mtd: nand: move raw NAND related code to the raw/ subdir

As part of the process of sharing more code between different NAND
based devices, we need to move all raw NAND related code to the raw/
subdirectory.

Signed-off-by: Boris Brezillon <boris.brezillon@bootlin.com>

6 files changed, 4325 insertions, 0 deletions

diff --git a/drivers/mtd/nand/raw/gpmi-nand/Makefile b/drivers/mtd/nand/raw/gpmi-nand/Makefile
new file mode 100644
index 0000000..3a46248
--- /dev/null
+++ b/drivers/mtd/nand/raw/gpmi-nand/Makefile
@@ -0,0 +1,3 @@
+obj-$(CONFIG_MTD_NAND_GPMI_NAND) += gpmi_nand.o
+gpmi_nand-objs += gpmi-nand.o
+gpmi_nand-objs += gpmi-lib.o
diff --git a/drivers/mtd/nand/raw/gpmi-nand/bch-regs.h b/drivers/mtd/nand/raw/gpmi-nand/bch-regs.h
new file mode 100644
index 0000000..05bb91f
--- /dev/null
+++ b/drivers/mtd/nand/raw/gpmi-nand/bch-regs.h
@@ -0,0 +1,128 @@
+/*
+ * Freescale GPMI NAND Flash Driver
+ *
+ * Copyright 2008-2011 Freescale Semiconductor, Inc.
+ * Copyright 2008 Embedded Alley Solutions, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+#ifndef __GPMI_NAND_BCH_REGS_H
+#define __GPMI_NAND_BCH_REGS_H
+
+#define HW_BCH_CTRL				0x00000000
+#define HW_BCH_CTRL_SET				0x00000004
+#define HW_BCH_CTRL_CLR				0x00000008
+#define HW_BCH_CTRL_TOG				0x0000000c
+
+#define BM_BCH_CTRL_COMPLETE_IRQ_EN		(1 << 8)
+#define BM_BCH_CTRL_COMPLETE_IRQ		(1 << 0)
+
+#define HW_BCH_STATUS0				0x00000010
+#define HW_BCH_MODE				0x00000020
+#define HW_BCH_ENCODEPTR			0x00000030
+#define HW_BCH_DATAPTR				0x00000040
+#define HW_BCH_METAPTR				0x00000050
+#define HW_BCH_LAYOUTSELECT			0x00000070
+
+#define HW_BCH_FLASH0LAYOUT0			0x00000080
+
+#define BP_BCH_FLASH0LAYOUT0_NBLOCKS		24
+#define BM_BCH_FLASH0LAYOUT0_NBLOCKS	(0xff << BP_BCH_FLASH0LAYOUT0_NBLOCKS)
+#define BF_BCH_FLASH0LAYOUT0_NBLOCKS(v)		\
+	(((v) << BP_BCH_FLASH0LAYOUT0_NBLOCKS) & BM_BCH_FLASH0LAYOUT0_NBLOCKS)
+
+#define BP_BCH_FLASH0LAYOUT0_META_SIZE		16
+#define BM_BCH_FLASH0LAYOUT0_META_SIZE	(0xff << BP_BCH_FLASH0LAYOUT0_META_SIZE)
+#define BF_BCH_FLASH0LAYOUT0_META_SIZE(v)	\
+	(((v) << BP_BCH_FLASH0LAYOUT0_META_SIZE)\
+					 & BM_BCH_FLASH0LAYOUT0_META_SIZE)
+
+#define BP_BCH_FLASH0LAYOUT0_ECC0		12
+#define BM_BCH_FLASH0LAYOUT0_ECC0	(0xf << BP_BCH_FLASH0LAYOUT0_ECC0)
+#define MX6Q_BP_BCH_FLASH0LAYOUT0_ECC0		11
+#define MX6Q_BM_BCH_FLASH0LAYOUT0_ECC0	(0x1f << MX6Q_BP_BCH_FLASH0LAYOUT0_ECC0)
+#define BF_BCH_FLASH0LAYOUT0_ECC0(v, x)				\
+	(GPMI_IS_MX6(x)					\
+		? (((v) << MX6Q_BP_BCH_FLASH0LAYOUT0_ECC0)	\
+			& MX6Q_BM_BCH_FLASH0LAYOUT0_ECC0)	\
+		: (((v) << BP_BCH_FLASH0LAYOUT0_ECC0)		\
+			& BM_BCH_FLASH0LAYOUT0_ECC0)		\
+	)
+
+#define MX6Q_BP_BCH_FLASH0LAYOUT0_GF_13_14	10
+#define MX6Q_BM_BCH_FLASH0LAYOUT0_GF_13_14			\
+				(0x1 << MX6Q_BP_BCH_FLASH0LAYOUT0_GF_13_14)
+#define BF_BCH_FLASH0LAYOUT0_GF(v, x)				\
+	((GPMI_IS_MX6(x) && ((v) == 14))			\
+		? (((1) << MX6Q_BP_BCH_FLASH0LAYOUT0_GF_13_14)	\
+			& MX6Q_BM_BCH_FLASH0LAYOUT0_GF_13_14)	\
+		: 0						\
+	)
+
+#define BP_BCH_FLASH0LAYOUT0_DATA0_SIZE		0
+#define BM_BCH_FLASH0LAYOUT0_DATA0_SIZE		\
+			(0xfff << BP_BCH_FLASH0LAYOUT0_DATA0_SIZE)
+#define MX6Q_BM_BCH_FLASH0LAYOUT0_DATA0_SIZE	\
+			(0x3ff << BP_BCH_FLASH0LAYOUT0_DATA0_SIZE)
+#define BF_BCH_FLASH0LAYOUT0_DATA0_SIZE(v, x)				\
+	(GPMI_IS_MX6(x)						\
+		? (((v) >> 2) & MX6Q_BM_BCH_FLASH0LAYOUT0_DATA0_SIZE)	\
+		: ((v) & BM_BCH_FLASH0LAYOUT0_DATA0_SIZE)		\
+	)
+
+#define HW_BCH_FLASH0LAYOUT1			0x00000090
+
+#define BP_BCH_FLASH0LAYOUT1_PAGE_SIZE		16
+#define BM_BCH_FLASH0LAYOUT1_PAGE_SIZE		\
+			(0xffff << BP_BCH_FLASH0LAYOUT1_PAGE_SIZE)
+#define BF_BCH_FLASH0LAYOUT1_PAGE_SIZE(v)	\
+	(((v) << BP_BCH_FLASH0LAYOUT1_PAGE_SIZE) \
+					 & BM_BCH_FLASH0LAYOUT1_PAGE_SIZE)
+
+#define BP_BCH_FLASH0LAYOUT1_ECCN		12
+#define BM_BCH_FLASH0LAYOUT1_ECCN	(0xf << BP_BCH_FLASH0LAYOUT1_ECCN)
+#define MX6Q_BP_BCH_FLASH0LAYOUT1_ECCN		11
+#define MX6Q_BM_BCH_FLASH0LAYOUT1_ECCN	(0x1f << MX6Q_BP_BCH_FLASH0LAYOUT1_ECCN)
+#define BF_BCH_FLASH0LAYOUT1_ECCN(v, x)				\
+	(GPMI_IS_MX6(x)					\
+		? (((v) << MX6Q_BP_BCH_FLASH0LAYOUT1_ECCN)	\
+			& MX6Q_BM_BCH_FLASH0LAYOUT1_ECCN)	\
+		: (((v) << BP_BCH_FLASH0LAYOUT1_ECCN)		\
+			& BM_BCH_FLASH0LAYOUT1_ECCN)		\
+	)
+
+#define MX6Q_BP_BCH_FLASH0LAYOUT1_GF_13_14	10
+#define MX6Q_BM_BCH_FLASH0LAYOUT1_GF_13_14			\
+				(0x1 << MX6Q_BP_BCH_FLASH0LAYOUT1_GF_13_14)
+#define BF_BCH_FLASH0LAYOUT1_GF(v, x)				\
+	((GPMI_IS_MX6(x) && ((v) == 14))			\
+		? (((1) << MX6Q_BP_BCH_FLASH0LAYOUT1_GF_13_14)	\
+			& MX6Q_BM_BCH_FLASH0LAYOUT1_GF_13_14)	\
+		: 0						\
+	)
+
+#define BP_BCH_FLASH0LAYOUT1_DATAN_SIZE		0
+#define BM_BCH_FLASH0LAYOUT1_DATAN_SIZE		\
+			(0xfff << BP_BCH_FLASH0LAYOUT1_DATAN_SIZE)
+#define MX6Q_BM_BCH_FLASH0LAYOUT1_DATAN_SIZE	\
+			(0x3ff << BP_BCH_FLASH0LAYOUT1_DATAN_SIZE)
+#define BF_BCH_FLASH0LAYOUT1_DATAN_SIZE(v, x)				\
+	(GPMI_IS_MX6(x)						\
+		? (((v) >> 2) & MX6Q_BM_BCH_FLASH0LAYOUT1_DATAN_SIZE)	\
+		: ((v) & BM_BCH_FLASH0LAYOUT1_DATAN_SIZE)		\
+	)
+
+#define HW_BCH_VERSION				0x00000160
+#endif
diff --git a/drivers/mtd/nand/raw/gpmi-nand/gpmi-lib.c b/drivers/mtd/nand/raw/gpmi-nand/gpmi-lib.c
new file mode 100644
index 0000000..9778724
--- /dev/null
+++ b/drivers/mtd/nand/raw/gpmi-nand/gpmi-lib.c
@@ -0,0 +1,1510 @@
+/*
+ * Freescale GPMI NAND Flash Driver
+ *
+ * Copyright (C) 2008-2011 Freescale Semiconductor, Inc.
+ * Copyright (C) 2008 Embedded Alley Solutions, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+#include <linux/delay.h>
+#include <linux/clk.h>
+#include <linux/slab.h>
+
+#include "gpmi-nand.h"
+#include "gpmi-regs.h"
+#include "bch-regs.h"
+
+static struct timing_threshold timing_default_threshold = {
+	.max_data_setup_cycles       = (BM_GPMI_TIMING0_DATA_SETUP >>
+						BP_GPMI_TIMING0_DATA_SETUP),
+	.internal_data_setup_in_ns   = 0,
+	.max_sample_delay_factor     = (BM_GPMI_CTRL1_RDN_DELAY >>
+						BP_GPMI_CTRL1_RDN_DELAY),
+	.max_dll_clock_period_in_ns  = 32,
+	.max_dll_delay_in_ns         = 16,
+};
+
+#define MXS_SET_ADDR		0x4
+#define MXS_CLR_ADDR		0x8
+/*
+ * Clear the bit and poll it cleared.  This is usually called with
+ * a reset address and mask being either SFTRST(bit 31) or CLKGATE
+ * (bit 30).
+ */
+static int clear_poll_bit(void __iomem *addr, u32 mask)
+{
+	int timeout = 0x400;
+
+	/* clear the bit */
+	writel(mask, addr + MXS_CLR_ADDR);
+
+	/*
+	 * SFTRST needs 3 GPMI clocks to settle, the reference manual
+	 * recommends to wait 1us.
+	 */
+	udelay(1);
+
+	/* poll the bit becoming clear */
+	while ((readl(addr) & mask) && --timeout)
+		/* nothing */;
+
+	return !timeout;
+}
+
+#define MODULE_CLKGATE		(1 << 30)
+#define MODULE_SFTRST		(1 << 31)
+/*
+ * The current mxs_reset_block() will do two things:
+ *  [1] enable the module.
+ *  [2] reset the module.
+ *
+ * In most of the cases, it's ok.
+ * But in MX23, there is a hardware bug in the BCH block (see erratum #2847).
+ * If you try to soft reset the BCH block, it becomes unusable until
+ * the next hard reset. This case occurs in the NAND boot mode. When the board
+ * boots by NAND, the ROM of the chip will initialize the BCH blocks itself.
+ * So If the driver tries to reset the BCH again, the BCH will not work anymore.
+ * You will see a DMA timeout in this case. The bug has been fixed
+ * in the following chips, such as MX28.
+ *
+ * To avoid this bug, just add a new parameter `just_enable` for
+ * the mxs_reset_block(), and rewrite it here.
+ */
+static int gpmi_reset_block(void __iomem *reset_addr, bool just_enable)
+{
+	int ret;
+	int timeout = 0x400;
+
+	/* clear and poll SFTRST */
+	ret = clear_poll_bit(reset_addr, MODULE_SFTRST);
+	if (unlikely(ret))
+		goto error;
+
+	/* clear CLKGATE */
+	writel(MODULE_CLKGATE, reset_addr + MXS_CLR_ADDR);
+
+	if (!just_enable) {
+		/* set SFTRST to reset the block */
+		writel(MODULE_SFTRST, reset_addr + MXS_SET_ADDR);
+		udelay(1);
+
+		/* poll CLKGATE becoming set */
+		while ((!(readl(reset_addr) & MODULE_CLKGATE)) && --timeout)
+			/* nothing */;
+		if (unlikely(!timeout))
+			goto error;
+	}
+
+	/* clear and poll SFTRST */
+	ret = clear_poll_bit(reset_addr, MODULE_SFTRST);
+	if (unlikely(ret))
+		goto error;
+
+	/* clear and poll CLKGATE */
+	ret = clear_poll_bit(reset_addr, MODULE_CLKGATE);
+	if (unlikely(ret))
+		goto error;
+
+	return 0;
+
+error:
+	pr_err("%s(%p): module reset timeout\n", __func__, reset_addr);
+	return -ETIMEDOUT;
+}
+
+static int __gpmi_enable_clk(struct gpmi_nand_data *this, bool v)
+{
+	struct clk *clk;
+	int ret;
+	int i;
+
+	for (i = 0; i < GPMI_CLK_MAX; i++) {
+		clk = this->resources.clock[i];
+		if (!clk)
+			break;
+
+		if (v) {
+			ret = clk_prepare_enable(clk);
+			if (ret)
+				goto err_clk;
+		} else {
+			clk_disable_unprepare(clk);
+		}
+	}
+	return 0;
+
+err_clk:
+	for (; i > 0; i--)
+		clk_disable_unprepare(this->resources.clock[i - 1]);
+	return ret;
+}
+
+#define gpmi_enable_clk(x) __gpmi_enable_clk(x, true)
+#define gpmi_disable_clk(x) __gpmi_enable_clk(x, false)
+
+int gpmi_init(struct gpmi_nand_data *this)
+{
+	struct resources *r = &this->resources;
+	int ret;
+
+	ret = gpmi_enable_clk(this);
+	if (ret)
+		return ret;
+	ret = gpmi_reset_block(r->gpmi_regs, false);
+	if (ret)
+		goto err_out;
+
+	/*
+	 * Reset BCH here, too. We got failures otherwise :(
+	 * See later BCH reset for explanation of MX23 handling
+	 */
+	ret = gpmi_reset_block(r->bch_regs, GPMI_IS_MX23(this));
+	if (ret)
+		goto err_out;
+
+
+	/* Choose NAND mode. */
+	writel(BM_GPMI_CTRL1_GPMI_MODE, r->gpmi_regs + HW_GPMI_CTRL1_CLR);
+
+	/* Set the IRQ polarity. */
+	writel(BM_GPMI_CTRL1_ATA_IRQRDY_POLARITY,
+				r->gpmi_regs + HW_GPMI_CTRL1_SET);
+
+	/* Disable Write-Protection. */
+	writel(BM_GPMI_CTRL1_DEV_RESET, r->gpmi_regs + HW_GPMI_CTRL1_SET);
+
+	/* Select BCH ECC. */
+	writel(BM_GPMI_CTRL1_BCH_MODE, r->gpmi_regs + HW_GPMI_CTRL1_SET);
+
+	/*
+	 * Decouple the chip select from dma channel. We use dma0 for all
+	 * the chips.
+	 */
+	writel(BM_GPMI_CTRL1_DECOUPLE_CS, r->gpmi_regs + HW_GPMI_CTRL1_SET);
+
+	gpmi_disable_clk(this);
+	return 0;
+err_out:
+	gpmi_disable_clk(this);
+	return ret;
+}
+
+/* This function is very useful. It is called only when the bug occur. */
+void gpmi_dump_info(struct gpmi_nand_data *this)
+{
+	struct resources *r = &this->resources;
+	struct bch_geometry *geo = &this->bch_geometry;
+	u32 reg;
+	int i;
+
+	dev_err(this->dev, "Show GPMI registers :\n");
+	for (i = 0; i <= HW_GPMI_DEBUG / 0x10 + 1; i++) {
+		reg = readl(r->gpmi_regs + i * 0x10);
+		dev_err(this->dev, "offset 0x%.3x : 0x%.8x\n", i * 0x10, reg);
+	}
+
+	/* start to print out the BCH info */
+	dev_err(this->dev, "Show BCH registers :\n");
+	for (i = 0; i <= HW_BCH_VERSION / 0x10 + 1; i++) {
+		reg = readl(r->bch_regs + i * 0x10);
+		dev_err(this->dev, "offset 0x%.3x : 0x%.8x\n", i * 0x10, reg);
+	}
+	dev_err(this->dev, "BCH Geometry :\n"
+		"GF length              : %u\n"
+		"ECC Strength           : %u\n"
+		"Page Size in Bytes     : %u\n"
+		"Metadata Size in Bytes : %u\n"
+		"ECC Chunk Size in Bytes: %u\n"
+		"ECC Chunk Count        : %u\n"
+		"Payload Size in Bytes  : %u\n"
+		"Auxiliary Size in Bytes: %u\n"
+		"Auxiliary Status Offset: %u\n"
+		"Block Mark Byte Offset : %u\n"
+		"Block Mark Bit Offset  : %u\n",
+		geo->gf_len,
+		geo->ecc_strength,
+		geo->page_size,
+		geo->metadata_size,
+		geo->ecc_chunk_size,
+		geo->ecc_chunk_count,
+		geo->payload_size,
+		geo->auxiliary_size,
+		geo->auxiliary_status_offset,
+		geo->block_mark_byte_offset,
+		geo->block_mark_bit_offset);
+}
+
+/* Configures the geometry for BCH.  */
+int bch_set_geometry(struct gpmi_nand_data *this)
+{
+	struct resources *r = &this->resources;
+	struct bch_geometry *bch_geo = &this->bch_geometry;
+	unsigned int block_count;
+	unsigned int block_size;
+	unsigned int metadata_size;
+	unsigned int ecc_strength;
+	unsigned int page_size;
+	unsigned int gf_len;
+	int ret;
+
+	if (common_nfc_set_geometry(this))
+		return !0;
+
+	block_count   = bch_geo->ecc_chunk_count - 1;
+	block_size    = bch_geo->ecc_chunk_size;
+	metadata_size = bch_geo->metadata_size;
+	ecc_strength  = bch_geo->ecc_strength >> 1;
+	page_size     = bch_geo->page_size;
+	gf_len        = bch_geo->gf_len;
+
+	ret = gpmi_enable_clk(this);
+	if (ret)
+		return ret;
+
+	/*
+	* Due to erratum #2847 of the MX23, the BCH cannot be soft reset on this
+	* chip, otherwise it will lock up. So we skip resetting BCH on the MX23.
+	* On the other hand, the MX28 needs the reset, because one case has been
+	* seen where the BCH produced ECC errors constantly after 10000
+	* consecutive reboots. The latter case has not been seen on the MX23
+	* yet, still we don't know if it could happen there as well.
+	*/
+	ret = gpmi_reset_block(r->bch_regs, GPMI_IS_MX23(this));
+	if (ret)
+		goto err_out;
+
+	/* Configure layout 0. */
+	writel(BF_BCH_FLASH0LAYOUT0_NBLOCKS(block_count)
+			| BF_BCH_FLASH0LAYOUT0_META_SIZE(metadata_size)
+			| BF_BCH_FLASH0LAYOUT0_ECC0(ecc_strength, this)
+			| BF_BCH_FLASH0LAYOUT0_GF(gf_len, this)
+			| BF_BCH_FLASH0LAYOUT0_DATA0_SIZE(block_size, this),
+			r->bch_regs + HW_BCH_FLASH0LAYOUT0);
+
+	writel(BF_BCH_FLASH0LAYOUT1_PAGE_SIZE(page_size)
+			| BF_BCH_FLASH0LAYOUT1_ECCN(ecc_strength, this)
+			| BF_BCH_FLASH0LAYOUT1_GF(gf_len, this)
+			| BF_BCH_FLASH0LAYOUT1_DATAN_SIZE(block_size, this),
+			r->bch_regs + HW_BCH_FLASH0LAYOUT1);
+
+	/* Set *all* chip selects to use layout 0. */
+	writel(0, r->bch_regs + HW_BCH_LAYOUTSELECT);
+
+	/* Enable interrupts. */
+	writel(BM_BCH_CTRL_COMPLETE_IRQ_EN,
+				r->bch_regs + HW_BCH_CTRL_SET);
+
+	gpmi_disable_clk(this);
+	return 0;
+err_out:
+	gpmi_disable_clk(this);
+	return ret;
+}
+
+/* Converts time in nanoseconds to cycles. */
+static unsigned int ns_to_cycles(unsigned int time,
+			unsigned int period, unsigned int min)
+{
+	unsigned int k;
+
+	k = (time + period - 1) / period;
+	return max(k, min);
+}
+
+#define DEF_MIN_PROP_DELAY	5
+#define DEF_MAX_PROP_DELAY	9
+/* Apply timing to current hardware conditions. */
+static int gpmi_nfc_compute_hardware_timing(struct gpmi_nand_data *this,
+					struct gpmi_nfc_hardware_timing *hw)
+{
+	struct timing_threshold *nfc = &timing_default_threshold;
+	struct resources *r = &this->resources;
+	struct nand_chip *nand = &this->nand;
+	struct nand_timing target = this->timing;
+	bool improved_timing_is_available;
+	unsigned long clock_frequency_in_hz;
+	unsigned int clock_period_in_ns;
+	bool dll_use_half_periods;
+	unsigned int dll_delay_shift;
+	unsigned int max_sample_delay_in_ns;
+	unsigned int address_setup_in_cycles;
+	unsigned int data_setup_in_ns;
+	unsigned int data_setup_in_cycles;
+	unsigned int data_hold_in_cycles;
+	int ideal_sample_delay_in_ns;
+	unsigned int sample_delay_factor;
+	int tEYE;
+	unsigned int min_prop_delay_in_ns = DEF_MIN_PROP_DELAY;
+	unsigned int max_prop_delay_in_ns = DEF_MAX_PROP_DELAY;
+
+	/*
+	 * If there are multiple chips, we need to relax the timings to allow
+	 * for signal distortion due to higher capacitance.
+	 */
+	if (nand->numchips > 2) {
+		target.data_setup_in_ns    += 10;
+		target.data_hold_in_ns     += 10;
+		target.address_setup_in_ns += 10;
+	} else if (nand->numchips > 1) {
+		target.data_setup_in_ns    += 5;
+		target.data_hold_in_ns     += 5;
+		target.address_setup_in_ns += 5;
+	}
+
+	/* Check if improved timing information is available. */
+	improved_timing_is_available =
+		(target.tREA_in_ns  >= 0) &&
+		(target.tRLOH_in_ns >= 0) &&
+		(target.tRHOH_in_ns >= 0);
+
+	/* Inspect the clock. */
+	nfc->clock_frequency_in_hz = clk_get_rate(r->clock[0]);
+	clock_frequency_in_hz = nfc->clock_frequency_in_hz;
+	clock_period_in_ns    = NSEC_PER_SEC / clock_frequency_in_hz;
+
+	/*
+	 * The NFC quantizes setup and hold parameters in terms of clock cycles.
+	 * Here, we quantize the setup and hold timing parameters to the
+	 * next-highest clock period to make sure we apply at least the
+	 * specified times.
+	 *
+	 * For data setup and data hold, the hardware interprets a value of zero
+	 * as the largest possible delay. This is not what's intended by a zero
+	 * in the input parameter, so we impose a minimum of one cycle.
+	 */
+	data_setup_in_cycles    = ns_to_cycles(target.data_setup_in_ns,
+							clock_period_in_ns, 1);
+	data_hold_in_cycles     = ns_to_cycles(target.data_hold_in_ns,
+							clock_period_in_ns, 1);
+	address_setup_in_cycles = ns_to_cycles(target.address_setup_in_ns,
+							clock_period_in_ns, 0);
+
+	/*
+	 * The clock's period affects the sample delay in a number of ways:
+	 *
+	 * (1) The NFC HAL tells us the maximum clock period the sample delay
+	 *     DLL can tolerate. If the clock period is greater than half that
+	 *     maximum, we must configure the DLL to be driven by half periods.
+	 *
+	 * (2) We need to convert from an ideal sample delay, in ns, to a
+	 *     "sample delay factor," which the NFC uses. This factor depends on
+	 *     whether we're driving the DLL with full or half periods.
+	 *     Paraphrasing the reference manual:
+	 *
+	 *         AD = SDF x 0.125 x RP
+	 *
+	 * where:
+	 *
+	 *     AD   is the applied delay, in ns.
+	 *     SDF  is the sample delay factor, which is dimensionless.
+	 *     RP   is the reference period, in ns, which is a full clock period
+	 *          if the DLL is being driven by full periods, or half that if
+	 *          the DLL is being driven by half periods.
+	 *
+	 * Let's re-arrange this in a way that's more useful to us:
+	 *
+	 *                        8
+	 *         SDF  =  AD x ----
+	 *                       RP
+	 *
+	 * The reference period is either the clock period or half that, so this
+	 * is:
+	 *
+	 *                        8       AD x DDF
+	 *         SDF  =  AD x -----  =  --------
+	 *                      f x P        P
+	 *
+	 * where:
+	 *
+	 *       f  is 1 or 1/2, depending on how we're driving the DLL.
+	 *       P  is the clock period.
+	 *     DDF  is the DLL Delay Factor, a dimensionless value that
+	 *          incorporates all the constants in the conversion.
+	 *
+	 * DDF will be either 8 or 16, both of which are powers of two. We can
+	 * reduce the cost of this conversion by using bit shifts instead of
+	 * multiplication or division. Thus:
+	 *
+	 *                 AD << DDS
+	 *         SDF  =  ---------
+	 *                     P
+	 *
+	 *     or
+	 *
+	 *         AD  =  (SDF >> DDS) x P
+	 *
+	 * where:
+	 *
+	 *     DDS  is the DLL Delay Shift, the logarithm to base 2 of the DDF.
+	 */
+	if (clock_period_in_ns > (nfc->max_dll_clock_period_in_ns >> 1)) {
+		dll_use_half_periods = true;
+		dll_delay_shift      = 3 + 1;
+	} else {
+		dll_use_half_periods = false;
+		dll_delay_shift      = 3;
+	}
+
+	/*
+	 * Compute the maximum sample delay the NFC allows, under current
+	 * conditions. If the clock is running too slowly, no sample delay is
+	 * possible.
+	 */
+	if (clock_period_in_ns > nfc->max_dll_clock_period_in_ns)
+		max_sample_delay_in_ns = 0;
+	else {
+		/*
+		 * Compute the delay implied by the largest sample delay factor
+		 * the NFC allows.
+		 */
+		max_sample_delay_in_ns =
+			(nfc->max_sample_delay_factor * clock_period_in_ns) >>
+								dll_delay_shift;
+
+		/*
+		 * Check if the implied sample delay larger than the NFC
+		 * actually allows.
+		 */
+		if (max_sample_delay_in_ns > nfc->max_dll_delay_in_ns)
+			max_sample_delay_in_ns = nfc->max_dll_delay_in_ns;
+	}
+
+	/*
+	 * Check if improved timing information is available. If not, we have to
+	 * use a less-sophisticated algorithm.
+	 */
+	if (!improved_timing_is_available) {
+		/*
+		 * Fold the read setup time required by the NFC into the ideal
+		 * sample delay.
+		 */
+		ideal_sample_delay_in_ns = target.gpmi_sample_delay_in_ns +
+						nfc->internal_data_setup_in_ns;
+
+		/*
+		 * The ideal sample delay may be greater than the maximum
+		 * allowed by the NFC. If so, we can trade off sample delay time
+		 * for more data setup time.
+		 *
+		 * In each iteration of the following loop, we add a cycle to
+		 * the data setup time and subtract a corresponding amount from
+		 * the sample delay until we've satisified the constraints or
+		 * can't do any better.
+		 */
+		while ((ideal_sample_delay_in_ns > max_sample_delay_in_ns) &&
+			(data_setup_in_cycles < nfc->max_data_setup_cycles)) {
+
+			data_setup_in_cycles++;
+			ideal_sample_delay_in_ns -= clock_period_in_ns;
+
+			if (ideal_sample_delay_in_ns < 0)
+				ideal_sample_delay_in_ns = 0;
+
+		}
+
+		/*
+		 * Compute the sample delay factor that corresponds most closely
+		 * to the ideal sample delay. If the result is too large for the
+		 * NFC, use the maximum value.
+		 *
+		 * Notice that we use the ns_to_cycles function to compute the
+		 * sample delay factor. We do this because the form of the
+		 * computation is the same as that for calculating cycles.
+		 */
+		sample_delay_factor =
+			ns_to_cycles(
+				ideal_sample_delay_in_ns << dll_delay_shift,
+							clock_period_in_ns, 0);
+
+		if (sample_delay_factor > nfc->max_sample_delay_factor)
+			sample_delay_factor = nfc->max_sample_delay_factor;
+
+		/* Skip to the part where we return our results. */
+		goto return_results;
+	}
+
+	/*
+	 * If control arrives here, we have more detailed timing information,
+	 * so we can use a better algorithm.
+	 */
+
+	/*
+	 * Fold the read setup time required by the NFC into the maximum
+	 * propagation delay.
+	 */
+	max_prop_delay_in_ns += nfc->internal_data_setup_in_ns;
+
+	/*
+	 * Earlier, we computed the number of clock cycles required to satisfy
+	 * the data setup time. Now, we need to know the actual nanoseconds.
+	 */
+	data_setup_in_ns = clock_period_in_ns * data_setup_in_cycles;
+
+	/*
+	 * Compute tEYE, the width of the data eye when reading from the NAND
+	 * Flash. The eye width is fundamentally determined by the data setup
+	 * time, perturbed by propagation delays and some characteristics of the
+	 * NAND Flash device.
+	 *
+	 * start of the eye = max_prop_delay + tREA
+	 * end of the eye   = min_prop_delay + tRHOH + data_setup
+	 */
+	tEYE = (int)min_prop_delay_in_ns + (int)target.tRHOH_in_ns +
+							(int)data_setup_in_ns;
+
+	tEYE -= (int)max_prop_delay_in_ns + (int)target.tREA_in_ns;
+
+	/*
+	 * The eye must be open. If it's not, we can try to open it by
+	 * increasing its main forcer, the data setup time.
+	 *
+	 * In each iteration of the following loop, we increase the data setup
+	 * time by a single clock cycle. We do this until either the eye is
+	 * open or we run into NFC limits.
+	 */
+	while ((tEYE <= 0) &&
+			(data_setup_in_cycles < nfc->max_data_setup_cycles)) {
+		/* Give a cycle to data setup. */
+		data_setup_in_cycles++;
+		/* Synchronize the data setup time with the cycles. */
+		data_setup_in_ns += clock_period_in_ns;
+		/* Adjust tEYE accordingly. */
+		tEYE += clock_period_in_ns;
+	}
+
+	/*
+	 * When control arrives here, the eye is open. The ideal time to sample
+	 * the data is in the center of the eye:
+	 *
+	 *     end of the eye + start of the eye
+	 *     ---------------------------------  -  data_setup
+	 *                    2
+	 *
+	 * After some algebra, this simplifies to the code immediately below.
+	 */
+	ideal_sample_delay_in_ns =
+		((int)max_prop_delay_in_ns +
+			(int)target.tREA_in_ns +
+				(int)min_prop_delay_in_ns +
+					(int)target.tRHOH_in_ns -
+						(int)data_setup_in_ns) >> 1;
+
+	/*
+	 * The following figure illustrates some aspects of a NAND Flash read:
+	 *
+	 *
+	 *           __                   _____________________________________
+	 * RDN         \_________________/
+	 *
+	 *                                         <---- tEYE ----->
+	 *                                        /-----------------\
+	 * Read Data ----------------------------<                   >---------
+	 *                                        \-----------------/
+	 *             ^                 ^                 ^              ^
+	 *             |                 |                 |              |
+	 *             |<--Data Setup -->|<--Delay Time -->|              |
+	 *             |                 |                 |              |
+	 *             |                 |                                |
+	 *             |                 |<--   Quantized Delay Time   -->|
+	 *             |                 |                                |
+	 *
+	 *
+	 * We have some issues we must now address:
+	 *
+	 * (1) The *ideal* sample delay time must not be negative. If it is, we
+	 *     jam it to zero.
+	 *
+	 * (2) The *ideal* sample delay time must not be greater than that
+	 *     allowed by the NFC. If it is, we can increase the data setup
+	 *     time, which will reduce the delay between the end of the data
+	 *     setup and the center of the eye. It will also make the eye
+	 *     larger, which might help with the next issue...
+	 *
+	 * (3) The *quantized* sample delay time must not fall either before the
+	 *     eye opens or after it closes (the latter is the problem
+	 *     illustrated in the above figure).
+	 */
+
+	/* Jam a negative ideal sample delay to zero. */
+	if (ideal_sample_delay_in_ns < 0)
+		ideal_sample_delay_in_ns = 0;
+
+	/*
+	 * Extend the data setup as needed to reduce the ideal sample delay
+	 * below the maximum permitted by the NFC.
+	 */
+	while ((ideal_sample_delay_in_ns > max_sample_delay_in_ns) &&
+			(data_setup_in_cycles < nfc->max_data_setup_cycles)) {
+
+		/* Give a cycle to data setup. */
+		data_setup_in_cycles++;
+		/* Synchronize the data setup time with the cycles. */
+		data_setup_in_ns += clock_period_in_ns;
+		/* Adjust tEYE accordingly. */
+		tEYE += clock_period_in_ns;
+
+		/*
+		 * Decrease the ideal sample delay by one half cycle, to keep it
+		 * in the middle of the eye.
+		 */
+		ideal_sample_delay_in_ns -= (clock_period_in_ns >> 1);
+
+		/* Jam a negative ideal sample delay to zero. */
+		if (ideal_sample_delay_in_ns < 0)
+			ideal_sample_delay_in_ns = 0;
+	}
+
+	/*
+	 * Compute the sample delay factor that corresponds to the ideal sample
+	 * delay. If the result is too large, then use the maximum allowed
+	 * value.
+	 *
+	 * Notice that we use the ns_to_cycles function to compute the sample
+	 * delay factor. We do this because the form of the computation is the
+	 * same as that for calculating cycles.
+	 */
+	sample_delay_factor =
+		ns_to_cycles(ideal_sample_delay_in_ns << dll_delay_shift,
+							clock_period_in_ns, 0);
+
+	if (sample_delay_factor > nfc->max_sample_delay_factor)
+		sample_delay_factor = nfc->max_sample_delay_factor;
+
+	/*
+	 * These macros conveniently encapsulate a computation we'll use to
+	 * continuously evaluate whether or not the data sample delay is inside
+	 * the eye.
+	 */
+	#define IDEAL_DELAY  ((int) ideal_sample_delay_in_ns)
+
+	#define QUANTIZED_DELAY  \
+		((int) ((sample_delay_factor * clock_period_in_ns) >> \
+							dll_delay_shift))
+
+	#define DELAY_ERROR  (abs(QUANTIZED_DELAY - IDEAL_DELAY))
+
+	#define SAMPLE_IS_NOT_WITHIN_THE_EYE  (DELAY_ERROR > (tEYE >> 1))
+
+	/*
+	 * While the quantized sample time falls outside the eye, reduce the
+	 * sample delay or extend the data setup to move the sampling point back
+	 * toward the eye. Do not allow the number of data setup cycles to
+	 * exceed the maximum allowed by the NFC.
+	 */
+	while (SAMPLE_IS_NOT_WITHIN_THE_EYE &&
+			(data_setup_in_cycles < nfc->max_data_setup_cycles)) {
+		/*
+		 * If control arrives here, the quantized sample delay falls
+		 * outside the eye. Check if it's before the eye opens, or after
+		 * the eye closes.
+		 */
+		if (QUANTIZED_DELAY > IDEAL_DELAY) {
+			/*
+			 * If control arrives here, the quantized sample delay
+			 * falls after the eye closes. Decrease the quantized
+			 * delay time and then go back to re-evaluate.
+			 */
+			if (sample_delay_factor != 0)
+				sample_delay_factor--;
+			continue;
+		}
+
+		/*
+		 * If control arrives here, the quantized sample delay falls
+		 * before the eye opens. Shift the sample point by increasing
+		 * data setup time. This will also make the eye larger.
+		 */
+
+		/* Give a cycle to data setup. */
+		data_setup_in_cycles++;
+		/* Synchronize the data setup time with the cycles. */
+		data_setup_in_ns += clock_period_in_ns;
+		/* Adjust tEYE accordingly. */
+		tEYE += clock_period_in_ns;
+
+		/*
+		 * Decrease the ideal sample delay by one half cycle, to keep it
+		 * in the middle of the eye.
+		 */
+		ideal_sample_delay_in_ns -= (clock_period_in_ns >> 1);
+
+		/* ...and one less period for the delay time. */
+		ideal_sample_delay_in_ns -= clock_period_in_ns;
+
+		/* Jam a negative ideal sample delay to zero. */
+		if (ideal_sample_delay_in_ns < 0)
+			ideal_sample_delay_in_ns = 0;
+
+		/*
+		 * We have a new ideal sample delay, so re-compute the quantized
+		 * delay.
+		 */
+		sample_delay_factor =
+			ns_to_cycles(
+				ideal_sample_delay_in_ns << dll_delay_shift,
+							clock_period_in_ns, 0);
+
+		if (sample_delay_factor > nfc->max_sample_delay_factor)
+			sample_delay_factor = nfc->max_sample_delay_factor;
+	}
+
+	/* Control arrives here when we're ready to return our results. */
+return_results:
+	hw->data_setup_in_cycles    = data_setup_in_cycles;
+	hw->data_hold_in_cycles     = data_hold_in_cycles;
+	hw->address_setup_in_cycles = address_setup_in_cycles;
+	hw->use_half_periods        = dll_use_half_periods;
+	hw->sample_delay_factor     = sample_delay_factor;
+	hw->device_busy_timeout     = GPMI_DEFAULT_BUSY_TIMEOUT;
+	hw->wrn_dly_sel             = BV_GPMI_CTRL1_WRN_DLY_SEL_4_TO_8NS;
+
+	/* Return success. */
+	return 0;
+}
+
+/*
+ * <1> Firstly, we should know what's the GPMI-clock means.
+ *     The GPMI-clock is the internal clock in the gpmi nand controller.
+ *     If you set 100MHz to gpmi nand controller, the GPMI-clock's period
+ *     is 10ns. Mark the GPMI-clock's period as GPMI-clock-period.
+ *
+ * <2> Secondly, we should know what's the frequency on the nand chip pins.
+ *     The frequency on the nand chip pins is derived from the GPMI-clock.
+ *     We can get it from the following equation:
+ *
+ *         F = G / (DS + DH)
+ *
+ *         F  : the frequency on the nand chip pins.
+ *         G  : the GPMI clock, such as 100MHz.
+ *         DS : GPMI_HW_GPMI_TIMING0:DATA_SETUP
+ *         DH : GPMI_HW_GPMI_TIMING0:DATA_HOLD
+ *
+ * <3> Thirdly, when the frequency on the nand chip pins is above 33MHz,
+ *     the nand EDO(extended Data Out) timing could be applied.
+ *     The GPMI implements a feedback read strobe to sample the read data.
+ *     The feedback read strobe can be delayed to support the nand EDO timing
+ *     where the read strobe may deasserts before the read data is valid, and
+ *     read data is valid for some time after read strobe.
+ *
+ *     The following figure illustrates some aspects of a NAND Flash read:
+ *
+ *                   |<---tREA---->|
+ *                   |             |
+ *                   |         |   |
+ *                   |<--tRP-->|   |
+ *                   |         |   |
+ *                  __          ___|__________________________________
+ *     RDN            \________/   |
+ *                                 |
+ *                                 /---------\
+ *     Read Data    --------------<           >---------
+ *                                 \---------/
+ *                                |     |
+ *                                |<-D->|
+ *     FeedbackRDN  ________             ____________
+ *                          \___________/
+ *
+ *          D stands for delay, set in the HW_GPMI_CTRL1:RDN_DELAY.
+ *
+ *
+ * <4> Now, we begin to describe how to compute the right RDN_DELAY.
+ *
+ *  4.1) From the aspect of the nand chip pins:
+ *        Delay = (tREA + C - tRP)               {1}
+ *
+ *        tREA : the maximum read access time. From the ONFI nand standards,
+ *               we know that tREA is 16ns in mode 5, tREA is 20ns is mode 4.
+ *               Please check it in : www.onfi.org
+ *        C    : a constant for adjust the delay. default is 4.
+ *        tRP  : the read pulse width.
+ *               Specified by the HW_GPMI_TIMING0:DATA_SETUP:
+ *                    tRP = (GPMI-clock-period) * DATA_SETUP
+ *
+ *  4.2) From the aspect of the GPMI nand controller:
+ *         Delay = RDN_DELAY * 0.125 * RP        {2}
+ *
+ *         RP   : the DLL reference period.
+ *            if (GPMI-clock-period > DLL_THRETHOLD)
+ *                   RP = GPMI-clock-period / 2;
+ *            else
+ *                   RP = GPMI-clock-period;
+ *
+ *            Set the HW_GPMI_CTRL1:HALF_PERIOD if GPMI-clock-period
+ *            is greater DLL_THRETHOLD. In other SOCs, the DLL_THRETHOLD
+ *            is 16ns, but in mx6q, we use 12ns.
+ *
+ *  4.3) since {1} equals {2}, we get:
+ *
+ *                    (tREA + 4 - tRP) * 8
+ *         RDN_DELAY = ---------------------     {3}
+ *                           RP
+ *
+ *  4.4) We only support the fastest asynchronous mode of ONFI nand.
+ *       For some ONFI nand, the mode 4 is the fastest mode;
+ *       while for some ONFI nand, the mode 5 is the fastest mode.
+ *       So we only support the mode 4 and mode 5. It is no need to
+ *       support other modes.
+ */
+static void gpmi_compute_edo_timing(struct gpmi_nand_data *this,
+			struct gpmi_nfc_hardware_timing *hw)
+{
+	struct resources *r = &this->resources;
+	unsigned long rate = clk_get_rate(r->clock[0]);
+	int mode = this->timing_mode;
+	int dll_threshold = this->devdata->max_chain_delay;
+	unsigned long delay;
+	unsigned long clk_period;
+	int t_rea;
+	int c = 4;
+	int t_rp;
+	int rp;
+
+	/*
+	 * [1] for GPMI_HW_GPMI_TIMING0:
+	 *     The async mode requires 40MHz for mode 4, 50MHz for mode 5.
+	 *     The GPMI can support 100MHz at most. So if we want to
+	 *     get the 40MHz or 50MHz, we have to set DS=1, DH=1.
+	 *     Set the ADDRESS_SETUP to 0 in mode 4.
+	 */
+	hw->data_setup_in_cycles = 1;
+	hw->data_hold_in_cycles = 1;
+	hw->address_setup_in_cycles = ((mode == 5) ? 1 : 0);
+
+	/* [2] for GPMI_HW_GPMI_TIMING1 */
+	hw->device_busy_timeout = 0x9000;
+
+	/* [3] for GPMI_HW_GPMI_CTRL1 */
+	hw->wrn_dly_sel = BV_GPMI_CTRL1_WRN_DLY_SEL_NO_DELAY;
+
+	/*
+	 * Enlarge 10 times for the numerator and denominator in {3}.
+	 * This make us to get more accurate result.
+	 */
+	clk_period = NSEC_PER_SEC / (rate / 10);
+	dll_threshold *= 10;
+	t_rea = ((mode == 5) ? 16 : 20) * 10;
+	c *= 10;
+
+	t_rp = clk_period * 1; /* DATA_SETUP is 1 */
+
+	if (clk_period > dll_threshold) {
+		hw->use_half_periods = 1;
+		rp = clk_period / 2;
+	} else {
+		hw->use_half_periods = 0;
+		rp = clk_period;
+	}
+
+	/*
+	 * Multiply the numerator with 10, we could do a round off:
+	 *      7.8 round up to 8; 7.4 round down to 7.
+	 */
+	delay  = (((t_rea + c - t_rp) * 8) * 10) / rp;
+	delay = (delay + 5) / 10;
+
+	hw->sample_delay_factor = delay;
+}
+
+static int enable_edo_mode(struct gpmi_nand_data *this, int mode)
+{
+	struct resources  *r = &this->resources;
+	struct nand_chip *nand = &this->nand;
+	struct mtd_info	 *mtd = nand_to_mtd(nand);
+	uint8_t *feature;
+	unsigned long rate;
+	int ret;
+
+	feature = kzalloc(ONFI_SUBFEATURE_PARAM_LEN, GFP_KERNEL);
+	if (!feature)
+		return -ENOMEM;
+
+	nand->select_chip(mtd, 0);
+
+	/* [1] send SET FEATURE command to NAND */
+	feature[0] = mode;
+	ret = nand->onfi_set_features(mtd, nand,
+				ONFI_FEATURE_ADDR_TIMING_MODE, feature);
+	if (ret)
+		goto err_out;
+
+	/* [2] send GET FEATURE command to double-check the timing mode */
+	memset(feature, 0, ONFI_SUBFEATURE_PARAM_LEN);
+	ret = nand->onfi_get_features(mtd, nand,
+				ONFI_FEATURE_ADDR_TIMING_MODE, feature);
+	if (ret || feature[0] != mode)
+		goto err_out;
+
+	nand->select_chip(mtd, -1);
+
+	/* [3] set the main IO clock, 100MHz for mode 5, 80MHz for mode 4. */
+	rate = (mode == 5) ? 100000000 : 80000000;
+	clk_set_rate(r->clock[0], rate);
+
+	/* Let the gpmi_begin() re-compute the timing again. */
+	this->flags &= ~GPMI_TIMING_INIT_OK;
+
+	this->flags |= GPMI_ASYNC_EDO_ENABLED;
+	this->timing_mode = mode;
+	kfree(feature);
+	dev_info(this->dev, "enable the asynchronous EDO mode %d\n", mode);
+	return 0;
+
+err_out:
+	nand->select_chip(mtd, -1);
+	kfree(feature);
+	dev_err(this->dev, "mode:%d ,failed in set feature.\n", mode);
+	return -EINVAL;
+}
+
+int gpmi_extra_init(struct gpmi_nand_data *this)
+{
+	struct nand_chip *chip = &this->nand;
+
+	/* Enable the asynchronous EDO feature. */
+	if (GPMI_IS_MX6(this) && chip->onfi_version) {
+		int mode = onfi_get_async_timing_mode(chip);
+
+		/* We only support the timing mode 4 and mode 5. */
+		if (mode & ONFI_TIMING_MODE_5)
+			mode = 5;
+		else if (mode & ONFI_TIMING_MODE_4)
+			mode = 4;
+		else
+			return 0;
+
+		return enable_edo_mode(this, mode);
+	}
+	return 0;
+}
+
+/* Begin the I/O */
+void gpmi_begin(struct gpmi_nand_data *this)
+{
+	struct resources *r = &this->resources;
+	void __iomem *gpmi_regs = r->gpmi_regs;
+	unsigned int   clock_period_in_ns;
+	uint32_t       reg;
+	unsigned int   dll_wait_time_in_us;
+	struct gpmi_nfc_hardware_timing  hw;
+	int ret;
+
+	/* Enable the clock. */
+	ret = gpmi_enable_clk(this);
+	if (ret) {
+		dev_err(this->dev, "We failed in enable the clk\n");
+		goto err_out;
+	}
+
+	/* Only initialize the timing once */
+	if (this->flags & GPMI_TIMING_INIT_OK)
+		return;
+	this->flags |= GPMI_TIMING_INIT_OK;
+
+	if (this->flags & GPMI_ASYNC_EDO_ENABLED)
+		gpmi_compute_edo_timing(this, &hw);
+	else
+		gpmi_nfc_compute_hardware_timing(this, &hw);
+
+	/* [1] Set HW_GPMI_TIMING0 */
+	reg = BF_GPMI_TIMING0_ADDRESS_SETUP(hw.address_setup_in_cycles) |
+		BF_GPMI_TIMING0_DATA_HOLD(hw.data_hold_in_cycles)         |
+		BF_GPMI_TIMING0_DATA_SETUP(hw.data_setup_in_cycles);
+
+	writel(reg, gpmi_regs + HW_GPMI_TIMING0);
+
+	/* [2] Set HW_GPMI_TIMING1 */
+	writel(BF_GPMI_TIMING1_BUSY_TIMEOUT(hw.device_busy_timeout),
+		gpmi_regs + HW_GPMI_TIMING1);
+
+	/* [3] The following code is to set the HW_GPMI_CTRL1. */
+
+	/* Set the WRN_DLY_SEL */
+	writel(BM_GPMI_CTRL1_WRN_DLY_SEL, gpmi_regs + HW_GPMI_CTRL1_CLR);
+	writel(BF_GPMI_CTRL1_WRN_DLY_SEL(hw.wrn_dly_sel),
+					gpmi_regs + HW_GPMI_CTRL1_SET);
+
+	/* DLL_ENABLE must be set to 0 when setting RDN_DELAY or HALF_PERIOD. */
+	writel(BM_GPMI_CTRL1_DLL_ENABLE, gpmi_regs + HW_GPMI_CTRL1_CLR);
+
+	/* Clear out the DLL control fields. */
+	reg = BM_GPMI_CTRL1_RDN_DELAY | BM_GPMI_CTRL1_HALF_PERIOD;
+	writel(reg, gpmi_regs + HW_GPMI_CTRL1_CLR);
+
+	/* If no sample delay is called for, return immediately. */
+	if (!hw.sample_delay_factor)
+		return;
+
+	/* Set RDN_DELAY or HALF_PERIOD. */
+	reg = ((hw.use_half_periods) ? BM_GPMI_CTRL1_HALF_PERIOD : 0)
+		| BF_GPMI_CTRL1_RDN_DELAY(hw.sample_delay_factor);
+
+	writel(reg, gpmi_regs + HW_GPMI_CTRL1_SET);
+
+	/* At last, we enable the DLL. */
+	writel(BM_GPMI_CTRL1_DLL_ENABLE, gpmi_regs + HW_GPMI_CTRL1_SET);
+
+	/*
+	 * After we enable the GPMI DLL, we have to wait 64 clock cycles before
+	 * we can use the GPMI. Calculate the amount of time we need to wait,
+	 * in microseconds.
+	 */
+	clock_period_in_ns = NSEC_PER_SEC / clk_get_rate(r->clock[0]);
+	dll_wait_time_in_us = (clock_period_in_ns * 64) / 1000;
+
+	if (!dll_wait_time_in_us)
+		dll_wait_time_in_us = 1;
+
+	/* Wait for the DLL to settle. */
+	udelay(dll_wait_time_in_us);
+
+err_out:
+	return;
+}
+
+void gpmi_end(struct gpmi_nand_data *this)
+{
+	gpmi_disable_clk(this);
+}
+
+/* Clears a BCH interrupt. */
+void gpmi_clear_bch(struct gpmi_nand_data *this)
+{
+	struct resources *r = &this->resources;
+	writel(BM_BCH_CTRL_COMPLETE_IRQ, r->bch_regs + HW_BCH_CTRL_CLR);
+}
+
+/* Returns the Ready/Busy status of the given chip. */
+int gpmi_is_ready(struct gpmi_nand_data *this, unsigned chip)
+{
+	struct resources *r = &this->resources;
+	uint32_t mask = 0;
+	uint32_t reg = 0;
+
+	if (GPMI_IS_MX23(this)) {
+		mask = MX23_BM_GPMI_DEBUG_READY0 << chip;
+		reg = readl(r->gpmi_regs + HW_GPMI_DEBUG);
+	} else if (GPMI_IS_MX28(this) || GPMI_IS_MX6(this)) {
+		/*
+		 * In the imx6, all the ready/busy pins are bound
+		 * together. So we only need to check chip 0.
+		 */
+		if (GPMI_IS_MX6(this))
+			chip = 0;
+
+		/* MX28 shares the same R/B register as MX6Q. */
+		mask = MX28_BF_GPMI_STAT_READY_BUSY(1 << chip);
+		reg = readl(r->gpmi_regs + HW_GPMI_STAT);
+	} else
+		dev_err(this->dev, "unknown arch.\n");
+	return reg & mask;
+}
+
+static inline void set_dma_type(struct gpmi_nand_data *this,
+					enum dma_ops_type type)
+{
+	this->last_dma_type = this->dma_type;
+	this->dma_type = type;
+}
+
+int gpmi_send_command(struct gpmi_nand_data *this)
+{
+	struct dma_chan *channel = get_dma_chan(this);
+	struct dma_async_tx_descriptor *desc;
+	struct scatterlist *sgl;
+	int chip = this->current_chip;
+	u32 pio[3];
+
+	/* [1] send out the PIO words */
+	pio[0] = BF_GPMI_CTRL0_COMMAND_MODE(BV_GPMI_CTRL0_COMMAND_MODE__WRITE)
+		| BM_GPMI_CTRL0_WORD_LENGTH
+		| BF_GPMI_CTRL0_CS(chip, this)
+		| BF_GPMI_CTRL0_LOCK_CS(LOCK_CS_ENABLE, this)
+		| BF_GPMI_CTRL0_ADDRESS(BV_GPMI_CTRL0_ADDRESS__NAND_CLE)
+		| BM_GPMI_CTRL0_ADDRESS_INCREMENT
+		| BF_GPMI_CTRL0_XFER_COUNT(this->command_length);
+	pio[1] = pio[2] = 0;
+	desc = dmaengine_prep_slave_sg(channel,
+					(struct scatterlist *)pio,
+					ARRAY_SIZE(pio), DMA_TRANS_NONE, 0);
+	if (!desc)
+		return -EINVAL;
+
+	/* [2] send out the COMMAND + ADDRESS string stored in @buffer */
+	sgl = &this->cmd_sgl;
+
+	sg_init_one(sgl, this->cmd_buffer, this->command_length);
+	dma_map_sg(this->dev, sgl, 1, DMA_TO_DEVICE);
+	desc = dmaengine_prep_slave_sg(channel,
+				sgl, 1, DMA_MEM_TO_DEV,
+				DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+	if (!desc)
+		return -EINVAL;
+
+	/* [3] submit the DMA */
+	set_dma_type(this, DMA_FOR_COMMAND);
+	return start_dma_without_bch_irq(this, desc);
+}
+
+int gpmi_send_data(struct gpmi_nand_data *this)
+{
+	struct dma_async_tx_descriptor *desc;
+	struct dma_chan *channel = get_dma_chan(this);
+	int chip = this->current_chip;
+	uint32_t command_mode;
+	uint32_t address;
+	u32 pio[2];
+
+	/* [1] PIO */
+	command_mode = BV_GPMI_CTRL0_COMMAND_MODE__WRITE;
+	address      = BV_GPMI_CTRL0_ADDRESS__NAND_DATA;
+
+	pio[0] = BF_GPMI_CTRL0_COMMAND_MODE(command_mode)
+		| BM_GPMI_CTRL0_WORD_LENGTH
+		| BF_GPMI_CTRL0_CS(chip, this)
+		| BF_GPMI_CTRL0_LOCK_CS(LOCK_CS_ENABLE, this)
+		| BF_GPMI_CTRL0_ADDRESS(address)
+		| BF_GPMI_CTRL0_XFER_COUNT(this->upper_len);
+	pio[1] = 0;
+	desc = dmaengine_prep_slave_sg(channel, (struct scatterlist *)pio,
+					ARRAY_SIZE(pio), DMA_TRANS_NONE, 0);
+	if (!desc)
+		return -EINVAL;
+
+	/* [2] send DMA request */
+	prepare_data_dma(this, DMA_TO_DEVICE);
+	desc = dmaengine_prep_slave_sg(channel, &this->data_sgl,
+					1, DMA_MEM_TO_DEV,
+					DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+	if (!desc)
+		return -EINVAL;
+
+	/* [3] submit the DMA */
+	set_dma_type(this, DMA_FOR_WRITE_DATA);
+	return start_dma_without_bch_irq(this, desc);
+}
+
+int gpmi_read_data(struct gpmi_nand_data *this)
+{
+	struct dma_async_tx_descriptor *desc;
+	struct dma_chan *channel = get_dma_chan(this);
+	int chip = this->current_chip;
+	u32 pio[2];
+
+	/* [1] : send PIO */
+	pio[0] = BF_GPMI_CTRL0_COMMAND_MODE(BV_GPMI_CTRL0_COMMAND_MODE__READ)
+		| BM_GPMI_CTRL0_WORD_LENGTH
+		| BF_GPMI_CTRL0_CS(chip, this)
+		| BF_GPMI_CTRL0_LOCK_CS(LOCK_CS_ENABLE, this)
+		| BF_GPMI_CTRL0_ADDRESS(BV_GPMI_CTRL0_ADDRESS__NAND_DATA)
+		| BF_GPMI_CTRL0_XFER_COUNT(this->upper_len);
+	pio[1] = 0;
+	desc = dmaengine_prep_slave_sg(channel,
+					(struct scatterlist *)pio,
+					ARRAY_SIZE(pio), DMA_TRANS_NONE, 0);
+	if (!desc)
+		return -EINVAL;
+
+	/* [2] : send DMA request */
+	prepare_data_dma(this, DMA_FROM_DEVICE);
+	desc = dmaengine_prep_slave_sg(channel, &this->data_sgl,
+					1, DMA_DEV_TO_MEM,
+					DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+	if (!desc)
+		return -EINVAL;
+
+	/* [3] : submit the DMA */
+	set_dma_type(this, DMA_FOR_READ_DATA);
+	return start_dma_without_bch_irq(this, desc);
+}
+
+int gpmi_send_page(struct gpmi_nand_data *this,
+			dma_addr_t payload, dma_addr_t auxiliary)
+{
+	struct bch_geometry *geo = &this->bch_geometry;
+	uint32_t command_mode;
+	uint32_t address;
+	uint32_t ecc_command;
+	uint32_t buffer_mask;
+	struct dma_async_tx_descriptor *desc;
+	struct dma_chan *channel = get_dma_chan(this);
+	int chip = this->current_chip;
+	u32 pio[6];
+
+	/* A DMA descriptor that does an ECC page read. */
+	command_mode = BV_GPMI_CTRL0_COMMAND_MODE__WRITE;
+	address      = BV_GPMI_CTRL0_ADDRESS__NAND_DATA;
+	ecc_command  = BV_GPMI_ECCCTRL_ECC_CMD__BCH_ENCODE;
+	buffer_mask  = BV_GPMI_ECCCTRL_BUFFER_MASK__BCH_PAGE |
+				BV_GPMI_ECCCTRL_BUFFER_MASK__BCH_AUXONLY;
+
+	pio[0] = BF_GPMI_CTRL0_COMMAND_MODE(command_mode)
+		| BM_GPMI_CTRL0_WORD_LENGTH
+		| BF_GPMI_CTRL0_CS(chip, this)
+		| BF_GPMI_CTRL0_LOCK_CS(LOCK_CS_ENABLE, this)
+		| BF_GPMI_CTRL0_ADDRESS(address)
+		| BF_GPMI_CTRL0_XFER_COUNT(0);
+	pio[1] = 0;
+	pio[2] = BM_GPMI_ECCCTRL_ENABLE_ECC
+		| BF_GPMI_ECCCTRL_ECC_CMD(ecc_command)
+		| BF_GPMI_ECCCTRL_BUFFER_MASK(buffer_mask);
+	pio[3] = geo->page_size;
+	pio[4] = payload;
+	pio[5] = auxiliary;
+
+	desc = dmaengine_prep_slave_sg(channel,
+					(struct scatterlist *)pio,
+					ARRAY_SIZE(pio), DMA_TRANS_NONE,
+					DMA_CTRL_ACK);
+	if (!desc)
+		return -EINVAL;
+
+	set_dma_type(this, DMA_FOR_WRITE_ECC_PAGE);
+	return start_dma_with_bch_irq(this, desc);
+}
+
+int gpmi_read_page(struct gpmi_nand_data *this,
+				dma_addr_t payload, dma_addr_t auxiliary)
+{
+	struct bch_geometry *geo = &this->bch_geometry;
+	uint32_t command_mode;
+	uint32_t address;
+	uint32_t ecc_command;
+	uint32_t buffer_mask;
+	struct dma_async_tx_descriptor *desc;
+	struct dma_chan *channel = get_dma_chan(this);
+	int chip = this->current_chip;
+	u32 pio[6];
+
+	/* [1] Wait for the chip to report ready. */
+	command_mode = BV_GPMI_CTRL0_COMMAND_MODE__WAIT_FOR_READY;
+	address      = BV_GPMI_CTRL0_ADDRESS__NAND_DATA;
+
+	pio[0] =  BF_GPMI_CTRL0_COMMAND_MODE(command_mode)
+		| BM_GPMI_CTRL0_WORD_LENGTH
+		| BF_GPMI_CTRL0_CS(chip, this)
+		| BF_GPMI_CTRL0_LOCK_CS(LOCK_CS_ENABLE, this)
+		| BF_GPMI_CTRL0_ADDRESS(address)
+		| BF_GPMI_CTRL0_XFER_COUNT(0);
+	pio[1] = 0;
+	desc = dmaengine_prep_slave_sg(channel,
+				(struct scatterlist *)pio, 2,
+				DMA_TRANS_NONE, 0);
+	if (!desc)
+		return -EINVAL;
+
+	/* [2] Enable the BCH block and read. */
+	command_mode = BV_GPMI_CTRL0_COMMAND_MODE__READ;
+	address      = BV_GPMI_CTRL0_ADDRESS__NAND_DATA;
+	ecc_command  = BV_GPMI_ECCCTRL_ECC_CMD__BCH_DECODE;
+	buffer_mask  = BV_GPMI_ECCCTRL_BUFFER_MASK__BCH_PAGE
+			| BV_GPMI_ECCCTRL_BUFFER_MASK__BCH_AUXONLY;
+
+	pio[0] =  BF_GPMI_CTRL0_COMMAND_MODE(command_mode)
+		| BM_GPMI_CTRL0_WORD_LENGTH
+		| BF_GPMI_CTRL0_CS(chip, this)
+		| BF_GPMI_CTRL0_LOCK_CS(LOCK_CS_ENABLE, this)
+		| BF_GPMI_CTRL0_ADDRESS(address)
+		| BF_GPMI_CTRL0_XFER_COUNT(geo->page_size);
+
+	pio[1] = 0;
+	pio[2] =  BM_GPMI_ECCCTRL_ENABLE_ECC
+		| BF_GPMI_ECCCTRL_ECC_CMD(ecc_command)
+		| BF_GPMI_ECCCTRL_BUFFER_MASK(buffer_mask);
+	pio[3] = geo->page_size;
+	pio[4] = payload;
+	pio[5] = auxiliary;
+	desc = dmaengine_prep_slave_sg(channel,
+					(struct scatterlist *)pio,
+					ARRAY_SIZE(pio), DMA_TRANS_NONE,
+					DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+	if (!desc)
+		return -EINVAL;
+
+	/* [3] Disable the BCH block */
+	command_mode = BV_GPMI_CTRL0_COMMAND_MODE__WAIT_FOR_READY;
+	address      = BV_GPMI_CTRL0_ADDRESS__NAND_DATA;
+
+	pio[0] = BF_GPMI_CTRL0_COMMAND_MODE(command_mode)
+		| BM_GPMI_CTRL0_WORD_LENGTH
+		| BF_GPMI_CTRL0_CS(chip, this)
+		| BF_GPMI_CTRL0_LOCK_CS(LOCK_CS_ENABLE, this)
+		| BF_GPMI_CTRL0_ADDRESS(address)
+		| BF_GPMI_CTRL0_XFER_COUNT(geo->page_size);
+	pio[1] = 0;
+	pio[2] = 0; /* clear GPMI_HW_GPMI_ECCCTRL, disable the BCH. */
+	desc = dmaengine_prep_slave_sg(channel,
+				(struct scatterlist *)pio, 3,
+				DMA_TRANS_NONE,
+				DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+	if (!desc)
+		return -EINVAL;
+
+	/* [4] submit the DMA */
+	set_dma_type(this, DMA_FOR_READ_ECC_PAGE);
+	return start_dma_with_bch_irq(this, desc);
+}
+
+/**
+ * gpmi_copy_bits - copy bits from one memory region to another
+ * @dst: destination buffer
+ * @dst_bit_off: bit offset we're starting to write at
+ * @src: source buffer
+ * @src_bit_off: bit offset we're starting to read from
+ * @nbits: number of bits to copy
+ *
+ * This functions copies bits from one memory region to another, and is used by
+ * the GPMI driver to copy ECC sections which are not guaranteed to be byte
+ * aligned.
+ *
+ * src and dst should not overlap.
+ *
+ */
+void gpmi_copy_bits(u8 *dst, size_t dst_bit_off,
+		    const u8 *src, size_t src_bit_off,
+		    size_t nbits)
+{
+	size_t i;
+	size_t nbytes;
+	u32 src_buffer = 0;
+	size_t bits_in_src_buffer = 0;
+
+	if (!nbits)
+		return;
+
+	/*
+	 * Move src and dst pointers to the closest byte pointer and store bit
+	 * offsets within a byte.
+	 */
+	src += src_bit_off / 8;
+	src_bit_off %= 8;
+
+	dst += dst_bit_off / 8;
+	dst_bit_off %= 8;
+
+	/*
+	 * Initialize the src_buffer value with bits available in the first
+	 * byte of data so that we end up with a byte aligned src pointer.
+	 */
+	if (src_bit_off) {
+		src_buffer = src[0] >> src_bit_off;
+		if (nbits >= (8 - src_bit_off)) {
+			bits_in_src_buffer += 8 - src_bit_off;
+		} else {
+			src_buffer &= GENMASK(nbits - 1, 0);
+			bits_in_src_buffer += nbits;
+		}
+		nbits -= bits_in_src_buffer;
+		src++;
+	}
+
+	/* Calculate the number of bytes that can be copied from src to dst. */
+	nbytes = nbits / 8;
+
+	/* Try to align dst to a byte boundary. */
+	if (dst_bit_off) {
+		if (bits_in_src_buffer < (8 - dst_bit_off) && nbytes) {
+			src_buffer |= src[0] << bits_in_src_buffer;
+			bits_in_src_buffer += 8;
+			src++;
+			nbytes--;
+		}
+
+		if (bits_in_src_buffer >= (8 - dst_bit_off)) {
+			dst[0] &= GENMASK(dst_bit_off - 1, 0);
+			dst[0] |= src_buffer << dst_bit_off;
+			src_buffer >>= (8 - dst_bit_off);
+			bits_in_src_buffer -= (8 - dst_bit_off);
+			dst_bit_off = 0;
+			dst++;
+			if (bits_in_src_buffer > 7) {
+				bits_in_src_buffer -= 8;
+				dst[0] = src_buffer;
+				dst++;
+				src_buffer >>= 8;
+			}
+		}
+	}
+
+	if (!bits_in_src_buffer && !dst_bit_off) {
+		/*
+		 * Both src and dst pointers are byte aligned, thus we can
+		 * just use the optimized memcpy function.
+		 */
+		if (nbytes)
+			memcpy(dst, src, nbytes);
+	} else {
+		/*
+		 * src buffer is not byte aligned, hence we have to copy each
+		 * src byte to the src_buffer variable before extracting a byte
+		 * to store in dst.
+		 */
+		for (i = 0; i < nbytes; i++) {
+			src_buffer |= src[i] << bits_in_src_buffer;
+			dst[i] = src_buffer;
+			src_buffer >>= 8;
+		}
+	}
+	/* Update dst and src pointers */
+	dst += nbytes;
+	src += nbytes;
+
+	/*
+	 * nbits is the number of remaining bits. It should not exceed 8 as
+	 * we've already copied as much bytes as possible.
+	 */
+	nbits %= 8;
+
+	/*
+	 * If there's no more bits to copy to the destination and src buffer
+	 * was already byte aligned, then we're done.
+	 */
+	if (!nbits && !bits_in_src_buffer)
+		return;
+
+	/* Copy the remaining bits to src_buffer */
+	if (nbits)
+		src_buffer |= (*src & GENMASK(nbits - 1, 0)) <<
+			      bits_in_src_buffer;
+	bits_in_src_buffer += nbits;
+
+	/*
+	 * In case there were not enough bits to get a byte aligned dst buffer
+	 * prepare the src_buffer variable to match the dst organization (shift
+	 * src_buffer by dst_bit_off and retrieve the least significant bits
+	 * from dst).
+	 */
+	if (dst_bit_off)
+		src_buffer = (src_buffer << dst_bit_off) |
+			     (*dst & GENMASK(dst_bit_off - 1, 0));
+	bits_in_src_buffer += dst_bit_off;
+
+	/*
+	 * Keep most significant bits from dst if we end up with an unaligned
+	 * number of bits.
+	 */
+	nbytes = bits_in_src_buffer / 8;
+	if (bits_in_src_buffer % 8) {
+		src_buffer |= (dst[nbytes] &
+			       GENMASK(7, bits_in_src_buffer % 8)) <<
+			      (nbytes * 8);
+		nbytes++;
+	}
+
+	/* Copy the remaining bytes to dst */
+	for (i = 0; i < nbytes; i++) {
+		dst[i] = src_buffer;
+		src_buffer >>= 8;
+	}
+}
diff --git a/drivers/mtd/nand/raw/gpmi-nand/gpmi-nand.c b/drivers/mtd/nand/raw/gpmi-nand/gpmi-nand.c
new file mode 100644
index 0000000..61fdd73
--- /dev/null
+++ b/drivers/mtd/nand/raw/gpmi-nand/gpmi-nand.c
@@ -0,0 +1,2182 @@
+/*
+ * Freescale GPMI NAND Flash Driver
+ *
+ * Copyright (C) 2010-2015 Freescale Semiconductor, Inc.
+ * Copyright (C) 2008 Embedded Alley Solutions, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+#include <linux/clk.h>
+#include <linux/slab.h>
+#include <linux/sched/task_stack.h>
+#include <linux/interrupt.h>
+#include <linux/module.h>
+#include <linux/mtd/partitions.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include "gpmi-nand.h"
+#include "bch-regs.h"
+
+/* Resource names for the GPMI NAND driver. */
+#define GPMI_NAND_GPMI_REGS_ADDR_RES_NAME  "gpmi-nand"
+#define GPMI_NAND_BCH_REGS_ADDR_RES_NAME   "bch"
+#define GPMI_NAND_BCH_INTERRUPT_RES_NAME   "bch"
+
+/* add our owner bbt descriptor */
+static uint8_t scan_ff_pattern[] = { 0xff };
+static struct nand_bbt_descr gpmi_bbt_descr = {
+	.options	= 0,
+	.offs		= 0,
+	.len		= 1,
+	.pattern	= scan_ff_pattern
+};
+
+/*
+ * We may change the layout if we can get the ECC info from the datasheet,
+ * else we will use all the (page + OOB).
+ */
+static int gpmi_ooblayout_ecc(struct mtd_info *mtd, int section,
+			      struct mtd_oob_region *oobregion)
+{
+	struct nand_chip *chip = mtd_to_nand(mtd);
+	struct gpmi_nand_data *this = nand_get_controller_data(chip);
+	struct bch_geometry *geo = &this->bch_geometry;
+
+	if (section)
+		return -ERANGE;
+
+	oobregion->offset = 0;
+	oobregion->length = geo->page_size - mtd->writesize;
+
+	return 0;
+}
+
+static int gpmi_ooblayout_free(struct mtd_info *mtd, int section,
+			       struct mtd_oob_region *oobregion)
+{
+	struct nand_chip *chip = mtd_to_nand(mtd);
+	struct gpmi_nand_data *this = nand_get_controller_data(chip);
+	struct bch_geometry *geo = &this->bch_geometry;
+
+	if (section)
+		return -ERANGE;
+
+	/* The available oob size we have. */
+	if (geo->page_size < mtd->writesize + mtd->oobsize) {
+		oobregion->offset = geo->page_size - mtd->writesize;
+		oobregion->length = mtd->oobsize - oobregion->offset;
+	}
+
+	return 0;
+}
+
+static const char * const gpmi_clks_for_mx2x[] = {
+	"gpmi_io",
+};
+
+static const struct mtd_ooblayout_ops gpmi_ooblayout_ops = {
+	.ecc = gpmi_ooblayout_ecc,
+	.free = gpmi_ooblayout_free,
+};
+
+static const struct gpmi_devdata gpmi_devdata_imx23 = {
+	.type = IS_MX23,
+	.bch_max_ecc_strength = 20,
+	.max_chain_delay = 16,
+	.clks = gpmi_clks_for_mx2x,
+	.clks_count = ARRAY_SIZE(gpmi_clks_for_mx2x),
+};
+
+static const struct gpmi_devdata gpmi_devdata_imx28 = {
+	.type = IS_MX28,
+	.bch_max_ecc_strength = 20,
+	.max_chain_delay = 16,
+	.clks = gpmi_clks_for_mx2x,
+	.clks_count = ARRAY_SIZE(gpmi_clks_for_mx2x),
+};
+
+static const char * const gpmi_clks_for_mx6[] = {
+	"gpmi_io", "gpmi_apb", "gpmi_bch", "gpmi_bch_apb", "per1_bch",
+};
+
+static const struct gpmi_devdata gpmi_devdata_imx6q = {
+	.type = IS_MX6Q,
+	.bch_max_ecc_strength = 40,
+	.max_chain_delay = 12,
+	.clks = gpmi_clks_for_mx6,
+	.clks_count = ARRAY_SIZE(gpmi_clks_for_mx6),
+};
+
+static const struct gpmi_devdata gpmi_devdata_imx6sx = {
+	.type = IS_MX6SX,
+	.bch_max_ecc_strength = 62,
+	.max_chain_delay = 12,
+	.clks = gpmi_clks_for_mx6,
+	.clks_count = ARRAY_SIZE(gpmi_clks_for_mx6),
+};
+
+static const char * const gpmi_clks_for_mx7d[] = {
+	"gpmi_io", "gpmi_bch_apb",
+};
+
+static const struct gpmi_devdata gpmi_devdata_imx7d = {
+	.type = IS_MX7D,
+	.bch_max_ecc_strength = 62,
+	.max_chain_delay = 12,
+	.clks = gpmi_clks_for_mx7d,
+	.clks_count = ARRAY_SIZE(gpmi_clks_for_mx7d),
+};
+
+static irqreturn_t bch_irq(int irq, void *cookie)
+{
+	struct gpmi_nand_data *this = cookie;
+
+	gpmi_clear_bch(this);
+	complete(&this->bch_done);
+	return IRQ_HANDLED;
+}
+
+/*
+ *  Calculate the ECC strength by hand:
+ *	E : The ECC strength.
+ *	G : the length of Galois Field.
+ *	N : The chunk count of per page.
+ *	O : the oobsize of the NAND chip.
+ *	M : the metasize of per page.
+ *
+ *	The formula is :
+ *		E * G * N
+ *	      ------------ <= (O - M)
+ *                  8
+ *
+ *      So, we get E by:
+ *                    (O - M) * 8
+ *              E <= -------------
+ *                       G * N
+ */
+static inline int get_ecc_strength(struct gpmi_nand_data *this)
+{
+	struct bch_geometry *geo = &this->bch_geometry;
+	struct mtd_info	*mtd = nand_to_mtd(&this->nand);
+	int ecc_strength;
+
+	ecc_strength = ((mtd->oobsize - geo->metadata_size) * 8)
+			/ (geo->gf_len * geo->ecc_chunk_count);
+
+	/* We need the minor even number. */
+	return round_down(ecc_strength, 2);
+}
+
+static inline bool gpmi_check_ecc(struct gpmi_nand_data *this)
+{
+	struct bch_geometry *geo = &this->bch_geometry;
+
+	/* Do the sanity check. */
+	if (GPMI_IS_MX23(this) || GPMI_IS_MX28(this)) {
+		/* The mx23/mx28 only support the GF13. */
+		if (geo->gf_len == 14)
+			return false;
+	}
+	return geo->ecc_strength <= this->devdata->bch_max_ecc_strength;
+}
+
+/*
+ * If we can get the ECC information from the nand chip, we do not
+ * need to calculate them ourselves.
+ *
+ * We may have available oob space in this case.
+ */
+static int set_geometry_by_ecc_info(struct gpmi_nand_data *this)
+{
+	struct bch_geometry *geo = &this->bch_geometry;
+	struct nand_chip *chip = &this->nand;
+	struct mtd_info *mtd = nand_to_mtd(chip);
+	unsigned int block_mark_bit_offset;
+
+	if (!(chip->ecc_strength_ds > 0 && chip->ecc_step_ds > 0))
+		return -EINVAL;
+
+	switch (chip->ecc_step_ds) {
+	case SZ_512:
+		geo->gf_len = 13;
+		break;
+	case SZ_1K:
+		geo->gf_len = 14;
+		break;
+	default:
+		dev_err(this->dev,
+			"unsupported nand chip. ecc bits : %d, ecc size : %d\n",
+			chip->ecc_strength_ds, chip->ecc_step_ds);
+		return -EINVAL;
+	}
+	geo->ecc_chunk_size = chip->ecc_step_ds;
+	geo->ecc_strength = round_up(chip->ecc_strength_ds, 2);
+	if (!gpmi_check_ecc(this))
+		return -EINVAL;
+
+	/* Keep the C >= O */
+	if (geo->ecc_chunk_size < mtd->oobsize) {
+		dev_err(this->dev,
+			"unsupported nand chip. ecc size: %d, oob size : %d\n",
+			chip->ecc_step_ds, mtd->oobsize);
+		return -EINVAL;
+	}
+
+	/* The default value, see comment in the legacy_set_geometry(). */
+	geo->metadata_size = 10;
+
+	geo->ecc_chunk_count = mtd->writesize / geo->ecc_chunk_size;
+
+	/*
+	 * Now, the NAND chip with 2K page(data chunk is 512byte) shows below:
+	 *
+	 *    |                          P                            |
+	 *    |<----------------------------------------------------->|
+	 *    |                                                       |
+	 *    |                                        (Block Mark)   |
+	 *    |                      P'                      |      | |     |
+	 *    |<-------------------------------------------->|  D   | |  O' |
+	 *    |                                              |<---->| |<--->|
+	 *    V                                              V      V V     V
+	 *    +---+----------+-+----------+-+----------+-+----------+-+-----+
+	 *    | M |   data   |E|   data   |E|   data   |E|   data   |E|     |
+	 *    +---+----------+-+----------+-+----------+-+----------+-+-----+
+	 *                                                   ^              ^
+	 *                                                   |      O       |
+	 *                                                   |<------------>|
+	 *                                                   |              |
+	 *
+	 *	P : the page size for BCH module.
+	 *	E : The ECC strength.
+	 *	G : the length of Galois Field.
+	 *	N : The chunk count of per page.
+	 *	M : the metasize of per page.
+	 *	C : the ecc chunk size, aka the "data" above.
+	 *	P': the nand chip's page size.
+	 *	O : the nand chip's oob size.
+	 *	O': the free oob.
+	 *
+	 *	The formula for P is :
+	 *
+	 *	            E * G * N
+	 *	       P = ------------ + P' + M
+	 *                      8
+	 *
+	 * The position of block mark moves forward in the ECC-based view
+	 * of page, and the delta is:
+	 *
+	 *                   E * G * (N - 1)
+	 *             D = (---------------- + M)
+	 *                          8
+	 *
+	 * Please see the comment in legacy_set_geometry().
+	 * With the condition C >= O , we still can get same result.
+	 * So the bit position of the physical block mark within the ECC-based
+	 * view of the page is :
+	 *             (P' - D) * 8
+	 */
+	geo->page_size = mtd->writesize + geo->metadata_size +
+		(geo->gf_len * geo->ecc_strength * geo->ecc_chunk_count) / 8;
+
+	geo->payload_size = mtd->writesize;
+
+	geo->auxiliary_status_offset = ALIGN(geo->metadata_size, 4);
+	geo->auxiliary_size = ALIGN(geo->metadata_size, 4)
+				+ ALIGN(geo->ecc_chunk_count, 4);
+
+	if (!this->swap_block_mark)
+		return 0;
+
+	/* For bit swap. */
+	block_mark_bit_offset = mtd->writesize * 8 -
+		(geo->ecc_strength * geo->gf_len * (geo->ecc_chunk_count - 1)
+				+ geo->metadata_size * 8);
+
+	geo->block_mark_byte_offset = block_mark_bit_offset / 8;
+	geo->block_mark_bit_offset  = block_mark_bit_offset % 8;
+	return 0;
+}
+
+static int legacy_set_geometry(struct gpmi_nand_data *this)
+{
+	struct bch_geometry *geo = &this->bch_geometry;
+	struct mtd_info *mtd = nand_to_mtd(&this->nand);
+	unsigned int metadata_size;
+	unsigned int status_size;
+	unsigned int block_mark_bit_offset;
+
+	/*
+	 * The size of the metadata can be changed, though we set it to 10
+	 * bytes now. But it can't be too large, because we have to save
+	 * enough space for BCH.
+	 */
+	geo->metadata_size = 10;
+
+	/* The default for the length of Galois Field. */
+	geo->gf_len = 13;
+
+	/* The default for chunk size. */
+	geo->ecc_chunk_size = 512;
+	while (geo->ecc_chunk_size < mtd->oobsize) {
+		geo->ecc_chunk_size *= 2; /* keep C >= O */
+		geo->gf_len = 14;
+	}
+
+	geo->ecc_chunk_count = mtd->writesize / geo->ecc_chunk_size;
+
+	/* We use the same ECC strength for all chunks. */
+	geo->ecc_strength = get_ecc_strength(this);
+	if (!gpmi_check_ecc(this)) {
+		dev_err(this->dev,
+			"ecc strength: %d cannot be supported by the controller (%d)\n"
+			"try to use minimum ecc strength that NAND chip required\n",
+			geo->ecc_strength,
+			this->devdata->bch_max_ecc_strength);
+		return -EINVAL;
+	}
+
+	geo->page_size = mtd->writesize + geo->metadata_size +
+		(geo->gf_len * geo->ecc_strength * geo->ecc_chunk_count) / 8;
+	geo->payload_size = mtd->writesize;
+
+	/*
+	 * The auxiliary buffer contains the metadata and the ECC status. The
+	 * metadata is padded to the nearest 32-bit boundary. The ECC status
+	 * contains one byte for every ECC chunk, and is also padded to the
+	 * nearest 32-bit boundary.
+	 */
+	metadata_size = ALIGN(geo->metadata_size, 4);
+	status_size   = ALIGN(geo->ecc_chunk_count, 4);
+
+	geo->auxiliary_size = metadata_size + status_size;
+	geo->auxiliary_status_offset = metadata_size;
+
+	if (!this->swap_block_mark)
+		return 0;
+
+	/*
+	 * We need to compute the byte and bit offsets of
+	 * the physical block mark within the ECC-based view of the page.
+	 *
+	 * NAND chip with 2K page shows below:
+	 *                                             (Block Mark)
+	 *                                                   |      |
+	 *                                                   |  D   |
+	 *                                                   |<---->|
+	 *                                                   V      V
+	 *    +---+----------+-+----------+-+----------+-+----------+-+
+	 *    | M |   data   |E|   data   |E|   data   |E|   data   |E|
+	 *    +---+----------+-+----------+-+----------+-+----------+-+
+	 *
+	 * The position of block mark moves forward in the ECC-based view
+	 * of page, and the delta is:
+	 *
+	 *                   E * G * (N - 1)
+	 *             D = (---------------- + M)
+	 *                          8
+	 *
+	 * With the formula to compute the ECC strength, and the condition
+	 *       : C >= O         (C is the ecc chunk size)
+	 *
+	 * It's easy to deduce to the following result:
+	 *
+	 *         E * G       (O - M)      C - M         C - M
+	 *      ----------- <= ------- <=  --------  <  ---------
+	 *           8            N           N          (N - 1)
+	 *
+	 *  So, we get:
+	 *
+	 *                   E * G * (N - 1)
+	 *             D = (---------------- + M) < C
+	 *                          8
+	 *
+	 *  The above inequality means the position of block mark
+	 *  within the ECC-based view of the page is still in the data chunk,
+	 *  and it's NOT in the ECC bits of the chunk.
+	 *
+	 *  Use the following to compute the bit position of the
+	 *  physical block mark within the ECC-based view of the page:
+	 *          (page_size - D) * 8
+	 *
+	 *  --Huang Shijie
+	 */
+	block_mark_bit_offset = mtd->writesize * 8 -
+		(geo->ecc_strength * geo->gf_len * (geo->ecc_chunk_count - 1)
+				+ geo->metadata_size * 8);
+
+	geo->block_mark_byte_offset = block_mark_bit_offset / 8;
+	geo->block_mark_bit_offset  = block_mark_bit_offset % 8;
+	return 0;
+}
+
+int common_nfc_set_geometry(struct gpmi_nand_data *this)
+{
+	if ((of_property_read_bool(this->dev->of_node, "fsl,use-minimum-ecc"))
+				|| legacy_set_geometry(this))
+		return set_geometry_by_ecc_info(this);
+
+	return 0;
+}
+
+struct dma_chan *get_dma_chan(struct gpmi_nand_data *this)
+{
+	/* We use the DMA channel 0 to access all the nand chips. */
+	return this->dma_chans[0];
+}
+
+/* Can we use the upper's buffer directly for DMA? */
+void prepare_data_dma(struct gpmi_nand_data *this, enum dma_data_direction dr)
+{
+	struct scatterlist *sgl = &this->data_sgl;
+	int ret;
+
+	/* first try to map the upper buffer directly */
+	if (virt_addr_valid(this->upper_buf) &&
+		!object_is_on_stack(this->upper_buf)) {
+		sg_init_one(sgl, this->upper_buf, this->upper_len);
+		ret = dma_map_sg(this->dev, sgl, 1, dr);
+		if (ret == 0)
+			goto map_fail;
+
+		this->direct_dma_map_ok = true;
+		return;
+	}
+
+map_fail:
+	/* We have to use our own DMA buffer. */
+	sg_init_one(sgl, this->data_buffer_dma, this->upper_len);
+
+	if (dr == DMA_TO_DEVICE)
+		memcpy(this->data_buffer_dma, this->upper_buf, this->upper_len);
+
+	dma_map_sg(this->dev, sgl, 1, dr);
+
+	this->direct_dma_map_ok = false;
+}
+
+/* This will be called after the DMA operation is finished. */
+static void dma_irq_callback(void *param)
+{
+	struct gpmi_nand_data *this = param;
+	struct completion *dma_c = &this->dma_done;
+
+	switch (this->dma_type) {
+	case DMA_FOR_COMMAND:
+		dma_unmap_sg(this->dev, &this->cmd_sgl, 1, DMA_TO_DEVICE);
+		break;
+
+	case DMA_FOR_READ_DATA:
+		dma_unmap_sg(this->dev, &this->data_sgl, 1, DMA_FROM_DEVICE);
+		if (this->direct_dma_map_ok == false)
+			memcpy(this->upper_buf, this->data_buffer_dma,
+				this->upper_len);
+		break;
+
+	case DMA_FOR_WRITE_DATA:
+		dma_unmap_sg(this->dev, &this->data_sgl, 1, DMA_TO_DEVICE);
+		break;
+
+	case DMA_FOR_READ_ECC_PAGE:
+	case DMA_FOR_WRITE_ECC_PAGE:
+		/* We have to wait the BCH interrupt to finish. */
+		break;
+
+	default:
+		dev_err(this->dev, "in wrong DMA operation.\n");
+	}
+
+	complete(dma_c);
+}
+
+int start_dma_without_bch_irq(struct gpmi_nand_data *this,
+				struct dma_async_tx_descriptor *desc)
+{
+	struct completion *dma_c = &this->dma_done;
+	unsigned long timeout;
+
+	init_completion(dma_c);
+
+	desc->callback		= dma_irq_callback;
+	desc->callback_param	= this;
+	dmaengine_submit(desc);
+	dma_async_issue_pending(get_dma_chan(this));
+
+	/* Wait for the interrupt from the DMA block. */
+	timeout = wait_for_completion_timeout(dma_c, msecs_to_jiffies(1000));
+	if (!timeout) {
+		dev_err(this->dev, "DMA timeout, last DMA :%d\n",
+			this->last_dma_type);
+		gpmi_dump_info(this);
+		return -ETIMEDOUT;
+	}
+	return 0;
+}
+
+/*
+ * This function is used in BCH reading or BCH writing pages.
+ * It will wait for the BCH interrupt as long as ONE second.
+ * Actually, we must wait for two interrupts :
+ *	[1] firstly the DMA interrupt and
+ *	[2] secondly the BCH interrupt.
+ */
+int start_dma_with_bch_irq(struct gpmi_nand_data *this,
+			struct dma_async_tx_descriptor *desc)
+{
+	struct completion *bch_c = &this->bch_done;
+	unsigned long timeout;
+
+	/* Prepare to receive an interrupt from the BCH block. */
+	init_completion(bch_c);
+
+	/* start the DMA */
+	start_dma_without_bch_irq(this, desc);
+
+	/* Wait for the interrupt from the BCH block. */
+	timeout = wait_for_completion_timeout(bch_c, msecs_to_jiffies(1000));
+	if (!timeout) {
+		dev_err(this->dev, "BCH timeout, last DMA :%d\n",
+			this->last_dma_type);
+		gpmi_dump_info(this);
+		return -ETIMEDOUT;
+	}
+	return 0;
+}
+
+static int acquire_register_block(struct gpmi_nand_data *this,
+				  const char *res_name)
+{
+	struct platform_device *pdev = this->pdev;
+	struct resources *res = &this->resources;
+	struct resource *r;
+	void __iomem *p;
+
+	r = platform_get_resource_byname(pdev, IORESOURCE_MEM, res_name);
+	p = devm_ioremap_resource(&pdev->dev, r);
+	if (IS_ERR(p))
+		return PTR_ERR(p);
+
+	if (!strcmp(res_name, GPMI_NAND_GPMI_REGS_ADDR_RES_NAME))
+		res->gpmi_regs = p;
+	else if (!strcmp(res_name, GPMI_NAND_BCH_REGS_ADDR_RES_NAME))
+		res->bch_regs = p;
+	else
+		dev_err(this->dev, "unknown resource name : %s\n", res_name);
+
+	return 0;
+}
+
+static int acquire_bch_irq(struct gpmi_nand_data *this, irq_handler_t irq_h)
+{
+	struct platform_device *pdev = this->pdev;
+	const char *res_name = GPMI_NAND_BCH_INTERRUPT_RES_NAME;
+	struct resource *r;
+	int err;
+
+	r = platform_get_resource_byname(pdev, IORESOURCE_IRQ, res_name);
+	if (!r) {
+		dev_err(this->dev, "Can't get resource for %s\n", res_name);
+		return -ENODEV;
+	}
+
+	err = devm_request_irq(this->dev, r->start, irq_h, 0, res_name, this);
+	if (err)
+		dev_err(this->dev, "error requesting BCH IRQ\n");
+
+	return err;
+}
+
+static void release_dma_channels(struct gpmi_nand_data *this)
+{
+	unsigned int i;
+	for (i = 0; i < DMA_CHANS; i++)
+		if (this->dma_chans[i]) {
+			dma_release_channel(this->dma_chans[i]);
+			this->dma_chans[i] = NULL;
+		}
+}
+
+static int acquire_dma_channels(struct gpmi_nand_data *this)
+{
+	struct platform_device *pdev = this->pdev;
+	struct dma_chan *dma_chan;
+
+	/* request dma channel */
+	dma_chan = dma_request_slave_channel(&pdev->dev, "rx-tx");
+	if (!dma_chan) {
+		dev_err(this->dev, "Failed to request DMA channel.\n");
+		goto acquire_err;
+	}
+
+	this->dma_chans[0] = dma_chan;
+	return 0;
+
+acquire_err:
+	release_dma_channels(this);
+	return -EINVAL;
+}
+
+static int gpmi_get_clks(struct gpmi_nand_data *this)
+{
+	struct resources *r = &this->resources;
+	struct clk *clk;
+	int err, i;
+
+	for (i = 0; i < this->devdata->clks_count; i++) {
+		clk = devm_clk_get(this->dev, this->devdata->clks[i]);
+		if (IS_ERR(clk)) {
+			err = PTR_ERR(clk);
+			goto err_clock;
+		}
+
+		r->clock[i] = clk;
+	}
+
+	if (GPMI_IS_MX6(this))
+		/*
+		 * Set the default value for the gpmi clock.
+		 *
+		 * If you want to use the ONFI nand which is in the
+		 * Synchronous Mode, you should change the clock as you need.
+		 */
+		clk_set_rate(r->clock[0], 22000000);
+
+	return 0;
+
+err_clock:
+	dev_dbg(this->dev, "failed in finding the clocks.\n");
+	return err;
+}
+
+static int acquire_resources(struct gpmi_nand_data *this)
+{
+	int ret;
+
+	ret = acquire_register_block(this, GPMI_NAND_GPMI_REGS_ADDR_RES_NAME);
+	if (ret)
+		goto exit_regs;
+
+	ret = acquire_register_block(this, GPMI_NAND_BCH_REGS_ADDR_RES_NAME);
+	if (ret)
+		goto exit_regs;
+
+	ret = acquire_bch_irq(this, bch_irq);
+	if (ret)
+		goto exit_regs;
+
+	ret = acquire_dma_channels(this);
+	if (ret)
+		goto exit_regs;
+
+	ret = gpmi_get_clks(this);
+	if (ret)
+		goto exit_clock;
+	return 0;
+
+exit_clock:
+	release_dma_channels(this);
+exit_regs:
+	return ret;
+}
+
+static void release_resources(struct gpmi_nand_data *this)
+{
+	release_dma_channels(this);
+}
+
+static int init_hardware(struct gpmi_nand_data *this)
+{
+	int ret;
+
+	/*
+	 * This structure contains the "safe" GPMI timing that should succeed
+	 * with any NAND Flash device
+	 * (although, with less-than-optimal performance).
+	 */
+	struct nand_timing  safe_timing = {
+		.data_setup_in_ns        = 80,
+		.data_hold_in_ns         = 60,
+		.address_setup_in_ns     = 25,
+		.gpmi_sample_delay_in_ns =  6,
+		.tREA_in_ns              = -1,
+		.tRLOH_in_ns             = -1,
+		.tRHOH_in_ns             = -1,
+	};
+
+	/* Initialize the hardwares. */
+	ret = gpmi_init(this);
+	if (ret)
+		return ret;
+
+	this->timing = safe_timing;
+	return 0;
+}
+
+static int read_page_prepare(struct gpmi_nand_data *this,
+			void *destination, unsigned length,
+			void *alt_virt, dma_addr_t alt_phys, unsigned alt_size,
+			void **use_virt, dma_addr_t *use_phys)
+{
+	struct device *dev = this->dev;
+
+	if (virt_addr_valid(destination)) {
+		dma_addr_t dest_phys;
+
+		dest_phys = dma_map_single(dev, destination,
+						length, DMA_FROM_DEVICE);
+		if (dma_mapping_error(dev, dest_phys)) {
+			if (alt_size < length) {
+				dev_err(dev, "Alternate buffer is too small\n");
+				return -ENOMEM;
+			}
+			goto map_failed;
+		}
+		*use_virt = destination;
+		*use_phys = dest_phys;
+		this->direct_dma_map_ok = true;
+		return 0;
+	}
+
+map_failed:
+	*use_virt = alt_virt;
+	*use_phys = alt_phys;
+	this->direct_dma_map_ok = false;
+	return 0;
+}
+
+static inline void read_page_end(struct gpmi_nand_data *this,
+			void *destination, unsigned length,
+			void *alt_virt, dma_addr_t alt_phys, unsigned alt_size,
+			void *used_virt, dma_addr_t used_phys)
+{
+	if (this->direct_dma_map_ok)
+		dma_unmap_single(this->dev, used_phys, length, DMA_FROM_DEVICE);
+}
+
+static inline void read_page_swap_end(struct gpmi_nand_data *this,
+			void *destination, unsigned length,
+			void *alt_virt, dma_addr_t alt_phys, unsigned alt_size,
+			void *used_virt, dma_addr_t used_phys)
+{
+	if (!this->direct_dma_map_ok)
+		memcpy(destination, alt_virt, length);
+}
+
+static int send_page_prepare(struct gpmi_nand_data *this,
+			const void *source, unsigned length,
+			void *alt_virt, dma_addr_t alt_phys, unsigned alt_size,
+			const void **use_virt, dma_addr_t *use_phys)
+{
+	struct device *dev = this->dev;
+
+	if (virt_addr_valid(source)) {
+		dma_addr_t source_phys;
+
+		source_phys = dma_map_single(dev, (void *)source, length,
+						DMA_TO_DEVICE);
+		if (dma_mapping_error(dev, source_phys)) {
+			if (alt_size < length) {
+				dev_err(dev, "Alternate buffer is too small\n");
+				return -ENOMEM;
+			}
+			goto map_failed;
+		}
+		*use_virt = source;
+		*use_phys = source_phys;
+		return 0;
+	}
+map_failed:
+	/*
+	 * Copy the content of the source buffer into the alternate
+	 * buffer and set up the return values accordingly.
+	 */
+	memcpy(alt_virt, source, length);
+
+	*use_virt = alt_virt;
+	*use_phys = alt_phys;
+	return 0;
+}
+
+static void send_page_end(struct gpmi_nand_data *this,
+			const void *source, unsigned length,
+			void *alt_virt, dma_addr_t alt_phys, unsigned alt_size,
+			const void *used_virt, dma_addr_t used_phys)
+{
+	struct device *dev = this->dev;
+	if (used_virt == source)
+		dma_unmap_single(dev, used_phys, length, DMA_TO_DEVICE);
+}
+
+static void gpmi_free_dma_buffer(struct gpmi_nand_data *this)
+{
+	struct device *dev = this->dev;
+
+	if (this->page_buffer_virt && virt_addr_valid(this->page_buffer_virt))
+		dma_free_coherent(dev, this->page_buffer_size,
+					this->page_buffer_virt,
+					this->page_buffer_phys);
+	kfree(this->cmd_buffer);
+	kfree(this->data_buffer_dma);
+	kfree(this->raw_buffer);
+
+	this->cmd_buffer	= NULL;
+	this->data_buffer_dma	= NULL;
+	this->raw_buffer	= NULL;
+	this->page_buffer_virt	= NULL;
+	this->page_buffer_size	=  0;
+}
+
+/* Allocate the DMA buffers */
+static int gpmi_alloc_dma_buffer(struct gpmi_nand_data *this)
+{
+	struct bch_geometry *geo = &this->bch_geometry;
+	struct device *dev = this->dev;
+	struct mtd_info *mtd = nand_to_mtd(&this->nand);
+
+	/* [1] Allocate a command buffer. PAGE_SIZE is enough. */
+	this->cmd_buffer = kzalloc(PAGE_SIZE, GFP_DMA | GFP_KERNEL);
+	if (this->cmd_buffer == NULL)
+		goto error_alloc;
+
+	/*
+	 * [2] Allocate a read/write data buffer.
+	 *     The gpmi_alloc_dma_buffer can be called twice.
+	 *     We allocate a PAGE_SIZE length buffer if gpmi_alloc_dma_buffer
+	 *     is called before the nand_scan_ident; and we allocate a buffer
+	 *     of the real NAND page size when the gpmi_alloc_dma_buffer is
+	 *     called after the nand_scan_ident.
+	 */
+	this->data_buffer_dma = kzalloc(mtd->writesize ?: PAGE_SIZE,
+					GFP_DMA | GFP_KERNEL);
+	if (this->data_buffer_dma == NULL)
+		goto error_alloc;
+
+	/*
+	 * [3] Allocate the page buffer.
+	 *
+	 * Both the payload buffer and the auxiliary buffer must appear on
+	 * 32-bit boundaries. We presume the size of the payload buffer is a
+	 * power of two and is much larger than four, which guarantees the
+	 * auxiliary buffer will appear on a 32-bit boundary.
+	 */
+	this->page_buffer_size = geo->payload_size + geo->auxiliary_size;
+	this->page_buffer_virt = dma_alloc_coherent(dev, this->page_buffer_size,
+					&this->page_buffer_phys, GFP_DMA);
+	if (!this->page_buffer_virt)
+		goto error_alloc;
+
+	this->raw_buffer = kzalloc(mtd->writesize + mtd->oobsize, GFP_KERNEL);
+	if (!this->raw_buffer)
+		goto error_alloc;
+
+	/* Slice up the page buffer. */
+	this->payload_virt = this->page_buffer_virt;
+	this->payload_phys = this->page_buffer_phys;
+	this->auxiliary_virt = this->payload_virt + geo->payload_size;
+	this->auxiliary_phys = this->payload_phys + geo->payload_size;
+	return 0;
+
+error_alloc:
+	gpmi_free_dma_buffer(this);
+	return -ENOMEM;
+}
+
+static void gpmi_cmd_ctrl(struct mtd_info *mtd, int data, unsigned int ctrl)
+{
+	struct nand_chip *chip = mtd_to_nand(mtd);
+	struct gpmi_nand_data *this = nand_get_controller_data(chip);
+	int ret;
+
+	/*
+	 * Every operation begins with a command byte and a series of zero or
+	 * more address bytes. These are distinguished by either the Address
+	 * Latch Enable (ALE) or Command Latch Enable (CLE) signals being
+	 * asserted. When MTD is ready to execute the command, it will deassert
+	 * both latch enables.
+	 *
+	 * Rather than run a separate DMA operation for every single byte, we
+	 * queue them up and run a single DMA operation for the entire series
+	 * of command and data bytes. NAND_CMD_NONE means the END of the queue.
+	 */
+	if ((ctrl & (NAND_ALE | NAND_CLE))) {
+		if (data != NAND_CMD_NONE)
+			this->cmd_buffer[this->command_length++] = data;
+		return;
+	}
+
+	if (!this->command_length)
+		return;
+
+	ret = gpmi_send_command(this);
+	if (ret)
+		dev_err(this->dev, "Chip: %u, Error %d\n",
+			this->current_chip, ret);
+
+	this->command_length = 0;
+}
+
+static int gpmi_dev_ready(struct mtd_info *mtd)
+{
+	struct nand_chip *chip = mtd_to_nand(mtd);
+	struct gpmi_nand_data *this = nand_get_controller_data(chip);
+
+	return gpmi_is_ready(this, this->current_chip);
+}
+
+static void gpmi_select_chip(struct mtd_info *mtd, int chipnr)
+{
+	struct nand_chip *chip = mtd_to_nand(mtd);
+	struct gpmi_nand_data *this = nand_get_controller_data(chip);
+
+	if ((this->current_chip < 0) && (chipnr >= 0))
+		gpmi_begin(this);
+	else if ((this->current_chip >= 0) && (chipnr < 0))
+		gpmi_end(this);
+
+	this->current_chip = chipnr;
+}
+
+static void gpmi_read_buf(struct mtd_info *mtd, uint8_t *buf, int len)
+{
+	struct nand_chip *chip = mtd_to_nand(mtd);
+	struct gpmi_nand_data *this = nand_get_controller_data(chip);
+
+	dev_dbg(this->dev, "len is %d\n", len);
+	this->upper_buf	= buf;
+	this->upper_len	= len;
+
+	gpmi_read_data(this);
+}
+
+static void gpmi_write_buf(struct mtd_info *mtd, const uint8_t *buf, int len)
+{
+	struct nand_chip *chip = mtd_to_nand(mtd);
+	struct gpmi_nand_data *this = nand_get_controller_data(chip);
+
+	dev_dbg(this->dev, "len is %d\n", len);
+	this->upper_buf	= (uint8_t *)buf;
+	this->upper_len	= len;
+
+	gpmi_send_data(this);
+}
+
+static uint8_t gpmi_read_byte(struct mtd_info *mtd)
+{
+	struct nand_chip *chip = mtd_to_nand(mtd);
+	struct gpmi_nand_data *this = nand_get_controller_data(chip);
+	uint8_t *buf = this->data_buffer_dma;
+
+	gpmi_read_buf(mtd, buf, 1);
+	return buf[0];
+}
+
+/*
+ * Handles block mark swapping.
+ * It can be called in swapping the block mark, or swapping it back,
+ * because the the operations are the same.
+ */
+static void block_mark_swapping(struct gpmi_nand_data *this,
+				void *payload, void *auxiliary)
+{
+	struct bch_geometry *nfc_geo = &this->bch_geometry;
+	unsigned char *p;
+	unsigned char *a;
+	unsigned int  bit;
+	unsigned char mask;
+	unsigned char from_data;
+	unsigned char from_oob;
+
+	if (!this->swap_block_mark)
+		return;
+
+	/*
+	 * If control arrives here, we're swapping. Make some convenience
+	 * variables.
+	 */
+	bit = nfc_geo->block_mark_bit_offset;
+	p   = payload + nfc_geo->block_mark_byte_offset;
+	a   = auxiliary;
+
+	/*
+	 * Get the byte from the data area that overlays the block mark. Since
+	 * the ECC engine applies its own view to the bits in the page, the
+	 * physical block mark won't (in general) appear on a byte boundary in
+	 * the data.
+	 */
+	from_data = (p[0] >> bit) | (p[1] << (8 - bit));
+
+	/* Get the byte from the OOB. */
+	from_oob = a[0];
+
+	/* Swap them. */
+	a[0] = from_data;
+
+	mask = (0x1 << bit) - 1;
+	p[0] = (p[0] & mask) | (from_oob << bit);
+
+	mask = ~0 << bit;
+	p[1] = (p[1] & mask) | (from_oob >> (8 - bit));
+}
+
+static int gpmi_ecc_read_page_data(struct nand_chip *chip,
+				   uint8_t *buf, int oob_required,
+				   int page)
+{
+	struct gpmi_nand_data *this = nand_get_controller_data(chip);
+	struct bch_geometry *nfc_geo = &this->bch_geometry;
+	struct mtd_info *mtd = nand_to_mtd(chip);
+	void          *payload_virt;
+	dma_addr_t    payload_phys;
+	void          *auxiliary_virt;
+	dma_addr_t    auxiliary_phys;
+	unsigned int  i;
+	unsigned char *status;
+	unsigned int  max_bitflips = 0;
+	int           ret;
+
+	dev_dbg(this->dev, "page number is : %d\n", page);
+	ret = read_page_prepare(this, buf, nfc_geo->payload_size,
+					this->payload_virt, this->payload_phys,
+					nfc_geo->payload_size,
+					&payload_virt, &payload_phys);
+	if (ret) {
+		dev_err(this->dev, "Inadequate DMA buffer\n");
+		ret = -ENOMEM;
+		return ret;
+	}
+	auxiliary_virt = this->auxiliary_virt;
+	auxiliary_phys = this->auxiliary_phys;
+
+	/* go! */
+	ret = gpmi_read_page(this, payload_phys, auxiliary_phys);
+	read_page_end(this, buf, nfc_geo->payload_size,
+			this->payload_virt, this->payload_phys,
+			nfc_geo->payload_size,
+			payload_virt, payload_phys);
+	if (ret) {
+		dev_err(this->dev, "Error in ECC-based read: %d\n", ret);
+		return ret;
+	}
+
+	/* Loop over status bytes, accumulating ECC status. */
+	status = auxiliary_virt + nfc_geo->auxiliary_status_offset;
+
+	read_page_swap_end(this, buf, nfc_geo->payload_size,
+			   this->payload_virt, this->payload_phys,
+			   nfc_geo->payload_size,
+			   payload_virt, payload_phys);
+
+	for (i = 0; i < nfc_geo->ecc_chunk_count; i++, status++) {
+		if ((*status == STATUS_GOOD) || (*status == STATUS_ERASED))
+			continue;
+
+		if (*status == STATUS_UNCORRECTABLE) {
+			int eccbits = nfc_geo->ecc_strength * nfc_geo->gf_len;
+			u8 *eccbuf = this->raw_buffer;
+			int offset, bitoffset;
+			int eccbytes;
+			int flips;
+
+			/* Read ECC bytes into our internal raw_buffer */
+			offset = nfc_geo->metadata_size * 8;
+			offset += ((8 * nfc_geo->ecc_chunk_size) + eccbits) * (i + 1);
+			offset -= eccbits;
+			bitoffset = offset % 8;
+			eccbytes = DIV_ROUND_UP(offset + eccbits, 8);
+			offset /= 8;
+			eccbytes -= offset;
+			nand_change_read_column_op(chip, offset, eccbuf,
+						   eccbytes, false);
+
+			/*
+			 * ECC data are not byte aligned and we may have
+			 * in-band data in the first and last byte of
+			 * eccbuf. Set non-eccbits to one so that
+			 * nand_check_erased_ecc_chunk() does not count them
+			 * as bitflips.
+			 */
+			if (bitoffset)
+				eccbuf[0] |= GENMASK(bitoffset - 1, 0);
+
+			bitoffset = (bitoffset + eccbits) % 8;
+			if (bitoffset)
+				eccbuf[eccbytes - 1] |= GENMASK(7, bitoffset);
+
+			/*
+			 * The ECC hardware has an uncorrectable ECC status
+			 * code in case we have bitflips in an erased page. As
+			 * nothing was written into this subpage the ECC is
+			 * obviously wrong and we can not trust it. We assume
+			 * at this point that we are reading an erased page and
+			 * try to correct the bitflips in buffer up to
+			 * ecc_strength bitflips. If this is a page with random
+			 * data, we exceed this number of bitflips and have a
+			 * ECC failure. Otherwise we use the corrected buffer.
+			 */
+			if (i == 0) {
+				/* The first block includes metadata */
+				flips = nand_check_erased_ecc_chunk(
+						buf + i * nfc_geo->ecc_chunk_size,
+						nfc_geo->ecc_chunk_size,
+						eccbuf, eccbytes,
+						auxiliary_virt,
+						nfc_geo->metadata_size,
+						nfc_geo->ecc_strength);
+			} else {
+				flips = nand_check_erased_ecc_chunk(
+						buf + i * nfc_geo->ecc_chunk_size,
+						nfc_geo->ecc_chunk_size,
+						eccbuf, eccbytes,
+						NULL, 0,
+						nfc_geo->ecc_strength);
+			}
+
+			if (flips > 0) {
+				max_bitflips = max_t(unsigned int, max_bitflips,
+						     flips);
+				mtd->ecc_stats.corrected += flips;
+				continue;
+			}
+
+			mtd->ecc_stats.failed++;
+			continue;
+		}
+
+		mtd->ecc_stats.corrected += *status;
+		max_bitflips = max_t(unsigned int, max_bitflips, *status);
+	}
+
+	/* handle the block mark swapping */
+	block_mark_swapping(this, buf, auxiliary_virt);
+
+	if (oob_required) {
+		/*
+		 * It's time to deliver the OOB bytes. See gpmi_ecc_read_oob()
+		 * for details about our policy for delivering the OOB.
+		 *
+		 * We fill the caller's buffer with set bits, and then copy the
+		 * block mark to th caller's buffer. Note that, if block mark
+		 * swapping was necessary, it has already been done, so we can
+		 * rely on the first byte of the auxiliary buffer to contain
+		 * the block mark.
+		 */
+		memset(chip->oob_poi, ~0, mtd->oobsize);
+		chip->oob_poi[0] = ((uint8_t *) auxiliary_virt)[0];
+	}
+
+	return max_bitflips;
+}
+
+static int gpmi_ecc_read_page(struct mtd_info *mtd, struct nand_chip *chip,
+			      uint8_t *buf, int oob_required, int page)
+{
+	nand_read_page_op(chip, page, 0, NULL, 0);
+
+	return gpmi_ecc_read_page_data(chip, buf, oob_required, page);
+}
+
+/* Fake a virtual small page for the subpage read */
+static int gpmi_ecc_read_subpage(struct mtd_info *mtd, struct nand_chip *chip,
+			uint32_t offs, uint32_t len, uint8_t *buf, int page)
+{
+	struct gpmi_nand_data *this = nand_get_controller_data(chip);
+	void __iomem *bch_regs = this->resources.bch_regs;
+	struct bch_geometry old_geo = this->bch_geometry;
+	struct bch_geometry *geo = &this->bch_geometry;
+	int size = chip->ecc.size; /* ECC chunk size */
+	int meta, n, page_size;
+	u32 r1_old, r2_old, r1_new, r2_new;
+	unsigned int max_bitflips;
+	int first, last, marker_pos;
+	int ecc_parity_size;
+	int col = 0;
+	int old_swap_block_mark = this->swap_block_mark;
+
+	/* The size of ECC parity */
+	ecc_parity_size = geo->gf_len * geo->ecc_strength / 8;
+
+	/* Align it with the chunk size */
+	first = offs / size;
+	last = (offs + len - 1) / size;
+
+	if (this->swap_block_mark) {
+		/*
+		 * Find the chunk which contains the Block Marker.
+		 * If this chunk is in the range of [first, last],
+		 * we have to read out the whole page.
+		 * Why? since we had swapped the data at the position of Block
+		 * Marker to the metadata which is bound with the chunk 0.
+		 */
+		marker_pos = geo->block_mark_byte_offset / size;
+		if (last >= marker_pos && first <= marker_pos) {
+			dev_dbg(this->dev,
+				"page:%d, first:%d, last:%d, marker at:%d\n",
+				page, first, last, marker_pos);
+			return gpmi_ecc_read_page(mtd, chip, buf, 0, page);
+		}
+	}
+
+	meta = geo->metadata_size;
+	if (first) {
+		col = meta + (size + ecc_parity_size) * first;
+		meta = 0;
+		buf = buf + first * size;
+	}
+
+	nand_read_page_op(chip, page, col, NULL, 0);
+
+	/* Save the old environment */
+	r1_old = r1_new = readl(bch_regs + HW_BCH_FLASH0LAYOUT0);
+	r2_old = r2_new = readl(bch_regs + HW_BCH_FLASH0LAYOUT1);
+
+	/* change the BCH registers and bch_geometry{} */
+	n = last - first + 1;
+	page_size = meta + (size + ecc_parity_size) * n;
+
+	r1_new &= ~(BM_BCH_FLASH0LAYOUT0_NBLOCKS |
+			BM_BCH_FLASH0LAYOUT0_META_SIZE);
+	r1_new |= BF_BCH_FLASH0LAYOUT0_NBLOCKS(n - 1)
+			| BF_BCH_FLASH0LAYOUT0_META_SIZE(meta);
+	writel(r1_new, bch_regs + HW_BCH_FLASH0LAYOUT0);
+
+	r2_new &= ~BM_BCH_FLASH0LAYOUT1_PAGE_SIZE;
+	r2_new |= BF_BCH_FLASH0LAYOUT1_PAGE_SIZE(page_size);
+	writel(r2_new, bch_regs + HW_BCH_FLASH0LAYOUT1);
+
+	geo->ecc_chunk_count = n;
+	geo->payload_size = n * size;
+	geo->page_size = page_size;
+	geo->auxiliary_status_offset = ALIGN(meta, 4);
+
+	dev_dbg(this->dev, "page:%d(%d:%d)%d, chunk:(%d:%d), BCH PG size:%d\n",
+		page, offs, len, col, first, n, page_size);
+
+	/* Read the subpage now */
+	this->swap_block_mark = false;
+	max_bitflips = gpmi_ecc_read_page_data(chip, buf, 0, page);
+
+	/* Restore */
+	writel(r1_old, bch_regs + HW_BCH_FLASH0LAYOUT0);
+	writel(r2_old, bch_regs + HW_BCH_FLASH0LAYOUT1);
+	this->bch_geometry = old_geo;
+	this->swap_block_mark = old_swap_block_mark;
+
+	return max_bitflips;
+}
+
+static int gpmi_ecc_write_page(struct mtd_info *mtd, struct nand_chip *chip,
+				const uint8_t *buf, int oob_required, int page)
+{
+	struct gpmi_nand_data *this = nand_get_controller_data(chip);
+	struct bch_geometry *nfc_geo = &this->bch_geometry;
+	const void *payload_virt;
+	dma_addr_t payload_phys;
+	const void *auxiliary_virt;
+	dma_addr_t auxiliary_phys;
+	int        ret;
+
+	dev_dbg(this->dev, "ecc write page.\n");
+
+	nand_prog_page_begin_op(chip, page, 0, NULL, 0);
+
+	if (this->swap_block_mark) {
+		/*
+		 * If control arrives here, we're doing block mark swapping.
+		 * Since we can't modify the caller's buffers, we must copy them
+		 * into our own.
+		 */
+		memcpy(this->payload_virt, buf, mtd->writesize);
+		payload_virt = this->payload_virt;
+		payload_phys = this->payload_phys;
+
+		memcpy(this->auxiliary_virt, chip->oob_poi,
+				nfc_geo->auxiliary_size);
+		auxiliary_virt = this->auxiliary_virt;
+		auxiliary_phys = this->auxiliary_phys;
+
+		/* Handle block mark swapping. */
+		block_mark_swapping(this,
+				(void *)payload_virt, (void *)auxiliary_virt);
+	} else {
+		/*
+		 * If control arrives here, we're not doing block mark swapping,
+		 * so we can to try and use the caller's buffers.
+		 */
+		ret = send_page_prepare(this,
+				buf, mtd->writesize,
+				this->payload_virt, this->payload_phys,
+				nfc_geo->payload_size,
+				&payload_virt, &payload_phys);
+		if (ret) {
+			dev_err(this->dev, "Inadequate payload DMA buffer\n");
+			return 0;
+		}
+
+		ret = send_page_prepare(this,
+				chip->oob_poi, mtd->oobsize,
+				this->auxiliary_virt, this->auxiliary_phys,
+				nfc_geo->auxiliary_size,
+				&auxiliary_virt, &auxiliary_phys);
+		if (ret) {
+			dev_err(this->dev, "Inadequate auxiliary DMA buffer\n");
+			goto exit_auxiliary;
+		}
+	}
+
+	/* Ask the NFC. */
+	ret = gpmi_send_page(this, payload_phys, auxiliary_phys);
+	if (ret)
+		dev_err(this->dev, "Error in ECC-based write: %d\n", ret);
+
+	if (!this->swap_block_mark) {
+		send_page_end(this, chip->oob_poi, mtd->oobsize,
+				this->auxiliary_virt, this->auxiliary_phys,
+				nfc_geo->auxiliary_size,
+				auxiliary_virt, auxiliary_phys);
+exit_auxiliary:
+		send_page_end(this, buf, mtd->writesize,
+				this->payload_virt, this->payload_phys,
+				nfc_geo->payload_size,
+				payload_virt, payload_phys);
+	}
+
+	if (ret)
+		return ret;
+
+	return nand_prog_page_end_op(chip);
+}
+
+/*
+ * There are several places in this driver where we have to handle the OOB and
+ * block marks. This is the function where things are the most complicated, so
+ * this is where we try to explain it all. All the other places refer back to
+ * here.
+ *
+ * These are the rules, in order of decreasing importance:
+ *
+ * 1) Nothing the caller does can be allowed to imperil the block mark.
+ *
+ * 2) In read operations, the first byte of the OOB we return must reflect the
+ *    true state of the block mark, no matter where that block mark appears in
+ *    the physical page.
+ *
+ * 3) ECC-based read operations return an OOB full of set bits (since we never
+ *    allow ECC-based writes to the OOB, it doesn't matter what ECC-based reads
+ *    return).
+ *
+ * 4) "Raw" read operations return a direct view of the physical bytes in the
+ *    page, using the conventional definition of which bytes are data and which
+ *    are OOB. This gives the caller a way to see the actual, physical bytes
+ *    in the page, without the distortions applied by our ECC engine.
+ *
+ *
+ * What we do for this specific read operation depends on two questions:
+ *
+ * 1) Are we doing a "raw" read, or an ECC-based read?
+ *
+ * 2) Are we using block mark swapping or transcription?
+ *
+ * There are four cases, illustrated by the following Karnaugh map:
+ *
+ *                    |           Raw           |         ECC-based       |
+ *       -------------+-------------------------+-------------------------+
+ *                    | Read the conventional   |                         |
+ *                    | OOB at the end of the   |                         |
+ *       Swapping     | page and return it. It  |                         |
+ *                    | contains exactly what   |                         |
+ *                    | we want.                | Read the block mark and |
+ *       -------------+-------------------------+ return it in a buffer   |
+ *                    | Read the conventional   | full of set bits.       |
+ *                    | OOB at the end of the   |                         |
+ *                    | page and also the block |                         |
+ *       Transcribing | mark in the metadata.   |                         |
+ *                    | Copy the block mark     |                         |
+ *                    | into the first byte of  |                         |
+ *                    | the OOB.                |                         |
+ *       -------------+-------------------------+-------------------------+
+ *
+ * Note that we break rule #4 in the Transcribing/Raw case because we're not
+ * giving an accurate view of the actual, physical bytes in the page (we're
+ * overwriting the block mark). That's OK because it's more important to follow
+ * rule #2.
+ *
+ * It turns out that knowing whether we want an "ECC-based" or "raw" read is not
+ * easy. When reading a page, for example, the NAND Flash MTD code calls our
+ * ecc.read_page or ecc.read_page_raw function. Thus, the fact that MTD wants an
+ * ECC-based or raw view of the page is implicit in which function it calls
+ * (there is a similar pair of ECC-based/raw functions for writing).
+ */
+static int gpmi_ecc_read_oob(struct mtd_info *mtd, struct nand_chip *chip,
+				int page)
+{
+	struct gpmi_nand_data *this = nand_get_controller_data(chip);
+
+	dev_dbg(this->dev, "page number is %d\n", page);
+	/* clear the OOB buffer */
+	memset(chip->oob_poi, ~0, mtd->oobsize);
+
+	/* Read out the conventional OOB. */
+	nand_read_page_op(chip, page, mtd->writesize, NULL, 0);
+	chip->read_buf(mtd, chip->oob_poi, mtd->oobsize);
+
+	/*
+	 * Now, we want to make sure the block mark is correct. In the
+	 * non-transcribing case (!GPMI_IS_MX23()), we already have it.
+	 * Otherwise, we need to explicitly read it.
+	 */
+	if (GPMI_IS_MX23(this)) {
+		/* Read the block mark into the first byte of the OOB buffer. */
+		nand_read_page_op(chip, page, 0, NULL, 0);
+		chip->oob_poi[0] = chip->read_byte(mtd);
+	}
+
+	return 0;
+}
+
+static int
+gpmi_ecc_write_oob(struct mtd_info *mtd, struct nand_chip *chip, int page)
+{
+	struct mtd_oob_region of = { };
+
+	/* Do we have available oob area? */
+	mtd_ooblayout_free(mtd, 0, &of);
+	if (!of.length)
+		return -EPERM;
+
+	if (!nand_is_slc(chip))
+		return -EPERM;
+
+	return nand_prog_page_op(chip, page, mtd->writesize + of.offset,
+				 chip->oob_poi + of.offset, of.length);
+}
+
+/*
+ * This function reads a NAND page without involving the ECC engine (no HW
+ * ECC correction).
+ * The tricky part in the GPMI/BCH controller is that it stores ECC bits
+ * inline (interleaved with payload DATA), and do not align data chunk on
+ * byte boundaries.
+ * We thus need to take care moving the payload data and ECC bits stored in the
+ * page into the provided buffers, which is why we're using gpmi_copy_bits.
+ *
+ * See set_geometry_by_ecc_info inline comments to have a full description
+ * of the layout used by the GPMI controller.
+ */
+static int gpmi_ecc_read_page_raw(struct mtd_info *mtd,
+				  struct nand_chip *chip, uint8_t *buf,
+				  int oob_required, int page)
+{
+	struct gpmi_nand_data *this = nand_get_controller_data(chip);
+	struct bch_geometry *nfc_geo = &this->bch_geometry;
+	int eccsize = nfc_geo->ecc_chunk_size;
+	int eccbits = nfc_geo->ecc_strength * nfc_geo->gf_len;
+	u8 *tmp_buf = this->raw_buffer;
+	size_t src_bit_off;
+	size_t oob_bit_off;
+	size_t oob_byte_off;
+	uint8_t *oob = chip->oob_poi;
+	int step;
+
+	nand_read_page_op(chip, page, 0, tmp_buf,
+			  mtd->writesize + mtd->oobsize);
+
+	/*
+	 * If required, swap the bad block marker and the data stored in the
+	 * metadata section, so that we don't wrongly consider a block as bad.
+	 *
+	 * See the layout description for a detailed explanation on why this
+	 * is needed.
+	 */
+	if (this->swap_block_mark)
+		swap(tmp_buf[0], tmp_buf[mtd->writesize]);
+
+	/*
+	 * Copy the metadata section into the oob buffer (this section is
+	 * guaranteed to be aligned on a byte boundary).
+	 */
+	if (oob_required)
+		memcpy(oob, tmp_buf, nfc_geo->metadata_size);
+
+	oob_bit_off = nfc_geo->metadata_size * 8;
+	src_bit_off = oob_bit_off;
+
+	/* Extract interleaved payload data and ECC bits */
+	for (step = 0; step < nfc_geo->ecc_chunk_count; step++) {
+		if (buf)
+			gpmi_copy_bits(buf, step * eccsize * 8,
+				       tmp_buf, src_bit_off,
+				       eccsize * 8);
+		src_bit_off += eccsize * 8;
+
+		/* Align last ECC block to align a byte boundary */
+		if (step == nfc_geo->ecc_chunk_count - 1 &&
+		    (oob_bit_off + eccbits) % 8)
+			eccbits += 8 - ((oob_bit_off + eccbits) % 8);
+
+		if (oob_required)
+			gpmi_copy_bits(oob, oob_bit_off,
+				       tmp_buf, src_bit_off,
+				       eccbits);
+
+		src_bit_off += eccbits;
+		oob_bit_off += eccbits;
+	}
+
+	if (oob_required) {
+		oob_byte_off = oob_bit_off / 8;
+
+		if (oob_byte_off < mtd->oobsize)
+			memcpy(oob + oob_byte_off,
+			       tmp_buf + mtd->writesize + oob_byte_off,
+			       mtd->oobsize - oob_byte_off);
+	}
+
+	return 0;
+}
+
+/*
+ * This function writes a NAND page without involving the ECC engine (no HW
+ * ECC generation).
+ * The tricky part in the GPMI/BCH controller is that it stores ECC bits
+ * inline (interleaved with payload DATA), and do not align data chunk on
+ * byte boundaries.
+ * We thus need to take care moving the OOB area at the right place in the
+ * final page, which is why we're using gpmi_copy_bits.
+ *
+ * See set_geometry_by_ecc_info inline comments to have a full description
+ * of the layout used by the GPMI controller.
+ */
+static int gpmi_ecc_write_page_raw(struct mtd_info *mtd,
+				   struct nand_chip *chip,
+				   const uint8_t *buf,
+				   int oob_required, int page)
+{
+	struct gpmi_nand_data *this = nand_get_controller_data(chip);
+	struct bch_geometry *nfc_geo = &this->bch_geometry;
+	int eccsize = nfc_geo->ecc_chunk_size;
+	int eccbits = nfc_geo->ecc_strength * nfc_geo->gf_len;
+	u8 *tmp_buf = this->raw_buffer;
+	uint8_t *oob = chip->oob_poi;
+	size_t dst_bit_off;
+	size_t oob_bit_off;
+	size_t oob_byte_off;
+	int step;
+
+	/*
+	 * Initialize all bits to 1 in case we don't have a buffer for the
+	 * payload or oob data in order to leave unspecified bits of data
+	 * to their initial state.
+	 */
+	if (!buf || !oob_required)
+		memset(tmp_buf, 0xff, mtd->writesize + mtd->oobsize);
+
+	/*
+	 * First copy the metadata section (stored in oob buffer) at the
+	 * beginning of the page, as imposed by the GPMI layout.
+	 */
+	memcpy(tmp_buf, oob, nfc_geo->metadata_size);
+	oob_bit_off = nfc_geo->metadata_size * 8;
+	dst_bit_off = oob_bit_off;
+
+	/* Interleave payload data and ECC bits */
+	for (step = 0; step < nfc_geo->ecc_chunk_count; step++) {
+		if (buf)
+			gpmi_copy_bits(tmp_buf, dst_bit_off,
+				       buf, step * eccsize * 8, eccsize * 8);
+		dst_bit_off += eccsize * 8;
+
+		/* Align last ECC block to align a byte boundary */
+		if (step == nfc_geo->ecc_chunk_count - 1 &&
+		    (oob_bit_off + eccbits) % 8)
+			eccbits += 8 - ((oob_bit_off + eccbits) % 8);
+
+		if (oob_required)
+			gpmi_copy_bits(tmp_buf, dst_bit_off,
+				       oob, oob_bit_off, eccbits);
+
+		dst_bit_off += eccbits;
+		oob_bit_off += eccbits;
+	}
+
+	oob_byte_off = oob_bit_off / 8;
+
+	if (oob_required && oob_byte_off < mtd->oobsize)
+		memcpy(tmp_buf + mtd->writesize + oob_byte_off,
+		       oob + oob_byte_off, mtd->oobsize - oob_byte_off);
+
+	/*
+	 * If required, swap the bad block marker and the first byte of the
+	 * metadata section, so that we don't modify the bad block marker.
+	 *
+	 * See the layout description for a detailed explanation on why this
+	 * is needed.
+	 */
+	if (this->swap_block_mark)
+		swap(tmp_buf[0], tmp_buf[mtd->writesize]);
+
+	return nand_prog_page_op(chip, page, 0, tmp_buf,
+				 mtd->writesize + mtd->oobsize);
+}
+
+static int gpmi_ecc_read_oob_raw(struct mtd_info *mtd, struct nand_chip *chip,
+				 int page)
+{
+	return gpmi_ecc_read_page_raw(mtd, chip, NULL, 1, page);
+}
+
+static int gpmi_ecc_write_oob_raw(struct mtd_info *mtd, struct nand_chip *chip,
+				 int page)
+{
+	return gpmi_ecc_write_page_raw(mtd, chip, NULL, 1, page);
+}
+
+static int gpmi_block_markbad(struct mtd_info *mtd, loff_t ofs)
+{
+	struct nand_chip *chip = mtd_to_nand(mtd);
+	struct gpmi_nand_data *this = nand_get_controller_data(chip);
+	int ret = 0;
+	uint8_t *block_mark;
+	int column, page, chipnr;
+
+	chipnr = (int)(ofs >> chip->chip_shift);
+	chip->select_chip(mtd, chipnr);
+
+	column = !GPMI_IS_MX23(this) ? mtd->writesize : 0;
+
+	/* Write the block mark. */
+	block_mark = this->data_buffer_dma;
+	block_mark[0] = 0; /* bad block marker */
+
+	/* Shift to get page */
+	page = (int)(ofs >> chip->page_shift);
+
+	ret = nand_prog_page_op(chip, page, column, block_mark, 1);
+
+	chip->select_chip(mtd, -1);
+
+	return ret;
+}
+
+static int nand_boot_set_geometry(struct gpmi_nand_data *this)
+{
+	struct boot_rom_geometry *geometry = &this->rom_geometry;
+
+	/*
+	 * Set the boot block stride size.
+	 *
+	 * In principle, we should be reading this from the OTP bits, since
+	 * that's where the ROM is going to get it. In fact, we don't have any
+	 * way to read the OTP bits, so we go with the default and hope for the
+	 * best.
+	 */
+	geometry->stride_size_in_pages = 64;
+
+	/*
+	 * Set the search area stride exponent.
+	 *
+	 * In principle, we should be reading this from the OTP bits, since
+	 * that's where the ROM is going to get it. In fact, we don't have any
+	 * way to read the OTP bits, so we go with the default and hope for the
+	 * best.
+	 */
+	geometry->search_area_stride_exponent = 2;
+	return 0;
+}
+
+static const char  *fingerprint = "STMP";
+static int mx23_check_transcription_stamp(struct gpmi_nand_data *this)
+{
+	struct boot_rom_geometry *rom_geo = &this->rom_geometry;
+	struct device *dev = this->dev;
+	struct nand_chip *chip = &this->nand;
+	struct mtd_info *mtd = nand_to_mtd(chip);
+	unsigned int search_area_size_in_strides;
+	unsigned int stride;
+	unsigned int page;
+	uint8_t *buffer = chip->data_buf;
+	int saved_chip_number;
+	int found_an_ncb_fingerprint = false;
+
+	/* Compute the number of strides in a search area. */
+	search_area_size_in_strides = 1 << rom_geo->search_area_stride_exponent;
+
+	saved_chip_number = this->current_chip;
+	chip->select_chip(mtd, 0);
+
+	/*
+	 * Loop through the first search area, looking for the NCB fingerprint.
+	 */
+	dev_dbg(dev, "Scanning for an NCB fingerprint...\n");
+
+	for (stride = 0; stride < search_area_size_in_strides; stride++) {
+		/* Compute the page addresses. */
+		page = stride * rom_geo->stride_size_in_pages;
+
+		dev_dbg(dev, "Looking for a fingerprint in page 0x%x\n", page);
+
+		/*
+		 * Read the NCB fingerprint. The fingerprint is four bytes long
+		 * and starts in the 12th byte of the page.
+		 */
+		nand_read_page_op(chip, page, 12, NULL, 0);
+		chip->read_buf(mtd, buffer, strlen(fingerprint));
+
+		/* Look for the fingerprint. */
+		if (!memcmp(buffer, fingerprint, strlen(fingerprint))) {
+			found_an_ncb_fingerprint = true;
+			break;
+		}
+
+	}
+
+	chip->select_chip(mtd, saved_chip_number);
+
+	if (found_an_ncb_fingerprint)
+		dev_dbg(dev, "\tFound a fingerprint\n");
+	else
+		dev_dbg(dev, "\tNo fingerprint found\n");
+	return found_an_ncb_fingerprint;
+}
+
+/* Writes a transcription stamp. */
+static int mx23_write_transcription_stamp(struct gpmi_nand_data *this)
+{
+	struct device *dev = this->dev;
+	struct boot_rom_geometry *rom_geo = &this->rom_geometry;
+	struct nand_chip *chip = &this->nand;
+	struct mtd_info *mtd = nand_to_mtd(chip);
+	unsigned int block_size_in_pages;
+	unsigned int search_area_size_in_strides;
+	unsigned int search_area_size_in_pages;
+	unsigned int search_area_size_in_blocks;
+	unsigned int block;
+	unsigned int stride;
+	unsigned int page;
+	uint8_t      *buffer = chip->data_buf;
+	int saved_chip_number;
+	int status;
+
+	/* Compute the search area geometry. */
+	block_size_in_pages = mtd->erasesize / mtd->writesize;
+	search_area_size_in_strides = 1 << rom_geo->search_area_stride_exponent;
+	search_area_size_in_pages = search_area_size_in_strides *
+					rom_geo->stride_size_in_pages;
+	search_area_size_in_blocks =
+		  (search_area_size_in_pages + (block_size_in_pages - 1)) /
+				    block_size_in_pages;
+
+	dev_dbg(dev, "Search Area Geometry :\n");
+	dev_dbg(dev, "\tin Blocks : %u\n", search_area_size_in_blocks);
+	dev_dbg(dev, "\tin Strides: %u\n", search_area_size_in_strides);
+	dev_dbg(dev, "\tin Pages  : %u\n", search_area_size_in_pages);
+
+	/* Select chip 0. */
+	saved_chip_number = this->current_chip;
+	chip->select_chip(mtd, 0);
+
+	/* Loop over blocks in the first search area, erasing them. */
+	dev_dbg(dev, "Erasing the search area...\n");
+
+	for (block = 0; block < search_area_size_in_blocks; block++) {
+		/* Erase this block. */
+		dev_dbg(dev, "\tErasing block 0x%x\n", block);
+		status = nand_erase_op(chip, block);
+		if (status)
+			dev_err(dev, "[%s] Erase failed.\n", __func__);
+	}
+
+	/* Write the NCB fingerprint into the page buffer. */
+	memset(buffer, ~0, mtd->writesize);
+	memcpy(buffer + 12, fingerprint, strlen(fingerprint));
+
+	/* Loop through the first search area, writing NCB fingerprints. */
+	dev_dbg(dev, "Writing NCB fingerprints...\n");
+	for (stride = 0; stride < search_area_size_in_strides; stride++) {
+		/* Compute the page addresses. */
+		page = stride * rom_geo->stride_size_in_pages;
+
+		/* Write the first page of the current stride. */
+		dev_dbg(dev, "Writing an NCB fingerprint in page 0x%x\n", page);
+
+		status = chip->ecc.write_page_raw(mtd, chip, buffer, 0, page);
+		if (status)
+			dev_err(dev, "[%s] Write failed.\n", __func__);
+	}
+
+	/* Deselect chip 0. */
+	chip->select_chip(mtd, saved_chip_number);
+	return 0;
+}
+
+static int mx23_boot_init(struct gpmi_nand_data  *this)
+{
+	struct device *dev = this->dev;
+	struct nand_chip *chip = &this->nand;
+	struct mtd_info *mtd = nand_to_mtd(chip);
+	unsigned int block_count;
+	unsigned int block;
+	int     chipnr;
+	int     page;
+	loff_t  byte;
+	uint8_t block_mark;
+	int     ret = 0;
+
+	/*
+	 * If control arrives here, we can't use block mark swapping, which
+	 * means we're forced to use transcription. First, scan for the
+	 * transcription stamp. If we find it, then we don't have to do
+	 * anything -- the block marks are already transcribed.
+	 */
+	if (mx23_check_transcription_stamp(this))
+		return 0;
+
+	/*
+	 * If control arrives here, we couldn't find a transcription stamp, so
+	 * so we presume the block marks are in the conventional location.
+	 */
+	dev_dbg(dev, "Transcribing bad block marks...\n");
+
+	/* Compute the number of blocks in the entire medium. */
+	block_count = chip->chipsize >> chip->phys_erase_shift;
+
+	/*
+	 * Loop over all the blocks in the medium, transcribing block marks as
+	 * we go.
+	 */
+	for (block = 0; block < block_count; block++) {
+		/*
+		 * Compute the chip, page and byte addresses for this block's
+		 * conventional mark.
+		 */
+		chipnr = block >> (chip->chip_shift - chip->phys_erase_shift);
+		page = block << (chip->phys_erase_shift - chip->page_shift);
+		byte = block <<  chip->phys_erase_shift;
+
+		/* Send the command to read the conventional block mark. */
+		chip->select_chip(mtd, chipnr);
+		nand_read_page_op(chip, page, mtd->writesize, NULL, 0);
+		block_mark = chip->read_byte(mtd);
+		chip->select_chip(mtd, -1);
+
+		/*
+		 * Check if the block is marked bad. If so, we need to mark it
+		 * again, but this time the result will be a mark in the
+		 * location where we transcribe block marks.
+		 */
+		if (block_mark != 0xff) {
+			dev_dbg(dev, "Transcribing mark in block %u\n", block);
+			ret = chip->block_markbad(mtd, byte);
+			if (ret)
+				dev_err(dev,
+					"Failed to mark block bad with ret %d\n",
+					ret);
+		}
+	}
+
+	/* Write the stamp that indicates we've transcribed the block marks. */
+	mx23_write_transcription_stamp(this);
+	return 0;
+}
+
+static int nand_boot_init(struct gpmi_nand_data  *this)
+{
+	nand_boot_set_geometry(this);
+
+	/* This is ROM arch-specific initilization before the BBT scanning. */
+	if (GPMI_IS_MX23(this))
+		return mx23_boot_init(this);
+	return 0;
+}
+
+static int gpmi_set_geometry(struct gpmi_nand_data *this)
+{
+	int ret;
+
+	/* Free the temporary DMA memory for reading ID. */
+	gpmi_free_dma_buffer(this);
+
+	/* Set up the NFC geometry which is used by BCH. */
+	ret = bch_set_geometry(this);
+	if (ret) {
+		dev_err(this->dev, "Error setting BCH geometry : %d\n", ret);
+		return ret;
+	}
+
+	/* Alloc the new DMA buffers according to the pagesize and oobsize */
+	return gpmi_alloc_dma_buffer(this);
+}
+
+static int gpmi_init_last(struct gpmi_nand_data *this)
+{
+	struct nand_chip *chip = &this->nand;
+	struct mtd_info *mtd = nand_to_mtd(chip);
+	struct nand_ecc_ctrl *ecc = &chip->ecc;
+	struct bch_geometry *bch_geo = &this->bch_geometry;
+	int ret;
+
+	/* Set up the medium geometry */
+	ret = gpmi_set_geometry(this);
+	if (ret)
+		return ret;
+
+	/* Init the nand_ecc_ctrl{} */
+	ecc->read_page	= gpmi_ecc_read_page;
+	ecc->write_page	= gpmi_ecc_write_page;
+	ecc->read_oob	= gpmi_ecc_read_oob;
+	ecc->write_oob	= gpmi_ecc_write_oob;
+	ecc->read_page_raw = gpmi_ecc_read_page_raw;
+	ecc->write_page_raw = gpmi_ecc_write_page_raw;
+	ecc->read_oob_raw = gpmi_ecc_read_oob_raw;
+	ecc->write_oob_raw = gpmi_ecc_write_oob_raw;
+	ecc->mode	= NAND_ECC_HW;
+	ecc->size	= bch_geo->ecc_chunk_size;
+	ecc->strength	= bch_geo->ecc_strength;
+	mtd_set_ooblayout(mtd, &gpmi_ooblayout_ops);
+
+	/*
+	 * We only enable the subpage read when:
+	 *  (1) the chip is imx6, and
+	 *  (2) the size of the ECC parity is byte aligned.
+	 */
+	if (GPMI_IS_MX6(this) &&
+		((bch_geo->gf_len * bch_geo->ecc_strength) % 8) == 0) {
+		ecc->read_subpage = gpmi_ecc_read_subpage;
+		chip->options |= NAND_SUBPAGE_READ;
+	}
+
+	/*
+	 * Can we enable the extra features? such as EDO or Sync mode.
+	 *
+	 * We do not check the return value now. That's means if we fail in
+	 * enable the extra features, we still can run in the normal way.
+	 */
+	gpmi_extra_init(this);
+
+	return 0;
+}
+
+static int gpmi_nand_init(struct gpmi_nand_data *this)
+{
+	struct nand_chip *chip = &this->nand;
+	struct mtd_info  *mtd = nand_to_mtd(chip);
+	int ret;
+
+	/* init current chip */
+	this->current_chip	= -1;
+
+	/* init the MTD data structures */
+	mtd->name		= "gpmi-nand";
+	mtd->dev.parent		= this->dev;
+
+	/* init the nand_chip{}, we don't support a 16-bit NAND Flash bus. */
+	nand_set_controller_data(chip, this);
+	nand_set_flash_node(chip, this->pdev->dev.of_node);
+	chip->select_chip	= gpmi_select_chip;
+	chip->cmd_ctrl		= gpmi_cmd_ctrl;
+	chip->dev_ready		= gpmi_dev_ready;
+	chip->read_byte		= gpmi_read_byte;
+	chip->read_buf		= gpmi_read_buf;
+	chip->write_buf		= gpmi_write_buf;
+	chip->badblock_pattern	= &gpmi_bbt_descr;
+	chip->block_markbad	= gpmi_block_markbad;
+	chip->options		|= NAND_NO_SUBPAGE_WRITE;
+
+	/* Set up swap_block_mark, must be set before the gpmi_set_geometry() */
+	this->swap_block_mark = !GPMI_IS_MX23(this);
+
+	/*
+	 * Allocate a temporary DMA buffer for reading ID in the
+	 * nand_scan_ident().
+	 */
+	this->bch_geometry.payload_size = 1024;
+	this->bch_geometry.auxiliary_size = 128;
+	ret = gpmi_alloc_dma_buffer(this);
+	if (ret)
+		goto err_out;
+
+	ret = nand_scan_ident(mtd, GPMI_IS_MX6(this) ? 2 : 1, NULL);
+	if (ret)
+		goto err_out;
+
+	if (chip->bbt_options & NAND_BBT_USE_FLASH) {
+		chip->bbt_options |= NAND_BBT_NO_OOB;
+
+		if (of_property_read_bool(this->dev->of_node,
+						"fsl,no-blockmark-swap"))
+			this->swap_block_mark = false;
+	}
+	dev_dbg(this->dev, "Blockmark swapping %sabled\n",
+		this->swap_block_mark ? "en" : "dis");
+
+	ret = gpmi_init_last(this);
+	if (ret)
+		goto err_out;
+
+	chip->options |= NAND_SKIP_BBTSCAN;
+	ret = nand_scan_tail(mtd);
+	if (ret)
+		goto err_out;
+
+	ret = nand_boot_init(this);
+	if (ret)
+		goto err_nand_cleanup;
+	ret = chip->scan_bbt(mtd);
+	if (ret)
+		goto err_nand_cleanup;
+
+	ret = mtd_device_register(mtd, NULL, 0);
+	if (ret)
+		goto err_nand_cleanup;
+	return 0;
+
+err_nand_cleanup:
+	nand_cleanup(chip);
+err_out:
+	gpmi_free_dma_buffer(this);
+	return ret;
+}
+
+static const struct of_device_id gpmi_nand_id_table[] = {
+	{
+		.compatible = "fsl,imx23-gpmi-nand",
+		.data = &gpmi_devdata_imx23,
+	}, {
+		.compatible = "fsl,imx28-gpmi-nand",
+		.data = &gpmi_devdata_imx28,
+	}, {
+		.compatible = "fsl,imx6q-gpmi-nand",
+		.data = &gpmi_devdata_imx6q,
+	}, {
+		.compatible = "fsl,imx6sx-gpmi-nand",
+		.data = &gpmi_devdata_imx6sx,
+	}, {
+		.compatible = "fsl,imx7d-gpmi-nand",
+		.data = &gpmi_devdata_imx7d,
+	}, {}
+};
+MODULE_DEVICE_TABLE(of, gpmi_nand_id_table);
+
+static int gpmi_nand_probe(struct platform_device *pdev)
+{
+	struct gpmi_nand_data *this;
+	const struct of_device_id *of_id;
+	int ret;
+
+	this = devm_kzalloc(&pdev->dev, sizeof(*this), GFP_KERNEL);
+	if (!this)
+		return -ENOMEM;
+
+	of_id = of_match_device(gpmi_nand_id_table, &pdev->dev);
+	if (of_id) {
+		this->devdata = of_id->data;
+	} else {
+		dev_err(&pdev->dev, "Failed to find the right device id.\n");
+		return -ENODEV;
+	}
+
+	platform_set_drvdata(pdev, this);
+	this->pdev  = pdev;
+	this->dev   = &pdev->dev;
+
+	ret = acquire_resources(this);
+	if (ret)
+		goto exit_acquire_resources;
+
+	ret = init_hardware(this);
+	if (ret)
+		goto exit_nfc_init;
+
+	ret = gpmi_nand_init(this);
+	if (ret)
+		goto exit_nfc_init;
+
+	dev_info(this->dev, "driver registered.\n");
+
+	return 0;
+
+exit_nfc_init:
+	release_resources(this);
+exit_acquire_resources:
+
+	return ret;
+}
+
+static int gpmi_nand_remove(struct platform_device *pdev)
+{
+	struct gpmi_nand_data *this = platform_get_drvdata(pdev);
+
+	nand_release(nand_to_mtd(&this->nand));
+	gpmi_free_dma_buffer(this);
+	release_resources(this);
+	return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int gpmi_pm_suspend(struct device *dev)
+{
+	struct gpmi_nand_data *this = dev_get_drvdata(dev);
+
+	release_dma_channels(this);
+	return 0;
+}
+
+static int gpmi_pm_resume(struct device *dev)
+{
+	struct gpmi_nand_data *this = dev_get_drvdata(dev);
+	int ret;
+
+	ret = acquire_dma_channels(this);
+	if (ret < 0)
+		return ret;
+
+	/* re-init the GPMI registers */
+	this->flags &= ~GPMI_TIMING_INIT_OK;
+	ret = gpmi_init(this);
+	if (ret) {
+		dev_err(this->dev, "Error setting GPMI : %d\n", ret);
+		return ret;
+	}
+
+	/* re-init the BCH registers */
+	ret = bch_set_geometry(this);
+	if (ret) {
+		dev_err(this->dev, "Error setting BCH : %d\n", ret);
+		return ret;
+	}
+
+	/* re-init others */
+	gpmi_extra_init(this);
+
+	return 0;
+}
+#endif /* CONFIG_PM_SLEEP */
+
+static const struct dev_pm_ops gpmi_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(gpmi_pm_suspend, gpmi_pm_resume)
+};
+
+static struct platform_driver gpmi_nand_driver = {
+	.driver = {
+		.name = "gpmi-nand",
+		.pm = &gpmi_pm_ops,
+		.of_match_table = gpmi_nand_id_table,
+	},
+	.probe   = gpmi_nand_probe,
+	.remove  = gpmi_nand_remove,
+};
+module_platform_driver(gpmi_nand_driver);
+
+MODULE_AUTHOR("Freescale Semiconductor, Inc.");
+MODULE_DESCRIPTION("i.MX GPMI NAND Flash Controller Driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/mtd/nand/raw/gpmi-nand/gpmi-nand.h b/drivers/mtd/nand/raw/gpmi-nand/gpmi-nand.h
new file mode 100644
index 0000000..06c1f99
--- /dev/null
+++ b/drivers/mtd/nand/raw/gpmi-nand/gpmi-nand.h
@@ -0,0 +1,315 @@
+/*
+ * Freescale GPMI NAND Flash Driver
+ *
+ * Copyright (C) 2010-2011 Freescale Semiconductor, Inc.
+ * Copyright (C) 2008 Embedded Alley Solutions, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+#ifndef __DRIVERS_MTD_NAND_GPMI_NAND_H
+#define __DRIVERS_MTD_NAND_GPMI_NAND_H
+
+#include <linux/mtd/rawnand.h>
+#include <linux/platform_device.h>
+#include <linux/dma-mapping.h>
+#include <linux/dmaengine.h>
+
+#define GPMI_CLK_MAX 5 /* MX6Q needs five clocks */
+struct resources {
+	void __iomem  *gpmi_regs;
+	void __iomem  *bch_regs;
+	unsigned int  dma_low_channel;
+	unsigned int  dma_high_channel;
+	struct clk    *clock[GPMI_CLK_MAX];
+};
+
+/**
+ * struct bch_geometry - BCH geometry description.
+ * @gf_len:                   The length of Galois Field. (e.g., 13 or 14)
+ * @ecc_strength:             A number that describes the strength of the ECC
+ *                            algorithm.
+ * @page_size:                The size, in bytes, of a physical page, including
+ *                            both data and OOB.
+ * @metadata_size:            The size, in bytes, of the metadata.
+ * @ecc_chunk_size:           The size, in bytes, of a single ECC chunk. Note
+ *                            the first chunk in the page includes both data and
+ *                            metadata, so it's a bit larger than this value.
+ * @ecc_chunk_count:          The number of ECC chunks in the page,
+ * @payload_size:             The size, in bytes, of the payload buffer.
+ * @auxiliary_size:           The size, in bytes, of the auxiliary buffer.
+ * @auxiliary_status_offset:  The offset into the auxiliary buffer at which
+ *                            the ECC status appears.
+ * @block_mark_byte_offset:   The byte offset in the ECC-based page view at
+ *                            which the underlying physical block mark appears.
+ * @block_mark_bit_offset:    The bit offset into the ECC-based page view at
+ *                            which the underlying physical block mark appears.
+ */
+struct bch_geometry {
+	unsigned int  gf_len;
+	unsigned int  ecc_strength;
+	unsigned int  page_size;
+	unsigned int  metadata_size;
+	unsigned int  ecc_chunk_size;
+	unsigned int  ecc_chunk_count;
+	unsigned int  payload_size;
+	unsigned int  auxiliary_size;
+	unsigned int  auxiliary_status_offset;
+	unsigned int  block_mark_byte_offset;
+	unsigned int  block_mark_bit_offset;
+};
+
+/**
+ * struct boot_rom_geometry - Boot ROM geometry description.
+ * @stride_size_in_pages:        The size of a boot block stride, in pages.
+ * @search_area_stride_exponent: The logarithm to base 2 of the size of a
+ *                               search area in boot block strides.
+ */
+struct boot_rom_geometry {
+	unsigned int  stride_size_in_pages;
+	unsigned int  search_area_stride_exponent;
+};
+
+/* DMA operations types */
+enum dma_ops_type {
+	DMA_FOR_COMMAND = 1,
+	DMA_FOR_READ_DATA,
+	DMA_FOR_WRITE_DATA,
+	DMA_FOR_READ_ECC_PAGE,
+	DMA_FOR_WRITE_ECC_PAGE
+};
+
+/**
+ * struct nand_timing - Fundamental timing attributes for NAND.
+ * @data_setup_in_ns:         The data setup time, in nanoseconds. Usually the
+ *                            maximum of tDS and tWP. A negative value
+ *                            indicates this characteristic isn't known.
+ * @data_hold_in_ns:          The data hold time, in nanoseconds. Usually the
+ *                            maximum of tDH, tWH and tREH. A negative value
+ *                            indicates this characteristic isn't known.
+ * @address_setup_in_ns:      The address setup time, in nanoseconds. Usually
+ *                            the maximum of tCLS, tCS and tALS. A negative
+ *                            value indicates this characteristic isn't known.
+ * @gpmi_sample_delay_in_ns:  A GPMI-specific timing parameter. A negative value
+ *                            indicates this characteristic isn't known.
+ * @tREA_in_ns:               tREA, in nanoseconds, from the data sheet. A
+ *                            negative value indicates this characteristic isn't
+ *                            known.
+ * @tRLOH_in_ns:              tRLOH, in nanoseconds, from the data sheet. A
+ *                            negative value indicates this characteristic isn't
+ *                            known.
+ * @tRHOH_in_ns:              tRHOH, in nanoseconds, from the data sheet. A
+ *                            negative value indicates this characteristic isn't
+ *                            known.
+ */
+struct nand_timing {
+	int8_t  data_setup_in_ns;
+	int8_t  data_hold_in_ns;
+	int8_t  address_setup_in_ns;
+	int8_t  gpmi_sample_delay_in_ns;
+	int8_t  tREA_in_ns;
+	int8_t  tRLOH_in_ns;
+	int8_t  tRHOH_in_ns;
+};
+
+enum gpmi_type {
+	IS_MX23,
+	IS_MX28,
+	IS_MX6Q,
+	IS_MX6SX,
+	IS_MX7D,
+};
+
+struct gpmi_devdata {
+	enum gpmi_type type;
+	int bch_max_ecc_strength;
+	int max_chain_delay; /* See the async EDO mode */
+	const char * const *clks;
+	const int clks_count;
+};
+
+struct gpmi_nand_data {
+	/* flags */
+#define GPMI_ASYNC_EDO_ENABLED	(1 << 0)
+#define GPMI_TIMING_INIT_OK	(1 << 1)
+	int			flags;
+	const struct gpmi_devdata *devdata;
+
+	/* System Interface */
+	struct device		*dev;
+	struct platform_device	*pdev;
+
+	/* Resources */
+	struct resources	resources;
+
+	/* Flash Hardware */
+	struct nand_timing	timing;
+	int			timing_mode;
+
+	/* BCH */
+	struct bch_geometry	bch_geometry;
+	struct completion	bch_done;
+
+	/* NAND Boot issue */
+	bool			swap_block_mark;
+	struct boot_rom_geometry rom_geometry;
+
+	/* MTD / NAND */
+	struct nand_chip	nand;
+
+	/* General-use Variables */
+	int			current_chip;
+	unsigned int		command_length;
+
+	/* passed from upper layer */
+	uint8_t			*upper_buf;
+	int			upper_len;
+
+	/* for DMA operations */
+	bool			direct_dma_map_ok;
+
+	struct scatterlist	cmd_sgl;
+	char			*cmd_buffer;
+
+	struct scatterlist	data_sgl;
+	char			*data_buffer_dma;
+
+	void			*page_buffer_virt;
+	dma_addr_t		page_buffer_phys;
+	unsigned int		page_buffer_size;
+
+	void			*payload_virt;
+	dma_addr_t		payload_phys;
+
+	void			*auxiliary_virt;
+	dma_addr_t		auxiliary_phys;
+
+	void			*raw_buffer;
+
+	/* DMA channels */
+#define DMA_CHANS		8
+	struct dma_chan		*dma_chans[DMA_CHANS];
+	enum dma_ops_type	last_dma_type;
+	enum dma_ops_type	dma_type;
+	struct completion	dma_done;
+
+	/* private */
+	void			*private;
+};
+
+/**
+ * struct gpmi_nfc_hardware_timing - GPMI hardware timing parameters.
+ * @data_setup_in_cycles:      The data setup time, in cycles.
+ * @data_hold_in_cycles:       The data hold time, in cycles.
+ * @address_setup_in_cycles:   The address setup time, in cycles.
+ * @device_busy_timeout:       The timeout waiting for NAND Ready/Busy,
+ *                             this value is the number of cycles multiplied
+ *                             by 4096.
+ * @use_half_periods:          Indicates the clock is running slowly, so the
+ *                             NFC DLL should use half-periods.
+ * @sample_delay_factor:       The sample delay factor.
+ * @wrn_dly_sel:               The delay on the GPMI write strobe.
+ */
+struct gpmi_nfc_hardware_timing {
+	/* for HW_GPMI_TIMING0 */
+	uint8_t  data_setup_in_cycles;
+	uint8_t  data_hold_in_cycles;
+	uint8_t  address_setup_in_cycles;
+
+	/* for HW_GPMI_TIMING1 */
+	uint16_t device_busy_timeout;
+#define GPMI_DEFAULT_BUSY_TIMEOUT	0x500 /* default busy timeout value.*/
+
+	/* for HW_GPMI_CTRL1 */
+	bool     use_half_periods;
+	uint8_t  sample_delay_factor;
+	uint8_t  wrn_dly_sel;
+};
+
+/**
+ * struct timing_threshold - Timing threshold
+ * @max_data_setup_cycles:       The maximum number of data setup cycles that
+ *                               can be expressed in the hardware.
+ * @internal_data_setup_in_ns:   The time, in ns, that the NFC hardware requires
+ *                               for data read internal setup. In the Reference
+ *                               Manual, see the chapter "High-Speed NAND
+ *                               Timing" for more details.
+ * @max_sample_delay_factor:     The maximum sample delay factor that can be
+ *                               expressed in the hardware.
+ * @max_dll_clock_period_in_ns:  The maximum period of the GPMI clock that the
+ *                               sample delay DLL hardware can possibly work
+ *                               with (the DLL is unusable with longer periods).
+ *                               If the full-cycle period is greater than HALF
+ *                               this value, the DLL must be configured to use
+ *                               half-periods.
+ * @max_dll_delay_in_ns:         The maximum amount of delay, in ns, that the
+ *                               DLL can implement.
+ * @clock_frequency_in_hz:       The clock frequency, in Hz, during the current
+ *                               I/O transaction. If no I/O transaction is in
+ *                               progress, this is the clock frequency during
+ *                               the most recent I/O transaction.
+ */
+struct timing_threshold {
+	const unsigned int      max_chip_count;
+	const unsigned int      max_data_setup_cycles;
+	const unsigned int      internal_data_setup_in_ns;
+	const unsigned int      max_sample_delay_factor;
+	const unsigned int      max_dll_clock_period_in_ns;
+	const unsigned int      max_dll_delay_in_ns;
+	unsigned long           clock_frequency_in_hz;
+
+};
+
+/* Common Services */
+int common_nfc_set_geometry(struct gpmi_nand_data *);
+struct dma_chan *get_dma_chan(struct gpmi_nand_data *);
+void prepare_data_dma(struct gpmi_nand_data *,
+		      enum dma_data_direction dr);
+int start_dma_without_bch_irq(struct gpmi_nand_data *,
+			      struct dma_async_tx_descriptor *);
+int start_dma_with_bch_irq(struct gpmi_nand_data *,
+			   struct dma_async_tx_descriptor *);
+
+/* GPMI-NAND helper function library */
+int gpmi_init(struct gpmi_nand_data *);
+int gpmi_extra_init(struct gpmi_nand_data *);
+void gpmi_clear_bch(struct gpmi_nand_data *);
+void gpmi_dump_info(struct gpmi_nand_data *);
+int bch_set_geometry(struct gpmi_nand_data *);
+int gpmi_is_ready(struct gpmi_nand_data *, unsigned chip);
+int gpmi_send_command(struct gpmi_nand_data *);
+void gpmi_begin(struct gpmi_nand_data *);
+void gpmi_end(struct gpmi_nand_data *);
+int gpmi_read_data(struct gpmi_nand_data *);
+int gpmi_send_data(struct gpmi_nand_data *);
+int gpmi_send_page(struct gpmi_nand_data *,
+		   dma_addr_t payload, dma_addr_t auxiliary);
+int gpmi_read_page(struct gpmi_nand_data *,
+		   dma_addr_t payload, dma_addr_t auxiliary);
+
+void gpmi_copy_bits(u8 *dst, size_t dst_bit_off,
+		    const u8 *src, size_t src_bit_off,
+		    size_t nbits);
+
+/* BCH : Status Block Completion Codes */
+#define STATUS_GOOD		0x00
+#define STATUS_ERASED		0xff
+#define STATUS_UNCORRECTABLE	0xfe
+
+/* Use the devdata to distinguish different Archs. */
+#define GPMI_IS_MX23(x)		((x)->devdata->type == IS_MX23)
+#define GPMI_IS_MX28(x)		((x)->devdata->type == IS_MX28)
+#define GPMI_IS_MX6Q(x)		((x)->devdata->type == IS_MX6Q)
+#define GPMI_IS_MX6SX(x)	((x)->devdata->type == IS_MX6SX)
+#define GPMI_IS_MX7D(x)		((x)->devdata->type == IS_MX7D)
+
+#define GPMI_IS_MX6(x)		(GPMI_IS_MX6Q(x) || GPMI_IS_MX6SX(x) || \
+				 GPMI_IS_MX7D(x))
+#endif
diff --git a/drivers/mtd/nand/raw/gpmi-nand/gpmi-regs.h b/drivers/mtd/nand/raw/gpmi-nand/gpmi-regs.h
new file mode 100644
index 0000000..82114cd
--- /dev/null
+++ b/drivers/mtd/nand/raw/gpmi-nand/gpmi-regs.h
@@ -0,0 +1,187 @@
+/*
+ * Freescale GPMI NAND Flash Driver
+ *
+ * Copyright 2008-2011 Freescale Semiconductor, Inc.
+ * Copyright 2008 Embedded Alley Solutions, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+#ifndef __GPMI_NAND_GPMI_REGS_H
+#define __GPMI_NAND_GPMI_REGS_H
+
+#define HW_GPMI_CTRL0					0x00000000
+#define HW_GPMI_CTRL0_SET				0x00000004
+#define HW_GPMI_CTRL0_CLR				0x00000008
+#define HW_GPMI_CTRL0_TOG				0x0000000c
+
+#define BP_GPMI_CTRL0_COMMAND_MODE			24
+#define BM_GPMI_CTRL0_COMMAND_MODE	(3 << BP_GPMI_CTRL0_COMMAND_MODE)
+#define BF_GPMI_CTRL0_COMMAND_MODE(v)	\
+	(((v) << BP_GPMI_CTRL0_COMMAND_MODE) & BM_GPMI_CTRL0_COMMAND_MODE)
+#define BV_GPMI_CTRL0_COMMAND_MODE__WRITE		0x0
+#define BV_GPMI_CTRL0_COMMAND_MODE__READ		0x1
+#define BV_GPMI_CTRL0_COMMAND_MODE__READ_AND_COMPARE	0x2
+#define BV_GPMI_CTRL0_COMMAND_MODE__WAIT_FOR_READY	0x3
+
+#define BM_GPMI_CTRL0_WORD_LENGTH			(1 << 23)
+#define BV_GPMI_CTRL0_WORD_LENGTH__16_BIT		0x0
+#define BV_GPMI_CTRL0_WORD_LENGTH__8_BIT		0x1
+
+/*
+ *  Difference in LOCK_CS between imx23 and imx28 :
+ *  This bit may impact the _POWER_ consumption. So some chips
+ *  do not set it.
+ */
+#define MX23_BP_GPMI_CTRL0_LOCK_CS			22
+#define MX28_BP_GPMI_CTRL0_LOCK_CS			27
+#define LOCK_CS_ENABLE					0x1
+#define BF_GPMI_CTRL0_LOCK_CS(v, x)			0x0
+
+/* Difference in CS between imx23 and imx28 */
+#define BP_GPMI_CTRL0_CS				20
+#define MX23_BM_GPMI_CTRL0_CS		(3 << BP_GPMI_CTRL0_CS)
+#define MX28_BM_GPMI_CTRL0_CS		(7 << BP_GPMI_CTRL0_CS)
+#define BF_GPMI_CTRL0_CS(v, x)		(((v) << BP_GPMI_CTRL0_CS) & \
+						(GPMI_IS_MX23((x)) \
+						? MX23_BM_GPMI_CTRL0_CS	\
+						: MX28_BM_GPMI_CTRL0_CS))
+
+#define BP_GPMI_CTRL0_ADDRESS				17
+#define BM_GPMI_CTRL0_ADDRESS		(3 << BP_GPMI_CTRL0_ADDRESS)
+#define BF_GPMI_CTRL0_ADDRESS(v)	\
+		(((v) << BP_GPMI_CTRL0_ADDRESS) & BM_GPMI_CTRL0_ADDRESS)
+#define BV_GPMI_CTRL0_ADDRESS__NAND_DATA		0x0
+#define BV_GPMI_CTRL0_ADDRESS__NAND_CLE			0x1
+#define BV_GPMI_CTRL0_ADDRESS__NAND_ALE			0x2
+
+#define BM_GPMI_CTRL0_ADDRESS_INCREMENT			(1 << 16)
+#define BV_GPMI_CTRL0_ADDRESS_INCREMENT__DISABLED	0x0
+#define BV_GPMI_CTRL0_ADDRESS_INCREMENT__ENABLED	0x1
+
+#define BP_GPMI_CTRL0_XFER_COUNT			0
+#define BM_GPMI_CTRL0_XFER_COUNT	(0xffff << BP_GPMI_CTRL0_XFER_COUNT)
+#define BF_GPMI_CTRL0_XFER_COUNT(v)	\
+		(((v) << BP_GPMI_CTRL0_XFER_COUNT) & BM_GPMI_CTRL0_XFER_COUNT)
+
+#define HW_GPMI_COMPARE					0x00000010
+
+#define HW_GPMI_ECCCTRL					0x00000020
+#define HW_GPMI_ECCCTRL_SET				0x00000024
+#define HW_GPMI_ECCCTRL_CLR				0x00000028
+#define HW_GPMI_ECCCTRL_TOG				0x0000002c
+
+#define BP_GPMI_ECCCTRL_ECC_CMD				13
+#define BM_GPMI_ECCCTRL_ECC_CMD		(3 << BP_GPMI_ECCCTRL_ECC_CMD)
+#define BF_GPMI_ECCCTRL_ECC_CMD(v)	\
+		(((v) << BP_GPMI_ECCCTRL_ECC_CMD) & BM_GPMI_ECCCTRL_ECC_CMD)
+#define BV_GPMI_ECCCTRL_ECC_CMD__BCH_DECODE		0x0
+#define BV_GPMI_ECCCTRL_ECC_CMD__BCH_ENCODE		0x1
+
+#define BM_GPMI_ECCCTRL_ENABLE_ECC			(1 << 12)
+#define BV_GPMI_ECCCTRL_ENABLE_ECC__ENABLE		0x1
+#define BV_GPMI_ECCCTRL_ENABLE_ECC__DISABLE		0x0
+
+#define BP_GPMI_ECCCTRL_BUFFER_MASK			0
+#define BM_GPMI_ECCCTRL_BUFFER_MASK	(0x1ff << BP_GPMI_ECCCTRL_BUFFER_MASK)
+#define BF_GPMI_ECCCTRL_BUFFER_MASK(v)	\
+	(((v) << BP_GPMI_ECCCTRL_BUFFER_MASK) & BM_GPMI_ECCCTRL_BUFFER_MASK)
+#define BV_GPMI_ECCCTRL_BUFFER_MASK__BCH_AUXONLY	0x100
+#define BV_GPMI_ECCCTRL_BUFFER_MASK__BCH_PAGE		0x1FF
+
+#define HW_GPMI_ECCCOUNT				0x00000030
+#define HW_GPMI_PAYLOAD					0x00000040
+#define HW_GPMI_AUXILIARY				0x00000050
+#define HW_GPMI_CTRL1					0x00000060
+#define HW_GPMI_CTRL1_SET				0x00000064
+#define HW_GPMI_CTRL1_CLR				0x00000068
+#define HW_GPMI_CTRL1_TOG				0x0000006c
+
+#define BP_GPMI_CTRL1_DECOUPLE_CS			24
+#define BM_GPMI_CTRL1_DECOUPLE_CS	(1 << BP_GPMI_CTRL1_DECOUPLE_CS)
+
+#define BP_GPMI_CTRL1_WRN_DLY_SEL			22
+#define BM_GPMI_CTRL1_WRN_DLY_SEL	(0x3 << BP_GPMI_CTRL1_WRN_DLY_SEL)
+#define BF_GPMI_CTRL1_WRN_DLY_SEL(v)  \
+	(((v) << BP_GPMI_CTRL1_WRN_DLY_SEL) & BM_GPMI_CTRL1_WRN_DLY_SEL)
+#define BV_GPMI_CTRL1_WRN_DLY_SEL_4_TO_8NS		0x0
+#define BV_GPMI_CTRL1_WRN_DLY_SEL_6_TO_10NS		0x1
+#define BV_GPMI_CTRL1_WRN_DLY_SEL_7_TO_12NS		0x2
+#define BV_GPMI_CTRL1_WRN_DLY_SEL_NO_DELAY		0x3
+
+#define BM_GPMI_CTRL1_BCH_MODE				(1 << 18)
+
+#define BP_GPMI_CTRL1_DLL_ENABLE			17
+#define BM_GPMI_CTRL1_DLL_ENABLE	(1 << BP_GPMI_CTRL1_DLL_ENABLE)
+
+#define BP_GPMI_CTRL1_HALF_PERIOD			16
+#define BM_GPMI_CTRL1_HALF_PERIOD	(1 << BP_GPMI_CTRL1_HALF_PERIOD)
+
+#define BP_GPMI_CTRL1_RDN_DELAY				12
+#define BM_GPMI_CTRL1_RDN_DELAY		(0xf << BP_GPMI_CTRL1_RDN_DELAY)
+#define BF_GPMI_CTRL1_RDN_DELAY(v)	\
+		(((v) << BP_GPMI_CTRL1_RDN_DELAY) & BM_GPMI_CTRL1_RDN_DELAY)
+
+#define BM_GPMI_CTRL1_DEV_RESET				(1 << 3)
+#define BV_GPMI_CTRL1_DEV_RESET__ENABLED		0x0
+#define BV_GPMI_CTRL1_DEV_RESET__DISABLED		0x1
+
+#define BM_GPMI_CTRL1_ATA_IRQRDY_POLARITY		(1 << 2)
+#define BV_GPMI_CTRL1_ATA_IRQRDY_POLARITY__ACTIVELOW	0x0
+#define BV_GPMI_CTRL1_ATA_IRQRDY_POLARITY__ACTIVEHIGH	0x1
+
+#define BM_GPMI_CTRL1_CAMERA_MODE			(1 << 1)
+#define BV_GPMI_CTRL1_GPMI_MODE__NAND			0x0
+#define BV_GPMI_CTRL1_GPMI_MODE__ATA			0x1
+
+#define BM_GPMI_CTRL1_GPMI_MODE				(1 << 0)
+
+#define HW_GPMI_TIMING0					0x00000070
+
+#define BP_GPMI_TIMING0_ADDRESS_SETUP			16
+#define BM_GPMI_TIMING0_ADDRESS_SETUP	(0xff << BP_GPMI_TIMING0_ADDRESS_SETUP)
+#define BF_GPMI_TIMING0_ADDRESS_SETUP(v)	\
+	(((v) << BP_GPMI_TIMING0_ADDRESS_SETUP) & BM_GPMI_TIMING0_ADDRESS_SETUP)
+
+#define BP_GPMI_TIMING0_DATA_HOLD			8
+#define BM_GPMI_TIMING0_DATA_HOLD	(0xff << BP_GPMI_TIMING0_DATA_HOLD)
+#define BF_GPMI_TIMING0_DATA_HOLD(v)		\
+	(((v) << BP_GPMI_TIMING0_DATA_HOLD) & BM_GPMI_TIMING0_DATA_HOLD)
+
+#define BP_GPMI_TIMING0_DATA_SETUP			0
+#define BM_GPMI_TIMING0_DATA_SETUP	(0xff << BP_GPMI_TIMING0_DATA_SETUP)
+#define BF_GPMI_TIMING0_DATA_SETUP(v)		\
+	(((v) << BP_GPMI_TIMING0_DATA_SETUP) & BM_GPMI_TIMING0_DATA_SETUP)
+
+#define HW_GPMI_TIMING1					0x00000080
+#define BP_GPMI_TIMING1_BUSY_TIMEOUT			16
+#define BM_GPMI_TIMING1_BUSY_TIMEOUT	(0xffff << BP_GPMI_TIMING1_BUSY_TIMEOUT)
+#define BF_GPMI_TIMING1_BUSY_TIMEOUT(v)		\
+	(((v) << BP_GPMI_TIMING1_BUSY_TIMEOUT) & BM_GPMI_TIMING1_BUSY_TIMEOUT)
+
+#define HW_GPMI_TIMING2					0x00000090
+#define HW_GPMI_DATA					0x000000a0
+
+/* MX28 uses this to detect READY. */
+#define HW_GPMI_STAT					0x000000b0
+#define MX28_BP_GPMI_STAT_READY_BUSY			24
+#define MX28_BM_GPMI_STAT_READY_BUSY	(0xff << MX28_BP_GPMI_STAT_READY_BUSY)
+#define MX28_BF_GPMI_STAT_READY_BUSY(v)		\
+	(((v) << MX28_BP_GPMI_STAT_READY_BUSY) & MX28_BM_GPMI_STAT_READY_BUSY)
+
+/* MX23 uses this to detect READY. */
+#define HW_GPMI_DEBUG					0x000000c0
+#define MX23_BP_GPMI_DEBUG_READY0			28
+#define MX23_BM_GPMI_DEBUG_READY0	(1 << MX23_BP_GPMI_DEBUG_READY0)
+#endif