Add the rt linux 4.1.3-rt3 as base
[kvmfornfv.git] / kernel / drivers / dma / mpc512x_dma.c
diff --git a/kernel/drivers/dma/mpc512x_dma.c b/kernel/drivers/dma/mpc512x_dma.c
new file mode 100644 (file)
index 0000000..e6281e7
--- /dev/null
@@ -0,0 +1,1089 @@
+/*
+ * Copyright (C) Freescale Semicondutor, Inc. 2007, 2008.
+ * Copyright (C) Semihalf 2009
+ * Copyright (C) Ilya Yanok, Emcraft Systems 2010
+ * Copyright (C) Alexander Popov, Promcontroller 2014
+ *
+ * Written by Piotr Ziecik <kosmo@semihalf.com>. Hardware description
+ * (defines, structures and comments) was taken from MPC5121 DMA driver
+ * written by Hongjun Chen <hong-jun.chen@freescale.com>.
+ *
+ * Approved as OSADL project by a majority of OSADL members and funded
+ * by OSADL membership fees in 2009;  for details see www.osadl.org.
+ *
+ * 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.
+ *
+ * The full GNU General Public License is included in this distribution in the
+ * file called COPYING.
+ */
+
+/*
+ * MPC512x and MPC8308 DMA driver. It supports
+ * memory to memory data transfers (tested using dmatest module) and
+ * data transfers between memory and peripheral I/O memory
+ * by means of slave scatter/gather with these limitations:
+ *  - chunked transfers (described by s/g lists with more than one item)
+ *     are refused as long as proper support for scatter/gather is missing;
+ *  - transfers on MPC8308 always start from software as this SoC appears
+ *     not to have external request lines for peripheral flow control;
+ *  - only peripheral devices with 4-byte FIFO access register are supported;
+ *  - minimal memory <-> I/O memory transfer chunk is 4 bytes and consequently
+ *     source and destination addresses must be 4-byte aligned
+ *     and transfer size must be aligned on (4 * maxburst) boundary;
+ */
+
+#include <linux/module.h>
+#include <linux/dmaengine.h>
+#include <linux/dma-mapping.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/slab.h>
+#include <linux/of_address.h>
+#include <linux/of_device.h>
+#include <linux/of_irq.h>
+#include <linux/of_dma.h>
+#include <linux/of_platform.h>
+
+#include <linux/random.h>
+
+#include "dmaengine.h"
+
+/* Number of DMA Transfer descriptors allocated per channel */
+#define MPC_DMA_DESCRIPTORS    64
+
+/* Macro definitions */
+#define MPC_DMA_TCD_OFFSET     0x1000
+
+/*
+ * Maximum channel counts for individual hardware variants
+ * and the maximum channel count over all supported controllers,
+ * used for data structure size
+ */
+#define MPC8308_DMACHAN_MAX    16
+#define MPC512x_DMACHAN_MAX    64
+#define MPC_DMA_CHANNELS       64
+
+/* Arbitration mode of group and channel */
+#define MPC_DMA_DMACR_EDCG     (1 << 31)
+#define MPC_DMA_DMACR_ERGA     (1 << 3)
+#define MPC_DMA_DMACR_ERCA     (1 << 2)
+
+/* Error codes */
+#define MPC_DMA_DMAES_VLD      (1 << 31)
+#define MPC_DMA_DMAES_GPE      (1 << 15)
+#define MPC_DMA_DMAES_CPE      (1 << 14)
+#define MPC_DMA_DMAES_ERRCHN(err) \
+                               (((err) >> 8) & 0x3f)
+#define MPC_DMA_DMAES_SAE      (1 << 7)
+#define MPC_DMA_DMAES_SOE      (1 << 6)
+#define MPC_DMA_DMAES_DAE      (1 << 5)
+#define MPC_DMA_DMAES_DOE      (1 << 4)
+#define MPC_DMA_DMAES_NCE      (1 << 3)
+#define MPC_DMA_DMAES_SGE      (1 << 2)
+#define MPC_DMA_DMAES_SBE      (1 << 1)
+#define MPC_DMA_DMAES_DBE      (1 << 0)
+
+#define MPC_DMA_DMAGPOR_SNOOP_ENABLE   (1 << 6)
+
+#define MPC_DMA_TSIZE_1                0x00
+#define MPC_DMA_TSIZE_2                0x01
+#define MPC_DMA_TSIZE_4                0x02
+#define MPC_DMA_TSIZE_16       0x04
+#define MPC_DMA_TSIZE_32       0x05
+
+/* MPC5121 DMA engine registers */
+struct __attribute__ ((__packed__)) mpc_dma_regs {
+       /* 0x00 */
+       u32 dmacr;              /* DMA control register */
+       u32 dmaes;              /* DMA error status */
+       /* 0x08 */
+       u32 dmaerqh;            /* DMA enable request high(channels 63~32) */
+       u32 dmaerql;            /* DMA enable request low(channels 31~0) */
+       u32 dmaeeih;            /* DMA enable error interrupt high(ch63~32) */
+       u32 dmaeeil;            /* DMA enable error interrupt low(ch31~0) */
+       /* 0x18 */
+       u8 dmaserq;             /* DMA set enable request */
+       u8 dmacerq;             /* DMA clear enable request */
+       u8 dmaseei;             /* DMA set enable error interrupt */
+       u8 dmaceei;             /* DMA clear enable error interrupt */
+       /* 0x1c */
+       u8 dmacint;             /* DMA clear interrupt request */
+       u8 dmacerr;             /* DMA clear error */
+       u8 dmassrt;             /* DMA set start bit */
+       u8 dmacdne;             /* DMA clear DONE status bit */
+       /* 0x20 */
+       u32 dmainth;            /* DMA interrupt request high(ch63~32) */
+       u32 dmaintl;            /* DMA interrupt request low(ch31~0) */
+       u32 dmaerrh;            /* DMA error high(ch63~32) */
+       u32 dmaerrl;            /* DMA error low(ch31~0) */
+       /* 0x30 */
+       u32 dmahrsh;            /* DMA hw request status high(ch63~32) */
+       u32 dmahrsl;            /* DMA hardware request status low(ch31~0) */
+       union {
+               u32 dmaihsa;    /* DMA interrupt high select AXE(ch63~32) */
+               u32 dmagpor;    /* (General purpose register on MPC8308) */
+       };
+       u32 dmailsa;            /* DMA interrupt low select AXE(ch31~0) */
+       /* 0x40 ~ 0xff */
+       u32 reserve0[48];       /* Reserved */
+       /* 0x100 */
+       u8 dchpri[MPC_DMA_CHANNELS];
+       /* DMA channels(0~63) priority */
+};
+
+struct __attribute__ ((__packed__)) mpc_dma_tcd {
+       /* 0x00 */
+       u32 saddr;              /* Source address */
+
+       u32 smod:5;             /* Source address modulo */
+       u32 ssize:3;            /* Source data transfer size */
+       u32 dmod:5;             /* Destination address modulo */
+       u32 dsize:3;            /* Destination data transfer size */
+       u32 soff:16;            /* Signed source address offset */
+
+       /* 0x08 */
+       u32 nbytes;             /* Inner "minor" byte count */
+       u32 slast;              /* Last source address adjustment */
+       u32 daddr;              /* Destination address */
+
+       /* 0x14 */
+       u32 citer_elink:1;      /* Enable channel-to-channel linking on
+                                * minor loop complete
+                                */
+       u32 citer_linkch:6;     /* Link channel for minor loop complete */
+       u32 citer:9;            /* Current "major" iteration count */
+       u32 doff:16;            /* Signed destination address offset */
+
+       /* 0x18 */
+       u32 dlast_sga;          /* Last Destination address adjustment/scatter
+                                * gather address
+                                */
+
+       /* 0x1c */
+       u32 biter_elink:1;      /* Enable channel-to-channel linking on major
+                                * loop complete
+                                */
+       u32 biter_linkch:6;
+       u32 biter:9;            /* Beginning "major" iteration count */
+       u32 bwc:2;              /* Bandwidth control */
+       u32 major_linkch:6;     /* Link channel number */
+       u32 done:1;             /* Channel done */
+       u32 active:1;           /* Channel active */
+       u32 major_elink:1;      /* Enable channel-to-channel linking on major
+                                * loop complete
+                                */
+       u32 e_sg:1;             /* Enable scatter/gather processing */
+       u32 d_req:1;            /* Disable request */
+       u32 int_half:1;         /* Enable an interrupt when major counter is
+                                * half complete
+                                */
+       u32 int_maj:1;          /* Enable an interrupt when major iteration
+                                * count completes
+                                */
+       u32 start:1;            /* Channel start */
+};
+
+struct mpc_dma_desc {
+       struct dma_async_tx_descriptor  desc;
+       struct mpc_dma_tcd              *tcd;
+       dma_addr_t                      tcd_paddr;
+       int                             error;
+       struct list_head                node;
+       int                             will_access_peripheral;
+};
+
+struct mpc_dma_chan {
+       struct dma_chan                 chan;
+       struct list_head                free;
+       struct list_head                prepared;
+       struct list_head                queued;
+       struct list_head                active;
+       struct list_head                completed;
+       struct mpc_dma_tcd              *tcd;
+       dma_addr_t                      tcd_paddr;
+
+       /* Settings for access to peripheral FIFO */
+       dma_addr_t                      src_per_paddr;
+       u32                             src_tcd_nunits;
+       dma_addr_t                      dst_per_paddr;
+       u32                             dst_tcd_nunits;
+
+       /* Lock for this structure */
+       spinlock_t                      lock;
+};
+
+struct mpc_dma {
+       struct dma_device               dma;
+       struct tasklet_struct           tasklet;
+       struct mpc_dma_chan             channels[MPC_DMA_CHANNELS];
+       struct mpc_dma_regs __iomem     *regs;
+       struct mpc_dma_tcd __iomem      *tcd;
+       int                             irq;
+       int                             irq2;
+       uint                            error_status;
+       int                             is_mpc8308;
+
+       /* Lock for error_status field in this structure */
+       spinlock_t                      error_status_lock;
+};
+
+#define DRV_NAME       "mpc512x_dma"
+
+/* Convert struct dma_chan to struct mpc_dma_chan */
+static inline struct mpc_dma_chan *dma_chan_to_mpc_dma_chan(struct dma_chan *c)
+{
+       return container_of(c, struct mpc_dma_chan, chan);
+}
+
+/* Convert struct dma_chan to struct mpc_dma */
+static inline struct mpc_dma *dma_chan_to_mpc_dma(struct dma_chan *c)
+{
+       struct mpc_dma_chan *mchan = dma_chan_to_mpc_dma_chan(c);
+       return container_of(mchan, struct mpc_dma, channels[c->chan_id]);
+}
+
+/*
+ * Execute all queued DMA descriptors.
+ *
+ * Following requirements must be met while calling mpc_dma_execute():
+ *     a) mchan->lock is acquired,
+ *     b) mchan->active list is empty,
+ *     c) mchan->queued list contains at least one entry.
+ */
+static void mpc_dma_execute(struct mpc_dma_chan *mchan)
+{
+       struct mpc_dma *mdma = dma_chan_to_mpc_dma(&mchan->chan);
+       struct mpc_dma_desc *first = NULL;
+       struct mpc_dma_desc *prev = NULL;
+       struct mpc_dma_desc *mdesc;
+       int cid = mchan->chan.chan_id;
+
+       while (!list_empty(&mchan->queued)) {
+               mdesc = list_first_entry(&mchan->queued,
+                                               struct mpc_dma_desc, node);
+               /*
+                * Grab either several mem-to-mem transfer descriptors
+                * or one peripheral transfer descriptor,
+                * don't mix mem-to-mem and peripheral transfer descriptors
+                * within the same 'active' list.
+                */
+               if (mdesc->will_access_peripheral) {
+                       if (list_empty(&mchan->active))
+                               list_move_tail(&mdesc->node, &mchan->active);
+                       break;
+               } else {
+                       list_move_tail(&mdesc->node, &mchan->active);
+               }
+       }
+
+       /* Chain descriptors into one transaction */
+       list_for_each_entry(mdesc, &mchan->active, node) {
+               if (!first)
+                       first = mdesc;
+
+               if (!prev) {
+                       prev = mdesc;
+                       continue;
+               }
+
+               prev->tcd->dlast_sga = mdesc->tcd_paddr;
+               prev->tcd->e_sg = 1;
+               mdesc->tcd->start = 1;
+
+               prev = mdesc;
+       }
+
+       prev->tcd->int_maj = 1;
+
+       /* Send first descriptor in chain into hardware */
+       memcpy_toio(&mdma->tcd[cid], first->tcd, sizeof(struct mpc_dma_tcd));
+
+       if (first != prev)
+               mdma->tcd[cid].e_sg = 1;
+
+       if (mdma->is_mpc8308) {
+               /* MPC8308, no request lines, software initiated start */
+               out_8(&mdma->regs->dmassrt, cid);
+       } else if (first->will_access_peripheral) {
+               /* Peripherals involved, start by external request signal */
+               out_8(&mdma->regs->dmaserq, cid);
+       } else {
+               /* Memory to memory transfer, software initiated start */
+               out_8(&mdma->regs->dmassrt, cid);
+       }
+}
+
+/* Handle interrupt on one half of DMA controller (32 channels) */
+static void mpc_dma_irq_process(struct mpc_dma *mdma, u32 is, u32 es, int off)
+{
+       struct mpc_dma_chan *mchan;
+       struct mpc_dma_desc *mdesc;
+       u32 status = is | es;
+       int ch;
+
+       while ((ch = fls(status) - 1) >= 0) {
+               status &= ~(1 << ch);
+               mchan = &mdma->channels[ch + off];
+
+               spin_lock(&mchan->lock);
+
+               out_8(&mdma->regs->dmacint, ch + off);
+               out_8(&mdma->regs->dmacerr, ch + off);
+
+               /* Check error status */
+               if (es & (1 << ch))
+                       list_for_each_entry(mdesc, &mchan->active, node)
+                               mdesc->error = -EIO;
+
+               /* Execute queued descriptors */
+               list_splice_tail_init(&mchan->active, &mchan->completed);
+               if (!list_empty(&mchan->queued))
+                       mpc_dma_execute(mchan);
+
+               spin_unlock(&mchan->lock);
+       }
+}
+
+/* Interrupt handler */
+static irqreturn_t mpc_dma_irq(int irq, void *data)
+{
+       struct mpc_dma *mdma = data;
+       uint es;
+
+       /* Save error status register */
+       es = in_be32(&mdma->regs->dmaes);
+       spin_lock(&mdma->error_status_lock);
+       if ((es & MPC_DMA_DMAES_VLD) && mdma->error_status == 0)
+               mdma->error_status = es;
+       spin_unlock(&mdma->error_status_lock);
+
+       /* Handle interrupt on each channel */
+       if (mdma->dma.chancnt > 32) {
+               mpc_dma_irq_process(mdma, in_be32(&mdma->regs->dmainth),
+                                       in_be32(&mdma->regs->dmaerrh), 32);
+       }
+       mpc_dma_irq_process(mdma, in_be32(&mdma->regs->dmaintl),
+                                       in_be32(&mdma->regs->dmaerrl), 0);
+
+       /* Schedule tasklet */
+       tasklet_schedule(&mdma->tasklet);
+
+       return IRQ_HANDLED;
+}
+
+/* process completed descriptors */
+static void mpc_dma_process_completed(struct mpc_dma *mdma)
+{
+       dma_cookie_t last_cookie = 0;
+       struct mpc_dma_chan *mchan;
+       struct mpc_dma_desc *mdesc;
+       struct dma_async_tx_descriptor *desc;
+       unsigned long flags;
+       LIST_HEAD(list);
+       int i;
+
+       for (i = 0; i < mdma->dma.chancnt; i++) {
+               mchan = &mdma->channels[i];
+
+               /* Get all completed descriptors */
+               spin_lock_irqsave(&mchan->lock, flags);
+               if (!list_empty(&mchan->completed))
+                       list_splice_tail_init(&mchan->completed, &list);
+               spin_unlock_irqrestore(&mchan->lock, flags);
+
+               if (list_empty(&list))
+                       continue;
+
+               /* Execute callbacks and run dependencies */
+               list_for_each_entry(mdesc, &list, node) {
+                       desc = &mdesc->desc;
+
+                       if (desc->callback)
+                               desc->callback(desc->callback_param);
+
+                       last_cookie = desc->cookie;
+                       dma_run_dependencies(desc);
+               }
+
+               /* Free descriptors */
+               spin_lock_irqsave(&mchan->lock, flags);
+               list_splice_tail_init(&list, &mchan->free);
+               mchan->chan.completed_cookie = last_cookie;
+               spin_unlock_irqrestore(&mchan->lock, flags);
+       }
+}
+
+/* DMA Tasklet */
+static void mpc_dma_tasklet(unsigned long data)
+{
+       struct mpc_dma *mdma = (void *)data;
+       unsigned long flags;
+       uint es;
+
+       spin_lock_irqsave(&mdma->error_status_lock, flags);
+       es = mdma->error_status;
+       mdma->error_status = 0;
+       spin_unlock_irqrestore(&mdma->error_status_lock, flags);
+
+       /* Print nice error report */
+       if (es) {
+               dev_err(mdma->dma.dev,
+                       "Hardware reported following error(s) on channel %u:\n",
+                                                     MPC_DMA_DMAES_ERRCHN(es));
+
+               if (es & MPC_DMA_DMAES_GPE)
+                       dev_err(mdma->dma.dev, "- Group Priority Error\n");
+               if (es & MPC_DMA_DMAES_CPE)
+                       dev_err(mdma->dma.dev, "- Channel Priority Error\n");
+               if (es & MPC_DMA_DMAES_SAE)
+                       dev_err(mdma->dma.dev, "- Source Address Error\n");
+               if (es & MPC_DMA_DMAES_SOE)
+                       dev_err(mdma->dma.dev, "- Source Offset"
+                                               " Configuration Error\n");
+               if (es & MPC_DMA_DMAES_DAE)
+                       dev_err(mdma->dma.dev, "- Destination Address"
+                                                               " Error\n");
+               if (es & MPC_DMA_DMAES_DOE)
+                       dev_err(mdma->dma.dev, "- Destination Offset"
+                                               " Configuration Error\n");
+               if (es & MPC_DMA_DMAES_NCE)
+                       dev_err(mdma->dma.dev, "- NBytes/Citter"
+                                               " Configuration Error\n");
+               if (es & MPC_DMA_DMAES_SGE)
+                       dev_err(mdma->dma.dev, "- Scatter/Gather"
+                                               " Configuration Error\n");
+               if (es & MPC_DMA_DMAES_SBE)
+                       dev_err(mdma->dma.dev, "- Source Bus Error\n");
+               if (es & MPC_DMA_DMAES_DBE)
+                       dev_err(mdma->dma.dev, "- Destination Bus Error\n");
+       }
+
+       mpc_dma_process_completed(mdma);
+}
+
+/* Submit descriptor to hardware */
+static dma_cookie_t mpc_dma_tx_submit(struct dma_async_tx_descriptor *txd)
+{
+       struct mpc_dma_chan *mchan = dma_chan_to_mpc_dma_chan(txd->chan);
+       struct mpc_dma_desc *mdesc;
+       unsigned long flags;
+       dma_cookie_t cookie;
+
+       mdesc = container_of(txd, struct mpc_dma_desc, desc);
+
+       spin_lock_irqsave(&mchan->lock, flags);
+
+       /* Move descriptor to queue */
+       list_move_tail(&mdesc->node, &mchan->queued);
+
+       /* If channel is idle, execute all queued descriptors */
+       if (list_empty(&mchan->active))
+               mpc_dma_execute(mchan);
+
+       /* Update cookie */
+       cookie = dma_cookie_assign(txd);
+       spin_unlock_irqrestore(&mchan->lock, flags);
+
+       return cookie;
+}
+
+/* Alloc channel resources */
+static int mpc_dma_alloc_chan_resources(struct dma_chan *chan)
+{
+       struct mpc_dma *mdma = dma_chan_to_mpc_dma(chan);
+       struct mpc_dma_chan *mchan = dma_chan_to_mpc_dma_chan(chan);
+       struct mpc_dma_desc *mdesc;
+       struct mpc_dma_tcd *tcd;
+       dma_addr_t tcd_paddr;
+       unsigned long flags;
+       LIST_HEAD(descs);
+       int i;
+
+       /* Alloc DMA memory for Transfer Control Descriptors */
+       tcd = dma_alloc_coherent(mdma->dma.dev,
+                       MPC_DMA_DESCRIPTORS * sizeof(struct mpc_dma_tcd),
+                                                       &tcd_paddr, GFP_KERNEL);
+       if (!tcd)
+               return -ENOMEM;
+
+       /* Alloc descriptors for this channel */
+       for (i = 0; i < MPC_DMA_DESCRIPTORS; i++) {
+               mdesc = kzalloc(sizeof(struct mpc_dma_desc), GFP_KERNEL);
+               if (!mdesc) {
+                       dev_notice(mdma->dma.dev, "Memory allocation error. "
+                                       "Allocated only %u descriptors\n", i);
+                       break;
+               }
+
+               dma_async_tx_descriptor_init(&mdesc->desc, chan);
+               mdesc->desc.flags = DMA_CTRL_ACK;
+               mdesc->desc.tx_submit = mpc_dma_tx_submit;
+
+               mdesc->tcd = &tcd[i];
+               mdesc->tcd_paddr = tcd_paddr + (i * sizeof(struct mpc_dma_tcd));
+
+               list_add_tail(&mdesc->node, &descs);
+       }
+
+       /* Return error only if no descriptors were allocated */
+       if (i == 0) {
+               dma_free_coherent(mdma->dma.dev,
+                       MPC_DMA_DESCRIPTORS * sizeof(struct mpc_dma_tcd),
+                                                               tcd, tcd_paddr);
+               return -ENOMEM;
+       }
+
+       spin_lock_irqsave(&mchan->lock, flags);
+       mchan->tcd = tcd;
+       mchan->tcd_paddr = tcd_paddr;
+       list_splice_tail_init(&descs, &mchan->free);
+       spin_unlock_irqrestore(&mchan->lock, flags);
+
+       /* Enable Error Interrupt */
+       out_8(&mdma->regs->dmaseei, chan->chan_id);
+
+       return 0;
+}
+
+/* Free channel resources */
+static void mpc_dma_free_chan_resources(struct dma_chan *chan)
+{
+       struct mpc_dma *mdma = dma_chan_to_mpc_dma(chan);
+       struct mpc_dma_chan *mchan = dma_chan_to_mpc_dma_chan(chan);
+       struct mpc_dma_desc *mdesc, *tmp;
+       struct mpc_dma_tcd *tcd;
+       dma_addr_t tcd_paddr;
+       unsigned long flags;
+       LIST_HEAD(descs);
+
+       spin_lock_irqsave(&mchan->lock, flags);
+
+       /* Channel must be idle */
+       BUG_ON(!list_empty(&mchan->prepared));
+       BUG_ON(!list_empty(&mchan->queued));
+       BUG_ON(!list_empty(&mchan->active));
+       BUG_ON(!list_empty(&mchan->completed));
+
+       /* Move data */
+       list_splice_tail_init(&mchan->free, &descs);
+       tcd = mchan->tcd;
+       tcd_paddr = mchan->tcd_paddr;
+
+       spin_unlock_irqrestore(&mchan->lock, flags);
+
+       /* Free DMA memory used by descriptors */
+       dma_free_coherent(mdma->dma.dev,
+                       MPC_DMA_DESCRIPTORS * sizeof(struct mpc_dma_tcd),
+                                                               tcd, tcd_paddr);
+
+       /* Free descriptors */
+       list_for_each_entry_safe(mdesc, tmp, &descs, node)
+               kfree(mdesc);
+
+       /* Disable Error Interrupt */
+       out_8(&mdma->regs->dmaceei, chan->chan_id);
+}
+
+/* Send all pending descriptor to hardware */
+static void mpc_dma_issue_pending(struct dma_chan *chan)
+{
+       /*
+        * We are posting descriptors to the hardware as soon as
+        * they are ready, so this function does nothing.
+        */
+}
+
+/* Check request completion status */
+static enum dma_status
+mpc_dma_tx_status(struct dma_chan *chan, dma_cookie_t cookie,
+              struct dma_tx_state *txstate)
+{
+       return dma_cookie_status(chan, cookie, txstate);
+}
+
+/* Prepare descriptor for memory to memory copy */
+static struct dma_async_tx_descriptor *
+mpc_dma_prep_memcpy(struct dma_chan *chan, dma_addr_t dst, dma_addr_t src,
+                                       size_t len, unsigned long flags)
+{
+       struct mpc_dma *mdma = dma_chan_to_mpc_dma(chan);
+       struct mpc_dma_chan *mchan = dma_chan_to_mpc_dma_chan(chan);
+       struct mpc_dma_desc *mdesc = NULL;
+       struct mpc_dma_tcd *tcd;
+       unsigned long iflags;
+
+       /* Get free descriptor */
+       spin_lock_irqsave(&mchan->lock, iflags);
+       if (!list_empty(&mchan->free)) {
+               mdesc = list_first_entry(&mchan->free, struct mpc_dma_desc,
+                                                                       node);
+               list_del(&mdesc->node);
+       }
+       spin_unlock_irqrestore(&mchan->lock, iflags);
+
+       if (!mdesc) {
+               /* try to free completed descriptors */
+               mpc_dma_process_completed(mdma);
+               return NULL;
+       }
+
+       mdesc->error = 0;
+       mdesc->will_access_peripheral = 0;
+       tcd = mdesc->tcd;
+
+       /* Prepare Transfer Control Descriptor for this transaction */
+       memset(tcd, 0, sizeof(struct mpc_dma_tcd));
+
+       if (IS_ALIGNED(src | dst | len, 32)) {
+               tcd->ssize = MPC_DMA_TSIZE_32;
+               tcd->dsize = MPC_DMA_TSIZE_32;
+               tcd->soff = 32;
+               tcd->doff = 32;
+       } else if (!mdma->is_mpc8308 && IS_ALIGNED(src | dst | len, 16)) {
+               /* MPC8308 doesn't support 16 byte transfers */
+               tcd->ssize = MPC_DMA_TSIZE_16;
+               tcd->dsize = MPC_DMA_TSIZE_16;
+               tcd->soff = 16;
+               tcd->doff = 16;
+       } else if (IS_ALIGNED(src | dst | len, 4)) {
+               tcd->ssize = MPC_DMA_TSIZE_4;
+               tcd->dsize = MPC_DMA_TSIZE_4;
+               tcd->soff = 4;
+               tcd->doff = 4;
+       } else if (IS_ALIGNED(src | dst | len, 2)) {
+               tcd->ssize = MPC_DMA_TSIZE_2;
+               tcd->dsize = MPC_DMA_TSIZE_2;
+               tcd->soff = 2;
+               tcd->doff = 2;
+       } else {
+               tcd->ssize = MPC_DMA_TSIZE_1;
+               tcd->dsize = MPC_DMA_TSIZE_1;
+               tcd->soff = 1;
+               tcd->doff = 1;
+       }
+
+       tcd->saddr = src;
+       tcd->daddr = dst;
+       tcd->nbytes = len;
+       tcd->biter = 1;
+       tcd->citer = 1;
+
+       /* Place descriptor in prepared list */
+       spin_lock_irqsave(&mchan->lock, iflags);
+       list_add_tail(&mdesc->node, &mchan->prepared);
+       spin_unlock_irqrestore(&mchan->lock, iflags);
+
+       return &mdesc->desc;
+}
+
+static struct dma_async_tx_descriptor *
+mpc_dma_prep_slave_sg(struct dma_chan *chan, struct scatterlist *sgl,
+               unsigned int sg_len, enum dma_transfer_direction direction,
+               unsigned long flags, void *context)
+{
+       struct mpc_dma *mdma = dma_chan_to_mpc_dma(chan);
+       struct mpc_dma_chan *mchan = dma_chan_to_mpc_dma_chan(chan);
+       struct mpc_dma_desc *mdesc = NULL;
+       dma_addr_t per_paddr;
+       u32 tcd_nunits;
+       struct mpc_dma_tcd *tcd;
+       unsigned long iflags;
+       struct scatterlist *sg;
+       size_t len;
+       int iter, i;
+
+       /* Currently there is no proper support for scatter/gather */
+       if (sg_len != 1)
+               return NULL;
+
+       if (!is_slave_direction(direction))
+               return NULL;
+
+       for_each_sg(sgl, sg, sg_len, i) {
+               spin_lock_irqsave(&mchan->lock, iflags);
+
+               mdesc = list_first_entry(&mchan->free,
+                                               struct mpc_dma_desc, node);
+               if (!mdesc) {
+                       spin_unlock_irqrestore(&mchan->lock, iflags);
+                       /* Try to free completed descriptors */
+                       mpc_dma_process_completed(mdma);
+                       return NULL;
+               }
+
+               list_del(&mdesc->node);
+
+               if (direction == DMA_DEV_TO_MEM) {
+                       per_paddr = mchan->src_per_paddr;
+                       tcd_nunits = mchan->src_tcd_nunits;
+               } else {
+                       per_paddr = mchan->dst_per_paddr;
+                       tcd_nunits = mchan->dst_tcd_nunits;
+               }
+
+               spin_unlock_irqrestore(&mchan->lock, iflags);
+
+               if (per_paddr == 0 || tcd_nunits == 0)
+                       goto err_prep;
+
+               mdesc->error = 0;
+               mdesc->will_access_peripheral = 1;
+
+               /* Prepare Transfer Control Descriptor for this transaction */
+               tcd = mdesc->tcd;
+
+               memset(tcd, 0, sizeof(struct mpc_dma_tcd));
+
+               if (!IS_ALIGNED(sg_dma_address(sg), 4))
+                       goto err_prep;
+
+               if (direction == DMA_DEV_TO_MEM) {
+                       tcd->saddr = per_paddr;
+                       tcd->daddr = sg_dma_address(sg);
+                       tcd->soff = 0;
+                       tcd->doff = 4;
+               } else {
+                       tcd->saddr = sg_dma_address(sg);
+                       tcd->daddr = per_paddr;
+                       tcd->soff = 4;
+                       tcd->doff = 0;
+               }
+
+               tcd->ssize = MPC_DMA_TSIZE_4;
+               tcd->dsize = MPC_DMA_TSIZE_4;
+
+               len = sg_dma_len(sg);
+               tcd->nbytes = tcd_nunits * 4;
+               if (!IS_ALIGNED(len, tcd->nbytes))
+                       goto err_prep;
+
+               iter = len / tcd->nbytes;
+               if (iter >= 1 << 15) {
+                       /* len is too big */
+                       goto err_prep;
+               }
+               /* citer_linkch contains the high bits of iter */
+               tcd->biter = iter & 0x1ff;
+               tcd->biter_linkch = iter >> 9;
+               tcd->citer = tcd->biter;
+               tcd->citer_linkch = tcd->biter_linkch;
+
+               tcd->e_sg = 0;
+               tcd->d_req = 1;
+
+               /* Place descriptor in prepared list */
+               spin_lock_irqsave(&mchan->lock, iflags);
+               list_add_tail(&mdesc->node, &mchan->prepared);
+               spin_unlock_irqrestore(&mchan->lock, iflags);
+       }
+
+       return &mdesc->desc;
+
+err_prep:
+       /* Put the descriptor back */
+       spin_lock_irqsave(&mchan->lock, iflags);
+       list_add_tail(&mdesc->node, &mchan->free);
+       spin_unlock_irqrestore(&mchan->lock, iflags);
+
+       return NULL;
+}
+
+static int mpc_dma_device_config(struct dma_chan *chan,
+                                struct dma_slave_config *cfg)
+{
+       struct mpc_dma_chan *mchan = dma_chan_to_mpc_dma_chan(chan);
+       unsigned long flags;
+
+       /*
+        * Software constraints:
+        *  - only transfers between a peripheral device and
+        *     memory are supported;
+        *  - only peripheral devices with 4-byte FIFO access register
+        *     are supported;
+        *  - minimal transfer chunk is 4 bytes and consequently
+        *     source and destination addresses must be 4-byte aligned
+        *     and transfer size must be aligned on (4 * maxburst)
+        *     boundary;
+        *  - during the transfer RAM address is being incremented by
+        *     the size of minimal transfer chunk;
+        *  - peripheral port's address is constant during the transfer.
+        */
+
+       if (cfg->src_addr_width != DMA_SLAVE_BUSWIDTH_4_BYTES ||
+           cfg->dst_addr_width != DMA_SLAVE_BUSWIDTH_4_BYTES ||
+           !IS_ALIGNED(cfg->src_addr, 4) ||
+           !IS_ALIGNED(cfg->dst_addr, 4)) {
+               return -EINVAL;
+       }
+
+       spin_lock_irqsave(&mchan->lock, flags);
+
+       mchan->src_per_paddr = cfg->src_addr;
+       mchan->src_tcd_nunits = cfg->src_maxburst;
+       mchan->dst_per_paddr = cfg->dst_addr;
+       mchan->dst_tcd_nunits = cfg->dst_maxburst;
+
+       /* Apply defaults */
+       if (mchan->src_tcd_nunits == 0)
+               mchan->src_tcd_nunits = 1;
+       if (mchan->dst_tcd_nunits == 0)
+               mchan->dst_tcd_nunits = 1;
+
+       spin_unlock_irqrestore(&mchan->lock, flags);
+
+       return 0;
+}
+
+static int mpc_dma_device_terminate_all(struct dma_chan *chan)
+{
+       struct mpc_dma_chan *mchan = dma_chan_to_mpc_dma_chan(chan);
+       struct mpc_dma *mdma = dma_chan_to_mpc_dma(chan);
+       unsigned long flags;
+
+       /* Disable channel requests */
+       spin_lock_irqsave(&mchan->lock, flags);
+
+       out_8(&mdma->regs->dmacerq, chan->chan_id);
+       list_splice_tail_init(&mchan->prepared, &mchan->free);
+       list_splice_tail_init(&mchan->queued, &mchan->free);
+       list_splice_tail_init(&mchan->active, &mchan->free);
+
+       spin_unlock_irqrestore(&mchan->lock, flags);
+
+       return 0;
+}
+
+static int mpc_dma_probe(struct platform_device *op)
+{
+       struct device_node *dn = op->dev.of_node;
+       struct device *dev = &op->dev;
+       struct dma_device *dma;
+       struct mpc_dma *mdma;
+       struct mpc_dma_chan *mchan;
+       struct resource res;
+       ulong regs_start, regs_size;
+       int retval, i;
+       u8 chancnt;
+
+       mdma = devm_kzalloc(dev, sizeof(struct mpc_dma), GFP_KERNEL);
+       if (!mdma) {
+               dev_err(dev, "Memory exhausted!\n");
+               retval = -ENOMEM;
+               goto err;
+       }
+
+       mdma->irq = irq_of_parse_and_map(dn, 0);
+       if (mdma->irq == NO_IRQ) {
+               dev_err(dev, "Error mapping IRQ!\n");
+               retval = -EINVAL;
+               goto err;
+       }
+
+       if (of_device_is_compatible(dn, "fsl,mpc8308-dma")) {
+               mdma->is_mpc8308 = 1;
+               mdma->irq2 = irq_of_parse_and_map(dn, 1);
+               if (mdma->irq2 == NO_IRQ) {
+                       dev_err(dev, "Error mapping IRQ!\n");
+                       retval = -EINVAL;
+                       goto err_dispose1;
+               }
+       }
+
+       retval = of_address_to_resource(dn, 0, &res);
+       if (retval) {
+               dev_err(dev, "Error parsing memory region!\n");
+               goto err_dispose2;
+       }
+
+       regs_start = res.start;
+       regs_size = resource_size(&res);
+
+       if (!devm_request_mem_region(dev, regs_start, regs_size, DRV_NAME)) {
+               dev_err(dev, "Error requesting memory region!\n");
+               retval = -EBUSY;
+               goto err_dispose2;
+       }
+
+       mdma->regs = devm_ioremap(dev, regs_start, regs_size);
+       if (!mdma->regs) {
+               dev_err(dev, "Error mapping memory region!\n");
+               retval = -ENOMEM;
+               goto err_dispose2;
+       }
+
+       mdma->tcd = (struct mpc_dma_tcd *)((u8 *)(mdma->regs)
+                                                       + MPC_DMA_TCD_OFFSET);
+
+       retval = request_irq(mdma->irq, &mpc_dma_irq, 0, DRV_NAME, mdma);
+       if (retval) {
+               dev_err(dev, "Error requesting IRQ!\n");
+               retval = -EINVAL;
+               goto err_dispose2;
+       }
+
+       if (mdma->is_mpc8308) {
+               retval = request_irq(mdma->irq2, &mpc_dma_irq, 0,
+                                                       DRV_NAME, mdma);
+               if (retval) {
+                       dev_err(dev, "Error requesting IRQ2!\n");
+                       retval = -EINVAL;
+                       goto err_free1;
+               }
+       }
+
+       spin_lock_init(&mdma->error_status_lock);
+
+       dma = &mdma->dma;
+       dma->dev = dev;
+       dma->device_alloc_chan_resources = mpc_dma_alloc_chan_resources;
+       dma->device_free_chan_resources = mpc_dma_free_chan_resources;
+       dma->device_issue_pending = mpc_dma_issue_pending;
+       dma->device_tx_status = mpc_dma_tx_status;
+       dma->device_prep_dma_memcpy = mpc_dma_prep_memcpy;
+       dma->device_prep_slave_sg = mpc_dma_prep_slave_sg;
+       dma->device_config = mpc_dma_device_config;
+       dma->device_terminate_all = mpc_dma_device_terminate_all;
+
+       INIT_LIST_HEAD(&dma->channels);
+       dma_cap_set(DMA_MEMCPY, dma->cap_mask);
+       dma_cap_set(DMA_SLAVE, dma->cap_mask);
+
+       if (mdma->is_mpc8308)
+               chancnt = MPC8308_DMACHAN_MAX;
+       else
+               chancnt = MPC512x_DMACHAN_MAX;
+
+       for (i = 0; i < chancnt; i++) {
+               mchan = &mdma->channels[i];
+
+               mchan->chan.device = dma;
+               dma_cookie_init(&mchan->chan);
+
+               INIT_LIST_HEAD(&mchan->free);
+               INIT_LIST_HEAD(&mchan->prepared);
+               INIT_LIST_HEAD(&mchan->queued);
+               INIT_LIST_HEAD(&mchan->active);
+               INIT_LIST_HEAD(&mchan->completed);
+
+               spin_lock_init(&mchan->lock);
+               list_add_tail(&mchan->chan.device_node, &dma->channels);
+       }
+
+       tasklet_init(&mdma->tasklet, mpc_dma_tasklet, (unsigned long)mdma);
+
+       /*
+        * Configure DMA Engine:
+        * - Dynamic clock,
+        * - Round-robin group arbitration,
+        * - Round-robin channel arbitration.
+        */
+       if (mdma->is_mpc8308) {
+               /* MPC8308 has 16 channels and lacks some registers */
+               out_be32(&mdma->regs->dmacr, MPC_DMA_DMACR_ERCA);
+
+               /* enable snooping */
+               out_be32(&mdma->regs->dmagpor, MPC_DMA_DMAGPOR_SNOOP_ENABLE);
+               /* Disable error interrupts */
+               out_be32(&mdma->regs->dmaeeil, 0);
+
+               /* Clear interrupts status */
+               out_be32(&mdma->regs->dmaintl, 0xFFFF);
+               out_be32(&mdma->regs->dmaerrl, 0xFFFF);
+       } else {
+               out_be32(&mdma->regs->dmacr, MPC_DMA_DMACR_EDCG |
+                                       MPC_DMA_DMACR_ERGA | MPC_DMA_DMACR_ERCA);
+
+               /* Disable hardware DMA requests */
+               out_be32(&mdma->regs->dmaerqh, 0);
+               out_be32(&mdma->regs->dmaerql, 0);
+
+               /* Disable error interrupts */
+               out_be32(&mdma->regs->dmaeeih, 0);
+               out_be32(&mdma->regs->dmaeeil, 0);
+
+               /* Clear interrupts status */
+               out_be32(&mdma->regs->dmainth, 0xFFFFFFFF);
+               out_be32(&mdma->regs->dmaintl, 0xFFFFFFFF);
+               out_be32(&mdma->regs->dmaerrh, 0xFFFFFFFF);
+               out_be32(&mdma->regs->dmaerrl, 0xFFFFFFFF);
+
+               /* Route interrupts to IPIC */
+               out_be32(&mdma->regs->dmaihsa, 0);
+               out_be32(&mdma->regs->dmailsa, 0);
+       }
+
+       /* Register DMA engine */
+       dev_set_drvdata(dev, mdma);
+       retval = dma_async_device_register(dma);
+       if (retval)
+               goto err_free2;
+
+       /* Register with OF helpers for DMA lookups (nonfatal) */
+       if (dev->of_node) {
+               retval = of_dma_controller_register(dev->of_node,
+                                               of_dma_xlate_by_chan_id, mdma);
+               if (retval)
+                       dev_warn(dev, "Could not register for OF lookup\n");
+       }
+
+       return 0;
+
+err_free2:
+       if (mdma->is_mpc8308)
+               free_irq(mdma->irq2, mdma);
+err_free1:
+       free_irq(mdma->irq, mdma);
+err_dispose2:
+       if (mdma->is_mpc8308)
+               irq_dispose_mapping(mdma->irq2);
+err_dispose1:
+       irq_dispose_mapping(mdma->irq);
+err:
+       return retval;
+}
+
+static int mpc_dma_remove(struct platform_device *op)
+{
+       struct device *dev = &op->dev;
+       struct mpc_dma *mdma = dev_get_drvdata(dev);
+
+       if (dev->of_node)
+               of_dma_controller_free(dev->of_node);
+       dma_async_device_unregister(&mdma->dma);
+       if (mdma->is_mpc8308) {
+               free_irq(mdma->irq2, mdma);
+               irq_dispose_mapping(mdma->irq2);
+       }
+       free_irq(mdma->irq, mdma);
+       irq_dispose_mapping(mdma->irq);
+
+       return 0;
+}
+
+static const struct of_device_id mpc_dma_match[] = {
+       { .compatible = "fsl,mpc5121-dma", },
+       { .compatible = "fsl,mpc8308-dma", },
+       {},
+};
+
+static struct platform_driver mpc_dma_driver = {
+       .probe          = mpc_dma_probe,
+       .remove         = mpc_dma_remove,
+       .driver = {
+               .name = DRV_NAME,
+               .of_match_table = mpc_dma_match,
+       },
+};
+
+module_platform_driver(mpc_dma_driver);
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Piotr Ziecik <kosmo@semihalf.com>");