Merge branch 'heads/soc2' into boards2-base

34 files changed, 1260 insertions, 146 deletions

diff --git a/arch/arm/boot/dts/r7s72100.dtsi b/arch/arm/boot/dts/r7s72100.dtsi
new file mode 100644
index 00000000000..46b82aa7dc4
--- /dev/null
+++ b/arch/arm/boot/dts/r7s72100.dtsi
@@ -0,0 +1,36 @@
+/*
+ * Device Tree Source for the r7s72100 SoC
+ *
+ * Copyright (C) 2013 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.
+ */
+
+/ {
+	compatible = "renesas,r7s72100";
+	interrupt-parent = <&gic>;
+	#address-cells = <1>;
+	#size-cells = <1>;
+
+	cpus {
+		#address-cells = <1>;
+		#size-cells = <0>;
+
+		cpu@0 {
+			device_type = "cpu";
+			compatible = "arm,cortex-a9";
+			reg = <0>;
+		};
+	};
+
+	gic: interrupt-controller@e8201000 {
+		compatible = "arm,cortex-a9-gic";
+		#interrupt-cells = <3>;
+		#address-cells = <0>;
+		interrupt-controller;
+		reg = <0xe8201000 0x1000>,
+			<0xe8202000 0x1000>;
+	};
+};
diff --git a/arch/arm/mach-shmobile/Kconfig b/arch/arm/mach-shmobile/Kconfig
index eda28579496..6e7d0a9b0be 100644
--- a/arch/arm/mach-shmobile/Kconfig
+++ b/arch/arm/mach-shmobile/Kconfig
@@ -113,6 +113,12 @@ config ARCH_EMEV2
 	select ARM_GIC
 	select CPU_V7
 
+config ARCH_R7S72100
+	bool "RZ/A1H (R7S72100)"
+	select ARM_GIC
+	select CPU_V7
+	select SH_CLK_CPG
+
 comment "SH-Mobile Board Type"
 
 config MACH_APE6EVM
diff --git a/arch/arm/mach-shmobile/Makefile b/arch/arm/mach-shmobile/Makefile
index e552e84b1fa..663f8941153 100644
--- a/arch/arm/mach-shmobile/Makefile
+++ b/arch/arm/mach-shmobile/Makefile
@@ -15,8 +15,10 @@ obj-$(CONFIG_ARCH_R8A7740)	+= setup-r8a7740.o
 obj-$(CONFIG_ARCH_R8A7778)	+= setup-r8a7778.o
 obj-$(CONFIG_ARCH_R8A7779)	+= setup-r8a7779.o
 obj-$(CONFIG_ARCH_R8A7790)	+= setup-r8a7790.o
-obj-$(CONFIG_ARCH_R8A7791)	+= setup-r8a7791.o
+obj-$(CONFIG_ARCH_R8A7790)	+= setup-r8a7790.o setup-rcar-gen2.o
+obj-$(CONFIG_ARCH_R8A7791)	+= setup-r8a7791.o setup-rcar-gen2.o
 obj-$(CONFIG_ARCH_EMEV2)	+= setup-emev2.o
+obj-$(CONFIG_ARCH_R7S72100)	+= setup-r7s72100.o
 
 # Clock objects
 ifndef CONFIG_COMMON_CLK
@@ -30,12 +32,15 @@ obj-$(CONFIG_ARCH_R8A7779)	+= clock-r8a7779.o
 obj-$(CONFIG_ARCH_R8A7790)	+= clock-r8a7790.o
 obj-$(CONFIG_ARCH_R8A7791)	+= clock-r8a7791.o
 obj-$(CONFIG_ARCH_EMEV2)	+= clock-emev2.o
+obj-$(CONFIG_ARCH_R7S72100)	+= clock-r7s72100.o
 endif
 
 # SMP objects
 smp-y				:= platsmp.o headsmp.o
 smp-$(CONFIG_ARCH_SH73A0)	+= smp-sh73a0.o headsmp-scu.o platsmp-scu.o
 smp-$(CONFIG_ARCH_R8A7779)	+= smp-r8a7779.o headsmp-scu.o platsmp-scu.o
+smp-$(CONFIG_ARCH_R8A7790)	+= smp-r8a7790.o platsmp-apmu.o
+smp-$(CONFIG_ARCH_R8A7791)	+= smp-r8a7791.o platsmp-apmu.o
 smp-$(CONFIG_ARCH_EMEV2)	+= smp-emev2.o headsmp-scu.o platsmp-scu.o
 
 # IRQ objects
diff --git a/arch/arm/mach-shmobile/board-ape6evm-reference.c b/arch/arm/mach-shmobile/board-ape6evm-reference.c
index a23fa714f7a..3276afcf3cc 100644
--- a/arch/arm/mach-shmobile/board-ape6evm-reference.c
+++ b/arch/arm/mach-shmobile/board-ape6evm-reference.c
@@ -57,7 +57,7 @@ static const char *ape6evm_boards_compat_dt[] __initdata = {
 };
 
 DT_MACHINE_START(APE6EVM_DT, "ape6evm")
-	.init_early	= r8a73a4_init_delay,
+	.init_early	= r8a73a4_init_early,
 	.init_machine	= ape6evm_add_standard_devices,
 	.dt_compat	= ape6evm_boards_compat_dt,
 MACHINE_END
diff --git a/arch/arm/mach-shmobile/board-ape6evm.c b/arch/arm/mach-shmobile/board-ape6evm.c
index 0365d2e2e73..7627385f516 100644
--- a/arch/arm/mach-shmobile/board-ape6evm.c
+++ b/arch/arm/mach-shmobile/board-ape6evm.c
@@ -279,7 +279,7 @@ static const char *ape6evm_boards_compat_dt[] __initdata = {
 };
 
 DT_MACHINE_START(APE6EVM_DT, "ape6evm")
-	.init_early	= r8a73a4_init_delay,
+	.init_early	= r8a73a4_init_early,
 	.init_machine	= ape6evm_add_standard_devices,
 	.dt_compat	= ape6evm_boards_compat_dt,
 MACHINE_END
diff --git a/arch/arm/mach-shmobile/board-lager-reference.c b/arch/arm/mach-shmobile/board-lager-reference.c
index 9c316a1b2e3..1a1a4a88863 100644
--- a/arch/arm/mach-shmobile/board-lager-reference.c
+++ b/arch/arm/mach-shmobile/board-lager-reference.c
@@ -38,8 +38,9 @@ static const char *lager_boards_compat_dt[] __initdata = {
 };
 
 DT_MACHINE_START(LAGER_DT, "lager")
-	.init_early	= r8a7790_init_delay,
+	.smp		= smp_ops(r8a7790_smp_ops),
+	.init_early	= r8a7790_init_early,
+	.init_time	= rcar_gen2_timer_init,
 	.init_machine	= lager_add_standard_devices,
-	.init_time	= r8a7790_timer_init,
 	.dt_compat	= lager_boards_compat_dt,
 MACHINE_END
diff --git a/arch/arm/mach-shmobile/board-lager.c b/arch/arm/mach-shmobile/board-lager.c
index 1e231423b2b..59951340b55 100644
--- a/arch/arm/mach-shmobile/board-lager.c
+++ b/arch/arm/mach-shmobile/board-lager.c
@@ -229,8 +229,9 @@ static const char *lager_boards_compat_dt[] __initdata = {
 };
 
 DT_MACHINE_START(LAGER_DT, "lager")
-	.init_early	= r8a7790_init_delay,
-	.init_time	= r8a7790_timer_init,
+	.smp		= smp_ops(r8a7790_smp_ops),
+	.init_early	= r8a7790_init_early,
+	.init_time	= rcar_gen2_timer_init,
 	.init_machine	= lager_add_standard_devices,
 	.dt_compat	= lager_boards_compat_dt,
 MACHINE_END
diff --git a/arch/arm/mach-shmobile/clock-r7s72100.c b/arch/arm/mach-shmobile/clock-r7s72100.c
new file mode 100644
index 00000000000..4aba20ca127
--- /dev/null
+++ b/arch/arm/mach-shmobile/clock-r7s72100.c
@@ -0,0 +1,202 @@
+/*
+ * r7a72100 clock framework support
+ *
+ * Copyright (C) 2013  Renesas Solutions Corp.
+ * Copyright (C) 2012  Phil Edworthy
+ * Copyright (C) 2011  Magnus Damm
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; version 2 of the License.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+#include <linux/init.h>
+#include <linux/kernel.h>
+#include <linux/io.h>
+#include <linux/sh_clk.h>
+#include <linux/clkdev.h>
+#include <mach/common.h>
+#include <mach/r7s72100.h>
+
+/* registers */
+#define FRQCR		0xfcfe0010
+#define FRQCR2		0xfcfe0014
+#define STBCR3		0xfcfe0420
+#define STBCR4		0xfcfe0424
+
+#define PLL_RATE 30
+
+static struct clk_mapping cpg_mapping = {
+	.phys	= 0xfcfe0000,
+	.len	= 0x1000,
+};
+
+/* Fixed 32 KHz root clock for RTC */
+static struct clk r_clk = {
+	.rate           = 32768,
+};
+
+/*
+ * Default rate for the root input clock, reset this with clk_set_rate()
+ * from the platform code.
+ */
+static struct clk extal_clk = {
+	.rate		= 13330000,
+	.mapping	= &cpg_mapping,
+};
+
+static unsigned long pll_recalc(struct clk *clk)
+{
+	return clk->parent->rate * PLL_RATE;
+}
+
+static struct sh_clk_ops pll_clk_ops = {
+	.recalc		= pll_recalc,
+};
+
+static struct clk pll_clk = {
+	.ops		= &pll_clk_ops,
+	.parent		= &extal_clk,
+	.flags		= CLK_ENABLE_ON_INIT,
+};
+
+static unsigned long bus_recalc(struct clk *clk)
+{
+	return clk->parent->rate * 2 / 3;
+}
+
+static struct sh_clk_ops bus_clk_ops = {
+	.recalc		= bus_recalc,
+};
+
+static struct clk bus_clk = {
+	.ops		= &bus_clk_ops,
+	.parent		= &pll_clk,
+	.flags		= CLK_ENABLE_ON_INIT,
+};
+
+static unsigned long peripheral0_recalc(struct clk *clk)
+{
+	return clk->parent->rate / 12;
+}
+
+static struct sh_clk_ops peripheral0_clk_ops = {
+	.recalc		= peripheral0_recalc,
+};
+
+static struct clk peripheral0_clk = {
+	.ops		= &peripheral0_clk_ops,
+	.parent		= &pll_clk,
+	.flags		= CLK_ENABLE_ON_INIT,
+};
+
+static unsigned long peripheral1_recalc(struct clk *clk)
+{
+	return clk->parent->rate / 6;
+}
+
+static struct sh_clk_ops peripheral1_clk_ops = {
+	.recalc		= peripheral1_recalc,
+};
+
+static struct clk peripheral1_clk = {
+	.ops		= &peripheral1_clk_ops,
+	.parent		= &pll_clk,
+	.flags		= CLK_ENABLE_ON_INIT,
+};
+
+struct clk *main_clks[] = {
+	&r_clk,
+	&extal_clk,
+	&pll_clk,
+	&bus_clk,
+	&peripheral0_clk,
+	&peripheral1_clk,
+};
+
+static int div2[] = { 1, 3, 0, 3 }; /* 1, 2/3, reserve, 1/3 */
+static int multipliers[] = { 1, 2, 1, 1 };
+
+static struct clk_div_mult_table div4_div_mult_table = {
+	.divisors = div2,
+	.nr_divisors = ARRAY_SIZE(div2),
+	.multipliers = multipliers,
+	.nr_multipliers = ARRAY_SIZE(multipliers),
+};
+
+static struct clk_div4_table div4_table = {
+	.div_mult_table = &div4_div_mult_table,
+};
+
+enum { DIV4_I,
+	DIV4_NR };
+
+#define DIV4(_reg, _bit, _mask, _flags) \
+	SH_CLK_DIV4(&pll_clk, _reg, _bit, _mask, _flags)
+
+/* The mask field specifies the div2 entries that are valid */
+struct clk div4_clks[DIV4_NR] = {
+	[DIV4_I]  = DIV4(FRQCR, 8, 0xB, CLK_ENABLE_REG_16BIT
+					| CLK_ENABLE_ON_INIT),
+};
+
+enum { MSTP47, MSTP46, MSTP45, MSTP44, MSTP43, MSTP42, MSTP41, MSTP40,
+	MSTP33,	MSTP_NR };
+
+static struct clk mstp_clks[MSTP_NR] = {
+	[MSTP47] = SH_CLK_MSTP8(&peripheral1_clk, STBCR4, 7, 0), /* SCIF0 */
+	[MSTP46] = SH_CLK_MSTP8(&peripheral1_clk, STBCR4, 6, 0), /* SCIF1 */
+	[MSTP45] = SH_CLK_MSTP8(&peripheral1_clk, STBCR4, 5, 0), /* SCIF2 */
+	[MSTP44] = SH_CLK_MSTP8(&peripheral1_clk, STBCR4, 4, 0), /* SCIF3 */
+	[MSTP43] = SH_CLK_MSTP8(&peripheral1_clk, STBCR4, 3, 0), /* SCIF4 */
+	[MSTP42] = SH_CLK_MSTP8(&peripheral1_clk, STBCR4, 2, 0), /* SCIF5 */
+	[MSTP41] = SH_CLK_MSTP8(&peripheral1_clk, STBCR4, 1, 0), /* SCIF6 */
+	[MSTP40] = SH_CLK_MSTP8(&peripheral1_clk, STBCR4, 0, 0), /* SCIF7 */
+	[MSTP33] = SH_CLK_MSTP8(&peripheral0_clk, STBCR3, 3, 0), /* MTU2 */
+};
+
+static struct clk_lookup lookups[] = {
+	/* main clocks */
+	CLKDEV_CON_ID("rclk", &r_clk),
+	CLKDEV_CON_ID("extal", &extal_clk),
+	CLKDEV_CON_ID("pll_clk", &pll_clk),
+	CLKDEV_CON_ID("peripheral_clk", &peripheral1_clk),
+
+	/* DIV4 clocks */
+	CLKDEV_CON_ID("cpu_clk", &div4_clks[DIV4_I]),
+
+	/* MSTP clocks */
+	CLKDEV_ICK_ID("sci_fck", "sh-sci.0", &mstp_clks[MSTP47]),
+	CLKDEV_ICK_ID("sci_fck", "sh-sci.1", &mstp_clks[MSTP46]),
+	CLKDEV_ICK_ID("sci_fck", "sh-sci.2", &mstp_clks[MSTP45]),
+	CLKDEV_ICK_ID("sci_fck", "sh-sci.3", &mstp_clks[MSTP44]),
+	CLKDEV_ICK_ID("sci_fck", "sh-sci.4", &mstp_clks[MSTP43]),
+	CLKDEV_ICK_ID("sci_fck", "sh-sci.5", &mstp_clks[MSTP42]),
+	CLKDEV_ICK_ID("sci_fck", "sh-sci.6", &mstp_clks[MSTP41]),
+	CLKDEV_ICK_ID("sci_fck", "sh-sci.7", &mstp_clks[MSTP40]),
+};
+
+void __init r7s72100_clock_init(void)
+{
+	int k, ret = 0;
+
+	for (k = 0; !ret && (k < ARRAY_SIZE(main_clks)); k++)
+		ret = clk_register(main_clks[k]);
+
+	clkdev_add_table(lookups, ARRAY_SIZE(lookups));
+
+	if (!ret)
+		ret = sh_clk_div4_register(div4_clks, DIV4_NR, &div4_table);
+
+	if (!ret)
+		ret = sh_clk_mstp_register(mstp_clks, MSTP_NR);
+
+	if (!ret)
+		shmobile_clk_init();
+	else
+		panic("failed to setup rza1 clocks\n");
+}
diff --git a/arch/arm/mach-shmobile/clock-r8a73a4.c b/arch/arm/mach-shmobile/clock-r8a73a4.c
index 8ea5ef6c79c..74841edf04e 100644
--- a/arch/arm/mach-shmobile/clock-r8a73a4.c
+++ b/arch/arm/mach-shmobile/clock-r8a73a4.c
@@ -504,7 +504,7 @@ static struct clk div6_clks[DIV6_NR] = {
 
 /* MSTP */
 enum {
-	MSTP217, MSTP216, MSTP207, MSTP206, MSTP204, MSTP203,
+	MSTP218, MSTP217, MSTP216, MSTP207, MSTP206, MSTP204, MSTP203,
 	MSTP329, MSTP323, MSTP318, MSTP317, MSTP316,
 	MSTP315, MSTP314, MSTP313, MSTP312, MSTP305, MSTP300,
 	MSTP411, MSTP410, MSTP409,
@@ -519,6 +519,7 @@ static struct clk mstp_clks[MSTP_NR] = {
 	[MSTP207] = SH_CLK_MSTP32(&div6_clks[DIV6_MP],	SMSTPCR2, 7, 0), /* SCIFB1 */
 	[MSTP216] = SH_CLK_MSTP32(&div6_clks[DIV6_MP],	SMSTPCR2, 16, 0), /* SCIFB2 */
 	[MSTP217] = SH_CLK_MSTP32(&div6_clks[DIV6_MP],	SMSTPCR2, 17, 0), /* SCIFB3 */
+	[MSTP218] = SH_CLK_MSTP32(&div4_clks[DIV4_HP],	SMSTPCR2, 18, 0), /* DMAC */
 	[MSTP300] = SH_CLK_MSTP32(&div4_clks[DIV4_HP],	SMSTPCR3, 0, 0), /* IIC2 */
 	[MSTP305] = SH_CLK_MSTP32(&div6_clks[DIV6_MMC1],SMSTPCR3, 5, 0), /* MMCIF1 */
 	[MSTP312] = SH_CLK_MSTP32(&div6_clks[DIV6_SDHI2],SMSTPCR3, 12, 0), /* SDHI2 */
@@ -578,6 +579,8 @@ static struct clk_lookup lookups[] = {
 	CLKDEV_DEV_ID("sh-sci.3", &mstp_clks[MSTP207]),
 	CLKDEV_DEV_ID("sh-sci.4", &mstp_clks[MSTP216]),
 	CLKDEV_DEV_ID("sh-sci.5", &mstp_clks[MSTP217]),
+	CLKDEV_DEV_ID("sh-dma-engine.0", &mstp_clks[MSTP218]),
+	CLKDEV_DEV_ID("e6700020.dma-controller", &mstp_clks[MSTP218]),
 	CLKDEV_DEV_ID("rcar_thermal", &mstp_clks[MSTP522]),
 	CLKDEV_DEV_ID("e6520000.i2c", &mstp_clks[MSTP300]),
 	CLKDEV_DEV_ID("sh_mmcif.1", &mstp_clks[MSTP305]),
diff --git a/arch/arm/mach-shmobile/clock-r8a7790.c b/arch/arm/mach-shmobile/clock-r8a7790.c
index d99b87bc76e..a64f965c7da 100644
--- a/arch/arm/mach-shmobile/clock-r8a7790.c
+++ b/arch/arm/mach-shmobile/clock-r8a7790.c
@@ -52,6 +52,7 @@
 #define SMSTPCR5 0xe6150144
 #define SMSTPCR7 0xe615014c
 #define SMSTPCR8 0xe6150990
+#define SMSTPCR9 0xe6150994
 
 #define SDCKCR		0xE6150074
 #define SD2CKCR		0xE6150078
@@ -181,6 +182,7 @@ static struct clk div6_clks[DIV6_NR] = {
 
 /* MSTP */
 enum {
+	MSTP931, MSTP930, MSTP929, MSTP928,
 	MSTP813,
 	MSTP726, MSTP725, MSTP724, MSTP723, MSTP722, MSTP721, MSTP720,
 	MSTP717, MSTP716,
@@ -192,6 +194,10 @@ enum {
 };
 
 static struct clk mstp_clks[MSTP_NR] = {
+	[MSTP931] = SH_CLK_MSTP32(&hp_clk, SMSTPCR9, 31, 0), /* I2C0 */
+	[MSTP930] = SH_CLK_MSTP32(&hp_clk, SMSTPCR9, 30, 0), /* I2C1 */
+	[MSTP929] = SH_CLK_MSTP32(&hp_clk, SMSTPCR9, 29, 0), /* I2C2 */
+	[MSTP928] = SH_CLK_MSTP32(&hp_clk, SMSTPCR9, 28, 0), /* I2C3 */
 	[MSTP813] = SH_CLK_MSTP32(&p_clk, SMSTPCR8, 13, 0), /* Ether */
 	[MSTP726] = SH_CLK_MSTP32(&zx_clk, SMSTPCR7, 26, 0), /* LVDS0 */
 	[MSTP725] = SH_CLK_MSTP32(&zx_clk, SMSTPCR7, 25, 0), /* LVDS1 */
@@ -271,6 +277,10 @@ static struct clk_lookup lookups[] = {
 	CLKDEV_DEV_ID("sh-sci.7", &mstp_clks[MSTP720]),
 	CLKDEV_DEV_ID("sh-sci.8", &mstp_clks[MSTP717]),
 	CLKDEV_DEV_ID("sh-sci.9", &mstp_clks[MSTP716]),
+	CLKDEV_DEV_ID("e6508000.i2c", &mstp_clks[MSTP931]),
+	CLKDEV_DEV_ID("e6518000.i2c", &mstp_clks[MSTP930]),
+	CLKDEV_DEV_ID("e6530000.i2c", &mstp_clks[MSTP929]),
+	CLKDEV_DEV_ID("e6540000.i2c", &mstp_clks[MSTP928]),
 	CLKDEV_DEV_ID("r8a7790-ether", &mstp_clks[MSTP813]),
 	CLKDEV_DEV_ID("rcar_thermal", &mstp_clks[MSTP522]),
 	CLKDEV_DEV_ID("ee200000.mmcif", &mstp_clks[MSTP315]),
@@ -300,7 +310,7 @@ static struct clk_lookup lookups[] = {
 
 void __init r8a7790_clock_init(void)
 {
-	u32 mode = r8a7790_read_mode_pins();
+	u32 mode = rcar_gen2_read_mode_pins();
 	int k, ret = 0;
 
 	switch (mode & (MD(14) | MD(13))) {
diff --git a/arch/arm/mach-shmobile/headsmp.S b/arch/arm/mach-shmobile/headsmp.S
index f93751caf5c..e5be5c88644 100644
--- a/arch/arm/mach-shmobile/headsmp.S
+++ b/arch/arm/mach-shmobile/headsmp.S
@@ -40,6 +40,9 @@ shmobile_boot_fn:
 	.globl	shmobile_boot_arg
 shmobile_boot_arg:
 2:	.space	4
+	.globl	shmobile_boot_size
+shmobile_boot_size:
+	.long	. - shmobile_boot_vector
 
 /*
  * Per-CPU SMP boot function/argument selection code based on MPIDR
diff --git a/arch/arm/mach-shmobile/include/mach/common.h b/arch/arm/mach-shmobile/include/mach/common.h
index 7b938681e75..e31980590eb 100644
--- a/arch/arm/mach-shmobile/include/mach/common.h
+++ b/arch/arm/mach-shmobile/include/mach/common.h
@@ -9,16 +9,23 @@ extern void shmobile_setup_console(void);
 extern void shmobile_boot_vector(void);
 extern unsigned long shmobile_boot_fn;
 extern unsigned long shmobile_boot_arg;
+extern unsigned long shmobile_boot_size;
 extern void shmobile_smp_boot(void);
 extern void shmobile_smp_sleep(void);
 extern void shmobile_smp_hook(unsigned int cpu, unsigned long fn,
 			      unsigned long arg);
+extern int shmobile_smp_cpu_disable(unsigned int cpu);
+extern void shmobile_invalidate_start(void);
 extern void shmobile_boot_scu(void);
 extern void shmobile_smp_scu_prepare_cpus(unsigned int max_cpus);
-extern int shmobile_smp_scu_boot_secondary(unsigned int cpu,
-					   struct task_struct *idle);
 extern void shmobile_smp_scu_cpu_die(unsigned int cpu);
 extern int shmobile_smp_scu_cpu_kill(unsigned int cpu);
+extern void shmobile_smp_apmu_prepare_cpus(unsigned int max_cpus);
+extern int shmobile_smp_apmu_boot_secondary(unsigned int cpu,
+					    struct task_struct *idle);
+extern void shmobile_smp_apmu_cpu_die(unsigned int cpu);
+extern int shmobile_smp_apmu_cpu_kill(unsigned int cpu);
+extern void shmobile_invalidate_start(void);
 struct clk;
 extern int shmobile_clk_init(void);
 extern void shmobile_handle_irq_intc(struct pt_regs *);
@@ -39,7 +46,6 @@ static inline int shmobile_cpuidle_init(void) { return 0; }
 #endif
 
 extern void __iomem *shmobile_scu_base;
-extern void shmobile_smp_init_cpus(unsigned int ncores);
 
 static inline void __init shmobile_init_late(void)
 {
diff --git a/arch/arm/mach-shmobile/include/mach/r7s72100.h b/arch/arm/mach-shmobile/include/mach/r7s72100.h
new file mode 100644
index 00000000000..5f34b20ecd4
--- /dev/null
+++ b/arch/arm/mach-shmobile/include/mach/r7s72100.h
@@ -0,0 +1,8 @@
+#ifndef __ASM_R7S72100_H__
+#define __ASM_R7S72100_H__
+
+void r7s72100_add_dt_devices(void);
+void r7s72100_clock_init(void);
+void r7s72100_init_early(void);
+
+#endif /* __ASM_R7S72100_H__ */
diff --git a/arch/arm/mach-shmobile/include/mach/r8a73a4.h b/arch/arm/mach-shmobile/include/mach/r8a73a4.h
index f3a9b702da5..ce8bdd1d8a8 100644
--- a/arch/arm/mach-shmobile/include/mach/r8a73a4.h
+++ b/arch/arm/mach-shmobile/include/mach/r8a73a4.h
@@ -1,10 +1,19 @@
 #ifndef __ASM_R8A73A4_H__
 #define __ASM_R8A73A4_H__
 
+/* DMA slave IDs */
+enum {
+	SHDMA_SLAVE_INVALID,
+	SHDMA_SLAVE_MMCIF0_TX,
+	SHDMA_SLAVE_MMCIF0_RX,
+	SHDMA_SLAVE_MMCIF1_TX,
+	SHDMA_SLAVE_MMCIF1_RX,
+};
+
 void r8a73a4_add_standard_devices(void);
 void r8a73a4_add_dt_devices(void);
 void r8a73a4_clock_init(void);
 void r8a73a4_pinmux_init(void);
-void r8a73a4_init_delay(void);
+void r8a73a4_init_early(void);
 
 #endif /* __ASM_R8A73A4_H__ */
diff --git a/arch/arm/mach-shmobile/include/mach/r8a7778.h b/arch/arm/mach-shmobile/include/mach/r8a7778.h
index ea1dca6880f..dbe221a484d 100644
--- a/arch/arm/mach-shmobile/include/mach/r8a7778.h
+++ b/arch/arm/mach-shmobile/include/mach/r8a7778.h
@@ -1,6 +1,7 @@
 /*
  * Copyright (C) 2013  Renesas Solutions Corp.
  * Copyright (C) 2013  Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
+ * Copyright (C) 2013  Cogent Embedded, Inc.
  *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
@@ -21,6 +22,13 @@
 #include <linux/sh_eth.h>
 #include <linux/platform_data/camera-rcar.h>
 
+/* HPB-DMA slave IDs */
+enum {
+	HPBDMA_SLAVE_DUMMY,
+	HPBDMA_SLAVE_SDHI0_TX,
+	HPBDMA_SLAVE_SDHI0_RX,
+};
+
 extern void r8a7778_add_standard_devices(void);
 extern void r8a7778_add_standard_devices_dt(void);
 extern void r8a7778_add_ether_device(struct sh_eth_plat_data *pdata);
@@ -33,6 +41,7 @@ extern void r8a7778_init_delay(void);
 extern void r8a7778_init_irq_dt(void);
 extern void r8a7778_clock_init(void);
 extern void r8a7778_init_irq_extpin(int irlm);
+extern void r8a7778_init_irq_extpin_dt(int irlm);
 extern void r8a7778_pinmux_init(void);
 
 extern int r8a7778_usb_phy_power(bool enable);
diff --git a/arch/arm/mach-shmobile/include/mach/r8a7779.h b/arch/arm/mach-shmobile/include/mach/r8a7779.h
index 11c740047e1..17af34ed89c 100644
--- a/arch/arm/mach-shmobile/include/mach/r8a7779.h
+++ b/arch/arm/mach-shmobile/include/mach/r8a7779.h
@@ -6,6 +6,13 @@
 #include <linux/sh_eth.h>
 #include <linux/platform_data/camera-rcar.h>
 
+/* HPB-DMA slave IDs */
+enum {
+	HPBDMA_SLAVE_DUMMY,
+	HPBDMA_SLAVE_SDHI0_TX,
+	HPBDMA_SLAVE_SDHI0_RX,
+};
+
 struct platform_device;
 
 struct r8a7779_pm_ch {
@@ -26,6 +33,7 @@ static inline struct r8a7779_pm_ch *to_r8a7779_ch(struct generic_pm_domain *d)
 
 extern void r8a7779_init_delay(void);
 extern void r8a7779_init_irq_extpin(int irlm);
+extern void r8a7779_init_irq_extpin_dt(int irlm);
 extern void r8a7779_init_irq_dt(void);
 extern void r8a7779_map_io(void);
 extern void r8a7779_earlytimer_init(void);
diff --git a/arch/arm/mach-shmobile/include/mach/r8a7790.h b/arch/arm/mach-shmobile/include/mach/r8a7790.h
index 788d5595209..5fbfa28b40b 100644
--- a/arch/arm/mach-shmobile/include/mach/r8a7790.h
+++ b/arch/arm/mach-shmobile/include/mach/r8a7790.h
@@ -1,14 +1,13 @@
 #ifndef __ASM_R8A7790_H__
 #define __ASM_R8A7790_H__
 
+#include <mach/rcar-gen2.h>
+
 void r8a7790_add_standard_devices(void);
 void r8a7790_add_dt_devices(void);
 void r8a7790_clock_init(void);
 void r8a7790_pinmux_init(void);
-void r8a7790_init_delay(void);
-void r8a7790_timer_init(void);
-
-#define MD(nr) BIT(nr)
-u32 r8a7790_read_mode_pins(void);
+void r8a7790_init_early(void);
+extern struct smp_operations r8a7790_smp_ops;
 
 #endif /* __ASM_R8A7790_H__ */
diff --git a/arch/arm/mach-shmobile/include/mach/r8a7791.h b/arch/arm/mach-shmobile/include/mach/r8a7791.h
index 2e6d6613108..051ead3c286 100644
--- a/arch/arm/mach-shmobile/include/mach/r8a7791.h
+++ b/arch/arm/mach-shmobile/include/mach/r8a7791.h
@@ -1,8 +1,10 @@
 #ifndef __ASM_R8A7791_H__
 #define __ASM_R8A7791_H__
 
+void r8a7791_add_standard_devices(void);
 void r8a7791_add_dt_devices(void);
 void r8a7791_clock_init(void);
 void r8a7791_init_early(void);
+extern struct smp_operations r8a7791_smp_ops;
 
 #endif /* __ASM_R8A7791_H__ */
diff --git a/arch/arm/mach-shmobile/include/mach/rcar-gen2.h b/arch/arm/mach-shmobile/include/mach/rcar-gen2.h
new file mode 100644
index 00000000000..43f606eb2d8
--- /dev/null
+++ b/arch/arm/mach-shmobile/include/mach/rcar-gen2.h
@@ -0,0 +1,8 @@
+#ifndef __ASM_RCAR_GEN2_H__
+#define __ASM_RCAR_GEN2_H__
+
+void rcar_gen2_timer_init(void);
+#define MD(nr) BIT(nr)
+u32 rcar_gen2_read_mode_pins(void);
+
+#endif /* __ASM_RCAR_GEN2_H__ */
diff --git a/arch/arm/mach-shmobile/platsmp-apmu.c b/arch/arm/mach-shmobile/platsmp-apmu.c
new file mode 100644
index 00000000000..1da5a72d964
--- /dev/null
+++ b/arch/arm/mach-shmobile/platsmp-apmu.c
@@ -0,0 +1,195 @@
+/*
+ * SMP support for SoCs with APMU
+ *
+ * Copyright (C) 2013  Magnus Damm
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/delay.h>
+#include <linux/init.h>
+#include <linux/io.h>
+#include <linux/ioport.h>
+#include <linux/of_address.h>
+#include <linux/smp.h>
+#include <asm/cacheflush.h>
+#include <asm/cp15.h>
+#include <asm/smp_plat.h>
+#include <mach/common.h>
+
+static struct {
+	void __iomem *iomem;
+	int bit;
+} apmu_cpus[CONFIG_NR_CPUS];
+
+#define WUPCR_OFFS 0x10
+#define PSTR_OFFS 0x40
+#define CPUNCR_OFFS(n) (0x100 + (0x10 * (n)))
+
+static int apmu_power_on(void __iomem *p, int bit)
+{
+	/* request power on */
+	writel_relaxed(BIT(bit), p + WUPCR_OFFS);
+
+	/* wait for APMU to finish */
+	while (readl_relaxed(p + WUPCR_OFFS) != 0)
+		;
+
+	return 0;
+}
+
+static int apmu_power_off(void __iomem *p, int bit)
+{
+	/* request Core Standby for next WFI */
+	writel_relaxed(3, p + CPUNCR_OFFS(bit));
+	return 0;
+}
+
+static int apmu_power_off_poll(void __iomem *p, int bit)
+{
+	int k;
+
+	for (k = 0; k < 1000; k++) {
+		if (((readl_relaxed(p + PSTR_OFFS) >> (bit * 4)) & 0x03) == 3)
+			return 1;
+
+		mdelay(1);
+	}
+
+	return 0;
+}
+
+static int apmu_wrap(int cpu, int (*fn)(void __iomem *p, int cpu))
+{
+	void __iomem *p = apmu_cpus[cpu].iomem;
+
+	return p ? fn(p, apmu_cpus[cpu].bit) : -EINVAL;
+}
+
+static void apmu_init_cpu(struct resource *res, int cpu, int bit)
+{
+	if (apmu_cpus[cpu].iomem)
+		return;
+
+	apmu_cpus[cpu].iomem = ioremap_nocache(res->start, resource_size(res));
+	apmu_cpus[cpu].bit = bit;
+
+	pr_debug("apmu ioremap %d %d 0x%08x 0x%08x\n", cpu, bit,
+		 res->start, resource_size(res));
+}
+
+static struct {
+	struct resource iomem;
+	int cpus[4];
+} apmu_config[] = {
+	{
+		.iomem = DEFINE_RES_MEM(0xe6152000, 0x88),
+		.cpus = { 0, 1, 2, 3 },
+	},
+	{
+		.iomem = DEFINE_RES_MEM(0xe6151000, 0x88),
+		.cpus = { 0x100, 0x101, 0x102, 0x103 },
+	}
+};
+
+static void apmu_parse_cfg(void (*fn)(struct resource *res, int cpu, int bit))
+{
+	u32 id;
+	int k;
+	int bit, index;
+	bool is_allowed;
+
+	for (k = 0; k < ARRAY_SIZE(apmu_config); k++) {
+		/* only enable the cluster that includes the boot CPU */
+		is_allowed = false;
+		for (bit = 0; bit < ARRAY_SIZE(apmu_config[k].cpus); bit++) {
+			id = apmu_config[k].cpus[bit];
+			if (id >= 0) {
+				if (id == cpu_logical_map(0))
+					is_allowed = true;
+			}
+		}
+		if (!is_allowed)
+			continue;
+
+		for (bit = 0; bit < ARRAY_SIZE(apmu_config[k].cpus); bit++) {
+			id = apmu_config[k].cpus[bit];
+			if (id >= 0) {
+				index = get_logical_index(id);
+				if (index >= 0)
+					fn(&apmu_config[k].iomem, index, bit);
+			}
+		}
+	}
+}
+
+void __init shmobile_smp_apmu_prepare_cpus(unsigned int max_cpus)
+{
+	/* install boot code shared by all CPUs */
+	shmobile_boot_fn = virt_to_phys(s