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, 897 insertions, 0 deletions

diff --git a/drivers/fsi/fsi-master-gpio.c b/drivers/fsi/fsi-master-gpio.c
new file mode 100644
index 000000000..7d5f29b4b
--- /dev/null
+++ b/drivers/fsi/fsi-master-gpio.c
@@ -0,0 +1,897 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * A FSI master controller, using a simple GPIO bit-banging interface
+ */
+
+#include <linux/crc4.h>
+#include <linux/delay.h>
+#include <linux/device.h>
+#include <linux/fsi.h>
+#include <linux/gpio/consumer.h>
+#include <linux/io.h>
+#include <linux/irqflags.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+
+#include "fsi-master.h"
+
+#define	FSI_GPIO_STD_DLY	1	/* Standard pin delay in nS */
+#define LAST_ADDR_INVALID		0x1
+
+struct fsi_master_gpio {
+	struct fsi_master	master;
+	struct device		*dev;
+	struct mutex		cmd_lock;	/* mutex for command ordering */
+	struct gpio_desc	*gpio_clk;
+	struct gpio_desc	*gpio_data;
+	struct gpio_desc	*gpio_trans;	/* Voltage translator */
+	struct gpio_desc	*gpio_enable;	/* FSI enable */
+	struct gpio_desc	*gpio_mux;	/* Mux control */
+	bool			external_mode;
+	bool			no_delays;
+	uint32_t		last_addr;
+	uint8_t			t_send_delay;
+	uint8_t			t_echo_delay;
+};
+
+#define CREATE_TRACE_POINTS
+#include <trace/events/fsi_master_gpio.h>
+
+#define to_fsi_master_gpio(m) container_of(m, struct fsi_master_gpio, master)
+
+struct fsi_gpio_msg {
+	uint64_t	msg;
+	uint8_t		bits;
+};
+
+static void clock_toggle(struct fsi_master_gpio *master, int count)
+{
+	int i;
+
+	for (i = 0; i < count; i++) {
+		if (!master->no_delays)
+			ndelay(FSI_GPIO_STD_DLY);
+		gpiod_set_value(master->gpio_clk, 0);
+		if (!master->no_delays)
+			ndelay(FSI_GPIO_STD_DLY);
+		gpiod_set_value(master->gpio_clk, 1);
+	}
+}
+
+static int sda_clock_in(struct fsi_master_gpio *master)
+{
+	int in;
+
+	if (!master->no_delays)
+		ndelay(FSI_GPIO_STD_DLY);
+	gpiod_set_value(master->gpio_clk, 0);
+
+	/* Dummy read to feed the synchronizers */
+	gpiod_get_value(master->gpio_data);
+
+	/* Actual data read */
+	in = gpiod_get_value(master->gpio_data);
+	if (!master->no_delays)
+		ndelay(FSI_GPIO_STD_DLY);
+	gpiod_set_value(master->gpio_clk, 1);
+	return in ? 1 : 0;
+}
+
+static void sda_out(struct fsi_master_gpio *master, int value)
+{
+	gpiod_set_value(master->gpio_data, value);
+}
+
+static void set_sda_input(struct fsi_master_gpio *master)
+{
+	gpiod_direction_input(master->gpio_data);
+	gpiod_set_value(master->gpio_trans, 0);
+}
+
+static void set_sda_output(struct fsi_master_gpio *master, int value)
+{
+	gpiod_set_value(master->gpio_trans, 1);
+	gpiod_direction_output(master->gpio_data, value);
+}
+
+static void clock_zeros(struct fsi_master_gpio *master, int count)
+{
+	trace_fsi_master_gpio_clock_zeros(master, count);
+	set_sda_output(master, 1);
+	clock_toggle(master, count);
+}
+
+static void echo_delay(struct fsi_master_gpio *master)
+{
+	clock_zeros(master, master->t_echo_delay);
+}
+
+
+static void serial_in(struct fsi_master_gpio *master, struct fsi_gpio_msg *msg,
+			uint8_t num_bits)
+{
+	uint8_t bit, in_bit;
+
+	set_sda_input(master);
+
+	for (bit = 0; bit < num_bits; bit++) {
+		in_bit = sda_clock_in(master);
+		msg->msg <<= 1;
+		msg->msg |= ~in_bit & 0x1;	/* Data is active low */
+	}
+	msg->bits += num_bits;
+
+	trace_fsi_master_gpio_in(master, num_bits, msg->msg);
+}
+
+static void serial_out(struct fsi_master_gpio *master,
+			const struct fsi_gpio_msg *cmd)
+{
+	uint8_t bit;
+	uint64_t msg = ~cmd->msg;	/* Data is active low */
+	uint64_t sda_mask = 0x1ULL << (cmd->bits - 1);
+	uint64_t last_bit = ~0;
+	int next_bit;
+
+	trace_fsi_master_gpio_out(master, cmd->bits, cmd->msg);
+
+	if (!cmd->bits) {
+		dev_warn(master->dev, "trying to output 0 bits\n");
+		return;
+	}
+	set_sda_output(master, 0);
+
+	/* Send the start bit */
+	sda_out(master, 0);
+	clock_toggle(master, 1);
+
+	/* Send the message */
+	for (bit = 0; bit < cmd->bits; bit++) {
+		next_bit = (msg & sda_mask) >> (cmd->bits - 1);
+		if (last_bit ^ next_bit) {
+			sda_out(master, next_bit);
+			last_bit = next_bit;
+		}
+		clock_toggle(master, 1);
+		msg <<= 1;
+	}
+}
+
+static void msg_push_bits(struct fsi_gpio_msg *msg, uint64_t data, int bits)
+{
+	msg->msg <<= bits;
+	msg->msg |= data & ((1ull << bits) - 1);
+	msg->bits += bits;
+}
+
+static void msg_push_crc(struct fsi_gpio_msg *msg)
+{
+	uint8_t crc;
+	int top;
+
+	top = msg->bits & 0x3;
+
+	/* start bit, and any non-aligned top bits */
+	crc = crc4(0, 1 << top | msg->msg >> (msg->bits - top), top + 1);
+
+	/* aligned bits */
+	crc = crc4(crc, msg->msg, msg->bits - top);
+
+	msg_push_bits(msg, crc, 4);
+}
+
+static bool check_same_address(struct fsi_master_gpio *master, int id,
+		uint32_t addr)
+{
+	/* this will also handle LAST_ADDR_INVALID */
+	return master->last_addr == (((id & 0x3) << 21) | (addr & ~0x3));
+}
+
+static bool check_relative_address(struct fsi_master_gpio *master, int id,
+		uint32_t addr, uint32_t *rel_addrp)
+{
+	uint32_t last_addr = master->last_addr;
+	int32_t rel_addr;
+
+	if (last_addr == LAST_ADDR_INVALID)
+		return false;
+
+	/* We may be in 23-bit addressing mode, which uses the id as the
+	 * top two address bits. So, if we're referencing a different ID,
+	 * use absolute addresses.
+	 */
+	if (((last_addr >> 21) & 0x3) != id)
+		return false;
+
+	/* remove the top two bits from any 23-bit addressing */
+	last_addr &= (1 << 21) - 1;
+
+	/* We know that the addresses are limited to 21 bits, so this won't
+	 * overflow the signed rel_addr */
+	rel_addr = addr - last_addr;
+	if (rel_addr > 255 || rel_addr < -256)
+		return false;
+
+	*rel_addrp = (uint32_t)rel_addr;
+
+	return true;
+}
+
+static void last_address_update(struct fsi_master_gpio *master,
+		int id, bool valid, uint32_t addr)
+{
+	if (!valid)
+		master->last_addr = LAST_ADDR_INVALID;
+	else
+		master->last_addr = ((id & 0x3) << 21) | (addr & ~0x3);
+}
+
+/*
+ * Encode an Absolute/Relative/Same Address command
+ */
+static void build_ar_command(struct fsi_master_gpio *master,
+		struct fsi_gpio_msg *cmd, uint8_t id,
+		uint32_t addr, size_t size, const void *data)
+{
+	int i, addr_bits, opcode_bits;
+	bool write = !!data;
+	uint8_t ds, opcode;
+	uint32_t rel_addr;
+
+	cmd->bits = 0;
+	cmd->msg = 0;
+
+	/* we have 21 bits of address max */
+	addr &= ((1 << 21) - 1);
+
+	/* cmd opcodes are variable length - SAME_AR is only two bits */
+	opcode_bits = 3;
+
+	if (check_same_address(master, id, addr)) {
+		/* we still address the byte offset within the word */
+		addr_bits = 2;
+		opcode_bits = 2;
+		opcode = FSI_CMD_SAME_AR;
+		trace_fsi_master_gpio_cmd_same_addr(master);
+
+	} else if (check_relative_address(master, id, addr, &rel_addr)) {
+		/* 8 bits plus sign */
+		addr_bits = 9;
+		addr = rel_addr;
+		opcode = FSI_CMD_REL_AR;
+		trace_fsi_master_gpio_cmd_rel_addr(master, rel_addr);
+
+	} else {
+		addr_bits = 21;
+		opcode = FSI_CMD_ABS_AR;
+		trace_fsi_master_gpio_cmd_abs_addr(master, addr);
+	}
+
+	/*
+	 * The read/write size is encoded in the lower bits of the address
+	 * (as it must be naturally-aligned), and the following ds bit.
+	 *
+	 *	size	addr:1	addr:0	ds
+	 *	1	x	x	0
+	 *	2	x	0	1
+	 *	4	0	1	1
+	 *
+	 */
+	ds = size > 1 ? 1 : 0;
+	addr &= ~(size - 1);
+	if (size == 4)
+		addr |= 1;
+
+	msg_push_bits(cmd, id, 2);
+	msg_push_bits(cmd, opcode, opcode_bits);
+	msg_push_bits(cmd, write ? 0 : 1, 1);
+	msg_push_bits(cmd, addr, addr_bits);
+	msg_push_bits(cmd, ds, 1);
+	for (i = 0; write && i < size; i++)
+		msg_push_bits(cmd, ((uint8_t *)data)[i], 8);
+
+	msg_push_crc(cmd);
+}
+
+static void build_dpoll_command(struct fsi_gpio_msg *cmd, uint8_t slave_id)
+{
+	cmd->bits = 0;
+	cmd->msg = 0;
+
+	msg_push_bits(cmd, slave_id, 2);
+	msg_push_bits(cmd, FSI_CMD_DPOLL, 3);
+	msg_push_crc(cmd);
+}
+
+static void build_epoll_command(struct fsi_gpio_msg *cmd, uint8_t slave_id)
+{
+	cmd->bits = 0;
+	cmd->msg = 0;
+
+	msg_push_bits(cmd, slave_id, 2);
+	msg_push_bits(cmd, FSI_CMD_EPOLL, 3);
+	msg_push_crc(cmd);
+}
+
+static void build_term_command(struct fsi_gpio_msg *cmd, uint8_t slave_id)
+{
+	cmd->bits = 0;
+	cmd->msg = 0;
+
+	msg_push_bits(cmd, slave_id, 2);
+	msg_push_bits(cmd, FSI_CMD_TERM, 6);
+	msg_push_crc(cmd);
+}
+
+/*
+ * Note: callers rely specifically on this returning -EAGAIN for
+ * a CRC error detected in the response. Use other error code
+ * for other situations. It will be converted to something else
+ * higher up the stack before it reaches userspace.
+ */
+static int read_one_response(struct fsi_master_gpio *master,
+		uint8_t data_size, struct fsi_gpio_msg *msgp, uint8_t *tagp)
+{
+	struct fsi_gpio_msg msg;
+	unsigned long flags;
+	uint32_t crc;
+	uint8_t tag;
+	int i;
+
+	local_irq_save(flags);
+
+	/* wait for the start bit */
+	for (i = 0; i < FSI_MASTER_MTOE_COUNT; i++) {
+		msg.bits = 0;
+		msg.msg = 0;
+		serial_in(master, &msg, 1);
+		if (msg.msg)
+			break;
+	}
+	if (i == FSI_MASTER_MTOE_COUNT) {
+		dev_dbg(master->dev,
+			"Master time out waiting for response\n");
+		local_irq_restore(flags);
+		return -ETIMEDOUT;
+	}
+
+	msg.bits = 0;
+	msg.msg = 0;
+
+	/* Read slave ID & response tag */
+	serial_in(master, &msg, 4);
+
+	tag = msg.msg & 0x3;
+
+	/* If we have an ACK and we're expecting data, clock the data in too */
+	if (tag == FSI_RESP_ACK && data_size)
+		serial_in(master, &msg, data_size * 8);
+
+	/* read CRC */
+	serial_in(master, &msg, FSI_CRC_SIZE);
+
+	local_irq_restore(flags);
+
+	/* we have a whole message now; check CRC */
+	crc = crc4(0, 1, 1);
+	crc = crc4(crc, msg.msg, msg.bits);
+	if (crc) {
+		/* Check if it's all 1's, that probably means the host is off */
+		if (((~msg.msg) & ((1ull << msg.bits) - 1)) == 0)
+			return -ENODEV;
+		dev_dbg(master->dev, "ERR response CRC msg: 0x%016llx (%d bits)\n",
+			msg.msg, msg.bits);
+		return -EAGAIN;
+	}
+
+	if (msgp)
+		*msgp = msg;
+	if (tagp)
+		*tagp = tag;
+
+	return 0;
+}
+
+static int issue_term(struct fsi_master_gpio *master, uint8_t slave)
+{
+	struct fsi_gpio_msg cmd;
+	unsigned long flags;
+	uint8_t tag;
+	int rc;
+
+	build_term_command(&cmd, slave);
+
+	local_irq_save(flags);
+	serial_out(master, &cmd);
+	echo_delay(master);
+	local_irq_restore(flags);
+
+	rc = read_one_response(master, 0, NULL, &tag);
+	if (rc < 0) {
+		dev_err(master->dev,
+				"TERM failed; lost communication with slave\n");
+		return -EIO;
+	} else if (tag != FSI_RESP_ACK) {
+		dev_err(master->dev, "TERM failed; response %d\n", tag);
+		return -EIO;
+	}
+
+	return 0;
+}
+
+static int poll_for_response(struct fsi_master_gpio *master,
+		uint8_t slave, uint8_t size, void *data)
+{
+	struct fsi_gpio_msg response, cmd;
+	int busy_count = 0, rc, i;
+	unsigned long flags;
+	uint8_t tag;
+	uint8_t *data_byte = data;
+	int crc_err_retries = 0;
+retry:
+	rc = read_one_response(master, size, &response, &tag);
+
+	/* Handle retries on CRC errors */
+	if (rc == -EAGAIN) {
+		/* Too many retries ? */
+		if (crc_err_retries++ > FSI_CRC_ERR_RETRIES) {
+			/*
+			 * Pass it up as a -EIO otherwise upper level will retry
+			 * the whole command which isn't what we want here.
+			 */
+			rc = -EIO;
+			goto fail;
+		}
+		dev_dbg(master->dev,
+			 "CRC error retry %d\n", crc_err_retries);
+		trace_fsi_master_gpio_crc_rsp_error(master);
+		build_epoll_command(&cmd, slave);
+		local_irq_save(flags);
+		clock_zeros(master, FSI_MASTER_EPOLL_CLOCKS);
+		serial_out(master, &cmd);
+		echo_delay(master);
+		local_irq_restore(flags);
+		goto retry;
+	} else if (rc)
+		goto fail;
+
+	switch (tag) {
+	case FSI_RESP_ACK:
+		if (size && data) {
+			uint64_t val = response.msg;
+			/* clear crc & mask */
+			val >>= 4;
+			val &= (1ull << (size * 8)) - 1;
+
+			for (i = 0; i < size; i++) {
+				data_byte[size-i-1] = val;
+				val >>= 8;
+			}
+		}
+		break;
+	case FSI_RESP_BUSY:
+		/*
+		 * Its necessary to clock slave before issuing
+		 * d-poll, not indicated in the hardware protocol
+		 * spec. < 20 clocks causes slave to hang, 21 ok.
+		 */
+		if (busy_count++ < FSI_MASTER_MAX_BUSY) {
+			build_dpoll_command(&cmd, slave);
+			local_irq_save(flags);
+			clock_zeros(master, FSI_MASTER_DPOLL_CLOCKS);
+			serial_out(master, &cmd);
+			echo_delay(master);
+			local_irq_restore(flags);
+			goto retry;
+		}
+		dev_warn(master->dev,
+			"ERR slave is stuck in busy state, issuing TERM\n");
+		local_irq_save(flags);
+		clock_zeros(master, FSI_MASTER_DPOLL_CLOCKS);
+		local_irq_restore(flags);
+		issue_term(master, slave);
+		rc = -EIO;
+		break;
+
+	case FSI_RESP_ERRA:
+		dev_dbg(master->dev, "ERRA received: 0x%x\n", (int)response.msg);
+		rc = -EIO;
+		break;
+	case FSI_RESP_ERRC:
+		dev_dbg(master->dev, "ERRC received: 0x%x\n", (int)response.msg);
+		trace_fsi_master_gpio_crc_cmd_error(master);
+		rc = -EAGAIN;
+		break;
+	}
+
+	if (busy_count > 0)
+		trace_fsi_master_gpio_poll_response_busy(master, busy_count);
+ fail:
+	/*
+	 * tSendDelay clocks, avoids signal reflections when switching
+	 * from receive of response back to send of data.
+	 */
+	local_irq_save(flags);
+	clock_zeros(master, master->t_send_delay);
+	local_irq_restore(flags);
+
+	return rc;
+}
+
+static int send_request(struct fsi_master_gpio *master,
+		struct fsi_gpio_msg *cmd)
+{
+	unsigned long flags;
+
+	if (master->external_mode)
+		return -EBUSY;
+
+	local_irq_save(flags);
+	serial_out(master, cmd);
+	echo_delay(master);
+	local_irq_restore(flags);
+
+	return 0;
+}
+
+static int fsi_master_gpio_xfer(struct fsi_master_gpio *master, uint8_t slave,
+		struct fsi_gpio_msg *cmd, size_t resp_len, void *resp)
+{
+	int rc = -EAGAIN, retries = 0;
+
+	while ((retries++) < FSI_CRC_ERR_RETRIES) {
+		rc = send_request(master, cmd);
+		if (rc)
+			break;
+		rc = poll_for_response(master, slave, resp_len, resp);
+		if (rc != -EAGAIN)
+			break;
+		rc = -EIO;
+		dev_warn(master->dev, "ECRC retry %d\n", retries);
+
+		/* Pace it a bit before retry */
+		msleep(1);
+	}
+
+	return rc;
+}
+
+static int fsi_master_gpio_read(struct fsi_master *_master, int link,
+		uint8_t id, uint32_t addr, void *val, size_t size)
+{
+	struct fsi_master_gpio *master = to_fsi_master_gpio(_master);
+	struct fsi_gpio_msg cmd;
+	int rc;
+
+	if (link != 0)
+		return -ENODEV;
+
+	mutex_lock(&master->cmd_lock);
+	build_ar_command(master, &cmd, id, addr, size, NULL);
+	rc = fsi_master_gpio_xfer(master, id, &cmd, size, val);
+	last_address_update(master, id, rc == 0, addr);
+	mutex_unlock(&master->cmd_lock);
+
+	return rc;
+}
+
+static int fsi_master_gpio_write(struct fsi_master *_master, int link,
+		uint8_t id, uint32_t addr, const void *val, size_t size)
+{
+	struct fsi_master_gpio *master = to_fsi_master_gpio(_master);
+	struct fsi_gpio_msg cmd;
+	int rc;
+
+	if (link != 0)
+		return -ENODEV;
+
+	mutex_lock(&master->cmd_lock);
+	build_ar_command(master, &cmd, id, addr, size, val);
+	rc = fsi_master_gpio_xfer(master, id, &cmd, 0, NULL);
+	last_address_update(master, id, rc == 0, addr);
+	mutex_unlock(&master->cmd_lock);
+
+	return rc;
+}
+
+static int fsi_master_gpio_term(struct fsi_master *_master,
+		int link, uint8_t id)
+{
+	struct fsi_master_gpio *master = to_fsi_master_gpio(_master);
+	struct fsi_gpio_msg cmd;
+	int rc;
+
+	if (link != 0)
+		return -ENODEV;
+
+	mutex_lock(&master->cmd_lock);
+	build_term_command(&cmd, id);
+	rc = fsi_master_gpio_xfer(master, id, &cmd, 0, NULL);
+	last_address_update(master, id, false, 0);
+	mutex_unlock(&master->cmd_lock);
+
+	return rc;
+}
+
+static int fsi_master_gpio_break(struct fsi_master *_master, int link)
+{
+	struct fsi_master_gpio *master = to_fsi_master_gpio(_master);
+	unsigned long flags;
+
+	if (link != 0)
+		return -ENODEV;
+
+	trace_fsi_master_gpio_break(master);
+
+	mutex_lock(&master->cmd_lock);
+	if (master->external_mode) {
+		mutex_unlock(&master->cmd_lock);
+		return -EBUSY;
+	}
+
+	local_irq_save(flags);
+
+	set_sda_output(master, 1);
+	sda_out(master, 1);
+	clock_toggle(master, FSI_PRE_BREAK_CLOCKS);
+	sda_out(master, 0);
+	clock_toggle(master, FSI_BREAK_CLOCKS);
+	echo_delay(master);
+	sda_out(master, 1);
+	clock_toggle(master, FSI_POST_BREAK_CLOCKS);
+
+	local_irq_restore(flags);
+
+	last_address_update(master, 0, false, 0);
+	mutex_unlock(&master->cmd_lock);
+
+	/* Wait for logic reset to take effect */
+	udelay(200);
+
+	return 0;
+}
+
+static void fsi_master_gpio_init(struct fsi_master_gpio *master)
+{
+	unsigned long flags;
+
+	gpiod_direction_output(master->gpio_mux, 1);
+	gpiod_direction_output(master->gpio_trans, 1);
+	gpiod_direction_output(master->gpio_enable, 1);
+	gpiod_direction_output(master->gpio_clk, 1);
+	gpiod_direction_output(master->gpio_data, 1);
+
+	/* todo: evaluate if clocks can be reduced */
+	local_irq_save(flags);
+	clock_zeros(master, FSI_INIT_CLOCKS);
+	local_irq_restore(flags);
+}
+
+static void fsi_master_gpio_init_external(struct fsi_master_gpio *master)
+{
+	gpiod_direction_output(master->gpio_mux, 0);
+	gpiod_direction_output(master->gpio_trans, 0);
+	gpiod_direction_output(master->gpio_enable, 1);
+	gpiod_direction_input(master->gpio_clk);
+	gpiod_direction_input(master->gpio_data);
+}
+
+static int fsi_master_gpio_link_enable(struct fsi_master *_master, int link,
+				       bool enable)
+{
+	struct fsi_master_gpio *master = to_fsi_master_gpio(_master);
+	int rc = -EBUSY;
+
+	if (link != 0)
+		return -ENODEV;
+
+	mutex_lock(&master->cmd_lock);
+	if (!master->external_mode) {
+		gpiod_set_value(master->gpio_enable, enable ? 1 : 0);
+		rc = 0;
+	}
+	mutex_unlock(&master->cmd_lock);
+
+	return rc;
+}
+
+static int fsi_master_gpio_link_config(struct fsi_master *_master, int link,
+				       u8 t_send_delay, u8 t_echo_delay)
+{
+	struct fsi_master_gpio *master = to_fsi_master_gpio(_master);
+
+	if (link != 0)
+		return -ENODEV;
+
+	mutex_lock(&master->cmd_lock);
+	master->t_send_delay = t_send_delay;
+	master->t_echo_delay = t_echo_delay;
+	mutex_unlock(&master->cmd_lock);
+
+	return 0;
+}
+
+static ssize_t external_mode_show(struct device *dev,
+		struct device_attribute *attr, char *buf)
+{
+	struct fsi_master_gpio *master = dev_get_drvdata(dev);
+
+	return snprintf(buf, PAGE_SIZE - 1, "%u\n",
+			master->external_mode ? 1 : 0);
+}
+
+static ssize_t external_mode_store(struct device *dev,
+		struct device_attribute *attr, const char *buf, size_t count)
+{
+	struct fsi_master_gpio *master = dev_get_drvdata(dev);
+	unsigned long val;
+	bool external_mode;
+	int err;
+
+	err = kstrtoul(buf, 0, &val);
+	if (err)
+		return err;
+
+	external_mode = !!val;
+
+	mutex_lock(&master->cmd_lock);
+
+	if (external_mode == master->external_mode) {
+		mutex_unlock(&master->cmd_lock);
+		return count;
+	}
+
+	master->external_mode = external_mode;
+	if (master->external_mode)
+		fsi_master_gpio_init_external(master);
+	else
+		fsi_master_gpio_init(master);
+
+	mutex_unlock(&master->cmd_lock);
+
+	fsi_master_rescan(&master->master);
+
+	return count;
+}
+
+static DEVICE_ATTR(external_mode, 0664,
+		external_mode_show, external_mode_store);
+
+static void fsi_master_gpio_release(struct device *dev)
+{
+	struct fsi_master_gpio *master = to_fsi_master_gpio(dev_to_fsi_master(dev));
+
+	of_node_put(dev_of_node(master->dev));
+
+	kfree(master);
+}
+
+static int fsi_master_gpio_probe(struct platform_device *pdev)
+{
+	struct fsi_master_gpio *master;
+	struct gpio_desc *gpio;
+	int rc;
+
+	master = kzalloc(sizeof(*master), GFP_KERNEL);
+	if (!master)
+		return -ENOMEM;
+
+	master->dev = &pdev->dev;
+	master->master.dev.parent = master->dev;
+	master->master.dev.of_node = of_node_get(dev_of_node(master->dev));
+	master->master.dev.release = fsi_master_gpio_release;
+	master->last_addr = LAST_ADDR_INVALID;
+
+	gpio = devm_gpiod_get(&pdev->dev, "clock", 0);
+	if (IS_ERR(gpio)) {
+		dev_err(&pdev->dev, "failed to get clock gpio\n");
+		rc = PTR_ERR(gpio);
+		goto err_free;
+	}
+	master->gpio_clk = gpio;
+
+	gpio = devm_gpiod_get(&pdev->dev, "data", 0);
+	if (IS_ERR(gpio)) {
+		dev_err(&pdev->dev, "failed to get data gpio\n");
+		rc = PTR_ERR(gpio);
+		goto err_free;
+	}
+	master->gpio_data = gpio;
+
+	/* Optional GPIOs */
+	gpio = devm_gpiod_get_optional(&pdev->dev, "trans", 0);
+	if (IS_ERR(gpio)) {
+		dev_err(&pdev->dev, "failed to get trans gpio\n");
+		rc = PTR_ERR(gpio);
+		goto err_free;
+	}
+	master->gpio_trans = gpio;
+
+	gpio = devm_gpiod_get_optional(&pdev->dev, "enable", 0);
+	if (IS_ERR(gpio)) {
+		dev_err(&pdev->dev, "failed to get enable gpio\n");
+		rc = PTR_ERR(gpio);
+		goto err_free;
+	}
+	master->gpio_enable = gpio;
+
+	gpio = devm_gpiod_get_optional(&pdev->dev, "mux", 0);
+	if (IS_ERR(gpio)) {
+		dev_err(&pdev->dev, "failed to get mux gpio\n");
+		rc = PTR_ERR(gpio);
+		goto err_free;
+	}
+	master->gpio_mux = gpio;
+
+	/*
+	 * Check if GPIO block is slow enought that no extra delays
+	 * are necessary. This improves performance on ast2500 by
+	 * an order of magnitude.
+	 */
+	master->no_delays = device_property_present(&pdev->dev, "no-gpio-delays");
+
+	/* Default FSI command delays */
+	master->t_send_delay = FSI_SEND_DELAY_CLOCKS;
+	master->t_echo_delay = FSI_ECHO_DELAY_CLOCKS;
+
+	master->master.n_links = 1;
+	master->master.flags = FSI_MASTER_FLAG_SWCLOCK;
+	master->master.read = fsi_master_gpio_read;
+	master->master.write = fsi_master_gpio_write;
+	master->master.term = fsi_master_gpio_term;
+	master->master.send_break = fsi_master_gpio_break;
+	master->master.link_enable = fsi_master_gpio_link_enable;
+	master->master.link_config = fsi_master_gpio_link_config;
+	platform_set_drvdata(pdev, master);
+	mutex_init(&master->cmd_lock);
+
+	fsi_master_gpio_init(master);
+
+	rc = device_create_file(&pdev->dev, &dev_attr_external_mode);
+	if (rc)
+		goto err_free;
+
+	rc = fsi_master_register(&master->master);
+	if (rc) {
+		device_remove_file(&pdev->dev, &dev_attr_external_mode);
+		put_device(&master->master.dev);
+		return rc;
+	}
+	return 0;
+ err_free:
+	kfree(master);
+	return rc;
+}
+
+
+
+static int fsi_master_gpio_remove(struct platform_device *pdev)
+{
+	struct fsi_master_gpio *master = platform_get_drvdata(pdev);
+
+	device_remove_file(&pdev->dev, &dev_attr_external_mode);
+
+	fsi_master_unregister(&master->master);
+
+	return 0;
+}
+
+static const struct of_device_id fsi_master_gpio_match[] = {
+	{ .compatible = "fsi-master-gpio" },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, fsi_master_gpio_match);
+
+static struct platform_driver fsi_master_gpio_driver = {
+	.driver = {
+		.name		= "fsi-master-gpio",
+		.of_match_table	= fsi_master_gpio_match,
+	},
+	.probe	= fsi_master_gpio_probe,
+	.remove = fsi_master_gpio_remove,
+};
+
+module_platform_driver(fsi_master_gpio_driver);
+MODULE_LICENSE("GPL");