Merge branch 'board' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/renesas into next/boards

* 'board' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/renesas:
  ARM: mach-shmobile: bonito: make sure static function
  ARM: mach-shmobile: sh7372 CEU supports up to 8188x8188 images
  ARM: mach-shmobile: mackerel: Add FSI DMAEngine support

34 files changed, 616 insertions, 165 deletions

diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
index cf006d40342c..36586dba6fa6 100644
--- a/arch/arm/Kconfig
+++ b/arch/arm/Kconfig
@@ -1186,6 +1186,15 @@ if !MMU
 source "arch/arm/Kconfig-nommu"
 endif
 
+config ARM_ERRATA_326103
+	bool "ARM errata: FSR write bit incorrect on a SWP to read-only memory"
+	depends on CPU_V6
+	help
+	  Executing a SWP instruction to read-only memory does not set bit 11
+	  of the FSR on the ARM 1136 prior to r1p0. This causes the kernel to
+	  treat the access as a read, preventing a COW from occurring and
+	  causing the faulting task to livelock.
+
 config ARM_ERRATA_411920
 	bool "ARM errata: Invalidation of the Instruction Cache operation can fail"
 	depends on CPU_V6 || CPU_V6K
diff --git a/arch/arm/boot/dts/sh7372.dtsi b/arch/arm/boot/dts/sh7372.dtsi
new file mode 100644
index 000000000000..677fc603f8b3
--- /dev/null
+++ b/arch/arm/boot/dts/sh7372.dtsi
@@ -0,0 +1,21 @@
+/*
+ * Device Tree Source for the sh7372 SoC
+ *
+ * Copyright (C) 2012 Renesas Solutions Corp.
+ *
+ * This file is licensed under the terms of the GNU General Public License
+ * version 2.  This program is licensed "as is" without any warranty of any
+ * kind, whether express or implied.
+ */
+
+/include/ "skeleton.dtsi"
+
+/ {
+	compatible = "renesas,sh7372";
+
+	cpus {
+		cpu@0 {
+			compatible = "arm,cortex-a8";
+		};
+	};
+};
diff --git a/arch/arm/boot/dts/versatile-ab.dts b/arch/arm/boot/dts/versatile-ab.dts
index 0b32925f2147..e2fe3195c0d1 100644
--- a/arch/arm/boot/dts/versatile-ab.dts
+++ b/arch/arm/boot/dts/versatile-ab.dts
@@ -173,7 +173,7 @@
 			mmc@5000 {
 				compatible = "arm,primecell";
 				reg = < 0x5000 0x1000>;
-				interrupts = <22>;
+				interrupts = <22 34>;
 			};
 			kmi@6000 {
 				compatible = "arm,pl050", "arm,primecell";
diff --git a/arch/arm/boot/dts/versatile-pb.dts b/arch/arm/boot/dts/versatile-pb.dts
index 166461073b78..7e8175269064 100644
--- a/arch/arm/boot/dts/versatile-pb.dts
+++ b/arch/arm/boot/dts/versatile-pb.dts
@@ -41,7 +41,7 @@
 			mmc@b000 {
 				compatible = "arm,primecell";
 				reg = <0xb000 0x1000>;
-				interrupts = <23>;
+				interrupts = <23 34>;
 			};
 		};
 	};
diff --git a/arch/arm/include/asm/thread_info.h b/arch/arm/include/asm/thread_info.h
index d4c24d412a8d..0f04d84582e1 100644
--- a/arch/arm/include/asm/thread_info.h
+++ b/arch/arm/include/asm/thread_info.h
@@ -118,6 +118,13 @@ extern void iwmmxt_task_switch(struct thread_info *);
 extern void vfp_sync_hwstate(struct thread_info *);
 extern void vfp_flush_hwstate(struct thread_info *);
 
+struct user_vfp;
+struct user_vfp_exc;
+
+extern int vfp_preserve_user_clear_hwstate(struct user_vfp __user *,
+					   struct user_vfp_exc __user *);
+extern int vfp_restore_user_hwstate(struct user_vfp __user *,
+				    struct user_vfp_exc __user *);
 #endif
 
 /*
diff --git a/arch/arm/include/asm/tls.h b/arch/arm/include/asm/tls.h
index 60843eb0f61c..73409e6c0251 100644
--- a/arch/arm/include/asm/tls.h
+++ b/arch/arm/include/asm/tls.h
@@ -7,6 +7,8 @@
 
 	.macro set_tls_v6k, tp, tmp1, tmp2
 	mcr	p15, 0, \tp, c13, c0, 3		@ set TLS register
+	mov	\tmp1, #0
+	mcr	p15, 0, \tmp1, c13, c0, 2	@ clear user r/w TLS register
 	.endm
 
 	.macro set_tls_v6, tp, tmp1, tmp2
@@ -15,6 +17,8 @@
 	mov	\tmp2, #0xffff0fff
 	tst	\tmp1, #HWCAP_TLS		@ hardware TLS available?
 	mcrne	p15, 0, \tp, c13, c0, 3		@ yes, set TLS register
+	movne	\tmp1, #0
+	mcrne	p15, 0, \tmp1, c13, c0, 2	@ clear user r/w TLS register
 	streq	\tp, [\tmp2, #-15]		@ set TLS value at 0xffff0ff0
 	.endm
 
diff --git a/arch/arm/kernel/irq.c b/arch/arm/kernel/irq.c
index 71ccdbfed662..8349d4e97e2b 100644
--- a/arch/arm/kernel/irq.c
+++ b/arch/arm/kernel/irq.c
@@ -155,10 +155,10 @@ static bool migrate_one_irq(struct irq_desc *desc)
 	}
 
 	c = irq_data_get_irq_chip(d);
-	if (c->irq_set_affinity)
-		c->irq_set_affinity(d, affinity, true);
-	else
+	if (!c->irq_set_affinity)
 		pr_debug("IRQ%u: unable to set affinity\n", d->irq);
+	else if (c->irq_set_affinity(d, affinity, true) == IRQ_SET_MASK_OK && ret)
+		cpumask_copy(d->affinity, affinity);
 
 	return ret;
 }
diff --git a/arch/arm/kernel/signal.c b/arch/arm/kernel/signal.c
index 7cb532fc8aa4..d68d1b694680 100644
--- a/arch/arm/kernel/signal.c
+++ b/arch/arm/kernel/signal.c
@@ -180,44 +180,23 @@ static int restore_iwmmxt_context(struct iwmmxt_sigframe *frame)
 
 static int preserve_vfp_context(struct vfp_sigframe __user *frame)
 {
-	struct thread_info *thread = current_thread_info();
-	struct vfp_hard_struct *h = &thread->vfpstate.hard;
 	const unsigned long magic = VFP_MAGIC;
 	const unsigned long size = VFP_STORAGE_SIZE;
 	int err = 0;
 
-	vfp_sync_hwstate(thread);
 	__put_user_error(magic, &frame->magic, err);
 	__put_user_error(size, &frame->size, err);
 
-	/*
-	 * Copy the floating point registers. There can be unused
-	 * registers see asm/hwcap.h for details.
-	 */
-	err |= __copy_to_user(&frame->ufp.fpregs, &h->fpregs,
-			      sizeof(h->fpregs));
-	/*
-	 * Copy the status and control register.
-	 */
-	__put_user_error(h->fpscr, &frame->ufp.fpscr, err);
-
-	/*
-	 * Copy the exception registers.
-	 */
-	__put_user_error(h->fpexc, &frame->ufp_exc.fpexc, err);
-	__put_user_error(h->fpinst, &frame->ufp_exc.fpinst, err);
-	__put_user_error(h->fpinst2, &frame->ufp_exc.fpinst2, err);
+	if (err)
+		return -EFAULT;
 
-	return err ? -EFAULT : 0;
+	return vfp_preserve_user_clear_hwstate(&frame->ufp, &frame->ufp_exc);
 }
 
 static int restore_vfp_context(struct vfp_sigframe __user *frame)
 {
-	struct thread_info *thread = current_thread_info();
-	struct vfp_hard_struct *h = &thread->vfpstate.hard;
 	unsigned long magic;
 	unsigned long size;
-	unsigned long fpexc;
 	int err = 0;
 
 	__get_user_error(magic, &frame->magic, err);
@@ -228,33 +207,7 @@ static int restore_vfp_context(struct vfp_sigframe __user *frame)
 	if (magic != VFP_MAGIC || size != VFP_STORAGE_SIZE)
 		return -EINVAL;
 
-	vfp_flush_hwstate(thread);
-
-	/*
-	 * Copy the floating point registers. There can be unused
-	 * registers see asm/hwcap.h for details.
-	 */
-	err |= __copy_from_user(&h->fpregs, &frame->ufp.fpregs,
-				sizeof(h->fpregs));
-	/*
-	 * Copy the status and control register.
-	 */
-	__get_user_error(h->fpscr, &frame->ufp.fpscr, err);
-
-	/*
-	 * Sanitise and restore the exception registers.
-	 */
-	__get_user_error(fpexc, &frame->ufp_exc.fpexc, err);
-	/* Ensure the VFP is enabled. */
-	fpexc |= FPEXC_EN;
-	/* Ensure FPINST2 is invalid and the exception flag is cleared. */
-	fpexc &= ~(FPEXC_EX | FPEXC_FP2V);
-	h->fpexc = fpexc;
-
-	__get_user_error(h->fpinst, &frame->ufp_exc.fpinst, err);
-	__get_user_error(h->fpinst2, &frame->ufp_exc.fpinst2, err);
-
-	return err ? -EFAULT : 0;
+	return vfp_restore_user_hwstate(&frame->ufp, &frame->ufp_exc);
 }
 
 #endif
diff --git a/arch/arm/kernel/smp.c b/arch/arm/kernel/smp.c
index addbbe8028c2..f6a4d32b0421 100644
--- a/arch/arm/kernel/smp.c
+++ b/arch/arm/kernel/smp.c
@@ -510,10 +510,6 @@ static void ipi_cpu_stop(unsigned int cpu)
 	local_fiq_disable();
 	local_irq_disable();
 
-#ifdef CONFIG_HOTPLUG_CPU
-	platform_cpu_kill(cpu);
-#endif
-
 	while (1)
 		cpu_relax();
 }
@@ -576,17 +572,25 @@ void smp_send_reschedule(int cpu)
 	smp_cross_call(cpumask_of(cpu), IPI_RESCHEDULE);
 }
 
+#ifdef CONFIG_HOTPLUG_CPU
+static void smp_kill_cpus(cpumask_t *mask)
+{
+	unsigned int cpu;
+	for_each_cpu(cpu, mask)
+		platform_cpu_kill(cpu);
+}
+#else
+static void smp_kill_cpus(cpumask_t *mask) { }
+#endif
+
 void smp_send_stop(void)
 {
 	unsigned long timeout;
+	struct cpumask mask;
 
-	if (num_online_cpus() > 1) {
-		struct cpumask mask;
-		cpumask_copy(&mask, cpu_online_mask);
-		cpumask_clear_cpu(smp_processor_id(), &mask);
-
-		smp_cross_call(&mask, IPI_CPU_STOP);
-	}
+	cpumask_copy(&mask, cpu_online_mask);
+	cpumask_clear_cpu(smp_processor_id(), &mask);
+	smp_cross_call(&mask, IPI_CPU_STOP);
 
 	/* Wait up to one second for other CPUs to stop */
 	timeout = USEC_PER_SEC;
@@ -595,6 +599,8 @@ void smp_send_stop(void)
 
 	if (num_online_cpus() > 1)
 		pr_warning("SMP: failed to stop secondary CPUs\n");
+
+	smp_kill_cpus(&mask);
 }
 
 /*
diff --git a/arch/arm/mach-shmobile/board-ag5evm.c b/arch/arm/mach-shmobile/board-ag5evm.c
index cb224a344af0..0891ec6e27f5 100644
--- a/arch/arm/mach-shmobile/board-ag5evm.c
+++ b/arch/arm/mach-shmobile/board-ag5evm.c
@@ -365,23 +365,13 @@ static struct platform_device mipidsi0_device = {
 };
 
 /* SDHI0 */
-static irqreturn_t ag5evm_sdhi0_gpio_cd(int irq, void *arg)
-{
-	struct device *dev = arg;
-	struct sh_mobile_sdhi_info *info = dev->platform_data;
-	struct tmio_mmc_data *pdata = info->pdata;
-
-	tmio_mmc_cd_wakeup(pdata);
-
-	return IRQ_HANDLED;
-}
-
 static struct sh_mobile_sdhi_info sdhi0_info = {
 	.dma_slave_tx	= SHDMA_SLAVE_SDHI0_TX,
 	.dma_slave_rx	= SHDMA_SLAVE_SDHI0_RX,
-	.tmio_flags	= TMIO_MMC_HAS_IDLE_WAIT,
+	.tmio_flags	= TMIO_MMC_HAS_IDLE_WAIT | TMIO_MMC_USE_GPIO_CD,
 	.tmio_caps	= MMC_CAP_SD_HIGHSPEED,
 	.tmio_ocr_mask	= MMC_VDD_27_28 | MMC_VDD_28_29,
+	.cd_gpio	= GPIO_PORT251,
 };
 
 static struct resource sdhi0_resources[] = {
@@ -557,7 +547,6 @@ static void __init ag5evm_init(void)
 	lcd_backlight_reset();
 
 	/* enable SDHI0 on CN15 [SD I/F] */
-	gpio_request(GPIO_FN_SDHICD0, NULL);
 	gpio_request(GPIO_FN_SDHIWP0, NULL);
 	gpio_request(GPIO_FN_SDHICMD0, NULL);
 	gpio_request(GPIO_FN_SDHICLK0, NULL);
@@ -566,13 +555,6 @@ static void __init ag5evm_init(void)
 	gpio_request(GPIO_FN_SDHID0_1, NULL);
 	gpio_request(GPIO_FN_SDHID0_0, NULL);
 
-	if (!request_irq(intcs_evt2irq(0x3c0), ag5evm_sdhi0_gpio_cd,
-			 IRQF_TRIGGER_FALLING | IRQF_TRIGGER_RISING,
-			 "sdhi0 cd", &sdhi0_device.dev))
-		sdhi0_info.tmio_flags |= TMIO_MMC_HAS_COLD_CD;
-	else
-		pr_warn("Unable to setup SDHI0 GPIO IRQ\n");
-
 	/* enable SDHI1 on CN4 [WLAN I/F] */
 	gpio_request(GPIO_FN_SDHICLK1, NULL);
 	gpio_request(GPIO_FN_SDHICMD1_PU, NULL);
diff --git a/arch/arm/mach-shmobile/board-ap4evb.c b/arch/arm/mach-shmobile/board-ap4evb.c
index b56dde2732bb..0c3caeba2f3e 100644
--- a/arch/arm/mach-shmobile/board-ap4evb.c
+++ b/arch/arm/mach-shmobile/board-ap4evb.c
@@ -997,6 +997,8 @@ static struct sh_mobile_ceu_companion csi2 = {
 
 static struct sh_mobile_ceu_info sh_mobile_ceu_info = {
 	.flags = SH_CEU_FLAG_USE_8BIT_BUS,
+	.max_width = 8188,
+	.max_height = 8188,
 	.csi2 = &csi2,
 };
 
diff --git a/arch/arm/mach-shmobile/board-bonito.c b/arch/arm/mach-shmobile/board-bonito.c
index 81fd95f7f52a..63ab7062bee3 100644
--- a/arch/arm/mach-shmobile/board-bonito.c
+++ b/arch/arm/mach-shmobile/board-bonito.c
@@ -486,7 +486,7 @@ static void __init bonito_earlytimer_init(void)
 	shmobile_earlytimer_init();
 }
 
-void __init bonito_add_early_devices(void)
+static void __init bonito_add_early_devices(void)
 {
 	r8a7740_add_early_devices();
 
diff --git a/arch/arm/mach-shmobile/board-mackerel.c b/arch/arm/mach-shmobile/board-mackerel.c
index f49e28abe0ab..aae2e24fde46 100644
--- a/arch/arm/mach-shmobile/board-mackerel.c
+++ b/arch/arm/mach-shmobile/board-mackerel.c
@@ -908,6 +908,8 @@ fsi_set_rate_end:
 static struct sh_fsi_platform_info fsi_info = {
 	.port_a = {
 		.flags = SH_FSI_BRS_INV,
+		.tx_id = SHDMA_SLAVE_FSIA_TX,
+		.rx_id = SHDMA_SLAVE_FSIA_RX,
 	},
 	.port_b = {
 		.flags = SH_FSI_BRS_INV	|
@@ -920,9 +922,11 @@ static struct sh_fsi_platform_info fsi_info = {
 
 static struct resource fsi_resources[] = {
 	[0] = {
+		/* we need 0xFE1F0000 to access DMA
+		 * instead of 0xFE3C0000 */
 		.name	= "FSI",
-		.start	= 0xFE3C0000,
-		.end	= 0xFE3C0400 - 1,
+		.start  = 0xFE1F0000,
+		.end    = 0xFE1F0400 - 1,
 		.flags	= IORESOURCE_MEM,
 	},
 	[1] = {
@@ -1011,21 +1015,12 @@ static int slot_cn7_get_cd(struct platform_device *pdev)
 }
 
 /* SDHI0 */
-static irqreturn_t mackerel_sdhi0_gpio_cd(int irq, void *arg)
-{
-	struct device *dev = arg;
-	struct sh_mobile_sdhi_info *info = dev->platform_data;
-	struct tmio_mmc_data *pdata = info->pdata;
-
-	tmio_mmc_cd_wakeup(pdata);
-
-	return IRQ_HANDLED;
-}
-
 static struct sh_mobile_sdhi_info sdhi0_info = {
 	.dma_slave_tx	= SHDMA_SLAVE_SDHI0_TX,
 	.dma_slave_rx	= SHDMA_SLAVE_SDHI0_RX,
+	.tmio_flags	= TMIO_MMC_USE_GPIO_CD,
 	.tmio_caps	= MMC_CAP_SD_HIGHSPEED | MMC_CAP_SDIO_IRQ,
+	.cd_gpio	= GPIO_PORT172,
 };
 
 static struct resource sdhi0_resources[] = {
@@ -1257,6 +1252,8 @@ static void mackerel_camera_del(struct soc_camera_device *icd)
 
 static struct sh_mobile_ceu_info sh_mobile_ceu_info = {
 	.flags = SH_CEU_FLAG_USE_8BIT_BUS,
+	.max_width = 8188,
+	.max_height = 8188,
 };
 
 static struct resource ceu_resources[] = {
@@ -1384,7 +1381,6 @@ static void __init mackerel_init(void)
 {
 	u32 srcr4;
 	struct clk *clk;
-	int ret;
 
 	/* External clock source */
 	clk_set_rate(&sh7372_dv_clki_clk, 27000000);
@@ -1481,7 +1477,6 @@ static void __init mackerel_init(void)
 	irq_set_irq_type(IRQ21, IRQ_TYPE_LEVEL_HIGH);
 
 	/* enable SDHI0 */
-	gpio_request(GPIO_FN_SDHICD0, NULL);
 	gpio_request(GPIO_FN_SDHIWP0, NULL);
 	gpio_request(GPIO_FN_SDHICMD0, NULL);
 	gpio_request(GPIO_FN_SDHICLK0, NULL);
@@ -1490,13 +1485,6 @@ static void __init mackerel_init(void)
 	gpio_request(GPIO_FN_SDHID0_1, NULL);
 	gpio_request(GPIO_FN_SDHID0_0, NULL);
 
-	ret = request_irq(evt2irq(0x3340), mackerel_sdhi0_gpio_cd,
-			  IRQF_TRIGGER_FALLING, "sdhi0 cd", &sdhi0_device.dev);
-	if (!ret)
-		sdhi0_info.tmio_flags |= TMIO_MMC_HAS_COLD_CD;
-	else
-		pr_err("Cannot get IRQ #%d: %d\n", evt2irq(0x3340), ret);
-
 #if !defined(CONFIG_MMC_SH_MMCIF) && !defined(CONFIG_MMC_SH_MMCIF_MODULE)
 	/* enable SDHI1 */
 	gpio_request(GPIO_FN_SDHICMD1, NULL);
diff --git a/arch/arm/mach-shmobile/clock-r8a7740.c b/arch/arm/mach-shmobile/clock-r8a7740.c
index 99c4d743a99c..81b54a6af20f 100644
--- a/arch/arm/mach-shmobile/clock-r8a7740.c
+++ b/arch/arm/mach-shmobile/clock-r8a7740.c
@@ -47,6 +47,7 @@
 #define PLLC01CR	0xe6150028
 
 #define SUBCKCR		0xe6150080
+#define USBCKCR		0xe615008c
 
 #define MSTPSR0		0xe6150030
 #define MSTPSR1		0xe6150038
@@ -181,6 +182,95 @@ static struct clk pllc1_div2_clk = {
 	.parent		= &pllc1_clk,
 };
 
+/* USB clock */
+static struct clk *usb24s_parents[] = {
+	[0] = &system_clk,
+	[1] = &extal2_clk
+};
+
+static int usb24s_enable(struct clk *clk)
+{
+	__raw_writel(__raw_readl(USBCKCR) & ~(1 << 8), USBCKCR);
+
+	return 0;
+}
+
+static void usb24s_disable(struct clk *clk)
+{
+	__raw_writel(__raw_readl(USBCKCR) | (1 << 8), USBCKCR);
+}
+
+static int usb24s_set_parent(struct clk *clk, struct clk *parent)
+{
+	int i, ret;
+	u32 val;
+
+	if (!clk->parent_table || !clk->parent_num)
+		return -EINVAL;
+
+	/* Search the parent */
+	for (i = 0; i < clk->parent_num; i++)
+		if (clk->parent_table[i] == parent)
+			break;
+
+	if (i == clk->parent_num)
+		return -ENODEV;
+
+	ret = clk_reparent(clk, parent);
+	if (ret < 0)
+		return ret;
+
+	val = __raw_readl(USBCKCR);
+	val &= ~(1 << 7);
+	val |= i << 7;
+	__raw_writel(val, USBCKCR);
+
+	return 0;
+}
+
+static struct sh_clk_ops usb24s_clk_ops = {
+	.recalc		= followparent_recalc,
+	.enable		= usb24s_enable,
+	.disable	= usb24s_disable,
+	.set_parent	= usb24s_set_parent,
+};
+
+static struct clk usb24s_clk = {
+	.ops		= &usb24s_clk_ops,
+	.parent_table	= usb24s_parents,
+	.parent_num	= ARRAY_SIZE(usb24s_parents),
+	.parent		= &system_clk,
+};
+
+static unsigned long usb24_recalc(struct clk *clk)
+{
+	return clk->parent->rate /
+		((__raw_readl(USBCKCR) & (1 << 6)) ? 1 : 2);
+};
+
+static int usb24_set_rate(struct clk *clk, unsigned long rate)
+{
+	u32 val;
+
+	/* closer to which ? parent->rate or parent->rate/2 */
+	val = __raw_readl(USBCKCR);
+	val &= ~(1 << 6);
+	val |= (rate > (clk->parent->rate / 4) * 3) << 6;
+	__raw_writel(val, USBCKCR);
+
+	return 0;
+}
+
+static struct sh_clk_ops usb24_clk_ops = {
+	.recalc		= usb24_recalc,
+	.set_rate	= usb24_set_rate,
+};
+
+static struct clk usb24_clk = {
+	.ops		= &usb24_clk_ops,
+	.parent		= &usb24s_clk,
+};
+
 struct clk *main_clks[] = {
 	&extalr_clk,
 	&extal1_clk,
@@ -196,6 +286,8 @@ struct clk *main_clks[] = {
 	&pllc0_clk,
 	&pllc1_clk,
 	&pllc1_div2_clk,
+	&usb24s_clk,
+	&usb24_clk,
 };
 
 static void div4_kick(struct clk *clk)
@@ -223,7 +315,7 @@ static struct clk_div4_table div4_table = {
 
 enum {
 	DIV4_I, DIV4_ZG, DIV4_B, DIV4_M1, DIV4_HP,
-	DIV4_HPP, DIV4_S, DIV4_ZB, DIV4_M3, DIV4_CP,
+	DIV4_HPP, DIV4_USBP, DIV4_S, DIV4_ZB, DIV4_M3, DIV4_CP,
 	DIV4_NR
 };
 
@@ -234,6 +326,7 @@ struct clk div4_clks[DIV4_NR] = {
 	[DIV4_M1]	= SH_CLK_DIV4(&pllc1_clk, FRQCRA,  4, 0x6fff, CLK_ENABLE_ON_INIT),
 	[DIV4_HP]	= SH_CLK_DIV4(&pllc1_clk, FRQCRB,  4, 0x6fff, 0),
 	[DIV4_HPP]	= SH_CLK_DIV4(&pllc1_clk, FRQCRC, 20, 0x6fff, 0),
+	[DIV4_USBP]	= SH_CLK_DIV4(&pllc1_clk, FRQCRC, 16, 0x6fff, 0),
 	[DIV4_S]	= SH_CLK_DIV4(&pllc1_clk, FRQCRC, 12, 0x6fff, 0),
 	[DIV4_ZB]	= SH_CLK_DIV4(&pllc1_clk, FRQCRC,  8, 0x6fff, 0),
 	[DIV4_M3]	= SH_CLK_DIV4(&pllc1_clk, FRQCRC,  4, 0x6fff, 0),
@@ -257,7 +350,10 @@ enum {
 	MSTP222,
 	MSTP207, MSTP206, MSTP204, MSTP203, MSTP202, MSTP201, MSTP200,
 
-	MSTP329, MSTP323,
+	MSTP329, MSTP328, MSTP323, MSTP320,
+	MSTP314, MSTP313, MSTP312,
+
+	MSTP416, MSTP415, MSTP407, MSTP406,
 
 	MSTP_NR
 };
@@ -280,7 +376,17 @@ static struct clk mstp_clks[MSTP_NR] = {
 	[MSTP200] = SH_CLK_MSTP32(&div6_clks[DIV6_SUB],	SMSTPCR2,  0, 0), /* SCIFA4 */
 
 	[MSTP329] = SH_CLK_MSTP32(&r_clk,		SMSTPCR3, 29, 0), /* CMT10 */
+	[MSTP328] = SH_CLK_MSTP32(&div4_clks[DIV4_HP],	SMSTPCR3, 28, 0), /* FSI */
 	[MSTP323] = SH_CLK_MSTP32(&div6_clks[DIV6_SUB],	SMSTPCR3, 23, 0), /* IIC1 */
+	[MSTP320] = SH_CLK_MSTP32(&div4_clks[DIV4_HP],	SMSTPCR3, 20, 0), /* USBF */
+	[MSTP314] = SH_CLK_MSTP32(&div4_clks[DIV4_HP],	SMSTPCR3, 14, 0), /* SDHI0 */
+	[MSTP313] = SH_CLK_MSTP32(&div4_clks[DIV4_HP],	SMSTPCR3, 13, 0), /* SDHI1 */
+	[MSTP312] = SH_CLK_MSTP32(&div4_clks[DIV4_HP],	SMSTPCR3, 12, 0), /* MMC */
+
+	[MSTP416] = SH_CLK_MSTP32(&div4_clks[DIV4_HP],	SMSTPCR4, 16, 0), /* USBHOST */
+	[MSTP415] = SH_CLK_MSTP32(&div4_clks[DIV4_HP],	SMSTPCR4, 15, 0), /* SDHI2 */
+	[MSTP407] = SH_CLK_MSTP32(&div4_clks[DIV4_HP],	SMSTPCR4,  7, 0), /* USB-Func */
+	[MSTP406] = SH_CLK_MSTP32(&div4_clks[DIV4_HP],	SMSTPCR4,  6, 0), /* USB Phy */
 };
 
 static struct clk_lookup lookups[] = {
@@ -299,6 +405,7 @@ static struct clk_lookup lookups[] = {
 	CLKDEV_CON_ID("pllc0_clk",		&pllc0_clk),
 	CLKDEV_CON_ID("pllc1_clk",		&pllc1_clk),
 	CLKDEV_CON_ID("pllc1_div2_clk",		&pllc1_div2_clk),
+	CLKDEV_CON_ID("usb24s",			&usb24s_clk),
 
 	/* DIV4 clocks */
 	CLKDEV_CON_ID("i_clk",			&div4_clks[DIV4_I]),
@@ -334,7 +441,21 @@ static struct clk_lookup lookups[] = {
 	CLKDEV_DEV_ID("sh-sci.6",		&mstp_clks[MSTP230]),
 
 	CLKDEV_DEV_ID("sh_cmt.10",		&mstp_clks[MSTP329]),
+	CLKDEV_DEV_ID("sh_fsi2",		&mstp_clks[MSTP328]),
 	CLKDEV_DEV_ID("i2c-sh_mobile.1",	&mstp_clks[MSTP323]),
+	CLKDEV_DEV_ID("renesas_usbhs",		&mstp_clks[MSTP320]),
+	CLKDEV_DEV_ID("sh_mobile_sdhi.0",	&mstp_clks[MSTP314]),
+	CLKDEV_DEV_ID("sh_mobile_sdhi.1",	&mstp_clks[MSTP313]),
+	CLKDEV_DEV_ID("sh_mmcif",		&mstp_clks[MSTP312]),
+
+	CLKDEV_DEV_ID("sh_mobile_sdhi.2",	&mstp_clks[MSTP415]),
+
+	/* ICK */
+	CLKDEV_ICK_ID("host",	"renesas_usbhs",	&mstp_clks[MSTP416]),
+	CLKDEV_ICK_ID("func",	"renesas_usbhs",	&mstp_clks[MSTP407]),
+	CLKDEV_ICK_ID("phy",	"renesas_usbhs",	&mstp_clks[MSTP406]),
+	CLKDEV_ICK_ID("pci",	"renesas_usbhs",	&div4_clks[DIV4_USBP]),
+	CLKDEV_ICK_ID("usb24",	"renesas_usbhs",	&usb24_clk),
 };
 
 void __init r8a7740_clock_init(u8 md_ck)
diff --git a/arch/arm/mach-shmobile/headsmp.S b/arch/arm/mach-shmobile/headsmp.S
index 6ac015c89206..b202c1272526 100644
--- a/arch/arm/mach-shmobile/headsmp.S
+++ b/arch/arm/mach-shmobile/headsmp.S
@@ -16,6 +16,59 @@
 
 	__CPUINIT
 
+/* Cache invalidation nicked from arch/arm/mach-imx/head-v7.S, thanks!
+ *
+ * The secondary kernel init calls v7_flush_dcache_all before it enables
+ * the L1; however, the L1 comes out of reset in an undefined state, so
+ * the clean + invalidate performed by v7_flush_dcache_all causes a bunch
+ * of cache lines with uninitialized data and uninitialized tags to get
+ * written out to memory, which does really unpleasant things to the main
+ * processor.  We fix this by performing an invalidate, rather than a
+ * clean + invalidate, before jumping into the kernel.
+ *
+ * This funciton is cloned from arch/arm/mach-tegra/headsmp.S, and needs
+ * to be called for both secondary cores startup and primary core resume
+ * procedures.  Ideally, it should be moved into arch/arm/mm/cache-v7.S.
+ */
+ENTRY(v7_invalidate_l1)
+	mov	r0, #0
+	mcr	p15, 0, r0, c7, c5, 0	@ invalidate I cache
+	mcr	p15, 2, r0, c0, c0, 0
+	mrc	p15, 1, r0, c0, c0, 0
+
+	ldr	r1, =0x7fff
+	and	r2, r1, r0, lsr #13
+
+	ldr	r1, =0x3ff
+
+	and	r3, r1, r0, lsr #3	@ NumWays - 1
+	add	r2, r2, #1		@ NumSets
+
+	and	r0, r0, #0x7
+	add	r0, r0, #4	@ SetShift
+
+	clz	r1, r3		@ WayShift
+	add	r4, r3, #1	@ NumWays
+1:	sub	r2, r2, #1	@ NumSets--
+	mov	r3, r4		@ Temp = NumWays
+2:	subs	r3, r3, #1	@ Temp--
+	mov	r5, r3, lsl r1
+	mov	r6, r2, lsl r0
+	orr	r5, r5, r6	@ Reg = (Temp<<WayShift)|(NumSets<<SetShift)
+	mcr	p15, 0, r5, c7, c6, 2
+	bgt	2b
+	cmp	r2, #0
+	bgt	1b
+	dsb
+	isb
+	mov	pc, lr
+ENDPROC(v7_invalidate_l1)
+
+ENTRY(shmobile_invalidate_start)
+	bl	v7_invalidate_l1
+	b	secondary_startup
+ENDPROC(shmobile_invalidate_start)
+
 /*
  * Reset vector for secondary CPUs.
  * This will be mapped at address 0 by SBAR register.
@@ -24,4 +77,5 @@
 	.align  12
 ENTRY(shmobile_secondary_vector)
 	ldr     pc, 1f
-1:	.long   secondary_startup - PAGE_OFFSET + PLAT_PHYS_OFFSET
+1:	.long   shmobile_invalidate_start - PAGE_OFFSET + PLAT_PHYS_OFFSET
+ENDPROC(shmobile_secondary_vector)
diff --git a/arch/arm/mach-shmobile/include/mach/common.h b/arch/arm/mach-shmobile/include/mach/common.h
index 83ad3fe0a75f..ff5f12fd742f 100644
--- a/arch/arm/mach-shmobile/include/mach/common.h
+++ b/arch/arm/mach-shmobile/include/mach/common.h
@@ -3,8 +3,9 @@
 
 extern void shmobile_earlytimer_init(void);
 extern struct sys_timer shmobile_timer;
+extern void shmobile_setup_delay(unsigned int max_cpu_core_mhz,
+				 unsigned int mult, unsigned int div);
 struct twd_local_timer;
-void shmobile_twd_init(struct twd_local_timer *twd_local_timer);
 extern void shmobile_setup_console(void);
 extern void shmobile_secondary_vector(void);
 extern int shmobile_platform_cpu_kill(unsigned int cpu);
@@ -82,5 +83,6 @@ extern int r8a7779_platform_cpu_kill(unsigned int cpu);
 extern void r8a7779_secondary_init(unsigned int cpu);
 extern int r8a7779_boot_secondary(unsigned int cpu);
 extern void r8a7779_smp_prepare_cpus(void);
+extern void r8a7779_register_twd(void);
 
 #endif /* __ARCH_MACH_COMMON_H */
diff --git a/arch/arm/mach-shmobile/include/mach/intc.h b/arch/arm/mach-shmobile/include/mach/intc.h
index 8b22258c8caa..a5603c76cfe0 100644
--- a/arch/arm/mach-shmobile/include/mach/intc.h
+++ b/arch/arm/mach-shmobile/include/mach/intc.h
@@ -142,6 +142,50 @@ static struct intc_desc p ## _desc __initdata = {			\
 			     p ## _sense_registers, p ## _ack_registers) \
 }
 
+#define INTC_IRQ_PINS_16H(p, base, vect, str)				\
+									\
+static struct resource p ## _resources[] __initdata = {			\
+	[0] = {								\
+		.start	= base,						\
+		.end	= base + 0x64,					\
+		.flags	= IORESOURCE_MEM,				\
+	},								\
+};									\
+									\
+enum {									\
+	p ## _UNUSED = 0,						\
+	INTC_IRQ_PINS_ENUM_16H(p),					\
+};									\
+									\
+static struct intc_vect p ## _vectors[] __initdata = {			\
+	INTC_IRQ_PINS_VECT_16H(p, vect),				\
+};									\
+									\
+static struct intc_mask_reg p ## _mask_registers[] __initdata = {	\
+	INTC_IRQ_PINS_MASK_16H(p, base),				\
+};									\
+									\
+static struct intc_prio_reg p ## _prio_registers[] __initdata = {	\
+	INTC_IRQ_PINS_PRIO_16H(p, base),				\
+};									\
+									\
+static struct intc_sense_reg p ## _sense_registers[] __initdata = {	\
+	INTC_IRQ_PINS_SENSE_16H(p, base),				\
+};									\
+									\
+static struct intc_mask_reg p ## _ack_registers[] __initdata = {	\
+	INTC_IRQ_PINS_ACK_16H(p, base),					\
+};									\
+									\
+static struct intc_desc p ## _desc __initdata = {			\
+	.name = str,							\
+	.resource = p ## _resources,					\
+	.num_resources = ARRAY_SIZE(p ## _resources),			\
+	.hw = INTC_HW_DESC(p ## _vectors, NULL,				\
+			     p ## _mask_registers, p ## _prio_registers, \
+			     p ## _sense_registers, p ## _ack_registers) \
+}
+
 #define INTC_IRQ_PINS_32(p, base, vect, str)				\
 									\
 static struct resource p ## _resources[] __initdata = {			\
diff --git a/arch/arm/mach-shmobile/include/mach/irqs.h b/arch/arm/mach-shmobile/include/mach/irqs.h
index 4e686cc201fc..06a5da3c3050 100644
--- a/arch/arm/mach-shmobile/include/mach/irqs.h
+++ b/arch/arm/mach-shmobile/include/mach/irqs.h
@@ -7,7 +7,7 @@
 #define gic_spi(nr)		((nr) + 32)
 
 /* INTCS */
-#define INTCS_VECT_BASE		0x2200
+#define INTCS_VECT_BASE		0x3400
 #define INTCS_VECT(n, vect)	INTC_VECT((n), INTCS_VECT_BASE + (vect))
 #define intcs_evt2irq(evt)	evt2irq(INTCS_VECT_BASE + (evt))
 
diff --git a/arch/arm/mach-shmobile/include/mach/sh7372.h b/arch/arm/mach-shmobile/include/mach/sh7372.h
index 8254ab86f6cd..915d0093da08 100644
--- a/arch/arm/mach-shmobile/include/mach/sh7372.h
+++ b/arch/arm/mach-shmobile/include/mach/sh7372.h
@@ -457,6 +457,8 @@ enum {
 	SHDMA_SLAVE_SDHI1_TX,
 	SHDMA_SLAVE_SDHI2_RX,
 	SHDMA_SLAVE_SDHI2_TX,
+	SHDMA_SLAVE_FSIA_RX,
+	SHDMA_SLAVE_FSIA_TX,
 	SHDMA_SLAVE_MMCIF_RX,
 	SHDMA_SLAVE_MMCIF_TX,
 	SHDMA_SLAVE_USB0_TX,
diff --git a/arch/arm/mach-shmobile/intc-sh7372.c b/arch/arm/mach-shmobile/intc-sh7372.c
index 6447e0af52d4..2587a22842f2 100644
--- a/arch/arm/mach-shmobile/intc-sh7372.c
+++ b/arch/arm/mach-shmobile/intc-sh7372.c
@@ -19,6 +19,7 @@
 #include <linux/kernel.h>
 #include <linux/init.h>
 #include <linux/interrupt.h>
+#include <linux/module.h>
 #include <linux/irq.h>
 #include <linux/io.h>
 #include <linux/sh_intc.h>
@@ -305,14 +306,16 @@ static DECLARE_INTC_DESC(intca_desc, "sh7372-intca",
 			 intca_mask_registers, intca_prio_registers,
 			 NULL);
 
-INTC_IRQ_PINS_32(intca_irq_pins, 0xe6900000,
-		 INTC_VECT, "sh7372-intca-irq-pins");
+INTC_IRQ_PINS_16(intca_irq_pins_lo, 0xe6900000,
+		 INTC_VECT, "sh7372-intca-irq-lo");
+
+INTC_IRQ_PINS_16H(intca_irq_pins_hi, 0xe6900000,
+		 INTC_VECT, "sh7372-intca-irq-hi");
+
 enum {
 	UNUSED_INTCS = 0,
 	ENABLED_INTCS,
 
-	INTCS,
-
 	/* interrupt sources INTCS */
 
 	/* IRQ0S - IRQ31S */
@@ -426,8 +429,6 @@ static struct intc_vect intcs_vectors[] = {
 	INTCS_VECT(CPORTS2R, 0x1a20),
 	/* CEC */
 	INTCS_VECT(JPU6E, 0x1a80),
-
-	INTC_VECT(INTCS, 0xf80),
 };
 
 static struct intc_group intcs_groups[] __initdata = {
@@ -490,9 +491,6 @@ static struct intc_mask_reg intcs_mask_registers[] = {
 	{ 0xffd5019c, 0xffd501dc, 8, /* IMR7SA3 / IMCR7SA3 */
 	  { MFIS2_INTCS, CPORTS2R, 0, 0,
 	    JPU6E, 0, 0, 0 } },
-	{ 0xffd20104, 0, 16, /* INTAMASK */
-	  { 0, 0, 0, 0, 0, 0, 0, 0,
-	    0, 0, 0, 0, 0, 0, 0, INTCS } },
 };
 
 /* Priority is needed for INTCA to receive the INTCS interrupt */
@@ -557,18 +555,30 @@ static void __iomem *intcs_ffd5;
 void __init sh7372_init_irq(void)
 {
 	void __iomem *intevtsa;
+	int n;
 
 	intcs_ffd2 = ioremap_nocache(0xffd20000, PAGE_SIZE);
 	intevtsa = intcs_ffd2 + 0x100;
 	intcs_ffd5 = ioremap_nocache(0xffd50000, PAGE_SIZE);
 
 	register_intc_controller(&intca_desc);
-	register_intc_controller(&intca_irq_pins_desc);
+	register_intc_controller(&intca_irq_pins_lo_desc);
+	register_intc_controller(&intca_irq_pins_hi_desc);
 	register_intc_controller(&intcs_desc);
 
+	/* setup dummy cascade chip for INTCS */
+	n = evt2irq(0xf80);
+	irq_alloc_desc_at(n, numa_node_id());
+	irq_set_chip_and_handler_name(n, &dummy_irq_chip,
+				      handle_level_irq, "level");
+	set_irq_flags(n, IRQF_VALID); /* yuck */
+
 	/* demux using INTEVTSA */
-	irq_set_handler_data(evt2irq(0xf80), (void *)intevtsa);
-	irq_set_chained_handler(evt2irq(0xf80), intcs_demux);
+	irq_set_handler_data(n, (void *)intevtsa);
+	irq_set_chained_handler(n, intcs_demux);
+
+	/* unmask INTCS in INTAMASK */
+	iowrite16(0, intcs_ffd2 + 0x104);
 }
 
 static unsigned short ffd2[0x200];
diff --git a/arch/arm/mach-shmobile/pfc-r8a7740.c b/arch/arm/mach-shmobile/pfc-r8a7740.c
index a4fff6950b03..670fe1869dbc 100644
--- a/arch/arm/mach-shmobile/pfc-r8a7740.c
+++ b/arch/arm/mach-shmobile/pfc-r8a7740.c
@@ -22,6 +22,7 @@
 #include <linux/kernel.h>
 #include <linux/gpio.h>
 #include <mach/r8a7740.h>
+#include <mach/irqs.h>
 
 #define CPU_ALL_PORT(fn, pfx, sfx)					\
 	PORT_10(fn, pfx, sfx),		PORT_90(fn, pfx, sfx),		\
@@ -2527,6 +2528,41 @@ static struct pinmux_data_reg pinmux_data_regs[] = {
 	{ },
 };
 
+static struct pinmux_irq pinmux_irqs[] = {
+	PINMUX_IRQ(evt2irq(0x0200), PORT2_FN0,	 PORT13_FN0),	/* IRQ0A */
+	PINMUX_IRQ(evt2irq(0x0220), PORT20_FN0),		/* IRQ1A */
+	PINMUX_IRQ(evt2irq(0x0240), PORT11_FN0,	 PORT12_FN0),	/* IRQ2A */
+	PINMUX_IRQ(evt2irq(0x0260), PORT10_FN0,	 PORT14_FN0),	/* IRQ3A */
+	PINMUX_IRQ(evt2irq(0x0280), PORT15_FN0,	 PORT172_FN0),	/* IRQ4A */
+	PINMUX_IRQ(evt2irq(0x02A0), PORT0_FN0,	 PORT1_FN0),	/* IRQ5A */
+	PINMUX_IRQ(evt2irq(0x02C0), PORT121_FN0, PORT173_FN0),	/* IRQ6A */
+	PINMUX_IRQ(evt2irq(0x02E0), PORT120_FN0, PORT209_FN0),	/* IRQ7A */
+	PINMUX_IRQ(evt2irq(0x0300), PORT119_FN0),		/* IRQ8A */
+	PINMUX_IRQ(evt2irq(0x0320), PORT118_FN0, PORT210_FN0),	/* IRQ9A */
+	PINMUX_IRQ(evt2irq(0x0340), PORT19_FN0),		/* IRQ10A */
+	PINMUX_IRQ(evt2irq(0x0360), PORT104_FN0),		/* IRQ11A */
+	PINMUX_IRQ(evt2irq(0x0380), PORT42_FN0,	 PORT97_FN0),	/* IRQ12A */
+	PINMUX_IRQ(evt2irq(0x03A0), PORT64_FN0,	 PORT98_FN0),	/* IRQ13A */
+	PINMUX_IRQ(evt2irq(0x03C0), PORT63_FN0,	 PORT99_FN0),	/* IRQ14A */
+	PINMUX_IRQ(evt2irq(0x03E0), PORT62_FN0,	 PORT100_FN0),	/* IRQ15A */
+	PINMUX_IRQ(evt2irq(0x3200), PORT68_FN0,	 PORT211_FN0),	/* IRQ16A */
+	PINMUX_IRQ(evt2irq(0x3220), PORT69_FN0),		/* IRQ17A */
+	PINMUX_IRQ(evt2irq(0x3240), PORT70_FN0),		/* IRQ18A */
+	PINMUX_IRQ(evt2irq(0x3260), PORT71_FN0),		/* IRQ19A */
+	PINMUX_IRQ(evt2irq(0x3280), PORT67_FN0),		/* IRQ20A */
+	PINMUX_IRQ(evt2irq(0x32A0), PORT202_FN0),		/* IRQ21A */
+	PINMUX_IRQ(evt2irq(0x32C0), PORT95_FN0),		/* IRQ22A */
+	PINMUX_IRQ(evt2irq(0x32E0), PORT96_FN0),		/* IRQ23A */
+	PINMUX_IRQ(evt2irq(0x3300), PORT180_FN0),		/* IRQ24A */
+	PINMUX_IRQ(evt2irq(0x3320), PORT38_FN0),		/* IRQ25A */
+	PINMUX_IRQ(evt2irq(0x3340), PORT58_FN0,	 PORT81_FN0),	/* IRQ26A */
+	PINMUX_IRQ(evt2irq(0x3360), PORT57_FN0,	 PORT168_FN0),	/* IRQ27A */
+	PINMUX_IRQ(evt2irq(0x3380), PORT56_FN0,	 PORT169_FN0),	/* IRQ28A */
+	PINMUX_IRQ(evt2irq(0x33A0), PORT50_FN0,	 PORT170_FN0),	/* IRQ29A */
+	PINMUX_IRQ(evt2irq(0x33C0), PORT49_FN0,	 PORT171_FN0),	/* IRQ30A */
+	PINMUX_IRQ(evt2irq(0x33E0), PORT41_FN0,	 PORT167_FN0),	/* IRQ31A */
+};
+
 static struct pinmux_info r8a7740_pinmux_info = {
 	.name		= "r8a7740_pfc",
 	.reserved_id	= PINMUX_RESERVED,
@@ -2554,6 +2590,9 @@ static struct pinmux_info r8a7740_pinmux_info = {
 
 	.gpio_data	= pinmux_data,
 	.gpio_data_size	= ARRAY_SIZE(pinmux_data),
+
+	.gpio_irq	= pinmux_irqs,
+	.gpio_irq_size	= ARRAY_SIZE(pinmux_irqs),
 };
 
 void r8a7740_pinmux_init(void)
diff --git a/arch/arm/mach-shmobile/setup-r8a7740.c b/arch/arm/mach-shmobile/setup-r8a7740.c
index 14edb5cffa7f..34a1548dbda6 100644
--- a/arch/arm/mach-shmobile/setup-r8a7740.c
+++ b/arch/arm/mach-shmobile/setup-r8a7740.c
@@ -350,19 +350,19 @@ static void r8a7740_i2c_workaround(struct platform_device *pdev)
 	i2c_write(reg, ICSTART, i2c_read(reg, ICSTART) | 0x10);
 	i2c_read(reg, ICSTART); /* dummy read */
 
-	mdelay(100);
+	udelay(10);
 
 	i2c_write(reg, ICCR, 0x01);
-	i2c_read(reg, ICCR);
 	i2c_write(reg, ICSTART, 0x00);
-	i2c_read(reg, ICSTART);
+
+	udelay(10);
 
 	i2c_write(reg, ICCR, 0x10);
-	mdelay(100);
+	udelay(10);
 	i2c_write(reg, ICCR, 0x00);
-	mdelay(100);
+	udelay(10);
 	i2c_write(reg, ICCR, 0x10);
-	mdelay(100);
+	udelay(10);
 
 	iounmap(reg);
 }
diff --git a/arch/arm/mach-shmobile/setup-r8a7779.c b/arch/arm/mach-shmobile/setup-r8a7779.c
index 12c6f529ab89..e98e46f6cf55 100644
--- a/arch/arm/mach-shmobile/setup-r8a7779.c
+++ b/arch/arm/mach-shmobile/setup-r8a7779.c
@@ -262,10 +262,14 @@ void __init r8a7779_add_standard_devices(void)
 			    ARRAY_SIZE(r8a7779_late_devices));
 }
 
+/* do nothing for !CONFIG_SMP or !CONFIG_HAVE_TWD */
+void __init __weak r8a7779_register_twd(void) { }
+
 static void __init r8a7779_earlytimer_init(void)
 {
 	r8a7779_clock_init();
 	shmobile_earlytimer_init();
+	r8a7779_register_twd();
 }
 
 void __init r8a7779_add_early_devices(void)
diff --git a/arch/arm/mach-shmobile/setup-sh7372.c b/arch/arm/mach-shmobile/setup-sh7372.c
index 2fe8f83ca124..6a4bd582c028 100644
--- a/arch/arm/mach-shmobile/setup-sh7372.c
+++ b/arch/arm/mach-shmobile/setup-sh7372.c
@@ -22,6 +22,7 @@
 #include <linux/interrupt.h>
 #include <linux/irq.h>
 #include <linux/platform_device.h>
+#include <linux/of_platform.h>
 #include <linux/uio_driver.h>
 #include <linux/delay.h>
 #include <linux/input.h>
@@ -461,6 +462,16 @@ static const struct sh_dmae_slave_config sh7372_dmae_slaves[] = {
 		.chcr		= DM_INC | SM_FIX | 0x800 | TS_INDEX2VAL(XMIT_SZ_16BIT),
 		.mid_rid	= 0xce,
 	}, {
+		.slave_id	= SHDMA_SLAVE_FSIA_TX,
+		.addr		= 0xfe1f0024,
+		.chcr		= DM_FIX | SM_INC | 0x800 | TS_INDEX2VAL(XMIT_SZ_32BIT),
+		.mid_rid	= 0xb1,
+	}, {
+		.slave_id	= SHDMA_SLAVE_FSIA_RX,
+		.addr		= 0xfe1f0020,
+		.chcr		= DM_INC | SM_FIX | 0x800 | TS_INDEX2VAL(XMIT_SZ_32BIT),
+		.mid_rid	= 0xb2,
+	}, {
 		.slave_id	= SHDMA_SLAVE_MMCIF_TX,
 		.addr		= 0xe6bd0034,
 		.chcr		= DM_FIX | SM_INC | 0x800 | TS_INDEX2VAL(XMIT_SZ_32BIT),
@@ -1092,3 +1103,50 @@ void __init sh7372_add_early_devices(void)
 	/* override timer setup with soc-specific code */
 	shmobile_timer.init = sh7372_earlytimer_init;
 }
+
+#ifdef CONFIG_USE_OF
+
+void __init sh7372_add_early_devices_dt(void)
+{
+	shmobile_setup_delay(800, 1, 3); /* Cortex-A8 @ 800MHz */
+
+	early_platform_add_devices(sh7372_early_devices,
+				   ARRAY_SIZE(sh7372_early_devices));
+
+	/* setup early console here as well */
+	shmobile_setup_console();
+}
+
+static const struct of_dev_auxdata sh7372_auxdata_lookup[] __initconst = {
+	{ }
+};
+
+void __init sh7372_add_standard_devices_dt(void)
+{
+	/* clocks are setup late during boot in the case of DT */
+	sh7372_clock_init();
+
+	platform_add_devices(sh7372_early_devices,
+			    ARRAY_SIZE(sh7372_early_devices));
+
+	of_platform_populate(NULL, of_default_bus_match_table,
+			     sh7372_auxdata_lookup, NULL);
+}
+
+static const char *sh7372_boards_compat_dt[] __initdata = {
+	"renesas,sh7372",
+	NULL,
+};
+
+DT_MACHINE_START(SH7372_DT, "Generic SH7372 (Flattened Device Tree)")
+	.map_io		= sh7372_map_io,
+	.init_early	= sh7372_add_early_devices_dt,
+	.nr_irqs	= NR_IRQS_LEGACY,
+	.init_irq	= sh7372_init_irq,
+	.handle_irq	= shmobile_handle_irq_intc,
+	.init_machine	= sh7372_add_standard_devices_dt,
+	.timer		= &shmobile_timer,
+	.dt_compat	= sh7372_boards_compat_dt,
+MACHINE_END
+
+#endif /* CONFIG_USE_OF */
diff --git a/arch/arm/mach-shmobile/setup-sh73a0.c b/arch/arm/mach-shmobile/setup-sh73a0.c
index 5bebffc10455..04a0dfe75493 100644
--- a/arch/arm/mach-shmobile/setup-sh73a0.c
+++ b/arch/arm/mach-shmobile/setup-sh73a0.c
@@ -688,10 +688,14 @@ void __init sh73a0_add_standard_devices(void)
 			    ARRAY_SIZE(sh73a0_late_devices));
 }
 
+/* do nothing for !CONFIG_SMP or !CONFIG_HAVE_TWD */
+void __init __weak sh73a0_register_twd(void) { }
+
 static void __init sh73a0_earlytimer_init(void)
 {
 	sh73a0_clock_init();
 	shmobile_earlytimer_init();
+	sh73a0_register_twd();
 }
 
 void __init sh73a0_add_early_devices(void)
diff --git a/arch/arm/mach-shmobile/smp-r8a7779.c b/arch/arm/mach-shmobile/smp-r8a7779.c
index b62e19d4c9af..6d1d0238cbf7 100644
--- a/arch/arm/mach-shmobile/smp-r8a7779.c
+++ b/arch/arm/mach-shmobile/smp-r8a7779.c
@@ -64,8 +64,15 @@ static void __iomem *scu_base_addr(void)
 static DEFINE_SPINLOCK(scu_lock);
 static unsigned long tmp;
 
+#ifdef CONFIG_HAVE_ARM_TWD
 static DEFINE_TWD_LOCAL_TIMER(twd_local_timer, 0xf0000600, 29);
 
+void __init r8a7779_register_twd(void)
+{
+	twd_local_timer_register(&twd_local_timer);
+}
+#endif
+
 static void modify_scu_cpu_psr(unsigned long set, unsigned long clr)
 {
 	void __iomem *scu_base = scu_base_addr();
@@ -84,7 +91,6 @@ unsigned int __init r8a7779_get_core_count(void)
 {
 	void __iomem *scu_base = scu_base_addr();
 
-	shmobile_twd_init(&twd_local_timer);
 	return scu_get_core_count(scu_base);
 }
 
diff --git a/arch/arm/mach-shmobile/smp-sh73a0.c b/arch/arm/mach-shmobile/smp-sh73a0.c
index 14ad8b052f1a..e36c41c4ab40 100644
--- a/arch/arm/mach-shmobile/smp-sh73a0.c
+++ b/arch/arm/mach-shmobile/smp-sh73a0.c
@@ -42,7 +42,13 @@ static void __iomem *scu_base_addr(void)
 static DEFINE_SPINLOCK(scu_lock);
 static unsigned long tmp;
 
+#ifdef CONFIG_HAVE_ARM_TWD
 static DEFINE_TWD_LOCAL_TIMER(twd_local_timer, 0xf0000600, 29);
+void __init sh73a0_register_twd(void)
+{
+	twd_local_timer_register(&twd_local_timer);
+}
+#endif
 
 static void modify_scu_cpu_psr(unsigned long set, unsigned long clr)
 {
@@ -62,7 +68,6 @@ unsigned int __init sh73a0_get_core_count(void)
 {
 	void __iomem *scu_base = scu_base_addr();
 
-	shmobile_twd_init(&twd_local_timer);
 	return scu_get_core_count(scu_base);
 }
 
diff --git a/arch/arm/mach-shmobile/timer.c b/arch/arm/mach-shmobile/timer.c
index 2fba5f3d1c8a..a68919727e24 100644
--- a/arch/arm/mach-shmobile/timer.c
+++ b/arch/arm/mach-shmobile/timer.c
@@ -19,9 +19,27 @@
  *
  */
 #include <linux/platform_device.h>
+#include <linux/delay.h>
 #include <asm/mach/time.h>
 #include <asm/smp_twd.h>
 
+void __init shmobile_setup_delay(unsigned int max_cpu_core_mhz,
+				 unsigned int mult, unsigned int div)
+{
+	/* calculate a worst-case loops-per-jiffy value
+	 * based on maximum cpu core mhz setting and the
+	 * __delay() implementation in arch/arm/lib/delay.S
+	 *
+	 * this will result in a longer delay than expected
+	 * when the cpu core runs on lower frequencies.
+	 */
+
+	unsigned int value = (1000000 * mult) / (HZ * div);
+
+	if (!preset_lpj)
+		preset_lpj = max_cpu_core_mhz * value;
+}
+
 static void __init shmobile_late_time_init(void)
 {
 	/*
@@ -46,15 +64,6 @@ static void __init shmobile_timer_init(void)
 {
 }
 
-void __init shmobile_twd_init(struct twd_local_timer *twd_local_timer)
-{
-#ifdef CONFIG_HAVE_ARM_TWD
-	int err = twd_local_timer_register(twd_local_timer);
-	if (err)
-		pr_err("twd_local_timer_register failed %d\n", err);
-#endif
-}
-
 struct sys_timer shmobile_timer = {
 	.init		= shmobile_timer_init,
 };
diff --git a/arch/arm/mm/abort-ev6.S b/arch/arm/mm/abort-ev6.S
index ff1f7cc11f87..80741992a9fc 100644
--- a/arch/arm/mm/abort-ev6.S
+++ b/arch/arm/mm/abort-ev6.S
@@ -26,18 +26,23 @@ ENTRY(v6_early_abort)
 	mrc	p15, 0, r1, c5, c0, 0		@ get FSR
 	mrc	p15, 0, r0, c6, c0, 0		@ get FAR
 /*
- * Faulty SWP instruction on 1136 doesn't set bit 11 in DFSR (erratum 326103).
- * The test below covers all the write situations, including Java bytecodes
+ * Faulty SWP instruction on 1136 doesn't set bit 11 in DFSR.
  */
-	bic	r1, r1, #1 << 11		@ clear bit 11 of FSR
+#ifdef CONFIG_ARM_ERRATA_326103
+	ldr	ip, =0x4107b36
+	mrc	p15, 0, r3, c0, c0, 0		@ get processor id
+	teq	ip, r3, lsr #4			@ r0 ARM1136?
+	bne	do_DataAbort
 	tst	r5, #PSR_J_BIT			@ Java?
+	tsteq	r5, #PSR_T_BIT			@ Thumb?
 	bne	do_DataAbort
-	do_thumb_abort fsr=r1, pc=r4, psr=r5, tmp=r3
-	ldreq	r3, [r4]			@ read aborted ARM instruction
+	bic	r1, r1, #1 << 11		@ clear bit 11 of FSR
+	ldr	r3, [r4]			@ read aborted ARM instruction
 #ifdef CONFIG_CPU_ENDIAN_BE8
-	reveq	r3, r3
+	rev	r3, r3
 #endif
 	do_ldrd_abort tmp=ip, insn=r3
 	tst	r3, #1 << 20			@ L = 0 -> write
 	orreq	r1, r1, #1 << 11		@ yes.
+#endif
 	b	do_DataAbort
diff --git a/arch/arm/mm/cache-l2x0.c b/arch/arm/mm/cache-l2x0.c
index a53fd2aaa2f4..2a8e380501e8 100644
--- a/arch/arm/mm/cache-l2x0.c
+++ b/arch/arm/mm/cache-l2x0.c
@@ -32,6 +32,7 @@ static void __iomem *l2x0_base;
 static DEFINE_RAW_SPINLOCK(l2x0_lock);
 static u32 l2x0_way_mask;	/* Bitmask of active ways */
 static u32 l2x0_size;
+static unsigned long sync_reg_offset = L2X0_CACHE_SYNC;
 
 struct l2x0_regs l2x0_saved_regs;
 
@@ -61,12 +62,7 @@ static inline void cache_sync(void)
 {
 	void __iomem *base = l2x0_base;
 
-#ifdef CONFIG_PL310_ERRATA_753970
-	/* write to an unmmapped register */
-	writel_relaxed(0, base + L2X0_DUMMY_REG);
-#else
-	writel_relaxed(0, base + L2X0_CACHE_SYNC);
-#endif
+	writel_relaxed(0, base + sync_reg_offset);
 	cache_wait(base + L2X0_CACHE_SYNC, 1);
 }
 
@@ -85,10 +81,13 @@ static inline void l2x0_inv_line(unsigned long addr)
 }
 
 #if defined(CONFIG_PL310_ERRATA_588369) || defined(CONFIG_PL310_ERRATA_727915)
+static inline void debug_writel(unsigned long val)
+{
+	if (outer_cache.set_debug)
+		outer_cache.set_debug(val);
+}
 
-#define debug_writel(val)	outer_cache.set_debug(val)
-
-static void l2x0_set_debug(unsigned long val)
+static void pl310_set_debug(unsigned long val)
 {
 	writel_relaxed(val, l2x0_base + L2X0_DEBUG_CTRL);
 }
@@ -98,7 +97,7 @@ static inline void debug_writel(unsigned long val)
 {
 }
 
-#define l2x0_set_debug	NULL
+#define pl310_set_debug	NULL
 #endif
 
 #ifdef CONFIG_PL310_ERRATA_588369
@@ -331,6 +330,11 @@ void __init l2x0_init(void __iomem *base, u32 aux_val, u32 aux_mask)
 		else
 			ways = 8;
 		type = "L310";
+#ifdef CONFIG_PL310_ERRATA_753970
+		/* Unmapped register. */
+		sync_reg_offset = L2X0_DUMMY_REG;
+#endif
+		outer_cache.set_debug = pl310_set_debug;
 		break;
 	case L2X0_CACHE_ID_PART_L210:
 		ways = (aux >> 13) & 0xf;
@@ -379,7 +383,6 @@ void __init l2x0_init(void __iomem *base, u32 aux_val, u32 aux_mask)
 	outer_cache.flush_all = l2x0_flush_all;
 	outer_cache.inv_all = l2x0_inv_all;
 	outer_cache.disable = l2x0_disable;
-	outer_cache.set_debug = l2x0_set_debug;
 
 	printk(KERN_INFO "%s cache controller enabled\n", type);
 	printk(KERN_INFO "l2x0: %d ways, CACHE_ID 0x%08x, AUX_CTRL 0x%08x, Cache size: %d B\n",
diff --git a/arch/arm/mm/init.c b/arch/arm/mm/init.c
index 595079fa9d1d..8f5813bbffb5 100644
--- a/arch/arm/mm/init.c
+++ b/arch/arm/mm/init.c
@@ -293,11 +293,11 @@ EXPORT_SYMBOL(pfn_valid);
 #endif
 
 #ifndef CONFIG_SPARSEMEM
-static void arm_memory_present(void)
+static void __init arm_memory_present(void)
 {
 }
 #else
-static void arm_memory_present(void)
+static void __init arm_memory_present(void)
 {
 	struct memblock_region *reg;
 
diff --git a/arch/arm/mm/mmu.c b/arch/arm/mm/mmu.c
index b86f8933ff91..2c7cf2f9c837 100644
--- a/arch/arm/mm/mmu.c
+++ b/arch/arm/mm/mmu.c
@@ -618,8 +618,8 @@ static void __init alloc_init_section(pud_t *pud, unsigned long addr,
 	}
 }
 
-static void alloc_init_pud(pgd_t *pgd, unsigned long addr, unsigned long end,
-	unsigned long phys, const struct mem_type *type)
+static void __init alloc_init_pud(pgd_t *pgd, unsigned long addr,
+	unsigned long end, unsigned long phys, const struct mem_type *type)
 {
 	pud_t *pud = pud_offset(pgd, addr);
 	unsigned long next;
diff --git a/arch/arm/plat-omap/dma.c b/arch/arm/plat-omap/dma.c
index ecdb3da0dea9..c58d896cd5c3 100644
--- a/arch/arm/plat-omap/dma.c
+++ b/arch/arm/plat-omap/dma.c
@@ -916,6 +916,13 @@ void omap_start_dma(int lch)
 			l |= OMAP_DMA_CCR_BUFFERING_DISABLE;
 	l |= OMAP_DMA_CCR_EN;
 
+	/*
+	 * As dma_write() uses IO accessors which are weakly ordered, there
+	 * is no guarantee that data in coherent DMA memory will be visible
+	 * to the DMA device.  Add a memory barrier here to ensure that any
+	 * such data is visible prior to enabling DMA.
+	 */
+	mb();
 	p->dma_write(l, CCR, lch);
 
 	dma_chan[lch].flags |= OMAP_DMA_ACTIVE;
@@ -965,6 +972,13 @@ void omap_stop_dma(int lch)
 		p->dma_write(l, CCR, lch);
 	}
 
+	/*
+	 * Ensure that data transferred by DMA is visible to any access
+	 * after DMA has been disabled.  This is important for coherent
+	 * DMA regions.
+	 */
+	mb();
+
 	if (!omap_dma_in_1510_mode() && dma_chan[lch].next_lch != -1) {
 		int next_lch, cur_lch = lch;
 		char dma_chan_link_map[dma_lch_count];
diff --git a/arch/arm/vfp/vfpmodule.c b/arch/arm/vfp/vfpmodule.c
index 858748eaa144..bc683b8219b5 100644
--- a/arch/arm/vfp/vfpmodule.c
+++ b/arch/arm/vfp/vfpmodule.c
@@ -17,6 +17,8 @@
 #include <linux/sched.h>
 #include <linux/smp.h>
 #include <linux/init.h>
+#include <linux/uaccess.h>
+#include <linux/user.h>
 
 #include <asm/cp15.h>
 #include <asm/cputype.h>
@@ -529,6 +531,103 @@ void vfp_flush_hwstate(struct thread_info *thread)
 }
 
 /*
+ * Save the current VFP state into the provided structures and prepare
+ * for entry into a new function (signal handler).
+ */
+int vfp_preserve_user_clear_hwstate(struct user_vfp __user *ufp,
+				    struct user_vfp_exc __user *ufp_exc)
+{
+	struct thread_info *thread = current_thread_info();
+	struct vfp_hard_struct *hwstate = &thread->vfpstate.hard;
+	int err = 0;
+
+	/* Ensure that the saved hwstate is up-to-date. */
+	vfp_sync_hwstate(thread);
+
+	/*
+	 * Copy the floating point registers. There can be unused
+	 * registers see asm/hwcap.h for details.
+	 */
+	err |= __copy_to_user(&ufp->fpregs, &hwstate->fpregs,
+			      sizeof(hwstate->fpregs));
+	/*
+	 * Copy the status and control register.
+	 */
+	__put_user_error(hwstate->fpscr, &ufp->fpscr, err);
+
+	/*
+	 * Copy the exception registers.
+	 */
+	__put_user_error(hwstate->fpexc, &ufp_exc->fpexc, err);
+	__put_user_error(hwstate->fpinst, &ufp_exc->fpinst, err);
+	__put_user_error(hwstate->fpinst2, &ufp_exc->fpinst2, err);
+
+	if (err)
+		return -EFAULT;
+
+	/* Ensure that VFP is disabled. */
+	vfp_flush_hwstate(thread);
+
+	/*
+	 * As per the PCS, clear the length and stride bits for function
+	 * entry.
+	 */
+	hwstate->fpscr &= ~(FPSCR_LENGTH_MASK | FPSCR_STRIDE_MASK);
+
+	/*
+	 * Disable VFP in the hwstate so that we can detect if it gets
+	 * used.
+	 */
+	hwstate->fpexc &= ~FPEXC_EN;
+	return 0;
+}
+
+/* Sanitise and restore the current VFP state from the provided structures. */
+int vfp_restore_user_hwstate(struct user_vfp __user *ufp,
+			     struct user_vfp_exc __user *ufp_exc)
+{
+	struct thread_info *thread = current_thread_info();
+	struct vfp_hard_struct *hwstate = &thread->vfpstate.hard;
+	unsigned long fpexc;
+	int err = 0;
+
+	/*
+	 * If VFP has been used, then disable it to avoid corrupting
+	 * the new thread state.
+	 */
+	if (hwstate->fpexc & FPEXC_EN)
+		vfp_flush_hwstate(thread);
+
+	/*
+	 * Copy the floating point registers. There can be unused
+	 * registers see asm/hwcap.h for details.
+	 */
+	err |= __copy_from_user(&hwstate->fpregs, &ufp->fpregs,
+				sizeof(hwstate->fpregs));
+	/*
+	 * Copy the status and control register.
+	 */
+	__get_user_error(hwstate->fpscr, &ufp->fpscr, err);
+
+	/*
+	 * Sanitise and restore the exception registers.
+	 */
+	__get_user_error(fpexc, &ufp_exc->fpexc, err);
+
+	/* Ensure the VFP is enabled. */
+	fpexc |= FPEXC_EN;
+
+	/* Ensure FPINST2 is invalid and the exception flag is cleared. */
+	fpexc &= ~(FPEXC_EX | FPEXC_FP2V);
+	hwstate->fpexc = fpexc;
+
+	__get_user_error(hwstate->fpinst, &ufp_exc->fpinst, err);
+	__get_user_error(hwstate->fpinst2, &ufp_exc->fpinst2, err);
+
+	return err ? -EFAULT : 0;
+}
+
+/*
  * VFP hardware can lose all context when a CPU goes offline.
  * As we will be running in SMP mode with CPU hotplug, we will save the
  * hardware state at every thread switch.  We clear our held state when