fsi: master-ast-cf: Add new FSI master using Aspeed ColdFire

The Aspeed AST2x00 can contain a ColdFire v1 coprocessor which
is currently unused on OpenPower systems.

This adds an alternative to the fsi-master-gpio driver that
uses that coprocessor instead of bit banging from the ARM
core itself. The end result is about 4 times faster.

The firmware for the coprocessor and its source code can be
found at https://github.com/ozbenh/cf-fsi and is system specific.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

4 files changed, 1605 insertions, 0 deletions

diff --git a/drivers/fsi/Kconfig b/drivers/fsi/Kconfig
index 9c08f467a7bb..8d82b1e60514 100644
--- a/drivers/fsi/Kconfig
+++ b/drivers/fsi/Kconfig
@@ -27,6 +27,15 @@ config FSI_MASTER_HUB
 	allow chaining of FSI links to an arbitrary depth.  This allows for
 	a high target device fanout.
 
+config FSI_MASTER_AST_CF
+	tristate "FSI master based on Aspeed ColdFire coprocessor"
+	depends on GPIOLIB
+	depends on GPIO_ASPEED
+	---help---
+	This option enables a FSI master using the AST2400 and AST2500 GPIO
+	lines driven by the internal ColdFire coprocessor. This requires
+	the corresponding machine specific ColdFire firmware to be available.
+
 config FSI_SCOM
 	tristate "SCOM FSI client device driver"
 	---help---
diff --git a/drivers/fsi/Makefile b/drivers/fsi/Makefile
index 851182e1cd9e..a50d6ce22fb3 100644
--- a/drivers/fsi/Makefile
+++ b/drivers/fsi/Makefile
@@ -2,5 +2,6 @@
 obj-$(CONFIG_FSI) += fsi-core.o
 obj-$(CONFIG_FSI_MASTER_HUB) += fsi-master-hub.o
 obj-$(CONFIG_FSI_MASTER_GPIO) += fsi-master-gpio.o
+obj-$(CONFIG_FSI_MASTER_AST_CF) += fsi-master-ast-cf.o
 obj-$(CONFIG_FSI_SCOM) += fsi-scom.o
 obj-$(CONFIG_FSI_SBEFIFO) += fsi-sbefifo.o
diff --git a/drivers/fsi/cf-fsi-fw.h b/drivers/fsi/cf-fsi-fw.h
new file mode 100644
index 000000000000..712df0461911
--- /dev/null
+++ b/drivers/fsi/cf-fsi-fw.h
@@ -0,0 +1,157 @@
+// SPDX-License-Identifier: GPL-2.0+
+#ifndef __CF_FSI_FW_H
+#define __CF_FSI_FW_H
+
+/*
+ * uCode file layout
+ *
+ * 0000...03ff : m68k exception vectors
+ * 0400...04ff : Header info & boot config block
+ * 0500....... : Code & stack
+ */
+
+/*
+ * Header info & boot config area
+ *
+ * The Header info is built into the ucode and provide version and
+ * platform information.
+ *
+ * the Boot config needs to be adjusted by the ARM prior to starting
+ * the ucode if the Command/Status area isn't at 0x320000 in CF space
+ * (ie. beginning of SRAM).
+ */
+
+#define HDR_OFFSET	        0x400
+
+/* Info: Signature & version */
+#define HDR_SYS_SIG		0x00	/* 2 bytes system signature */
+#define  SYS_SIG_SHARED		0x5348
+#define  SYS_SIG_SPLIT		0x5350
+#define HDR_FW_VERS		0x02	/* 2 bytes Major.Minor */
+#define HDR_API_VERS		0x04	/* 2 bytes Major.Minor */
+#define  API_VERSION_MAJ	2	/* Current version */
+#define  API_VERSION_MIN	1
+#define HDR_FW_OPTIONS		0x08	/* 4 bytes option flags */
+#define  FW_OPTION_TRACE_EN	0x00000001	/* FW tracing enabled */
+#define	 FW_OPTION_CONT_CLOCK	0x00000002	/* Continuous clocking supported */
+#define HDR_FW_SIZE		0x10	/* 4 bytes size for combo image */
+
+/* Boot Config: Address of Command/Status area */
+#define HDR_CMD_STAT_AREA	0x80	/* 4 bytes CF address */
+#define HDR_FW_CONTROL		0x84	/* 4 bytes control flags */
+#define	 FW_CONTROL_CONT_CLOCK	0x00000002	/* Continuous clocking enabled */
+#define	 FW_CONTROL_DUMMY_RD	0x00000004	/* Extra dummy read (AST2400) */
+#define	 FW_CONTROL_USE_STOP	0x00000008	/* Use STOP instructions */
+#define HDR_CLOCK_GPIO_VADDR	0x90	/* 2 bytes offset from GPIO base */
+#define HDR_CLOCK_GPIO_DADDR	0x92	/* 2 bytes offset from GPIO base */
+#define HDR_DATA_GPIO_VADDR	0x94	/* 2 bytes offset from GPIO base */
+#define HDR_DATA_GPIO_DADDR	0x96	/* 2 bytes offset from GPIO base */
+#define HDR_TRANS_GPIO_VADDR	0x98	/* 2 bytes offset from GPIO base */
+#define HDR_TRANS_GPIO_DADDR	0x9a	/* 2 bytes offset from GPIO base */
+#define HDR_CLOCK_GPIO_BIT	0x9c	/* 1 byte bit number */
+#define HDR_DATA_GPIO_BIT	0x9d	/* 1 byte bit number */
+#define HDR_TRANS_GPIO_BIT	0x9e	/* 1 byte bit number */
+
+/*
+ *  Command/Status area layout: Main part
+ */
+
+/* Command/Status register:
+ *
+ * +---------------------------+
+ * | STAT | RLEN | CLEN | CMD  |
+ * |   8  |   8  |   8  |   8  |
+ * +---------------------------+
+ *    |       |      |      |
+ *    status  |      |      |
+ * Response len      |      |
+ * (in bits)         |      |
+ *                   |      |
+ *         Command len      |
+ *         (in bits)        |
+ *                          |
+ *               Command code
+ *
+ * Due to the big endian layout, that means that a byte read will
+ * return the status byte
+ */
+#define	CMD_STAT_REG	        0x00
+#define  CMD_REG_CMD_MASK	0x000000ff
+#define  CMD_REG_CMD_SHIFT	0
+#define	  CMD_NONE		0x00
+#define	  CMD_COMMAND		0x01
+#define	  CMD_BREAK		0x02
+#define	  CMD_IDLE_CLOCKS	0x03 /* clen = #clocks */
+#define   CMD_INVALID		0xff
+#define  CMD_REG_CLEN_MASK	0x0000ff00
+#define  CMD_REG_CLEN_SHIFT	8
+#define  CMD_REG_RLEN_MASK	0x00ff0000
+#define  CMD_REG_RLEN_SHIFT	16
+#define  CMD_REG_STAT_MASK	0xff000000
+#define  CMD_REG_STAT_SHIFT	24
+#define	  STAT_WORKING		0x00
+#define	  STAT_COMPLETE		0x01
+#define	  STAT_ERR_INVAL_CMD	0x80
+#define	  STAT_ERR_INVAL_IRQ	0x81
+#define	  STAT_ERR_MTOE		0x82
+
+/* Response tag & CRC */
+#define	STAT_RTAG		0x04
+
+/* Response CRC */
+#define	STAT_RCRC		0x05
+
+/* Echo and Send delay */
+#define	ECHO_DLY_REG		0x08
+#define	SEND_DLY_REG		0x09
+
+/* Command data area
+ *
+ * Last byte of message must be left aligned
+ */
+#define	CMD_DATA		0x10 /* 64 bit of data */
+
+/* Response data area, right aligned, unused top bits are 1 */
+#define	RSP_DATA		0x20 /* 32 bit of data */
+
+/* Misc */
+#define	INT_CNT			0x30 /* 32-bit interrupt count */
+#define	BAD_INT_VEC		0x34 /* 32-bit bad interrupt vector # */
+#define	CF_STARTED		0x38 /* byte, set to -1 when copro started */
+#define	CLK_CNT			0x3c /* 32-bit, clock count (debug only) */
+
+/*
+ *  SRAM layout: GPIO arbitration part
+ */
+#define ARB_REG			0x40
+#define  ARB_ARM_REQ		0x01
+#define  ARB_ARM_ACK		0x02
+
+/* Misc2 */
+#define CF_RESET_D0		0x50
+#define CF_RESET_D1		0x54
+#define BAD_INT_S0		0x58
+#define BAD_INT_S1		0x5c
+#define STOP_CNT		0x60
+
+/* Internal */
+
+/*
+ * SRAM layout: Trace buffer (debug builds only)
+ */
+#define	TRACEBUF		0x100
+#define	  TR_CLKOBIT0		0xc0
+#define	  TR_CLKOBIT1		0xc1
+#define	  TR_CLKOSTART		0x82
+#define	  TR_OLEN		0x83 /* + len */
+#define	  TR_CLKZ		0x84 /* + count */
+#define	  TR_CLKWSTART		0x85
+#define	  TR_CLKTAG		0x86 /* + tag */
+#define	  TR_CLKDATA		0x87 /* + len */
+#define	  TR_CLKCRC		0x88 /* + raw crc */
+#define	  TR_CLKIBIT0		0x90
+#define	  TR_CLKIBIT1		0x91
+#define	  TR_END		0xff
+
+#endif /* __CF_FSI_FW_H */
+
diff --git a/drivers/fsi/fsi-master-ast-cf.c b/drivers/fsi/fsi-master-ast-cf.c
new file mode 100644
index 000000000000..57afaae0b691
--- /dev/null
+++ b/drivers/fsi/fsi-master-ast-cf.c
@@ -0,0 +1,1438 @@
+// SPDX-License-Identifier: GPL-2.0+
+// Copyright 2018 IBM Corp
+/*
+ * 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 <linux/regmap.h>
+#include <linux/firmware.h>
+#include <linux/gpio/aspeed.h>
+#include <linux/mfd/syscon.h>
+#include <linux/of_address.h>
+#include <linux/genalloc.h>
+
+#include "fsi-master.h"
+#include "cf-fsi-fw.h"
+
+#define FW_FILE_NAME	"cf-fsi-fw.bin"
+
+/* Common SCU based coprocessor control registers */
+#define SCU_COPRO_CTRL			0x100
+#define   SCU_COPRO_RESET			0x00000002
+#define   SCU_COPRO_CLK_EN			0x00000001
+
+/* AST2500 specific ones */
+#define SCU_2500_COPRO_SEG0		0x104
+#define SCU_2500_COPRO_SEG1		0x108
+#define SCU_2500_COPRO_SEG2		0x10c
+#define SCU_2500_COPRO_SEG3		0x110
+#define SCU_2500_COPRO_SEG4		0x114
+#define SCU_2500_COPRO_SEG5		0x118
+#define SCU_2500_COPRO_SEG6		0x11c
+#define SCU_2500_COPRO_SEG7		0x120
+#define SCU_2500_COPRO_SEG8		0x124
+#define   SCU_2500_COPRO_SEG_SWAP		0x00000001
+#define SCU_2500_COPRO_CACHE_CTL	0x128
+#define   SCU_2500_COPRO_CACHE_EN		0x00000001
+#define   SCU_2500_COPRO_SEG0_CACHE_EN		0x00000002
+#define   SCU_2500_COPRO_SEG1_CACHE_EN		0x00000004
+#define   SCU_2500_COPRO_SEG2_CACHE_EN		0x00000008
+#define   SCU_2500_COPRO_SEG3_CACHE_EN		0x00000010
+#define   SCU_2500_COPRO_SEG4_CACHE_EN		0x00000020
+#define   SCU_2500_COPRO_SEG5_CACHE_EN		0x00000040
+#define   SCU_2500_COPRO_SEG6_CACHE_EN		0x00000080
+#define   SCU_2500_COPRO_SEG7_CACHE_EN		0x00000100
+#define   SCU_2500_COPRO_SEG8_CACHE_EN		0x00000200
+
+#define SCU_2400_COPRO_SEG0		0x104
+#define SCU_2400_COPRO_SEG2		0x108
+#define SCU_2400_COPRO_SEG4		0x10c
+#define SCU_2400_COPRO_SEG6		0x110
+#define SCU_2400_COPRO_SEG8		0x114
+#define   SCU_2400_COPRO_SEG_SWAP		0x80000000
+#define SCU_2400_COPRO_CACHE_CTL	0x118
+#define   SCU_2400_COPRO_CACHE_EN		0x00000001
+#define   SCU_2400_COPRO_SEG0_CACHE_EN		0x00000002
+#define   SCU_2400_COPRO_SEG2_CACHE_EN		0x00000004
+#define   SCU_2400_COPRO_SEG4_CACHE_EN		0x00000008
+#define   SCU_2400_COPRO_SEG6_CACHE_EN		0x00000010
+#define   SCU_2400_COPRO_SEG8_CACHE_EN		0x00000020
+
+/* CVIC registers */
+#define CVIC_EN_REG			0x10
+#define CVIC_TRIG_REG			0x18
+
+/*
+ * System register base address (needed for configuring the
+ * coldfire maps)
+ */
+#define SYSREG_BASE			0x1e600000
+
+/* Amount of SRAM required */
+#define SRAM_SIZE			0x1000
+
+#define LAST_ADDR_INVALID		0x1
+
+struct fsi_master_acf {
+	struct fsi_master	master;
+	struct device		*dev;
+	struct regmap		*scu;
+	struct mutex		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 */
+	uint16_t		gpio_clk_vreg;
+	uint16_t		gpio_clk_dreg;
+	uint16_t       		gpio_dat_vreg;
+	uint16_t       		gpio_dat_dreg;
+	uint16_t       		gpio_tra_vreg;
+	uint16_t       		gpio_tra_dreg;
+	uint8_t			gpio_clk_bit;
+	uint8_t			gpio_dat_bit;
+	uint8_t			gpio_tra_bit;
+	uint32_t		cf_mem_addr;
+	size_t			cf_mem_size;
+	void __iomem		*cf_mem;
+	void __iomem		*cvic;
+	struct gen_pool		*sram_pool;
+	void __iomem		*sram;
+	bool			is_ast2500;
+	bool			external_mode;
+	bool			trace_enabled;
+	uint32_t		last_addr;
+	uint8_t			t_send_delay;
+	uint8_t			t_echo_delay;
+	uint32_t		cvic_sw_irq;
+};
+#define to_fsi_master_acf(m) container_of(m, struct fsi_master_acf, master)
+
+struct fsi_msg {
+	uint64_t	msg;
+	uint8_t		bits;
+};
+
+#define CREATE_TRACE_POINTS
+#include <trace/events/fsi_master_ast_cf.h>
+
+static void msg_push_bits(struct fsi_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_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 void msg_finish_cmd(struct fsi_msg *cmd)
+{
+	/* Left align message */
+	cmd->msg <<= (64 - cmd->bits);
+}
+
+static bool check_same_address(struct fsi_master_acf *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_acf *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_acf *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_acf *master,
+			     struct fsi_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_acf_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_acf_cmd_rel_addr(master, rel_addr);
+
+	} else {
+		addr_bits = 21;
+		opcode = FSI_CMD_ABS_AR;
+		trace_fsi_master_acf_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);
+	msg_finish_cmd(cmd);
+}
+
+static void build_dpoll_command(struct fsi_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);
+	msg_finish_cmd(cmd);
+}
+
+static void build_epoll_command(struct fsi_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);
+	msg_finish_cmd(cmd);
+}
+
+static void build_term_command(struct fsi_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);
+	msg_finish_cmd(cmd);
+}
+
+static int do_copro_command(struct fsi_master_acf *master, uint32_t op)
+{
+	uint32_t timeout = 10000000;
+	uint8_t stat;
+
+	trace_fsi_master_acf_copro_command(master, op);
+
+	/* Send command */
+	iowrite32be(op, master->sram + CMD_STAT_REG);
+
+	/* Ring doorbell if any */
+	if (master->cvic)
+		iowrite32(0x2, master->cvic + CVIC_TRIG_REG);
+
+	/* Wait for status to indicate completion (or error) */
+	do {
+		if (timeout-- == 0) {
+			dev_warn(master->dev,
+				 "Timeout waiting for coprocessor completion\n");
+			return -ETIMEDOUT;
+		}
+		stat = ioread8(master->sram + CMD_STAT_REG);
+	} while(stat < STAT_COMPLETE || stat == 0xff);
+
+	if (stat == STAT_COMPLETE)
+		return 0;
+	switch(stat) {
+	case STAT_ERR_INVAL_CMD:
+		return -EINVAL;
+	case STAT_ERR_INVAL_IRQ:
+		return -EIO;
+	case STAT_ERR_MTOE:
+		return -ESHUTDOWN;
+	}
+	return -ENXIO;
+}
+
+static int clock_zeros(struct fsi_master_acf *master, int count)
+{
+	while (count) {
+		int rc, lcnt = min(count, 255);
+
+		rc = do_copro_command(master,
+				      CMD_IDLE_CLOCKS | (lcnt << CMD_REG_CLEN_SHIFT));
+		if (rc)
+			return rc;
+		count -= lcnt;
+	}
+	return 0;
+}
+
+static int send_request(struct fsi_master_acf *master, struct fsi_msg *cmd,
+			unsigned int resp_bits)
+{
+	uint32_t op;
+
+	trace_fsi_master_acf_send_request(master, cmd, resp_bits);
+
+	/* Store message into SRAM */
+	iowrite32be((cmd->msg >> 32), master->sram + CMD_DATA);
+	iowrite32be((cmd->msg & 0xffffffff), master->sram + CMD_DATA + 4);
+
+	op = CMD_COMMAND;
+	op |= cmd->bits << CMD_REG_CLEN_SHIFT;
+	if (resp_bits)
+		op |= resp_bits << CMD_REG_RLEN_SHIFT;
+
+	return do_copro_command(master, op);
+}
+
+static int read_copro_response(struct fsi_master_acf *master, uint8_t size,
+			       uint32_t *response, u8 *tag)
+{
+	uint8_t rtag = ioread8(master->sram + STAT_RTAG);
+	uint8_t rcrc = ioread8(master->sram + STAT_RCRC);
+	uint32_t rdata = 0;
+	uint32_t crc;
+	uint8_t ack;
+
+	*tag = ack = rtag & 3;
+
+	/* we have a whole message now; check CRC */
+	crc = crc4(0, 1, 1);
+	crc = crc4(crc, rtag, 4);
+	if (ack == FSI_RESP_ACK && size) {
+		rdata = ioread32be(master->sram + RSP_DATA);
+		crc = crc4(crc, rdata, size);
+		if (response)
+			*response = rdata;
+	}
+	crc = crc4(crc, rcrc, 4);
+
+	trace_fsi_master_acf_copro_response(master, rtag, rcrc, rdata, crc == 0);
+
+	if (crc) {
+		/*
+		 * Check if it's all 1's or all 0's, that probably means
+		 * the host is off
+		 */
+		if ((rtag == 0xf && rcrc == 0xf) || (rtag == 0 && rcrc == 0))
+			return -ENODEV;
+		dev_dbg(master->dev, "Bad response CRC !\n");
+		return -EAGAIN;
+	}
+	return 0;
+}
+
+static int send_term(struct fsi_master_acf *master, uint8_t slave)
+{
+	struct fsi_msg cmd;
+	uint8_t tag;
+	int rc;
+
+	build_term_command(&cmd, slave);
+
+	rc = send_request(master, &cmd, 0);
+	if (rc) {
+		dev_warn(master->dev, "Error %d sending term\n", rc);
+		return rc;
+	}
+
+	rc = read_copro_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 void dump_trace(struct fsi_master_acf *master)
+{
+	char trbuf[52];
+	char *p;
+	int i;
+
+	dev_dbg(master->dev,
+		"CMDSTAT:%08x RTAG=%02x RCRC=%02x RDATA=%02x #INT=%08x\n",
+		ioread32be(master->sram + CMD_STAT_REG),
+		ioread8(master->sram + STAT_RTAG),
+		ioread8(master->sram + STAT_RCRC),
+		ioread32be(master->sram + RSP_DATA),
+		ioread32be(master->sram + INT_CNT));
+
+	for (i = 0; i < 512; i++) {
+		uint8_t v;
+		if ((i % 16) == 0)
+			p = trbuf;
+		v = ioread8(master->sram + TRACEBUF + i);
+		p += sprintf(p, "%02x ", v);
+		if (((i % 16) == 15) || v == TR_END)
+			dev_dbg(master->dev, "%s\n", trbuf);
+		if (v == TR_END)
+			break;
+	}
+}
+
+static int handle_response(struct fsi_master_acf *master,
+			   uint8_t slave, uint8_t size, void *data)
+{
+	int busy_count = 0, rc;
+	int crc_err_retries = 0;
+	struct fsi_msg cmd;
+	uint32_t response;
+	uint8_t tag;
+retry:
+	rc = read_copro_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 bail;
+		}
+		trace_fsi_master_acf_crc_rsp_error(master, crc_err_retries);
+		if (master->trace_enabled)
+			dump_trace(master);
+		rc = clock_zeros(master, FSI_MASTER_EPOLL_CLOCKS);
+		if (rc) {
+			dev_warn(master->dev,
+				 "Error %d clocking zeros for E_POLL\n", rc);
+			return rc;
+		}
+		build_epoll_command(&cmd, slave);
+		rc = send_request(master, &cmd, size);
+		if (rc) {
+			dev_warn(master->dev, "Error %d sending E_POLL\n", rc);
+			return -EIO;
+		}
+		goto retry;
+	}
+	if (rc)
+		return rc;
+
+	switch (tag) {
+	case FSI_RESP_ACK:
+		if (size && data) {
+			if (size == 32)
+				*(__be32 *)data = cpu_to_be32(response);
+			else if (size == 16)
+				*(__be16 *)data = cpu_to_be16(response);
+			else
+				*(u8 *)data = response;
+		}
+		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.
+		 */
+		dev_dbg(master->dev, "Busy, retrying...\n");
+		if (master->trace_enabled)
+			dump_trace(master);
+		rc = clock_zeros(master, FSI_MASTER_DPOLL_CLOCKS);
+		if (rc) {
+			dev_warn(master->dev,
+				 "Error %d clocking zeros for D_POLL\n", rc);
+			break;
+		}
+		if (busy_count++ < FSI_MASTER_MAX_BUSY) {
+			build_dpoll_command(&cmd, slave);
+			rc = send_request(master, &cmd, size);
+			if (rc) {
+				dev_warn(master->dev, "Error %d sending D_POLL\n", rc);
+				break;
+			}
+			goto retry;
+		}
+		dev_dbg(master->dev,
+			"ERR slave is stuck in busy state, issuing TERM\n");
+		send_term(master, slave);
+		rc = -EIO;
+		break;
+
+	case FSI_RESP_ERRA:
+		dev_dbg(master->dev, "ERRA received\n");
+		if (master->trace_enabled)
+			dump_trace(master);
+		rc = -EIO;
+		break;
+	case FSI_RESP_ERRC:
+		dev_dbg(master->dev, "ERRC received\n");
+		if (master->trace_enabled)
+			dump_trace(master);
+		rc = -EAGAIN;
+		break;
+	}
+ bail:
+	if (busy_count > 0) {
+		trace_fsi_master_acf_poll_response_busy(master, busy_count);
+	}
+
+	return rc;
+}
+
+static int fsi_master_acf_xfer(struct fsi_master_acf *master, uint8_t slave,
+			       struct fsi_msg *cmd, size_t resp_len, void *resp)
+{
+	int rc = -EAGAIN, retries = 0;
+
+	resp_len <<= 3;
+	while ((retries++) < FSI_CRC_ERR_RETRIES) {
+		rc = send_request(master, cmd, resp_len);
+		if (rc) {
+			if (rc != -ESHUTDOWN)
+				dev_warn(master->dev, "Error %d sending command\n", rc);
+			break;
+		}
+		rc = handle_response(master, slave, resp_len, resp);
+		if (rc != -EAGAIN)
+			break;
+		rc = -EIO;
+		dev_dbg(master->dev, "ECRC retry %d\n", retries);
+
+		/* Pace it a bit before retry */
+		msleep(1);
+	}
+
+	return rc;
+}
+
+static int fsi_master_acf_read(struct fsi_master *_master, int link,
+			       uint8_t id, uint32_t addr, void *val,
+			       size_t size)
+{
+	struct fsi_master_acf *master = to_fsi_master_acf(_master);
+	struct fsi_msg cmd;
+	int rc;
+
+	if (link != 0)
+		return -ENODEV;
+
+	mutex_lock(&master->lock);
+	dev_dbg(master->dev, "read id %d addr %x size %ud\n", id, addr, size);
+	build_ar_command(master, &cmd, id, addr, size, NULL);
+	rc = fsi_master_acf_xfer(master, id, &cmd, size, val);
+	last_address_update(master, id, rc == 0, addr);
+	if (rc)
+		dev_dbg(master->dev, "read id %d addr 0x%08x err: %d\n",
+			id, addr, rc);
+	mutex_unlock(&master->lock);
+
+	return rc;
+}
+
+static int fsi_master_acf_write(struct fsi_master *_master, int link,
+				uint8_t id, uint32_t addr, const void *val,
+				size_t size)
+{
+	struct fsi_master_acf *master = to_fsi_master_acf(_master);
+	struct fsi_msg cmd;
+	int rc;
+
+	if (link != 0)
+		return -ENODEV;
+
+	mutex_lock(&master->lock);
+	build_ar_command(master, &cmd, id, addr, size, val);
+	dev_dbg(master->dev, "write id %d addr %x size %ud raw_data: %08x\n",
+		id, addr, size, *(uint32_t *)val);
+	rc = fsi_master_acf_xfer(master, id, &cmd, 0, NULL);
+	last_address_update(master, id, rc == 0, addr);
+	if (rc)
+		dev_dbg(master->dev, "write id %d addr 0x%08x err: %d\n",
+			id, addr, rc);
+	mutex_unlock(&master->lock);
+
+	return rc;
+}
+
+static int fsi_master_acf_term(struct fsi_master *_master,
+			       int link, uint8_t id)
+{
+	struct fsi_master_acf *master = to_fsi_master_acf(_master);
+	struct fsi_msg cmd;
+	int rc;
+
+	if (link != 0)
+		return -ENODEV;
+
+	mutex_lock(&master->lock);
+	build_term_command(&cmd, id);
+	dev_dbg(master->dev, "term id %d\n", id);
+	rc = fsi_master_acf_xfer(master, id, &cmd, 0, NULL);
+	last_address_update(master, id, false, 0);
+	mutex_unlock(&master->lock);
+
+	return rc;
+}
+
+static int fsi_master_acf_break(struct fsi_master *_master, int link)
+{
+	struct fsi_master_acf *master = to_fsi_master_acf(_master);
+	int rc;
+
+	if (link != 0)
+		return -ENODEV;
+
+	mutex_lock(&master->lock);
+	if (master->external_mode) {
+		mutex_unlock(&master->lock);
+		return -EBUSY;
+	}
+	dev_dbg(master->dev, "sending BREAK\n");
+	rc = do_copro_command(master, CMD_BREAK);
+	last_address_update(master, 0, false, 0);
+	mutex_unlock(&master->lock);
+
+	/* Wait for logic reset to take effect */
+	udelay(200);
+
+	return rc;
+}
+
+static void reset_cf(struct fsi_master_acf *master)
+{
+	regmap_write(master->scu, SCU_COPRO_CTRL, SCU_COPRO_RESET);
+	usleep_range(20,20);
+	regmap_write(master->scu, SCU_COPRO_CTRL, 0);
+	usleep_range(20,20);
+}
+
+static void start_cf(struct fsi_master_acf *master)
+{
+	regmap_write(master->scu, SCU_COPRO_CTRL, SCU_COPRO_CLK_EN);
+}
+
+static void setup_ast2500_cf_maps(struct fsi_master_acf *master)
+{
+	/*
+	 * Note about byteswap setting: the bus is wired backwards,
+	 * so setting the byteswap bit actually makes the ColdFire
+	 * work "normally" for a BE processor, ie, put the MSB in
+	 * the lowest address byte.
+	 *
+	 * We thus need to set the bit for our main memory which
+	 * contains our program code. We create two mappings for
+	 * the register, one with each setting.
+	 *
+	 * Segments 2 and 3 has a "swapped" mapping (BE)
+	 * and 6 and 7 have a non-swapped mapping (LE) which allows
+	 * us to avoid byteswapping register accesses since the
+	 * registers are all LE.
+	 */
+
+	/* Setup segment 0 to our memory region */
+	regmap_write(master->scu, SCU_2500_COPRO_SEG0, master->cf_mem_addr |
+		     SCU_2500_COPRO_SEG_SWAP);
+
+	/* Segments 2 and 3 to sysregs with byteswap (for SRAM) */
+	regmap_write(master->scu, SCU_2500_COPRO_SEG2, SYSREG_BASE |
+		     SCU_2500_COPRO_SEG_SWAP);
+	regmap_write(master->scu, SCU_2500_COPRO_SEG3, SYSREG_BASE | 0x100000 |
+		     SCU_2500_COPRO_SEG_SWAP);
+
+	/* And segment 6 and 7 to sysregs no byteswap */
+	regmap_write(master->scu, SCU_2500_COPRO_SEG6, SYSREG_BASE);
+	regmap_write(master->scu, SCU_2500_COPRO_SEG7, SYSREG_BASE | 0x100000);
+
+	/* Memory cachable, regs and SRAM not cachable */
+	regmap_write(master->scu, SCU_2500_COPRO_CACHE_CTL,
+		     SCU_2500_COPRO_SEG0_CACHE_EN | SCU_2500_COPRO_CACHE_EN);
+}
+
+static void setup_ast2400_cf_maps(struct fsi_master_acf *master)
+{
+	/* Setup segment 0 to our memory region */
+	regmap_write(master->scu, SCU_2400_COPRO_SEG0, master->cf_mem_addr |
+		     SCU_2400_COPRO_SEG_SWAP);
+
+	/* Segments 2 to sysregs with byteswap (for SRAM) */
+	regmap_write(master->scu, SCU_2400_COPRO_SEG2, SYSREG_BASE |
+		     SCU_2400_COPRO_SEG_SWAP);
+
+	/* And segment 6 to sysregs no byteswap */
+	regmap_write(master->scu, SCU_2400_COPRO_SEG6, SYSREG_BASE);
+
+	/* Memory cachable, regs and SRAM not cachable */
+	regmap_write(master->scu, SCU_2400_COPRO_CACHE_CTL,
+		     SCU_2400_COPRO_SEG0_CACHE_EN | SCU_2400_COPRO_CACHE_EN);
+}
+
+static void setup_common_fw_config(struct fsi_master_acf *master,
+				   void __iomem *base)
+{
+	iowrite16be(master->gpio_clk_vreg, base + HDR_CLOCK_GPIO_VADDR);
+	iowrite16be(master->gpio_clk_dreg, base + HDR_CLOCK_GPIO_DADDR);
+	iowrite16be(master->gpio_dat_vreg, base + HDR_DATA_GPIO_VADDR);
+	iowrite16be(master->gpio_dat_dreg, base + HDR_DATA_GPIO_DADDR);
+	iowrite16be(master->gpio_tra_vreg, base + HDR_TRANS_GPIO_VADDR);
+	iowrite16be(master->gpio_tra_dreg, base + HDR_TRANS_GPIO_DADDR);
+	iowrite8(master->gpio_clk_bit, base + HDR_CLOCK_GPIO_BIT);
+	iowrite8(master->gpio_dat_bit, base + HDR_DATA_GPIO_BIT);
+	iowrite8(master->gpio_tra_bit, base + HDR_TRANS_GPIO_BIT);
+}
+
+static void setup_ast2500_fw_config(struct fsi_master_acf *master)
+{
+	void __iomem *base = master->cf_mem + HDR_OFFSET;
+
+	setup_common_fw_config(master, base);
+	iowrite32be(FW_CONTROL_USE_STOP, base + HDR_FW_CONTROL);
+}
+
+static void setup_ast2400_fw_config(struct fsi_master_acf *master)
+{
+	void __iomem *base = master->cf_mem + HDR_OFFSET;
+
+	setup_common_fw_config(master, base);
+	iowrite32be(FW_CONTROL_CONT_CLOCK|FW_CONTROL_DUMMY_RD, base + HDR_FW_CONTROL);
+}
+
+static int setup_gpios_for_copro(struct fsi_master_acf *master)
+{
+
+	int rc;
+
+	/* This aren't under ColdFire control, just set them up appropriately */
+	gpiod_direction_output(master->gpio_mux, 1);
+	gpiod_direction_output(master->gpio_enable, 1);
+
+	/* Those are under ColdFire control, let it configure them */
+	rc = aspeed_gpio_copro_grab_gpio(master->gpio_clk, &master->gpio_clk_vreg,
+					 &master->gpio_clk_dreg, &master->gpio_clk_bit);
+	if (rc) {
+		dev_err(master->dev, "failed to assign clock gpio to coprocessor\n");
+		return rc;
+	}
+	rc = aspeed_gpio_copro_grab_gpio(master->gpio_data, &master->gpio_dat_vreg,
+					 &master->gpio_dat_dreg, &master->gpio_dat_bit);
+	if (rc) {
+		dev_err(master->dev, "failed to assign data gpio to coprocessor\n");
+		aspeed_gpio_copro_release_gpio(master->gpio_clk);
+		return rc;
+	}
+	rc = aspeed_gpio_copro_grab_gpio(master->gpio_trans, &master->gpio_tra_vreg,
+					 &master->gpio_tra_dreg, &master->gpio_tra_bit);
+	if (rc) {
+		dev_err(master->dev, "failed to assign trans gpio to coprocessor\n");
+		aspeed_gpio_copro_release_gpio(master->gpio_clk);
+		aspeed_gpio_copro_release_gpio(master->gpio_data);
+		return rc;
+	}
+	return 0;
+}
+
+static void release_copro_gpios(struct fsi_master_acf *master)
+{
+	aspeed_gpio_copro_release_gpio(master->gpio_clk);
+	aspeed_gpio_copro_release_gpio(master->gpio_data);
+	aspeed_gpio_copro_release_gpio(master->gpio_trans);
+}
+
+static int load_copro_firmware(struct fsi_master_acf *master)
+{
+	const struct firmware *fw;
+	uint16_t sig = 0, wanted_sig;
+	const u8 *data;
+	size_t size = 0;
+	int rc;
+
+	/* Get the binary */
+	rc = request_firmware(&fw, FW_FILE_NAME, master->dev);
+	if (rc) {
+		dev_err(
+			master->dev, "Error %d to load firwmare '%s' !\n",
+			rc, FW_FILE_NAME);
+		return rc;
+	}
+
+	/* Which image do we want ? (shared vs. split clock/data GPIOs) */
+	if (master->gpio_clk_vreg == master->gpio_dat_vreg)
+		wanted_sig = SYS_SIG_SHARED;
+	else
+		wanted_sig = SYS_SIG_SPLIT;
+	dev_dbg(master->dev, "Looking for image sig %04x\n", wanted_sig);
+
+	/* Try to find it */
+	for (data = fw->data; data < (fw->data + fw->size);) {
+		sig = be16_to_cpup((__be16 *)(data + HDR_OFFSET + HDR_SYS_SIG));
+		size = be32_to_cpup((__be32 *)(data + HDR_OFFSET + HDR_FW_SIZE));
+		if (sig == wanted_sig)
+			break;
+		data += size;
+	}
+	if (sig != wanted_sig) {
+		dev_err(master->dev, "Failed to locate image sig %04x in FW blob\n",
+			wanted_sig);
+		return -ENODEV;
+	}
+	if (size > master->cf_mem_size) {
+		dev_err(master->dev, "FW size (%zd) bigger than memory reserve (%zd)\n",
+			fw->size, master->cf_mem_size);
+		rc = -ENOMEM;
+	} else {
+		memcpy_toio(master->cf_mem, data, size);
+	}
+	release_firmware(fw);
+
+	return rc;
+}
+
+static int check_firmware_image(struct fsi_master_acf *master)
+{
+	uint32_t fw_vers, fw_api, fw_options;
+
+	fw_vers = ioread16be(master->cf_mem + HDR_OFFSET + HDR_FW_VERS);
+	fw_api = ioread16be(master->cf_mem + HDR_OFFSET + HDR_API_VERS);
+	fw_options = ioread32be(master->cf_mem + HDR_OFFSET + HDR_FW_OPTIONS);
+	master->trace_enabled = !!(fw_options & FW_OPTION_TRACE_EN);
+
+	/* Check version and signature */
+	dev_info(master->dev, "ColdFire initialized, firmware v%d API v%d.%d (trace %s)\n",
+		 fw_vers, fw_api >> 8, fw_api & 0xff,
+		 master->trace_enabled ? "enabled" : "disabled");
+
+	if ((fw_api >> 8) != API_VERSION_MAJ) {
+		dev_err(master->dev, "Unsupported coprocessor API version !\n");
+		return -ENODEV;
+	}
+
+	return 0;
+}
+
+static int copro_enable_sw_irq(struct fsi_master_acf *master)
+{
+	int timeout;
+	uint32_t val;
+
+	/*
+	 * Enable coprocessor interrupt input. I've had problems getting the
+	 * value to stick, so try in a loop
+	 */
+	for (timeout = 0; timeout < 10; timeout++) {
+		iowrite32(0x2, master->cvic + CVIC_EN_REG);
+		val = ioread32(master->cvic + CVIC_EN_REG);
+		if (val & 2)
+			break;
+		msleep(1);
+	}
+	if (!(val & 2)) {
+		dev_err(master->dev, "Failed to enable coprocessor interrupt !\n");
+		return -ENODEV;
+	}
+	return 0;
+}
+
+static int fsi_master_acf_setup(struct fsi_master_acf *master)
+{
+	int timeout, rc;
+	uint32_t val;
+
+	/* Make sure the ColdFire is stopped  */
+	reset_cf(master);
+
+	/*
+	 * Clear SRAM. This needs to happen before we setup the GPIOs
+	 * as we might start trying to arbitrate as soon as that happens.
+	 */
+	memset_io(master->sram, 0, SRAM_SIZE);
+
+	/* Configure GPIOs */
+	rc = setup_gpios_for_copro(master);
+	if (rc)
+		return rc;
+
+	/* Load the firmware into the reserved memory */
+	rc = load_copro_firmware(master);
+	if (rc)
+		return rc;
+
+	/* Read signature and check versions */
+	rc = check_firmware_image(master);
+	if (rc)
+		return rc;
+
+	/* Setup coldfire memory map */
+	if (master->is_ast2500) {
+		setup_ast2500_cf_maps(master);
+		setup_ast2500_fw_config(master);
+	} else {
+		setup_ast2400_cf_maps(master);
+		setup_ast2400_fw_config(master);
+	}
+
+	/* Start the ColdFire */
+	start_cf(master);
+
+	/* Wait for status register to indicate command completion
+	 * which signals the initialization is complete
+	 */
+	for (timeout = 0; timeout < 10; timeout++) {
+		val = ioread8(master->sram + CF_STARTED);
+		if (val)
+			break;
+		msleep(1);
+	}
+	if (!val) {
+		dev_err(master->dev, "Coprocessor startup timeout !\n");
+		rc = -ENODEV;
+		goto err;
+	}
+
+	/* Configure echo & send delay */
+	iowrite8(master->t_send_delay, master->sram + SEND_DLY_REG);
+	iowrite8(master->t_echo_delay, master->sram + ECHO_DLY_REG);
+
+	/* Enable SW interrupt to copro if any */
+	if (master->cvic) {
+		rc = copro_enable_sw_irq(master);
+		if (rc)
+			goto err;
+	}
+	return 0;
+ err:
+	/* An error occurred, don't leave the coprocessor running */
+	reset_cf(master);
+
+	/* Release the GPIOs */
+	release_copro_gpios(master);
+
+	return rc;
+}
+
+
+static void fsi_master_acf_terminate(struct fsi_master_acf *master)
+{
+	unsigned long flags;
+
+	/*
+	 * A GPIO arbitration requestion could come in while this is
+	 * happening. To avoid problems, we disable interrupts so it
+	 * cannot preempt us on this CPU
+	 */
+
+	local_irq_save(flags);
+
+	/* Stop the coprocessor */
+	reset_cf(master);
+
+	/* We mark the copro not-started */
+	iowrite32(0, master->sram + CF_STARTED);
+
+	/* We mark the ARB register as having given up arbitration to
+	 * deal with a potential race with the arbitration request
+	 */
+	iowrite8(ARB_ARM_ACK, master->sram + ARB_REG);
+
+	local_irq_restore(flags);
+
+	/* Return the GPIOs to the ARM */
+	release_copro_gpios(master);
+}
+
+static void fsi_master_acf_setup_external(struct fsi_master_acf *master)
+{
+	/* Setup GPIOs for external FSI master (FSP box) */
+	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_acf_link_enable(struct fsi_master *_master, int link)
+{
+	struct fsi_master_acf *master = to_fsi_master_acf(_master);
+	int rc = -EBUSY;
+
+	if (link != 0)
+		return -ENODEV;
+
+	mutex_lock(&master->lock);
+	if (!master->external_mode) {
+		gpiod_set_value(master->gpio_enable, 1);
+		rc = 0;
+	}
+	mutex_unlock(&master->lock);
+
+	return rc;
+}
+
+static int fsi_master_acf_link_config(struct fsi_master *_master, int link,
+				      u8 t_send_delay, u8 t_echo_delay)
+{
+	struct fsi_master_acf *master = to_fsi_master_acf(_master);
+
+	if (link != 0)
+		return -ENODEV;
+
+	mutex_lock(&master->lock);
+	master->t_send_delay = t_send_delay;
+	master->t_echo_delay = t_echo_delay;
+	dev_dbg(master->dev, "Changing delays: send=%d echo=%d\n",
+		t_send_delay, t_echo_delay);
+	iowrite8(master->t_send_delay, master->sram + SEND_DLY_REG);
+	iowrite8(master->t_echo_delay, master->sram + ECHO_DLY_REG);
+	mutex_unlock(&master->lock);
+
+	return 0;
+}
+
+static ssize_t external_mode_show(struct device *dev,
+				  struct device_attribute *attr, char *buf)
+{
+	struct fsi_master_acf *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_acf *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->lock);
+
+	if (external_mode == master->external_mode) {
+		mutex_unlock(&master->lock);
+		return count;
+	}
+
+	master->external_mode = external_mode;
+	if (master->external_mode) {
+		fsi_master_acf_terminate(master);
+		fsi_master_acf_setup_external(master);
+	} else
+		fsi_master_acf_setup(master);
+
+	mutex_unlock(&master->lock);
+
+	fsi_master_rescan(&master->master);
+
+	return count;
+}
+
+static DEVICE_ATTR(external_mode, 0664,
+		external_mode_show, external_mode_store);
+
+static int fsi_master_acf_gpio_request(void *data)
+{
+	struct fsi_master_acf *master = data;
+	int timeout;
+	u8 val;
+
+	/* Note: This doesn't require holding out mutex */
+
+	/* Write reqest */
+	iowrite8(ARB_ARM_REQ, master->sram + ARB_REG);
+
+	/*
+	 * There is a race (which does happen at boot time) when we get an
+	 * arbitration request as we are either about to or just starting
+	 * the coprocessor.
+	 *
+	 * To handle it, we first check if we are running. If not yet we
+	 * check whether the copro is started in the SCU.
+	 *
+	 * If it's not started, we can basically just assume we have arbitration
+	 * and return. Otherwise, we wait normally expecting for the arbitration
+	 * to eventually complete.
+	 */
+	if (ioread32(master->sram + CF_STARTED) == 0) {
+		unsigned int reg = 0;
+
+		regmap_read(master->scu, SCU_COPRO_CTRL, &reg);
+		if (!(reg & SCU_COPRO_CLK_EN))
+			return 0;
+	}
+
+	/* Ring doorbell if any */
+	if (master->cvic)
+		iowrite32(0x2, master->cvic + CVIC_TRIG_REG);
+
+	for (timeout = 0; timeout < 10000; timeout++) {
+		val = ioread8(master->sram + ARB_REG);
+		if (val != ARB_ARM_REQ)
+			break;
+		udelay(1);
+	}
+
+	/* If it failed, override anyway */
+	if (val != ARB_ARM_ACK)
+		dev_warn(master->dev, "GPIO request arbitration timeout\n");
+
+	return 0;
+}
+
+static int fsi_master_acf_gpio_release(void *data)
+{
+	struct fsi_master_acf *master = data;
+
+	/* Write release */
+	iowrite8(0, master->sram + ARB_REG);
+
+	/* Ring doorbell if any */
+	if (master->cvic)
+		iowrite32(0x2, master->cvic + CVIC_TRIG_REG);
+
+	return 0;
+}
+
+static void fsi_master_acf_release(struct device *dev)
+{
+	struct fsi_master_acf *master = to_fsi_master_acf(dev_to_fsi_master(dev));
+
+	/* Cleanup, stop coprocessor */
+	mutex_lock(&master->lock);
+	fsi_master_acf_terminate(master);
+	aspeed_gpio_copro_set_ops(NULL, NULL);
+	mutex_unlock(&master->lock);
+
+	/* Free resources */
+	gen_pool_free(master->sram_pool, (unsigned long)master->sram, SRAM_SIZE);
+	of_node_put(dev_of_node(master->dev));
+
+	kfree(master);
+}
+
+static const struct aspeed_gpio_copro_ops fsi_master_acf_gpio_ops = {
+	.request_access = fsi_master_acf_gpio_request,
+	.release_access = fsi_master_acf_gpio_release,
+};
+
+static int fsi_master_acf_probe(struct platform_device *pdev)
+{
+	struct device_node *np, *mnode = dev_of_node(&pdev->dev);
+	struct genpool_data_fixed gpdf;
+	struct fsi_master_acf *master;
+	struct gpio_desc *gpio;
+	struct resource res;
+	uint32_t cf_mem_align;
+	int rc;
+
+	master = kzalloc(sizeof(*master), GFP_KERNEL);
+	if (!master)
+		return -ENOMEM;
+
+	master->dev = &pdev->dev;
+	master->master.dev.parent = master->dev;
+	master->last_addr = LAST_ADDR_INVALID;
+
+	/* AST2400 vs. AST2500 */
+	master->is_ast2500 = of_device_is_compatible(mnode, "aspeed,ast2500-cf-fsi-master");
+
+	/* Grab the SCU, we'll need to access it to configure the coprocessor */
+	if (master->is_ast2500)
+		master->scu = syscon_regmap_lookup_by_compatible("aspeed,ast2500-scu");
+	else
+		master->scu = syscon_regmap_lookup_by_compatible("aspeed,ast2400-scu");
+	if (IS_ERR(master->scu)) {
+		dev_err(&pdev->dev, "failed to find SCU regmap\n");
+		rc = PTR_ERR(master->scu);
+		goto err_free;
+	}
+
+	/* Grab all the GPIOs we need */
+	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;
+
+	/* Grab the reserved memory region (use DMA API instead ?) */
+	np = of_parse_phandle(mnode, "memory-region", 0);
+	if (!np) {
+		dev_err(&pdev->dev, "Didn't find reserved memory\n");
+		rc = -EINVAL;
+		goto err_free;
+	}
+	rc = of_address_to_resource(np, 0, &res);
+	of_node_put(np);
+	if (rc) {
+		dev_err(&pdev->dev, "Couldn't address to resource for reserved memory\n");
+		rc = -ENOMEM;
+		goto err_free;
+	}
+	master->cf_mem_size = resource_size(&res);
+	master->cf_mem_addr = (uint32_t)res.start;
+	cf_mem_align = master->is_ast2500 ? 0x00100000 : 0x00200000;
+	if (master->cf_mem_addr & (cf_mem_align - 1)) {
+		dev_err(&pdev->dev, "Reserved memory has insufficient alignment\n");
+		rc = -ENOMEM;
+		goto err_free;
+	}
+	master->cf_mem = devm_ioremap_resource(&pdev->dev, &res);
+ 	if (IS_ERR(master->cf_mem)) {
+		rc = PTR_ERR(master->cf_mem);
+		dev_err(&pdev->dev, "Error %d mapping coldfire memory\n", rc);
+ 		goto err_free;
+	}
+	dev_dbg(&pdev->dev, "DRAM allocation @%x\n", master->cf_mem_addr);
+
+	/* AST2500 has a SW interrupt to the coprocessor */
+	if (master->is_ast2500) {
+		/* Grab the CVIC (ColdFire interrupts controller) */
+		np = of_parse_phandle(mnode, "aspeed,cvic", 0);
+		if (!np) {
+			dev_err(&pdev->dev, "Didn't find CVIC\n");
+			rc = -EINVAL;
+			goto err_free;
+		}
+		master->cvic = devm_of_iomap(&pdev->dev, np, 0, NULL);
+		if (IS_ERR(master->cvic)) {
+			rc = PTR_ERR(master->cvic);
+			dev_err(&pdev->dev, "Error %d mapping CVIC\n", rc);
+			goto err_free;
+		}
+		rc = of_property_read_u32(np, "copro-sw-interrupts",
+					  &master->cvic_sw_irq);
+		if (rc) {
+			dev_err(&pdev->dev, "Can't find coprocessor SW interrupt\n");
+			goto err_free;
+		}
+	}
+
+	/* Grab the SRAM */
+	master->sram_pool = of_gen_pool_get(dev_of_node(&pdev->dev), "aspeed,sram", 0);
+	if (!master->sram_pool) {
+		rc = -ENODEV;
+		dev_err(&pdev->dev, "Can't find sram pool\n");
+		goto err_free;
+	}
+
+	/* Current microcode only deals with fixed location in SRAM */
+	gpdf.offset = 0;
+	master->sram = (void __iomem *)gen_pool_alloc_algo(master->sram_pool, SRAM_SIZE,
+							   gen_pool_fixed_alloc, &gpdf);
+	if (!master->sram) {
+		rc = -ENOMEM;
+		dev_err(&pdev->dev, "Failed to allocate sram from pool\n");
+		goto err_free;
+	}
+	dev_dbg(&pdev->dev, "SRAM allocation @%lx\n",
+		(unsigned long)gen_pool_virt_to_phys(master->sram_pool,
+						     (unsigned long)master->sram));
+
+	/*
+	 * Hookup with the GPIO driver for arbitration of GPIO banks
+	 * ownership.
+	 */
+	aspeed_gpio_copro_set_ops(&fsi_master_acf_gpio_ops, master);
+
+	/* 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;
+	if (master->is_ast2500)
+		master->master.flags = FSI_MASTER_FLAG_SWCLOCK;
+	master->master.read = fsi_master_acf_read;
+	master->master.write = fsi_master_acf_write;
+	master->master.term = fsi_master_acf_term;
+	master->master.send_break = fsi_master_acf_break;
+	master->master.link_enable = fsi_master_acf_link_enable;
+	master->master.link_config = fsi_master_acf_link_config;
+	master->master.dev.of_node = of_node_get(dev_of_node(master->dev));
+	master->master.dev.release = fsi_master_acf_release;
+	platform_set_drvdata(pdev, master);
+	mutex_init(&master->lock);
+
+	mutex_lock(&master->lock);
+	rc = fsi_master_acf_setup(master);
+	mutex_unlock(&master->lock);
+	if (rc)
+		goto release_of_dev;
+
+	rc = device_create_file(&pdev->dev, &dev_attr_external_mode);
+	if (rc)
+		goto stop_copro;
+
+	rc = fsi_master_register(&master->master);
+	if (!rc)
+		return 0;
+
+	device_remove_file(master->dev, &dev_attr_external_mode);
+	put_device(&master->master.dev);
+	return rc;
+
+ stop_copro:
+	fsi_master_acf_terminate(master);
+ release_of_dev:
+	aspeed_gpio_copro_set_ops(NULL, NULL);
+	gen_pool_free(master->sram_pool, (unsigned long)master->sram, SRAM_SIZE);
+	of_node_put(dev_of_node(master->dev));
+ err_free:
+	kfree(master);
+	return rc;
+}
+
+
+static int fsi_master_acf_remove(struct platform_device *pdev)
+{
+	struct fsi_master_acf *master = platform_get_drvdata(pdev);
+
+	device_remove_file(master->dev, &dev_attr_external_mode);
+
+	fsi_master_unregister(&master->master);
+
+	return 0;
+}
+
+static const struct of_device_id fsi_master_acf_match[] = {
+	{ .compatible = "aspeed,ast2400-cf-fsi-master" },
+	{ .compatible = "aspeed,ast2500-cf-fsi-master" },
+	{ },
+};
+
+static struct platform_driver fsi_master_acf = {
+	.driver = {
+		.name		= "fsi-master-acf",
+		.of_match_table	= fsi_master_acf_match,
+	},
+	.probe	= fsi_master_acf_probe,
+	.remove = fsi_master_acf_remove,
+};
+
+module_platform_driver(fsi_master_acf);
+MODULE_LICENSE("GPL");