-		/* Unmap the SGT buffer */
-		dma_unmap_single(dev, fd_addr, swa->sg.sgt_size,
-				 DMA_BIDIRECTIONAL);
+			/* Unmap the SGT buffer */
+			dma_unmap_single(dev, fd_addr, swa->sg.sgt_size,
+					 DMA_BIDIRECTIONAL);
+		} else {
+			skb = swa->single.skb;
+
+			/* Unmap the SGT Buffer */
+			dma_unmap_single(dev, fd_addr, swa->single.sgt_size,
+					 DMA_BIDIRECTIONAL);
+
+			sgt = (struct dpaa2_sg_entry *)(buffer_start +
+							priv->tx_data_offset);
+			sg_addr = dpaa2_sg_get_addr(sgt);
+			dma_unmap_single(dev, sg_addr, skb->len, DMA_BIDIRECTIONAL);
+		}
 	} else {
 		netdev_dbg(priv->net_dev, "Invalid FD format\n");
 		return;
@@ -804,8 +905,17 @@ static void free_tx_fd(const struct dpaa2_eth_priv *priv,
 	}
 
 	/* Free SGT buffer allocated on tx */
-	if (fd_format != dpaa2_fd_single)
-		skb_free_frag(buffer_start);
+	if (fd_format != dpaa2_fd_single) {
+		sgt_cache = this_cpu_ptr(priv->sgt_cache);
+		if (swa->type == DPAA2_ETH_SWA_SG) {
+			skb_free_frag(buffer_start);
+		} else {
+			if (sgt_cache->count >= DPAA2_ETH_SGT_CACHE_SIZE)
+				kfree(buffer_start);
+			else
+				sgt_cache->buf[sgt_cache->count++] = buffer_start;
+		}
+	}
 
 	/* Move on with skb release */
 	napi_consume_skb(skb, in_napi);
@@ -829,22 +939,6 @@ static netdev_tx_t dpaa2_eth_tx(struct sk_buff *skb, struct net_device *net_dev)
 	percpu_extras = this_cpu_ptr(priv->percpu_extras);
 
 	needed_headroom = dpaa2_eth_needed_headroom(priv, skb);
-	if (skb_headroom(skb) < needed_headroom) {
-		struct sk_buff *ns;
-
-		ns = skb_realloc_headroom(skb, needed_headroom);
-		if (unlikely(!ns)) {
-			percpu_stats->tx_dropped++;
-			goto err_alloc_headroom;
-		}
-		percpu_extras->tx_reallocs++;
-
-		if (skb->sk)
-			skb_set_owner_w(ns, skb->sk);
-
-		dev_kfree_skb(skb);
-		skb = ns;
-	}
 
 	/* We'll be holding a back-reference to the skb until Tx Confirmation;
 	 * we don't want that overwritten by a concurrent Tx with a cloned skb.
@@ -863,6 +957,12 @@ static netdev_tx_t dpaa2_eth_tx(struct sk_buff *skb, struct net_device *net_dev)
 		err = build_sg_fd(priv, skb, &fd);
 		percpu_extras->tx_sg_frames++;
 		percpu_extras->tx_sg_bytes += skb->len;
+	} else if (skb_headroom(skb) < needed_headroom) {
+		err = build_sg_fd_single_buf(priv, skb, &fd);
+		percpu_extras->tx_sg_frames++;
+		percpu_extras->tx_sg_bytes += skb->len;
+		percpu_extras->tx_converted_sg_frames++;
+		percpu_extras->tx_converted_sg_bytes += skb->len;
 	} else {
 		err = build_single_fd(priv, skb, &fd);
 	}
@@ -920,7 +1020,6 @@ static netdev_tx_t dpaa2_eth_tx(struct sk_buff *skb, struct net_device *net_dev)
 	return NETDEV_TX_OK;
 
 err_build_fd:
-err_alloc_headroom:
 	dev_kfree_skb(skb);
 
 	return NETDEV_TX_OK;
@@ -1109,7 +1208,7 @@ static void drain_bufs(struct dpaa2_eth_priv *priv, int count)
 					       buf_array, count);
 		if (ret < 0) {
 			if (ret == -EBUSY &&
-			    retries++ >= DPAA2_ETH_SWP_BUSY_RETRIES)
+			    retries++ < DPAA2_ETH_SWP_BUSY_RETRIES)
 				continue;
 			netdev_err(priv->net_dev, "dpaa2_io_service_acquire() failed\n");
 			return;
@@ -1157,6 +1256,22 @@ static int refill_pool(struct dpaa2_eth_priv *priv,
 	return 0;
 }
 
+static void dpaa2_eth_sgt_cache_drain(struct dpaa2_eth_priv *priv)
+{
+	struct dpaa2_eth_sgt_cache *sgt_cache;
+	u16 count;
+	int k, i;
+
+	for_each_possible_cpu(k) {
+		sgt_cache = per_cpu_ptr(priv->sgt_cache, k);
+		count = sgt_cache->count;
+
+		for (i = 0; i < count; i++)
+			kfree(sgt_cache->buf[i]);
+		sgt_cache->count = 0;
+	}
+}
+
 static int pull_channel(struct dpaa2_eth_channel *ch)
 {
 	int err;
@@ -1558,6 +1673,9 @@ static int dpaa2_eth_stop(struct net_device *net_dev)
 	/* Empty the buffer pool */
 	drain_pool(priv);
 
+	/* Empty the Scatter-Gather Buffer cache */
+	dpaa2_eth_sgt_cache_drain(priv);
+
 	return 0;
 }
 
@@ -1959,14 +2077,9 @@ out_err:
 
 static int dpaa2_eth_xdp(struct net_device *dev, struct netdev_bpf *xdp)
 {
-	struct dpaa2_eth_priv *priv = netdev_priv(dev);
-
 	switch (xdp->command) {
 	case XDP_SETUP_PROG:
 		return setup_xdp(dev, xdp->prog);
-	case XDP_QUERY_PROG:
-		xdp->prog_id = priv->xdp_prog ? priv->xdp_prog->aux->id : 0;
-		break;
 	default:
 		return -EINVAL;
 	}
@@ -2096,17 +2209,13 @@ static int update_xps(struct dpaa2_eth_priv *priv)
 	return err;
 }
 
-static int dpaa2_eth_setup_tc(struct net_device *net_dev,
-			      enum tc_setup_type type, void *type_data)
+static int dpaa2_eth_setup_mqprio(struct net_device *net_dev,
+				  struct tc_mqprio_qopt *mqprio)
 {
 	struct dpaa2_eth_priv *priv = netdev_priv(net_dev);
-	struct tc_mqprio_qopt *mqprio = type_data;
 	u8 num_tc, num_queues;
 	int i;
 
-	if (type != TC_SETUP_QDISC_MQPRIO)
-		return -EOPNOTSUPP;
-
 	mqprio->hw = TC_MQPRIO_HW_OFFLOAD_TCS;
 	num_queues = dpaa2_eth_queue_count(priv);
 	num_tc = mqprio->num_tc;
@@ -2138,6 +2247,60 @@ out:
 	return 0;
 }
 
+#define bps_to_mbits(rate) (div_u64((rate), 1000000) * 8)
+
+static int dpaa2_eth_setup_tbf(struct net_device *net_dev, struct tc_tbf_qopt_offload *p)
+{
+	struct tc_tbf_qopt_offload_replace_params *cfg = &p->replace_params;
+	struct dpaa2_eth_priv *priv = netdev_priv(net_dev);
+	struct dpni_tx_shaping_cfg tx_cr_shaper = { 0 };
+	struct dpni_tx_shaping_cfg tx_er_shaper = { 0 };
+	int err;
+
+	if (p->command == TC_TBF_STATS)
+		return -EOPNOTSUPP;
+
+	/* Only per port Tx shaping */
+	if (p->parent != TC_H_ROOT)
+		return -EOPNOTSUPP;
+
+	if (p->command == TC_TBF_REPLACE) {
+		if (cfg->max_size > DPAA2_ETH_MAX_BURST_SIZE) {
+			netdev_err(net_dev, "burst size cannot be greater than %d\n",
+				   DPAA2_ETH_MAX_BURST_SIZE);
+			return -EINVAL;
+		}
+
+		tx_cr_shaper.max_burst_size = cfg->max_size;
+		/* The TBF interface is in bytes/s, whereas DPAA2 expects the
+		 * rate in Mbits/s
+		 */
+		tx_cr_shaper.rate_limit = bps_to_mbits(cfg->rate.rate_bytes_ps);
+	}
+
+	err = dpni_set_tx_shaping(priv->mc_io, 0, priv->mc_token, &tx_cr_shaper,
+				  &tx_er_shaper, 0);
+	if (err) {
+		netdev_err(net_dev, "dpni_set_tx_shaping() = %d\n", err);
+		return err;
+	}
+
+	return 0;
+}
+
+static int dpaa2_eth_setup_tc(struct net_device *net_dev,
+			      enum tc_setup_type type, void *type_data)
+{
+	switch (type) {
+	case TC_SETUP_QDISC_MQPRIO:
+		return dpaa2_eth_setup_mqprio(net_dev, type_data);
+	case TC_SETUP_QDISC_TBF:
+		return dpaa2_eth_setup_tbf(net_dev, type_data);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
 static const struct net_device_ops dpaa2_eth_ops = {
 	.ndo_open = dpaa2_eth_open,
 	.ndo_start_xmit = dpaa2_eth_tx,
@@ -2162,7 +2325,7 @@ static void cdan_cb(struct dpaa2_io_notification_ctx *ctx)
 	/* Update NAPI statistics */
 	ch->stats.cdan++;
 
-	napi_schedule_irqoff(&ch->napi);
+	napi_schedule(&ch->napi);
 }
 
 /* Allocate and configure a DPCON object */
@@ -2207,7 +2370,7 @@ close:
 free:
 	fsl_mc_object_free(dpcon);
 
-	return NULL;
+	return ERR_PTR(err);
 }
 
 static void free_dpcon(struct dpaa2_eth_priv *priv,
@@ -2231,8 +2394,8 @@ alloc_channel(struct dpaa2_eth_priv *priv)
 		return NULL;
 
 	channel->dpcon = setup_dpcon(priv);
-	if (IS_ERR_OR_NULL(channel->dpcon)) {
-		err = PTR_ERR_OR_ZERO(channel->dpcon);
+	if (IS_ERR(channel->dpcon)) {
+		err = PTR_ERR(channel->dpcon);
 		goto err_setup;
 	}
 
@@ -3602,7 +3765,7 @@ static int netdev_init(struct net_device *net_dev)
 	net_dev->features = NETIF_F_RXCSUM |
 			    NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM |
 			    NETIF_F_SG | NETIF_F_HIGHDMA |
-			    NETIF_F_LLTX;
+			    NETIF_F_LLTX | NETIF_F_HW_TC;
 	net_dev->hw_features = net_dev->features;
 
 	return 0;
@@ -3842,6 +4005,13 @@ static int dpaa2_eth_probe(struct fsl_mc_device *dpni_dev)
 		goto err_alloc_percpu_extras;
 	}
 
+	priv->sgt_cache = alloc_percpu(*priv->sgt_cache);
+	if (!priv->sgt_cache) {
+		dev_err(dev, "alloc_percpu(sgt_cache) failed\n");
+		err = -ENOMEM;
+		goto err_alloc_sgt_cache;
+	}
+
 	err = netdev_init(net_dev);
 	if (err)
 		goto err_netdev_init;
@@ -3910,6 +4080,8 @@ err_poll_thread:
 err_alloc_rings:
 err_csum:
 err_netdev_init:
+	free_percpu(priv->sgt_cache);
+err_alloc_sgt_cache:
 	free_percpu(priv->percpu_extras);
 err_alloc_percpu_extras:
 	free_percpu(priv->percpu_stats);
@@ -3955,6 +4127,7 @@ static int dpaa2_eth_remove(struct fsl_mc_device *ls_dev)
 		fsl_mc_free_irqs(ls_dev);
 
 	free_rings(priv);
+	free_percpu(priv->sgt_cache);
 	free_percpu(priv->percpu_stats);
 	free_percpu(priv->percpu_extras);
 
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
index 2d7ada0f0dbd..7f3c41dc98f2 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.h
@@ -43,6 +43,9 @@
  */
 #define DPAA2_ETH_FQ_TAILDROP_THRESH	(1024 * 1024)
 
+/* Maximum burst size value for Tx shaping */
+#define DPAA2_ETH_MAX_BURST_SIZE	0xF7FF
+
 /* Maximum number of Tx confirmation frames to be processed
  * in a single NAPI call
  */
@@ -125,6 +128,7 @@ struct dpaa2_eth_swa {
 	union {
 		struct {
 			struct sk_buff *skb;
+			int sgt_size;
 		} single;
 		struct {
 			struct sk_buff *skb;
@@ -282,9 +286,11 @@ struct dpaa2_eth_drv_stats {
 	__u64	tx_conf_bytes;
 	__u64	tx_sg_frames;
 	__u64	tx_sg_bytes;
-	__u64	tx_reallocs;
 	__u64	rx_sg_frames;
 	__u64	rx_sg_bytes;
+	/* Linear skbs sent as a S/G FD due to insufficient headroom */
+	__u64	tx_converted_sg_frames;
+	__u64	tx_converted_sg_bytes;
 	/* Enqueues retried due to portal busy */
 	__u64	tx_portal_busy;
 };
@@ -395,6 +401,12 @@ struct dpaa2_eth_cls_rule {
 	u8 in_use;
 };
 
+#define DPAA2_ETH_SGT_CACHE_SIZE	256
+struct dpaa2_eth_sgt_cache {
+	void *buf[DPAA2_ETH_SGT_CACHE_SIZE];
+	u16 count;
+};
+
 /* Driver private data */
 struct dpaa2_eth_priv {
 	struct net_device *net_dev;
@@ -409,6 +421,7 @@ struct dpaa2_eth_priv {
 
 	u8 num_channels;
 	struct dpaa2_eth_channel *channel[DPAA2_ETH_MAX_DPCONS];
+	struct dpaa2_eth_sgt_cache __percpu *sgt_cache;
 
 	struct dpni_attr dpni_attrs;
 	u16 dpni_ver_major;
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-ethtool.c b/drivers/net/ethernet/freescale/dpaa2/dpaa2-ethtool.c
index e88269fe3de7..8356f1fbbee1 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-ethtool.c
+++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-ethtool.c
@@ -43,9 +43,10 @@ static char dpaa2_ethtool_extras[][ETH_GSTRING_LEN] = {
 	"[drv] tx conf bytes",
 	"[drv] tx sg frames",
 	"[drv] tx sg bytes",
-	"[drv] tx realloc frames",
 	"[drv] rx sg frames",
 	"[drv] rx sg bytes",
+	"[drv] tx converted sg frames",
+	"[drv] tx converted sg bytes",
 	"[drv] enqueue portal busy",
 	/* Channel stats */
 	"[drv] dequeue portal busy",
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpni-cmd.h b/drivers/net/ethernet/freescale/dpaa2/dpni-cmd.h
index fd069f67be9b..593e3812af93 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpni-cmd.h
+++ b/drivers/net/ethernet/freescale/dpaa2/dpni-cmd.h
@@ -626,4 +626,17 @@ struct dpni_cmd_set_congestion_notification {
 	__le32 threshold_exit;
 };
 
+#define DPNI_COUPLED_SHIFT	0
+#define DPNI_COUPLED_SIZE	1
+
+struct dpni_cmd_set_tx_shaping {
+	__le16 tx_cr_max_burst_size;
+	__le16 tx_er_max_burst_size;
+	__le32 pad;
+	__le32 tx_cr_rate_limit;
+	__le32 tx_er_rate_limit;
+	/* from LSB: coupled:1 */
+	u8 coupled;
+};
+
 #endif /* _FSL_DPNI_CMD_H */
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpni.c b/drivers/net/ethernet/freescale/dpaa2/dpni.c
index 6b479ba66465..68ed4c41b282 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpni.c
+++ b/drivers/net/ethernet/freescale/dpaa2/dpni.c
@@ -1558,10 +1558,10 @@ int dpni_get_statistics(struct fsl_mc_io *mc_io,
  * @cmd_flags:	Command flags; one or more of 'MC_CMD_FLAG_'
  * @token:	Token of DPNI object
  * @cg_point:	Congestion point
- * @q_type:	Queue type on which the taildrop is configured.
+ * @qtype:	Queue type on which the taildrop is configured.
  *		Only Rx queues are supported for now
  * @tc:		Traffic class to apply this taildrop to
- * @q_index:	Index of the queue if the DPNI supports multiple queues for
+ * @index:	Index of the queue if the DPNI supports multiple queues for
  *		traffic distribution. Ignored if CONGESTION_POINT is not 0.
  * @taildrop:	Taildrop structure
  *
@@ -1602,10 +1602,10 @@ int dpni_set_taildrop(struct fsl_mc_io *mc_io,
  * @cmd_flags:	Command flags; one or more of 'MC_CMD_FLAG_'
  * @token:	Token of DPNI object
  * @cg_point:	Congestion point
- * @q_type:	Queue type on which the taildrop is configured.
+ * @qtype:	Queue type on which the taildrop is configured.
  *		Only Rx queues are supported for now
  * @tc:		Traffic class to apply this taildrop to
- * @q_index:	Index of the queue if the DPNI supports multiple queues for
+ * @index:	Index of the queue if the DPNI supports multiple queues for
  *		traffic distribution. Ignored if CONGESTION_POINT is not 0.
  * @taildrop:	Taildrop structure
  *
@@ -1963,3 +1963,39 @@ int dpni_clear_qos_table(struct fsl_mc_io *mc_io,
 	/* send command to mc*/
 	return mc_send_command(mc_io, &cmd);
 }
+
+/**
+ * dpni_set_tx_shaping() - Set the transmit shaping
+ * @mc_io:		Pointer to MC portal's I/O object
+ * @cmd_flags:		Command flags; one or more of 'MC_CMD_FLAG_'
+ * @token:		Token of DPNI object
+ * @tx_cr_shaper:	TX committed rate shaping configuration
+ * @tx_er_shaper:	TX excess rate shaping configuration
+ * @coupled:		Committed and excess rate shapers are coupled
+ *
+ * Return:	'0' on Success; Error code otherwise.
+ */
+int dpni_set_tx_shaping(struct fsl_mc_io *mc_io,
+			u32 cmd_flags,
+			u16 token,
+			const struct dpni_tx_shaping_cfg *tx_cr_shaper,
+			const struct dpni_tx_shaping_cfg *tx_er_shaper,
+			int coupled)
+{
+	struct dpni_cmd_set_tx_shaping *cmd_params;
+	struct fsl_mc_command cmd = { 0 };
+
+	/* prepare command */
+	cmd.header = mc_encode_cmd_header(DPNI_CMDID_SET_TX_SHAPING,
+					  cmd_flags,
+					  token);
+	cmd_params = (struct dpni_cmd_set_tx_shaping *)cmd.params;
+	cmd_params->tx_cr_max_burst_size = cpu_to_le16(tx_cr_shaper->max_burst_size);
+	cmd_params->tx_er_max_burst_size = cpu_to_le16(tx_er_shaper->max_burst_size);
+	cmd_params->tx_cr_rate_limit = cpu_to_le32(tx_cr_shaper->rate_limit);
+	cmd_params->tx_er_rate_limit = cpu_to_le32(tx_er_shaper->rate_limit);
+	dpni_set_field(cmd_params->coupled, COUPLED, coupled);
+
+	/* send command to mc*/
+	return mc_send_command(mc_io, &cmd);
+}
diff --git a/drivers/net/ethernet/freescale/dpaa2/dpni.h b/drivers/net/ethernet/freescale/dpaa2/dpni.h
index e874d8084142..39387991a1f9 100644
--- a/drivers/net/ethernet/freescale/dpaa2/dpni.h
+++ b/drivers/net/ethernet/freescale/dpaa2/dpni.h
@@ -1062,5 +1062,21 @@ int dpni_get_api_version(struct fsl_mc_io *mc_io,
 			 u32 cmd_flags,
 			 u16 *major_ver,
 			 u16 *minor_ver);
+/**
+ * struct dpni_tx_shaping - Structure representing DPNI tx shaping configuration
+ * @rate_limit:		Rate in Mbps
+ * @max_burst_size:	Burst size in bytes (up to 64KB)
+ */
+struct dpni_tx_shaping_cfg {
+	u32 rate_limit;
+	u16 max_burst_size;
+};
+
+int dpni_set_tx_shaping(struct fsl_mc_io *mc_io,
+			u32 cmd_flags,
+			u16 token,
+			const struct dpni_tx_shaping_cfg *tx_cr_shaper,
+			const struct dpni_tx_shaping_cfg *tx_er_shaper,
+			int coupled);
 
 #endif /* __FSL_DPNI_H */