Merge git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6: (127 commits)
  sh: update defconfigs.
  sh: Fix up the NUMA build for recent LMB changes.
  sh64: provide a stub per_cpu_trap_init() definition.
  sh: fix up CONFIG_KEXEC=n build.
  sh: fixup the docbook paths for clock framework shuffling.
  driver core: Early dev_name() depends on slab_is_available().
  sh: simplify WARN usage in SH clock driver
  sh: Check return value of clk_get on ms7724
  sh: Check return value of clk_get on ecovec24
  sh: move sh clock-cpg.c contents to drivers/sh/clk-cpg.c
  sh: move sh clock.c contents to drivers/sh/clk.
  sh: move sh asm/clock.h contents to linux/sh_clk.h V2
  sh: remove unused clock lookup
  sh: switch boards to clkdev
  sh: switch sh4-202 to clkdev
  sh: switch shx3 to clkdev
  sh: switch sh7757 to clkdev
  sh: switch sh7763 to clkdev
  sh: switch sh7780 to clkdev
  sh: switch sh7786 to clkdev
  ...

12 files changed, 1396 insertions, 167 deletions

diff --git a/drivers/base/platform.c b/drivers/base/platform.c
index c5fbe198fbd..765bcf0df3b 100644
--- a/drivers/base/platform.c
+++ b/drivers/base/platform.c
@@ -1254,6 +1254,26 @@ static int __init early_platform_driver_probe_id(char *class_str,
 		}
 
 		if (match) {
+			/*
+			 * Set up a sensible init_name to enable
+			 * dev_name() and others to be used before the
+			 * rest of the driver core is initialized.
+			 */
+			if (!match->dev.init_name && slab_is_available()) {
+				if (match->id != -1)
+					match->dev.init_name =
+						kasprintf(GFP_KERNEL, "%s.%d",
+							  match->name,
+							  match->id);
+				else
+					match->dev.init_name =
+						kasprintf(GFP_KERNEL, "%s",
+							  match->name);
+
+				if (!match->dev.init_name)
+					return -ENOMEM;
+			}
+
 			if (epdrv->pdrv->probe(match))
 				pr_warning("%s: unable to probe %s early.\n",
 					   class_str, match->name);
diff --git a/drivers/clocksource/sh_cmt.c b/drivers/clocksource/sh_cmt.c
index 744f748cc84..f6677cb1978 100644
--- a/drivers/clocksource/sh_cmt.c
+++ b/drivers/clocksource/sh_cmt.c
@@ -150,13 +150,12 @@ static void sh_cmt_start_stop_ch(struct sh_cmt_priv *p, int start)
 
 static int sh_cmt_enable(struct sh_cmt_priv *p, unsigned long *rate)
 {
-	struct sh_timer_config *cfg = p->pdev->dev.platform_data;
 	int ret;
 
 	/* enable clock */
 	ret = clk_enable(p->clk);
 	if (ret) {
-		pr_err("sh_cmt: cannot enable clock \"%s\"\n", cfg->clk);
+		dev_err(&p->pdev->dev, "cannot enable clock\n");
 		return ret;
 	}
 
@@ -279,7 +278,7 @@ static void sh_cmt_clock_event_program_verify(struct sh_cmt_priv *p,
 			delay = 1;
 
 		if (!delay)
-			pr_warning("sh_cmt: too long delay\n");
+			dev_warn(&p->pdev->dev, "too long delay\n");
 
 	} while (delay);
 }
@@ -289,7 +288,7 @@ static void sh_cmt_set_next(struct sh_cmt_priv *p, unsigned long delta)
 	unsigned long flags;
 
 	if (delta > p->max_match_value)
-		pr_warning("sh_cmt: delta out of range\n");
+		dev_warn(&p->pdev->dev, "delta out of range\n");
 
 	spin_lock_irqsave(&p->lock, flags);
 	p->next_match_value = delta;
@@ -451,7 +450,7 @@ static int sh_cmt_register_clocksource(struct sh_cmt_priv *p,
 	cs->resume = sh_cmt_clocksource_resume;
 	cs->mask = CLOCKSOURCE_MASK(sizeof(unsigned long) * 8);
 	cs->flags = CLOCK_SOURCE_IS_CONTINUOUS;
-	pr_info("sh_cmt: %s used as clock source\n", cs->name);
+	dev_info(&p->pdev->dev, "used as clock source\n");
 	clocksource_register(cs);
 	return 0;
 }
@@ -497,13 +496,11 @@ static void sh_cmt_clock_event_mode(enum clock_event_mode mode,
 
 	switch (mode) {
 	case CLOCK_EVT_MODE_PERIODIC:
-		pr_info("sh_cmt: %s used for periodic clock events\n",
-			ced->name);
+		dev_info(&p->pdev->dev, "used for periodic clock events\n");
 		sh_cmt_clock_event_start(p, 1);
 		break;
 	case CLOCK_EVT_MODE_ONESHOT:
-		pr_info("sh_cmt: %s used for oneshot clock events\n",
-			ced->name);
+		dev_info(&p->pdev->dev, "used for oneshot clock events\n");
 		sh_cmt_clock_event_start(p, 0);
 		break;
 	case CLOCK_EVT_MODE_SHUTDOWN:
@@ -544,7 +541,7 @@ static void sh_cmt_register_clockevent(struct sh_cmt_priv *p,
 	ced->set_next_event = sh_cmt_clock_event_next;
 	ced->set_mode = sh_cmt_clock_event_mode;
 
-	pr_info("sh_cmt: %s used for clock events\n", ced->name);
+	dev_info(&p->pdev->dev, "used for clock events\n");
 	clockevents_register_device(ced);
 }
 
@@ -601,22 +598,27 @@ static int sh_cmt_setup(struct sh_cmt_priv *p, struct platform_device *pdev)
 	/* map memory, let mapbase point to our channel */
 	p->mapbase = ioremap_nocache(res->start, resource_size(res));
 	if (p->mapbase == NULL) {
-		pr_err("sh_cmt: failed to remap I/O memory\n");
+		dev_err(&p->pdev->dev, "failed to remap I/O memory\n");
 		goto err0;
 	}
 
 	/* request irq using setup_irq() (too early for request_irq()) */
-	p->irqaction.name = cfg->name;
+	p->irqaction.name = dev_name(&p->pdev->dev);
 	p->irqaction.handler = sh_cmt_interrupt;
 	p->irqaction.dev_id = p;
-	p->irqaction.flags = IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL;
+	p->irqaction.flags = IRQF_DISABLED | IRQF_TIMER | \
+			     IRQF_IRQPOLL  | IRQF_NOBALANCING;
 
 	/* get hold of clock */
-	p->clk = clk_get(&p->pdev->dev, cfg->clk);
+	p->clk = clk_get(&p->pdev->dev, "cmt_fck");
 	if (IS_ERR(p->clk)) {
-		pr_err("sh_cmt: cannot get clock \"%s\"\n", cfg->clk);
-		ret = PTR_ERR(p->clk);
-		goto err1;
+		dev_warn(&p->pdev->dev, "using deprecated clock lookup\n");
+		p->clk = clk_get(&p->pdev->dev, cfg->clk);
+		if (IS_ERR(p->clk)) {
+			dev_err(&p->pdev->dev, "cannot get clock\n");
+			ret = PTR_ERR(p->clk);
+			goto err1;
+		}
 	}
 
 	if (resource_size(res) == 6) {
@@ -629,17 +631,17 @@ static int sh_cmt_setup(struct sh_cmt_priv *p, struct platform_device *pdev)
 		p->clear_bits = ~0xc000;
 	}
 
-	ret = sh_cmt_register(p, cfg->name,
+	ret = sh_cmt_register(p, (char *)dev_name(&p->pdev->dev),
 			      cfg->clockevent_rating,
 			      cfg->clocksource_rating);
 	if (ret) {
-		pr_err("sh_cmt: registration failed\n");
+		dev_err(&p->pdev->dev, "registration failed\n");
 		goto err1;
 	}
 
 	ret = setup_irq(irq, &p->irqaction);
 	if (ret) {
-		pr_err("sh_cmt: failed to request irq %d\n", irq);
+		dev_err(&p->pdev->dev, "failed to request irq %d\n", irq);
 		goto err1;
 	}
 
@@ -654,11 +656,10 @@ err0:
 static int __devinit sh_cmt_probe(struct platform_device *pdev)
 {
 	struct sh_cmt_priv *p = platform_get_drvdata(pdev);
-	struct sh_timer_config *cfg = pdev->dev.platform_data;
 	int ret;
 
 	if (p) {
-		pr_info("sh_cmt: %s kept as earlytimer\n", cfg->name);
+		dev_info(&pdev->dev, "kept as earlytimer\n");
 		return 0;
 	}
 
diff --git a/drivers/clocksource/sh_mtu2.c b/drivers/clocksource/sh_mtu2.c
index 5fb78bfd73b..ef7a5be8a09 100644
--- a/drivers/clocksource/sh_mtu2.c
+++ b/drivers/clocksource/sh_mtu2.c
@@ -119,13 +119,12 @@ static void sh_mtu2_start_stop_ch(struct sh_mtu2_priv *p, int start)
 
 static int sh_mtu2_enable(struct sh_mtu2_priv *p)
 {
-	struct sh_timer_config *cfg = p->pdev->dev.platform_data;
 	int ret;
 
 	/* enable clock */
 	ret = clk_enable(p->clk);
 	if (ret) {
-		pr_err("sh_mtu2: cannot enable clock \"%s\"\n", cfg->clk);
+		dev_err(&p->pdev->dev, "cannot enable clock\n");
 		return ret;
 	}
 
@@ -194,8 +193,7 @@ static void sh_mtu2_clock_event_mode(enum clock_event_mode mode,
 
 	switch (mode) {
 	case CLOCK_EVT_MODE_PERIODIC:
-		pr_info("sh_mtu2: %s used for periodic clock events\n",
-			ced->name);
+		dev_info(&p->pdev->dev, "used for periodic clock events\n");
 		sh_mtu2_enable(p);
 		break;
 	case CLOCK_EVT_MODE_UNUSED:
@@ -222,13 +220,13 @@ static void sh_mtu2_register_clockevent(struct sh_mtu2_priv *p,
 	ced->cpumask = cpumask_of(0);
 	ced->set_mode = sh_mtu2_clock_event_mode;
 
-	pr_info("sh_mtu2: %s used for clock events\n", ced->name);
+	dev_info(&p->pdev->dev, "used for clock events\n");
 	clockevents_register_device(ced);
 
 	ret = setup_irq(p->irqaction.irq, &p->irqaction);
 	if (ret) {
-		pr_err("sh_mtu2: failed to request irq %d\n",
-		       p->irqaction.irq);
+		dev_err(&p->pdev->dev, "failed to request irq %d\n",
+			p->irqaction.irq);
 		return;
 	}
 }
@@ -274,26 +272,32 @@ static int sh_mtu2_setup(struct sh_mtu2_priv *p, struct platform_device *pdev)
 	/* map memory, let mapbase point to our channel */
 	p->mapbase = ioremap_nocache(res->start, resource_size(res));
 	if (p->mapbase == NULL) {
-		pr_err("sh_mtu2: failed to remap I/O memory\n");
+		dev_err(&p->pdev->dev, "failed to remap I/O memory\n");
 		goto err0;
 	}
 
 	/* setup data for setup_irq() (too early for request_irq()) */
-	p->irqaction.name = cfg->name;
+	p->irqaction.name = dev_name(&p->pdev->dev);
 	p->irqaction.handler = sh_mtu2_interrupt;
 	p->irqaction.dev_id = p;
 	p->irqaction.irq = irq;
-	p->irqaction.flags = IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL;
+	p->irqaction.flags = IRQF_DISABLED | IRQF_TIMER | \
+			     IRQF_IRQPOLL  | IRQF_NOBALANCING;
 
 	/* get hold of clock */
-	p->clk = clk_get(&p->pdev->dev, cfg->clk);
+	p->clk = clk_get(&p->pdev->dev, "mtu2_fck");
 	if (IS_ERR(p->clk)) {
-		pr_err("sh_mtu2: cannot get clock \"%s\"\n", cfg->clk);
-		ret = PTR_ERR(p->clk);
-		goto err1;
+		dev_warn(&p->pdev->dev, "using deprecated clock lookup\n");
+		p->clk = clk_get(&p->pdev->dev, cfg->clk);
+		if (IS_ERR(p->clk)) {
+			dev_err(&p->pdev->dev, "cannot get clock\n");
+			ret = PTR_ERR(p->clk);
+			goto err1;
+		}
 	}
 
-	return sh_mtu2_register(p, cfg->name, cfg->clockevent_rating);
+	return sh_mtu2_register(p, (char *)dev_name(&p->pdev->dev),
+				cfg->clockevent_rating);
  err1:
 	iounmap(p->mapbase);
  err0:
@@ -303,11 +307,10 @@ static int sh_mtu2_setup(struct sh_mtu2_priv *p, struct platform_device *pdev)
 static int __devinit sh_mtu2_probe(struct platform_device *pdev)
 {
 	struct sh_mtu2_priv *p = platform_get_drvdata(pdev);
-	struct sh_timer_config *cfg = pdev->dev.platform_data;
 	int ret;
 
 	if (p) {
-		pr_info("sh_mtu2: %s kept as earlytimer\n", cfg->name);
+		dev_info(&pdev->dev, "kept as earlytimer\n");
 		return 0;
 	}
 
diff --git a/drivers/clocksource/sh_tmu.c b/drivers/clocksource/sh_tmu.c
index fc9ff1e5b77..8e44e14ec4c 100644
--- a/drivers/clocksource/sh_tmu.c
+++ b/drivers/clocksource/sh_tmu.c
@@ -107,13 +107,12 @@ static void sh_tmu_start_stop_ch(struct sh_tmu_priv *p, int start)
 
 static int sh_tmu_enable(struct sh_tmu_priv *p)
 {
-	struct sh_timer_config *cfg = p->pdev->dev.platform_data;
 	int ret;
 
 	/* enable clock */
 	ret = clk_enable(p->clk);
 	if (ret) {
-		pr_err("sh_tmu: cannot enable clock \"%s\"\n", cfg->clk);
+		dev_err(&p->pdev->dev, "cannot enable clock\n");
 		return ret;
 	}
 
@@ -229,7 +228,7 @@ static int sh_tmu_register_clocksource(struct sh_tmu_priv *p,
 	cs->disable = sh_tmu_clocksource_disable;
 	cs->mask = CLOCKSOURCE_MASK(32);
 	cs->flags = CLOCK_SOURCE_IS_CONTINUOUS;
-	pr_info("sh_tmu: %s used as clock source\n", cs->name);
+	dev_info(&p->pdev->dev, "used as clock source\n");
 	clocksource_register(cs);
 	return 0;
 }
@@ -277,13 +276,11 @@ static void sh_tmu_clock_event_mode(enum clock_event_mode mode,
 
 	switch (mode) {
 	case CLOCK_EVT_MODE_PERIODIC:
-		pr_info("sh_tmu: %s used for periodic clock events\n",
-			ced->name);
+		dev_info(&p->pdev->dev, "used for periodic clock events\n");
 		sh_tmu_clock_event_start(p, 1);
 		break;
 	case CLOCK_EVT_MODE_ONESHOT:
-		pr_info("sh_tmu: %s used for oneshot clock events\n",
-			ced->name);
+		dev_info(&p->pdev->dev, "used for oneshot clock events\n");
 		sh_tmu_clock_event_start(p, 0);
 		break;
 	case CLOCK_EVT_MODE_UNUSED:
@@ -324,13 +321,13 @@ static void sh_tmu_register_clockevent(struct sh_tmu_priv *p,
 	ced->set_next_event = sh_tmu_clock_event_next;
 	ced->set_mode = sh_tmu_clock_event_mode;
 
-	pr_info("sh_tmu: %s used for clock events\n", ced->name);
+	dev_info(&p->pdev->dev, "used for clock events\n");
 	clockevents_register_device(ced);
 
 	ret = setup_irq(p->irqaction.irq, &p->irqaction);
 	if (ret) {
-		pr_err("sh_tmu: failed to request irq %d\n",
-		       p->irqaction.irq);
+		dev_err(&p->pdev->dev, "failed to request irq %d\n",
+			p->irqaction.irq);
 		return;
 	}
 }
@@ -379,26 +376,31 @@ static int sh_tmu_setup(struct sh_tmu_priv *p, struct platform_device *pdev)
 	/* map memory, let mapbase point to our channel */
 	p->mapbase = ioremap_nocache(res->start, resource_size(res));
 	if (p->mapbase == NULL) {
-		pr_err("sh_tmu: failed to remap I/O memory\n");
+		dev_err(&p->pdev->dev, "failed to remap I/O memory\n");
 		goto err0;
 	}
 
 	/* setup data for setup_irq() (too early for request_irq()) */
-	p->irqaction.name = cfg->name;
+	p->irqaction.name = dev_name(&p->pdev->dev);
 	p->irqaction.handler = sh_tmu_interrupt;
 	p->irqaction.dev_id = p;
 	p->irqaction.irq = irq;
-	p->irqaction.flags = IRQF_DISABLED | IRQF_TIMER | IRQF_IRQPOLL;
+	p->irqaction.flags = IRQF_DISABLED | IRQF_TIMER | \
+			     IRQF_IRQPOLL  | IRQF_NOBALANCING;
 
 	/* get hold of clock */
-	p->clk = clk_get(&p->pdev->dev, cfg->clk);
+	p->clk = clk_get(&p->pdev->dev, "tmu_fck");
 	if (IS_ERR(p->clk)) {
-		pr_err("sh_tmu: cannot get clock \"%s\"\n", cfg->clk);
-		ret = PTR_ERR(p->clk);
-		goto err1;
+		dev_warn(&p->pdev->dev, "using deprecated clock lookup\n");
+		p->clk = clk_get(&p->pdev->dev, cfg->clk);
+		if (IS_ERR(p->clk)) {
+			dev_err(&p->pdev->dev, "cannot get clock\n");
+			ret = PTR_ERR(p->clk);
+			goto err1;
+		}
 	}
 
-	return sh_tmu_register(p, cfg->name,
+	return sh_tmu_register(p, (char *)dev_name(&p->pdev->dev),
 			       cfg->clockevent_rating,
 			       cfg->clocksource_rating);
  err1:
@@ -410,11 +412,10 @@ static int sh_tmu_setup(struct sh_tmu_priv *p, struct platform_device *pdev)
 static int __devinit sh_tmu_probe(struct platform_device *pdev)
 {
 	struct sh_tmu_priv *p = platform_get_drvdata(pdev);
-	struct sh_timer_config *cfg = pdev->dev.platform_data;
 	int ret;
 
 	if (p) {
-		pr_info("sh_tmu: %s kept as earlytimer\n", cfg->name);
+		dev_info(&pdev->dev, "kept as earlytimer\n");
 		return 0;
 	}
 
diff --git a/drivers/dma/shdma.c b/drivers/dma/shdma.c
index 6f25a20de99..323afef7780 100644
--- a/drivers/dma/shdma.c
+++ b/drivers/dma/shdma.c
@@ -26,8 +26,7 @@
 #include <linux/dma-mapping.h>
 #include <linux/platform_device.h>
 #include <linux/pm_runtime.h>
-
-#include <asm/dmaengine.h>
+#include <linux/sh_dma.h>
 
 #include "shdma.h"
 
@@ -45,7 +44,7 @@ enum sh_dmae_desc_status {
 #define LOG2_DEFAULT_XFER_SIZE	2
 
 /* A bitmask with bits enough for enum sh_dmae_slave_chan_id */
-static unsigned long sh_dmae_slave_used[BITS_TO_LONGS(SHDMA_SLAVE_NUMBER)];
+static unsigned long sh_dmae_slave_used[BITS_TO_LONGS(SH_DMA_SLAVE_NUMBER)];
 
 static void sh_dmae_chan_ld_cleanup(struct sh_dmae_chan *sh_chan, bool all);
 
@@ -190,7 +189,7 @@ static int dmae_set_dmars(struct sh_dmae_chan *sh_chan, u16 val)
 	struct sh_dmae_device *shdev = container_of(sh_chan->common.device,
 						struct sh_dmae_device, common);
 	struct sh_dmae_pdata *pdata = shdev->pdata;
-	struct sh_dmae_channel *chan_pdata = &pdata->channel[sh_chan->id];
+	const struct sh_dmae_channel *chan_pdata = &pdata->channel[sh_chan->id];
 	u16 __iomem *addr = shdev->dmars + chan_pdata->dmars / sizeof(u16);
 	int shift = chan_pdata->dmars_bit;
 
@@ -266,8 +265,8 @@ static struct sh_desc *sh_dmae_get_desc(struct sh_dmae_chan *sh_chan)
 	return NULL;
 }
 
-static struct sh_dmae_slave_config *sh_dmae_find_slave(
-	struct sh_dmae_chan *sh_chan, enum sh_dmae_slave_chan_id slave_id)
+static const struct sh_dmae_slave_config *sh_dmae_find_slave(
+	struct sh_dmae_chan *sh_chan, struct sh_dmae_slave *param)
 {
 	struct dma_device *dma_dev = sh_chan->common.device;
 	struct sh_dmae_device *shdev = container_of(dma_dev,
@@ -275,11 +274,11 @@ static struct sh_dmae_slave_config *sh_dmae_find_slave(
 	struct sh_dmae_pdata *pdata = shdev->pdata;
 	int i;
 
-	if ((unsigned)slave_id >= SHDMA_SLAVE_NUMBER)
+	if (param->slave_id >= SH_DMA_SLAVE_NUMBER)
 		return NULL;
 
 	for (i = 0; i < pdata->slave_num; i++)
-		if (pdata->slave[i].slave_id == slave_id)
+		if (pdata->slave[i].slave_id == param->slave_id)
 			return pdata->slave + i;
 
 	return NULL;
@@ -299,9 +298,9 @@ static int sh_dmae_alloc_chan_resources(struct dma_chan *chan)
 	 * never runs concurrently with itself or free_chan_resources.
 	 */
 	if (param) {
-		struct sh_dmae_slave_config *cfg;
+		const struct sh_dmae_slave_config *cfg;
 
-		cfg = sh_dmae_find_slave(sh_chan, param->slave_id);
+		cfg = sh_dmae_find_slave(sh_chan, param);
 		if (!cfg) {
 			ret = -EINVAL;
 			goto efindslave;
@@ -574,12 +573,14 @@ static struct dma_async_tx_descriptor *sh_dmae_prep_slave_sg(
 {
 	struct sh_dmae_slave *param;
 	struct sh_dmae_chan *sh_chan;
+	dma_addr_t slave_addr;
 
 	if (!chan)
 		return NULL;
 
 	sh_chan = to_sh_chan(chan);
 	param = chan->private;
+	slave_addr = param->config->addr;
 
 	/* Someone calling slave DMA on a public channel? */
 	if (!param || !sg_len) {
@@ -592,7 +593,7 @@ static struct dma_async_tx_descriptor *sh_dmae_prep_slave_sg(
 	 * if (param != NULL), this is a successfully requested slave channel,
 	 * therefore param->config != NULL too.
 	 */
-	return sh_dmae_prep_sg(sh_chan, sgl, sg_len, &param->config->addr,
+	return sh_dmae_prep_sg(sh_chan, sgl, sg_len, &slave_addr,
 			       direction, flags);
 }
 
@@ -873,7 +874,7 @@ static int __devinit sh_dmae_chan_probe(struct sh_dmae_device *shdev, int id,
 					int irq, unsigned long flags)
 {
 	int err;
-	struct sh_dmae_channel *chan_pdata = &shdev->pdata->channel[id];
+	const struct sh_dmae_channel *chan_pdata = &shdev->pdata->channel[id];
 	struct platform_device *pdev = to_platform_device(shdev->common.dev);
 	struct sh_dmae_chan *new_sh_chan;
 
diff --git a/drivers/dma/shdma.h b/drivers/dma/shdma.h
index 153609a1e96..4021275a0a4 100644
--- a/drivers/dma/shdma.h
+++ b/drivers/dma/shdma.h
@@ -17,8 +17,8 @@
 #include <linux/interrupt.h>
 #include <linux/list.h>
 
-#include <asm/dmaengine.h>
-
+#define SH_DMAC_MAX_CHANNELS 6
+#define SH_DMA_SLAVE_NUMBER 256
 #define SH_DMA_TCR_MAX 0x00FFFFFF	/* 16MB */
 
 struct device;
diff --git a/drivers/serial/sh-sci.c b/drivers/serial/sh-sci.c
index 8eb094c1f61..8d993c4ccea 100644
--- a/drivers/serial/sh-sci.c
+++ b/drivers/serial/sh-sci.c
@@ -83,16 +83,16 @@ struct sci_port {
 
 	/* Interface clock */
 	struct clk		*iclk;
-	/* Data clock */
-	struct clk		*dclk;
+	/* Function clock */
+	struct clk		*fclk;
 
 	struct list_head	node;
 	struct dma_chan			*chan_tx;
 	struct dma_chan			*chan_rx;
 #ifdef CONFIG_SERIAL_SH_SCI_DMA
 	struct device			*dma_dev;
-	enum sh_dmae_slave_chan_id	slave_tx;
-	enum sh_dmae_slave_chan_id	slave_rx;
+	unsigned int			slave_tx;
+	unsigned int			slave_rx;
 	struct dma_async_tx_descriptor	*desc_tx;
 	struct dma_async_tx_descriptor	*desc_rx[2];
 	dma_cookie_t			cookie_tx;
@@ -107,6 +107,7 @@ struct sci_port {
 	struct work_struct		work_tx;
 	struct work_struct		work_rx;
 	struct timer_list		rx_timer;
+	unsigned int			rx_timeout;
 #endif
 };
 
@@ -674,22 +675,22 @@ static irqreturn_t sci_rx_interrupt(int irq, void *ptr)
 	struct sci_port *s = to_sci_port(port);
 
 	if (s->chan_rx) {
-		unsigned long tout;
 		u16 scr = sci_in(port, SCSCR);
 		u16 ssr = sci_in(port, SCxSR);
 
 		/* Disable future Rx interrupts */
-		sci_out(port, SCSCR, scr & ~SCI_CTRL_FLAGS_RIE);
+		if (port->type == PORT_SCIFA) {
+			disable_irq_nosync(irq);
+			scr |= 0x4000;
+		} else {
+			scr &= ~SCI_CTRL_FLAGS_RIE;
+		}
+		sci_out(port, SCSCR, scr);
 		/* Clear current interrupt */
 		sci_out(port, SCxSR, ssr & ~(1 | SCxSR_RDxF(port)));
-		/* Calculate delay for 1.5 DMA buffers */
-		tout = (port->timeout - HZ / 50) * s->buf_len_rx * 3 /
-			port->fifosize / 2;
-		dev_dbg(port->dev, "Rx IRQ: setup timeout in %lu ms\n",
-			tout * 1000 / HZ);
-		if (tout < 2)
-			tout = 2;
-		mod_timer(&s->rx_timer, jiffies + tout);
+		dev_dbg(port->dev, "Rx IRQ %lu: setup t-out in %u jiffies\n",
+			jiffies, s->rx_timeout);
+		mod_timer(&s->rx_timer, jiffies + s->rx_timeout);
 
 		return IRQ_HANDLED;
 	}
@@ -799,7 +800,7 @@ static int sci_notifier(struct notifier_block *self,
 	    (phase == CPUFREQ_RESUMECHANGE)) {
 		spin_lock_irqsave(&priv->lock, flags);
 		list_for_each_entry(sci_port, &priv->ports, node)
-			sci_port->port.uartclk = clk_get_rate(sci_port->dclk);
+			sci_port->port.uartclk = clk_get_rate(sci_port->iclk);
 		spin_unlock_irqrestore(&priv->lock, flags);
 	}
 
@@ -810,21 +811,17 @@ static void sci_clk_enable(struct uart_port *port)
 {
 	struct sci_port *sci_port = to_sci_port(port);
 
-	clk_enable(sci_port->dclk);
-	sci_port->port.uartclk = clk_get_rate(sci_port->dclk);
-
-	if (sci_port->iclk)
-		clk_enable(sci_port->iclk);
+	clk_enable(sci_port->iclk);
+	sci_port->port.uartclk = clk_get_rate(sci_port->iclk);
+	clk_enable(sci_port->fclk);
 }
 
 static void sci_clk_disable(struct uart_port *port)
 {
 	struct sci_port *sci_port = to_sci_port(port);
 
-	if (sci_port->iclk)
-		clk_disable(sci_port->iclk);
-
-	clk_disable(sci_port->dclk);
+	clk_disable(sci_port->fclk);
+	clk_disable(sci_port->iclk);
 }
 
 static int sci_request_irq(struct sci_port *port)
@@ -913,22 +910,26 @@ static void sci_dma_tx_complete(void *arg)
 
 	spin_lock_irqsave(&port->lock, flags);
 
-	xmit->tail += s->sg_tx.length;
+	xmit->tail += sg_dma_len(&s->sg_tx);
 	xmit->tail &= UART_XMIT_SIZE - 1;
 
-	port->icount.tx += s->sg_tx.length;
+	port->icount.tx += sg_dma_len(&s->sg_tx);
 
 	async_tx_ack(s->desc_tx);
 	s->cookie_tx = -EINVAL;
 	s->desc_tx = NULL;
 
-	spin_unlock_irqrestore(&port->lock, flags);
-
 	if (uart_circ_chars_pending(xmit) < WAKEUP_CHARS)
 		uart_write_wakeup(port);
 
-	if (uart_circ_chars_pending(xmit))
+	if (!uart_circ_empty(xmit)) {
 		schedule_work(&s->work_tx);
+	} else if (port->type == PORT_SCIFA) {
+		u16 ctrl = sci_in(port, SCSCR);
+		sci_out(port, SCSCR, ctrl & ~SCI_CTRL_FLAGS_TIE);
+	}
+
+	spin_unlock_irqrestore(&port->lock, flags);
 }
 
 /* Locking: called with port lock held */
@@ -972,13 +973,13 @@ static void sci_dma_rx_complete(void *arg)
 	unsigned long flags;
 	int count;
 
-	dev_dbg(port->dev, "%s(%d)\n", __func__, port->line);
+	dev_dbg(port->dev, "%s(%d) active #%d\n", __func__, port->line, s->active_rx);
 
 	spin_lock_irqsave(&port->lock, flags);
 
 	count = sci_dma_rx_push(s, tty, s->buf_len_rx);
 
-	mod_timer(&s->rx_timer, jiffies + msecs_to_jiffies(5));
+	mod_timer(&s->rx_timer, jiffies + s->rx_timeout);
 
 	spin_unlock_irqrestore(&port->lock, flags);
 
@@ -1050,6 +1051,8 @@ static void sci_submit_rx(struct sci_port *s)
 			sci_rx_dma_release(s, true);
 			return;
 		}
+		dev_dbg(s->port.dev, "%s(): cookie %d to #%d\n", __func__,
+			s->cookie_rx[i], i);
 	}
 
 	s->active_rx = s->cookie_rx[0];
@@ -1107,10 +1110,10 @@ static void work_fn_rx(struct work_struct *work)
 		return;
 	}
 
-	dev_dbg(port->dev, "%s: cookie %d #%d\n", __func__,
-		s->cookie_rx[new], new);
-
 	s->active_rx = s->cookie_rx[!new];
+
+	dev_dbg(port->dev, "%s: cookie %d #%d, new active #%d\n", __func__,
+		s->cookie_rx[new], new, s->active_rx);
 }
 
 static void work_fn_tx(struct work_struct *work)
@@ -1131,14 +1134,13 @@ static void work_fn_tx(struct work_struct *work)
 	 */
 	spin_lock_irq(&port->lock);
 	sg->offset = xmit->tail & (UART_XMIT_SIZE - 1);
-	sg->dma_address = (sg_dma_address(sg) & ~(UART_XMIT_SIZE - 1)) +
+	sg_dma_address(sg) = (sg_dma_address(sg) & ~(UART_XMIT_SIZE - 1)) +
 		sg->offset;
-	sg->length = min((int)CIRC_CNT(xmit->head, xmit->tail, UART_XMIT_SIZE),
+	sg_dma_len(sg) = min((int)CIRC_CNT(xmit->head, xmit->tail, UART_XMIT_SIZE),
 		CIRC_CNT_TO_END(xmit->head, xmit->tail, UART_XMIT_SIZE));
-	sg->dma_length = sg->length;
 	spin_unlock_irq(&port->lock);
 
-	BUG_ON(!sg->length);
+	BUG_ON(!sg_dma_len(sg));
 
 	desc = chan->device->device_prep_slave_sg(chan,
 			sg, s->sg_len_tx, DMA_TO_DEVICE,
@@ -1173,23 +1175,28 @@ static void work_fn_tx(struct work_struct *work)
 
 static void sci_start_tx(struct uart_port *port)
 {
+	struct sci_port *s = to_sci_port(port);
 	unsigned short ctrl;
 
 #ifdef CONFIG_SERIAL_SH_SCI_DMA
-	struct sci_port *s = to_sci_port(port);
-
-	if (s->chan_tx) {
-		if (!uart_circ_empty(&s->port.state->xmit) && s->cookie_tx < 0)
-			schedule_work(&s->work_tx);
-
-		return;
+	if (port->type == PORT_SCIFA) {
+		u16 new, scr = sci_in(port, SCSCR);
+		if (s->chan_tx)
+			new = scr | 0x8000;
+		else
+			new = scr & ~0x8000;
+		if (new != scr)
+			sci_out(port, SCSCR, new);
 	}
+	if (s->chan_tx && !uart_circ_empty(&s->port.state->xmit) &&
+	    s->cookie_tx < 0)
+		schedule_work(&s->work_tx);
 #endif
-
-	/* Set TIE (Transmit Interrupt Enable) bit in SCSCR */
-	ctrl = sci_in(port, SCSCR);
-	ctrl |= SCI_CTRL_FLAGS_TIE;
-	sci_out(port, SCSCR, ctrl);
+	if (!s->chan_tx || port->type == PORT_SCIFA) {
+		/* Set TIE (Transmit Interrupt Enable) bit in SCSCR */
+		ctrl = sci_in(port, SCSCR);
+		sci_out(port, SCSCR, ctrl | SCI_CTRL_FLAGS_TIE);
+	}
 }
 
 static void sci_stop_tx(struct uart_port *port)
@@ -1198,6 +1205,8 @@ static void sci_stop_tx(struct uart_port *port)
 
 	/* Clear TIE (Transmit Interrupt Enable) bit in SCSCR */
 	ctrl = sci_in(port, SCSCR);
+	if (port->type == PORT_SCIFA)
+		ctrl &= ~0x8000;
 	ctrl &= ~SCI_CTRL_FLAGS_TIE;
 	sci_out(port, SCSCR, ctrl);
 }
@@ -1208,6 +1217,8 @@ static void sci_start_rx(struct uart_port *port)
 
 	/* Set RIE (Receive Interrupt Enable) bit in SCSCR */
 	ctrl |= sci_in(port, SCSCR);
+	if (port->type == PORT_SCIFA)
+		ctrl &= ~0x4000;
 	sci_out(port, SCSCR, ctrl);
 }
 
@@ -1217,6 +1228,8 @@ static void sci_stop_rx(struct uart_port *port)
 
 	/* Clear RIE (Receive Interrupt Enable) bit in SCSCR */
 	ctrl = sci_in(port, SCSCR);
+	if (port->type == PORT_SCIFA)
+		ctrl &= ~0x4000;
 	ctrl &= ~(SCI_CTRL_FLAGS_RIE | SCI_CTRL_FLAGS_REIE);
 	sci_out(port, SCSCR, ctrl);
 }
@@ -1251,8 +1264,12 @@ static void rx_timer_fn(unsigned long arg)
 {
 	struct sci_port *s = (struct sci_port *)arg;
 	struct uart_port *port = &s->port;
-
 	u16 scr = sci_in(port, SCSCR);
+
+	if (port->type == PORT_SCIFA) {
+		scr &= ~0x4000;
+		enable_irq(s->irqs[1]);
+	}
 	sci_out(port, SCSCR, scr | SCI_CTRL_FLAGS_RIE);
 	dev_dbg(port->dev, "DMA Rx timed out\n");
 	schedule_work(&s->work_rx);
@@ -1339,8 +1356,7 @@ static void sci_request_dma(struct uart_port *port)
 			sg_init_table(sg, 1);
 			sg_set_page(sg, virt_to_page(buf[i]), s->buf_len_rx,
 				    (int)buf[i] & ~PAGE_MASK);
-			sg->dma_address = dma[i];
-			sg->dma_length = sg->length;
+			sg_dma_address(sg) = dma[i];
 		}
 
 		INIT_WORK(&s->work_rx, work_fn_rx);
@@ -1403,8 +1419,12 @@ static void sci_shutdown(struct uart_port *port)
 static void sci_set_termios(struct uart_port *port, struct ktermios *termios,
 			    struct ktermios *old)
 {
+#ifdef CONFIG_SERIAL_SH_SCI_DMA
+	struct sci_port *s = to_sci_port(port);
+#endif
 	unsigned int status, baud, smr_val, max_baud;
 	int t = -1;
+	u16 scfcr = 0;
 
 	/*
 	 * earlyprintk comes here early on with port->uartclk set to zero.
@@ -1427,7 +1447,7 @@ static void sci_set_termios(struct uart_port *port, struct ktermios *termios,
 	sci_out(port, SCSCR, 0x00);	/* TE=0, RE=0, CKE1=0 */
 
 	if (port->type != PORT_SCI)
-		sci_out(port, SCFCR, SCFCR_RFRST | SCFCR_TFRST);
+		sci_out(port, SCFCR, scfcr | SCFCR_RFRST | SCFCR_TFRST);
 
 	smr_val = sci_in(port, SCSMR) & 3;
 	if ((termios->c_cflag & CSIZE) == CS7)
@@ -1458,10 +1478,32 @@ static void sci_set_termios(struct uart_port *port, struct ktermios *termios,
 	}
 
 	sci_init_pins(port, termios->c_cflag);
-	sci_out(port, SCFCR, (termios->c_cflag & CRTSCTS) ? SCFCR_MCE : 0);
+	sci_out(port, SCFCR, scfcr | ((termios->c_cflag & CRTSCTS) ? SCFCR_MCE : 0));
 
 	sci_out(port, SCSCR, SCSCR_INIT(port));
 
+#ifdef CONFIG_SERIAL_SH_SCI_DMA
+	/*
+	 * Calculate delay for 1.5 DMA buffers: see
+	 * drivers/serial/serial_core.c::uart_update_timeout(). With 10 bits
+	 * (CS8), 250Hz, 115200 baud and 64 bytes FIFO, the above function
+	 * calculates 1 jiffie for the data plus 5 jiffies for the "slop(e)."
+	 * Then below we calculate 3 jiffies (12ms) for 1.5 DMA buffers (3 FIFO
+	 * sizes), but it has been found out experimentally, that this is not
+	 * enough: the driver too often needlessly runs on a DMA timeout. 20ms
+	 * as a minimum seem to work perfectly.
+	 */
+	if (s->chan_rx) {
+		s->rx_timeout = (port->timeout - HZ / 50) * s->buf_len_rx * 3 /
+			port->fifosize / 2;
+		dev_dbg(port->dev,
+			"DMA Rx t-out %ums, tty t-out %u jiffies\n",
+			s->rx_timeout * 1000 / HZ, port->timeout);
+		if (s->rx_timeout < msecs_to_jiffies(20))
+			s->rx_timeout = msecs_to_jiffies(20);
+	}
+#endif
+
 	if ((termios->c_cflag & CREAD) != 0)
 		sci_start_rx(port);
 }
@@ -1553,10 +1595,10 @@ static struct uart_ops sci_uart_ops = {
 #endif
 };
 
-static void __devinit sci_init_single(struct platform_device *dev,
-				      struct sci_port *sci_port,
-				      unsigned int index,
-				      struct plat_sci_port *p)
+static int __devinit sci_init_single(struct platform_device *dev,
+				     struct sci_port *sci_port,
+				     unsigned int index,
+				     struct plat_sci_port *p)
 {
 	struct uart_port *port = &sci_port->port;
 
@@ -1577,8 +1619,23 @@ static void __devinit sci_init_single(struct platform_device *dev,
 	}
 
 	if (dev) {
-		sci_port->iclk = p->clk ? clk_get(&dev->dev, p->clk) : NULL;
-		sci_port->dclk = clk_get(&dev->dev, "peripheral_clk");
+		sci_port->iclk = clk_get(&dev->dev, "sci_ick");
+		if (IS_ERR(sci_port->iclk)) {
+			sci_port->iclk = clk_get(&dev->dev, "peripheral_clk");
+			if (IS_ERR(sci_port->iclk)) {
+				dev_err(&dev->dev, "can't get iclk\n");
+				return PTR_ERR(sci_port->iclk);
+			}
+		}
+
+		/*
+		 * The function clock is optional, ignore it if we can't
+		 * find it.
+		 */
+		sci_port->fclk = clk_get(&dev->dev, "sci_fck");
+		if (IS_ERR(sci_port->fclk))
+			sci_port->fclk = NULL;
+
 		sci_port->enable = sci_clk_enable;
 		sci_port->disable = sci_clk_disable;
 		port->dev = &dev->dev;
@@ -1605,6 +1662,7 @@ static void __devinit sci_init_single(struct platform_device *dev,
 #endif
 
 	memcpy(&sci_port->irqs, &p->irqs, sizeof(p->irqs));
+	return 0;
 }
 
 #ifdef CONFIG_SERIAL_SH_SCI_CONSOLE
@@ -1754,8 +1812,11 @@ static int sci_remove(struct platform_device *dev)
 	cpufreq_unregister_notifier(&priv->clk_nb, CPUFREQ_TRANSITION_NOTIFIER);
 
 	spin_lock_irqsave(&priv->lock, flags);
-	list_for_each_entry(p, &priv->ports, node)
+	list_for_each_entry(p, &priv->ports, node) {
 		uart_remove_one_port(&sci_uart_driver, &p->port);
+		clk_put(p->iclk);
+		clk_put(p->fclk);
+	}
 	spin_unlock_irqrestore(&priv->lock, flags);
 
 	kfree(priv);
@@ -1781,7 +1842,9 @@ static int __devinit sci_probe_single(struct platform_device *dev,
 		return 0;
 	}
 
-	sci_init_single(dev, sciport, index, p);
+	ret = sci_init_single(dev, sciport, index, p);
+	if (ret)
+		return ret;
 
 	ret = uart_add_one_port(&sci_uart_driver, &sciport->port);
 	if (ret)
diff --git a/drivers/sh/Kconfig b/drivers/sh/Kconfig
new file mode 100644
index 00000000000..a54de0b9b3d
--- /dev/null
+++ b/drivers/sh/Kconfig
@@ -0,0 +1,24 @@
+config INTC_USERIMASK
+	bool "Userspace interrupt masking support"
+	depends on ARCH_SHMOBILE || (SUPERH && CPU_SH4A)
+	help
+	  This enables support for hardware-assisted userspace hardirq
+	  masking.
+
+	  SH-4A and newer interrupt blocks all support a special shadowed
+	  page with all non-masking registers obscured when mapped in to
+	  userspace. This is primarily for use by userspace device
+	  drivers that are using special priority levels.
+
+	  If in doubt, say N.
+
+config INTC_BALANCING
+	bool "Hardware IRQ balancing support"
+	depends on SMP && SUPERH && CPU_SUBTYPE_SH7786
+	help
+	  This enables support for IRQ auto-distribution mode on SH-X3
+	  SMP parts. All of the balancing and CPU wakeup decisions are
+	  taken care of automatically by hardware for distributed
+	  vectors.
+
+	  If in doubt, say N.
diff --git a/drivers/sh/Makefile b/drivers/sh/Makefile
index 4956bf1f213..78bb5127abd 100644
--- a/drivers/sh/Makefile
+++ b/drivers/sh/Makefile
@@ -4,4 +4,6 @@
 obj-$(CONFIG_SUPERHYWAY)	+= superhyway/
 obj-$(CONFIG_MAPLE)		+= maple/
 obj-$(CONFIG_GENERIC_GPIO)	+= pfc.o
+obj-$(CONFIG_SUPERH)		+= clk.o
+obj-$(CONFIG_SH_CLK_CPG)	+= clk-cpg.o
 obj-y				+= intc.o
diff --git a/drivers/sh/clk-cpg.c b/drivers/sh/clk-cpg.c
new file mode 100644
index 00000000000..f5c80ba9ab1
--- /dev/null
+++ b/drivers/sh/clk-cpg.c
@@ -0,0 +1,298 @@
+#include <linux/clk.h>
+#include <linux/compiler.h>
+#include <linux/slab.h>
+#include <linux/io.h>
+#include <linux/sh_clk.h>
+
+static int sh_clk_mstp32_enable(struct clk *clk)
+{
+	__raw_writel(__raw_readl(clk->enable_reg) & ~(1 << clk->enable_bit),
+		     clk->enable_reg);
+	return 0;
+}
+
+static void sh_clk_mstp32_disable(struct clk *clk)
+{
+	__raw_writel(__raw_readl(clk->enable_reg) | (1 << clk->enable_bit),
+		     clk->enable_reg);
+}
+
+static struct clk_ops sh_clk_mstp32_clk_ops = {
+	.enable		= sh_clk_mstp32_enable,
+	.disable	= sh_clk_mstp32_disable,
+	.recalc		= followparent_recalc,
+};
+
+int __init sh_clk_mstp32_register(struct clk *clks, int nr)
+{
+	struct clk *clkp;
+	int ret = 0;
+	int k;
+
+	for (k = 0; !ret && (k < nr); k++) {
+		clkp = clks + k;
+		clkp->ops = &sh_clk_mstp32_clk_ops;
+		ret |= clk_register(clkp);
+	}
+
+	return ret;
+}
+
+static long sh_clk_div_round_rate(struct clk *clk, unsigned long rate)
+{
+	return clk_rate_table_round(clk, clk->freq_table, rate);
+}
+
+static int sh_clk_div6_divisors[64] = {
+	1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16,
+	17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32,
+	33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48,
+	49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64
+};
+
+static struct clk_div_mult_table sh_clk_div6_table = {
+	.divisors = sh_clk_div6_divisors,
+	.nr_divisors = ARRAY_SIZE(sh_clk_div6_divisors),
+};
+
+static unsigned long sh_clk_div6_recalc(struct clk *clk)
+{
+	struct clk_div_mult_table *table = &sh_clk_div6_table;
+	unsigned int idx;
+
+	clk_rate_table_build(clk, clk->freq_table, table->nr_divisors,
+			     table, NULL);
+
+	idx = __raw_readl(clk->enable_reg) & 0x003f;
+
+	return clk->freq_table[idx].frequency;
+}
+
+static int sh_clk_div6_set_rate(struct clk *clk,
+				unsigned long rate, int algo_id)
+{
+	unsigned long value;
+	int idx;
+
+	idx = clk_rate_table_find(clk, clk->freq_table, rate);
+	if (idx < 0)
+		return idx;
+
+	value = __raw_readl(clk->enable_reg);
+	value &= ~0x3f;
+	value |= idx;
+	__raw_writel(value, clk->enable_reg);
+	return 0;
+}
+
+static int sh_clk_div6_enable(struct clk *clk)
+{
+	unsigned long value;
+	int ret;
+
+	ret = sh_clk_div6_set_rate(clk, clk->rate, 0);
+	if (ret == 0) {
+		value = __raw_readl(clk->enable_reg);
+		value &= ~0x100; /* clear stop bit to enable clock */
+		__raw_writel(value, clk->enable_reg);
+	}
+	return ret;
+}
+
+static void sh_clk_div6_disable(struct clk *clk)
+{
+	unsigned long value;
+
+	value = __raw_readl(clk->enable_reg);
+	value |= 0x100; /* stop clock */
+	value |= 0x3f; /* VDIV bits must be non-zero, overwrite divider */
+	__raw_writel(value, clk->enable_reg);
+}
+
+static struct clk_ops sh_clk_div6_clk_ops = {
+	.recalc		= sh_clk_div6_recalc,
+	.round_rate	= sh_clk_div_round_rate,
+	.set_rate	= sh_clk_div6_set_rate,
+	.enable		= sh_clk_div6_enable,
+	.disable	= sh_clk_div6_disable,
+};
+
+int __init sh_clk_div6_register(struct clk *clks, int nr)
+{
+	struct clk *clkp;
+	void *freq_table;
+	int nr_divs = sh_clk_div6_table.nr_divisors;
+	int freq_table_size = sizeof(struct cpufreq_frequency_table);
+	int ret = 0;
+	int k;
+
+	freq_table_size *= (nr_divs + 1);
+	freq_table = kzalloc(freq_table_size * nr, GFP_KERNEL);
+	if (!freq_table) {
+		pr_err("sh_clk_div6_register: unable to alloc memory\n");
+		return -ENOMEM;
+	}
+
+	for (k = 0; !ret && (k < nr); k++) {
+		clkp = clks + k;
+
+		clkp->ops = &sh_clk_div6_clk_ops;
+		clkp->id = -1;
+		clkp->freq_table = freq_table + (k * freq_table_size);
+		clkp->freq_table[nr_divs].frequency = CPUFREQ_TABLE_END;
+
+		ret = clk_register(clkp);
+	}
+
+	return ret;
+}
+
+static unsigned long sh_clk_div4_recalc(struct clk *clk)
+{
+	struct clk_div4_table *d4t = clk->priv;
+	struct clk_div_mult_table *table = d4t->div_mult_table;
+	unsigned int idx;
+
+	clk_rate_table_build(clk, clk->freq_table, table->nr_divisors,
+			     table, &clk->arch_flags);
+
+	idx = (__raw_readl(clk->enable_reg) >> clk->enable_bit) & 0x000f;
+
+	return clk->freq_table[idx].frequency;
+}
+
+static int sh_clk_div4_set_parent(struct clk *clk, struct clk *parent)
+{
+	struct clk_div4_table *d4t = clk->priv;
+	struct clk_div_mult_table *table = d4t->div_mult_table;
+	u32 value;
+	int ret;
+
+	/* we really need a better way to determine parent index, but for
+	 * now assume internal parent comes with CLK_ENABLE_ON_INIT set,
+	 * no CLK_ENABLE_ON_INIT means external clock...
+	 */
+
+	if (parent->flags & CLK_ENABLE_ON_INIT)
+		value = __raw_readl(clk->enable_reg) & ~(1 << 7);
+	else
+		value = __raw_readl(clk->enable_reg) | (1 << 7);
+
+	ret = clk_reparent(clk, parent);
+	if (ret < 0)
+		return ret;
+
+	__raw_writel(value, clk->enable_reg);
+
+	/* Rebiuld the frequency table */
+	clk_rate_table_build(clk, clk->freq_table, table->nr_divisors,
+			     table, &clk->arch_flags);
+
+	return 0;
+}
+
+static int sh_clk_div4_set_rate(struct clk *clk, unsigned long rate, int algo_id)
+{
+	struct clk_div4_table *d4t = clk->priv;
+	unsigned long value;
+	int idx = clk_rate_table_find(clk, clk->freq_table, rate);
+	if (idx < 0)
+		return idx;
+
+	value = __raw_readl(clk->enable_reg);
+	value &= ~(0xf << clk->enable_bit);
+	value |= (idx << clk->enable_bit);
+	__raw_writel(value, clk->enable_reg);
+
+	if (d4t->kick)
+		d4t->kick(clk);
+
+	return 0;
+}
+
+static int sh_clk_div4_enable(struct clk *clk)
+{
+	__raw_writel(__raw_readl(clk->enable_reg) & ~(1 << 8), clk->enable_reg);
+	return 0;
+}
+
+static void sh_clk_div4_disable(struct clk *clk)
+{
+	__raw_writel(__raw_readl(clk->enable_reg) | (1 << 8), clk->enable_reg);
+}
+
+static struct clk_ops sh_clk_div4_clk_ops = {
+	.recalc		= sh_clk_div4_recalc,
+	.set_rate	= sh_clk_div4_set_rate,
+	.round_rate	= sh_clk_div_round_rate,
+};
+
+static struct clk_ops sh_clk_div4_enable_clk_ops = {
+	.recalc		= sh_clk_div4_recalc,
+	.set_rate	= sh_clk_div4_set_rate,
+	.round_rate	= sh_clk_div_round_rate,
+	.enable		= sh_clk_div4_enable,
+	.disable	= sh_clk_div4_disable,
+};
+
+static struct clk_ops sh_clk_div4_reparent_clk_ops = {
+	.recalc		= sh_clk_div4_recalc,
+	.set_rate	= sh_clk_div4_set_rate,
+	.round_rate	= sh_clk_div_round_rate,
+	.enable		= sh_clk_div4_enable,
+	.disable	= sh_clk_div4_disable,
+	.set_parent	= sh_clk_div4_set_parent,
+};
+
+static int __init sh_clk_div4_register_ops(struct clk *clks, int nr,
+			struct clk_div4_table *table, struct clk_ops *ops)
+{
+	struct clk *clkp;
+	void *freq_table;
+	int nr_divs = table->div_mult_table->nr_divisors;
+	int freq_table_size = sizeof(struct cpufreq_frequency_table);
+	int ret = 0;
+	int k;
+
+	freq_table_size *= (nr_divs + 1);
+	freq_table = kzalloc(freq_table_size * nr, GFP_KERNEL);
+	if (!freq_table) {
+		pr_err("sh_clk_div4_register: unable to alloc memory\n");
+		return -ENOMEM;
+	}
+
+	for (k = 0; !ret && (k < nr); k++) {
+		clkp = clks + k;
+
+		clkp->ops = ops;
+		clkp->id = -1;
+		clkp->priv = table;
+
+		clkp->freq_table = freq_table + (k * freq_table_size);
+		clkp->freq_table[nr_divs].frequency = CPUFREQ_TABLE_END;
+
+		ret = clk_register(clkp);
+	}
+
+	return ret;
+}
+
+int __init sh_clk_div4_register(struct clk *clks, int nr,
+				struct clk_div4_table *table)
+{
+	return sh_clk_div4_register_ops(clks, nr, table, &sh_clk_div4_clk_ops);
+}
+
+int __init sh_clk_div4_enable_register(struct clk *clks, int nr,
+				struct clk_div4_table *table)
+{
+	return sh_clk_div4_register_ops(clks, nr, table,
+					&sh_clk_div4_enable_clk_ops);
+}
+
+int __init sh_clk_div4_reparent_register(struct clk *clks, int nr,
+				struct clk_div4_table *table)
+{
+	return sh_clk_div4_register_ops(clks, nr, table,
+					&sh_clk_div4_reparent_clk_ops);
+}
diff --git a/drivers/sh/clk.c b/drivers/sh/clk.c
new file mode 100644
index 00000000000..5d84adac9ec
--- /dev/null
+++ b/drivers/sh/clk.c
@@ -0,0 +1,545 @@
+/*
+ * drivers/sh/clk.c - SuperH clock framework
+ *
+ *  Copyright (C) 2005 - 2009  Paul Mundt
+ *
+ * This clock framework is derived from the OMAP version by:
+ *
+ *	Copyright (C) 2004 - 2008 Nokia Corporation
+ *	Written by Tuukka Tikkanen <tuukka.tikkanen@elektrobit.com>
+ *
+ *  Modified for omap shared clock framework by Tony Lindgren <tony@atomide.com>
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License.  See the file "COPYING" in the main directory of this archive
+ * for more details.
+ */
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/mutex.h>
+#include <linux/list.h>
+#include <linux/kobject.h>
+#include <linux/sysdev.h>
+#include <linux/seq_file.h>
+#include <linux/err.h>
+#include <linux/platform_device.h>
+#include <linux/debugfs.h>
+#include <linux/cpufreq.h>
+#include <linux/clk.h>
+#include <linux/sh_clk.h>
+
+static LIST_HEAD(clock_list);
+static DEFINE_SPINLOCK(clock_lock);
+static DEFINE_MUTEX(clock_list_sem);
+
+void clk_rate_table_build(struct clk *clk,
+			  struct cpufreq_frequency_table *freq_table,
+			  int nr_freqs,
+			  struct clk_div_mult_table *src_table,
+			  unsigned long *bitmap)
+{
+	unsigned long mult, div;
+	unsigned long freq;
+	int i;
+
+	for (i = 0; i < nr_freqs; i++) {
+		div = 1;
+		mult = 1;
+
+		if (src_table->divisors && i < src_table->nr_divisors)
+			div = src_table->divisors[i];
+
+		if (src_table->multipliers && i < src_table->nr_multipliers)
+			mult = src_table->multipliers[i];
+
+		if (!div || !mult || (bitmap && !test_bit(i, bitmap)))
+			freq = CPUFREQ_ENTRY_INVALID;
+		else
+			freq = clk->parent->rate * mult / div;
+
+		freq_table[i].index = i;
+		freq_table[i].frequency = freq;
+	}
+
+	/* Termination entry */
+	freq_table[i].index = i;
+	freq_table[i].frequency = CPUFREQ_TABLE_END;
+}
+
+long clk_rate_table_round(struct clk *clk,
+			  struct cpufreq_frequency_table *freq_table,
+			  unsigned long rate)
+{
+	unsigned long rate_error, rate_error_prev = ~0UL;
+	unsigned long rate_best_fit = rate;
+	unsigned long highest, lowest;
+	int i;
+
+	highest = lowest = 0;
+
+	for (i = 0; freq_table[i].frequency != CPUFREQ_TABLE_END; i++) {
+		unsigned long freq = freq_table[i].frequency;
+
+		if (freq == CPUFREQ_ENTRY_INVALID)
+			continue;
+
+		if (freq > highest)
+			highest = freq;
+		if (freq < lowest)
+			lowest = freq;
+
+		rate_error = abs(freq - rate);
+		if (rate_error < rate_error_prev) {
+			rate_best_fit = freq;
+			rate_error_prev = rate_error;
+		}
+
+		if (rate_error == 0)
+			break;
+	}
+
+	if (rate >= highest)
+		rate_best_fit = highest;
+	if (rate <= lowest)
+		rate_best_fit = lowest;
+
+	return rate_best_fit;
+}
+
+int clk_rate_table_find(struct clk *clk,
+			struct cpufreq_frequency_table *freq_table,
+			unsigned long rate)
+{
+	int i;
+
+	for (i = 0; freq_table[i].frequency != CPUFREQ_TABLE_END; i++) {
+		unsigned long freq = freq_table[i].frequency;
+
+		if (freq == CPUFREQ_ENTRY_INVALID)
+			continue;
+
+		if (freq == rate)
+			return i;
+	}
+
+	return -ENOENT;
+}
+
+/* Used for clocks that always have same value as the parent clock */
+unsigned long followparent_recalc(struct clk *clk)
+{
+	return clk->parent ? clk->parent->rate : 0;
+}
+
+int clk_reparent(struct clk *child, struct clk *parent)
+{
+	list_del_init(&child->sibling);
+	if (parent)
+		list_add(&child->sibling, &parent->children);
+	child->parent = parent;
+
+	/* now do the debugfs renaming to reattach the child
+	   to the proper parent */
+
+	return 0;
+}
+
+/* Propagate rate to children */
+void propagate_rate(struct clk *tclk)
+{
+	struct clk *clkp;
+
+	list_for_each_entry(clkp, &tclk->children, sibling) {
+		if (clkp->ops && clkp->ops->recalc)
+			clkp->rate = clkp->ops->recalc(clkp);
+
+		propagate_rate(clkp);
+	}
+}
+
+static void __clk_disable(struct clk *clk)
+{
+	if (WARN(!clk->usecount, "Trying to disable clock %s with 0 usecount\n",
+		 clk->name))
+		return;
+
+	if (!(--clk->usecount)) {
+		if (likely(clk->ops && clk->ops->disable))
+			clk->ops->disable(clk);
+		if (likely(clk->parent))
+			__clk_disable(clk->parent);
+	}
+}
+
+void clk_disable(struct clk *clk)
+{
+	unsigned long flags;
+
+	if (!clk)
+		return;
+
+	spin_lock_irqsave(&clock_lock, flags);
+	__clk_disable(clk);
+	spin_unlock_irqrestore(&clock_lock, flags);
+}
+EXPORT_SYMBOL_GPL(clk_disable);
+
+static int __clk_enable(struct clk *clk)
+{
+	int ret = 0;
+
+	if (clk->usecount++ == 0) {
+		if (clk->parent) {
+			ret = __clk_enable(clk->parent);
+			if (unlikely(ret))
+				goto err;
+		}
+
+		if (clk->ops && clk->ops->enable) {
+			ret = clk->ops->enable(clk);
+			if (ret) {
+				if (clk->parent)
+					__clk_disable(clk->parent);
+				goto err;
+			}
+		}
+	}
+
+	return ret;
+err:
+	clk->usecount--;
+	return ret;
+}
+
+int clk_enable(struct clk *clk)
+{
+	unsigned long flags;
+	int ret;
+
+	if (!clk)
+		return -EINVAL;
+
+	spin_lock_irqsave(&clock_lock, flags);
+	ret = __clk_enable(clk);
+	spin_unlock_irqrestore(&clock_lock, flags);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(clk_enable);
+
+static LIST_HEAD(root_clks);
+
+/**
+ * recalculate_root_clocks - recalculate and propagate all root clocks
+ *
+ * Recalculates all root clocks (clocks with no parent), which if the
+ * clock's .recalc is set correctly, should also propagate their rates.
+ * Called at init.
+ */
+void recalculate_root_clocks(void)
+{
+	struct clk *clkp;
+
+	list_for_each_entry(clkp, &root_clks, sibling) {
+		if (clkp->ops && clkp->ops->recalc)
+			clkp->rate = clkp->ops->recalc(clkp);
+		propagate_rate(clkp);
+	}
+}
+
+int clk_register(struct clk *clk)
+{
+	if (clk == NULL || IS_ERR(clk))
+		return -EINVAL;
+
+	/*
+	 * trap out already registered clocks
+	 */
+	if (clk->node.next || clk->node.prev)
+		return 0;
+
+	mutex_lock(&clock_list_sem);
+
+	INIT_LIST_HEAD(&clk->children);
+	clk->usecount = 0;
+
+	if (clk->parent)
+		list_add(&clk->sibling, &clk->parent->children);
+	else
+		list_add(&clk->sibling, &root_clks);
+
+	list_add(&clk->node, &clock_list);
+	if (clk->ops && clk->ops->init)
+		clk->ops->init(clk);
+	mutex_unlock(&clock_list_sem);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(clk_register);
+
+void clk_unregister(struct clk *clk)
+{
+	mutex_lock(&clock_list_sem);
+	list_del(&clk->sibling);
+	list_del(&clk->node);
+	mutex_unlock(&clock_list_sem);
+}
+EXPORT_SYMBOL_GPL(clk_unregister);
+
+void clk_enable_init_clocks(void)
+{
+	struct clk *clkp;
+
+	list_for_each_entry(clkp, &clock_list, node)
+		if (clkp->flags & CLK_ENABLE_ON_INIT)
+			clk_enable(clkp);
+}
+
+unsigned long clk_get_rate(struct clk *clk)
+{
+	return clk->rate;
+}
+EXPORT_SYMBOL_GPL(clk_get_rate);
+
+int clk_set_rate(struct clk *clk, unsigned long rate)
+{
+	return clk_set_rate_ex(clk, rate, 0);
+}
+EXPORT_SYMBOL_GPL(clk_set_rate);
+
+int clk_set_rate_ex(struct clk *clk, unsigned long rate, int algo_id)
+{
+	int ret = -EOPNOTSUPP;
+	unsigned long flags;
+
+	spin_lock_irqsave(&clock_lock, flags);
+
+	if (likely(clk->ops && clk->ops->set_rate)) {
+		ret = clk->ops->set_rate(clk, rate, algo_id);
+		if (ret != 0)
+			goto out_unlock;
+	} else {
+		clk->rate = rate;
+		ret = 0;
+	}
+
+	if (clk->ops && clk->ops->recalc)
+		clk->rate = clk->ops->recalc(clk);
+
+	propagate_rate(clk);
+
+out_unlock:
+	spin_unlock_irqrestore(&clock_lock, flags);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(clk_set_rate_ex);
+
+int clk_set_parent(struct clk *clk, struct clk *parent)
+{
+	unsigned long flags;
+	int ret = -EINVAL;
+
+	if (!parent || !clk)
+		return ret;
+	if (clk->parent == parent)
+		return 0;
+
+	spin_lock_irqsave(&clock_lock, flags);
+	if (clk->usecount == 0) {
+		if (clk->ops->set_parent)
+			ret = clk->ops->set_parent(clk, parent);
+		else
+			ret = clk_reparent(clk, parent);
+
+		if (ret == 0) {
+			pr_debug("clock: set parent of %s to %s (new rate %ld)\n",
+				 clk->name, clk->parent->name, clk->rate);
+			if (clk->ops->recalc)
+				clk->rate = clk->ops->recalc(clk);
+			propagate_rate(clk);
+		}
+	} else
+		ret = -EBUSY;
+	spin_unlock_irqrestore(&clock_lock, flags);
+
+	return ret;
+}
+EXPORT_SYMBOL_GPL(clk_set_parent);
+
+struct clk *clk_get_parent(struct clk *clk)
+{
+	return clk->parent;
+}
+EXPORT_SYMBOL_GPL(clk_get_parent);
+
+long clk_round_rate(struct clk *clk, unsigned long rate)
+{
+	if (likely(clk->ops && clk->ops->round_rate)) {
+		unsigned long flags, rounded;
+
+		spin_lock_irqsave(&clock_lock, flags);
+		rounded = clk->ops->round_rate(clk, rate);
+		spin_unlock_irqrestore(&clock_lock, flags);
+
+		return rounded;
+	}
+
+	return clk_get_rate(clk);
+}
+EXPORT_SYMBOL_GPL(clk_round_rate);
+
+#ifdef CONFIG_PM
+static int clks_sysdev_suspend(struct sys_device *dev, pm_message_t state)
+{
+	static pm_message_t prev_state;
+	struct clk *clkp;
+
+	switch (state.event) {
+	case PM_EVENT_ON:
+		/* Resumeing from hibernation */
+		if (prev_state.event != PM_EVENT_FREEZE)
+			break;
+
+		list_for_each_entry(clkp, &clock_list, node) {
+			if (likely(clkp->ops)) {
+				unsigned long rate = clkp->rate;
+
+				if (likely(clkp->ops->set_parent))
+					clkp->ops->set_parent(clkp,
+						clkp->parent);
+				if (likely(clkp->ops->set_rate))
+					clkp->ops->set_rate(clkp,
+						rate, NO_CHANGE);
+				else if (likely(clkp->ops->recalc))
+					clkp->rate = clkp->ops->recalc(clkp);
+			}
+		}
+		break;
+	case PM_EVENT_FREEZE:
+		break;
+	case PM_EVENT_SUSPEND:
+		break;
+	}
+
+	prev_state = state;
+	return 0;
+}
+
+static int clks_sysdev_resume(struct sys_device *dev)
+{
+	return clks_sysdev_suspend(dev, PMSG_ON);
+}
+
+static struct sysdev_class clks_sysdev_class = {
+	.name = "clks",
+};
+
+static struct sysdev_driver clks_sysdev_driver = {
+	.suspend = clks_sysdev_suspend,
+	.resume = clks_sysdev_resume,
+};
+
+static struct sys_device clks_sysdev_dev = {
+	.cls = &clks_sysdev_class,
+};
+
+static int __init clk_sysdev_init(void)
+{
+	sysdev_class_register(&clks_sysdev_class);
+	sysdev_driver_register(&clks_sysdev_class, &clks_sysdev_driver);
+	sysdev_register(&clks_sysdev_dev);
+
+	return 0;
+}
+subsys_initcall(clk_sysdev_init);
+#endif
+
+/*
+ *	debugfs support to trace clock tree hierarchy and attributes
+ */
+static struct dentry *clk_debugfs_root;
+
+static int clk_debugfs_register_one(struct clk *c)
+{
+	int err;
+	struct dentry *d, *child, *child_tmp;
+	struct clk *pa = c->parent;
+	char s[255];
+	char *p = s;
+
+	p += sprintf(p, "%s", c->name);
+	if (c->id >= 0)
+		sprintf(p, ":%d", c->id);
+	d = debugfs_create_dir(s, pa ? pa->dentry : clk_debugfs_root);
+	if (!d)
+		return -ENOMEM;
+	c->dentry = d;
+
+	d = debugfs_create_u8("usecount", S_IRUGO, c->dentry, (u8 *)&c->usecount);
+	if (!d) {
+		err = -ENOMEM;
+		goto err_out;
+	}
+	d = debugfs_create_u32("rate", S_IRUGO, c->dentry, (u32 *)&c->rate);
+	if (!d) {
+		err = -ENOMEM;
+		goto err_out;
+	}
+	d = debugfs_create_x32("flags", S_IRUGO, c->dentry, (u32 *)&c->flags);
+	if (!d) {
+		err = -ENOMEM;
+		goto err_out;
+	}
+	return 0;
+
+err_out:
+	d = c->dentry;
+	list_for_each_entry_safe(child, child_tmp, &d->d_subdirs, d_u.d_child)
+		debugfs_remove(child);
+	debugfs_remove(c->dentry);
+	return err;
+}
+
+static int clk_debugfs_register(struct clk *c)
+{
+	int err;
+	struct clk *pa = c->parent;
+
+	if (pa && !pa->dentry) {
+		err = clk_debugfs_register(pa);
+		if (err)
+			return err;
+	}
+
+	if (!c->dentry && c->name) {
+		err = clk_debugfs_register_one(c);
+		if (err)
+			return err;
+	}
+	return 0;
+}
+
+static int __init clk_debugfs_init(void)
+{
+	struct clk *c;
+	struct dentry *d;
+	int err;
+
+	d = debugfs_create_dir("clock", NULL);
+	if (!d)
+		return -ENOMEM;
+	clk_debugfs_root = d;
+
+	list_for_each_entry(c, &clock_list, node) {
+		err = clk_debugfs_register(c);
+		if (err)
+			goto err_out;
+	}
+	return 0;
+err_out:
+	debugfs_remove_recursive(clk_debugfs_root);
+	return err;
+}
+late_initcall(clk_debugfs_init);
diff --git a/drivers/sh/intc.c b/drivers/sh/intc.c
index 94ad6bd86a0..c585574b9ae 100644
--- a/drivers/sh/intc.c
+++ b/drivers/sh/intc.c
@@ -28,6 +28,7 @@
 #include <linux/topology.h>
 #include <linux/bitmap.h>
 #include <linux/cpumask.h>
+#include <asm/sizes.h>
 
 #define _INTC_MK(fn, mode, addr_e, addr_d, width, shift) \
 	((shift) | ((width) << 5) | ((fn) << 9) | ((mode) << 13) | \
@@ -45,6 +46,12 @@ struct intc_handle_int {
 	unsigned long handle;
 };
 
+struct intc_window {
+	phys_addr_t phys;
+	void __iomem *virt;
+	unsigned long size;
+};
+
 struct intc_desc_int {
 	struct list_head list;
 	struct sys_device sysdev;
@@ -58,6 +65,8 @@ struct intc_desc_int {
 	unsigned int nr_prio;
 	struct intc_handle_int *sense;
 	unsigned int nr_sense;
+	struct intc_window *window;
+	unsigned int nr_windows;
 	struct irq_chip chip;
 };
 
@@ -87,8 +96,12 @@ static DEFINE_SPINLOCK(vector_lock);
 #define SMP_NR(d, x) 1
 #endif
 
-static unsigned int intc_prio_level[NR_IRQS]; /* for now */
+static unsigned int intc_prio_level[NR_IRQS];	/* for now */
+static unsigned int default_prio_level = 2;	/* 2 - 16 */
 static unsigned long ack_handle[NR_IRQS];
+#ifdef CONFIG_INTC_BALANCING
+static unsigned long dist_handle[NR_IRQS];
+#endif
 
 static inline struct intc_desc_int *get_intc_desc(unsigned int irq)
 {
@@ -96,6 +109,47 @@ static inline struct intc_desc_int *get_intc_desc(unsigned int irq)
 	return container_of(chip, struct intc_desc_int, chip);
 }
 
+static unsigned long intc_phys_to_virt(struct intc_desc_int *d,
+				       unsigned long address)
+{
+	struct intc_window *window;
+	int k;
+
+	/* scan through physical windows and convert address */
+	for (k = 0; k < d->nr_windows; k++) {
+		window = d->window + k;
+
+		if (address < window->phys)
+			continue;
+
+		if (address >= (window->phys + window->size))
+			continue;
+
+		address -= window->phys;
+		address += (unsigned long)window->virt;
+
+		return address;
+	}
+
+	/* no windows defined, register must be 1:1 mapped virt:phys */
+	return address;
+}
+
+static unsigned int intc_get_reg(struct intc_desc_int *d, unsigned long address)
+{
+	unsigned int k;
+
+	address = intc_phys_to_virt(d, address);
+
+	for (k = 0; k < d->nr_reg; k++) {
+		if (d->reg[k] == address)
+			return k;
+	}
+
+	BUG();
+	return 0;
+}
+
 static inline unsigned int set_field(unsigned int value,
 				     unsigned int field_value,
 				     unsigned int handle)
@@ -229,6 +283,85 @@ static void (*intc_disable_fns[])(unsigned long addr,
 	[MODE_PCLR_REG] = intc_mode_field,
 };
 
+#ifdef CONFIG_INTC_BALANCING
+static inline void intc_balancing_enable(unsigned int irq)
+{
+	struct intc_desc_int *d = get_intc_desc(irq);
+	unsigned long handle = dist_handle[irq];
+	unsigned long addr;
+
+	if (irq_balancing_disabled(irq) || !handle)
+		return;
+
+	addr = INTC_REG(d, _INTC_ADDR_D(handle), 0);
+	intc_reg_fns[_INTC_FN(handle)](addr, handle, 1);
+}
+
+static inline void intc_balancing_disable(unsigned int irq)
+{
+	struct intc_desc_int *d = get_intc_desc(irq);
+	unsigned long handle = dist_handle[irq];
+	unsigned long addr;
+
+	if (irq_balancing_disabled(irq) || !handle)
+		return;
+
+	addr = INTC_REG(d, _INTC_ADDR_D(handle), 0);
+	intc_reg_fns[_INTC_FN(handle)](addr, handle, 0);
+}
+
+static unsigned int intc_dist_data(struct intc_desc *desc,
+				   struct intc_desc_int *d,
+				   intc_enum enum_id)
+{
+	struct intc_mask_reg *mr = desc->hw.mask_regs;
+	unsigned int i, j, fn, mode;
+	unsigned long reg_e, reg_d;
+
+	for (i = 0; mr && enum_id && i < desc->hw.nr_mask_regs; i++) {
+		mr = desc->hw.mask_regs + i;
+
+		/*
+		 * Skip this entry if there's no auto-distribution
+		 * register associated with it.
+		 */
+		if (!mr->dist_reg)
+			continue;
+
+		for (j = 0; j < ARRAY_SIZE(mr->enum_ids); j++) {
+			if (mr->enum_ids[j] != enum_id)
+				continue;
+
+			fn = REG_FN_MODIFY_BASE;
+			mode = MODE_ENABLE_REG;
+			reg_e = mr->dist_reg;
+			reg_d = mr->dist_reg;
+
+			fn += (mr->reg_width >> 3) - 1;
+			return _INTC_MK(fn, mode,
+					intc_get_reg(d, reg_e),
+					intc_get_reg(d, reg_d),
+					1,
+					(mr->reg_width - 1) - j);
+		}
+	}
+
+	/*
+	 * It's possible we've gotten here with no distribution options
+	 * available for the IRQ in question, so we just skip over those.
+	 */
+	return 0;
+}
+#else
+static inline void intc_balancing_enable(unsigned int irq)
+{
+}
+
+static inline void intc_balancing_disable(unsigned int irq)
+{
+}
+#endif
+
 static inline void _intc_enable(unsigned int irq, unsigned long handle)
 {
 	struct intc_desc_int *d = get_intc_desc(irq);
@@ -244,6 +377,8 @@ static inline void _intc_enable(unsigned int irq, unsigned long handle)
 		intc_enable_fns[_INTC_MODE(handle)](addr, handle, intc_reg_fns\
 						    [_INTC_FN(handle)], irq);
 	}
+
+	intc_balancing_enable(irq);
 }
 
 static void intc_enable(unsigned int irq)
@@ -254,10 +389,12 @@ static void intc_enable(unsigned int irq)
 static void intc_disable(unsigned int irq)
 {
 	struct intc_desc_int *d = get_intc_desc(irq);
-	unsigned long handle = (unsigned long) get_irq_chip_data(irq);
+	unsigned long handle = (unsigned long)get_irq_chip_data(irq);
 	unsigned long addr;
 	unsigned int cpu;
 
+	intc_balancing_disable(irq);
+
 	for (cpu = 0; cpu < SMP_NR(d, _INTC_ADDR_D(handle)); cpu++) {
 #ifdef CONFIG_SMP
 		if (!cpumask_test_cpu(cpu, irq_to_desc(irq)->affinity))
@@ -336,8 +473,7 @@ static void intc_mask_ack(unsigned int irq)
 
 	intc_disable(irq);
 
-	/* read register and write zero only to the assocaited bit */
-
+	/* read register and write zero only to the associated bit */
 	if (handle) {
 		addr = INTC_REG(d, _INTC_ADDR_D(handle), 0);
 		switch (_INTC_FN(handle)) {
@@ -366,7 +502,8 @@ static struct intc_handle_int *intc_find_irq(struct intc_handle_int *hp,
 {
 	int i;
 
-	/* this doesn't scale well, but...
+	/*
+	 * this doesn't scale well, but...
 	 *
 	 * this function should only be used for cerain uncommon
 	 * operations such as intc_set_priority() and intc_set_sense()
@@ -377,7 +514,6 @@ static struct intc_handle_int *intc_find_irq(struct intc_handle_int *hp,
 	 * memory footprint down is to make sure the array is sorted
 	 * and then perform a bisect to lookup the irq.
 	 */
-
 	for (i = 0; i < nr_hp; i++) {
 		if ((hp + i)->irq != irq)
 			continue;
@@ -408,7 +544,6 @@ int intc_set_priority(unsigned int irq, unsigned int prio)
 		 * primary masking method is using intc_prio_level[irq]
 		 * priority level will be set during next enable()
 		 */
-
 		if (_INTC_FN(ihp->handle) != REG_FN_ERR)
 			_intc_enable(irq, ihp->handle);
 	}
@@ -447,20 +582,6 @@ static int intc_set_sense(unsigned int irq, unsigned int type)
 	return 0;
 }
 
-static unsigned int __init intc_get_reg(struct intc_desc_int *d,
-				 unsigned long address)
-{
-	unsigned int k;
-
-	for (k = 0; k < d->nr_reg; k++) {
-		if (d->reg[k] == address)
-			return k;
-	}
-
-	BUG();
-	return 0;
-}
-
 static intc_enum __init intc_grp_id(struct intc_desc *desc,
 				    intc_enum enum_id)
 {
@@ -718,13 +839,14 @@ static void __init intc_register_irq(struct intc_desc *desc,
 	 */
 	set_bit(irq, intc_irq_map);
 
-	/* Prefer single interrupt source bitmap over other combinations:
+	/*
+	 * Prefer single interrupt source bitmap over other combinations:
+	 *
 	 * 1. bitmap, single interrupt source
 	 * 2. priority, single interrupt source
 	 * 3. bitmap, multiple interrupt sources (groups)
 	 * 4. priority, multiple interrupt sources (groups)
 	 */
-
 	data[0] = intc_mask_data(desc, d, enum_id, 0);
 	data[1] = intc_prio_data(desc, d, enum_id, 0);
 
@@ -749,10 +871,11 @@ static void __init intc_register_irq(struct intc_desc *desc,
 				      handle_level_irq, "level");
 	set_irq_chip_data(irq, (void *)data[primary]);
 
-	/* set priority level
+	/*
+	 * set priority level
 	 * - this needs to be at least 2 for 5-bit priorities on 7780
 	 */
-	intc_prio_level[irq] = 2;
+	intc_prio_level[irq] = default_prio_level;
 
 	/* enable secondary masking method if present */
 	if (data[!primary])
@@ -769,7 +892,6 @@ static void __init intc_register_irq(struct intc_desc *desc,
 			 * only secondary priority should access registers, so
 			 * set _INTC_FN(h) = REG_FN_ERR for intc_set_priority()
 			 */
-
 			hp->handle &= ~_INTC_MK(0x0f, 0, 0, 0, 0, 0);
 			hp->handle |= _INTC_MK(REG_FN_ERR, 0, 0, 0, 0, 0);
 		}
@@ -790,6 +912,11 @@ static void __init intc_register_irq(struct intc_desc *desc,
 	if (desc->hw.ack_regs)
 		ack_handle[irq] = intc_ack_data(desc, d, enum_id);
 
+#ifdef CONFIG_INTC_BALANCING
+	if (desc->hw.mask_regs)
+		dist_handle[irq] = intc_dist_data(desc, d, enum_id);
+#endif
+
 #ifdef CONFIG_ARM
 	set_irq_flags(irq, IRQF_VALID); /* Enable IRQ on ARM systems */
 #endif
@@ -801,6 +928,8 @@ static unsigned int __init save_reg(struct intc_desc_int *d,
 				    unsigned int smp)
 {
 	if (value) {
+		value = intc_phys_to_virt(d, value);
+
 		d->reg[cnt] = value;
 #ifdef CONFIG_SMP
 		d->smp[cnt] = smp;
@@ -816,25 +945,59 @@ static void intc_redirect_irq(unsigned int irq, struct irq_desc *desc)
 	generic_handle_irq((unsigned int)get_irq_data(irq));
 }
 
-void __init register_intc_controller(struct intc_desc *desc)
+int __init register_intc_controller(struct intc_desc *desc)
 {
 	unsigned int i, k, smp;
 	struct intc_hw_desc *hw = &desc->hw;
 	struct intc_desc_int *d;
+	struct resource *res;
+
+	pr_info("intc: Registered controller '%s' with %u IRQs\n",
+		desc->name, hw->nr_vectors);
 
 	d = kzalloc(sizeof(*d), GFP_NOWAIT);
+	if (!d)
+		goto err0;
 
 	INIT_LIST_HEAD(&d->list);
 	list_add(&d->list, &intc_list);
 
+	if (desc->num_resources) {
+		d->nr_windows = desc->num_resources;
+		d->window = kzalloc(d->nr_windows * sizeof(*d->window),
+				    GFP_NOWAIT);
+		if (!d->window)
+			goto err1;
+
+		for (k = 0; k < d->nr_windows; k++) {
+			res = desc->resource + k;
+			WARN_ON(resource_type(res) != IORESOURCE_MEM);
+			d->window[k].phys = res->start;
+			d->window[k].size = resource_size(res);
+			d->window[k].virt = ioremap_nocache(res->start,
+							 resource_size(res));
+			if (!d->window[k].virt)
+				goto err2;
+		}
+	}
+
 	d->nr_reg = hw->mask_regs ? hw->nr_mask_regs * 2 : 0;
+#ifdef CONFIG_INTC_BALANCING
+	if (d->nr_reg)
+		d->nr_reg += hw->nr_mask_regs;
+#endif
 	d->nr_reg += hw->prio_regs ? hw->nr_prio_regs * 2 : 0;
 	d->nr_reg += hw->sense_regs ? hw->nr_sense_regs : 0;
 	d->nr_reg += hw->ack_regs ? hw->nr_ack_regs : 0;
 
 	d->reg = kzalloc(d->nr_reg * sizeof(*d->reg), GFP_NOWAIT);
+	if (!d->reg)
+		goto err2;
+
 #ifdef CONFIG_SMP
 	d->smp = kzalloc(d->nr_reg * sizeof(*d->smp), GFP_NOWAIT);
+	if (!d->smp)
+		goto err3;
 #endif
 	k = 0;
 
@@ -843,12 +1006,17 @@ void __init register_intc_controller(struct intc_desc *desc)
 			smp = IS_SMP(hw->mask_regs[i]);
 			k += save_reg(d, k, hw->mask_regs[i].set_reg, smp);
 			k += save_reg(d, k, hw->mask_regs[i].clr_reg, smp);
+#ifdef CONFIG_INTC_BALANCING
+			k += save_reg(d, k, hw->mask_regs[i].dist_reg, 0);
+#endif
 		}
 	}
 
 	if (hw->prio_regs) {
 		d->prio = kzalloc(hw->nr_vectors * sizeof(*d->prio),
 				  GFP_NOWAIT);
+		if (!d->prio)
+			goto err4;
 
 		for (i = 0; i < hw->nr_prio_regs; i++) {
 			smp = IS_SMP(hw->prio_regs[i]);
@@ -860,6 +1028,8 @@ void __init register_intc_controller(struct intc_desc *desc)
 	if (hw->sense_regs) {
 		d->sense = kzalloc(hw->nr_vectors * sizeof(*d->sense),
 				   GFP_NOWAIT);
+		if (!d->sense)
+			goto err5;
 
 		for (i = 0; i < hw->nr_sense_regs; i++)
 			k += save_reg(d, k, hw->sense_regs[i].reg, 0);
@@ -906,7 +1076,7 @@ void __init register_intc_controller(struct intc_desc *desc)
 
 		irq_desc = irq_to_desc_alloc_node(irq, numa_node_id());
 		if (unlikely(!irq_desc)) {
-			pr_info("can't get irq_desc for %d\n", irq);
+			pr_err("can't get irq_desc for %d\n", irq);
 			continue;
 		}
 
@@ -926,7 +1096,7 @@ void __init register_intc_controller(struct intc_desc *desc)
 			 */
 			irq_desc = irq_to_desc_alloc_node(irq2, numa_node_id());
 			if (unlikely(!irq_desc)) {
-				pr_info("can't get irq_desc for %d\n", irq2);
+				pr_err("can't get irq_desc for %d\n", irq2);
 				continue;
 			}
 
@@ -942,8 +1112,100 @@ void __init register_intc_controller(struct intc_desc *desc)
 	/* enable bits matching force_enable after registering irqs */
 	if (desc->force_enable)
 		intc_enable_disable_enum(desc, d, desc->force_enable, 1);
+
+	return 0;
+err5:
+	kfree(d->prio);
+err4:
+#ifdef CONFIG_SMP
+	kfree(d->smp);
+err3:
+#endif
+	kfree(d->reg);
+err2:
+	for (k = 0; k < d->nr_windows; k++)
+		if (d->window[k].virt)
+			iounmap(d->window[k].virt);
+
+	kfree(d->window);
+err1:
+	kfree(d);
+err0:
+	pr_err("unable to allocate INTC memory\n");
+
+	return -ENOMEM;
+}
+
+#ifdef CONFIG_INTC_USERIMASK
+static void __iomem *uimask;
+
+int register_intc_userimask(unsigned long addr)
+{
+	if (unlikely(uimask))
+		return -EBUSY;
+
+	uimask = ioremap_nocache(addr, SZ_4K);
+	if (unlikely(!uimask))
+		return -ENOMEM;
+
+	pr_info("intc: userimask support registered for levels 0 -> %d\n",
+		default_prio_level - 1);
+
+	return 0;
+}
+
+static ssize_t
+show_intc_userimask(struct sysdev_class *cls,
+		    struct sysdev_class_attribute *attr, char *buf)
+{
+	return sprintf(buf, "%d\n", (__raw_readl(uimask) >> 4) & 0xf);
+}
+
+static ssize_t
+store_intc_userimask(struct sysdev_class *cls,
+		     struct sysdev_class_attribute *attr,
+		     const char *buf, size_t count)
+{
+	unsigned long level;
+
+	level = simple_strtoul(buf, NULL, 10);
+
+	/*
+	 * Minimal acceptable IRQ levels are in the 2 - 16 range, but
+	 * these are chomped so as to not interfere with normal IRQs.
+	 *
+	 * Level 1 is a special case on some CPUs in that it's not
+	 * directly settable, but given that USERIMASK cuts off below a
+	 * certain level, we don't care about this limitation here.
+	 * Level 0 on the other hand equates to user masking disabled.
+	 *
+	 * We use default_prio_level as a cut off so that only special
+	 * case opt-in IRQs can be mangled.
+	 */
+	if (level >= default_prio_level)
+		return -EINVAL;
+
+	__raw_writel(0xa5 << 24 | level << 4, uimask);
+
+	return count;
 }
 
+static SYSDEV_CLASS_ATTR(userimask, S_IRUSR | S_IWUSR,
+			 show_intc_userimask, store_intc_userimask);
+#endif
+
+static ssize_t
+show_intc_name(struct sys_device *dev, struct sysdev_attribute *attr, char *buf)
+{
+	struct intc_desc_int *d;
+
+	d = container_of(dev, struct intc_desc_int, sysdev);
+
+	return sprintf(buf, "%s\n", d->chip.name);
+}
+
+static SYSDEV_ATTR(name, S_IRUGO, show_intc_name, NULL);
+
 static int intc_suspend(struct sys_device *dev, pm_message_t state)
 {
 	struct intc_desc_int *d;
@@ -1003,19 +1265,28 @@ static int __init register_intc_sysdevs(void)
 	int id = 0;
 
 	error = sysdev_class_register(&intc_sysdev_class);
+#ifdef CONFIG_INTC_USERIMASK
+	if (!error && uimask)
+		error = sysdev_class_create_file(&intc_sysdev_class,
+						 &attr_userimask);
+#endif
 	if (!error) {
 		list_for_each_entry(d, &intc_list, list) {
 			d->sysdev.id = id;
 			d->sysdev.cls = &intc_sysdev_class;
 			error = sysdev_register(&d->sysdev);
+			if (error == 0)
+				error = sysdev_create_file(&d->sysdev,
+							   &attr_name);
 			if (error)
 				break;
+
 			id++;
 		}
 	}
 
 	if (error)
-		pr_warning("intc: sysdev registration error\n");
+		pr_err("intc: sysdev registration error\n");
 
 	return error;
 }
@@ -1048,7 +1319,7 @@ unsigned int create_irq_nr(unsigned int irq_want, int node)
 
 	desc = irq_to_desc_alloc_node(new, node);
 	if (unlikely(!desc)) {
-		pr_info("can't get irq_desc for %d\n", new);
+		pr_err("can't get irq_desc for %d\n", new);
 		goto out_unlock;
 	}