Merge branch 'for-linus' of git://ftp.arm.linux.org.uk/~rmk/linux-arm

Pull ARM changes from Russell King:

 - Perf updates from Will Deacon:
   - Support for Qualcomm Krait processors (run perf on your phone!)
   - Support for Cortex-A12 (run perf stat on your FPGA!)
   - Support for perf_sample_event_took, allowing us to automatically decrease
     the sample rate if we can't handle the PMU interrupts quickly enough
     (run perf record on your FPGA!).

 - Basic uprobes support from David Long:
     This patch series adds basic uprobes support to ARM. It is based on
     patches developed earlier by Rabin Vincent. That approach of adding
     hooks into the kprobes instruction parsing code was not well received.
     This approach separates the ARM instruction parsing code in kprobes out
     into a separate set of functions which can be used by both kprobes and
     uprobes. Both kprobes and uprobes then provide their own semantic action
     tables to process the results of the parsing.

 - ARMv7M (microcontroller) updates from Uwe Kleine-König

 - OMAP DMA updates (recently added Vinod's Ack even though they've been
   sitting in linux-next for a few months) to reduce the reliance of
   omap-dma on the code in arch/arm.

 - SA11x0 changes from Dmitry Eremin-Solenikov and Alexander Shiyan

 - Support for Cortex-A12 CPU

 - Align support for ARMv6 with ARMv7 so they can cooperate better in a
   single zImage.

 - Addition of first AT_HWCAP2 feature bits for ARMv8 crypto support.

 - Removal of IRQ_DISABLED from various ARM files

 - Improved efficiency of virt_to_page() for single zImage

 - Patch from Ulf Hansson to permit runtime PM callbacks to be available for
   AMBA devices for suspend/resume as well.

 - Finally kill asm/system.h on ARM.

* 'for-linus' of git://ftp.arm.linux.org.uk/~rmk/linux-arm: (89 commits)
  dmaengine: omap-dma: more consolidation of CCR register setup
  dmaengine: omap-dma: move IRQ handling to omap-dma
  dmaengine: omap-dma: move register read/writes into omap-dma.c
  ARM: omap: dma: get rid of 'p' allocation and clean up
  ARM: omap: move dma channel allocation into plat-omap code
  ARM: omap: dma: get rid of errata global
  ARM: omap: clean up DMA register accesses
  ARM: omap: remove almost-const variables
  ARM: omap: remove references to disable_irq_lch
  dmaengine: omap-dma: cleanup errata 3.3 handling
  dmaengine: omap-dma: provide register read/write functions
  dmaengine: omap-dma: use cached CCR value when enabling DMA
  dmaengine: omap-dma: move barrier to omap_dma_start_desc()
  dmaengine: omap-dma: move clnk_ctrl setting to preparation functions
  dmaengine: omap-dma: improve efficiency loading C.SA/C.EI/C.FI registers
  dmaengine: omap-dma: consolidate clearing channel status register
  dmaengine: omap-dma: move CCR buffering disable errata out of the fast path
  dmaengine: omap-dma: provide register definitions
  dmaengine: omap-dma: consolidate setup of CCR
  dmaengine: omap-dma: consolidate setup of CSDP
  ...

5 files changed, 600 insertions, 75 deletions

diff --git a/drivers/amba/bus.c b/drivers/amba/bus.c
index 9e6029105607..3cf61a127ee5 100644
--- a/drivers/amba/bus.c
+++ b/drivers/amba/bus.c
@@ -83,7 +83,7 @@ static struct device_attribute amba_dev_attrs[] = {
 	__ATTR_NULL,
 };
 
-#ifdef CONFIG_PM_RUNTIME
+#ifdef CONFIG_PM
 /*
  * Hooks to provide runtime PM of the pclk (bus clock).  It is safe to
  * enable/disable the bus clock at runtime PM suspend/resume as this
@@ -123,7 +123,7 @@ static const struct dev_pm_ops amba_pm = {
 	.thaw		= pm_generic_thaw,
 	.poweroff	= pm_generic_poweroff,
 	.restore	= pm_generic_restore,
-	SET_RUNTIME_PM_OPS(
+	SET_PM_RUNTIME_PM_OPS(
 		amba_pm_runtime_suspend,
 		amba_pm_runtime_resume,
 		NULL
diff --git a/drivers/clocksource/timer-u300.c b/drivers/clocksource/timer-u300.c
index e63d469661fd..5dcf756970e7 100644
--- a/drivers/clocksource/timer-u300.c
+++ b/drivers/clocksource/timer-u300.c
@@ -333,7 +333,7 @@ static irqreturn_t u300_timer_interrupt(int irq, void *dev_id)
 
 static struct irqaction u300_timer_irq = {
 	.name		= "U300 Timer Tick",
-	.flags		= IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL,
+	.flags		= IRQF_TIMER | IRQF_IRQPOLL,
 	.handler	= u300_timer_interrupt,
 };
 
diff --git a/drivers/dma/omap-dma.c b/drivers/dma/omap-dma.c
index 362e7c49f2e1..64ceca2920b8 100644
--- a/drivers/dma/omap-dma.c
+++ b/drivers/dma/omap-dma.c
@@ -5,6 +5,7 @@
  * it under the terms of the GNU General Public License version 2 as
  * published by the Free Software Foundation.
  */
+#include <linux/delay.h>
 #include <linux/dmaengine.h>
 #include <linux/dma-mapping.h>
 #include <linux/err.h>
@@ -26,11 +27,21 @@ struct omap_dmadev {
 	spinlock_t lock;
 	struct tasklet_struct task;
 	struct list_head pending;
+	void __iomem *base;
+	const struct omap_dma_reg *reg_map;
+	struct omap_system_dma_plat_info *plat;
+	bool legacy;
+	spinlock_t irq_lock;
+	uint32_t irq_enable_mask;
+	struct omap_chan *lch_map[32];
 };
 
 struct omap_chan {
 	struct virt_dma_chan vc;
 	struct list_head node;
+	void __iomem *channel_base;
+	const struct omap_dma_reg *reg_map;
+	uint32_t ccr;
 
 	struct dma_slave_config	cfg;
 	unsigned dma_sig;
@@ -54,19 +65,93 @@ struct omap_desc {
 	dma_addr_t dev_addr;
 
 	int16_t fi;		/* for OMAP_DMA_SYNC_PACKET */
-	uint8_t es;		/* OMAP_DMA_DATA_TYPE_xxx */
-	uint8_t sync_mode;	/* OMAP_DMA_SYNC_xxx */
-	uint8_t sync_type;	/* OMAP_DMA_xxx_SYNC* */
-	uint8_t periph_port;	/* Peripheral port */
+	uint8_t es;		/* CSDP_DATA_TYPE_xxx */
+	uint32_t ccr;		/* CCR value */
+	uint16_t clnk_ctrl;	/* CLNK_CTRL value */
+	uint16_t cicr;		/* CICR value */
+	uint32_t csdp;		/* CSDP value */
 
 	unsigned sglen;
 	struct omap_sg sg[0];
 };
 
+enum {
+	CCR_FS			= BIT(5),
+	CCR_READ_PRIORITY	= BIT(6),
+	CCR_ENABLE		= BIT(7),
+	CCR_AUTO_INIT		= BIT(8),	/* OMAP1 only */
+	CCR_REPEAT		= BIT(9),	/* OMAP1 only */
+	CCR_OMAP31_DISABLE	= BIT(10),	/* OMAP1 only */
+	CCR_SUSPEND_SENSITIVE	= BIT(8),	/* OMAP2+ only */
+	CCR_RD_ACTIVE		= BIT(9),	/* OMAP2+ only */
+	CCR_WR_ACTIVE		= BIT(10),	/* OMAP2+ only */
+	CCR_SRC_AMODE_CONSTANT	= 0 << 12,
+	CCR_SRC_AMODE_POSTINC	= 1 << 12,
+	CCR_SRC_AMODE_SGLIDX	= 2 << 12,
+	CCR_SRC_AMODE_DBLIDX	= 3 << 12,
+	CCR_DST_AMODE_CONSTANT	= 0 << 14,
+	CCR_DST_AMODE_POSTINC	= 1 << 14,
+	CCR_DST_AMODE_SGLIDX	= 2 << 14,
+	CCR_DST_AMODE_DBLIDX	= 3 << 14,
+	CCR_CONSTANT_FILL	= BIT(16),
+	CCR_TRANSPARENT_COPY	= BIT(17),
+	CCR_BS			= BIT(18),
+	CCR_SUPERVISOR		= BIT(22),
+	CCR_PREFETCH		= BIT(23),
+	CCR_TRIGGER_SRC		= BIT(24),
+	CCR_BUFFERING_DISABLE	= BIT(25),
+	CCR_WRITE_PRIORITY	= BIT(26),
+	CCR_SYNC_ELEMENT	= 0,
+	CCR_SYNC_FRAME		= CCR_FS,
+	CCR_SYNC_BLOCK		= CCR_BS,
+	CCR_SYNC_PACKET		= CCR_BS | CCR_FS,
+
+	CSDP_DATA_TYPE_8	= 0,
+	CSDP_DATA_TYPE_16	= 1,
+	CSDP_DATA_TYPE_32	= 2,
+	CSDP_SRC_PORT_EMIFF	= 0 << 2, /* OMAP1 only */
+	CSDP_SRC_PORT_EMIFS	= 1 << 2, /* OMAP1 only */
+	CSDP_SRC_PORT_OCP_T1	= 2 << 2, /* OMAP1 only */
+	CSDP_SRC_PORT_TIPB	= 3 << 2, /* OMAP1 only */
+	CSDP_SRC_PORT_OCP_T2	= 4 << 2, /* OMAP1 only */
+	CSDP_SRC_PORT_MPUI	= 5 << 2, /* OMAP1 only */
+	CSDP_SRC_PACKED		= BIT(6),
+	CSDP_SRC_BURST_1	= 0 << 7,
+	CSDP_SRC_BURST_16	= 1 << 7,
+	CSDP_SRC_BURST_32	= 2 << 7,
+	CSDP_SRC_BURST_64	= 3 << 7,
+	CSDP_DST_PORT_EMIFF	= 0 << 9, /* OMAP1 only */
+	CSDP_DST_PORT_EMIFS	= 1 << 9, /* OMAP1 only */
+	CSDP_DST_PORT_OCP_T1	= 2 << 9, /* OMAP1 only */
+	CSDP_DST_PORT_TIPB	= 3 << 9, /* OMAP1 only */
+	CSDP_DST_PORT_OCP_T2	= 4 << 9, /* OMAP1 only */
+	CSDP_DST_PORT_MPUI	= 5 << 9, /* OMAP1 only */
+	CSDP_DST_PACKED		= BIT(13),
+	CSDP_DST_BURST_1	= 0 << 14,
+	CSDP_DST_BURST_16	= 1 << 14,
+	CSDP_DST_BURST_32	= 2 << 14,
+	CSDP_DST_BURST_64	= 3 << 14,
+
+	CICR_TOUT_IE		= BIT(0),	/* OMAP1 only */
+	CICR_DROP_IE		= BIT(1),
+	CICR_HALF_IE		= BIT(2),
+	CICR_FRAME_IE		= BIT(3),
+	CICR_LAST_IE		= BIT(4),
+	CICR_BLOCK_IE		= BIT(5),
+	CICR_PKT_IE		= BIT(7),	/* OMAP2+ only */
+	CICR_TRANS_ERR_IE	= BIT(8),	/* OMAP2+ only */
+	CICR_SUPERVISOR_ERR_IE	= BIT(10),	/* OMAP2+ only */
+	CICR_MISALIGNED_ERR_IE	= BIT(11),	/* OMAP2+ only */
+	CICR_DRAIN_IE		= BIT(12),	/* OMAP2+ only */
+	CICR_SUPER_BLOCK_IE	= BIT(14),	/* OMAP2+ only */
+
+	CLNK_CTRL_ENABLE_LNK	= BIT(15),
+};
+
 static const unsigned es_bytes[] = {
-	[OMAP_DMA_DATA_TYPE_S8] = 1,
-	[OMAP_DMA_DATA_TYPE_S16] = 2,
-	[OMAP_DMA_DATA_TYPE_S32] = 4,
+	[CSDP_DATA_TYPE_8] = 1,
+	[CSDP_DATA_TYPE_16] = 2,
+	[CSDP_DATA_TYPE_32] = 4,
 };
 
 static struct of_dma_filter_info omap_dma_info = {
@@ -93,28 +178,214 @@ static void omap_dma_desc_free(struct virt_dma_desc *vd)
 	kfree(container_of(vd, struct omap_desc, vd));
 }
 
+static void omap_dma_write(uint32_t val, unsigned type, void __iomem *addr)
+{
+	switch (type) {
+	case OMAP_DMA_REG_16BIT:
+		writew_relaxed(val, addr);
+		break;
+	case OMAP_DMA_REG_2X16BIT:
+		writew_relaxed(val, addr);
+		writew_relaxed(val >> 16, addr + 2);
+		break;
+	case OMAP_DMA_REG_32BIT:
+		writel_relaxed(val, addr);
+		break;
+	default:
+		WARN_ON(1);
+	}
+}
+
+static unsigned omap_dma_read(unsigned type, void __iomem *addr)
+{
+	unsigned val;
+
+	switch (type) {
+	case OMAP_DMA_REG_16BIT:
+		val = readw_relaxed(addr);
+		break;
+	case OMAP_DMA_REG_2X16BIT:
+		val = readw_relaxed(addr);
+		val |= readw_relaxed(addr + 2) << 16;
+		break;
+	case OMAP_DMA_REG_32BIT:
+		val = readl_relaxed(addr);
+		break;
+	default:
+		WARN_ON(1);
+		val = 0;
+	}
+
+	return val;
+}
+
+static void omap_dma_glbl_write(struct omap_dmadev *od, unsigned reg, unsigned val)
+{
+	const struct omap_dma_reg *r = od->reg_map + reg;
+
+	WARN_ON(r->stride);
+
+	omap_dma_write(val, r->type, od->base + r->offset);
+}
+
+static unsigned omap_dma_glbl_read(struct omap_dmadev *od, unsigned reg)
+{
+	const struct omap_dma_reg *r = od->reg_map + reg;
+
+	WARN_ON(r->stride);
+
+	return omap_dma_read(r->type, od->base + r->offset);
+}
+
+static void omap_dma_chan_write(struct omap_chan *c, unsigned reg, unsigned val)
+{
+	const struct omap_dma_reg *r = c->reg_map + reg;
+
+	omap_dma_write(val, r->type, c->channel_base + r->offset);
+}
+
+static unsigned omap_dma_chan_read(struct omap_chan *c, unsigned reg)
+{
+	const struct omap_dma_reg *r = c->reg_map + reg;
+
+	return omap_dma_read(r->type, c->channel_base + r->offset);
+}
+
+static void omap_dma_clear_csr(struct omap_chan *c)
+{
+	if (dma_omap1())
+		omap_dma_chan_read(c, CSR);
+	else
+		omap_dma_chan_write(c, CSR, ~0);
+}
+
+static unsigned omap_dma_get_csr(struct omap_chan *c)
+{
+	unsigned val = omap_dma_chan_read(c, CSR);
+
+	if (!dma_omap1())
+		omap_dma_chan_write(c, CSR, val);
+
+	return val;
+}
+
+static void omap_dma_assign(struct omap_dmadev *od, struct omap_chan *c,
+	unsigned lch)
+{
+	c->channel_base = od->base + od->plat->channel_stride * lch;
+
+	od->lch_map[lch] = c;
+}
+
+static void omap_dma_start(struct omap_chan *c, struct omap_desc *d)
+{
+	struct omap_dmadev *od = to_omap_dma_dev(c->vc.chan.device);
+
+	if (__dma_omap15xx(od->plat->dma_attr))
+		omap_dma_chan_write(c, CPC, 0);
+	else
+		omap_dma_chan_write(c, CDAC, 0);
+
+	omap_dma_clear_csr(c);
+
+	/* Enable interrupts */
+	omap_dma_chan_write(c, CICR, d->cicr);
+
+	/* Enable channel */
+	omap_dma_chan_write(c, CCR, d->ccr | CCR_ENABLE);
+}
+
+static void omap_dma_stop(struct omap_chan *c)
+{
+	struct omap_dmadev *od = to_omap_dma_dev(c->vc.chan.device);
+	uint32_t val;
+
+	/* disable irq */
+	omap_dma_chan_write(c, CICR, 0);
+
+	omap_dma_clear_csr(c);
+
+	val = omap_dma_chan_read(c, CCR);
+	if (od->plat->errata & DMA_ERRATA_i541 && val & CCR_TRIGGER_SRC) {
+		uint32_t sysconfig;
+		unsigned i;
+
+		sysconfig = omap_dma_glbl_read(od, OCP_SYSCONFIG);
+		val = sysconfig & ~DMA_SYSCONFIG_MIDLEMODE_MASK;
+		val |= DMA_SYSCONFIG_MIDLEMODE(DMA_IDLEMODE_NO_IDLE);
+		omap_dma_glbl_write(od, OCP_SYSCONFIG, val);
+
+		val = omap_dma_chan_read(c, CCR);
+		val &= ~CCR_ENABLE;
+		omap_dma_chan_write(c, CCR, val);
+
+		/* Wait for sDMA FIFO to drain */
+		for (i = 0; ; i++) {
+			val = omap_dma_chan_read(c, CCR);
+			if (!(val & (CCR_RD_ACTIVE | CCR_WR_ACTIVE)))
+				break;
+
+			if (i > 100)
+				break;
+
+			udelay(5);
+		}
+
+		if (val & (CCR_RD_ACTIVE | CCR_WR_ACTIVE))
+			dev_err(c->vc.chan.device->dev,
+				"DMA drain did not complete on lch %d\n",
+			        c->dma_ch);
+
+		omap_dma_glbl_write(od, OCP_SYSCONFIG, sysconfig);
+	} else {
+		val &= ~CCR_ENABLE;
+		omap_dma_chan_write(c, CCR, val);
+	}
+
+	mb();
+
+	if (!__dma_omap15xx(od->plat->dma_attr) && c->cyclic) {
+		val = omap_dma_chan_read(c, CLNK_CTRL);
+
+		if (dma_omap1())
+			val |= 1 << 14; /* set the STOP_LNK bit */
+		else
+			val &= ~CLNK_CTRL_ENABLE_LNK;
+
+		omap_dma_chan_write(c, CLNK_CTRL, val);
+	}
+}
+
 static void omap_dma_start_sg(struct omap_chan *c, struct omap_desc *d,
 	unsigned idx)
 {
 	struct omap_sg *sg = d->sg + idx;
+	unsigned cxsa, cxei, cxfi;
 
-	if (d->dir == DMA_DEV_TO_MEM)
-		omap_set_dma_dest_params(c->dma_ch, OMAP_DMA_PORT_EMIFF,
-			OMAP_DMA_AMODE_POST_INC, sg->addr, 0, 0);
-	else
-		omap_set_dma_src_params(c->dma_ch, OMAP_DMA_PORT_EMIFF,
-			OMAP_DMA_AMODE_POST_INC, sg->addr, 0, 0);
+	if (d->dir == DMA_DEV_TO_MEM) {
+		cxsa = CDSA;
+		cxei = CDEI;
+		cxfi = CDFI;
+	} else {
+		cxsa = CSSA;
+		cxei = CSEI;
+		cxfi = CSFI;
+	}
 
-	omap_set_dma_transfer_params(c->dma_ch, d->es, sg->en, sg->fn,
-		d->sync_mode, c->dma_sig, d->sync_type);
+	omap_dma_chan_write(c, cxsa, sg->addr);
+	omap_dma_chan_write(c, cxei, 0);
+	omap_dma_chan_write(c, cxfi, 0);
+	omap_dma_chan_write(c, CEN, sg->en);
+	omap_dma_chan_write(c, CFN, sg->fn);
 
-	omap_start_dma(c->dma_ch);
+	omap_dma_start(c, d);
 }
 
 static void omap_dma_start_desc(struct omap_chan *c)
 {
 	struct virt_dma_desc *vd = vchan_next_desc(&c->vc);
 	struct omap_desc *d;
+	unsigned cxsa, cxei, cxfi;
 
 	if (!vd) {
 		c->desc = NULL;
@@ -126,12 +397,32 @@ static void omap_dma_start_desc(struct omap_chan *c)
 	c->desc = d = to_omap_dma_desc(&vd->tx);
 	c->sgidx = 0;
 
-	if (d->dir == DMA_DEV_TO_MEM)
-		omap_set_dma_src_params(c->dma_ch, d->periph_port,
-			OMAP_DMA_AMODE_CONSTANT, d->dev_addr, 0, d->fi);
-	else
-		omap_set_dma_dest_params(c->dma_ch, d->periph_port,
-			OMAP_DMA_AMODE_CONSTANT, d->dev_addr, 0, d->fi);
+	/*
+	 * This provides the necessary barrier to ensure data held in
+	 * DMA coherent memory is visible to the DMA engine prior to
+	 * the transfer starting.
+	 */
+	mb();
+
+	omap_dma_chan_write(c, CCR, d->ccr);
+	if (dma_omap1())
+		omap_dma_chan_write(c, CCR2, d->ccr >> 16);
+
+	if (d->dir == DMA_DEV_TO_MEM) {
+		cxsa = CSSA;
+		cxei = CSEI;
+		cxfi = CSFI;
+	} else {
+		cxsa = CDSA;
+		cxei = CDEI;
+		cxfi = CDFI;
+	}
+
+	omap_dma_chan_write(c, cxsa, d->dev_addr);
+	omap_dma_chan_write(c, cxei, 0);
+	omap_dma_chan_write(c, cxfi, d->fi);
+	omap_dma_chan_write(c, CSDP, d->csdp);
+	omap_dma_chan_write(c, CLNK_CTRL, d->clnk_ctrl);
 
 	omap_dma_start_sg(c, d, 0);
 }
@@ -186,24 +477,118 @@ static void omap_dma_sched(unsigned long data)
 	}
 }
 
+static irqreturn_t omap_dma_irq(int irq, void *devid)
+{
+	struct omap_dmadev *od = devid;
+	unsigned status, channel;
+
+	spin_lock(&od->irq_lock);
+
+	status = omap_dma_glbl_read(od, IRQSTATUS_L1);
+	status &= od->irq_enable_mask;
+	if (status == 0) {
+		spin_unlock(&od->irq_lock);
+		return IRQ_NONE;
+	}
+
+	while ((channel = ffs(status)) != 0) {
+		unsigned mask, csr;
+		struct omap_chan *c;
+
+		channel -= 1;
+		mask = BIT(channel);
+		status &= ~mask;
+
+		c = od->lch_map[channel];
+		if (c == NULL) {
+			/* This should never happen */
+			dev_err(od->ddev.dev, "invalid channel %u\n", channel);
+			continue;
+		}
+
+		csr = omap_dma_get_csr(c);
+		omap_dma_glbl_write(od, IRQSTATUS_L1, mask);
+
+		omap_dma_callback(channel, csr, c);
+	}
+
+	spin_unlock(&od->irq_lock);
+
+	return IRQ_HANDLED;
+}
+
 static int omap_dma_alloc_chan_resources(struct dma_chan *chan)
 {
+	struct omap_dmadev *od = to_omap_dma_dev(chan->device);
 	struct omap_chan *c = to_omap_dma_chan(chan);
+	int ret;
+
+	if (od->legacy) {
+		ret = omap_request_dma(c->dma_sig, "DMA engine",
+				       omap_dma_callback, c, &c->dma_ch);
+	} else {
+		ret = omap_request_dma(c->dma_sig, "DMA engine", NULL, NULL,
+				       &c->dma_ch);
+	}
+
+	dev_dbg(od->ddev.dev, "allocating channel %u for %u\n",
+		c->dma_ch, c->dma_sig);
 
-	dev_dbg(c->vc.chan.device->dev, "allocating channel for %u\n", c->dma_sig);
+	if (ret >= 0) {
+		omap_dma_assign(od, c, c->dma_ch);
 
-	return omap_request_dma(c->dma_sig, "DMA engine",
-		omap_dma_callback, c, &c->dma_ch);
+		if (!od->legacy) {
+			unsigned val;
+
+			spin_lock_irq(&od->irq_lock);
+			val = BIT(c->dma_ch);
+			omap_dma_glbl_write(od, IRQSTATUS_L1, val);
+			od->irq_enable_mask |= val;
+			omap_dma_glbl_write(od, IRQENABLE_L1, od->irq_enable_mask);
+
+			val = omap_dma_glbl_read(od, IRQENABLE_L0);
+			val &= ~BIT(c->dma_ch);
+			omap_dma_glbl_write(od, IRQENABLE_L0, val);
+			spin_unlock_irq(&od->irq_lock);
+		}
+	}
+
+	if (dma_omap1()) {
+		if (__dma_omap16xx(od->plat->dma_attr)) {
+			c->ccr = CCR_OMAP31_DISABLE;
+			/* Duplicate what plat-omap/dma.c does */
+			c->ccr |= c->dma_ch + 1;
+		} else {
+			c->ccr = c->dma_sig & 0x1f;
+		}
+	} else {
+		c->ccr = c->dma_sig & 0x1f;
+		c->ccr |= (c->dma_sig & ~0x1f) << 14;
+	}
+	if (od->plat->errata & DMA_ERRATA_IFRAME_BUFFERING)
+		c->ccr |= CCR_BUFFERING_DISABLE;
+
+	return ret;
 }
 
 static void omap_dma_free_chan_resources(struct dma_chan *chan)
 {
+	struct omap_dmadev *od = to_omap_dma_dev(chan->device);
 	struct omap_chan *c = to_omap_dma_chan(chan);
 
+	if (!od->legacy) {
+		spin_lock_irq(&od->irq_lock);
+		od->irq_enable_mask &= ~BIT(c->dma_ch);
+		omap_dma_glbl_write(od, IRQENABLE_L1, od->irq_enable_mask);
+		spin_unlock_irq(&od->irq_lock);
+	}
+
+	c->channel_base = NULL;
+	od->lch_map[c->dma_ch] = NULL;
 	vchan_free_chan_resources(&c->vc);
 	omap_free_dma(c->dma_ch);
 
-	dev_dbg(c->vc.chan.device->dev, "freeing channel for %u\n", c->dma_sig);
+	dev_dbg(od->ddev.dev, "freeing channel for %u\n", c->dma_sig);
 }
 
 static size_t omap_dma_sg_size(struct omap_sg *sg)
@@ -239,6 +624,74 @@ static size_t omap_dma_desc_size_pos(struct omap_desc *d, dma_addr_t addr)
 	return size;
 }
 
+/*
+ * OMAP 3.2/3.3 erratum: sometimes 0 is returned if CSAC/CDAC is
+ * read before the DMA controller finished disabling the channel.
+ */
+static uint32_t omap_dma_chan_read_3_3(struct omap_chan *c, unsigned reg)
+{
+	struct omap_dmadev *od = to_omap_dma_dev(c->vc.chan.device);
+	uint32_t val;
+
+	val = omap_dma_chan_read(c, reg);
+	if (val == 0 && od->plat->errata & DMA_ERRATA_3_3)
+		val = omap_dma_chan_read(c, reg);
+
+	return val;
+}
+
+static dma_addr_t omap_dma_get_src_pos(struct omap_chan *c)
+{
+	struct omap_dmadev *od = to_omap_dma_dev(c->vc.chan.device);
+	dma_addr_t addr, cdac;
+
+	if (__dma_omap15xx(od->plat->dma_attr)) {
+		addr = omap_dma_chan_read(c, CPC);
+	} else {
+		addr = omap_dma_chan_read_3_3(c, CSAC);
+		cdac = omap_dma_chan_read_3_3(c, CDAC);
+
+		/*
+		 * CDAC == 0 indicates that the DMA transfer on the channel has
+		 * not been started (no data has been transferred so far).
+		 * Return the programmed source start address in this case.
+		 */
+		if (cdac == 0)
+			addr = omap_dma_chan_read(c, CSSA);
+	}
+
+	if (dma_omap1())
+		addr |= omap_dma_chan_read(c, CSSA) & 0xffff0000;
+
+	return addr;
+}
+
+static dma_addr_t omap_dma_get_dst_pos(struct omap_chan *c)
+{
+	struct omap_dmadev *od = to_omap_dma_dev(c->vc.chan.device);
+	dma_addr_t addr;
+
+	if (__dma_omap15xx(od->plat->dma_attr)) {
+		addr = omap_dma_chan_read(c, CPC);
+	} else {
+		addr = omap_dma_chan_read_3_3(c, CDAC);
+
+		/*
+		 * CDAC == 0 indicates that the DMA transfer on the channel
+		 * has not been started (no data has been transferred so
+		 * far).  Return the programmed destination start address in
+		 * this case.
+		 */
+		if (addr == 0)
+			addr = omap_dma_chan_read(c, CDSA);
+	}
+
+	if (dma_omap1())
+		addr |= omap_dma_chan_read(c, CDSA) & 0xffff0000;
+
+	return addr;
+}
+
 static enum dma_status omap_dma_tx_status(struct dma_chan *chan,
 	dma_cookie_t cookie, struct dma_tx_state *txstate)
 {
@@ -260,9 +713,9 @@ static enum dma_status omap_dma_tx_status(struct dma_chan *chan,
 		dma_addr_t pos;
 
 		if (d->dir == DMA_MEM_TO_DEV)
-			pos = omap_get_dma_src_pos(c->dma_ch);
+			pos = omap_dma_get_src_pos(c);
 		else if (d->dir == DMA_DEV_TO_MEM)
-			pos = omap_get_dma_dst_pos(c->dma_ch);
+			pos = omap_dma_get_dst_pos(c);
 		else
 			pos = 0;
 
@@ -304,24 +757,23 @@ static struct dma_async_tx_descriptor *omap_dma_prep_slave_sg(
 	struct dma_chan *chan, struct scatterlist *sgl, unsigned sglen,
 	enum dma_transfer_direction dir, unsigned long tx_flags, void *context)
 {
+	struct omap_dmadev *od = to_omap_dma_dev(chan->device);
 	struct omap_chan *c = to_omap_dma_chan(chan);
 	enum dma_slave_buswidth dev_width;
 	struct scatterlist *sgent;
 	struct omap_desc *d;
 	dma_addr_t dev_addr;
-	unsigned i, j = 0, es, en, frame_bytes, sync_type;
+	unsigned i, j = 0, es, en, frame_bytes;
 	u32 burst;
 
 	if (dir == DMA_DEV_TO_MEM) {
 		dev_addr = c->cfg.src_addr;
 		dev_width = c->cfg.src_addr_width;
 		burst = c->cfg.src_maxburst;
-		sync_type = OMAP_DMA_SRC_SYNC;
 	} else if (dir == DMA_MEM_TO_DEV) {
 		dev_addr = c->cfg.dst_addr;
 		dev_width = c->cfg.dst_addr_width;
 		burst = c->cfg.dst_maxburst;
-		sync_type = OMAP_DMA_DST_SYNC;
 	} else {
 		dev_err(chan->device->dev, "%s: bad direction?\n", __func__);
 		return NULL;
@@ -330,13 +782,13 @@ static struct dma_async_tx_descriptor *omap_dma_prep_slave_sg(
 	/* Bus width translates to the element size (ES) */
 	switch (dev_width) {
 	case DMA_SLAVE_BUSWIDTH_1_BYTE:
-		es = OMAP_DMA_DATA_TYPE_S8;
+		es = CSDP_DATA_TYPE_8;
 		break;
 	case DMA_SLAVE_BUSWIDTH_2_BYTES:
-		es = OMAP_DMA_DATA_TYPE_S16;
+		es = CSDP_DATA_TYPE_16;
 		break;
 	case DMA_SLAVE_BUSWIDTH_4_BYTES:
-		es = OMAP_DMA_DATA_TYPE_S32;
+		es = CSDP_DATA_TYPE_32;
 		break;
 	default: /* not reached */
 		return NULL;
@@ -350,9 +802,31 @@ static struct dma_async_tx_descriptor *omap_dma_prep_slave_sg(
 	d->dir = dir;
 	d->dev_addr = dev_addr;
 	d->es = es;
-	d->sync_mode = OMAP_DMA_SYNC_FRAME;
-	d->sync_type = sync_type;
-	d->periph_port = OMAP_DMA_PORT_TIPB;
+
+	d->ccr = c->ccr | CCR_SYNC_FRAME;
+	if (dir == DMA_DEV_TO_MEM)
+		d->ccr |= CCR_DST_AMODE_POSTINC | CCR_SRC_AMODE_CONSTANT;
+	else
+		d->ccr |= CCR_DST_AMODE_CONSTANT | CCR_SRC_AMODE_POSTINC;
+
+	d->cicr = CICR_DROP_IE | CICR_BLOCK_IE;
+	d->csdp = es;
+
+	if (dma_omap1()) {
+		d->cicr |= CICR_TOUT_IE;
+
+		if (dir == DMA_DEV_TO_MEM)
+			d->csdp |= CSDP_DST_PORT_EMIFF | CSDP_SRC_PORT_TIPB;
+		else
+			d->csdp |= CSDP_DST_PORT_TIPB | CSDP_SRC_PORT_EMIFF;
+	} else {
+		if (dir == DMA_DEV_TO_MEM)
+			d->ccr |= CCR_TRIGGER_SRC;
+
+		d->cicr |= CICR_MISALIGNED_ERR_IE | CICR_TRANS_ERR_IE;
+	}
+	if (od->plat->errata & DMA_ERRATA_PARALLEL_CHANNELS)
+		d->clnk_ctrl = c->dma_ch;
 
 	/*
 	 * Build our scatterlist entries: each contains the address,
@@ -382,23 +856,22 @@ static struct dma_async_tx_descriptor *omap_dma_prep_dma_cyclic(
 	size_t period_len, enum dma_transfer_direction dir, unsigned long flags,
 	void *context)
 {
+	struct omap_dmadev *od = to_omap_dma_dev(chan->device);
 	struct omap_chan *c = to_omap_dma_chan(chan);
 	enum dma_slave_buswidth dev_width;
 	struct omap_desc *d;
 	dma_addr_t dev_addr;
-	unsigned es, sync_type;
+	unsigned es;
 	u32 burst;
 
 	if (dir == DMA_DEV_TO_MEM) {
 		dev_addr = c->cfg.src_addr;
 		dev_width = c->cfg.src_addr_width;
 		burst = c->cfg.src_maxburst;
-		sync_type = OMAP_DMA_SRC_SYNC;
 	} else if (dir == DMA_MEM_TO_DEV) {
 		dev_addr = c->cfg.dst_addr;
 		dev_width = c->cfg.dst_addr_width;
 		burst = c->cfg.dst_maxburst;
-		sync_type = OMAP_DMA_DST_SYNC;
 	} else {
 		dev_err(chan->device->dev, "%s: bad direction?\n", __func__);
 		return NULL;
@@ -407,13 +880,13 @@ static struct dma_async_tx_descriptor *omap_dma_prep_dma_cyclic(
 	/* Bus width translates to the element size (ES) */
 	switch (dev_width) {
 	case DMA_SLAVE_BUSWIDTH_1_BYTE:
-		es = OMAP_DMA_DATA_TYPE_S8;
+		es = CSDP_DATA_TYPE_8;
 		break;
 	case DMA_SLAVE_BUSWIDTH_2_BYTES:
-		es = OMAP_DMA_DATA_TYPE_S16;
+		es = CSDP_DATA_TYPE_16;
 		break;
 	case DMA_SLAVE_BUSWIDTH_4_BYTES:
-		es = OMAP_DMA_DATA_TYPE_S32;
+		es = CSDP_DATA_TYPE_32;
 		break;
 	default: /* not reached */
 		return NULL;
@@ -428,32 +901,51 @@ static struct dma_async_tx_descriptor *omap_dma_prep_dma_cyclic(
 	d->dev_addr = dev_addr;
 	d->fi = burst;
 	d->es = es;
-	if (burst)
-		d->sync_mode = OMAP_DMA_SYNC_PACKET;
-	else
-		d->sync_mode = OMAP_DMA_SYNC_ELEMENT;
-	d->sync_type = sync_type;
-	d->periph_port = OMAP_DMA_PORT_MPUI;
 	d->sg[0].addr = buf_addr;
 	d->sg[0].en = period_len / es_bytes[es];
 	d->sg[0].fn = buf_len / period_len;
 	d->sglen = 1;
 
-	if (!c->cyclic) {
-		c->cyclic = true;
-		omap_dma_link_lch(c->dma_ch, c->dma_ch);
+	d->ccr = c->ccr;
+	if (dir == DMA_DEV_TO_MEM)
+		d->ccr |= CCR_DST_AMODE_POSTINC | CCR_SRC_AMODE_CONSTANT;
+	else
+		d->ccr |= CCR_DST_AMODE_CONSTANT | CCR_SRC_AMODE_POSTINC;
 
-		if (flags & DMA_PREP_INTERRUPT)
-			omap_enable_dma_irq(c->dma_ch, OMAP_DMA_FRAME_IRQ);
+	d->cicr = CICR_DROP_IE;
+	if (flags & DMA_PREP_INTERRUPT)
+		d->cicr |= CICR_FRAME_IE;
 
-		omap_disable_dma_irq(c->dma_ch, OMAP_DMA_BLOCK_IRQ);
-	}
+	d->csdp = es;
+
+	if (dma_omap1()) {
+		d->cicr |= CICR_TOUT_IE;
+
+		if (dir == DMA_DEV_TO_MEM)
+			d->csdp |= CSDP_DST_PORT_EMIFF | CSDP_SRC_PORT_MPUI;
+		else
+			d->csdp |= CSDP_DST_PORT_MPUI | CSDP_SRC_PORT_EMIFF;
+	} else {
+		if (burst)
+			d->ccr |= CCR_SYNC_PACKET;
+		else
+			d->ccr |= CCR_SYNC_ELEMENT;
+
+		if (dir == DMA_DEV_TO_MEM)
+			d->ccr |= CCR_TRIGGER_SRC;
+
+		d->cicr |= CICR_MISALIGNED_ERR_IE | CICR_TRANS_ERR_IE;
 
-	if (dma_omap2plus()) {
-		omap_set_dma_src_burst_mode(c->dma_ch, OMAP_DMA_DATA_BURST_16);
-		omap_set_dma_dest_burst_mode(c->dma_ch, OMAP_DMA_DATA_BURST_16);
+		d->csdp |= CSDP_DST_BURST_64 | CSDP_SRC_BURST_64;
 	}
 
+	if (__dma_omap15xx(od->plat->dma_attr))
+		d->ccr |= CCR_AUTO_INIT | CCR_REPEAT;
+	else
+		d->clnk_ctrl = c->dma_ch | CLNK_CTRL_ENABLE_LNK;
+
+	c->cyclic = true;
+
 	return vchan_tx_prep(&c->vc, &d->vd, flags);
 }
 
@@ -483,20 +975,19 @@ static int omap_dma_terminate_all(struct omap_chan *c)
 
 	/*
 	 * Stop DMA activity: we assume the callback will not be called
-	 * after omap_stop_dma() returns (even if it does, it will see
+	 * after omap_dma_stop() returns (even if it does, it will see
 	 * c->desc is NULL and exit.)
 	 */
 	if (c->desc) {
 		c->desc = NULL;
 		/* Avoid stopping the dma twice */
 		if (!c->paused)
-			omap_stop_dma(c->dma_ch);
+			omap_dma_stop(c);
 	}
 
 	if (c->cyclic) {
 		c->cyclic = false;
 		c->paused = false;
-		omap_dma_unlink_lch(c->dma_ch, c->dma_ch);
 	}
 
 	vchan_get_all_descriptors(&c->vc, &head);
@@ -513,7 +1004,7 @@ static int omap_dma_pause(struct omap_chan *c)
 		return -EINVAL;
 
 	if (!c->paused) {
-		omap_stop_dma(c->dma_ch);
+		omap_dma_stop(c);
 		c->paused = true;
 	}
 
@@ -527,7 +1018,7 @@ static int omap_dma_resume(struct omap_chan *c)
 		return -EINVAL;
 
 	if (c->paused) {
-		omap_start_dma(c->dma_ch);
+		omap_dma_start(c, c->desc);
 		c->paused = false;
 	}
 
@@ -573,6 +1064,7 @@ static int omap_dma_chan_init(struct omap_dmadev *od, int dma_sig)
 	if (!c)
 		return -ENOMEM;
 
+	c->reg_map = od->reg_map;
 	c->dma_sig = dma_sig;
 	c->vc.desc_free = omap_dma_desc_free;
 	vchan_init(&c->vc, &od->ddev);
@@ -594,18 +1086,29 @@ static void omap_dma_free(struct omap_dmadev *od)
 		tasklet_kill(&c->vc.task);
 		kfree(c);
 	}
-	kfree(od);
 }
 
 static int omap_dma_probe(struct platform_device *pdev)
 {
 	struct omap_dmadev *od;
-	int rc, i;
+	struct resource *res;
+	int rc, i, irq;
 
-	od = kzalloc(sizeof(*od), GFP_KERNEL);
+	od = devm_kzalloc(&pdev->dev, sizeof(*od), GFP_KERNEL);
 	if (!od)
 		return -ENOMEM;
 
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	od->base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(od->base))
+		return PTR_ERR(od->base);
+
+	od->plat = omap_get_plat_info();
+	if (!od->plat)
+		return -EPROBE_DEFER;
+
+	od->reg_map = od->plat->reg_map;
+
 	dma_cap_set(DMA_SLAVE, od->ddev.cap_mask);
 	dma_cap_set(DMA_CYCLIC, od->ddev.cap_mask);
 	od->ddev.device_alloc_chan_resources = omap_dma_alloc_chan_resources;
@@ -619,6 +1122,7 @@ static int omap_dma_probe(struct platform_device *pdev)
 	INIT_LIST_HEAD(&od->ddev.channels);
 	INIT_LIST_HEAD(&od->pending);
 	spin_lock_init(&od->lock);
+	spin_lock_init(&od->irq_lock);
 
 	tasklet_init(&od->task, omap_dma_sched, (unsigned long)od);
 
@@ -630,6 +1134,21 @@ static int omap_dma_probe(struct platform_device *pdev)
 		}
 	}
 
+	irq = platform_get_irq(pdev, 1);
+	if (irq <= 0) {
+		dev_info(&pdev->dev, "failed to get L1 IRQ: %d\n", irq);
+		od->legacy = true;
+	} else {
+		/* Disable all interrupts */
+		od->irq_enable_mask = 0;
+		omap_dma_glbl_write(od, IRQENABLE_L1, 0);
+
+		rc = devm_request_irq(&pdev->dev, irq, omap_dma_irq,
+				      IRQF_SHARED, "omap-dma-engine", od);
+		if (rc)
+			return rc;
+	}
+
 	rc = dma_async_device_register(&od->ddev);
 	if (rc) {
 		pr_warn("OMAP-DMA: failed to register slave DMA engine device: %d\n",
@@ -666,6 +1185,12 @@ static int omap_dma_remove(struct platform_device *pdev)
 		of_dma_controller_free(pdev->dev.of_node);
 
 	dma_async_device_unregister(&od->ddev);
+
+	if (!od->legacy) {
+		/* Disable all interrupts */
+		omap_dma_glbl_write(od, IRQENABLE_L0, 0);
+	}
+
 	omap_dma_free(od);
 
 	return 0;
diff --git a/drivers/mmc/host/mmci.h b/drivers/mmc/host/mmci.h
index 168bc72f7a94..84c0e59b792a 100644
--- a/drivers/mmc/host/mmci.h
+++ b/drivers/mmc/host/mmci.h
@@ -38,10 +38,11 @@
 #define MCI_CPSM_INTERRUPT	(1 << 8)
 #define MCI_CPSM_PENDING	(1 << 9)
 #define MCI_CPSM_ENABLE		(1 << 10)
-#define MCI_SDIO_SUSP		(1 << 11)
-#define MCI_ENCMD_COMPL		(1 << 12)
-#define MCI_NIEN		(1 << 13)
-#define MCI_CE_ATACMD		(1 << 14)
+/* Argument flag extenstions in the ST Micro versions */
+#define MCI_ST_SDIO_SUSP	(1 << 11)
+#define MCI_ST_ENCMD_COMPL	(1 << 12)
+#define MCI_ST_NIEN		(1 << 13)
+#define MCI_ST_CE_ATACMD	(1 << 14)
 
 #define MMCIRESPCMD		0x010
 #define MMCIRESPONSE0		0x014
diff --git a/drivers/usb/gadget/lpc32xx_udc.c b/drivers/usb/gadget/lpc32xx_udc.c
index a139894c600f..e471580a2a3b 100644
--- a/drivers/usb/gadget/lpc32xx_udc.c
+++ b/drivers/usb/gadget/lpc32xx_udc.c
@@ -55,7 +55,6 @@
 #include <mach/hardware.h>
 #include <linux/io.h>
 #include <asm/irq.h>
-#include <asm/system.h>
 
 #include <mach/platform.h>
 #include <mach/irqs.h>