Merge tag 'net-next-6.3' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-nextgrafted

Pull networking updates from Jakub Kicinski:
 "Core:

   - Add dedicated kmem_cache for typical/small skb->head, avoid having
     to access struct page at kfree time, and improve memory use.

   - Introduce sysctl to set default RPS configuration for new netdevs.

   - Define Netlink protocol specification format which can be used to
     describe messages used by each family and auto-generate parsers.
     Add tools for generating kernel data structures and uAPI headers.

   - Expose all net/core sysctls inside netns.

   - Remove 4s sleep in netpoll if carrier is instantly detected on
     boot.

   - Add configurable limit of MDB entries per port, and port-vlan.

   - Continue populating drop reasons throughout the stack.

   - Retire a handful of legacy Qdiscs and classifiers.

  Protocols:

   - Support IPv4 big TCP (TSO frames larger than 64kB).

   - Add IP_LOCAL_PORT_RANGE socket option, to control local port range
     on socket by socket basis.

   - Track and report in procfs number of MPTCP sockets used.

   - Support mixing IPv4 and IPv6 flows in the in-kernel MPTCP path
     manager.

   - IPv6: don't check net.ipv6.route.max_size and rely on garbage
     collection to free memory (similarly to IPv4).

   - Support Penultimate Segment Pop (PSP) flavor in SRv6 (RFC8986).

   - ICMP: add per-rate limit counters.

   - Add support for user scanning requests in ieee802154.

   - Remove static WEP support.

   - Support minimal Wi-Fi 7 Extremely High Throughput (EHT) rate
     reporting.

   - WiFi 7 EHT channel puncturing support (client & AP).

  BPF:

   - Add a rbtree data structure following the "next-gen data structure"
     precedent set by recently added linked list, that is, by using
     kfunc + kptr instead of adding a new BPF map type.

   - Expose XDP hints via kfuncs with initial support for RX hash and
     timestamp metadata.

   - Add BPF_F_NO_TUNNEL_KEY extension to bpf_skb_set_tunnel_key to
     better support decap on GRE tunnel devices not operating in collect
     metadata.

   - Improve x86 JIT's codegen for PROBE_MEM runtime error checks.

   - Remove the need for trace_printk_lock for bpf_trace_printk and
     bpf_trace_vprintk helpers.

   - Extend libbpf's bpf_tracing.h support for tracing arguments of
     kprobes/uprobes and syscall as a special case.

   - Significantly reduce the search time for module symbols by
     livepatch and BPF.

   - Enable cpumasks to be used as kptrs, which is useful for tracing
     programs tracking which tasks end up running on which CPUs in
     different time intervals.

   - Add support for BPF trampoline on s390x and riscv64.

   - Add capability to export the XDP features supported by the NIC.

   - Add __bpf_kfunc tag for marking kernel functions as kfuncs.

   - Add cgroup.memory=nobpf kernel parameter option to disable BPF
     memory accounting for container environments.

  Netfilter:

   - Remove the CLUSTERIP target. It has been marked as obsolete for
     years, and we still have WARN splats wrt races of the out-of-band
     /proc interface installed by this target.

   - Add 'destroy' commands to nf_tables. They are identical to the
     existing 'delete' commands, but do not return an error if the
     referenced object (set, chain, rule...) did not exist.

  Driver API:

   - Improve cpumask_local_spread() locality to help NICs set the right
     IRQ affinity on AMD platforms.

   - Separate C22 and C45 MDIO bus transactions more clearly.

   - Introduce new DCB table to control DSCP rewrite on egress.

   - Support configuration of Physical Layer Collision Avoidance (PLCA)
     Reconciliation Sublayer (RS) (802.3cg-2019). Modern version of
     shared medium Ethernet.

   - Support for MAC Merge layer (IEEE 802.3-2018 clause 99). Allowing
     preemption of low priority frames by high priority frames.

   - Add support for controlling MACSec offload using netlink SET.

   - Rework devlink instance refcounts to allow registration and
     de-registration under the instance lock. Split the code into
     multiple files, drop some of the unnecessarily granular locks and
     factor out common parts of netlink operation handling.

   - Add TX frame aggregation parameters (for USB drivers).

   - Add a new attr TCA_EXT_WARN_MSG to report TC (offload) warning
     messages with notifications for debug.

   - Allow offloading of UDP NEW connections via act_ct.

   - Add support for per action HW stats in TC.

   - Support hardware miss to TC action (continue processing in SW from
     a specific point in the action chain).

   - Warn if old Wireless Extension user space interface is used with
     modern cfg80211/mac80211 drivers. Do not support Wireless
     Extensions for Wi-Fi 7 devices at all. Everyone should switch to
     using nl80211 interface instead.

   - Improve the CAN bit timing configuration. Use extack to return
     error messages directly to user space, update the SJW handling,
     including the definition of a new default value that will benefit
     CAN-FD controllers, by increasing their oscillator tolerance.

  New hardware / drivers:

   - Ethernet:
      - nVidia BlueField-3 support (control traffic driver)
      - Ethernet support for imx93 SoCs
      - Motorcomm yt8531 gigabit Ethernet PHY
      - onsemi NCN26000 10BASE-T1S PHY (with support for PLCA)
      - Microchip LAN8841 PHY (incl. cable diagnostics and PTP)
      - Amlogic gxl MDIO mux

   - WiFi:
      - RealTek RTL8188EU (rtl8xxxu)
      - Qualcomm Wi-Fi 7 devices (ath12k)

   - CAN:
      - Renesas R-Car V4H

  Drivers:

   - Bluetooth:
      - Set Per Platform Antenna Gain (PPAG) for Intel controllers.

   - Ethernet NICs:
      - Intel (1G, igc):
         - support TSN / Qbv / packet scheduling features of i226 model
      - Intel (100G, ice):
         - use GNSS subsystem instead of TTY
         - multi-buffer XDP support
         - extend support for GPIO pins to E823 devices
      - nVidia/Mellanox:
         - update the shared buffer configuration on PFC commands
         - implement PTP adjphase function for HW offset control
         - TC support for Geneve and GRE with VF tunnel offload
         - more efficient crypto key management method
         - multi-port eswitch support
      - Netronome/Corigine:
         - add DCB IEEE support
         - support IPsec offloading for NFP3800
      - Freescale/NXP (enetc):
         - support XDP_REDIRECT for XDP non-linear buffers
         - improve reconfig, avoid link flap and waiting for idle
         - support MAC Merge layer
      - Other NICs:
         - sfc/ef100: add basic devlink support for ef100
         - ionic: rx_push mode operation (writing descriptors via MMIO)
         - bnxt: use the auxiliary bus abstraction for RDMA
         - r8169: disable ASPM and reset bus in case of tx timeout
         - cpsw: support QSGMII mode for J721e CPSW9G
         - cpts: support pulse-per-second output
         - ngbe: add an mdio bus driver
         - usbnet: optimize usbnet_bh() by avoiding unnecessary queuing
         - r8152: handle devices with FW with NCM support
         - amd-xgbe: support 10Mbps, 2.5GbE speeds and rx-adaptation
         - virtio-net: support multi buffer XDP
         - virtio/vsock: replace virtio_vsock_pkt with sk_buff
         - tsnep: XDP support

   - Ethernet high-speed switches:
      - nVidia/Mellanox (mlxsw):
         - add support for latency TLV (in FW control messages)
      - Microchip (sparx5):
         - separate explicit and implicit traffic forwarding rules, make
           the implicit rules always active
         - add support for egress DSCP rewrite
         - IS0 VCAP support (Ingress Classification)
         - IS2 VCAP filters (protos, L3 addrs, L4 ports, flags, ToS
           etc.)
         - ES2 VCAP support (Egress Access Control)
         - support for Per-Stream Filtering and Policing (802.1Q,
           8.6.5.1)

   - Ethernet embedded switches:
      - Marvell (mv88e6xxx):
         - add MAB (port auth) offload support
         - enable PTP receive for mv88e6390
      - NXP (ocelot):
         - support MAC Merge layer
         - support for the the vsc7512 internal copper phys
      - Microchip:
         - lan9303: convert to PHYLINK
         - lan966x: support TC flower filter statistics
         - lan937x: PTP support for KSZ9563/KSZ8563 and LAN937x
         - lan937x: support Credit Based Shaper configuration
         - ksz9477: support Energy Efficient Ethernet
      - other:
         - qca8k: convert to regmap read/write API, use bulk operations
         - rswitch: Improve TX timestamp accuracy

   - Intel WiFi (iwlwifi):
      - EHT (Wi-Fi 7) rate reporting
      - STEP equalizer support: transfer some STEP (connection to radio
        on platforms with integrated wifi) related parameters from the
        BIOS to the firmware.

   - Qualcomm 802.11ax WiFi (ath11k):
      - IPQ5018 support
      - Fine Timing Measurement (FTM) responder role support
      - channel 177 support

   - MediaTek WiFi (mt76):
      - per-PHY LED support
      - mt7996: EHT (Wi-Fi 7) support
      - Wireless Ethernet Dispatch (WED) reset support
      - switch to using page pool allocator

   - RealTek WiFi (rtw89):
      - support new version of Bluetooth co-existance

   - Mobile:
      - rmnet: support TX aggregation"

* tag 'net-next-6.3' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next: (1872 commits)
  page_pool: add a comment explaining the fragment counter usage
  net: ethtool: fix __ethtool_dev_mm_supported() implementation
  ethtool: pse-pd: Fix double word in comments
  xsk: add linux/vmalloc.h to xsk.c
  sefltests: netdevsim: wait for devlink instance after netns removal
  selftest: fib_tests: Always cleanup before exit
  net/mlx5e: Align IPsec ASO result memory to be as required by hardware
  net/mlx5e: TC, Set CT miss to the specific ct action instance
  net/mlx5e: Rename CHAIN_TO_REG to MAPPED_OBJ_TO_REG
  net/mlx5: Refactor tc miss handling to a single function
  net/mlx5: Kconfig: Make tc offload depend on tc skb extension
  net/sched: flower: Support hardware miss to tc action
  net/sched: flower: Move filter handle initialization earlier
  net/sched: cls_api: Support hardware miss to tc action
  net/sched: Rename user cookie and act cookie
  sfc: fix builds without CONFIG_RTC_LIB
  sfc: clean up some inconsistent indentings
  net/mlx4_en: Introduce flexible array to silence overflow warning
  net: lan966x: Fix possible deadlock inside PTP
  net/ulp: Remove redundant ->clone() test in inet_clone_ulp().
  ...

1 files changed, 893 insertions, 0 deletions

diff --git a/drivers/net/ethernet/mscc/ocelot_fdma.c b/drivers/net/ethernet/mscc/ocelot_fdma.c
new file mode 100644
index 000000000..8e3894cf5
--- /dev/null
+++ b/drivers/net/ethernet/mscc/ocelot_fdma.c
@@ -0,0 +1,893 @@
+// SPDX-License-Identifier: (GPL-2.0 OR MIT)
+/*
+ * Microsemi SoCs FDMA driver
+ *
+ * Copyright (c) 2021 Microchip
+ *
+ * Page recycling code is mostly taken from gianfar driver.
+ */
+
+#include <linux/align.h>
+#include <linux/bitops.h>
+#include <linux/dmapool.h>
+#include <linux/dsa/ocelot.h>
+#include <linux/netdevice.h>
+#include <linux/of_platform.h>
+#include <linux/skbuff.h>
+
+#include "ocelot_fdma.h"
+#include "ocelot_qs.h"
+
+DEFINE_STATIC_KEY_FALSE(ocelot_fdma_enabled);
+
+static void ocelot_fdma_writel(struct ocelot *ocelot, u32 reg, u32 data)
+{
+	regmap_write(ocelot->targets[FDMA], reg, data);
+}
+
+static u32 ocelot_fdma_readl(struct ocelot *ocelot, u32 reg)
+{
+	u32 retval;
+
+	regmap_read(ocelot->targets[FDMA], reg, &retval);
+
+	return retval;
+}
+
+static dma_addr_t ocelot_fdma_idx_dma(dma_addr_t base, u16 idx)
+{
+	return base + idx * sizeof(struct ocelot_fdma_dcb);
+}
+
+static u16 ocelot_fdma_dma_idx(dma_addr_t base, dma_addr_t dma)
+{
+	return (dma - base) / sizeof(struct ocelot_fdma_dcb);
+}
+
+static u16 ocelot_fdma_idx_next(u16 idx, u16 ring_sz)
+{
+	return unlikely(idx == ring_sz - 1) ? 0 : idx + 1;
+}
+
+static u16 ocelot_fdma_idx_prev(u16 idx, u16 ring_sz)
+{
+	return unlikely(idx == 0) ? ring_sz - 1 : idx - 1;
+}
+
+static int ocelot_fdma_rx_ring_free(struct ocelot_fdma *fdma)
+{
+	struct ocelot_fdma_rx_ring *rx_ring = &fdma->rx_ring;
+
+	if (rx_ring->next_to_use >= rx_ring->next_to_clean)
+		return OCELOT_FDMA_RX_RING_SIZE -
+		       (rx_ring->next_to_use - rx_ring->next_to_clean) - 1;
+	else
+		return rx_ring->next_to_clean - rx_ring->next_to_use - 1;
+}
+
+static int ocelot_fdma_tx_ring_free(struct ocelot_fdma *fdma)
+{
+	struct ocelot_fdma_tx_ring *tx_ring = &fdma->tx_ring;
+
+	if (tx_ring->next_to_use >= tx_ring->next_to_clean)
+		return OCELOT_FDMA_TX_RING_SIZE -
+		       (tx_ring->next_to_use - tx_ring->next_to_clean) - 1;
+	else
+		return tx_ring->next_to_clean - tx_ring->next_to_use - 1;
+}
+
+static bool ocelot_fdma_tx_ring_empty(struct ocelot_fdma *fdma)
+{
+	struct ocelot_fdma_tx_ring *tx_ring = &fdma->tx_ring;
+
+	return tx_ring->next_to_clean == tx_ring->next_to_use;
+}
+
+static void ocelot_fdma_activate_chan(struct ocelot *ocelot, dma_addr_t dma,
+				      int chan)
+{
+	ocelot_fdma_writel(ocelot, MSCC_FDMA_DCB_LLP(chan), dma);
+	/* Barrier to force memory writes to DCB to be completed before starting
+	 * the channel.
+	 */
+	wmb();
+	ocelot_fdma_writel(ocelot, MSCC_FDMA_CH_ACTIVATE, BIT(chan));
+}
+
+static u32 ocelot_fdma_read_ch_safe(struct ocelot *ocelot)
+{
+	return ocelot_fdma_readl(ocelot, MSCC_FDMA_CH_SAFE);
+}
+
+static int ocelot_fdma_wait_chan_safe(struct ocelot *ocelot, int chan)
+{
+	u32 safe;
+
+	return readx_poll_timeout_atomic(ocelot_fdma_read_ch_safe, ocelot, safe,
+					 safe & BIT(chan), 0,
+					 OCELOT_FDMA_CH_SAFE_TIMEOUT_US);
+}
+
+static void ocelot_fdma_dcb_set_data(struct ocelot_fdma_dcb *dcb,
+				     dma_addr_t dma_addr,
+				     size_t size)
+{
+	u32 offset = dma_addr & 0x3;
+
+	dcb->llp = 0;
+	dcb->datap = ALIGN_DOWN(dma_addr, 4);
+	dcb->datal = ALIGN_DOWN(size, 4);
+	dcb->stat = MSCC_FDMA_DCB_STAT_BLOCKO(offset);
+}
+
+static bool ocelot_fdma_rx_alloc_page(struct ocelot *ocelot,
+				      struct ocelot_fdma_rx_buf *rxb)
+{
+	dma_addr_t mapping;
+	struct page *page;
+
+	page = dev_alloc_page();
+	if (unlikely(!page))
+		return false;
+
+	mapping = dma_map_page(ocelot->dev, page, 0, PAGE_SIZE,
+			       DMA_FROM_DEVICE);
+	if (unlikely(dma_mapping_error(ocelot->dev, mapping))) {
+		__free_page(page);
+		return false;
+	}
+
+	rxb->page = page;
+	rxb->page_offset = 0;
+	rxb->dma_addr = mapping;
+
+	return true;
+}
+
+static int ocelot_fdma_alloc_rx_buffs(struct ocelot *ocelot, u16 alloc_cnt)
+{
+	struct ocelot_fdma *fdma = ocelot->fdma;
+	struct ocelot_fdma_rx_ring *rx_ring;
+	struct ocelot_fdma_rx_buf *rxb;
+	struct ocelot_fdma_dcb *dcb;
+	dma_addr_t dma_addr;
+	int ret = 0;
+	u16 idx;
+
+	rx_ring = &fdma->rx_ring;
+	idx = rx_ring->next_to_use;
+
+	while (alloc_cnt--) {
+		rxb = &rx_ring->bufs[idx];
+		/* try reuse page */
+		if (unlikely(!rxb->page)) {
+			if (unlikely(!ocelot_fdma_rx_alloc_page(ocelot, rxb))) {
+				dev_err_ratelimited(ocelot->dev,
+						    "Failed to allocate rx\n");
+				ret = -ENOMEM;
+				break;
+			}
+		}
+
+		dcb = &rx_ring->dcbs[idx];
+		dma_addr = rxb->dma_addr + rxb->page_offset;
+		ocelot_fdma_dcb_set_data(dcb, dma_addr, OCELOT_FDMA_RXB_SIZE);
+
+		idx = ocelot_fdma_idx_next(idx, OCELOT_FDMA_RX_RING_SIZE);
+		/* Chain the DCB to the next one */
+		dcb->llp = ocelot_fdma_idx_dma(rx_ring->dcbs_dma, idx);
+	}
+
+	rx_ring->next_to_use = idx;
+	rx_ring->next_to_alloc = idx;
+
+	return ret;
+}
+
+static bool ocelot_fdma_tx_dcb_set_skb(struct ocelot *ocelot,
+				       struct ocelot_fdma_tx_buf *tx_buf,
+				       struct ocelot_fdma_dcb *dcb,
+				       struct sk_buff *skb)
+{
+	dma_addr_t mapping;
+
+	mapping = dma_map_single(ocelot->dev, skb->data, skb->len,
+				 DMA_TO_DEVICE);
+	if (unlikely(dma_mapping_error(ocelot->dev, mapping)))
+		return false;
+
+	dma_unmap_addr_set(tx_buf, dma_addr, mapping);
+
+	ocelot_fdma_dcb_set_data(dcb, mapping, OCELOT_FDMA_RX_SIZE);
+	tx_buf->skb = skb;
+	dcb->stat |= MSCC_FDMA_DCB_STAT_BLOCKL(skb->len);
+	dcb->stat |= MSCC_FDMA_DCB_STAT_SOF | MSCC_FDMA_DCB_STAT_EOF;
+
+	return true;
+}
+
+static bool ocelot_fdma_check_stop_rx(struct ocelot *ocelot)
+{
+	u32 llp;
+
+	/* Check if the FDMA hits the DCB with LLP == NULL */
+	llp = ocelot_fdma_readl(ocelot, MSCC_FDMA_DCB_LLP(MSCC_FDMA_XTR_CHAN));
+	if (unlikely(llp))
+		return false;
+
+	ocelot_fdma_writel(ocelot, MSCC_FDMA_CH_DISABLE,
+			   BIT(MSCC_FDMA_XTR_CHAN));
+
+	return true;
+}
+
+static void ocelot_fdma_rx_set_llp(struct ocelot_fdma_rx_ring *rx_ring)
+{
+	struct ocelot_fdma_dcb *dcb;
+	unsigned int idx;
+
+	idx = ocelot_fdma_idx_prev(rx_ring->next_to_use,
+				   OCELOT_FDMA_RX_RING_SIZE);
+	dcb = &rx_ring->dcbs[idx];
+	dcb->llp = 0;
+}
+
+static void ocelot_fdma_rx_restart(struct ocelot *ocelot)
+{
+	struct ocelot_fdma *fdma = ocelot->fdma;
+	struct ocelot_fdma_rx_ring *rx_ring;
+	const u8 chan = MSCC_FDMA_XTR_CHAN;
+	dma_addr_t new_llp, dma_base;
+	unsigned int idx;
+	u32 llp_prev;
+	int ret;
+
+	rx_ring = &fdma->rx_ring;
+	ret = ocelot_fdma_wait_chan_safe(ocelot, chan);
+	if (ret) {
+		dev_err_ratelimited(ocelot->dev,
+				    "Unable to stop RX channel\n");
+		return;
+	}
+
+	ocelot_fdma_rx_set_llp(rx_ring);
+
+	/* FDMA stopped on the last DCB that contained a NULL LLP, since
+	 * we processed some DCBs in RX, there is free space, and  we must set
+	 * DCB_LLP to point to the next DCB
+	 */
+	llp_prev = ocelot_fdma_readl(ocelot, MSCC_FDMA_DCB_LLP_PREV(chan));
+	dma_base = rx_ring->dcbs_dma;
+
+	/* Get the next DMA addr located after LLP == NULL DCB */
+	idx = ocelot_fdma_dma_idx(dma_base, llp_prev);
+	idx = ocelot_fdma_idx_next(idx, OCELOT_FDMA_RX_RING_SIZE);
+	new_llp = ocelot_fdma_idx_dma(dma_base, idx);
+
+	/* Finally reactivate the channel */
+	ocelot_fdma_activate_chan(ocelot, new_llp, chan);
+}
+
+static bool ocelot_fdma_add_rx_frag(struct ocelot_fdma_rx_buf *rxb, u32 stat,
+				    struct sk_buff *skb, bool first)
+{
+	int size = MSCC_FDMA_DCB_STAT_BLOCKL(stat);
+	struct page *page = rxb->page;
+
+	if (likely(first)) {
+		skb_put(skb, size);
+	} else {
+		skb_add_rx_frag(skb, skb_shinfo(skb)->nr_frags, page,
+				rxb->page_offset, size, OCELOT_FDMA_RX_SIZE);
+	}
+
+	/* Try to reuse page */
+	if (unlikely(page_ref_count(page) != 1 || page_is_pfmemalloc(page)))
+		return false;
+
+	/* Change offset to the other half */
+	rxb->page_offset ^= OCELOT_FDMA_RX_SIZE;
+
+	page_ref_inc(page);
+
+	return true;
+}
+
+static void ocelot_fdma_reuse_rx_page(struct ocelot *ocelot,
+				      struct ocelot_fdma_rx_buf *old_rxb)
+{
+	struct ocelot_fdma_rx_ring *rx_ring = &ocelot->fdma->rx_ring;
+	struct ocelot_fdma_rx_buf *new_rxb;
+
+	new_rxb = &rx_ring->bufs[rx_ring->next_to_alloc];
+	rx_ring->next_to_alloc = ocelot_fdma_idx_next(rx_ring->next_to_alloc,
+						      OCELOT_FDMA_RX_RING_SIZE);
+
+	/* Copy page reference */
+	*new_rxb = *old_rxb;
+
+	/* Sync for use by the device */
+	dma_sync_single_range_for_device(ocelot->dev, old_rxb->dma_addr,
+					 old_rxb->page_offset,
+					 OCELOT_FDMA_RX_SIZE, DMA_FROM_DEVICE);
+}
+
+static struct sk_buff *ocelot_fdma_get_skb(struct ocelot *ocelot, u32 stat,
+					   struct ocelot_fdma_rx_buf *rxb,
+					   struct sk_buff *skb)
+{
+	bool first = false;
+
+	/* Allocate skb head and data */
+	if (likely(!skb)) {
+		void *buff_addr = page_address(rxb->page) +
+				  rxb->page_offset;
+
+		skb = build_skb(buff_addr, OCELOT_FDMA_SKBFRAG_SIZE);
+		if (unlikely(!skb)) {
+			dev_err_ratelimited(ocelot->dev,
+					    "build_skb failed !\n");
+			return NULL;
+		}
+		first = true;
+	}
+
+	dma_sync_single_range_for_cpu(ocelot->dev, rxb->dma_addr,
+				      rxb->page_offset, OCELOT_FDMA_RX_SIZE,
+				      DMA_FROM_DEVICE);
+
+	if (ocelot_fdma_add_rx_frag(rxb, stat, skb, first)) {
+		/* Reuse the free half of the page for the next_to_alloc DCB*/
+		ocelot_fdma_reuse_rx_page(ocelot, rxb);
+	} else {
+		/* page cannot be reused, unmap it */
+		dma_unmap_page(ocelot->dev, rxb->dma_addr, PAGE_SIZE,
+			       DMA_FROM_DEVICE);
+	}
+
+	/* clear rx buff content */
+	rxb->page = NULL;
+
+	return skb;
+}
+
+static bool ocelot_fdma_receive_skb(struct ocelot *ocelot, struct sk_buff *skb)
+{
+	struct net_device *ndev;
+	void *xfh = skb->data;
+	u64 timestamp;
+	u64 src_port;
+
+	skb_pull(skb, OCELOT_TAG_LEN);
+
+	ocelot_xfh_get_src_port(xfh, &src_port);
+	if (unlikely(src_port >= ocelot->num_phys_ports))
+		return false;
+
+	ndev = ocelot_port_to_netdev(ocelot, src_port);
+	if (unlikely(!ndev))
+		return false;
+
+	pskb_trim(skb, skb->len - ETH_FCS_LEN);
+
+	skb->dev = ndev;
+	skb->protocol = eth_type_trans(skb, skb->dev);
+	skb->dev->stats.rx_bytes += skb->len;
+	skb->dev->stats.rx_packets++;
+
+	if (ocelot->ptp) {
+		ocelot_xfh_get_rew_val(xfh, &timestamp);
+		ocelot_ptp_rx_timestamp(ocelot, skb, timestamp);
+	}
+
+	if (likely(!skb_defer_rx_timestamp(skb)))
+		netif_receive_skb(skb);
+
+	return true;
+}
+
+static int ocelot_fdma_rx_get(struct ocelot *ocelot, int budget)
+{
+	struct ocelot_fdma *fdma = ocelot->fdma;
+	struct ocelot_fdma_rx_ring *rx_ring;
+	struct ocelot_fdma_rx_buf *rxb;
+	struct ocelot_fdma_dcb *dcb;
+	struct sk_buff *skb;
+	int work_done = 0;
+	int cleaned_cnt;
+	u32 stat;
+	u16 idx;
+
+	cleaned_cnt = ocelot_fdma_rx_ring_free(fdma);
+	rx_ring = &fdma->rx_ring;
+	skb = rx_ring->skb;
+
+	while (budget--) {
+		idx = rx_ring->next_to_clean;
+		dcb = &rx_ring->dcbs[idx];
+		stat = dcb->stat;
+		if (MSCC_FDMA_DCB_STAT_BLOCKL(stat) == 0)
+			break;
+
+		/* New packet is a start of frame but we already got a skb set,
+		 * we probably lost an EOF packet, free skb
+		 */
+		if (unlikely(skb && (stat & MSCC_FDMA_DCB_STAT_SOF))) {
+			dev_kfree_skb(skb);
+			skb = NULL;
+		}
+
+		rxb = &rx_ring->bufs[idx];
+		/* Fetch next to clean buffer from the rx_ring */
+		skb = ocelot_fdma_get_skb(ocelot, stat, rxb, skb);
+		if (unlikely(!skb))
+			break;
+
+		work_done++;
+		cleaned_cnt++;
+
+		idx = ocelot_fdma_idx_next(idx, OCELOT_FDMA_RX_RING_SIZE);
+		rx_ring->next_to_clean = idx;
+
+		if (unlikely(stat & MSCC_FDMA_DCB_STAT_ABORT ||
+			     stat & MSCC_FDMA_DCB_STAT_PD)) {
+			dev_err_ratelimited(ocelot->dev,
+					    "DCB aborted or pruned\n");
+			dev_kfree_skb(skb);
+			skb = NULL;
+			continue;
+		}
+
+		/* We still need to process the other fragment of the packet
+		 * before delivering it to the network stack
+		 */
+		if (!(stat & MSCC_FDMA_DCB_STAT_EOF))
+			continue;
+
+		if (unlikely(!ocelot_fdma_receive_skb(ocelot, skb)))
+			dev_kfree_skb(skb);
+
+		skb = NULL;
+	}
+
+	rx_ring->skb = skb;
+
+	if (cleaned_cnt)
+		ocelot_fdma_alloc_rx_buffs(ocelot, cleaned_cnt);
+
+	return work_done;
+}
+
+static void ocelot_fdma_wakeup_netdev(struct ocelot *ocelot)
+{
+	struct ocelot_port_private *priv;
+	struct ocelot_port *ocelot_port;
+	struct net_device *dev;
+	int port;
+
+	for (port = 0; port < ocelot->num_phys_ports; port++) {
+		ocelot_port = ocelot->ports[port];
+		if (!ocelot_port)
+			continue;
+		priv = container_of(ocelot_port, struct ocelot_port_private,
+				    port);
+		dev = priv->dev;
+
+		if (unlikely(netif_queue_stopped(dev)))
+			netif_wake_queue(dev);
+	}
+}
+
+static void ocelot_fdma_tx_cleanup(struct ocelot *ocelot, int budget)
+{
+	struct ocelot_fdma *fdma = ocelot->fdma;
+	struct ocelot_fdma_tx_ring *tx_ring;
+	struct ocelot_fdma_tx_buf *buf;
+	unsigned int new_null_llp_idx;
+	struct ocelot_fdma_dcb *dcb;
+	bool end_of_list = false;
+	struct sk_buff *skb;
+	dma_addr_t dma;
+	u32 dcb_llp;
+	u16 ntc;
+	int ret;
+
+	tx_ring = &fdma->tx_ring;
+
+	/* Purge the TX packets that have been sent up to the NULL llp or the
+	 * end of done list.
+	 */
+	while (!ocelot_fdma_tx_ring_empty(fdma)) {
+		ntc = tx_ring->next_to_clean;
+		dcb = &tx_ring->dcbs[ntc];
+		if (!(dcb->stat & MSCC_FDMA_DCB_STAT_PD))
+			break;
+
+		buf = &tx_ring->bufs[ntc];
+		skb = buf->skb;
+		dma_unmap_single(ocelot->dev, dma_unmap_addr(buf, dma_addr),
+				 skb->len, DMA_TO_DEVICE);
+		napi_consume_skb(skb, budget);
+		dcb_llp = dcb->llp;
+
+		/* Only update after accessing all dcb fields */
+		tx_ring->next_to_clean = ocelot_fdma_idx_next(ntc,
+							      OCELOT_FDMA_TX_RING_SIZE);
+
+		/* If we hit the NULL LLP, stop, we might need to reload FDMA */
+		if (dcb_llp == 0) {
+			end_of_list = true;
+			break;
+		}
+	}
+
+	/* No need to try to wake if there were no TX cleaned_cnt up. */
+	if (ocelot_fdma_tx_ring_free(fdma))
+		ocelot_fdma_wakeup_netdev(ocelot);
+
+	/* If there is still some DCBs to be processed by the FDMA or if the
+	 * pending list is empty, there is no need to restart the FDMA.
+	 */
+	if (!end_of_list || ocelot_fdma_tx_ring_empty(fdma))
+		return;
+
+	ret = ocelot_fdma_wait_chan_safe(ocelot, MSCC_FDMA_INJ_CHAN);
+	if (ret) {
+		dev_warn(ocelot->dev,
+			 "Failed to wait for TX channel to stop\n");
+		return;
+	}
+
+	/* Set NULL LLP to be the last DCB used */
+	new_null_llp_idx = ocelot_fdma_idx_prev(tx_ring->next_to_use,
+						OCELOT_FDMA_TX_RING_SIZE);
+	dcb = &tx_ring->dcbs[new_null_llp_idx];
+	dcb->llp = 0;
+
+	dma = ocelot_fdma_idx_dma(tx_ring->dcbs_dma, tx_ring->next_to_clean);
+	ocelot_fdma_activate_chan(ocelot, dma, MSCC_FDMA_INJ_CHAN);
+}
+
+static int ocelot_fdma_napi_poll(struct napi_struct *napi, int budget)
+{
+	struct ocelot_fdma *fdma = container_of(napi, struct ocelot_fdma, napi);
+	struct ocelot *ocelot = fdma->ocelot;
+	int work_done = 0;
+	bool rx_stopped;
+
+	ocelot_fdma_tx_cleanup(ocelot, budget);
+
+	rx_stopped = ocelot_fdma_check_stop_rx(ocelot);
+
+	work_done = ocelot_fdma_rx_get(ocelot, budget);
+
+	if (rx_stopped)
+		ocelot_fdma_rx_restart(ocelot);
+
+	if (work_done < budget) {
+		napi_complete_done(&fdma->napi, work_done);
+		ocelot_fdma_writel(ocelot, MSCC_FDMA_INTR_ENA,
+				   BIT(MSCC_FDMA_INJ_CHAN) |
+				   BIT(MSCC_FDMA_XTR_CHAN));
+	}
+
+	return work_done;
+}
+
+static irqreturn_t ocelot_fdma_interrupt(int irq, void *dev_id)
+{
+	u32 ident, llp, frm, err, err_code;
+	struct ocelot *ocelot = dev_id;
+
+	ident = ocelot_fdma_readl(ocelot, MSCC_FDMA_INTR_IDENT);
+	frm = ocelot_fdma_readl(ocelot, MSCC_FDMA_INTR_FRM);
+	llp = ocelot_fdma_readl(ocelot, MSCC_FDMA_INTR_LLP);
+
+	ocelot_fdma_writel(ocelot, MSCC_FDMA_INTR_LLP, llp & ident);
+	ocelot_fdma_writel(ocelot, MSCC_FDMA_INTR_FRM, frm & ident);
+	if (frm || llp) {
+		ocelot_fdma_writel(ocelot, MSCC_FDMA_INTR_ENA, 0);
+		napi_schedule(&ocelot->fdma->napi);
+	}
+
+	err = ocelot_fdma_readl(ocelot, MSCC_FDMA_EVT_ERR);
+	if (unlikely(err)) {
+		err_code = ocelot_fdma_readl(ocelot, MSCC_FDMA_EVT_ERR_CODE);
+		dev_err_ratelimited(ocelot->dev,
+				    "Error ! chans mask: %#x, code: %#x\n",
+				    err, err_code);
+
+		ocelot_fdma_writel(ocelot, MSCC_FDMA_EVT_ERR, err);
+		ocelot_fdma_writel(ocelot, MSCC_FDMA_EVT_ERR_CODE, err_code);
+	}
+
+	return IRQ_HANDLED;
+}
+
+static void ocelot_fdma_send_skb(struct ocelot *ocelot,
+				 struct ocelot_fdma *fdma, struct sk_buff *skb)
+{
+	struct ocelot_fdma_tx_ring *tx_ring = &fdma->tx_ring;
+	struct ocelot_fdma_tx_buf *tx_buf;
+	struct ocelot_fdma_dcb *dcb;
+	dma_addr_t dma;
+	u16 next_idx;
+
+	dcb = &tx_ring->dcbs[tx_ring->next_to_use];
+	tx_buf = &tx_ring->bufs[tx_ring->next_to_use];
+	if (!ocelot_fdma_tx_dcb_set_skb(ocelot, tx_buf, dcb, skb)) {
+		dev_kfree_skb_any(skb);
+		return;
+	}
+
+	next_idx = ocelot_fdma_idx_next(tx_ring->next_to_use,
+					OCELOT_FDMA_TX_RING_SIZE);
+	skb_tx_timestamp(skb);
+
+	/* If the FDMA TX chan is empty, then enqueue the DCB directly */
+	if (ocelot_fdma_tx_ring_empty(fdma)) {
+		dma = ocelot_fdma_idx_dma(tx_ring->dcbs_dma,
+					  tx_ring->next_to_use);
+		ocelot_fdma_activate_chan(ocelot, dma, MSCC_FDMA_INJ_CHAN);
+	} else {
+		/* Chain the DCBs */
+		dcb->llp = ocelot_fdma_idx_dma(tx_ring->dcbs_dma, next_idx);
+	}
+
+	tx_ring->next_to_use = next_idx;
+}
+
+static int ocelot_fdma_prepare_skb(struct ocelot *ocelot, int port, u32 rew_op,
+				   struct sk_buff *skb, struct net_device *dev)
+{
+	int needed_headroom = max_t(int, OCELOT_TAG_LEN - skb_headroom(skb), 0);
+	int needed_tailroom = max_t(int, ETH_FCS_LEN - skb_tailroom(skb), 0);
+	void *ifh;
+	int err;
+
+	if (unlikely(needed_headroom || needed_tailroom ||
+		     skb_header_cloned(skb))) {
+		err = pskb_expand_head(skb, needed_headroom, needed_tailroom,
+				       GFP_ATOMIC);
+		if (unlikely(err)) {
+			dev_kfree_skb_any(skb);
+			return 1;
+		}
+	}
+
+	err = skb_linearize(skb);
+	if (err) {
+		net_err_ratelimited("%s: skb_linearize error (%d)!\n",
+				    dev->name, err);
+		dev_kfree_skb_any(skb);
+		return 1;
+	}
+
+	ifh = skb_push(skb, OCELOT_TAG_LEN);
+	skb_put(skb, ETH_FCS_LEN);
+	memset(ifh, 0, OCELOT_TAG_LEN);
+	ocelot_ifh_port_set(ifh, port, rew_op, skb_vlan_tag_get(skb));
+
+	return 0;
+}
+
+int ocelot_fdma_inject_frame(struct ocelot *ocelot, int port, u32 rew_op,
+			     struct sk_buff *skb, struct net_device *dev)
+{
+	struct ocelot_fdma *fdma = ocelot->fdma;
+	int ret = NETDEV_TX_OK;
+
+	spin_lock(&fdma->tx_ring.xmit_lock);
+
+	if (ocelot_fdma_tx_ring_free(fdma) == 0) {
+		netif_stop_queue(dev);
+		ret = NETDEV_TX_BUSY;
+		goto out;
+	}
+
+	if (ocelot_fdma_prepare_skb(ocelot, port, rew_op, skb, dev))
+		goto out;
+
+	ocelot_fdma_send_skb(ocelot, fdma, skb);
+
+out:
+	spin_unlock(&fdma->tx_ring.xmit_lock);
+
+	return ret;
+}
+
+static void ocelot_fdma_free_rx_ring(struct ocelot *ocelot)
+{
+	struct ocelot_fdma *fdma = ocelot->fdma;
+	struct ocelot_fdma_rx_ring *rx_ring;
+	struct ocelot_fdma_rx_buf *rxb;
+	u16 idx;
+
+	rx_ring = &fdma->rx_ring;
+	idx = rx_ring->next_to_clean;
+
+	/* Free the pages held in the RX ring */
+	while (idx != rx_ring->next_to_use) {
+		rxb = &rx_ring->bufs[idx];
+		dma_unmap_page(ocelot->dev, rxb->dma_addr, PAGE_SIZE,
+			       DMA_FROM_DEVICE);
+		__free_page(rxb->page);
+		idx = ocelot_fdma_idx_next(idx, OCELOT_FDMA_RX_RING_SIZE);
+	}
+
+	if (fdma->rx_ring.skb)
+		dev_kfree_skb_any(fdma->rx_ring.skb);
+}
+
+static void ocelot_fdma_free_tx_ring(struct ocelot *ocelot)
+{
+	struct ocelot_fdma *fdma = ocelot->fdma;
+	struct ocelot_fdma_tx_ring *tx_ring;
+	struct ocelot_fdma_tx_buf *txb;
+	struct sk_buff *skb;
+	u16 idx;
+
+	tx_ring = &fdma->tx_ring;
+	idx = tx_ring->next_to_clean;
+
+	while (idx != tx_ring->next_to_use) {
+		txb = &tx_ring->bufs[idx];
+		skb = txb->skb;
+		dma_unmap_single(ocelot->dev, dma_unmap_addr(txb, dma_addr),
+				 skb->len, DMA_TO_DEVICE);
+		dev_kfree_skb_any(skb);
+		idx = ocelot_fdma_idx_next(idx, OCELOT_FDMA_TX_RING_SIZE);
+	}
+}
+
+static int ocelot_fdma_rings_alloc(struct ocelot *ocelot)
+{
+	struct ocelot_fdma *fdma = ocelot->fdma;
+	struct ocelot_fdma_dcb *dcbs;
+	unsigned int adjust;
+	dma_addr_t dcbs_dma;
+	int ret;
+
+	/* Create a pool of consistent memory blocks for hardware descriptors */
+	fdma->dcbs_base = dmam_alloc_coherent(ocelot->dev,
+					      OCELOT_DCBS_HW_ALLOC_SIZE,
+					      &fdma->dcbs_dma_base, GFP_KERNEL);
+	if (!fdma->dcbs_base)
+		return -ENOMEM;
+
+	/* DCBs must be aligned on a 32bit boundary */
+	dcbs = fdma->dcbs_base;
+	dcbs_dma = fdma->dcbs_dma_base;
+	if (!IS_ALIGNED(dcbs_dma, 4)) {
+		adjust = dcbs_dma & 0x3;
+		dcbs_dma = ALIGN(dcbs_dma, 4);
+		dcbs = (void *)dcbs + adjust;
+	}
+
+	/* TX queue */
+	fdma->tx_ring.dcbs = dcbs;
+	fdma->tx_ring.dcbs_dma = dcbs_dma;
+	spin_lock_init(&fdma->tx_ring.xmit_lock);
+
+	/* RX queue */
+	fdma->rx_ring.dcbs = dcbs + OCELOT_FDMA_TX_RING_SIZE;
+	fdma->rx_ring.dcbs_dma = dcbs_dma + OCELOT_FDMA_TX_DCB_SIZE;
+	ret = ocelot_fdma_alloc_rx_buffs(ocelot,
+					 ocelot_fdma_tx_ring_free(fdma));
+	if (ret) {
+		ocelot_fdma_free_rx_ring(ocelot);
+		return ret;
+	}
+
+	/* Set the last DCB LLP as NULL, this is normally done when restarting
+	 * the RX chan, but this is for the first run
+	 */
+	ocelot_fdma_rx_set_llp(&fdma->rx_ring);
+
+	return 0;
+}
+
+void ocelot_fdma_netdev_init(struct ocelot *ocelot, struct net_device *dev)
+{
+	struct ocelot_fdma *fdma = ocelot->fdma;
+
+	dev->needed_headroom = OCELOT_TAG_LEN;
+	dev->needed_tailroom = ETH_FCS_LEN;
+
+	if (fdma->ndev)
+		return;
+
+	fdma->ndev = dev;
+	netif_napi_add_weight(dev, &fdma->napi, ocelot_fdma_napi_poll,
+			      OCELOT_FDMA_WEIGHT);
+}
+
+void ocelot_fdma_netdev_deinit(struct ocelot *ocelot, struct net_device *dev)
+{
+	struct ocelot_fdma *fdma = ocelot->fdma;
+
+	if (fdma->ndev == dev) {
+		netif_napi_del(&fdma->napi);
+		fdma->ndev = NULL;
+	}
+}
+
+void ocelot_fdma_init(struct platform_device *pdev, struct ocelot *ocelot)
+{
+	struct device *dev = ocelot->dev;
+	struct ocelot_fdma *fdma;
+	int ret;
+
+	fdma = devm_kzalloc(dev, sizeof(*fdma), GFP_KERNEL);
+	if (!fdma)
+		return;
+
+	ocelot->fdma = fdma;
+	ocelot->dev->coherent_dma_mask = DMA_BIT_MASK(32);
+
+	ocelot_fdma_writel(ocelot, MSCC_FDMA_INTR_ENA, 0);
+
+	fdma->ocelot = ocelot;
+	fdma->irq = platform_get_irq_byname(pdev, "fdma");
+	ret = devm_request_irq(dev, fdma->irq, ocelot_fdma_interrupt, 0,
+			       dev_name(dev), ocelot);
+	if (ret)
+		goto err_free_fdma;
+
+	ret = ocelot_fdma_rings_alloc(ocelot);
+	if (ret)
+		goto err_free_irq;
+
+	static_branch_enable(&ocelot_fdma_enabled);
+
+	return;
+
+err_free_irq:
+	devm_free_irq(dev, fdma->irq, fdma);
+err_free_fdma:
+	devm_kfree(dev, fdma);
+
+	ocelot->fdma = NULL;
+}
+
+void ocelot_fdma_start(struct ocelot *ocelot)
+{
+	struct ocelot_fdma *fdma = ocelot->fdma;
+
+	/* Reconfigure for extraction and injection using DMA */
+	ocelot_write_rix(ocelot, QS_INJ_GRP_CFG_MODE(2), QS_INJ_GRP_CFG, 0);
+	ocelot_write_rix(ocelot, QS_INJ_CTRL_GAP_SIZE(0), QS_INJ_CTRL, 0);
+
+	ocelot_write_rix(ocelot, QS_XTR_GRP_CFG_MODE(2), QS_XTR_GRP_CFG, 0);
+
+	ocelot_fdma_writel(ocelot, MSCC_FDMA_INTR_LLP, 0xffffffff);
+	ocelot_fdma_writel(ocelot, MSCC_FDMA_INTR_FRM, 0xffffffff);
+
+	ocelot_fdma_writel(ocelot, MSCC_FDMA_INTR_LLP_ENA,
+			   BIT(MSCC_FDMA_INJ_CHAN) | BIT(MSCC_FDMA_XTR_CHAN));
+	ocelot_fdma_writel(ocelot, MSCC_FDMA_INTR_FRM_ENA,
+			   BIT(MSCC_FDMA_XTR_CHAN));
+	ocelot_fdma_writel(ocelot, MSCC_FDMA_INTR_ENA,
+			   BIT(MSCC_FDMA_INJ_CHAN) | BIT(MSCC_FDMA_XTR_CHAN));
+
+	napi_enable(&fdma->napi);
+
+	ocelot_fdma_activate_chan(ocelot, ocelot->fdma->rx_ring.dcbs_dma,
+				  MSCC_FDMA_XTR_CHAN);
+}
+
+void ocelot_fdma_deinit(struct ocelot *ocelot)
+{
+	struct ocelot_fdma *fdma = ocelot->fdma;
+
+	ocelot_fdma_writel(ocelot, MSCC_FDMA_INTR_ENA, 0);
+	ocelot_fdma_writel(ocelot, MSCC_FDMA_CH_FORCEDIS,
+			   BIT(MSCC_FDMA_XTR_CHAN));
+	ocelot_fdma_writel(ocelot, MSCC_FDMA_CH_FORCEDIS,
+			   BIT(MSCC_FDMA_INJ_CHAN));
+	napi_synchronize(&fdma->napi);
+	napi_disable(&fdma->napi);
+
+	ocelot_fdma_free_rx_ring(ocelot);
+	ocelot_fdma_free_tx_ring(ocelot);
+}