MX2: Add DMA support for mx2 and (eventually) mx1

This patch adds DMA support for Freescale i.MX27 SoCs. It is derived
from the i.MX1 port and should (though currently untested) still be
working for the i.MX1.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

arch/arm/plat-mxc/Makefile

@@ -5,4 +5,4 @@
 # Common support
 obj-y := irq.o clock.o gpio.o time.o devices.o
 
-obj-$(CONFIG_ARCH_MX2) += iomux-mx1-mx2.o
+obj-$(CONFIG_ARCH_MX2) += iomux-mx1-mx2.o dma-mx1-mx2.o

arch/arm/plat-mxc/dma-mx1-mx2.c

+/*
+ *  linux/arch/arm/plat-mxc/dma-mx1-mx2.c
+ *
+ *  i.MX DMA registration and IRQ dispatching
+ *
+ * Copyright 2006 Pavel Pisa <pisa@cmp.felk.cvut.cz>
+ * Copyright 2008 Juergen Beisert, <kernel@pengutronix.de>
+ * Copyright 2008 Sascha Hauer, <s.hauer@pengutronix.de>
+ *
+ * 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/module.h>
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/interrupt.h>
+#include <linux/errno.h>
+#include <linux/clk.h>
+#include <linux/scatterlist.h>
+#include <linux/io.h>
+
+#include <asm/system.h>
+#include <asm/irq.h>
+#include <mach/hardware.h>
+#include <asm/dma.h>
+#include <mach/dma-mx1-mx2.h>
+
+#define DMA_DCR     0x00		/* Control Register */
+#define DMA_DISR    0x04		/* Interrupt status Register */
+#define DMA_DIMR    0x08		/* Interrupt mask Register */
+#define DMA_DBTOSR  0x0c		/* Burst timeout status Register */
+#define DMA_DRTOSR  0x10		/* Request timeout Register */
+#define DMA_DSESR   0x14		/* Transfer Error Status Register */
+#define DMA_DBOSR   0x18		/* Buffer overflow status Register */
+#define DMA_DBTOCR  0x1c		/* Burst timeout control Register */
+#define DMA_WSRA    0x40		/* W-Size Register A */
+#define DMA_XSRA    0x44		/* X-Size Register A */
+#define DMA_YSRA    0x48		/* Y-Size Register A */
+#define DMA_WSRB    0x4c		/* W-Size Register B */
+#define DMA_XSRB    0x50		/* X-Size Register B */
+#define DMA_YSRB    0x54		/* Y-Size Register B */
+#define DMA_SAR(x)  (0x80 + ((x) << 6))	/* Source Address Registers */
+#define DMA_DAR(x)  (0x84 + ((x) << 6))	/* Destination Address Registers */
+#define DMA_CNTR(x) (0x88 + ((x) << 6))	/* Count Registers */
+#define DMA_CCR(x)  (0x8c + ((x) << 6))	/* Control Registers */
+#define DMA_RSSR(x) (0x90 + ((x) << 6))	/* Request source select Registers */
+#define DMA_BLR(x)  (0x94 + ((x) << 6))	/* Burst length Registers */
+#define DMA_RTOR(x) (0x98 + ((x) << 6))	/* Request timeout Registers */
+#define DMA_BUCR(x) (0x98 + ((x) << 6))	/* Bus Utilization Registers */
+#define DMA_CCNR(x) (0x9C + ((x) << 6))	/* Channel counter Registers */
+
+#define DCR_DRST           (1<<1)
+#define DCR_DEN            (1<<0)
+#define DBTOCR_EN          (1<<15)
+#define DBTOCR_CNT(x)      ((x) & 0x7fff)
+#define CNTR_CNT(x)        ((x) & 0xffffff)
+#define CCR_ACRPT          (1<<14)
+#define CCR_DMOD_LINEAR    (0x0 << 12)
+#define CCR_DMOD_2D        (0x1 << 12)
+#define CCR_DMOD_FIFO      (0x2 << 12)
+#define CCR_DMOD_EOBFIFO   (0x3 << 12)
+#define CCR_SMOD_LINEAR    (0x0 << 10)
+#define CCR_SMOD_2D        (0x1 << 10)
+#define CCR_SMOD_FIFO      (0x2 << 10)
+#define CCR_SMOD_EOBFIFO   (0x3 << 10)
+#define CCR_MDIR_DEC       (1<<9)
+#define CCR_MSEL_B         (1<<8)
+#define CCR_DSIZ_32        (0x0 << 6)
+#define CCR_DSIZ_8         (0x1 << 6)
+#define CCR_DSIZ_16        (0x2 << 6)
+#define CCR_SSIZ_32        (0x0 << 4)
+#define CCR_SSIZ_8         (0x1 << 4)
+#define CCR_SSIZ_16        (0x2 << 4)
+#define CCR_REN            (1<<3)
+#define CCR_RPT            (1<<2)
+#define CCR_FRC            (1<<1)
+#define CCR_CEN            (1<<0)
+#define RTOR_EN            (1<<15)
+#define RTOR_CLK           (1<<14)
+#define RTOR_PSC           (1<<13)
+
+/*
+ * struct imx_dma_channel - i.MX specific DMA extension
+ * @name: name specified by DMA client
+ * @irq_handler: client callback for end of transfer
+ * @err_handler: client callback for error condition
+ * @data: clients context data for callbacks
+ * @dma_mode: direction of the transfer %DMA_MODE_READ or %DMA_MODE_WRITE
+ * @sg: pointer to the actual read/written chunk for scatter-gather emulation
+ * @resbytes: total residual number of bytes to transfer
+ *            (it can be lower or same as sum of SG mapped chunk sizes)
+ * @sgcount: number of chunks to be read/written
+ *
+ * Structure is used for IMX DMA processing. It would be probably good
+ * @struct dma_struct in the future for external interfacing and use
+ * @struct imx_dma_channel only as extension to it.
+ */
+
+struct imx_dma_channel {
+	const char *name;
+	void (*irq_handler) (int, void *);
+	void (*err_handler) (int, void *, int errcode);
+	void (*prog_handler) (int, void *, struct scatterlist *);
+	void *data;
+	dmamode_t  dma_mode;
+	struct scatterlist *sg;
+	unsigned int resbytes;
+	int dma_num;
+
+	int in_use;
+
+	u32 ccr_from_device;
+	u32 ccr_to_device;
+
+	struct timer_list watchdog;
+
+	int hw_chaining;
+};
+
+static struct imx_dma_channel imx_dma_channels[IMX_DMA_CHANNELS];
+
+static struct clk *dma_clk;
+
+static int imx_dma_hw_chain(struct imx_dma_channel *imxdma)
+{
+	if (cpu_is_mx27())
+		return imxdma->hw_chaining;
+	else
+		return 0;
+}
+
+
+/*
+ * imx_dma_sg_next - prepare next chunk for scatter-gather DMA emulation
+ */
+static inline int imx_dma_sg_next(int channel, struct scatterlist *sg)
+{
+	struct imx_dma_channel *imxdma = &imx_dma_channels[channel];
+	unsigned long now;
+
+	if (!imxdma->name) {
+		printk(KERN_CRIT "%s: called for  not allocated channel %d\n",
+		       __func__, channel);
+		return 0;
+	}
+
+	now = min(imxdma->resbytes, sg->length);
+	imxdma->resbytes -= now;
+
+	if ((imxdma->dma_mode & DMA_MODE_MASK) == DMA_MODE_READ)
+		__raw_writel(sg->dma_address, DMA_BASE + DMA_DAR(channel));
+	else
+		__raw_writel(sg->dma_address, DMA_BASE + DMA_SAR(channel));
+
+	__raw_writel(now, DMA_BASE + DMA_CNTR(channel));
+
+	pr_debug("imxdma%d: next sg chunk dst 0x%08x, src 0x%08x, "
+		"size 0x%08x\n", channel,
+		 __raw_readl(DMA_BASE + DMA_DAR(channel)),
+		 __raw_readl(DMA_BASE + DMA_SAR(channel)),
+		 __raw_readl(DMA_BASE + DMA_CNTR(channel)));
+
+	return now;
+}
+
+/**
+ * imx_dma_setup_single - setup i.MX DMA channel for linear memory to/from
+ * device transfer
+ *
+ * @channel: i.MX DMA channel number
+ * @dma_address: the DMA/physical memory address of the linear data block
+ *		to transfer
+ * @dma_length: length of the data block in bytes
+ * @dev_addr: physical device port address
+ * @dmamode: DMA transfer mode, %DMA_MODE_READ from the device to the memory
+ *           or %DMA_MODE_WRITE from memory to the device
+ *
+ * Return value: if incorrect parameters are provided -%EINVAL.
+ *		Zero indicates success.
+ */
+int
+imx_dma_setup_single(int channel, dma_addr_t dma_address,
+		     unsigned int dma_length, unsigned int dev_addr,
+		     dmamode_t dmamode)
+{
+	struct imx_dma_channel *imxdma = &imx_dma_channels[channel];
+
+	imxdma->sg = NULL;
+	imxdma->dma_mode = dmamode;
+
+	if (!dma_address) {
+		printk(KERN_ERR "imxdma%d: imx_dma_setup_single null address\n",
+		       channel);
+		return -EINVAL;
+	}
+
+	if (!dma_length) {
+		printk(KERN_ERR "imxdma%d: imx_dma_setup_single zero length\n",
+		       channel);
+		return -EINVAL;
+	}
+
+	if ((dmamode & DMA_MODE_MASK) == DMA_MODE_READ) {
+		pr_debug("imxdma%d: %s dma_addressg=0x%08x dma_length=%d "
+			"dev_addr=0x%08x for read\n",
+			channel, __func__, (unsigned int)dma_address,
+			dma_length, dev_addr);
+
+		__raw_writel(dev_addr, DMA_BASE + DMA_SAR(channel));
+		__raw_writel(dma_address, DMA_BASE + DMA_DAR(channel));
+		__raw_writel(imxdma->ccr_from_device,
+				DMA_BASE + DMA_CCR(channel));
+	} else if ((dmamode & DMA_MODE_MASK) == DMA_MODE_WRITE) {
+		pr_debug("imxdma%d: %s dma_addressg=0x%08x dma_length=%d "
+			"dev_addr=0x%08x for write\n",
+			channel, __func__, (unsigned int)dma_address,
+			dma_length, dev_addr);
+
+		__raw_writel(dma_address, DMA_BASE + DMA_SAR(channel));
+		__raw_writel(dev_addr, DMA_BASE + DMA_DAR(channel));
+		__raw_writel(imxdma->ccr_to_device,
+				DMA_BASE + DMA_CCR(channel));
+	} else {
+		printk(KERN_ERR "imxdma%d: imx_dma_setup_single bad dmamode\n",
+		       channel);
+		return -EINVAL;
+	}
+
+	__raw_writel(dma_length, DMA_BASE + DMA_CNTR(channel));
+
+	return 0;
+}
+EXPORT_SYMBOL(imx_dma_setup_single);
+
+/**
+ * imx_dma_setup_sg - setup i.MX DMA channel SG list to/from device transfer
+ * @channel: i.MX DMA channel number
+ * @sg: pointer to the scatter-gather list/vector
+ * @sgcount: scatter-gather list hungs count
+ * @dma_length: total length of the transfer request in bytes
+ * @dev_addr: physical device port address
+ * @dmamode: DMA transfer mode, %DMA_MODE_READ from the device to the memory
+ *           or %DMA_MODE_WRITE from memory to the device
+ *
+ * The function sets up DMA channel state and registers to be ready for
+ * transfer specified by provided parameters. The scatter-gather emulation
+ * is set up according to the parameters.
+ *
+ * The full preparation of the transfer requires setup of more register
+ * by the caller before imx_dma_enable() can be called.
+ *
+ * %BLR(channel) holds transfer burst length in bytes, 0 means 64 bytes
+ *
+ * %RSSR(channel) has to be set to the DMA request line source %DMA_REQ_xxx
+ *
+ * %CCR(channel) has to specify transfer parameters, the next settings is
+ * typical for linear or simple scatter-gather transfers if %DMA_MODE_READ is
+ * specified
+ *
+ * %CCR_DMOD_LINEAR | %CCR_DSIZ_32 | %CCR_SMOD_FIFO | %CCR_SSIZ_x
+ *
+ * The typical setup for %DMA_MODE_WRITE is specified by next options
+ * combination
+ *
+ * %CCR_SMOD_LINEAR | %CCR_SSIZ_32 | %CCR_DMOD_FIFO | %CCR_DSIZ_x
+ *
+ * Be careful here and do not mistakenly mix source and target device
+ * port sizes constants, they are really different:
+ * %CCR_SSIZ_8, %CCR_SSIZ_16, %CCR_SSIZ_32,
+ * %CCR_DSIZ_8, %CCR_DSIZ_16, %CCR_DSIZ_32
+ *
+ * Return value: if incorrect parameters are provided -%EINVAL.
+ * Zero indicates success.
+ */
+int
+imx_dma_setup_sg(int channel,
+		 struct scatterlist *sg, unsigned int sgcount,
+		 unsigned int dma_length, unsigned int dev_addr,
+		 dmamode_t dmamode)
+{
+	struct imx_dma_channel *imxdma = &imx_dma_channels[channel];
+
+	if (imxdma->in_use)
+		return -EBUSY;
+
+	imxdma->sg = sg;
+	imxdma->dma_mode = dmamode;
+	imxdma->resbytes = dma_length;
+
+	if (!sg || !sgcount) {
+		printk(KERN_ERR "imxdma%d: imx_dma_setup_sg epty sg list\n",
+		       channel);
+		return -EINVAL;
+	}
+
+	if (!sg->length) {
+		printk(KERN_ERR "imxdma%d: imx_dma_setup_sg zero length\n",
+		       channel);
+		return -EINVAL;
+	}
+
+	if ((dmamode & DMA_MODE_MASK) == DMA_MODE_READ) {
+		pr_debug("imxdma%d: %s sg=%p sgcount=%d total length=%d "
+			"dev_addr=0x%08x for read\n",
+			channel, __func__, sg, sgcount, dma_length, dev_addr);
+
+		__raw_writel(dev_addr, DMA_BASE + DMA_SAR(channel));
+		__raw_writel(imxdma->ccr_from_device,
+				DMA_BASE + DMA_CCR(channel));
+	} else if ((dmamode & DMA_MODE_MASK) == DMA_MODE_WRITE) {
+		pr_debug("imxdma%d: %s sg=%p sgcount=%d total length=%d "
+			"dev_addr=0x%08x for write\n",
+			channel, __func__, sg, sgcount, dma_length, dev_addr);
+
+		__raw_writel(dev_addr, DMA_BASE + DMA_DAR(channel));
+		__raw_writel(imxdma->ccr_to_device,
+				DMA_BASE + DMA_CCR(channel));
+	} else {
+		printk(KERN_ERR "imxdma%d: imx_dma_setup_sg bad dmamode\n",
+		       channel);
+		return -EINVAL;
+	}
+
+	imx_dma_sg_next(channel, sg);
+
+	return 0;
+}
+EXPORT_SYMBOL(imx_dma_setup_sg);
+
+int
+imx_dma_config_channel(int channel, unsigned int config_port,
+	unsigned int config_mem, unsigned int dmareq, int hw_chaining)
+{
+	struct imx_dma_channel *imxdma = &imx_dma_channels[channel];
+	u32 dreq = 0;
+
+	imxdma->hw_chaining = 0;
+
+	if (hw_chaining) {
+		imxdma->hw_chaining = 1;
+		if (!imx_dma_hw_chain(imxdma))
+			return -EINVAL;
+	}
+
+	if (dmareq)
+		dreq = CCR_REN;
+
+	imxdma->ccr_from_device = config_port | (config_mem << 2) | dreq;
+	imxdma->ccr_to_device = config_mem | (config_port << 2) | dreq;
+
+	__raw_writel(dmareq, DMA_BASE + DMA_RSSR(channel));
+
+	return 0;
+}
+EXPORT_SYMBOL(imx_dma_config_channel);
+
+void imx_dma_config_burstlen(int channel, unsigned int burstlen)
+{
+	__raw_writel(burstlen, DMA_BASE + DMA_BLR(channel));
+}
+EXPORT_SYMBOL(imx_dma_config_burstlen);
+
+/**
+ * imx_dma_setup_handlers - setup i.MX DMA channel end and error notification
+ * handlers
+ * @channel: i.MX DMA channel number
+ * @irq_handler: the pointer to the function called if the transfer
+ *		ends successfully
+ * @err_handler: the pointer to the function called if the premature
+ *		end caused by error occurs
+ * @data: user specified value to be passed to the handlers
+ */
+int
+imx_dma_setup_handlers(int channel,
+		       void (*irq_handler) (int, void *),
+		       void (*err_handler) (int, void *, int),
+		       void *data)
+{
+	struct imx_dma_channel *imxdma = &imx_dma_channels[channel];
+	unsigned long flags;
+
+	if (!imxdma->name) {
+		printk(KERN_CRIT "%s: called for  not allocated channel %d\n",
+		       __func__, channel);
+		return -ENODEV;
+	}
+
+	local_irq_save(flags);
+	__raw_writel(1 << channel, DMA_BASE + DMA_DISR);
+	imxdma->irq_handler = irq_handler;
+	imxdma->err_handler = err_handler;
+	imxdma->data = data;
+	local_irq_restore(flags);
+	return 0;
+}
+EXPORT_SYMBOL(imx_dma_setup_handlers);
+
+/**
+ * imx_dma_setup_progression_handler - setup i.MX DMA channel progression
+ * handlers
+ * @channel: i.MX DMA channel number
+ * @prog_handler: the pointer to the function called if the transfer progresses
+ */
+int
+imx_dma_setup_progression_handler(int channel,
+			void (*prog_handler) (int, void*, struct scatterlist*))
+{
+	struct imx_dma_channel *imxdma = &imx_dma_channels[channel];
+	unsigned long flags;
+
+	if (!imxdma->name) {
+		printk(KERN_CRIT "%s: called for  not allocated channel %d\n",
+		       __func__, channel);
+		return -ENODEV;
+	}
+
+	local_irq_save(flags);
+	imxdma->prog_handler = prog_handler;
+	local_irq_restore(flags);
+	return 0;
+}
+EXPORT_SYMBOL(imx_dma_setup_progression_handler);
+
+/**
+ * imx_dma_enable - function to start i.MX DMA channel operation
+ * @channel: i.MX DMA channel number
+ *
+ * The channel has to be allocated by driver through imx_dma_request()
+ * or imx_dma_request_by_prio() function.
+ * The transfer parameters has to be set to the channel registers through
+ * call of the imx_dma_setup_single() or imx_dma_setup_sg() function
+ * and registers %BLR(channel), %RSSR(channel) and %CCR(channel) has to
+ * be set prior this function call by the channel user.
+ */
+void imx_dma_enable(int channel)
+{
+	struct imx_dma_channel *imxdma = &imx_dma_channels[channel];
+	unsigned long flags;
+
+	pr_debug("imxdma%d: imx_dma_enable\n", channel);
+
+	if (!imxdma->name) {
+		printk(KERN_CRIT "%s: called for  not allocated channel %d\n",
+		       __func__, channel);
+		return;
+	}
+
+	if (imxdma->in_use)
+		return;
+
+	local_irq_save(flags);
+
+	__raw_writel(1 << channel, DMA_BASE + DMA_DISR);
+	__raw_writel(__raw_readl(DMA_BASE + DMA_DIMR) & ~(1 << channel),
+		DMA_BASE + DMA_DIMR);
+	__raw_writel(__raw_readl(DMA_BASE + DMA_CCR(channel)) | CCR_CEN |
+		CCR_ACRPT,
+		DMA_BASE + DMA_CCR(channel));
+
+#ifdef CONFIG_ARCH_MX2
+	if (imxdma->sg && imx_dma_hw_chain(imxdma)) {
+		imxdma->sg = sg_next(imxdma->sg);
+		if (imxdma->sg) {
+			u32 tmp;
+			imx_dma_sg_next(channel, imxdma->sg);
+			tmp = __raw_readl(DMA_BASE + DMA_CCR(channel));
+			__raw_writel(tmp | CCR_RPT | CCR_ACRPT,
+				DMA_BASE + DMA_CCR(channel));
+		}
+	}
+#endif
+	imxdma->in_use = 1;
+
+	local_irq_restore(flags);
+}
+EXPORT_SYMBOL(imx_dma_enable);
+
+/**
+ * imx_dma_disable - stop, finish i.MX DMA channel operatin
+ * @channel: i.MX DMA channel number
+ */
+void imx_dma_disable(int channel)
+{
+	struct imx_dma_channel *imxdma = &imx_dma_channels[channel];
+	unsigned long flags;
+
+	pr_debug("imxdma%d: imx_dma_disable\n", channel);
+
+	if (imx_dma_hw_chain(imxdma))
+		del_timer(&imxdma->watchdog);
+
+	local_irq_save(flags);
+	__raw_writel(__raw_readl(DMA_BASE + DMA_DIMR) | (1 << channel),
+		DMA_BASE + DMA_DIMR);
+	__raw_writel(__raw_readl(DMA_BASE + DMA_CCR(channel)) & ~CCR_CEN,
+		DMA_BASE + DMA_CCR(channel));
+	__raw_writel(1 << channel, DMA_BASE + DMA_DISR);
+	imxdma->in_use = 0;
+	local_irq_restore(flags);
+}
+EXPORT_SYMBOL(imx_dma_disable);
+
+static void imx_dma_watchdog(unsigned long chno)
+{
+	struct imx_dma_channel *imxdma = &imx_dma_channels[chno];
+
+	__raw_writel(0, DMA_BASE + DMA_CCR(chno));
+	imxdma->in_use = 0;
+	imxdma->sg = NULL;
+
+	if (imxdma->err_handler)
+		imxdma->err_handler(chno, imxdma->data, IMX_DMA_ERR_TIMEOUT);
+}
+
+static irqreturn_t dma_err_handler(int irq, void *dev_id)
+{
+	int i, disr;
+	struct imx_dma_channel *imxdma;
+	unsigned int err_mask;
+	int errcode;
+
+	disr = __raw_readl(DMA_BASE + DMA_DISR);
+
+	err_mask = __raw_readl(DMA_BASE + DMA_DBTOSR) |
+		   __raw_readl(DMA_BASE + DMA_DRTOSR) |
+		   __raw_readl(DMA_BASE + DMA_DSESR)  |
+		   __raw_readl(DMA_BASE + DMA_DBOSR);
+
+	if (!err_mask)
+		return IRQ_HANDLED;
+
+	__raw_writel(disr & err_mask, DMA_BASE + DMA_DISR);
+
+	for (i = 0; i < IMX_DMA_CHANNELS; i++) {
+		if (!(err_mask & (1 << i)))
+			continue;
+		imxdma = &imx_dma_channels[i];
+		errcode = 0;
+
+		if (__raw_readl(DMA_BASE + DMA_DBTOSR) & (1 << i)) {
+			__raw_writel(1 << i, DMA_BASE + DMA_DBTOSR);
+			errcode |= IMX_DMA_ERR_BURST;
+		}
+		if (__raw_readl(DMA_BASE + DMA_DRTOSR) & (1 << i)) {
+			__raw_writel(1 << i, DMA_BASE + DMA_DRTOSR);
+			errcode |= IMX_DMA_ERR_REQUEST;
+		}
+		if (__raw_readl(DMA_BASE + DMA_DSESR) & (1 << i)) {
+			__raw_writel(1 << i, DMA_BASE + DMA_DSESR);
+			errcode |= IMX_DMA_ERR_TRANSFER;
+		}
+		if (__raw_readl(DMA_BASE + DMA_DBOSR) & (1 << i)) {
+			__raw_writel(1 << i, DMA_BASE + DMA_DBOSR);
+			errcode |= IMX_DMA_ERR_BUFFER;
+		}
+		if (imxdma->name && imxdma->err_handler) {
+			imxdma->err_handler(i, imxdma->data, errcode);
+			continue;
+		}
+
+		imx_dma_channels[i].sg = NULL;
+
+		printk(KERN_WARNING
+		       "DMA timeout on channel %d (%s) -%s%s%s%s\n",
+		       i, imxdma->name,
+		       errcode & IMX_DMA_ERR_BURST ?    " burst" : "",
+		       errcode & IMX_DMA_ERR_REQUEST ?  " request" : "",
+		       errcode & IMX_DMA_ERR_TRANSFER ? " transfer" : "",
+		       errcode & IMX_DMA_ERR_BUFFER ?   " buffer" : "");
+	}
+	return IRQ_HANDLED;
+}
+
+static void dma_irq_handle_channel(int chno)
+{
+	struct imx_dma_channel *imxdma = &imx_dma_channels[chno];
+
+	if (!imxdma->name) {
+		/*
+		 * IRQ for an unregistered DMA channel:
+		 * let's clear the interrupts and disable it.
+		 */
+		printk(KERN_WARNING
+		       "spurious IRQ for DMA channel %d\n", chno);
+		return;
+	}
+
+	if (imxdma->sg) {
+		u32 tmp;
+		struct scatterlist *current_sg = imxdma->sg;
+		imxdma->sg = sg_next(imxdma->sg);
+
+		if (imxdma->sg) {
+			imx_dma_sg_next(chno, imxdma->sg);
+
+			tmp = __raw_readl(DMA_BASE + DMA_CCR(chno));
+
+			if (imx_dma_hw_chain(imxdma)) {
+				/* FIXME: The timeout should probably be
+				 * configurable
+				 */
+				mod_timer(&imxdma->watchdog,
+					jiffies + msecs_to_jiffies(500));
+
+				tmp |= CCR_CEN | CCR_RPT | CCR_ACRPT;
+				__raw_writel(tmp, DMA_BASE +
+						DMA_CCR(chno));
+			} else {
+				__raw_writel(tmp & ~CCR_CEN, DMA_BASE +
+						DMA_CCR(chno));
+				tmp |= CCR_CEN;
+			}
+
+			__raw_writel(tmp, DMA_BASE + DMA_CCR(chno));
+
+			if (imxdma->prog_handler)
+				imxdma->prog_handler(chno, imxdma->data,
+						current_sg);
+
+			return;
+		}
+
+		if (imx_dma_hw_chain(imxdma)) {
+			del_timer(&imxdma->watchdog);
+			return;
+		}
+	}
+
+	__raw_writel(0, DMA_BASE + DMA_CCR(chno));
+	imxdma->in_use = 0;
+	if (imxdma->irq_handler)
+		imxdma->irq_handler(chno, imxdma->data);
+}
+
+static irqreturn_t dma_irq_handler(int irq, void *dev_id)
+{
+	int i, disr;
+
+#ifdef CONFIG_ARCH_MX2
+	dma_err_handler(irq, dev_id);
+#endif
+
+	disr = __raw_readl(DMA_BASE + DMA_DISR);
+
+	pr_debug("imxdma: dma_irq_handler called, disr=0x%08x\n",
+		     disr);
+
+	__raw_writel(disr, DMA_BASE + DMA_DISR);
+	for (i = 0; i < IMX_DMA_CHANNELS; i++) {
+		if (disr & (1 << i))
+			dma_irq_handle_channel(i);
+	}
+
+	return IRQ_HANDLED;
+}
+
+/**
+ * imx_dma_request - request/allocate specified channel number
+ * @channel: i.MX DMA channel number
+ * @name: the driver/caller own non-%NULL identification
+ */
+int imx_dma_request(int channel, const char *name)
+{
+	struct imx_dma_channel *imxdma = &imx_dma_channels[channel];
+	unsigned long flags;
+	int ret;
+
+	/* basic sanity checks */
+	if (!name)
+		return -EINVAL;
+
+	if (channel >= IMX_DMA_CHANNELS) {
+		printk(KERN_CRIT "%s: called for  non-existed channel %d\n",
+		       __func__, channel);
+		return -EINVAL;
+	}
+
+	local_irq_save(flags);
+	if (imxdma->name) {
+		local_irq_restore(flags);
+		return -EBUSY;
+	}
+
+#ifdef CONFIG_ARCH_MX2
+	ret = request_irq(MXC_INT_DMACH0 + channel, dma_irq_handler, 0, "DMA",
+			NULL);
+	if (ret) {
+		printk(KERN_CRIT "Can't register IRQ %d for DMA channel %d\n",
+				MXC_INT_DMACH0 + channel, channel);
+		return ret;
+	}
+	init_timer(&imxdma->watchdog);
+	imxdma->watchdog.function = &imx_dma_watchdog;
+	imxdma->watchdog.data = channel;
+#endif
+
+	imxdma->name = name;
+	imxdma->irq_handler = NULL;
+	imxdma->err_handler = NULL;
+	imxdma->data = NULL;
+	imxdma->sg = NULL;
+
+	local_irq_restore(flags);
+	return 0;
+}
+EXPORT_SYMBOL(imx_dma_request);
+
+/**
+ * imx_dma_free - release previously acquired channel
+ * @channel: i.MX DMA channel number
+ */
+void imx_dma_free(int channel)
+{
+	unsigned long flags;
+	struct imx_dma_channel *imxdma = &imx_dma_channels[channel];
+
+	if (!imxdma->name) {
+		printk(KERN_CRIT
+		       "%s: trying to free free channel %d\n",
+		       __func__, channel);
+		return;
+	}
+
+	local_irq_save(flags);
+	/* Disable interrupts */
+	__raw_writel(__raw_readl(DMA_BASE + DMA_DIMR) | (1 << channel),
+		DMA_BASE + DMA_DIMR);
+	__raw_writel(__raw_readl(DMA_BASE + DMA_CCR(channel)) & ~CCR_CEN,
+		DMA_BASE + DMA_CCR(channel));
+	imxdma->name = NULL;
+
+#ifdef CONFIG_ARCH_MX2
+	free_irq(MXC_INT_DMACH0 + channel, NULL);
+#endif
+
+	local_irq_restore(flags);
+}
+EXPORT_SYMBOL(imx_dma_free);
+
+/**
+ * imx_dma_request_by_prio - find and request some of free channels best
+ * suiting requested priority
+ * @channel: i.MX DMA channel number
+ * @name: the driver/caller own non-%NULL identification
+ *
+ * This function tries to find a free channel in the specified priority group
+ * This function tries to find a free channel in the specified priority group
+ * if the priority cannot be achieved it tries to look for free channel
+ * in the higher and then even lower priority groups.
+ *
+ * Return value: If there is no free channel to allocate, -%ENODEV is returned.
+ *               On successful allocation channel is returned.
+ */
+int imx_dma_request_by_prio(const char *name, enum imx_dma_prio prio)
+{
+	int i;
+	int best;
+
+	switch (prio) {
+	case (DMA_PRIO_HIGH):
+		best = 8;
+		break;
+	case (DMA_PRIO_MEDIUM):
+		best = 4;
+		break;
+	case (DMA_PRIO_LOW):
+	default:
+		best = 0;
+		break;
+	}
+
+	for (i = best; i < IMX_DMA_CHANNELS; i++)
+		if (!imx_dma_request(i, name))
+			return i;
+
+	for (i = best - 1; i >= 0; i--)
+		if (!imx_dma_request(i, name))
+			return i;
+
+	printk(KERN_ERR "%s: no free DMA channel found\n", __func__);
+
+	return -ENODEV;
+}
+EXPORT_SYMBOL(imx_dma_request_by_prio);
+
+static int __init imx_dma_init(void)
+{
+	int ret = 0;
+	int i;
+
+	dma_clk = clk_get(NULL, "dma_clk");
+	clk_enable(dma_clk);
+
+	/* reset DMA module */
+	__raw_writel(DCR_DRST, DMA_BASE + DMA_DCR);
+
+#ifdef CONFIG_ARCH_MX1
+	ret = request_irq(DMA_INT, dma_irq_handler, 0, "DMA", NULL);
+	if (ret) {
+		printk(KERN_CRIT "Wow!  Can't register IRQ for DMA\n");
+		return ret;
+	}
+
+	ret = request_irq(DMA_ERR, dma_err_handler, 0, "DMA", NULL);
+	if (ret) {
+		printk(KERN_CRIT "Wow!  Can't register ERRIRQ for DMA\n");
+		free_irq(DMA_INT, NULL);
+		return ret;
+	}
+#endif
+	/* enable DMA module */
+	__raw_writel(DCR_DEN, DMA_BASE + DMA_DCR);
+
+	/* clear all interrupts */
+	__raw_writel((1 << IMX_DMA_CHANNELS) - 1, DMA_BASE + DMA_DISR);
+
+	/* disable interrupts */
+	__raw_writel((1 << IMX_DMA_CHANNELS) - 1, DMA_BASE + DMA_DIMR);
+
+	for (i = 0; i < IMX_DMA_CHANNELS; i++) {
+		imx_dma_channels[i].sg = NULL;
+		imx_dma_channels[i].dma_num = i;
+	}
+
+	return ret;
+}
+
+arch_initcall(imx_dma_init);

arch/arm/plat-mxc/include/mach/dma-mx1-mx2.h

+/*
+ *  linux/arch/arm/plat-mxc/include/mach/dma-mx1-mx2.h
+ *
+ *  i.MX DMA registration and IRQ dispatching
+ *
+ * Copyright 2006 Pavel Pisa <pisa@cmp.felk.cvut.cz>
+ * Copyright 2008 Juergen Beisert, <kernel@pengutronix.de>
+ * Copyright 2008 Sascha Hauer, <s.hauer@pengutronix.de>
+ *
+ * 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 <asm/dma.h>
+
+#ifndef __ASM_ARCH_MXC_DMA_H
+#define __ASM_ARCH_MXC_DMA_H
+
+#define IMX_DMA_CHANNELS  16
+
+#define DMA_BASE IO_ADDRESS(DMA_BASE_ADDR)
+
+#define IMX_DMA_MEMSIZE_32	(0 << 4)
+#define IMX_DMA_MEMSIZE_8	(1 << 4)
+#define IMX_DMA_MEMSIZE_16	(2 << 4)
+#define IMX_DMA_TYPE_LINEAR	(0 << 10)
+#define IMX_DMA_TYPE_2D		(1 << 10)
+#define IMX_DMA_TYPE_FIFO	(2 << 10)
+
+#define IMX_DMA_ERR_BURST     (1 << 0)
+#define IMX_DMA_ERR_REQUEST   (1 << 1)
+#define IMX_DMA_ERR_TRANSFER  (1 << 2)
+#define IMX_DMA_ERR_BUFFER    (1 << 3)
+#define IMX_DMA_ERR_TIMEOUT   (1 << 4)
+
+int
+imx_dma_config_channel(int channel, unsigned int config_port,
+	unsigned int config_mem, unsigned int dmareq, int hw_chaining);
+
+void
+imx_dma_config_burstlen(int channel, unsigned int burstlen);
+
+int
+imx_dma_setup_single(int channel, dma_addr_t dma_address,
+		unsigned int dma_length, unsigned int dev_addr,
+		dmamode_t dmamode);
+
+int
+imx_dma_setup_sg(int channel, struct scatterlist *sg,
+		unsigned int sgcount, unsigned int dma_length,
+		unsigned int dev_addr, dmamode_t dmamode);
+
+int
+imx_dma_setup_handlers(int channel,
+		void (*irq_handler) (int, void *),
+		void (*err_handler) (int, void *, int), void *data);
+
+int
+imx_dma_setup_progression_handler(int channel,
+		void (*prog_handler) (int, void*, struct scatterlist*));
+
+void imx_dma_enable(int channel);
+
+void imx_dma_disable(int channel);
+
+int imx_dma_request(int channel, const char *name);
+
+void imx_dma_free(int channel);
+
+enum imx_dma_prio {
+	DMA_PRIO_HIGH = 0,
+	DMA_PRIO_MEDIUM = 1,
+	DMA_PRIO_LOW = 2
+};
+
+int imx_dma_request_by_prio(const char *name, enum imx_dma_prio prio);
+
+#endif	/* _ASM_ARCH_MXC_DMA_H */