dmaengine: mmp-pdma support

1. virtual channel vs. physical channel
Virtual channel is managed by dmaengine
Physical channel handling resource, such as irq
Physical channel is alloced dynamically as descending priority,
freed immediately when irq done.
The availble highest priority physically channel will alwayes be alloced

Issue pending list -> alloc highest dma physically channel available -> dma done -> free physically channel

2. list: running list & pending list
submit: desc list -> pending list
issue_pending_list: if (IDLE) pending list -> running list; free pending list (RUN)
irq: free running list (IDLE)
     check pendlist -> pending list -> running list; free pending list (RUN)

3. irq:
Each list generate one irq, calling callback
One list may contain several desc chain, in such case, make sure only the last desc list generate irq.

4. async
Submit will add desc chain to pending list, which can be multi-called
If multi desc chain is submitted, only the last desc would generate irq -> call back
If IDLE, issue_pending_list start pending_list, transforming pendlist to running list
If RUN, irq will start pending list

5. test
5.1 pxa3xx_nand on pxa910
5.2 insmod dmatest.ko (threads_per_chan=y)
By default drivers/dma/dmatest.c test every channel and test memcpy with 1 threads per channel

Signed-off-by: Zhangfei Gao <zhangfei.gao@marvell.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Vinod Koul <vinod.koul@linux.intel.com>

1 files changed, 875 insertions, 0 deletions

diff --git a/drivers/dma/mmp_pdma.c b/drivers/dma/mmp_pdma.c
new file mode 100644
index 000000000000..14da1f403edf
--- /dev/null
+++ b/drivers/dma/mmp_pdma.c
@@ -0,0 +1,875 @@
+/*
+ * Copyright 2012 Marvell International Ltd.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/types.h>
+#include <linux/interrupt.h>
+#include <linux/dma-mapping.h>
+#include <linux/slab.h>
+#include <linux/dmaengine.h>
+#include <linux/platform_device.h>
+#include <linux/device.h>
+#include <linux/platform_data/mmp_dma.h>
+#include <linux/dmapool.h>
+#include <linux/of_device.h>
+#include <linux/of.h>
+
+#include "dmaengine.h"
+
+#define DCSR		0x0000
+#define DALGN		0x00a0
+#define DINT		0x00f0
+#define DDADR		0x0200
+#define DSADR		0x0204
+#define DTADR		0x0208
+#define DCMD		0x020c
+
+#define DCSR_RUN	(1 << 31)	/* Run Bit (read / write) */
+#define DCSR_NODESC	(1 << 30)	/* No-Descriptor Fetch (read / write) */
+#define DCSR_STOPIRQEN	(1 << 29)	/* Stop Interrupt Enable (read / write) */
+#define DCSR_REQPEND	(1 << 8)	/* Request Pending (read-only) */
+#define DCSR_STOPSTATE	(1 << 3)	/* Stop State (read-only) */
+#define DCSR_ENDINTR	(1 << 2)	/* End Interrupt (read / write) */
+#define DCSR_STARTINTR	(1 << 1)	/* Start Interrupt (read / write) */
+#define DCSR_BUSERR	(1 << 0)	/* Bus Error Interrupt (read / write) */
+
+#define DCSR_EORIRQEN	(1 << 28)       /* End of Receive Interrupt Enable (R/W) */
+#define DCSR_EORJMPEN	(1 << 27)       /* Jump to next descriptor on EOR */
+#define DCSR_EORSTOPEN	(1 << 26)       /* STOP on an EOR */
+#define DCSR_SETCMPST	(1 << 25)       /* Set Descriptor Compare Status */
+#define DCSR_CLRCMPST	(1 << 24)       /* Clear Descriptor Compare Status */
+#define DCSR_CMPST	(1 << 10)       /* The Descriptor Compare Status */
+#define DCSR_EORINTR	(1 << 9)        /* The end of Receive */
+
+#define DRCMR_MAPVLD	(1 << 7)	/* Map Valid (read / write) */
+#define DRCMR_CHLNUM	0x1f		/* mask for Channel Number (read / write) */
+
+#define DDADR_DESCADDR	0xfffffff0	/* Address of next descriptor (mask) */
+#define DDADR_STOP	(1 << 0)	/* Stop (read / write) */
+
+#define DCMD_INCSRCADDR	(1 << 31)	/* Source Address Increment Setting. */
+#define DCMD_INCTRGADDR	(1 << 30)	/* Target Address Increment Setting. */
+#define DCMD_FLOWSRC	(1 << 29)	/* Flow Control by the source. */
+#define DCMD_FLOWTRG	(1 << 28)	/* Flow Control by the target. */
+#define DCMD_STARTIRQEN	(1 << 22)	/* Start Interrupt Enable */
+#define DCMD_ENDIRQEN	(1 << 21)	/* End Interrupt Enable */
+#define DCMD_ENDIAN	(1 << 18)	/* Device Endian-ness. */
+#define DCMD_BURST8	(1 << 16)	/* 8 byte burst */
+#define DCMD_BURST16	(2 << 16)	/* 16 byte burst */
+#define DCMD_BURST32	(3 << 16)	/* 32 byte burst */
+#define DCMD_WIDTH1	(1 << 14)	/* 1 byte width */
+#define DCMD_WIDTH2	(2 << 14)	/* 2 byte width (HalfWord) */
+#define DCMD_WIDTH4	(3 << 14)	/* 4 byte width (Word) */
+#define DCMD_LENGTH	0x01fff		/* length mask (max = 8K - 1) */
+
+#define PDMA_ALIGNMENT		3
+#define PDMA_MAX_DESC_BYTES	0x1000
+
+struct mmp_pdma_desc_hw {
+	u32 ddadr;	/* Points to the next descriptor + flags */
+	u32 dsadr;	/* DSADR value for the current transfer */
+	u32 dtadr;	/* DTADR value for the current transfer */
+	u32 dcmd;	/* DCMD value for the current transfer */
+} __aligned(32);
+
+struct mmp_pdma_desc_sw {
+	struct mmp_pdma_desc_hw desc;
+	struct list_head node;
+	struct list_head tx_list;
+	struct dma_async_tx_descriptor async_tx;
+};
+
+struct mmp_pdma_phy;
+
+struct mmp_pdma_chan {
+	struct device *dev;
+	struct dma_chan chan;
+	struct dma_async_tx_descriptor desc;
+	struct mmp_pdma_phy *phy;
+	enum dma_transfer_direction dir;
+
+	/* channel's basic info */
+	struct tasklet_struct tasklet;
+	u32 dcmd;
+	u32 drcmr;
+	u32 dev_addr;
+
+	/* list for desc */
+	spinlock_t desc_lock;		/* Descriptor list lock */
+	struct list_head chain_pending;	/* Link descriptors queue for pending */
+	struct list_head chain_running;	/* Link descriptors queue for running */
+	bool idle;			/* channel statue machine */
+
+	struct dma_pool *desc_pool;	/* Descriptors pool */
+};
+
+struct mmp_pdma_phy {
+	int idx;
+	void __iomem *base;
+	struct mmp_pdma_chan *vchan;
+};
+
+struct mmp_pdma_device {
+	int				dma_channels;
+	void __iomem			*base;
+	struct device			*dev;
+	struct dma_device		device;
+	struct mmp_pdma_phy		*phy;
+};
+
+#define tx_to_mmp_pdma_desc(tx) container_of(tx, struct mmp_pdma_desc_sw, async_tx)
+#define to_mmp_pdma_desc(lh) container_of(lh, struct mmp_pdma_desc_sw, node)
+#define to_mmp_pdma_chan(dchan) container_of(dchan, struct mmp_pdma_chan, chan)
+#define to_mmp_pdma_dev(dmadev) container_of(dmadev, struct mmp_pdma_device, device)
+
+static void set_desc(struct mmp_pdma_phy *phy, dma_addr_t addr)
+{
+	u32 reg = (phy->idx << 4) + DDADR;
+
+	writel(addr, phy->base + reg);
+}
+
+static void enable_chan(struct mmp_pdma_phy *phy)
+{
+	u32 reg;
+
+	if (!phy->vchan)
+		return;
+
+	reg = phy->vchan->drcmr;
+	reg = (((reg) < 64) ? 0x0100 : 0x1100) + (((reg) & 0x3f) << 2);
+	writel(DRCMR_MAPVLD | phy->idx, phy->base + reg);
+
+	reg = (phy->idx << 2) + DCSR;
+	writel(readl(phy->base + reg) | DCSR_RUN,
+					phy->base + reg);
+}
+
+static void disable_chan(struct mmp_pdma_phy *phy)
+{
+	u32 reg;
+
+	if (phy) {
+		reg = (phy->idx << 2) + DCSR;
+		writel(readl(phy->base + reg) & ~DCSR_RUN,
+						phy->base + reg);
+	}
+}
+
+static int clear_chan_irq(struct mmp_pdma_phy *phy)
+{
+	u32 dcsr;
+	u32 dint = readl(phy->base + DINT);
+	u32 reg = (phy->idx << 2) + DCSR;
+
+	if (dint & BIT(phy->idx)) {
+		/* clear irq */
+		dcsr = readl(phy->base + reg);
+		writel(dcsr, phy->base + reg);
+		if ((dcsr & DCSR_BUSERR) && (phy->vchan))
+			dev_warn(phy->vchan->dev, "DCSR_BUSERR\n");
+		return 0;
+	}
+	return -EAGAIN;
+}
+
+static irqreturn_t mmp_pdma_chan_handler(int irq, void *dev_id)
+{
+	struct mmp_pdma_phy *phy = dev_id;
+
+	if (clear_chan_irq(phy) == 0) {
+		tasklet_schedule(&phy->vchan->tasklet);
+		return IRQ_HANDLED;
+	} else
+		return IRQ_NONE;
+}
+
+static irqreturn_t mmp_pdma_int_handler(int irq, void *dev_id)
+{
+	struct mmp_pdma_device *pdev = dev_id;
+	struct mmp_pdma_phy *phy;
+	u32 dint = readl(pdev->base + DINT);
+	int i, ret;
+	int irq_num = 0;
+
+	while (dint) {
+		i = __ffs(dint);
+		dint &= (dint - 1);
+		phy = &pdev->phy[i];
+		ret = mmp_pdma_chan_handler(irq, phy);
+		if (ret == IRQ_HANDLED)
+			irq_num++;
+	}
+
+	if (irq_num)
+		return IRQ_HANDLED;
+	else
+		return IRQ_NONE;
+}
+
+/* lookup free phy channel as descending priority */
+static struct mmp_pdma_phy *lookup_phy(struct mmp_pdma_chan *pchan)
+{
+	int prio, i;
+	struct mmp_pdma_device *pdev = to_mmp_pdma_dev(pchan->chan.device);
+	struct mmp_pdma_phy *phy;
+
+	/*
+	 * dma channel priorities
+	 * ch 0 - 3,  16 - 19  <--> (0)
+	 * ch 4 - 7,  20 - 23  <--> (1)
+	 * ch 8 - 11, 24 - 27  <--> (2)
+	 * ch 12 - 15, 28 - 31  <--> (3)
+	 */
+	for (prio = 0; prio <= (((pdev->dma_channels - 1) & 0xf) >> 2); prio++) {
+		for (i = 0; i < pdev->dma_channels; i++) {
+			if (prio != ((i & 0xf) >> 2))
+				continue;
+			phy = &pdev->phy[i];
+			if (!phy->vchan) {
+				phy->vchan = pchan;
+				return phy;
+			}
+		}
+	}
+
+	return NULL;
+}
+
+/* desc->tx_list ==> pending list */
+static void append_pending_queue(struct mmp_pdma_chan *chan,
+					struct mmp_pdma_desc_sw *desc)
+{
+	struct mmp_pdma_desc_sw *tail =
+				to_mmp_pdma_desc(chan->chain_pending.prev);
+
+	if (list_empty(&chan->chain_pending))
+		goto out_splice;
+
+	/* one irq per queue, even appended */
+	tail->desc.ddadr = desc->async_tx.phys;
+	tail->desc.dcmd &= ~DCMD_ENDIRQEN;
+
+	/* softly link to pending list */
+out_splice:
+	list_splice_tail_init(&desc->tx_list, &chan->chain_pending);
+}
+
+/**
+ * start_pending_queue - transfer any pending transactions
+ * pending list ==> running list
+ */
+static void start_pending_queue(struct mmp_pdma_chan *chan)
+{
+	struct mmp_pdma_desc_sw *desc;
+
+	/* still in running, irq will start the pending list */
+	if (!chan->idle) {
+		dev_dbg(chan->dev, "DMA controller still busy\n");
+		return;
+	}
+
+	if (list_empty(&chan->chain_pending)) {
+		/* chance to re-fetch phy channel with higher prio */
+		if (chan->phy) {
+			chan->phy->vchan = NULL;
+			chan->phy = NULL;
+		}
+		dev_dbg(chan->dev, "no pending list\n");
+		return;
+	}
+
+	if (!chan->phy) {
+		chan->phy = lookup_phy(chan);
+		if (!chan->phy) {
+			dev_dbg(chan->dev, "no free dma channel\n");
+			return;
+		}
+	}
+
+	/*
+	 * pending -> running
+	 * reintilize pending list
+	 */
+	desc = list_first_entry(&chan->chain_pending,
+				struct mmp_pdma_desc_sw, node);
+	list_splice_tail_init(&chan->chain_pending, &chan->chain_running);
+
+	/*
+	 * Program the descriptor's address into the DMA controller,
+	 * then start the DMA transaction
+	 */
+	set_desc(chan->phy, desc->async_tx.phys);
+	enable_chan(chan->phy);
+	chan->idle = false;
+}
+
+
+/* desc->tx_list ==> pending list */
+static dma_cookie_t mmp_pdma_tx_submit(struct dma_async_tx_descriptor *tx)
+{
+	struct mmp_pdma_chan *chan = to_mmp_pdma_chan(tx->chan);
+	struct mmp_pdma_desc_sw *desc = tx_to_mmp_pdma_desc(tx);
+	struct mmp_pdma_desc_sw *child;
+	unsigned long flags;
+	dma_cookie_t cookie = -EBUSY;
+
+	spin_lock_irqsave(&chan->desc_lock, flags);
+
+	list_for_each_entry(child, &desc->tx_list, node) {
+		cookie = dma_cookie_assign(&child->async_tx);
+	}
+
+	append_pending_queue(chan, desc);
+
+	spin_unlock_irqrestore(&chan->desc_lock, flags);
+
+	return cookie;
+}
+
+struct mmp_pdma_desc_sw *mmp_pdma_alloc_descriptor(struct mmp_pdma_chan *chan)
+{
+	struct mmp_pdma_desc_sw *desc;
+	dma_addr_t pdesc;
+
+	desc = dma_pool_alloc(chan->desc_pool, GFP_ATOMIC, &pdesc);
+	if (!desc) {
+		dev_err(chan->dev, "out of memory for link descriptor\n");
+		return NULL;
+	}
+
+	memset(desc, 0, sizeof(*desc));
+	INIT_LIST_HEAD(&desc->tx_list);
+	dma_async_tx_descriptor_init(&desc->async_tx, &chan->chan);
+	/* each desc has submit */
+	desc->async_tx.tx_submit = mmp_pdma_tx_submit;
+	desc->async_tx.phys = pdesc;
+
+	return desc;
+}
+
+/**
+ * mmp_pdma_alloc_chan_resources - Allocate resources for DMA channel.
+ *
+ * This function will create a dma pool for descriptor allocation.
+ * Request irq only when channel is requested
+ * Return - The number of allocated descriptors.
+ */
+
+static int mmp_pdma_alloc_chan_resources(struct dma_chan *dchan)
+{
+	struct mmp_pdma_chan *chan = to_mmp_pdma_chan(dchan);
+
+	if (chan->desc_pool)
+		return 1;
+
+	chan->desc_pool =
+		dma_pool_create(dev_name(&dchan->dev->device), chan->dev,
+				  sizeof(struct mmp_pdma_desc_sw),
+				  __alignof__(struct mmp_pdma_desc_sw), 0);
+	if (!chan->desc_pool) {
+		dev_err(chan->dev, "unable to allocate descriptor pool\n");
+		return -ENOMEM;
+	}
+	if (chan->phy) {
+		chan->phy->vchan = NULL;
+		chan->phy = NULL;
+	}
+	chan->idle = true;
+	chan->dev_addr = 0;
+	return 1;
+}
+
+static void mmp_pdma_free_desc_list(struct mmp_pdma_chan *chan,
+				  struct list_head *list)
+{
+	struct mmp_pdma_desc_sw *desc, *_desc;
+
+	list_for_each_entry_safe(desc, _desc, list, node) {
+		list_del(&desc->node);
+		dma_pool_free(chan->desc_pool, desc, desc->async_tx.phys);
+	}
+}
+
+static void mmp_pdma_free_chan_resources(struct dma_chan *dchan)
+{
+	struct mmp_pdma_chan *chan = to_mmp_pdma_chan(dchan);
+	unsigned long flags;
+
+	spin_lock_irqsave(&chan->desc_lock, flags);
+	mmp_pdma_free_desc_list(chan, &chan->chain_pending);
+	mmp_pdma_free_desc_list(chan, &chan->chain_running);
+	spin_unlock_irqrestore(&chan->desc_lock, flags);
+
+	dma_pool_destroy(chan->desc_pool);
+	chan->desc_pool = NULL;
+	chan->idle = true;
+	chan->dev_addr = 0;
+	if (chan->phy) {
+		chan->phy->vchan = NULL;
+		chan->phy = NULL;
+	}
+	return;
+}
+
+static struct dma_async_tx_descriptor *
+mmp_pdma_prep_memcpy(struct dma_chan *dchan,
+	dma_addr_t dma_dst, dma_addr_t dma_src,
+	size_t len, unsigned long flags)
+{
+	struct mmp_pdma_chan *chan;
+	struct mmp_pdma_desc_sw *first = NULL, *prev = NULL, *new;
+	size_t copy = 0;
+
+	if (!dchan)
+		return NULL;
+
+	if (!len)
+		return NULL;
+
+	chan = to_mmp_pdma_chan(dchan);
+
+	if (!chan->dir) {
+		chan->dir = DMA_MEM_TO_MEM;
+		chan->dcmd = DCMD_INCTRGADDR | DCMD_INCSRCADDR;
+		chan->dcmd |= DCMD_BURST32;
+	}
+
+	do {
+		/* Allocate the link descriptor from DMA pool */
+		new = mmp_pdma_alloc_descriptor(chan);
+		if (!new) {
+			dev_err(chan->dev, "no memory for desc\n");
+			goto fail;
+		}
+
+		copy = min_t(size_t, len, PDMA_MAX_DESC_BYTES);
+
+		new->desc.dcmd = chan->dcmd | (DCMD_LENGTH & copy);
+		new->desc.dsadr = dma_src;
+		new->desc.dtadr = dma_dst;
+
+		if (!first)
+			first = new;
+		else
+			prev->desc.ddadr = new->async_tx.phys;
+
+		new->async_tx.cookie = 0;
+		async_tx_ack(&new->async_tx);
+
+		prev = new;
+		len -= copy;
+
+		if (chan->dir == DMA_MEM_TO_DEV) {
+			dma_src += copy;
+		} else if (chan->dir == DMA_DEV_TO_MEM) {
+			dma_dst += copy;
+		} else if (chan->dir == DMA_MEM_TO_MEM) {
+			dma_src += copy;
+			dma_dst += copy;
+		}
+
+		/* Insert the link descriptor to the LD ring */
+		list_add_tail(&new->node, &first->tx_list);
+	} while (len);
+
+	first->async_tx.flags = flags; /* client is in control of this ack */
+	first->async_tx.cookie = -EBUSY;
+
+	/* last desc and fire IRQ */
+	new->desc.ddadr = DDADR_STOP;
+	new->desc.dcmd |= DCMD_ENDIRQEN;
+
+	return &first->async_tx;
+
+fail:
+	if (first)
+		mmp_pdma_free_desc_list(chan, &first->tx_list);
+	return NULL;
+}
+
+static struct dma_async_tx_descriptor *
+mmp_pdma_prep_slave_sg(struct dma_chan *dchan, struct scatterlist *sgl,
+			 unsigned int sg_len, enum dma_transfer_direction dir,
+			 unsigned long flags, void *context)
+{
+	struct mmp_pdma_chan *chan = to_mmp_pdma_chan(dchan);
+	struct mmp_pdma_desc_sw *first = NULL, *prev = NULL, *new = NULL;
+	size_t len, avail;
+	struct scatterlist *sg;
+	dma_addr_t addr;
+	int i;
+
+	if ((sgl == NULL) || (sg_len == 0))
+		return NULL;
+
+	for_each_sg(sgl, sg, sg_len, i) {
+		addr = sg_dma_address(sg);
+		avail = sg_dma_len(sgl);
+
+		do {
+			len = min_t(size_t, avail, PDMA_MAX_DESC_BYTES);
+
+			/* allocate and populate the descriptor */
+			new = mmp_pdma_alloc_descriptor(chan);
+			if (!new) {
+				dev_err(chan->dev, "no memory for desc\n");
+				goto fail;
+			}
+
+			new->desc.dcmd = chan->dcmd | (DCMD_LENGTH & len);
+			if (dir == DMA_MEM_TO_DEV) {
+				new->desc.dsadr = addr;
+				new->desc.dtadr = chan->dev_addr;
+			} else {
+				new->desc.dsadr = chan->dev_addr;
+				new->desc.dtadr = addr;
+			}
+
+			if (!first)
+				first = new;
+			else
+				prev->desc.ddadr = new->async_tx.phys;
+
+			new->async_tx.cookie = 0;
+			async_tx_ack(&new->async_tx);
+			prev = new;
+
+			/* Insert the link descriptor to the LD ring */
+			list_add_tail(&new->node, &first->tx_list);
+
+			/* update metadata */
+			addr += len;
+			avail -= len;
+		} while (avail);
+	}
+
+	first->async_tx.cookie = -EBUSY;
+	first->async_tx.flags = flags;
+
+	/* last desc and fire IRQ */
+	new->desc.ddadr = DDADR_STOP;
+	new->desc.dcmd |= DCMD_ENDIRQEN;
+
+	return &first->async_tx;
+
+fail:
+	if (first)
+		mmp_pdma_free_desc_list(chan, &first->tx_list);
+	return NULL;
+}
+
+static int mmp_pdma_control(struct dma_chan *dchan, enum dma_ctrl_cmd cmd,
+		unsigned long arg)
+{
+	struct mmp_pdma_chan *chan = to_mmp_pdma_chan(dchan);
+	struct dma_slave_config *cfg = (void *)arg;
+	unsigned long flags;
+	int ret = 0;
+	u32 maxburst = 0, addr = 0;
+	enum dma_slave_buswidth width = DMA_SLAVE_BUSWIDTH_UNDEFINED;
+
+	if (!dchan)
+		return -EINVAL;
+
+	switch (cmd) {
+	case DMA_TERMINATE_ALL:
+		disable_chan(chan->phy);
+		if (chan->phy) {
+			chan->phy->vchan = NULL;
+			chan->phy = NULL;
+		}
+		spin_lock_irqsave(&chan->desc_lock, flags);
+		mmp_pdma_free_desc_list(chan, &chan->chain_pending);
+		mmp_pdma_free_desc_list(chan, &chan->chain_running);
+		spin_unlock_irqrestore(&chan->desc_lock, flags);
+		chan->idle = true;
+		break;
+	case DMA_SLAVE_CONFIG:
+		if (cfg->direction == DMA_DEV_TO_MEM) {
+			chan->dcmd = DCMD_INCTRGADDR | DCMD_FLOWSRC;
+			maxburst = cfg->src_maxburst;
+			width = cfg->src_addr_width;
+			addr = cfg->src_addr;
+		} else if (cfg->direction == DMA_MEM_TO_DEV) {
+			chan->dcmd = DCMD_INCSRCADDR | DCMD_FLOWTRG;
+			maxburst = cfg->dst_maxburst;
+			width = cfg->dst_addr_width;
+			addr = cfg->dst_addr;
+		}
+
+		if (width == DMA_SLAVE_BUSWIDTH_1_BYTE)
+			chan->dcmd |= DCMD_WIDTH1;
+		else if (width == DMA_SLAVE_BUSWIDTH_2_BYTES)
+			chan->dcmd |= DCMD_WIDTH2;
+		else if (width == DMA_SLAVE_BUSWIDTH_4_BYTES)
+			chan->dcmd |= DCMD_WIDTH4;
+
+		if (maxburst == 8)
+			chan->dcmd |= DCMD_BURST8;
+		else if (maxburst == 16)
+			chan->dcmd |= DCMD_BURST16;
+		else if (maxburst == 32)
+			chan->dcmd |= DCMD_BURST32;
+
+		if (cfg) {
+			chan->dir = cfg->direction;
+			chan->drcmr = cfg->slave_id;
+		}
+		chan->dev_addr = addr;
+		break;
+	default:
+		return -ENOSYS;
+	}
+
+	return ret;
+}
+
+static enum dma_status mmp_pdma_tx_status(struct dma_chan *dchan,
+			dma_cookie_t cookie, struct dma_tx_state *txstate)
+{
+	struct mmp_pdma_chan *chan = to_mmp_pdma_chan(dchan);
+	enum dma_status ret;
+	unsigned long flags;
+
+	spin_lock_irqsave(&chan->desc_lock, flags);
+	ret = dma_cookie_status(dchan, cookie, txstate);
+	spin_unlock_irqrestore(&chan->desc_lock, flags);
+
+	return ret;
+}
+
+/**
+ * mmp_pdma_issue_pending - Issue the DMA start command
+ * pending list ==> running list
+ */
+static void mmp_pdma_issue_pending(struct dma_chan *dchan)
+{
+	struct mmp_pdma_chan *chan = to_mmp_pdma_chan(dchan);
+	unsigned long flags;
+
+	spin_lock_irqsave(&chan->desc_lock, flags);
+	start_pending_queue(chan);
+	spin_unlock_irqrestore(&chan->desc_lock, flags);
+}
+
+/*
+ * dma_do_tasklet
+ * Do call back
+ * Start pending list
+ */
+static void dma_do_tasklet(unsigned long data)
+{
+	struct mmp_pdma_chan *chan = (struct mmp_pdma_chan *)data;
+	struct mmp_pdma_desc_sw *desc, *_desc;
+	LIST_HEAD(chain_cleanup);
+	unsigned long flags;
+
+	/* submit pending list; callback for each desc; free desc */
+
+	spin_lock_irqsave(&chan->desc_lock, flags);
+
+	/* update the cookie if we have some descriptors to cleanup */
+	if (!list_empty(&chan->chain_running)) {
+		dma_cookie_t cookie;
+
+		desc = to_mmp_pdma_desc(chan->chain_running.prev);
+		cookie = desc->async_tx.cookie;
+		dma_cookie_complete(&desc->async_tx);
+
+		dev_dbg(chan->dev, "completed_cookie=%d\n", cookie);
+	}
+
+	/*
+	 * move the descriptors to a temporary list so we can drop the lock
+	 * during the entire cleanup operation
+	 */
+	list_splice_tail_init(&chan->chain_running, &chain_cleanup);
+
+	/* the hardware is now idle and ready for more */
+	chan->idle = true;
+
+	/* Start any pending transactions automatically */
+	start_pending_queue(chan);
+	spin_unlock_irqrestore(&chan->desc_lock, flags);
+
+	/* Run the callback for each descriptor, in order */
+	list_for_each_entry_safe(desc, _desc, &chain_cleanup, node) {
+		struct dma_async_tx_descriptor *txd = &desc->async_tx;
+
+		/* Remove from the list of transactions */
+		list_del(&desc->node);
+		/* Run the link descriptor callback function */
+		if (txd->callback)
+			txd->callback(txd->callback_param);
+
+		dma_pool_free(chan->desc_pool, desc, txd->phys);
+	}
+}
+
+static int __devexit mmp_pdma_remove(struct platform_device *op)
+{
+	struct mmp_pdma_device *pdev = platform_get_drvdata(op);
+
+	dma_async_device_unregister(&pdev->device);
+	return 0;
+}
+
+static int __devinit mmp_pdma_chan_init(struct mmp_pdma_device *pdev,
+							int idx, int irq)
+{
+	struct mmp_pdma_phy *phy  = &pdev->phy[idx];
+	struct mmp_pdma_chan *chan;
+	int ret;
+
+	chan = devm_kzalloc(pdev->dev,
+			sizeof(struct mmp_pdma_chan), GFP_KERNEL);
+	if (chan == NULL)
+		return -ENOMEM;
+
+	phy->idx = idx;
+	phy->base = pdev->base;
+
+	if (irq) {
+		ret = devm_request_irq(pdev->dev, irq,
+			mmp_pdma_chan_handler, IRQF_DISABLED, "pdma", phy);
+		if (ret) {
+			dev_err(pdev->dev, "channel request irq fail!\n");
+			return ret;
+		}
+	}
+
+	spin_lock_init(&chan->desc_lock);
+	chan->dev = pdev->dev;
+	chan->chan.device = &pdev->device;
+	tasklet_init(&chan->tasklet, dma_do_tasklet, (unsigned long)chan);
+	INIT_LIST_HEAD(&chan->chain_pending);
+	INIT_LIST_HEAD(&chan->chain_running);
+
+	/* register virt channel to dma engine */
+	list_add_tail(&chan->chan.device_node,
+			&pdev->device.channels);
+
+	return 0;
+}
+
+static struct of_device_id mmp_pdma_dt_ids[] = {
+	{ .compatible = "marvell,pdma-1.0", },
+	{}
+};
+MODULE_DEVICE_TABLE(of, mmp_pdma_dt_ids);
+
+static int __devinit mmp_pdma_probe(struct platform_device *op)
+{
+	struct mmp_pdma_device *pdev;
+	const struct of_device_id *of_id;
+	struct mmp_dma_platdata *pdata = dev_get_platdata(&op->dev);
+	struct resource *iores;
+	int i, ret, irq = 0;
+	int dma_channels = 0, irq_num = 0;
+
+	pdev = devm_kzalloc(&op->dev, sizeof(*pdev), GFP_KERNEL);
+	if (!pdev)
+		return -ENOMEM;
+	pdev->dev = &op->dev;
+
+	iores = platform_get_resource(op, IORESOURCE_MEM, 0);
+	if (!iores)
+		return -EINVAL;
+
+	pdev->base = devm_request_and_ioremap(pdev->dev, iores);
+	if (!pdev->base)
+		return -EADDRNOTAVAIL;
+
+	of_id = of_match_device(mmp_pdma_dt_ids, pdev->dev);
+	if (of_id)
+		of_property_read_u32(pdev->dev->of_node,
+				"#dma-channels", &dma_channels);
+	else if (pdata && pdata->dma_channels)
+		dma_channels = pdata->dma_channels;
+	else
+		dma_channels = 32;	/* default 32 channel */
+	pdev->dma_channels = dma_channels;
+
+	for (i = 0; i < dma_channels; i++) {
+		if (platform_get_irq(op, i) > 0)
+			irq_num++;
+	}
+
+	pdev->phy = devm_kzalloc(pdev->dev,
+		dma_channels * sizeof(struct mmp_pdma_chan), GFP_KERNEL);
+	if (pdev->phy == NULL)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&pdev->device.channels);
+
+	if (irq_num != dma_channels) {
+		/* all chan share one irq, demux inside */
+		irq = platform_get_irq(op, 0);
+		ret = devm_request_irq(pdev->dev, irq,
+			mmp_pdma_int_handler, IRQF_DISABLED, "pdma", pdev);
+		if (ret)
+			return ret;
+	}
+
+	for (i = 0; i < dma_channels; i++) {
+		irq = (irq_num != dma_channels) ? 0 : platform_get_irq(op, i);
+		ret = mmp_pdma_chan_init(pdev, i, irq);
+		if (ret)
+			return ret;
+	}
+
+	dma_cap_set(DMA_SLAVE, pdev->device.cap_mask);
+	dma_cap_set(DMA_MEMCPY, pdev->device.cap_mask);
+	dma_cap_set(DMA_SLAVE, pdev->device.cap_mask);
+	pdev->device.dev = &op->dev;
+	pdev->device.device_alloc_chan_resources = mmp_pdma_alloc_chan_resources;
+	pdev->device.device_free_chan_resources = mmp_pdma_free_chan_resources;
+	pdev->device.device_tx_status = mmp_pdma_tx_status;
+	pdev->device.device_prep_dma_memcpy = mmp_pdma_prep_memcpy;
+	pdev->device.device_prep_slave_sg = mmp_pdma_prep_slave_sg;
+	pdev->device.device_issue_pending = mmp_pdma_issue_pending;
+	pdev->device.device_control = mmp_pdma_control;
+	pdev->device.copy_align = PDMA_ALIGNMENT;
+
+	if (pdev->dev->coherent_dma_mask)
+		dma_set_mask(pdev->dev, pdev->dev->coherent_dma_mask);
+	else
+		dma_set_mask(pdev->dev, DMA_BIT_MASK(64));
+
+	ret = dma_async_device_register(&pdev->device);
+	if (ret) {
+		dev_err(pdev->device.dev, "unable to register\n");
+		return ret;
+	}
+
+	dev_info(pdev->device.dev, "initialized\n");
+	return 0;
+}
+
+static const struct platform_device_id mmp_pdma_id_table[] = {
+	{ "mmp-pdma", },
+	{ },
+};
+
+static struct platform_driver mmp_pdma_driver = {
+	.driver		= {
+		.name	= "mmp-pdma",
+		.owner  = THIS_MODULE,
+		.of_match_table = mmp_pdma_dt_ids,
+	},
+	.id_table	= mmp_pdma_id_table,
+	.probe		= mmp_pdma_probe,
+	.remove		= __devexit_p(mmp_pdma_remove),
+};
+
+module_platform_driver(mmp_pdma_driver);
+
+MODULE_DESCRIPTION("MARVELL MMP Periphera DMA Driver");
+MODULE_AUTHOR("Marvell International Ltd.");
+MODULE_LICENSE("GPL v2");