Merge branch 'master' into sh/hw-breakpoints

Conflict between FPU thread flag migration and debug
thread flag addition.

Conflicts:
	arch/sh/include/asm/thread_info.h
	arch/sh/include/asm/ubc.h
	arch/sh/kernel/process_32.c

129 files changed, 4299 insertions, 2882 deletions

diff --git a/arch/sh/Kconfig b/arch/sh/Kconfig
index d563884833e9..185e9af251cf 100644
--- a/arch/sh/Kconfig
+++ b/arch/sh/Kconfig
@@ -16,7 +16,9 @@ config SUPERH
         select HAVE_IOREMAP_PROT if MMU
         select HAVE_ARCH_TRACEHOOK
         select HAVE_DMA_API_DEBUG
+        select HAVE_DMA_ATTRS
         select HAVE_PERF_EVENTS
+        select PERF_USE_VMALLOC
         select HAVE_KERNEL_GZIP
         select HAVE_KERNEL_BZIP2
         select HAVE_KERNEL_LZMA
@@ -37,6 +39,7 @@ config SUPERH32
         select HAVE_FTRACE_MCOUNT_RECORD
         select HAVE_DYNAMIC_FTRACE
         select HAVE_FUNCTION_TRACE_MCOUNT_TEST
+        select HAVE_FTRACE_NMI_ENTER if DYNAMIC_FTRACE
         select HAVE_FUNCTION_GRAPH_TRACER
         select HAVE_ARCH_KGDB
         select HAVE_HW_BREAKPOINT if CPU_SH4A
@@ -171,6 +174,12 @@ config ARCH_HAS_CPU_IDLE_WAIT
 config IO_TRAPPED
         bool
 
+config DMA_COHERENT
+        bool
+
+config DMA_NONCOHERENT
+        def_bool !DMA_COHERENT
+
 source "init/Kconfig"
 
 source "kernel/Kconfig.freezer"
@@ -221,6 +230,7 @@ config CPU_SHX2
 
 config CPU_SHX3
         bool
+        select DMA_COHERENT
 
 config ARCH_SHMOBILE
         bool
@@ -762,17 +772,6 @@ config ENTRY_OFFSET
         default "0x00010000" if PAGE_SIZE_64KB
         default "0x00000000"
 
-config UBC_WAKEUP
-        bool "Wakeup UBC on startup"
-        depends on CPU_SH4 && !CPU_SH4A
-        help
-          Selecting this option will wakeup the User Break Controller (UBC) on
-          startup. Although the UBC is left in an awake state when the processor
-          comes up, some boot loaders misbehave by putting the UBC to sleep in a
-          power saving state, which causes issues with things like ptrace().
-
-          If unsure, say N.
-
 choice
         prompt "Kernel command line"
         optional
@@ -819,7 +818,13 @@ config MAPLE
          Dreamcast with a serial line terminal or a remote network
          connection.
 
-source "arch/sh/drivers/pci/Kconfig"
+config PCI
+        bool "PCI support"
+        depends on SYS_SUPPORTS_PCI
+        help
+          Find out whether you have a PCI motherboard. PCI is the name of a
+          bus system, i.e. the way the CPU talks to the other stuff inside
+          your box. If you have PCI, say Y, otherwise N.
 
 source "drivers/pci/pcie/Kconfig"
 
diff --git a/arch/sh/Makefile b/arch/sh/Makefile
index 66e40aabc600..ac17c5ac550e 100644
--- a/arch/sh/Makefile
+++ b/arch/sh/Makefile
@@ -78,6 +78,9 @@ defaultimage-$(CONFIG_SUPERH32)			:= zImage
 defaultimage-$(CONFIG_SH_SH7785LCR)             := uImage
 defaultimage-$(CONFIG_SH_RSK)                   := uImage
 defaultimage-$(CONFIG_SH_URQUELL)               := uImage
+defaultimage-$(CONFIG_SH_MIGOR)                 := uImage
+defaultimage-$(CONFIG_SH_AP325RXA)              := uImage
+defaultimage-$(CONFIG_SH_7724_SOLUTION_ENGINE)  := uImage
 defaultimage-$(CONFIG_SH_7206_SOLUTION_ENGINE)  := vmlinux
 defaultimage-$(CONFIG_SH_7619_SOLUTION_ENGINE)  := vmlinux
 
@@ -136,6 +139,7 @@ machdir-$(CONFIG_SH_7751_SYSTEMH)		+= mach-systemh
 machdir-$(CONFIG_SH_EDOSK7705)                  += mach-edosk7705
 machdir-$(CONFIG_SH_HIGHLANDER)                 += mach-highlander
 machdir-$(CONFIG_SH_MIGOR)                      += mach-migor
+machdir-$(CONFIG_SH_AP325RXA)                   += mach-ap325rxa
 machdir-$(CONFIG_SH_KFR2R09)                    += mach-kfr2r09
 machdir-$(CONFIG_SH_ECOVEC)                     += mach-ecovec24
 machdir-$(CONFIG_SH_SDK7780)                    += mach-sdk7780
diff --git a/arch/sh/boards/Makefile b/arch/sh/boards/Makefile
index 7baa21090231..ce0f26381784 100644
--- a/arch/sh/boards/Makefile
+++ b/arch/sh/boards/Makefile
@@ -1,7 +1,6 @@
 #
 # Specific board support, not covered by a mach group.
 #
-obj-$(CONFIG_SH_AP325RXA)       += board-ap325rxa.o
 obj-$(CONFIG_SH_MAGIC_PANEL_R2) += board-magicpanelr2.o
 obj-$(CONFIG_SH_SH7785LCR)      += board-sh7785lcr.o
 obj-$(CONFIG_SH_URQUELL)        += board-urquell.o
diff --git a/arch/sh/boards/mach-ap325rxa/Makefile b/arch/sh/boards/mach-ap325rxa/Makefile
new file mode 100644
index 000000000000..4cf1774d2613
--- /dev/null
+++ b/arch/sh/boards/mach-ap325rxa/Makefile
@@ -0,0 +1,2 @@
+obj-y    := setup.o sdram.o
+
diff --git a/arch/sh/boards/mach-ap325rxa/sdram.S b/arch/sh/boards/mach-ap325rxa/sdram.S
new file mode 100644
index 000000000000..db24fbed4fca
--- /dev/null
+++ b/arch/sh/boards/mach-ap325rxa/sdram.S
@@ -0,0 +1,69 @@
+/*
+ * AP325RXA sdram self/auto-refresh setup code
+ *
+ *  Copyright (C) 2009 Magnus Damm
+ *
+ * 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/sys.h>
+#include <linux/errno.h>
+#include <linux/linkage.h>
+#include <asm/asm-offsets.h>
+#include <asm/suspend.h>
+#include <asm/romimage-macros.h>
+
+/* code to enter and leave self-refresh. must be self-contained.
+ * this code will be copied to on-chip memory and executed from there.
+ */
+        .balign 4
+ENTRY(ap325rxa_sdram_enter_start)
+
+        /* SBSC: disable power down and put in self-refresh mode */
+        mov.l   1f, r4
+        mov.l   2f, r1
+        mov.l   @r4, r2
+        or      r1, r2
+        mov.l   3f, r3
+        and     r3, r2
+        mov.l   r2, @r4
+
+        rts
+         nop
+
+        .balign 4
+1:      .long   0xfe400008 /* SDCR0 */
+2:      .long   0x00000400
+3:      .long   0xffff7fff
+ENTRY(ap325rxa_sdram_enter_end)
+
+        .balign 4
+ENTRY(ap325rxa_sdram_leave_start)
+
+        /* SBSC: set auto-refresh mode */
+        mov.l   1f, r4
+        mov.l   @r4, r0
+        mov.l   4f, r1
+        and     r1, r0
+        mov.l   r0, @r4
+        mov.l   6f, r4
+        mov.l   8f, r0
+        mov.l   @r4, r1
+        mov     #-1, r4
+        add     r4, r1
+        or      r1, r0
+        mov.l   7f, r1
+        mov.l   r0, @r1
+
+        rts
+         nop
+
+        .balign 4
+1:      .long   0xfe400008 /* SDCR0 */
+4:      .long   0xfffffbff
+6:      .long   0xfe40001c /* RTCOR */
+7:      .long   0xfe400018 /* RTCNT */
+8:      .long   0xa55a0000
+ENTRY(ap325rxa_sdram_leave_end)
diff --git a/arch/sh/boards/board-ap325rxa.c b/arch/sh/boards/mach-ap325rxa/setup.c
index 2d080732a964..cf9dc12dfeb1 100644
--- a/arch/sh/boards/board-ap325rxa.c
+++ b/arch/sh/boards/mach-ap325rxa/setup.c
@@ -20,8 +20,6 @@
 #include <linux/i2c.h>
 #include <linux/smsc911x.h>
 #include <linux/gpio.h>
-#include <linux/spi/spi.h>
-#include <linux/spi/spi_gpio.h>
 #include <media/ov772x.h>
 #include <media/soc_camera.h>
 #include <media/soc_camera_platform.h>
@@ -29,6 +27,7 @@
 #include <video/sh_mobile_lcdc.h>
 #include <asm/io.h>
 #include <asm/clock.h>
+#include <asm/suspend.h>
 #include <cpu/sh7723.h>
 
 static struct smsc911x_platform_config smsc911x_config = {
@@ -409,17 +408,49 @@ static struct platform_device ceu_device = {
         },
 };
 
-struct spi_gpio_platform_data sdcard_cn3_platform_data = {
-        .sck = GPIO_PTD0,
-        .mosi = GPIO_PTD1,
-        .miso = GPIO_PTD2,
-        .num_chipselect = 1,
+static struct resource sdhi0_cn3_resources[] = {
+        [0] = {
+                .name   = "SDHI0",
+                .start  = 0x04ce0000,
+                .end    = 0x04ce01ff,
+                .flags  = IORESOURCE_MEM,
+        },
+        [1] = {
+                .start  = 101,
+                .flags  = IORESOURCE_IRQ,
+        },
 };
 
-static struct platform_device sdcard_cn3_device = {
-        .name           = "spi_gpio",
-        .dev    = {
-                .platform_data  = &sdcard_cn3_platform_data,
+static struct platform_device sdhi0_cn3_device = {
+        .name           = "sh_mobile_sdhi",
+        .id             = 0, /* "sdhi0" clock */
+        .num_resources  = ARRAY_SIZE(sdhi0_cn3_resources),
+        .resource       = sdhi0_cn3_resources,
+        .archdata = {
+                .hwblk_id = HWBLK_SDHI0,
+        },
+};
+
+static struct resource sdhi1_cn7_resources[] = {
+        [0] = {
+                .name   = "SDHI1",
+                .start  = 0x04cf0000,
+                .end    = 0x04cf01ff,
+                .flags  = IORESOURCE_MEM,
+        },
+        [1] = {
+                .start  = 24,
+                .flags  = IORESOURCE_IRQ,
+        },
+};
+
+static struct platform_device sdhi1_cn7_device = {
+        .name           = "sh_mobile_sdhi",
+        .id             = 1, /* "sdhi1" clock */
+        .num_resources  = ARRAY_SIZE(sdhi1_cn7_resources),
+        .resource       = sdhi1_cn7_resources,
+        .archdata = {
+                .hwblk_id = HWBLK_SDHI1,
         },
 };
 
@@ -470,22 +501,26 @@ static struct platform_device *ap325rxa_devices[] __initdata = {
         &lcdc_device,
         &ceu_device,
         &nand_flash_device,
-        &sdcard_cn3_device,
+        &sdhi0_cn3_device,
+        &sdhi1_cn7_device,
         &ap325rxa_camera[0],
         &ap325rxa_camera[1],
 };
 
-static struct spi_board_info ap325rxa_spi_devices[] = {
-        {
-                .modalias = "mmc_spi",
-                .max_speed_hz = 5000000,
-                .chip_select = 0,
-                .controller_data = (void *) GPIO_PTD5,
-        },
-};
+extern char ap325rxa_sdram_enter_start;
+extern char ap325rxa_sdram_enter_end;
+extern char ap325rxa_sdram_leave_start;
+extern char ap325rxa_sdram_leave_end;
 
 static int __init ap325rxa_devices_setup(void)
 {
+        /* register board specific self-refresh code */
+        sh_mobile_register_self_refresh(SUSP_SH_STANDBY | SUSP_SH_SF,
+                                        &ap325rxa_sdram_enter_start,
+                                        &ap325rxa_sdram_enter_end,
+                                        &ap325rxa_sdram_leave_start,
+                                        &ap325rxa_sdram_leave_end);
+
         /* LD3 and LD4 LEDs */
         gpio_request(GPIO_PTX5, NULL); /* RUN */
         gpio_direction_output(GPIO_PTX5, 1);
@@ -578,12 +613,28 @@ static int __init ap325rxa_devices_setup(void)
 
         platform_resource_setup_memory(&ceu_device, "ceu", 4 << 20);
 
+        /* SDHI0 - CN3 - SD CARD */
+        gpio_request(GPIO_FN_SDHI0CD_PTD, NULL);
+        gpio_request(GPIO_FN_SDHI0WP_PTD, NULL);
+        gpio_request(GPIO_FN_SDHI0D3_PTD, NULL);
+        gpio_request(GPIO_FN_SDHI0D2_PTD, NULL);
+        gpio_request(GPIO_FN_SDHI0D1_PTD, NULL);
+        gpio_request(GPIO_FN_SDHI0D0_PTD, NULL);
+        gpio_request(GPIO_FN_SDHI0CMD_PTD, NULL);
+        gpio_request(GPIO_FN_SDHI0CLK_PTD, NULL);
+
+        /* SDHI1 - CN7 - MICRO SD CARD */
+        gpio_request(GPIO_FN_SDHI1CD, NULL);
+        gpio_request(GPIO_FN_SDHI1D3, NULL);
+        gpio_request(GPIO_FN_SDHI1D2, NULL);
+        gpio_request(GPIO_FN_SDHI1D1, NULL);
+        gpio_request(GPIO_FN_SDHI1D0, NULL);
+        gpio_request(GPIO_FN_SDHI1CMD, NULL);
+        gpio_request(GPIO_FN_SDHI1CLK, NULL);
+
         i2c_register_board_info(0, ap325rxa_i2c_devices,
                                 ARRAY_SIZE(ap325rxa_i2c_devices));
 
-        spi_register_board_info(ap325rxa_spi_devices,
-                                ARRAY_SIZE(ap325rxa_spi_devices));
-
         return platform_add_devices(ap325rxa_devices,
                                 ARRAY_SIZE(ap325rxa_devices));
 }
diff --git a/arch/sh/boards/mach-ecovec24/Makefile b/arch/sh/boards/mach-ecovec24/Makefile
index 51f852151655..e69bc82208fc 100644
--- a/arch/sh/boards/mach-ecovec24/Makefile
+++ b/arch/sh/boards/mach-ecovec24/Makefile
@@ -6,4 +6,4 @@
 # for more details.
 #
 
-obj-y    := setup.o
\ No newline at end of file
+obj-y    := setup.o sdram.o
\ No newline at end of file
diff --git a/arch/sh/boards/mach-ecovec24/sdram.S b/arch/sh/boards/mach-ecovec24/sdram.S
new file mode 100644
index 000000000000..833440044407
--- /dev/null
+++ b/arch/sh/boards/mach-ecovec24/sdram.S
@@ -0,0 +1,52 @@
+/*
+ * Ecovec24 sdram self/auto-refresh setup code
+ *
+ *  Copyright (C) 2009 Magnus Damm
+ *
+ * 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/sys.h>
+#include <linux/errno.h>
+#include <linux/linkage.h>
+#include <asm/asm-offsets.h>
+#include <asm/suspend.h>
+#include <asm/romimage-macros.h>
+
+/* code to enter and leave self-refresh. must be self-contained.
+ * this code will be copied to on-chip memory and executed from there.
+ */
+        .balign 4
+ENTRY(ecovec24_sdram_enter_start)
+
+        /* DBSC: put memory in self-refresh mode */
+
+        ED 0xFD000010, 0x00000000 /* DBEN */
+        ED 0xFD000040, 0x00000000 /* DBRFPDN0 */
+        ED 0xFD000014, 0x00000002 /* DBCMDCNT (PALL) */
+        ED 0xFD000014, 0x00000004 /* DBCMDCNT (REF) */
+        ED 0xFD000040, 0x00000001 /* DBRFPDN0 */
+
+        rts
+         nop
+
+ENTRY(ecovec24_sdram_enter_end)
+
+        .balign 4
+ENTRY(ecovec24_sdram_leave_start)
+
+        /* DBSC: put memory in auto-refresh mode */
+
+        ED 0xFD000040, 0x00000000 /* DBRFPDN0 */
+        WAIT 1
+        ED 0xFD000014, 0x00000002 /* DBCMDCNT (PALL) */
+        ED 0xFD000014, 0x00000004 /* DBCMDCNT (REF) */
+        ED 0xFD000010, 0x00000001 /* DBEN */
+        ED 0xFD000040, 0x00010000 /* DBRFPDN0 */
+
+        rts
+         nop
+
+ENTRY(ecovec24_sdram_leave_end)
diff --git a/arch/sh/boards/mach-ecovec24/setup.c b/arch/sh/boards/mach-ecovec24/setup.c
index 3b1ceb46fa54..826e62326d51 100644
--- a/arch/sh/boards/mach-ecovec24/setup.c
+++ b/arch/sh/boards/mach-ecovec24/setup.c
@@ -20,12 +20,14 @@
 #include <linux/i2c.h>
 #include <linux/i2c/tsc2007.h>
 #include <linux/input.h>
+#include <linux/input/sh_keysc.h>
+#include <linux/mfd/sh_mobile_sdhi.h>
 #include <video/sh_mobile_lcdc.h>
 #include <media/sh_mobile_ceu.h>
 #include <asm/heartbeat.h>
 #include <asm/sh_eth.h>
-#include <asm/sh_keysc.h>
 #include <asm/clock.h>
+#include <asm/suspend.h>
 #include <cpu/sh7724.h>
 
 /*
@@ -147,6 +149,9 @@ static struct platform_device sh_eth_device = {
         },
         .num_resources = ARRAY_SIZE(sh_eth_resources),
         .resource = sh_eth_resources,
+        .archdata = {
+                .hwblk_id = HWBLK_ETHER,
+        },
 };
 
 /* USB0 host */
@@ -185,30 +190,18 @@ static struct platform_device usb0_host_device = {
         .resource       = usb0_host_resources,
 };
 
-/*
- * USB1
- *
- * CN5 can use both host/function,
- * and we can determine it by checking PTB[3]
- *
- * This time only USB1 host is supported.
- */
+/* USB1 host/function */
 void usb1_port_power(int port, int power)
 {
-        if (!gpio_get_value(GPIO_PTB3)) {
-                printk(KERN_ERR "USB1 function is not supported\n");
-                return;
-        }
-
         gpio_set_value(GPIO_PTB5, power);
 }
 
-static struct r8a66597_platdata usb1_host_data = {
+static struct r8a66597_platdata usb1_common_data = {
         .on_chip = 1,
         .port_power = usb1_port_power,
 };
 
-static struct resource usb1_host_resources[] = {
+static struct resource usb1_common_resources[] = {
         [0] = {
                 .start  = 0xa4d90000,
                 .end    = 0xa4d90124 - 1,
@@ -221,16 +214,16 @@ static struct resource usb1_host_resources[] = {
         },
 };
 
-static struct platform_device usb1_host_device = {
-        .name           = "r8a66597_hcd",
+static struct platform_device usb1_common_device = {
+        /* .name will be added in arch_setup */
         .id             = 1,
         .dev = {
                 .dma_mask               = NULL,         /*  not use dma */
                 .coherent_dma_mask      = 0xffffffff,
-                .platform_data          = &usb1_host_data,
+                .platform_data          = &usb1_common_data,
         },
-        .num_resources  = ARRAY_SIZE(usb1_host_resources),
-        .resource       = usb1_host_resources,
+        .num_resources  = ARRAY_SIZE(usb1_common_resources),
+        .resource       = usb1_common_resources,
 };
 
 /* LCDC */
@@ -428,16 +421,90 @@ static struct i2c_board_info ts_i2c_clients = {
         .irq            = IRQ0,
 };
 
+/* SHDI0 */
+static void sdhi0_set_pwr(struct platform_device *pdev, int state)
+{
+        gpio_set_value(GPIO_PTB6, state);
+}
+
+static struct sh_mobile_sdhi_info sdhi0_info = {
+        .set_pwr = sdhi0_set_pwr,
+};
+
+static struct resource sdhi0_resources[] = {
+        [0] = {
+                .name   = "SDHI0",
+                .start  = 0x04ce0000,
+                .end    = 0x04ce01ff,
+                .flags  = IORESOURCE_MEM,
+        },
+        [1] = {
+                .start  = 101,
+                .flags  = IORESOURCE_IRQ,
+        },
+};
+
+static struct platform_device sdhi0_device = {
+        .name           = "sh_mobile_sdhi",
+        .num_resources  = ARRAY_SIZE(sdhi0_resources),
+        .resource       = sdhi0_resources,
+        .id             = 0,
+        .dev    = {
+                .platform_data  = &sdhi0_info,
+        },
+        .archdata = {
+                .hwblk_id = HWBLK_SDHI0,
+        },
+};
+
+/* SHDI1 */
+static void sdhi1_set_pwr(struct platform_device *pdev, int state)
+{
+        gpio_set_value(GPIO_PTB7, state);
+}
+
+static struct sh_mobile_sdhi_info sdhi1_info = {
+        .set_pwr = sdhi1_set_pwr,
+};
+
+static struct resource sdhi1_resources[] = {
+        [0] = {
+                .name   = "SDHI1",
+                .start  = 0x04cf0000,
+                .end    = 0x04cf01ff,
+                .flags  = IORESOURCE_MEM,
+        },
+        [1] = {
+                .start  = 24,
+                .flags  = IORESOURCE_IRQ,
+        },
+};
+
+static struct platform_device sdhi1_device = {
+        .name           = "sh_mobile_sdhi",
+        .num_resources  = ARRAY_SIZE(sdhi1_resources),
+        .resource       = sdhi1_resources,
+        .id             = 1,
+        .dev    = {
+                .platform_data  = &sdhi1_info,
+        },
+        .archdata = {
+                .hwblk_id = HWBLK_SDHI1,
+        },
+};
+
 static struct platform_device *ecovec_devices[] __initdata = {
         &heartbeat_device,
         &nor_flash_device,
         &sh_eth_device,
         &usb0_host_device,
-        &usb1_host_device, /* USB1 host support */
+        &usb1_common_device,
         &lcdc_device,
         &ceu0_device,
         &ceu1_device,
         &keysc_device,
+        &sdhi0_device,
+        &sdhi1_device,
 };
 
 #define EEPROM_ADDR 0x50
@@ -466,12 +533,9 @@ static u8 mac_read(struct i2c_adapter *a, u8 command)
         return buf;
 }
 
-#define MAC_LEN 6
-static void __init sh_eth_init(void)
+static void __init sh_eth_init(struct sh_eth_plat_data *pd)
 {
         struct i2c_adapter *a = i2c_get_adapter(1);
-        struct clk *eth_clk;
-        u8 mac[MAC_LEN];
         int i;
 
         if (!a) {
@@ -479,39 +543,30 @@ static void __init sh_eth_init(void)
                 return;
         }
 
-        eth_clk = clk_get(NULL, "eth0");
-        if (!eth_clk) {
-                pr_err("can not get eth0 clk\n");
-                return;
-        }
-
         /* read MAC address frome EEPROM */
-        for (i = 0; i < MAC_LEN; i++) {
-                mac[i] = mac_read(a, 0x10 + i);
+        for (i = 0; i < sizeof(pd->mac_addr); i++) {
+                pd->mac_addr[i] = mac_read(a, 0x10 + i);
                 msleep(10);
         }
-
-        /* clock enable */
-        clk_enable(eth_clk);
-
-        /* reset sh-eth */
-        ctrl_outl(0x1, SH_ETH_ADDR + 0x0);
-
-        /* set MAC addr */
-        ctrl_outl((mac[0] << 24) |
-                  (mac[1] << 16) |
-                  (mac[2] <<  8) |
-                  (mac[3] <<  0), SH_ETH_MAHR);
-        ctrl_outl((mac[4] <<  8) |
-                  (mac[5] <<  0), SH_ETH_MALR);
-
-        clk_put(eth_clk);
 }
 
 #define PORT_HIZA 0xA4050158
 #define IODRIVEA  0xA405018A
+
+extern char ecovec24_sdram_enter_start;
+extern char ecovec24_sdram_enter_end;
+extern char ecovec24_sdram_leave_start;
+extern char ecovec24_sdram_leave_end;
+
 static int __init arch_setup(void)
 {
+        /* register board specific self-refresh code */
+        sh_mobile_register_self_refresh(SUSP_SH_STANDBY | SUSP_SH_SF,
+                                        &ecovec24_sdram_enter_start,
+                                        &ecovec24_sdram_enter_end,
+                                        &ecovec24_sdram_leave_start,
+                                        &ecovec24_sdram_leave_end);
+
         /* enable STATUS0, STATUS2 and PDSTATUS */
         gpio_request(GPIO_FN_STATUS0, NULL);
         gpio_request(GPIO_FN_STATUS2, NULL);
@@ -561,6 +616,14 @@ static int __init arch_setup(void)
         ctrl_outw(0x0600, 0xa40501d4);
         ctrl_outw(0x0600, 0xa4050192);
 
+        if (gpio_get_value(GPIO_PTB3)) {
+                printk(KERN_INFO "USB1 function is selected\n");
+                usb1_common_device.name = "r8a66597_udc";
+        } else {
+                printk(KERN_INFO "USB1 host is selected\n");
+                usb1_common_device.name = "r8a66597_hcd";
+        }
+
         /* enable LCDC */
         gpio_request(GPIO_FN_LCDD23,   NULL);
         gpio_request(GPIO_FN_LCDD22,   NULL);
@@ -603,8 +666,8 @@ static int __init arch_setup(void)
         gpio_direction_output(GPIO_PTR1, 0);
         gpio_direction_output(GPIO_PTA2, 0);
 
-        /* I/O buffer drive ability is low */
-        ctrl_outw((ctrl_inw(IODRIVEA) & ~0x00c0) | 0x0040 , IODRIVEA);
+        /* I/O buffer drive ability is high */
+        ctrl_outw((ctrl_inw(IODRIVEA) & ~0x00c0) | 0x0080 , IODRIVEA);
 
         if (gpio_get_value(GPIO_PTE6)) {
                 /* DVI */
@@ -710,6 +773,33 @@ static int __init arch_setup(void)
         gpio_direction_input(GPIO_PTR5);
         gpio_direction_input(GPIO_PTR6);
 
+        /* enable SDHI0 (needs DS2.4 set to ON) */
+        gpio_request(GPIO_FN_SDHI0CD,  NULL);
+        gpio_request(GPIO_FN_SDHI0WP,  NULL);
+        gpio_request(GPIO_FN_SDHI0CMD, NULL);
+        gpio_request(GPIO_FN_SDHI0CLK, NULL);
+        gpio_request(GPIO_FN_SDHI0D3,  NULL);
+        gpio_request(GPIO_FN_SDHI0D2,  NULL);
+        gpio_request(GPIO_FN_SDHI0D1,  NULL);
+        gpio_request(GPIO_FN_SDHI0D0,  NULL);
+        gpio_request(GPIO_PTB6, NULL);
+        gpio_direction_output(GPIO_PTB6, 0);
+
+        /* enable SDHI1 (needs DS2.6,7 set to ON,OFF) */
+        gpio_request(GPIO_FN_SDHI1CD,  NULL);
+        gpio_request(GPIO_FN_SDHI1WP,  NULL);
+        gpio_request(GPIO_FN_SDHI1CMD, NULL);
+        gpio_request(GPIO_FN_SDHI1CLK, NULL);
+        gpio_request(GPIO_FN_SDHI1D3,  NULL);
+        gpio_request(GPIO_FN_SDHI1D2,  NULL);
+        gpio_request(GPIO_FN_SDHI1D1,  NULL);
+        gpio_request(GPIO_FN_SDHI1D0,  NULL);
+        gpio_request(GPIO_PTB7, NULL);
+        gpio_direction_output(GPIO_PTB7, 0);
+
+        /* I/O buffer drive ability is high for SDHI1 */
+        ctrl_outw((ctrl_inw(IODRIVEA) & ~0x3000) | 0x2000 , IODRIVEA);
+
         /* enable I2C device */
         i2c_register_board_info(1, i2c1_devices,
                                 ARRAY_SIZE(i2c1_devices));
@@ -721,12 +811,11 @@ arch_initcall(arch_setup);
 
 static int __init devices_setup(void)
 {
-        sh_eth_init();
+        sh_eth_init(&sh_eth_plat);
         return 0;
 }
 device_initcall(devices_setup);
 
-
 static struct sh_machine_vector mv_ecovec __initmv = {
         .mv_name        = "R0P7724 (EcoVec)",
 };
diff --git a/arch/sh/boards/mach-highlander/setup.c b/arch/sh/boards/mach-highlander/setup.c
index 566e69d8d729..f663c14d8885 100644
--- a/arch/sh/boards/mach-highlander/setup.c
+++ b/arch/sh/boards/mach-highlander/setup.c
@@ -384,7 +384,7 @@ static unsigned char irl2irq[HL_NR_IRL];
 
 static int highlander_irq_demux(int irq)
 {
-        if (irq >= HL_NR_IRL || !irl2irq[irq])
+        if (irq >= HL_NR_IRL || irq < 0 || !irl2irq[irq])
                 return irq;
 
         return irl2irq[irq];
diff --git a/arch/sh/boards/mach-kfr2r09/Makefile b/arch/sh/boards/mach-kfr2r09/Makefile
index 5d5867826e3b..4e577a3bf658 100644
--- a/arch/sh/boards/mach-kfr2r09/Makefile
+++ b/arch/sh/boards/mach-kfr2r09/Makefile
@@ -1,2 +1,2 @@
-obj-y    := setup.o
+obj-y    := setup.o sdram.o
 obj-$(CONFIG_FB_SH_MOBILE_LCDC) +=  lcd_wqvga.o
diff --git a/arch/sh/boards/mach-kfr2r09/sdram.S b/arch/sh/boards/mach-kfr2r09/sdram.S
new file mode 100644
index 000000000000..0c9f55bec2fe
--- /dev/null
+++ b/arch/sh/boards/mach-kfr2r09/sdram.S
@@ -0,0 +1,80 @@
+/*
+ * KFR2R09 sdram self/auto-refresh setup code
+ *
+ *  Copyright (C) 2009 Magnus Damm
+ *
+ * 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/sys.h>
+#include <linux/errno.h>
+#include <linux/linkage.h>
+#include <asm/asm-offsets.h>
+#include <asm/suspend.h>
+#include <asm/romimage-macros.h>
+
+/* code to enter and leave self-refresh. must be self-contained.
+ * this code will be copied to on-chip memory and executed from there.
+ */
+        .balign 4
+ENTRY(kfr2r09_sdram_enter_start)
+
+        /* DBSC: put memory in self-refresh mode */
+
+        ED 0xFD000010, 0x00000000 /* DBEN */
+        ED 0xFD000040, 0x00000000 /* DBRFPDN0 */
+        ED 0xFD000014, 0x00000002 /* DBCMDCNT (PALL) */
+        ED 0xFD000014, 0x00000004 /* DBCMDCNT (REF) */
+        ED 0xFD000040, 0x00000001 /* DBRFPDN0 */
+
+        rts
+         nop
+
+ENTRY(kfr2r09_sdram_enter_end)
+
+        .balign 4
+ENTRY(kfr2r09_sdram_leave_start)
+
+        /* DBSC: put memory in auto-refresh mode */
+
+        mov.l   @(SH_SLEEP_MODE, r5), r0
+        tst     #SUSP_SH_RSTANDBY, r0
+        bf      resume_rstandby
+
+        ED 0xFD000040, 0x00000000 /* DBRFPDN0 */
+        WAIT 1
+        ED 0xFD000014, 0x00000002 /* DBCMDCNT (PALL) */
+        ED 0xFD000014, 0x00000004 /* DBCMDCNT (REF) */
+        ED 0xFD000010, 0x00000001 /* DBEN */
+        ED 0xFD000040, 0x00010000 /* DBRFPDN0 */
+
+        rts
+         nop
+
+resume_rstandby:
+
+        /* DBSC: re-initialize and put in auto-refresh */
+
+        ED 0xFD000108, 0x40000301 /* DBPDCNT0 */
+        ED 0xFD000020, 0x011B0002 /* DBCONF */
+        ED 0xFD000030, 0x03060E02 /* DBTR0 */
+        ED 0xFD000034, 0x01020102 /* DBTR1 */
+        ED 0xFD000038, 0x01090406 /* DBTR2 */
+        ED 0xFD000008, 0x00000004 /* DBKIND */
+        ED 0xFD000040, 0x00000001 /* DBRFPDN0 */
+        ED 0xFD000040, 0x00000000 /* DBRFPDN0 */
+        ED 0xFD000018, 0x00000001 /* DBCKECNT */
+        WAIT 1
+        ED 0xFD000010, 0x00000001 /* DBEN */
+        ED 0xFD000044, 0x000004AF /* DBRFPDN1 */
+        ED 0xFD000048, 0x20CF0037 /* DBRFPDN2 */
+        ED 0xFD000014, 0x00000004 /* DBCMDCNT (REF) */
+        ED 0xFD000108, 0x40000300 /* DBPDCNT0 */
+        ED 0xFD000040, 0x00010000 /* DBRFPDN0 */
+
+        rts
+         nop
+
+ENTRY(kfr2r09_sdram_leave_end)
diff --git a/arch/sh/boards/mach-kfr2r09/setup.c b/arch/sh/boards/mach-kfr2r09/setup.c
index c08d33fe2104..87438d6603d6 100644
--- a/arch/sh/boards/mach-kfr2r09/setup.c
+++ b/arch/sh/boards/mach-kfr2r09/setup.c
@@ -16,13 +16,16 @@
 #include <linux/clk.h>
 #include <linux/gpio.h>
 #include <linux/input.h>
+#include <linux/input/sh_keysc.h>
 #include <linux/i2c.h>
 #include <linux/usb/r8a66597.h>
+#include <media/soc_camera.h>
+#include <media/sh_mobile_ceu.h>
 #include <video/sh_mobile_lcdc.h>
+#include <asm/suspend.h>
 #include <asm/clock.h>
 #include <asm/machvec.h>
 #include <asm/io.h>
-#include <asm/sh_keysc.h>
 #include <cpu/sh7724.h>
 #include <mach/kfr2r09.h>
 
@@ -212,11 +215,154 @@ static struct platform_device kfr2r09_usb0_gadget_device = {
         .resource       = kfr2r09_usb0_gadget_resources,
 };
 
+static struct sh_mobile_ceu_info sh_mobile_ceu_info = {
+        .flags = SH_CEU_FLAG_USE_8BIT_BUS,
+};
+
+static struct resource kfr2r09_ceu_resources[] = {
+        [0] = {
+                .name   = "CEU",
+                .start  = 0xfe910000,
+                .end    = 0xfe91009f,
+                .flags  = IORESOURCE_MEM,
+        },
+        [1] = {
+                .start  = 52,
+                .end  = 52,
+                .flags  = IORESOURCE_IRQ,
+        },
+        [2] = {
+                /* place holder for contiguous memory */
+        },
+};
+
+static struct platform_device kfr2r09_ceu_device = {
+        .name           = "sh_mobile_ceu",
+        .id             = 0, /* "ceu0" clock */
+        .num_resources  = ARRAY_SIZE(kfr2r09_ceu_resources),
+        .resource       = kfr2r09_ceu_resources,
+        .dev    = {
+                .platform_data  = &sh_mobile_ceu_info,
+        },
+        .archdata = {
+                .hwblk_id = HWBLK_CEU0,
+        },
+};
+
+static struct i2c_board_info kfr2r09_i2c_camera = {
+        I2C_BOARD_INFO("rj54n1cb0c", 0x50),
+};
+
+static struct clk *camera_clk;
+
+#define DRVCRB 0xA405018C
+static int camera_power(struct device *dev, int mode)
+{
+        int ret;
+
+        if (mode) {
+                long rate;
+
+                camera_clk = clk_get(NULL, "video_clk");
+                if (IS_ERR(camera_clk))
+                        return PTR_ERR(camera_clk);
+
+                /* set VIO_CKO clock to 25MHz */
+                rate = clk_round_rate(camera_clk, 25000000);
+                ret = clk_set_rate(camera_clk, rate);
+                if (ret < 0)
+                        goto eclkrate;
+
+                /* set DRVCRB
+                 *
+                 * use 1.8 V for VccQ_VIO
+                 * use 2.85V for VccQ_SR
+                 */
+                ctrl_outw((ctrl_inw(DRVCRB) & ~0x0003) | 0x0001, DRVCRB);
+
+                /* reset clear */
+                ret = gpio_request(GPIO_PTB4, NULL);
+                if (ret < 0)
+                        goto eptb4;
+                ret = gpio_request(GPIO_PTB7, NULL);
+                if (ret < 0)
+                        goto eptb7;
+
+                ret = gpio_direction_output(GPIO_PTB4, 1);
+                if (!ret)
+                        ret = gpio_direction_output(GPIO_PTB7, 1);
+                if (ret < 0)
+                        goto egpioout;
+                msleep(1);
+
+                ret = clk_enable(camera_clk);   /* start VIO_CKO */
+                if (ret < 0)
+                        goto eclkon;
+
+                return 0;
+        }
+
+        ret = 0;
+
+        clk_disable(camera_clk);
+eclkon:
+        gpio_set_value(GPIO_PTB7, 0);
+egpioout:
+        gpio_set_value(GPIO_PTB4, 0);
+        gpio_free(GPIO_PTB7);
+eptb7:
+        gpio_free(GPIO_PTB4);
+eptb4:
+eclkrate:
+        clk_put(camera_clk);
+        return ret;
+}
+
+static struct soc_camera_link rj54n1_link = {
+        .power          = camera_power,
+        .board_info     = &kfr2r09_i2c_camera,
+        .i2c_adapter_id = 1,
+        .module_name    = "rj54n1cb0c",
+};
+
+static struct platform_device kfr2r09_camera = {
+        .name   = "soc-camera-pdrv",
+        .id     = 0,
+        .dev    = {
+                .platform_data = &rj54n1_link,
+        },
+};
+
+static struct resource kfr2r09_sh_sdhi0_resources[] = {
+        [0] = {
+                .name   = "SDHI0",
+                .start  = 0x04ce0000,
+                .end    = 0x04ce01ff,
+                .flags  = IORESOURCE_MEM,
+        },
+        [1] = {
+                .start  = 101,
+                .flags  = IORESOURCE_IRQ,
+        },
+};
+
+static struct platform_device kfr2r09_sh_sdhi0_device = {
+        .name           = "sh_mobile_sdhi",
+        .num_resources  = ARRAY_SIZE(kfr2r09_sh_sdhi0_resources),
+        .resource       = kfr2r09_sh_sdhi0_resources,
+        .archdata = {
+                .hwblk_id = HWBLK_SDHI0,
+        },
+};
+
 static struct platform_device *kfr2r09_devices[] __initdata = {
         &kfr2r09_nor_flash_device,
         &kfr2r09_nand_flash_device,
         &kfr2r09_sh_keysc_device,
         &kfr2r09_sh_lcdc_device,
+        &kfr2r09_ceu_device,
+        &kfr2r09_camera,
+        &kfr2r09_sh_sdhi0_device,
 };
 
 #define BSC_CS0BCR 0xfec10004
@@ -268,11 +414,59 @@ static int kfr2r09_usb0_gadget_i2c_setup(void)
 
         return 0;
 }
+
+static int kfr2r09_serial_i2c_setup(void)
+{
+        struct i2c_adapter *a;
+        struct i2c_msg msg;
+        unsigned char buf[2];
+        int ret;
+
+        a = i2c_get_adapter(0);
+        if (!a)
+                return -ENODEV;
+
+        /* set bit 6 (the 7th bit) of chip at 0x09, register 0x13 */
+        buf[0] = 0x13;
+        msg.addr = 0x09;
+        msg.buf = buf;
+        msg.len = 1;
+        msg.flags = 0;
+        ret = i2c_transfer(a, &msg, 1);
+        if (ret != 1)
+                return -ENODEV;
+
+        buf[0] = 0;
+        msg.addr = 0x09;
+        msg.buf = buf;
+        msg.len = 1;
+        msg.flags = I2C_M_RD;
+        ret = i2c_transfer(a, &msg, 1);
+        if (ret != 1)
+                return -ENODEV;
+
+        buf[1] = buf[0] | (1 << 6);
+        buf[0] = 0x13;
+        msg.addr = 0x09;
+        msg.buf = buf;
+        msg.len = 2;
+        msg.flags = 0;
+        ret = i2c_transfer(a, &msg, 1);
+        if (ret != 1)
+                return -ENODEV;
+
+        return 0;
+}
 #else
 static int kfr2r09_usb0_gadget_i2c_setup(void)
 {
         return -ENODEV;
 }
+
+static int kfr2r09_serial_i2c_setup(void)
+{
+        return -ENODEV;
+}
 #endif
 
 static int kfr2r09_usb0_gadget_setup(void)
@@ -299,11 +493,27 @@ static int kfr2r09_usb0_gadget_setup(void)
         return 0;
 }
 
+extern char kfr2r09_sdram_enter_start;
+extern char kfr2r09_sdram_enter_end;
+extern char kfr2r09_sdram_leave_start;
+extern char kfr2r09_sdram_leave_end;
+
 static int __init kfr2r09_devices_setup(void)
 {
+        /* register board specific self-refresh code */
+        sh_mobile_register_self_refresh(SUSP_SH_STANDBY | SUSP_SH_SF |
+                                        SUSP_SH_RSTANDBY,
+                                        &kfr2r09_sdram_enter_start,
+                                        &kfr2r09_sdram_enter_end,
+                                        &kfr2r09_sdram_leave_start,
+                                        &kfr2r09_sdram_leave_end);
+
         /* enable SCIF1 serial port for YC401 console support */
         gpio_request(GPIO_FN_SCIF1_RXD, NULL);
         gpio_request(GPIO_FN_SCIF1_TXD, NULL);
+        kfr2r09_serial_i2c_setup(); /* ECONTMSK(bit6=L10ONEN) set 1 */
+        gpio_request(GPIO_PTG3, NULL); /* HPON_ON */
+        gpio_direction_output(GPIO_PTG3, 1); /* HPON_ON = H */
 
         /* setup NOR flash at CS0 */
         ctrl_outl(0x36db0400, BSC_CS0BCR);
@@ -361,6 +571,32 @@ static int __init kfr2r09_devices_setup(void)
         if (kfr2r09_usb0_gadget_setup() == 0)
                 platform_device_register(&kfr2r09_usb0_gadget_device);
 
+        /* CEU */
+        gpio_request(GPIO_FN_VIO_CKO, NULL);
+        gpio_request(GPIO_FN_VIO0_CLK, NULL);
+        gpio_request(GPIO_FN_VIO0_VD, NULL);
+        gpio_request(GPIO_FN_VIO0_HD, NULL);
+        gpio_request(GPIO_FN_VIO0_FLD, NULL);
+        gpio_request(GPIO_FN_VIO0_D7, NULL);
+        gpio_request(GPIO_FN_VIO0_D6, NULL);
+        gpio_request(GPIO_FN_VIO0_D5, NULL);
+        gpio_request(GPIO_FN_VIO0_D4, NULL);
+        gpio_request(GPIO_FN_VIO0_D3, NULL);
+        gpio_request(GPIO_FN_VIO0_D2, NULL);
+        gpio_request(GPIO_FN_VIO0_D1, NULL);
+        gpio_request(GPIO_FN_VIO0_D0, NULL);
+
+        platform_resource_setup_memory(&kfr2r09_ceu_device, "ceu", 4 << 20);
+
+        /* SDHI0 connected to yc304 */
+        gpio_request(GPIO_FN_SDHI0CD, NULL);
+        gpio_request(GPIO_FN_SDHI0D3, NULL);
+        gpio_request(GPIO_FN_SDHI0D2, NULL);
+        gpio_request(GPIO_FN_SDHI0D1, NULL);
+        gpio_request(GPIO_FN_SDHI0D0, NULL);
+        gpio_request(GPIO_FN_SDHI0CMD, NULL);
+        gpio_request(GPIO_FN_SDHI0CLK, NULL);
+
         return platform_add_devices(kfr2r09_devices,
                                     ARRAY_SIZE(kfr2r09_devices));
 }
diff --git a/arch/sh/boards/mach-migor/Makefile b/arch/sh/boards/mach-migor/Makefile
index 5f231dd25c0e..4601a89e5ac7 100644
--- a/arch/sh/boards/mach-migor/Makefile
+++ b/arch/sh/boards/mach-migor/Makefile
@@ -1,2 +1,2 @@
-obj-y    := setup.o
+obj-y    := setup.o sdram.o
 obj-$(CONFIG_SH_MIGOR_QVGA)     +=  lcd_qvga.o
diff --git a/arch/sh/boards/mach-migor/sdram.S b/arch/sh/boards/mach-migor/sdram.S
new file mode 100644
index 000000000000..614aa3a1398c
--- /dev/null
+++ b/arch/sh/boards/mach-migor/sdram.S
@@ -0,0 +1,69 @@
+/*
+ * Migo-R sdram self/auto-refresh setup code
+ *
+ *  Copyright (C) 2009 Magnus Damm
+ *
+ * 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/sys.h>
+#include <linux/errno.h>
+#include <linux/linkage.h>
+#include <asm/asm-offsets.h>
+#include <asm/suspend.h>
+#include <asm/romimage-macros.h>
+
+/* code to enter and leave self-refresh. must be self-contained.
+ * this code will be copied to on-chip memory and executed from there.
+ */
+        .balign 4
+ENTRY(migor_sdram_enter_start)
+
+        /* SBSC: disable power down and put in self-refresh mode */
+        mov.l   1f, r4
+        mov.l   2f, r1
+        mov.l   @r4, r2
+        or      r1, r2
+        mov.l   3f, r3
+        and     r3, r2
+        mov.l   r2, @r4
+
+        rts
+         nop
+
+        .balign 4
+1:      .long   0xfe400008 /* SDCR0 */
+2:      .long   0x00000400
+3:      .long   0xffff7fff
+ENTRY(migor_sdram_enter_end)
+
+        .balign 4
+ENTRY(migor_sdram_leave_start)
+
+        /* SBSC: set auto-refresh mode */
+        mov.l   1f, r4
+        mov.l   @r4, r0
+        mov.l   4f, r1
+        and     r1, r0
+        mov.l   r0, @r4
+        mov.l   6f, r4
+        mov.l   8f, r0
+        mov.l   @r4, r1
+        mov     #-1, r4
+        add     r4, r1
+        or      r1, r0
+        mov.l   7f, r1
+        mov.l   r0, @r1
+
+        rts
+         nop
+
+        .balign 4
+1:      .long   0xfe400008 /* SDCR0 */
+4:      .long   0xfffffbff
+6:      .long   0xfe40001c /* RTCOR */
+7:      .long   0xfe400018 /* RTCNT */
+8:      .long   0xa55a0000
+ENTRY(migor_sdram_leave_end)
diff --git a/arch/sh/boards/mach-migor/setup.c b/arch/sh/boards/mach-migor/setup.c
index 6ed1fd32369e..9099b6da9957 100644
--- a/arch/sh/boards/mach-migor/setup.c
+++ b/arch/sh/boards/mach-migor/setup.c
@@ -11,6 +11,7 @@
 #include <linux/platform_device.h>
 #include <linux/interrupt.h>
 #include <linux/input.h>
+#include <linux/input/sh_keysc.h>
 #include <linux/mtd/physmap.h>
 #include <linux/mtd/nand.h>
 #include <linux/i2c.h>
@@ -18,8 +19,6 @@
 #include <linux/delay.h>
 #include <linux/clk.h>
 #include <linux/gpio.h>
-#include <linux/spi/spi.h>
-#include <linux/spi/spi_gpio.h>
 #include <video/sh_mobile_lcdc.h>
 #include <media/sh_mobile_ceu.h>
 #include <media/ov772x.h>
@@ -27,7 +26,7 @@
 #include <asm/clock.h>
 #include <asm/machvec.h>
 #include <asm/io.h>
-#include <asm/sh_keysc.h>
+#include <asm/suspend.h>
 #include <mach/migor.h>
 #include <cpu/sh7722.h>
 
@@ -390,17 +389,25 @@ static struct platform_device migor_ceu_device = {
         },
 };
 
-struct spi_gpio_platform_data sdcard_cn9_platform_data = {
-        .sck = GPIO_PTD0,
-        .mosi = GPIO_PTD1,
-        .miso = GPIO_PTD2,
-        .num_chipselect = 1,
+static struct resource sdhi_cn9_resources[] = {
+        [0] = {
+                .name   = "SDHI",
+                .start  = 0x04ce0000,
+                .end    = 0x04ce01ff,
+                .flags  = IORESOURCE_MEM,
+        },
+        [1] = {
+                .start  = 101,
+                .flags  = IORESOURCE_IRQ,
+        },
 };
 
-static struct platform_device sdcard_cn9_device = {
-        .name           = "spi_gpio",
-        .dev    = {
-                .platform_data  = &sdcard_cn9_platform_data,
+static struct platform_device sdhi_cn9_device = {
+        .name           = "sh_mobile_sdhi",
+        .num_resources  = ARRAY_SIZE(sdhi_cn9_resources),
+        .resource       = sdhi_cn9_resources,
+        .archdata = {
+                .hwblk_id = HWBLK_SDHI,
         },
 };
 
@@ -467,23 +474,24 @@ static struct platform_device *migor_devices[] __initdata = {
         &migor_ceu_device,
         &migor_nor_flash_device,
         &migor_nand_flash_device,
-        &sdcard_cn9_device,
+        &sdhi_cn9_device,
         &migor_camera[0],
         &migor_camera[1],
 };
 
-static struct spi_board_info migor_spi_devices[] = {
-        {
-                .modalias = "mmc_spi",
-                .max_speed_hz = 5000000,
-                .chip_select = 0,
-                .controller_data = (void *) GPIO_PTD5,
-        },
-};
+extern char migor_sdram_enter_start;
+extern char migor_sdram_enter_end;
+extern char migor_sdram_leave_start;
+extern char migor_sdram_leave_end;
 
 static int __init migor_devices_setup(void)
 {
-
+        /* register board specific self-refresh code */
+        sh_mobile_register_self_refresh(SUSP_SH_STANDBY | SUSP_SH_SF,
+                                        &migor_sdram_enter_start,
+                                        &migor_sdram_enter_end,
+                                        &migor_sdram_leave_start,
+                                        &migor_sdram_leave_end);
 #ifdef CONFIG_PM
         /* Let D11 LED show STATUS0 */
         gpio_request(GPIO_FN_STATUS0, NULL);
@@ -525,6 +533,16 @@ static int __init migor_devices_setup(void)
         gpio_request(GPIO_PTA1, NULL);
         gpio_direction_input(GPIO_PTA1);
 
+        /* SDHI */
+        gpio_request(GPIO_FN_SDHICD, NULL);
+        gpio_request(GPIO_FN_SDHIWP, NULL);
+        gpio_request(GPIO_FN_SDHID3, NULL);
+        gpio_request(GPIO_FN_SDHID2, NULL);
+        gpio_request(GPIO_FN_SDHID1, NULL);
+        gpio_request(GPIO_FN_SDHID0, NULL);
+        gpio_request(GPIO_FN_SDHICMD, NULL);
+        gpio_request(GPIO_FN_SDHICLK, NULL);
+
         /* Touch Panel */
         gpio_request(GPIO_FN_IRQ6, NULL);
 
@@ -612,9 +630,6 @@ static int __init migor_devices_setup(void)
         i2c_register_board_info(0, migor_i2c_devices,
                                 ARRAY_SIZE(migor_i2c_devices));
 
-        spi_register_board_info(migor_spi_devices,
-                                ARRAY_SIZE(migor_spi_devices));
-
         return platform_add_devices(migor_devices, ARRAY_SIZE(migor_devices));
 }
 arch_initcall(migor_devices_setup);
diff --git a/arch/sh/boards/mach-r2d/irq.c b/arch/sh/boards/mach-r2d/irq.c
index c70fecedcac4..78d7b27c80da 100644
--- a/arch/sh/boards/mach-r2d/irq.c
+++ b/arch/sh/boards/mach-r2d/irq.c
@@ -116,7 +116,7 @@ static unsigned char irl2irq[R2D_NR_IRL];
 
 int rts7751r2d_irq_demux(int irq)
 {
-        if (irq >= R2D_NR_IRL || !irl2irq[irq])
+        if (irq >= R2D_NR_IRL || irq < 0 || !irl2irq[irq])
                 return irq;
 
         return irl2irq[irq];
diff --git a/arch/sh/boards/mach-se/7722/irq.c b/arch/sh/boards/mach-se/7722/irq.c
index 02d21a3e2a8f..4eb31acfafef 100644
--- a/arch/sh/boards/mach-se/7722/irq.c
+++ b/arch/sh/boards/mach-se/7722/irq.c
@@ -16,15 +16,17 @@
 #include <asm/io.h>
 #include <mach-se/mach/se7722.h>
 
+unsigned int se7722_fpga_irq[SE7722_FPGA_IRQ_NR] = { 0, };
+
 static void disable_se7722_irq(unsigned int irq)
 {
-        unsigned int bit = irq - SE7722_FPGA_IRQ_BASE;
+        unsigned int bit = (unsigned int)get_irq_chip_data(irq);
         ctrl_outw(ctrl_inw(IRQ01_MASK) | 1 << bit, IRQ01_MASK);
 }
 
 static void enable_se7722_irq(unsigned int irq)
 {
-        unsigned int bit = irq - SE7722_FPGA_IRQ_BASE;
+        unsigned int bit = (unsigned int)get_irq_chip_data(irq);
         ctrl_outw(ctrl_inw(IRQ01_MASK) & ~(1 << bit), IRQ01_MASK);
 }
 
@@ -38,18 +40,15 @@ static struct irq_chip se7722_irq_chip __read_mostly = {
 static void se7722_irq_demux(unsigned int irq, struct irq_desc *desc)
 {
         unsigned short intv = ctrl_inw(IRQ01_STS);
-        struct irq_desc *ext_desc;
-        unsigned int ext_irq = SE7722_FPGA_IRQ_BASE;
+        unsigned int ext_irq = 0;
 
         intv &= (1 << SE7722_FPGA_IRQ_NR) - 1;
 
-        while (intv) {
-                if (intv & 1) {
-                        ext_desc = irq_desc + ext_irq;
-                        handle_level_irq(ext_irq, ext_desc);
-                }
-                intv >>= 1;
-                ext_irq++;
+        for (; intv; intv >>= 1, ext_irq++) {
+                if (!(intv & 1))
+                        continue;
+
+                generic_handle_irq(se7722_fpga_irq[ext_irq]);
         }
 }
 
@@ -63,11 +62,18 @@ void __init init_se7722_IRQ(void)
         ctrl_outw(0, IRQ01_MASK);       /* disable all irqs */
         ctrl_outw(0x2000, 0xb03fffec);  /* mrshpc irq enable */
 
-        for (i = 0; i < SE7722_FPGA_IRQ_NR; i++)
-                set_irq_chip_and_handler_name(SE7722_FPGA_IRQ_BASE + i,
+        for (i = 0; i < SE7722_FPGA_IRQ_NR; i++) {
+                se7722_fpga_irq[i] = create_irq();
+                if (se7722_fpga_irq[i] < 0)
+                        return;
+
+                set_irq_chip_and_handler_name(se7722_fpga_irq[i],
                                               &se7722_irq_chip,
                                               handle_level_irq, "level");
 
+                set_irq_chip_data(se7722_fpga_irq[i], (void *)i);
+        }
+
         set_irq_chained_handler(IRQ0_IRQ, se7722_irq_demux);
         set_irq_type(IRQ0_IRQ, IRQ_TYPE_LEVEL_LOW);
 
diff --git a/arch/sh/boards/mach-se/7722/setup.c b/arch/sh/boards/mach-se/7722/setup.c
index 36374078e521..b1cb9425b600 100644
--- a/arch/sh/boards/mach-se/7722/setup.c
+++ b/arch/sh/boards/mach-se/7722/setup.c
@@ -14,6 +14,7 @@
 #include <linux/platform_device.h>
 #include <linux/ata_platform.h>
 #include <linux/input.h>
+#include <linux/input/sh_keysc.h>
 #include <linux/smc91x.h>
 #include <mach-se/mach/se7722.h>
 #include <mach-se/mach/mrshpc.h>
@@ -21,7 +22,6 @@
 #include <asm/clock.h>
 #include <asm/io.h>
 #include <asm/heartbeat.h>
-#include <asm/sh_keysc.h>
 #include <cpu/sh7722.h>
 
 /* Heartbeat */
@@ -60,8 +60,7 @@ static struct resource smc91x_eth_resources[] = {
                 .flags  = IORESOURCE_MEM,
         },
         [1] = {
-                .start  = SMC_IRQ,
-                .end    = SMC_IRQ,
+                /* Filled in later */
                 .flags  = IORESOURCE_IRQ,
         },
 };
@@ -90,8 +89,7 @@ static struct resource cf_ide_resources[] = {
                 .flags  = IORESOURCE_IO,
         },
         [2] = {
-                .start  = MRSHPC_IRQ0,
-                .end    = MRSHPC_IRQ0,
+                /* Filled in later */
                 .flags  = IORESOURCE_IRQ,
         },
 };
@@ -153,6 +151,14 @@ static struct platform_device *se7722_devices[] __initdata = {
 static int __init se7722_devices_setup(void)
 {
         mrshpc_setup_windows();
+
+        /* Wire-up dynamic vectors */
+        cf_ide_resources[2].start = cf_ide_resources[2].end =
+                se7722_fpga_irq[SE7722_FPGA_IRQ_MRSHPC0];
+
+        smc91x_eth_resources[1].start = smc91x_eth_resources[1].end =
+                se7722_fpga_irq[SE7722_FPGA_IRQ_SMC];
+
         return platform_add_devices(se7722_devices, ARRAY_SIZE(se7722_devices));
 }
 device_initcall(se7722_devices_setup);
@@ -193,6 +199,5 @@ static void __init se7722_setup(char **cmdline_p)
 static struct sh_machine_vector mv_se7722 __initmv = {
         .mv_name                = "Solution Engine 7722" ,
         .mv_setup               = se7722_setup ,
-        .mv_nr_irqs             = SE7722_FPGA_IRQ_BASE + SE7722_FPGA_IRQ_NR,
         .mv_init_irq            = init_se7722_IRQ,
 };
diff --git a/arch/sh/boards/mach-se/7724/Makefile b/arch/sh/boards/mach-se/7724/Makefile
index 349cbd6ce82d..a08b36830f0e 100644
--- a/arch/sh/boards/mach-se/7724/Makefile
+++ b/arch/sh/boards/mach-se/7724/Makefile
@@ -7,4 +7,4 @@
 #
 #
 
-obj-y    := setup.o irq.o
\ No newline at end of file
+obj-y    := setup.o irq.o sdram.o
diff --git a/arch/sh/boards/mach-se/7724/sdram.S b/arch/sh/boards/mach-se/7724/sdram.S
new file mode 100644
index 000000000000..9040167d5022
--- /dev/null
+++ b/arch/sh/boards/mach-se/7724/sdram.S
@@ -0,0 +1,52 @@
+/*
+ * MS7724SE sdram self/auto-refresh setup code
+ *
+ *  Copyright (C) 2009 Magnus Damm
+ *
+ * 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/sys.h>
+#include <linux/errno.h>
+#include <linux/linkage.h>
+#include <asm/asm-offsets.h>
+#include <asm/suspend.h>
+#include <asm/romimage-macros.h>
+
+/* code to enter and leave self-refresh. must be self-contained.
+ * this code will be copied to on-chip memory and executed from there.
+ */
+        .balign 4
+ENTRY(ms7724se_sdram_enter_start)
+
+        /* DBSC: put memory in self-refresh mode */
+
+        ED 0xFD000010, 0x00000000 /* DBEN */
+        ED 0xFD000040, 0x00000000 /* DBRFPDN0 */
+        ED 0xFD000014, 0x00000002 /* DBCMDCNT (PALL) */
+        ED 0xFD000014, 0x00000004 /* DBCMDCNT (REF) */
+        ED 0xFD000040, 0x00000001 /* DBRFPDN0 */
+
+        rts
+         nop
+
+ENTRY(ms7724se_sdram_enter_end)
+
+        .balign 4
+ENTRY(ms7724se_sdram_leave_start)
+
+        /* DBSC: put memory in auto-refresh mode */
+
+        ED 0xFD000040, 0x00000000 /* DBRFPDN0 */
+        WAIT 1
+        ED 0xFD000014, 0x00000002 /* DBCMDCNT (PALL) */
+        ED 0xFD000014, 0x00000004 /* DBCMDCNT (REF) */
+        ED 0xFD000010, 0x00000001 /* DBEN */
+        ED 0xFD000040, 0x00010000 /* DBRFPDN0 */
+
+        rts
+         nop
+
+ENTRY(ms7724se_sdram_leave_end)
diff --git a/arch/sh/boards/mach-se/7724/setup.c b/arch/sh/boards/mach-se/7724/setup.c
index e78c3be8ad2f..da01fc0dc881 100644
--- a/arch/sh/boards/mach-se/7724/setup.c
+++ b/arch/sh/boards/mach-se/7724/setup.c
@@ -19,6 +19,7 @@
 #include <linux/smc91x.h>
 #include <linux/gpio.h>
 #include <linux/input.h>
+#include <linux/input/sh_keysc.h>
 #include <linux/usb/r8a66597.h>
 #include <video/sh_mobile_lcdc.h>
 #include <media/sh_mobile_ceu.h>
@@ -27,7 +28,7 @@
 #include <asm/heartbeat.h>
 #include <asm/sh_eth.h>
 #include <asm/clock.h>
-#include <asm/sh_keysc.h>
+#include <asm/suspend.h>
 #include <cpu/sh7724.h>
 #include <mach-se/mach/se7724.h>
 
@@ -448,6 +449,52 @@ static struct platform_device sh7724_usb1_gadget_device = {
         .resource       = sh7724_usb1_gadget_resources,
 };
 
+static struct resource sdhi0_cn7_resources[] = {
+        [0] = {
+                .name   = "SDHI0",
+                .start  = 0x04ce0000,
+                .end    = 0x04ce01ff,
+                .flags  = IORESOURCE_MEM,
+        },
+        [1] = {
+                .start  = 101,
+                .flags  = IORESOURCE_IRQ,
+        },
+};
+
+static struct platform_device sdhi0_cn7_device = {
+        .name           = "sh_mobile_sdhi",
+        .id             = 0,
+        .num_resources  = ARRAY_SIZE(sdhi0_cn7_resources),
+        .resource       = sdhi0_cn7_resources,
+        .archdata = {
+                .hwblk_id = HWBLK_SDHI0,
+        },
+};
+
+static struct resource sdhi1_cn8_resources[] = {
+        [0] = {
+                .name   = "SDHI1",
+                .start  = 0x04cf0000,
+                .end    = 0x04cf01ff,
+                .flags  = IORESOURCE_MEM,
+        },
+        [1] = {
+                .start  = 24,
+                .flags  = IORESOURCE_IRQ,
+        },
+};
+
+static struct platform_device sdhi1_cn8_device = {
+        .name           = "sh_mobile_sdhi",
+        .id             = 1,
+        .num_resources  = ARRAY_SIZE(sdhi1_cn8_resources),
+        .resource       = sdhi1_cn8_resources,
+        .archdata = {
+                .hwblk_id = HWBLK_SDHI1,
+        },
+};
+
 static struct platform_device *ms7724se_devices[] __initdata = {
         &heartbeat_device,
         &smc91x_eth_device,
@@ -460,6 +507,8 @@ static struct platform_device *ms7724se_devices[] __initdata = {
         &sh7724_usb0_host_device,
         &sh7724_usb1_gadget_device,
         &fsi_device,
+        &sdhi0_cn7_device,
+        &sdhi1_cn8_device,
 };
 
 #define EEPROM_OP   0xBA206000
@@ -484,7 +533,7 @@ static int __init sh_eth_is_eeprom_ready(void)
 static void __init sh_eth_init(void)
 {
         int i;
-        u16 mac[3];
+        u16 mac;
 
         /* check EEPROM status */
         if (!sh_eth_is_eeprom_ready())
@@ -498,16 +547,10 @@ static void __init sh_eth_init(void)
                 if (!sh_eth_is_eeprom_ready())
                         return;
 
-                mac[i] = ctrl_inw(EEPROM_DATA);
-                mac[i] = ((mac[i] & 0xFF) << 8) | (mac[i] >> 8); /* swap */
+                mac = ctrl_inw(EEPROM_DATA);
+                sh_eth_plat.mac_addr[i << 1] = mac & 0xff;
+                sh_eth_plat.mac_addr[(i << 1) + 1] = mac >> 8;
         }
-
-        /* reset sh-eth */
-        ctrl_outl(0x1, SH_ETH_ADDR + 0x0);
-
-        /* set MAC addr */
-        ctrl_outl(((mac[0] << 16) | (mac[1])), SH_ETH_MAHR);
-        ctrl_outl((mac[2]), SH_ETH_MALR);
 }
 
 #define SW4140    0xBA201000
@@ -524,11 +567,22 @@ static void __init sh_eth_init(void)
 #define SW41_G    0x4000
 #define SW41_H    0x8000
 
+extern char ms7724se_sdram_enter_start;
+extern char ms7724se_sdram_enter_end;
+extern char ms7724se_sdram_leave_start;
+extern char ms7724se_sdram_leave_end;
+
 static int __init devices_setup(void)
 {
         u16 sw = ctrl_inw(SW4140); /* select camera, monitor */
         struct clk *fsia_clk;
 
+        /* register board specific self-refresh code */
+        sh_mobile_register_self_refresh(SUSP_SH_STANDBY | SUSP_SH_SF,
+                                        &ms7724se_sdram_enter_start,
+                                        &ms7724se_sdram_enter_end,
+                                        &ms7724se_sdram_leave_start,
+                                        &ms7724se_sdram_leave_end);
         /* Reset Release */
         ctrl_outw(ctrl_inw(FPGA_OUT) &
                   ~((1 << 1)  | /* LAN */
@@ -698,6 +752,26 @@ static int __init devices_setup(void)
         clk_set_rate(&fsimcka_clk, 11000);
         clk_put(fsia_clk);
 
+        /* SDHI0 connected to cn7 */
+        gpio_request(GPIO_FN_SDHI0CD, NULL);
+        gpio_request(GPIO_FN_SDHI0WP, NULL);
+        gpio_request(GPIO_FN_SDHI0D3, NULL);
+        gpio_request(GPIO_FN_SDHI0D2, NULL);
+        gpio_request(GPIO_FN_SDHI0D1, NULL);
+        gpio_request(GPIO_FN_SDHI0D0, NULL);
+        gpio_request(GPIO_FN_SDHI0CMD, NULL);
+        gpio_request(GPIO_FN_SDHI0CLK, NULL);
+
+        /* SDHI1 connected to cn8 */
+        gpio_request(GPIO_FN_SDHI1CD, NULL);
+        gpio_request(GPIO_FN_SDHI1WP, NULL);
+        gpio_request(GPIO_FN_SDHI1D3, NULL);
+        gpio_request(GPIO_FN_SDHI1D2, NULL);
+        gpio_request(GPIO_FN_SDHI1D1, NULL);
+        gpio_request(GPIO_FN_SDHI1D0, NULL);
+        gpio_request(GPIO_FN_SDHI1CMD, NULL);
+        gpio_request(GPIO_FN_SDHI1CLK, NULL);
+
         /*
          * enable SH-Eth
          *
diff --git a/arch/sh/boot/compressed/misc.c b/arch/sh/boot/compressed/misc.c
index fd56a71ca9d9..b51b1fc4baae 100644
--- a/arch/sh/boot/compressed/misc.c
+++ b/arch/sh/boot/compressed/misc.c
@@ -131,7 +131,7 @@ void decompress_kernel(void)
 #ifdef CONFIG_SUPERH64
         output_addr = (CONFIG_MEMORY_START + 0x2000);
 #else
-        output_addr = PHYSADDR((unsigned long)&_text+PAGE_SIZE);
+        output_addr = __pa((unsigned long)&_text+PAGE_SIZE);
 #ifdef CONFIG_29BIT
         output_addr |= P2SEG;
 #endif
diff --git a/arch/sh/boot/romimage/Makefile b/arch/sh/boot/romimage/Makefile
index 5806eee84f6f..f473a24a2d92 100644
--- a/arch/sh/boot/romimage/Makefile
+++ b/arch/sh/boot/romimage/Makefile
@@ -4,16 +4,22 @@
 # create an image suitable for burning to flash from zImage
 #
 
-targets         := vmlinux head.o
+targets         := vmlinux head.o zeropage.bin piggy.o
 
 OBJECTS = $(obj)/head.o
-LDFLAGS_vmlinux := --oformat $(ld-bfd) -Ttext 0 -e romstart
+LDFLAGS_vmlinux := --oformat $(ld-bfd) -Ttext 0 -e romstart \
+                   -T $(obj)/../../kernel/vmlinux.lds
 
 $(obj)/vmlinux: $(OBJECTS) $(obj)/piggy.o FORCE
         $(call if_changed,ld)
         @:
 
+OBJCOPYFLAGS += -j .empty_zero_page
+
+$(obj)/zeropage.bin: vmlinux FORCE
+        $(call if_changed,objcopy)
+
 LDFLAGS_piggy.o := -r --format binary --oformat $(ld-bfd) -T
 
-$(obj)/piggy.o: $(obj)/vmlinux.scr arch/sh/boot/zImage FORCE
+$(obj)/piggy.o: $(obj)/vmlinux.scr $(obj)/zeropage.bin arch/sh/boot/zImage FORCE
         $(call if_changed,ld)
diff --git a/arch/sh/boot/romimage/head.S b/arch/sh/boot/romimage/head.S
index 219bc626dd71..93e779a405ec 100644
--- a/arch/sh/boot/romimage/head.S
+++ b/arch/sh/boot/romimage/head.S
@@ -5,6 +5,44 @@
  */
 
 .text
+        #include <asm/page.h>
+
         .global romstart
 romstart:
+        /* include board specific setup code */
 #include <mach/romimage.h>
+
+        /* copy the empty_zero_page contents to where vmlinux expects it */
+        mova    empty_zero_page_src, r0
+        mov.l   empty_zero_page_dst, r1
+        mov     #(PAGE_SHIFT - 4), r4
+        mov     #1, r3
+        shld    r4, r3 /* r3 = PAGE_SIZE / 16 */
+
+1:
+        mov.l   @r0, r4
+        mov.l   @(4, r0), r5
+        mov.l   @(8, r0), r6
+        mov.l   @(12, r0), r7
+        add     #16,r0
+        mov.l   r4, @r1
+        mov.l   r5, @(4, r1)
+        mov.l   r6, @(8, r1)
+        mov.l   r7, @(12, r1)
+        dt      r3
+        add     #16,r1
+        bf      1b
+
+        /* jump to the zImage entry point located after the zero page data */
+        mov     #PAGE_SHIFT, r4
+        mov     #1, r1
+        shld    r4, r1
+        mova    empty_zero_page_src, r0
+        add     r1, r0
+        jmp     @r0
+         nop
+
+        .align 2
+empty_zero_page_dst:
+        .long   _text
+empty_zero_page_src:
diff --git a/arch/sh/drivers/dma/dma-sysfs.c b/arch/sh/drivers/dma/dma-sysfs.c
index 347ee11351ec..1ee631d3725e 100644
--- a/arch/sh/drivers/dma/dma-sysfs.c
+++ b/arch/sh/drivers/dma/dma-sysfs.c
@@ -13,7 +13,6 @@
 #include <linux/init.h>
 #include <linux/sysdev.h>
 #include <linux/platform_device.h>
-#include <linux/module.h>
 #include <linux/err.h>
 #include <linux/string.h>
 #include <asm/dma.h>
@@ -21,7 +20,6 @@
 static struct sysdev_class dma_sysclass = {
         .name = "dma",
 };
-EXPORT_SYMBOL(dma_sysclass);
 
 static ssize_t dma_show_devices(struct sys_device *dev,
                                 struct sysdev_attribute *attr, char *buf)
diff --git a/arch/sh/drivers/pci/Kconfig b/arch/sh/drivers/pci/Kconfig
deleted file mode 100644
index e8db585a6638..000000000000
--- a/arch/sh/drivers/pci/Kconfig
+++ /dev/null
@@ -1,19 +0,0 @@
-config PCI
-        bool "PCI support"
-        depends on SYS_SUPPORTS_PCI
-        help
-          Find out whether you have a PCI motherboard. PCI is the name of a
-          bus system, i.e. the way the CPU talks to the other stuff inside
-          your box. If you have PCI, say Y, otherwise N.
-
-config SH_PCIDMA_NONCOHERENT
-        bool "Cache and PCI noncoherent"
-        depends on PCI
-        default y
-        help
-          Enable this option if your platform does not have a CPU cache which
-          remains coherent with PCI DMA. It is safest to say 'Y', although you
-          will see better performance if you can say 'N', because the PCI DMA
-          code will not have to flush the CPU's caches. If you have a PCI host
-          bridge integrated with your SH CPU, refer carefully to the chip specs
-          to see if you can say 'N' here. Otherwise, leave it as 'Y'.
diff --git a/arch/sh/include/asm/addrspace.h b/arch/sh/include/asm/addrspace.h
index 80d40813e057..99d6b3ecbe22 100644
--- a/arch/sh/include/asm/addrspace.h
+++ b/arch/sh/include/asm/addrspace.h
@@ -28,9 +28,6 @@
 /* Returns the privileged segment base of a given address  */
 #define PXSEG(a)        (((unsigned long)(a)) & 0xe0000000)
 
-/* Returns the physical address of a PnSEG (n=1,2) address   */
-#define PHYSADDR(a)     (((unsigned long)(a)) & 0x1fffffff)
-
 #if defined(CONFIG_29BIT) || defined(CONFIG_PMB_FIXED)
 /*
  * Map an address to a certain privileged segment
@@ -60,5 +57,11 @@
 #define P3_ADDR_MAX             P4SEG
 #endif
 
+#ifndef __ASSEMBLY__
+#ifdef CONFIG_PMB
+extern int __in_29bit_mode(void);
+#endif /* CONFIG_PMB */
+#endif /* __ASSEMBLY__ */
+
 #endif /* __KERNEL__ */
 #endif /* __ASM_SH_ADDRSPACE_H */
diff --git a/arch/sh/include/asm/atomic.h b/arch/sh/include/asm/atomic.h
index e8e78137c6f5..b16388d71954 100644
--- a/arch/sh/include/asm/atomic.h
+++ b/arch/sh/include/asm/atomic.h
@@ -78,11 +78,10 @@ static inline int atomic_add_unless(atomic_t *v, int a, int u)
 #define atomic_xchg(v, new) (xchg(&((v)->counter), new))
 #define atomic_inc_not_zero(v) atomic_add_unless((v), 1, 0)
 
-/* Atomic operations are already serializing on SH */
-#define smp_mb__before_atomic_dec()     barrier()
-#define smp_mb__after_atomic_dec()      barrier()
-#define smp_mb__before_atomic_inc()     barrier()
-#define smp_mb__after_atomic_inc()      barrier()
+#define smp_mb__before_atomic_dec()     smp_mb()
+#define smp_mb__after_atomic_dec()      smp_mb()
+#define smp_mb__before_atomic_inc()     smp_mb()
+#define smp_mb__after_atomic_inc()      smp_mb()
 
 #include <asm-generic/atomic-long.h>
 #include <asm-generic/atomic64.h>
diff --git a/arch/sh/include/asm/bitops.h b/arch/sh/include/asm/bitops.h
index ebe595b7ab1f..98511e4d28cb 100644
--- a/arch/sh/include/asm/bitops.h
+++ b/arch/sh/include/asm/bitops.h
@@ -26,8 +26,8 @@
 /*
  * clear_bit() doesn't provide any barrier for the compiler.
  */
-#define smp_mb__before_clear_bit()      barrier()
-#define smp_mb__after_clear_bit()       barrier()
+#define smp_mb__before_clear_bit()      smp_mb()
+#define smp_mb__after_clear_bit()       smp_mb()
 
 #ifdef CONFIG_SUPERH32
 static inline unsigned long ffz(unsigned long word)
diff --git a/arch/sh/include/asm/bugs.h b/arch/sh/include/asm/bugs.h
index 46260fcbdf4b..02a19a1c033a 100644
--- a/arch/sh/include/asm/bugs.h
+++ b/arch/sh/include/asm/bugs.h
@@ -14,11 +14,15 @@
 
 #include <asm/processor.h>
 
+extern void select_idle_routine(void);
+
 static void __init check_bugs(void)
 {
         extern unsigned long loops_per_jiffy;
         char *p = &init_utsname()->machine[2]; /* "sh" */
 
+        select_idle_routine();
+
         current_cpu_data.loops_per_jiffy = loops_per_jiffy;
 
         switch (current_cpu_data.family) {
diff --git a/arch/sh/include/asm/dma-mapping.h b/arch/sh/include/asm/dma-mapping.h
index 69d56dd4c968..87ced133a363 100644
--- a/arch/sh/include/asm/dma-mapping.h
+++ b/arch/sh/include/asm/dma-mapping.h
@@ -1,219 +1,108 @@
 #ifndef __ASM_SH_DMA_MAPPING_H
 #define __ASM_SH_DMA_MAPPING_H
 
-#include <linux/mm.h>
-#include <linux/scatterlist.h>
-#include <linux/dma-debug.h>
-#include <asm/cacheflush.h>
-#include <asm/io.h>
+extern struct dma_map_ops *dma_ops;
+extern void no_iommu_init(void);
+
+static inline struct dma_map_ops *get_dma_ops(struct device *dev)
+{
+        return dma_ops;
+}
+
 #include <asm-generic/dma-coherent.h>
+#include <asm-generic/dma-mapping-common.h>
+
+static inline int dma_supported(struct device *dev, u64 mask)
+{
+        struct dma_map_ops *ops = get_dma_ops(dev);
 
-extern struct bus_type pci_bus_type;
+        if (ops->dma_supported)
+                return ops->dma_supported(dev, mask);
 
-#define dma_supported(dev, mask)        (1)
+        return 1;
+}
 
 static inline int dma_set_mask(struct device *dev, u64 mask)
 {
+        struct dma_map_ops *ops = get_dma_ops(dev);
+
         if (!dev->dma_mask || !dma_supported(dev, mask))
                 return -EIO;
+        if (ops->set_dma_mask)
+                return ops->set_dma_mask(dev, mask);
 
         *dev->dma_mask = mask;
 
         return 0;
 }
 
-void *dma_alloc_coherent(struct device *dev, size_t size,
-                         dma_addr_t *dma_handle, gfp_t flag);
-
-void dma_free_coherent(struct device *dev, size_t size,
-                       void *vaddr, dma_addr_t dma_handle);
-
 void dma_cache_sync(struct device *dev, void *vaddr, size_t size,
                     enum dma_data_direction dir);
 
 #define dma_alloc_noncoherent(d, s, h, f) dma_alloc_coherent(d, s, h, f)
 #define dma_free_noncoherent(d, s, v, h) dma_free_coherent(d, s, v, h)
-#define dma_is_consistent(d, h) (1)
-
-static inline dma_addr_t dma_map_single(struct device *dev,
-                                        void *ptr, size_t size,
-                                        enum dma_data_direction dir)
-{
-        dma_addr_t addr = virt_to_phys(ptr);
-
-#if defined(CONFIG_PCI) && !defined(CONFIG_SH_PCIDMA_NONCOHERENT)
-        if (dev->bus == &pci_bus_type)
-                return addr;
-#endif
-        dma_cache_sync(dev, ptr, size, dir);
-
-        debug_dma_map_page(dev, virt_to_page(ptr),
-                           (unsigned long)ptr & ~PAGE_MASK, size,
-                           dir, addr, true);
-
-        return addr;
-}
-
-static inline void dma_unmap_single(struct device *dev, dma_addr_t addr,
-                                    size_t size, enum dma_data_direction dir)
-{
-        debug_dma_unmap_page(dev, addr, size, dir, true);
-}
 
-static inline int dma_map_sg(struct device *dev, struct scatterlist *sg,
-                             int nents, enum dma_data_direction dir)
-{
-        int i;
-
-        for (i = 0; i < nents; i++) {
-#if !defined(CONFIG_PCI) || defined(CONFIG_SH_PCIDMA_NONCOHERENT)
-                dma_cache_sync(dev, sg_virt(&sg[i]), sg[i].length, dir);
+#ifdef CONFIG_DMA_COHERENT
+#define dma_is_consistent(d, h) (1)
+#else
+#define dma_is_consistent(d, h) (0)
 #endif
-                sg[i].dma_address = sg_phys(&sg[i]);
-                sg[i].dma_length = sg[i].length;
-        }
 
-        debug_dma_map_sg(dev, sg, nents, i, dir);
-
-        return nents;
-}
-
-static inline void dma_unmap_sg(struct device *dev, struct scatterlist *sg,
-                                int nents, enum dma_data_direction dir)
-{
-        debug_dma_unmap_sg(dev, sg, nents, dir);
-}
-
-static inline dma_addr_t dma_map_page(struct device *dev, struct page *page,
-                                      unsigned long offset, size_t size,
-                                      enum dma_data_direction dir)
-{
-        return dma_map_single(dev, page_address(page) + offset, size, dir);
-}
-
-static inline void dma_unmap_page(struct device *dev, dma_addr_t dma_address,
-                                  size_t size, enum dma_data_direction dir)
-{
-        dma_unmap_single(dev, dma_address, size, dir);
-}
-
-static inline void __dma_sync_single(struct device *dev, dma_addr_t dma_handle,
-                                   size_t size, enum dma_data_direction dir)
+static inline int dma_get_cache_alignment(void)
 {
-#if defined(CONFIG_PCI) && !defined(CONFIG_SH_PCIDMA_NONCOHERENT)
-        if (dev->bus == &pci_bus_type)
-                return;
-#endif
-        dma_cache_sync(dev, phys_to_virt(dma_handle), size, dir);
+        /*
+         * Each processor family will define its own L1_CACHE_SHIFT,
+         * L1_CACHE_BYTES wraps to this, so this is always safe.
+         */
+        return L1_CACHE_BYTES;
 }
 
-static inline void dma_sync_single_range(struct device *dev,
-                                         dma_addr_t dma_handle,
-                                         unsigned long offset, size_t size,
-                                         enum dma_data_direction dir)
+static inline int dma_mapping_error(struct device *dev, dma_addr_t dma_addr)
 {
-#if defined(CONFIG_PCI) && !defined(CONFIG_SH_PCIDMA_NONCOHERENT)
-        if (dev->bus == &pci_bus_type)
-                return;
-#endif
-        dma_cache_sync(dev, phys_to_virt(dma_handle) + offset, size, dir);
-}
+        struct dma_map_ops *ops = get_dma_ops(dev);
 
-static inline void __dma_sync_sg(struct device *dev, struct scatterlist *sg,
-                               int nelems, enum dma_data_direction dir)
-{
-        int i;
+        if (ops->mapping_error)
+                return ops->mapping_error(dev, dma_addr);
 
-        for (i = 0; i < nelems; i++) {
-#if !defined(CONFIG_PCI) || defined(CONFIG_SH_PCIDMA_NONCOHERENT)
-                dma_cache_sync(dev, sg_virt(&sg[i]), sg[i].length, dir);
-#endif
-                sg[i].dma_address = sg_phys(&sg[i]);
-                sg[i].dma_length = sg[i].length;
-        }
+        return dma_addr == 0;
 }
 
-static inline void dma_sync_single_for_cpu(struct device *dev,
-                                           dma_addr_t dma_handle, size_t size,
-                                           enum dma_data_direction dir)
+static inline void *dma_alloc_coherent(struct device *dev, size_t size,
+                                       dma_addr_t *dma_handle, gfp_t gfp)
 {
-        __dma_sync_single(dev, dma_handle, size, dir);
-        debug_dma_sync_single_for_cpu(dev, dma_handle, size, dir);
-}
+        struct dma_map_ops *ops = get_dma_ops(dev);
+        void *memory;
 
-static inline void dma_sync_single_for_device(struct device *dev,
-                                              dma_addr_t dma_handle,
-                                              size_t size,
-                                              enum dma_data_direction dir)
-{
-        __dma_sync_single(dev, dma_handle, size, dir);
-        debug_dma_sync_single_for_device(dev, dma_handle, size, dir);
-}
+        if (dma_alloc_from_coherent(dev, size, dma_handle, &memory))
+                return memory;
+        if (!ops->alloc_coherent)
+                return NULL;
 
-static inline void dma_sync_single_range_for_cpu(struct device *dev,
-                                                 dma_addr_t dma_handle,
-                                                 unsigned long offset,
-                                                 size_t size,
-                                                 enum dma_data_direction direction)
-{
-        dma_sync_single_for_cpu(dev, dma_handle+offset, size, direction);
-        debug_dma_sync_single_range_for_cpu(dev, dma_handle,
-                                            offset, size, direction);
-}
+        memory = ops->alloc_coherent(dev, size, dma_handle, gfp);
+        debug_dma_alloc_coherent(dev, size, *dma_handle, memory);
 
-static inline void dma_sync_single_range_for_device(struct device *dev,
-                                                    dma_addr_t dma_handle,
-                                                    unsigned long offset,
-                                                    size_t size,
-                                                    enum dma_data_direction direction)
-{
-        dma_sync_single_for_device(dev, dma_handle+offset, size, direction);
-        debug_dma_sync_single_range_for_device(dev, dma_handle,
-                                               offset, size, direction);
+        return memory;
 }
 
-
-static inline void dma_sync_sg_for_cpu(struct device *dev,
-                                       struct scatterlist *sg, int nelems,
-                                       enum dma_data_direction dir)
+static inline void dma_free_coherent(struct device *dev, size_t size,
+                                     void *vaddr, dma_addr_t dma_handle)
 {
-        __dma_sync_sg(dev, sg, nelems, dir);
-        debug_dma_sync_sg_for_cpu(dev, sg, nelems, dir);
-}
+        struct dma_map_ops *ops = get_dma_ops(dev);
 
-static inline void dma_sync_sg_for_device(struct device *dev,
-                                          struct scatterlist *sg, int nelems,
-                                          enum dma_data_direction dir)
-{
-        __dma_sync_sg(dev, sg, nelems, dir);
-        debug_dma_sync_sg_for_device(dev, sg, nelems, dir);
-}
+        WARN_ON(irqs_disabled());       /* for portability */
 
-static inline int dma_get_cache_alignment(void)
-{
-        /*
-         * Each processor family will define its own L1_CACHE_SHIFT,
-         * L1_CACHE_BYTES wraps to this, so this is always safe.
-         */
-        return L1_CACHE_BYTES;
-}
+        if (dma_release_from_coherent(dev, get_order(size), vaddr))
+                return;
 
-static inline int dma_mapping_error(struct device *dev, dma_addr_t dma_addr)
-{
-        return dma_addr == 0;
+        debug_dma_free_coherent(dev, size, vaddr, dma_handle);
+        if (ops->free_coherent)
+                ops->free_coherent(dev, size, vaddr, dma_handle);
 }
 
-#define ARCH_HAS_DMA_DECLARE_COHERENT_MEMORY
-
-extern int
-dma_declare_coherent_memory(struct device *dev, dma_addr_t bus_addr,
-                            dma_addr_t device_addr, size_t size, int flags);
-
-extern void
-dma_release_declared_memory(struct device *dev);
-
-extern void *
-dma_mark_declared_memory_occupied(struct device *dev,
-                                  dma_addr_t device_addr, size_t size);
+/* arch/sh/mm/consistent.c */
+extern void *dma_generic_alloc_coherent(struct device *dev, size_t size,
+                                        dma_addr_t *dma_addr, gfp_t flag);
+extern void dma_generic_free_coherent(struct device *dev, size_t size,
+                                      void *vaddr, dma_addr_t dma_handle);
 
 #endif /* __ASM_SH_DMA_MAPPING_H */
diff --git a/arch/sh/include/asm/dwarf.h b/arch/sh/include/asm/dwarf.h
index ced6795891a6..bdccbbfdc0bd 100644
--- a/arch/sh/include/asm/dwarf.h
+++ b/arch/sh/include/asm/dwarf.h
@@ -194,6 +194,12 @@
 #define DWARF_ARCH_RA_REG       17
 
 #ifndef __ASSEMBLY__
+
+#include <linux/compiler.h>
+#include <linux/bug.h>
+#include <linux/list.h>
+#include <linux/module.h>
+
 /*
  * Read either the frame pointer (r14) or the stack pointer (r15).
  * NOTE: this MUST be inlined.
@@ -241,6 +247,12 @@ struct dwarf_cie {
 
         unsigned long flags;
 #define DWARF_CIE_Z_AUGMENTATION        (1 << 0)
+
+        /*
+         * 'mod' will be non-NULL if this CIE came from a module's
+         * .eh_frame section.
+         */
+        struct module *mod;
 };
 
 /**
@@ -255,6 +267,12 @@ struct dwarf_fde {
         unsigned char *instructions;
         unsigned char *end;
         struct list_head link;
+
+        /*
+         * 'mod' will be non-NULL if this FDE came from a module's
+         * .eh_frame section.
+         */
+        struct module *mod;
 };
 
 /**
@@ -364,6 +382,12 @@ static inline unsigned int DW_CFA_operand(unsigned long insn)
 
 extern struct dwarf_frame *dwarf_unwind_stack(unsigned long,
                                               struct dwarf_frame *);
+extern void dwarf_free_frame(struct dwarf_frame *);
+
+extern int module_dwarf_finalize(const Elf_Ehdr *, const Elf_Shdr *,
+                                 struct module *);
+extern void module_dwarf_cleanup(struct module *);
+
 #endif /* !__ASSEMBLY__ */
 
 #define CFI_STARTPROC   .cfi_startproc
@@ -391,6 +415,10 @@ extern struct dwarf_frame *dwarf_unwind_stack(unsigned long,
 static inline void dwarf_unwinder_init(void)
 {
 }
+
+#define module_dwarf_finalize(hdr, sechdrs, me) (0)
+#define module_dwarf_cleanup(mod)               do { } while (0)
+
 #endif
 
 #endif /* CONFIG_DWARF_UNWINDER */
diff --git a/arch/sh/include/asm/fixmap.h b/arch/sh/include/asm/fixmap.h
index 721fcc4d5e98..5ac1e40a511c 100644
--- a/arch/sh/include/asm/fixmap.h
+++ b/arch/sh/include/asm/fixmap.h
@@ -14,9 +14,9 @@
 #define _ASM_FIXMAP_H
 
 #include <linux/kernel.h>
+#include <linux/threads.h>
 #include <asm/page.h>
 #ifdef CONFIG_HIGHMEM
-#include <linux/threads.h>
 #include <asm/kmap_types.h>
 #endif
 
@@ -46,9 +46,15 @@
  * fix-mapped?
  */
 enum fixed_addresses {
-#define FIX_N_COLOURS 16
+        /*
+         * The FIX_CMAP entries are used by kmap_coherent() to get virtual
+         * addresses which are of a known color, and so their values are
+         * important. __fix_to_virt(FIX_CMAP_END - n) must give an address
+         * which is the same color as a page (n<<PAGE_SHIFT).
+         */
+#define FIX_N_COLOURS 8
         FIX_CMAP_BEGIN,
-        FIX_CMAP_END = FIX_CMAP_BEGIN + FIX_N_COLOURS,
+        FIX_CMAP_END = FIX_CMAP_BEGIN + (FIX_N_COLOURS * NR_CPUS) - 1,
         FIX_UNCACHED,
 #ifdef CONFIG_HIGHMEM
         FIX_KMAP_BEGIN, /* reserved pte's for temporary kernel mappings */
diff --git a/arch/sh/include/asm/fpu.h b/arch/sh/include/asm/fpu.h
index 1d3aee04b5cc..fb6bbb9b1cc8 100644
--- a/arch/sh/include/asm/fpu.h
+++ b/arch/sh/include/asm/fpu.h
@@ -18,16 +18,15 @@ static inline void grab_fpu(struct pt_regs *regs)
 
 struct task_struct;
 
-extern void save_fpu(struct task_struct *__tsk, struct pt_regs *regs);
+extern void save_fpu(struct task_struct *__tsk);
+void fpu_state_restore(struct pt_regs *regs);
 #else
 
+#define save_fpu(tsk)           do { } while (0)
 #define release_fpu(regs)       do { } while (0)
 #define grab_fpu(regs)          do { } while (0)
+#define fpu_state_restore(regs) do { } while (0)
 
-static inline void save_fpu(struct task_struct *tsk, struct pt_regs *regs)
-{
-        clear_tsk_thread_flag(tsk, TIF_USEDFPU);
-}
 #endif
 
 struct user_regset;
@@ -39,19 +38,28 @@ extern int fpregs_get(struct task_struct *target,
                       unsigned int pos, unsigned int count,
                       void *kbuf, void __user *ubuf);
 
+static inline void __unlazy_fpu(struct task_struct *tsk, struct pt_regs *regs)
+{
+        if (task_thread_info(tsk)->status & TS_USEDFPU) {
+                task_thread_info(tsk)->status &= ~TS_USEDFPU;
+                save_fpu(tsk);
+                release_fpu(regs);
+        } else
+                tsk->fpu_counter = 0;
+}
+
 static inline void unlazy_fpu(struct task_struct *tsk, struct pt_regs *regs)
 {
         preempt_disable();
-        if (test_tsk_thread_flag(tsk, TIF_USEDFPU))
-                save_fpu(tsk, regs);
+        __unlazy_fpu(tsk, regs);
         preempt_enable();
 }
 
 static inline void clear_fpu(struct task_struct *tsk, struct pt_regs *regs)
 {
         preempt_disable();
-        if (test_tsk_thread_flag(tsk, TIF_USEDFPU)) {
-                clear_tsk_thread_flag(tsk, TIF_USEDFPU);
+        if (task_thread_info(tsk)->status & TS_USEDFPU) {
+                task_thread_info(tsk)->status &= ~TS_USEDFPU;
                 release_fpu(regs);
         }
         preempt_enable();
diff --git a/arch/sh/include/asm/ftrace.h b/arch/sh/include/asm/ftrace.h
index 12f3a31f20af..13e9966464c2 100644
--- a/arch/sh/include/asm/ftrace.h
+++ b/arch/sh/include/asm/ftrace.h
@@ -35,4 +35,21 @@ static inline unsigned long ftrace_call_adjust(unsigned long addr)
 #endif /* __ASSEMBLY__ */
 #endif /* CONFIG_FUNCTION_TRACER */
 
+#ifndef __ASSEMBLY__
+
+/* arch/sh/kernel/return_address.c */
+extern void *return_address(unsigned int);
+
+#define HAVE_ARCH_CALLER_ADDR
+
+#define CALLER_ADDR0 ((unsigned long)__builtin_return_address(0))
+#define CALLER_ADDR1 ((unsigned long)return_address(1))
+#define CALLER_ADDR2 ((unsigned long)return_address(2))
+#define CALLER_ADDR3 ((unsigned long)return_address(3))
+#define CALLER_ADDR4 ((unsigned long)return_address(4))
+#define CALLER_ADDR5 ((unsigned long)return_address(5))
+#define CALLER_ADDR6 ((unsigned long)return_address(6))
+
+#endif /* __ASSEMBLY__ */
+
 #endif /* __ASM_SH_FTRACE_H */
diff --git a/arch/sh/include/asm/gpio.h b/arch/sh/include/asm/gpio.h
index 61f93da2c62e..f8d9a731e903 100644
--- a/arch/sh/include/asm/gpio.h
+++ b/arch/sh/include/asm/gpio.h
@@ -20,7 +20,7 @@
 #endif
 
 #define ARCH_NR_GPIOS 512
-#include <asm-generic/gpio.h>
+#include <linux/sh_pfc.h>
 
 #ifdef CONFIG_GPIOLIB
 
@@ -53,84 +53,4 @@ static inline int irq_to_gpio(unsigned int irq)
 
 #endif /* CONFIG_GPIOLIB */
 
-typedef unsigned short pinmux_enum_t;
-typedef unsigned short pinmux_flag_t;
-
-#define PINMUX_TYPE_NONE            0
-#define PINMUX_TYPE_FUNCTION        1
-#define PINMUX_TYPE_GPIO            2
-#define PINMUX_TYPE_OUTPUT          3
-#define PINMUX_TYPE_INPUT           4
-#define PINMUX_TYPE_INPUT_PULLUP    5
-#define PINMUX_TYPE_INPUT_PULLDOWN  6
-
-#define PINMUX_FLAG_TYPE            (0x7)
-#define PINMUX_FLAG_WANT_PULLUP     (1 << 3)
-#define PINMUX_FLAG_WANT_PULLDOWN   (1 << 4)
-
-#define PINMUX_FLAG_DBIT_SHIFT      5
-#define PINMUX_FLAG_DBIT            (0x1f << PINMUX_FLAG_DBIT_SHIFT)
-#define PINMUX_FLAG_DREG_SHIFT      10
-#define PINMUX_FLAG_DREG            (0x3f << PINMUX_FLAG_DREG_SHIFT)
-
-struct pinmux_gpio {
-        pinmux_enum_t enum_id;
-        pinmux_flag_t flags;
-};
-
-#define PINMUX_GPIO(gpio, data_or_mark) [gpio] = { data_or_mark }
-#define PINMUX_DATA(data_or_mark, ids...) data_or_mark, ids, 0
-
-struct pinmux_cfg_reg {
-        unsigned long reg, reg_width, field_width;
-        unsigned long *cnt;
-        pinmux_enum_t *enum_ids;
-};
-
-#define PINMUX_CFG_REG(name, r, r_width, f_width) \
-        .reg = r, .reg_width = r_width, .field_width = f_width,         \
-        .cnt = (unsigned long [r_width / f_width]) {}, \
-        .enum_ids = (pinmux_enum_t [(r_width / f_width) * (1 << f_width)]) \
-
-struct pinmux_data_reg {
-        unsigned long reg, reg_width, reg_shadow;
-        pinmux_enum_t *enum_ids;
-};
-
-#define PINMUX_DATA_REG(name, r, r_width) \
-        .reg = r, .reg_width = r_width, \
-        .enum_ids = (pinmux_enum_t [r_width]) \
-
-struct pinmux_range {
-        pinmux_enum_t begin;
-        pinmux_enum_t end;
-        pinmux_enum_t force;
-};
-
-struct pinmux_info {
-        char *name;
-        pinmux_enum_t reserved_id;
-        struct pinmux_range data;
-        struct pinmux_range input;
-        struct pinmux_range input_pd;
-        struct pinmux_range input_pu;
-        struct pinmux_range output;
-        struct pinmux_range mark;
-        struct pinmux_range function;
-
-        unsigned first_gpio, last_gpio;
-
-        struct pinmux_gpio *gpios;
-        struct pinmux_cfg_reg *cfg_regs;
-        struct pinmux_data_reg *data_regs;
-
-        pinmux_enum_t *gpio_data;
-        unsigned int gpio_data_size;
-
-        unsigned long *gpio_in_use;
-        struct gpio_chip chip;
-};
-
-int register_pinmux(struct pinmux_info *pip);
-
 #endif /* __ASM_SH_GPIO_H */
diff --git a/arch/sh/include/asm/hardirq.h b/arch/sh/include/asm/hardirq.h
index a5be4afa790b..48b191313a99 100644
--- a/arch/sh/include/asm/hardirq.h
+++ b/arch/sh/include/asm/hardirq.h
@@ -1,9 +1,16 @@
 #ifndef __ASM_SH_HARDIRQ_H
 #define __ASM_SH_HARDIRQ_H
 
-extern void ack_bad_irq(unsigned int irq);
-#define ack_bad_irq ack_bad_irq
+#include <linux/threads.h>
+#include <linux/irq.h>
+
+typedef struct {
+        unsigned int __softirq_pending;
+        unsigned int __nmi_count;               /* arch dependent */
+} ____cacheline_aligned irq_cpustat_t;
 
-#include <asm-generic/hardirq.h>
+#include <linux/irq_cpustat.h>  /* Standard mappings for irq_cpustat_t above */
+
+extern void ack_bad_irq(unsigned int irq);
 
 #endif /* __ASM_SH_HARDIRQ_H */
diff --git a/arch/sh/include/asm/io.h b/arch/sh/include/asm/io.h
index 5be45ea4dfec..512cd3e9d0ca 100644
--- a/arch/sh/include/asm/io.h
+++ b/arch/sh/include/asm/io.h
@@ -90,15 +90,11 @@
 #define ctrl_outl               __raw_writel
 #define ctrl_outq               __raw_writeq
 
+extern unsigned long generic_io_base;
+
 static inline void ctrl_delay(void)
 {
-#ifdef CONFIG_CPU_SH4
-        __raw_readw(CCN_PVR);
-#elif defined(P2SEG)
-        __raw_readw(P2SEG);
-#else
-#error "Need a dummy address for delay"
-#endif
+        __raw_readw(generic_io_base);
 }
 
 #define __BUILD_MEMORY_STRING(bwlq, type)                               \
@@ -186,8 +182,6 @@ __BUILD_MEMORY_STRING(q, u64)
 
 #define IO_SPACE_LIMIT 0xffffffff
 
-extern unsigned long generic_io_base;
-
 /*
  * This function provides a method for the generic case where a
  * board-specific ioport_map simply needs to return the port + some
@@ -246,7 +240,7 @@ void __iounmap(void __iomem *addr);
 static inline void __iomem *
 __ioremap_mode(unsigned long offset, unsigned long size, unsigned long flags)
 {
-#if defined(CONFIG_SUPERH32) && !defined(CONFIG_PMB_FIXED)
+#if defined(CONFIG_SUPERH32) && !defined(CONFIG_PMB_FIXED) && !defined(CONFIG_PMB)
         unsigned long last_addr = offset + size - 1;
 #endif
         void __iomem *ret;
@@ -255,7 +249,7 @@ __ioremap_mode(unsigned long offset, unsigned long size, unsigned long flags)
         if (ret)
                 return ret;
 
-#if defined(CONFIG_SUPERH32) && !defined(CONFIG_PMB_FIXED)
+#if defined(CONFIG_SUPERH32) && !defined(CONFIG_PMB_FIXED) && !defined(CONFIG_PMB)
         /*
          * For P1 and P2 space this is trivial, as everything is already
          * mapped. Uncached access for P1 addresses are done through P2.
diff --git a/arch/sh/include/asm/irqflags.h b/arch/sh/include/asm/irqflags.h
index 46e71da5be6b..a741153b41c2 100644
--- a/arch/sh/include/asm/irqflags.h
+++ b/arch/sh/include/asm/irqflags.h
@@ -1,34 +1,9 @@
 #ifndef __ASM_SH_IRQFLAGS_H
 #define __ASM_SH_IRQFLAGS_H
 
-#ifdef CONFIG_SUPERH32
-#include "irqflags_32.h"
-#else
-#include "irqflags_64.h"
-#endif
+#define RAW_IRQ_DISABLED        0xf0
+#define RAW_IRQ_ENABLED         0x00
 
-#define raw_local_save_flags(flags) \
-                do { (flags) = __raw_local_save_flags(); } while (0)
-
-static inline int raw_irqs_disabled_flags(unsigned long flags)
-{
-        return (flags != 0);
-}
-
-static inline int raw_irqs_disabled(void)
-{
-        unsigned long flags = __raw_local_save_flags();
-
-        return raw_irqs_disabled_flags(flags);
-}
-
-#define raw_local_irq_save(flags) \
-                do { (flags) = __raw_local_irq_save(); } while (0)
-
-static inline void raw_local_irq_restore(unsigned long flags)
-{
-        if ((flags & 0xf0) != 0xf0)
-                raw_local_irq_enable();
-}
+#include <asm-generic/irqflags.h>
 
 #endif /* __ASM_SH_IRQFLAGS_H */
diff --git a/arch/sh/include/asm/irqflags_32.h b/arch/sh/include/asm/irqflags_32.h
deleted file mode 100644
index 60218f541340..000000000000
--- a/arch/sh/include/asm/irqflags_32.h
+++ /dev/null
@@ -1,99 +0,0 @@
-#ifndef __ASM_SH_IRQFLAGS_32_H
-#define __ASM_SH_IRQFLAGS_32_H
-
-static inline void raw_local_irq_enable(void)
-{
-        unsigned long __dummy0, __dummy1;
-
-        __asm__ __volatile__ (
-                "stc    sr, %0\n\t"
-                "and    %1, %0\n\t"
-#ifdef CONFIG_CPU_HAS_SR_RB
-                "stc    r6_bank, %1\n\t"
-                "or     %1, %0\n\t"
-#endif
-                "ldc    %0, sr\n\t"
-                : "=&r" (__dummy0), "=r" (__dummy1)
-                : "1" (~0x000000f0)
-                : "memory"
-        );
-}
-
-static inline void raw_local_irq_disable(void)
-{
-        unsigned long flags;
-
-        __asm__ __volatile__ (
-                "stc    sr, %0\n\t"
-                "or     #0xf0, %0\n\t"
-                "ldc    %0, sr\n\t"
-                : "=&z" (flags)
-                : /* no inputs */
-                : "memory"
-        );
-}
-
-static inline void set_bl_bit(void)
-{
-        unsigned long __dummy0, __dummy1;
-
-        __asm__ __volatile__ (
-                "stc    sr, %0\n\t"
-                "or     %2, %0\n\t"
-                "and    %3, %0\n\t"
-                "ldc    %0, sr\n\t"
-                : "=&r" (__dummy0), "=r" (__dummy1)
-                : "r" (0x10000000), "r" (0xffffff0f)
-                : "memory"
-        );
-}
-
-static inline void clear_bl_bit(void)
-{
-        unsigned long __dummy0, __dummy1;
-
-        __asm__ __volatile__ (
-                "stc    sr, %0\n\t"
-                "and    %2, %0\n\t"
-                "ldc    %0, sr\n\t"
-                : "=&r" (__dummy0), "=r" (__dummy1)
-                : "1" (~0x10000000)
-                : "memory"
-        );
-}
-
-static inline unsigned long __raw_local_save_flags(void)
-{
-        unsigned long flags;
-
-        __asm__ __volatile__ (
-                "stc    sr, %0\n\t"
-                "and    #0xf0, %0\n\t"
-                : "=&z" (flags)
-                : /* no inputs */
-                : "memory"
-        );
-
-        return flags;
-}
-
-static inline unsigned long __raw_local_irq_save(void)
-{
-        unsigned long flags, __dummy;
-
-        __asm__ __volatile__ (
-                "stc    sr, %1\n\t"
-                "mov    %1, %0\n\t"
-                "or     #0xf0, %0\n\t"
-                "ldc    %0, sr\n\t"
-                "mov    %1, %0\n\t"
-                "and    #0xf0, %0\n\t"
-                : "=&z" (flags), "=&r" (__dummy)
-                : /* no inputs */
-                : "memory"
-        );
-
-        return flags;
-}
-
-#endif /* __ASM_SH_IRQFLAGS_32_H */
diff --git a/arch/sh/include/asm/irqflags_64.h b/arch/sh/include/asm/irqflags_64.h
deleted file mode 100644
index 88f65222c1d4..000000000000
--- a/arch/sh/include/asm/irqflags_64.h
+++ /dev/null
@@ -1,85 +0,0 @@
-#ifndef __ASM_SH_IRQFLAGS_64_H
-#define __ASM_SH_IRQFLAGS_64_H
-
-#include <cpu/registers.h>
-
-#define SR_MASK_LL      0x00000000000000f0LL
-#define SR_BL_LL        0x0000000010000000LL
-
-static inline void raw_local_irq_enable(void)
-{
-        unsigned long long __dummy0, __dummy1 = ~SR_MASK_LL;
-
-        __asm__ __volatile__("getcon    " __SR ", %0\n\t"
-                             "and       %0, %1, %0\n\t"
-                             "putcon    %0, " __SR "\n\t"
-                             : "=&r" (__dummy0)
-                             : "r" (__dummy1));
-}
-
-static inline void raw_local_irq_disable(void)
-{
-        unsigned long long __dummy0, __dummy1 = SR_MASK_LL;
-
-        __asm__ __volatile__("getcon    " __SR ", %0\n\t"
-                             "or        %0, %1, %0\n\t"
-                             "putcon    %0, " __SR "\n\t"
-                             : "=&r" (__dummy0)
-                             : "r" (__dummy1));
-}
-
-static inline void set_bl_bit(void)
-{
-        unsigned long long __dummy0, __dummy1 = SR_BL_LL;
-
-        __asm__ __volatile__("getcon    " __SR ", %0\n\t"
-                             "or        %0, %1, %0\n\t"
-                             "putcon    %0, " __SR "\n\t"
-                             : "=&r" (__dummy0)
-                             : "r" (__dummy1));
-
-}
-
-static inline void clear_bl_bit(void)
-{
-        unsigned long long __dummy0, __dummy1 = ~SR_BL_LL;
-
-        __asm__ __volatile__("getcon    " __SR ", %0\n\t"
-                             "and       %0, %1, %0\n\t"
-                             "putcon    %0, " __SR "\n\t"
-                             : "=&r" (__dummy0)
-                             : "r" (__dummy1));
-}
-
-static inline unsigned long __raw_local_save_flags(void)
-{
-        unsigned long long __dummy = SR_MASK_LL;
-        unsigned long flags;
-
-        __asm__ __volatile__ (
-                "getcon " __SR ", %0\n\t"
-                "and    %0, %1, %0"
-                : "=&r" (flags)
-                : "r" (__dummy));
-
-        return flags;
-}
-
-static inline unsigned long __raw_local_irq_save(void)
-{
-        unsigned long long __dummy0, __dummy1 = SR_MASK_LL;
-        unsigned long flags;
-
-        __asm__ __volatile__ (
-                "getcon " __SR ", %1\n\t"
-                "or     %1, r63, %0\n\t"
-                "or     %1, %2, %1\n\t"
-                "putcon %1, " __SR "\n\t"
-                "and    %0, %2, %0"
-                : "=&r" (flags), "=&r" (__dummy0)
-                : "r" (__dummy1));
-
-        return flags;
-}
-
-#endif /* __ASM_SH_IRQFLAGS_64_H */
diff --git a/arch/sh/include/asm/mmu.h b/arch/sh/include/asm/mmu.h
index f5963037c9d6..c7426ad9926e 100644
--- a/arch/sh/include/asm/mmu.h
+++ b/arch/sh/include/asm/mmu.h
@@ -7,12 +7,16 @@
 #define PMB_PASCR               0xff000070
 #define PMB_IRMCR               0xff000078
 
+#define PASCR_SE                0x80000000
+
 #define PMB_ADDR                0xf6100000
 #define PMB_DATA                0xf7100000
 #define PMB_ENTRY_MAX           16
 #define PMB_E_MASK              0x0000000f
 #define PMB_E_SHIFT             8
 
+#define PMB_PFN_MASK            0xff000000
+
 #define PMB_SZ_16M              0x00000000
 #define PMB_SZ_64M              0x00000010
 #define PMB_SZ_128M             0x00000080
@@ -62,17 +66,10 @@ struct pmb_entry {
 };
 
 /* arch/sh/mm/pmb.c */
-int __set_pmb_entry(unsigned long vpn, unsigned long ppn,
-                    unsigned long flags, int *entry);
-int set_pmb_entry(struct pmb_entry *pmbe);
-void clear_pmb_entry(struct pmb_entry *pmbe);
-struct pmb_entry *pmb_alloc(unsigned long vpn, unsigned long ppn,
-                            unsigned long flags);
-void pmb_free(struct pmb_entry *pmbe);
 long pmb_remap(unsigned long virt, unsigned long phys,
                unsigned long size, unsigned long flags);
 void pmb_unmap(unsigned long addr);
+int pmb_init(void);
 #endif /* __ASSEMBLY__ */
 
 #endif /* __MMU_H */
-
diff --git a/arch/sh/include/asm/pci.h b/arch/sh/include/asm/pci.h
index 4163950cd1c6..67f3999b544e 100644
--- a/arch/sh/include/asm/pci.h
+++ b/arch/sh/include/asm/pci.h
@@ -3,8 +3,6 @@
 
 #ifdef __KERNEL__
 
-#include <linux/dma-mapping.h>
-
 /* Can be used to override the logic in pci_scan_bus for skipping
    already-configured bus numbers - to be used for buggy BIOSes
    or architectures with incomplete PCI setup by the loader */
@@ -54,30 +52,18 @@ static inline void pcibios_penalize_isa_irq(int irq, int active)
  * address space.  The networking and block device layers use
  * this boolean for bounce buffer decisions.
  */
-#define PCI_DMA_BUS_IS_PHYS     (1)
-
-#include <linux/types.h>
-#include <linux/slab.h>
-#include <asm/scatterlist.h>
-#include <linux/string.h>
-#include <asm/io.h>
+#define PCI_DMA_BUS_IS_PHYS     (dma_ops->is_phys)
 
 /* pci_unmap_{single,page} being a nop depends upon the
  * configuration.
  */
-#ifdef CONFIG_SH_PCIDMA_NONCOHERENT
-#define DECLARE_PCI_UNMAP_ADDR(ADDR_NAME)       \
-        dma_addr_t ADDR_NAME;
-#define DECLARE_PCI_UNMAP_LEN(LEN_NAME)         \
-        __u32 LEN_NAME;
-#define pci_unmap_addr(PTR, ADDR_NAME)                  \
-        ((PTR)->ADDR_NAME)
-#define pci_unmap_addr_set(PTR, ADDR_NAME, VAL)         \
-        (((PTR)->ADDR_NAME) = (VAL))
-#define pci_unmap_len(PTR, LEN_NAME)                    \
-        ((PTR)->LEN_NAME)
-#define pci_unmap_len_set(PTR, LEN_NAME, VAL)           \
-        (((PTR)->LEN_NAME) = (VAL))
+#ifdef CONFIG_DMA_NONCOHERENT
+#define DECLARE_PCI_UNMAP_ADDR(ADDR_NAME)       dma_addr_t ADDR_NAME;
+#define DECLARE_PCI_UNMAP_LEN(LEN_NAME)         __u32 LEN_NAME;
+#define pci_unmap_addr(PTR, ADDR_NAME)          ((PTR)->ADDR_NAME)
+#define pci_unmap_addr_set(PTR, ADDR_NAME, VAL) (((PTR)->ADDR_NAME) = (VAL))
+#define pci_unmap_len(PTR, LEN_NAME)            ((PTR)->LEN_NAME)
+#define pci_unmap_len_set(PTR, LEN_NAME, VAL)   (((PTR)->LEN_NAME) = (VAL))
 #else
 #define DECLARE_PCI_UNMAP_ADDR(ADDR_NAME)
 #define DECLARE_PCI_UNMAP_LEN(LEN_NAME)
diff --git a/arch/sh/include/asm/perf_event.h b/arch/sh/include/asm/perf_event.h
index 11a302297ab7..3d0c9f36d150 100644
--- a/arch/sh/include/asm/perf_event.h
+++ b/arch/sh/include/asm/perf_event.h
@@ -1,8 +1,35 @@
 #ifndef __ASM_SH_PERF_EVENT_H
 #define __ASM_SH_PERF_EVENT_H
 
-/* SH only supports software events through this interface. */
-static inline void set_perf_event_pending(void) {}
+struct hw_perf_event;
+
+#define MAX_HWEVENTS    2
+
+struct sh_pmu {
+        const char      *name;
+        unsigned int    num_events;
+        void            (*disable_all)(void);
+        void            (*enable_all)(void);
+        void            (*enable)(struct hw_perf_event *, int);
+        void            (*disable)(struct hw_perf_event *, int);
+        u64             (*read)(int);
+        int             (*event_map)(int);
+        unsigned int    max_events;
+        unsigned long   raw_event_mask;
+        const int       (*cache_events)[PERF_COUNT_HW_CACHE_MAX]
+                                       [PERF_COUNT_HW_CACHE_OP_MAX]
+                                       [PERF_COUNT_HW_CACHE_RESULT_MAX];
+};
+
+/* arch/sh/kernel/perf_event.c */
+extern int register_sh_pmu(struct sh_pmu *);
+extern int reserve_pmc_hardware(void);
+extern void release_pmc_hardware(void);
+
+static inline void set_perf_event_pending(void)
+{
+        /* Nothing to see here, move along. */
+}
 
 #define PERF_EVENT_INDEX_OFFSET 0
 
diff --git a/arch/sh/include/asm/pgtable.h b/arch/sh/include/asm/pgtable.h
index 4f3efa7d5a64..ba3046e4f06f 100644
--- a/arch/sh/include/asm/pgtable.h
+++ b/arch/sh/include/asm/pgtable.h
@@ -75,13 +75,31 @@ static inline unsigned long long neff_sign_extend(unsigned long val)
 #define USER_PTRS_PER_PGD       (TASK_SIZE/PGDIR_SIZE)
 #define FIRST_USER_ADDRESS      0
 
-#ifdef CONFIG_32BIT
-#define PHYS_ADDR_MASK          0xffffffff
+#define PHYS_ADDR_MASK29                0x1fffffff
+#define PHYS_ADDR_MASK32                0xffffffff
+
+#ifdef CONFIG_PMB
+static inline unsigned long phys_addr_mask(void)
+{
+        /* Is the MMU in 29bit mode? */
+        if (__in_29bit_mode())
+                return PHYS_ADDR_MASK29;
+
+        return PHYS_ADDR_MASK32;
+}
+#elif defined(CONFIG_32BIT)
+static inline unsigned long phys_addr_mask(void)
+{
+        return PHYS_ADDR_MASK32;
+}
 #else
-#define PHYS_ADDR_MASK          0x1fffffff
+static inline unsigned long phys_addr_mask(void)
+{
+        return PHYS_ADDR_MASK29;
+}
 #endif
 
-#define PTE_PHYS_MASK           (PHYS_ADDR_MASK & PAGE_MASK)
+#define PTE_PHYS_MASK           (phys_addr_mask() & PAGE_MASK)
 #define PTE_FLAGS_MASK          (~(PTE_PHYS_MASK) << PAGE_SHIFT)
 
 #ifdef CONFIG_SUPERH32
diff --git a/arch/sh/include/asm/pgtable_32.h b/arch/sh/include/asm/pgtable_32.h
index c0d359ce337b..b35435516203 100644
--- a/arch/sh/include/asm/pgtable_32.h
+++ b/arch/sh/include/asm/pgtable_32.h
@@ -108,7 +108,7 @@ static inline unsigned long copy_ptea_attributes(unsigned long x)
 #define _PAGE_CLEAR_FLAGS       (_PAGE_PROTNONE | _PAGE_ACCESSED | _PAGE_FILE)
 #endif
 
-#define _PAGE_FLAGS_HARDWARE_MASK       (PHYS_ADDR_MASK & ~(_PAGE_CLEAR_FLAGS))
+#define _PAGE_FLAGS_HARDWARE_MASK       (phys_addr_mask() & ~(_PAGE_CLEAR_FLAGS))
 
 /* Hardware flags, page size encoding */
 #if !defined(CONFIG_MMU)
diff --git a/arch/sh/include/asm/processor_32.h b/arch/sh/include/asm/processor_32.h
index f4b54040dbc3..d60b28271a05 100644
--- a/arch/sh/include/asm/processor_32.h
+++ b/arch/sh/include/asm/processor_32.h
@@ -57,6 +57,7 @@ asmlinkage void __init sh_cpu_init(void);
 #define SR_DSP          0x00001000
 #define SR_IMASK        0x000000f0
 #define SR_FD           0x00008000
+#define SR_MD           0x40000000
 
 /*
  * DSP structure and data
@@ -134,7 +135,7 @@ struct mm_struct;
 extern void release_thread(struct task_struct *);
 
 /* Prepare to copy thread state - unlazy all lazy status */
-#define prepare_to_copy(tsk)    do { } while (0)
+void prepare_to_copy(struct task_struct *tsk);
 
 /*
  * create a kernel thread without removing it from tasklists
diff --git a/arch/sh/include/asm/scatterlist.h b/arch/sh/include/asm/scatterlist.h
index 327cc2e4c97b..e38d1d4c7f6f 100644
--- a/arch/sh/include/asm/scatterlist.h
+++ b/arch/sh/include/asm/scatterlist.h
@@ -1,7 +1,7 @@
 #ifndef __ASM_SH_SCATTERLIST_H
 #define __ASM_SH_SCATTERLIST_H
 
-#define ISA_DMA_THRESHOLD       PHYS_ADDR_MASK
+#define ISA_DMA_THRESHOLD       phys_addr_mask()
 
 #include <asm-generic/scatterlist.h>
 
diff --git a/arch/sh/include/asm/sh_keysc.h b/arch/sh/include/asm/sh_keysc.h
deleted file mode 100644
index 4a65b1e40eab..000000000000
--- a/arch/sh/include/asm/sh_keysc.h
+++ /dev/null
@@ -1,14 +0,0 @@
-#ifndef __ASM_KEYSC_H__
-#define __ASM_KEYSC_H__
-
-#define SH_KEYSC_MAXKEYS 30
-
-struct sh_keysc_info {
-        enum { SH_KEYSC_MODE_1, SH_KEYSC_MODE_2, SH_KEYSC_MODE_3 } mode;
-        int scan_timing; /* 0 -> 7, see KYCR1, SCN[2:0] */
-        int delay;
-        int kycr2_delay;
-        int keycodes[SH_KEYSC_MAXKEYS];
-};
-
-#endif /* __ASM_KEYSC_H__ */
diff --git a/arch/sh/include/asm/suspend.h b/arch/sh/include/asm/suspend.h
index 5c8ea28ff7a4..fe9c2a1ad047 100644
--- a/arch/sh/include/asm/suspend.h
+++ b/arch/sh/include/asm/suspend.h
@@ -2,6 +2,7 @@
 #define _ASM_SH_SUSPEND_H
 
 #ifndef __ASSEMBLY__
+#include <linux/notifier.h>
 static inline int arch_prepare_suspend(void) { return 0; }
 
 #include <asm/ptrace.h>
@@ -19,6 +20,69 @@ void sh_mobile_setup_cpuidle(void);
 static inline void sh_mobile_setup_cpuidle(void) {}
 #endif
 
+/* notifier chains for pre/post sleep hooks */
+extern struct atomic_notifier_head sh_mobile_pre_sleep_notifier_list;
+extern struct atomic_notifier_head sh_mobile_post_sleep_notifier_list;
+
+/* priority levels for notifiers */
+#define SH_MOBILE_SLEEP_BOARD   0
+#define SH_MOBILE_SLEEP_CPU     1
+#define SH_MOBILE_PRE(x)        (x)
+#define SH_MOBILE_POST(x)       (-(x))
+
+/* board code registration function for self-refresh assembly snippets */
+void sh_mobile_register_self_refresh(unsigned long flags,
+                                     void *pre_start, void *pre_end,
+                                     void *post_start, void *post_end);
+
+/* register structure for address/data information */
+struct sh_sleep_regs {
+        unsigned long stbcr;
+        unsigned long bar;
+
+        /* MMU */
+        unsigned long pteh;
+        unsigned long ptel;
+        unsigned long ttb;
+        unsigned long tea;
+        unsigned long mmucr;
+        unsigned long ptea;
+        unsigned long pascr;
+        unsigned long irmcr;
+
+        /* Cache */
+        unsigned long ccr;
+        unsigned long ramcr;
+};
+
+/* data area for low-level sleep code */
+struct sh_sleep_data {
+        /* current sleep mode (SUSP_SH_...) */
+        unsigned long mode;
+
+        /* addresses of board specific self-refresh snippets */
+        unsigned long sf_pre;
+        unsigned long sf_post;
+
+        /* address of resume code */
+        unsigned long resume;
+
+        /* register state saved and restored by the assembly code */
+        unsigned long vbr;
+        unsigned long spc;
+        unsigned long sr;
+        unsigned long sp;
+
+        /* structure for keeping register addresses */
+        struct sh_sleep_regs addr;
+
+        /* structure for saving/restoring register state */
+        struct sh_sleep_regs data;
+};
+
+/* a bitmap of supported sleep modes (SUSP_SH..) */
+extern unsigned long sh_mobile_sleep_supported;
+
 #endif
 
 /* flags passed to assembly suspend code */
@@ -27,5 +91,6 @@ static inline void sh_mobile_setup_cpuidle(void) {}
 #define SUSP_SH_RSTANDBY        (1 << 2) /* SH-Mobile R-standby mode */
 #define SUSP_SH_USTANDBY        (1 << 3) /* SH-Mobile U-standby mode */
 #define SUSP_SH_SF              (1 << 4) /* Enable self-refresh */
+#define SUSP_SH_MMU             (1 << 5) /* Save/restore MMU and cache */
 
 #endif /* _ASM_SH_SUSPEND_H */
diff --git a/arch/sh/include/asm/system.h b/arch/sh/include/asm/system.h
index 1014da8b3ed3..91b173fd37a6 100644
--- a/arch/sh/include/asm/system.h
+++ b/arch/sh/include/asm/system.h
@@ -169,10 +169,6 @@ BUILD_TRAP_HANDLER(fpu_error);
 BUILD_TRAP_HANDLER(fpu_state_restore);
 BUILD_TRAP_HANDLER(nmi);
 
-#ifdef CONFIG_BUG
-extern void handle_BUG(struct pt_regs *);
-#endif
-
 #define arch_align_stack(x) (x)
 
 struct mem_access {
diff --git a/arch/sh/include/asm/system_32.h b/arch/sh/include/asm/system_32.h
index 607d413f6168..06814f5b59c7 100644
--- a/arch/sh/include/asm/system_32.h
+++ b/arch/sh/include/asm/system_32.h
@@ -232,4 +232,33 @@ asmlinkage void do_exception_error(unsigned long r4, unsigned long r5,
                                    unsigned long r6, unsigned long r7,
                                    struct pt_regs __regs);
 
+static inline void set_bl_bit(void)
+{
+        unsigned long __dummy0, __dummy1;
+
+        __asm__ __volatile__ (
+                "stc    sr, %0\n\t"
+                "or     %2, %0\n\t"
+                "and    %3, %0\n\t"
+                "ldc    %0, sr\n\t"
+                : "=&r" (__dummy0), "=r" (__dummy1)
+                : "r" (0x10000000), "r" (0xffffff0f)
+                : "memory"
+        );
+}
+
+static inline void clear_bl_bit(void)
+{
+        unsigned long __dummy0, __dummy1;
+
+        __asm__ __volatile__ (
+                "stc    sr, %0\n\t"
+                "and    %2, %0\n\t"
+                "ldc    %0, sr\n\t"
+                : "=&r" (__dummy0), "=r" (__dummy1)
+                : "1" (~0x10000000)
+                : "memory"
+        );
+}
+
 #endif /* __ASM_SH_SYSTEM_32_H */
diff --git a/arch/sh/include/asm/system_64.h b/arch/sh/include/asm/system_64.h
index 8e4a03e7966c..ab1dd917ea87 100644
--- a/arch/sh/include/asm/system_64.h
+++ b/arch/sh/include/asm/system_64.h
@@ -12,6 +12,7 @@
  * License.  See the file "COPYING" in the main directory of this archive
  * for more details.
  */
+#include <cpu/registers.h>
 #include <asm/processor.h>
 
 /*
@@ -47,4 +48,29 @@ static inline reg_size_t register_align(void *val)
         return (unsigned long long)(signed long long)(signed long)val;
 }
 
+#define SR_BL_LL        0x0000000010000000LL
+
+static inline void set_bl_bit(void)
+{
+        unsigned long long __dummy0, __dummy1 = SR_BL_LL;
+
+        __asm__ __volatile__("getcon    " __SR ", %0\n\t"
+                             "or        %0, %1, %0\n\t"
+                             "putcon    %0, " __SR "\n\t"
+                             : "=&r" (__dummy0)
+                             : "r" (__dummy1));
+
+}
+
+static inline void clear_bl_bit(void)
+{
+        unsigned long long __dummy0, __dummy1 = ~SR_BL_LL;
+
+        __asm__ __volatile__("getcon    " __SR ", %0\n\t"
+                             "and       %0, %1, %0\n\t"
+                             "putcon    %0, " __SR "\n\t"
+                             : "=&r" (__dummy0)
+                             : "r" (__dummy1));
+}
+
 #endif /* __ASM_SH_SYSTEM_64_H */
diff --git a/arch/sh/include/asm/thread_info.h b/arch/sh/include/asm/thread_info.h
index 8ab9145bf50b..e0856b205c4e 100644
--- a/arch/sh/include/asm/thread_info.h
+++ b/arch/sh/include/asm/thread_info.h
@@ -19,6 +19,7 @@ struct thread_info {
         struct task_struct      *task;          /* main task structure */
         struct exec_domain      *exec_domain;   /* execution domain */
         unsigned long           flags;          /* low level flags */
+        __u32                   status;         /* thread synchronous flags */
         __u32                   cpu;
         int                     preempt_count; /* 0 => preemptable, <0 => BUG */
         mm_segment_t            addr_limit;     /* thread address space */
@@ -50,6 +51,7 @@ struct thread_info {
         .task           = &tsk,                 \
         .exec_domain    = &default_exec_domain, \
         .flags          = 0,                    \
+        .status         = 0,                    \
         .cpu            = 0,                    \
         .preempt_count  = INIT_PREEMPT_COUNT,   \
         .addr_limit     = KERNEL_DS,            \
@@ -111,14 +113,12 @@ extern void free_thread_info(struct thread_info *ti);
 #define TIF_SYSCALL_TRACE       0       /* syscall trace active */
 #define TIF_SIGPENDING          1       /* signal pending */
 #define TIF_NEED_RESCHED        2       /* rescheduling necessary */
-#define TIF_RESTORE_SIGMASK     3       /* restore signal mask in do_signal() */
 #define TIF_SINGLESTEP          4       /* singlestepping active */
 #define TIF_SYSCALL_AUDIT       5       /* syscall auditing active */
 #define TIF_SECCOMP             6       /* secure computing */
 #define TIF_NOTIFY_RESUME       7       /* callback before returning to user */
 #define TIF_SYSCALL_TRACEPOINT  8       /* for ftrace syscall instrumentation */
 #define TIF_DEBUG               9       /* uses UBC */
-#define TIF_USEDFPU             16      /* FPU was used by this task this quantum (SMP) */
 #define TIF_POLLING_NRFLAG      17      /* true if poll_idle() is polling TIF_NEED_RESCHED */
 #define TIF_MEMDIE              18
 #define TIF_FREEZE              19      /* Freezing for suspend */
@@ -126,14 +126,12 @@ extern void free_thread_info(struct thread_info *ti);
 #define _TIF_SYSCALL_TRACE      (1 << TIF_SYSCALL_TRACE)
 #define _TIF_SIGPENDING         (1 << TIF_SIGPENDING)
 #define _TIF_NEED_RESCHED       (1 << TIF_NEED_RESCHED)
-#define _TIF_RESTORE_SIGMASK    (1 << TIF_RESTORE_SIGMASK)
 #define _TIF_SINGLESTEP         (1 << TIF_SINGLESTEP)
 #define _TIF_SYSCALL_AUDIT      (1 << TIF_SYSCALL_AUDIT)
 #define _TIF_SECCOMP            (1 << TIF_SECCOMP)
 #define _TIF_NOTIFY_RESUME      (1 << TIF_NOTIFY_RESUME)
 #define _TIF_SYSCALL_TRACEPOINT (1 << TIF_SYSCALL_TRACEPOINT)
 #define _TIF_DEBUG              (1 << TIF_DEBUG)
-#define _TIF_USEDFPU            (1 << TIF_USEDFPU)
 #define _TIF_POLLING_NRFLAG     (1 << TIF_POLLING_NRFLAG)
 #define _TIF_FREEZE             (1 << TIF_FREEZE)
 
@@ -151,13 +149,33 @@ extern void free_thread_info(struct thread_info *ti);
 /* work to do on any return to u-space */
 #define _TIF_ALLWORK_MASK       (_TIF_SYSCALL_TRACE | _TIF_SIGPENDING      | \
                                  _TIF_NEED_RESCHED  | _TIF_SYSCALL_AUDIT   | \
-                                 _TIF_SINGLESTEP    | _TIF_RESTORE_SIGMASK | \
-                                 _TIF_NOTIFY_RESUME | _TIF_SYSCALL_TRACEPOINT)
+                                 _TIF_SINGLESTEP    | _TIF_NOTIFY_RESUME   | \
+                                 _TIF_SYSCALL_TRACEPOINT)
 
 /* work to do on interrupt/exception return */
 #define _TIF_WORK_MASK          (_TIF_ALLWORK_MASK & ~(_TIF_SYSCALL_TRACE | \
                                  _TIF_SYSCALL_AUDIT | _TIF_SINGLESTEP))
 
+/*
+ * Thread-synchronous status.
+ *
+ * This is different from the flags in that nobody else
+ * ever touches our thread-synchronous status, so we don't
+ * have to worry about atomic accesses.
+ */
+#define TS_RESTORE_SIGMASK      0x0001  /* restore signal mask in do_signal() */
+#define TS_USEDFPU              0x0002  /* FPU used by this task this quantum */
+
+#ifndef __ASSEMBLY__
+#define HAVE_SET_RESTORE_SIGMASK        1
+static inline void set_restore_sigmask(void)
+{
+        struct thread_info *ti = current_thread_info();
+        ti->status |= TS_RESTORE_SIGMASK;
+        set_bit(TIF_SIGPENDING, (unsigned long *)&ti->flags);
+}
+#endif  /* !__ASSEMBLY__ */
+
 #endif /* __KERNEL__ */
 
 #endif /* __ASM_SH_THREAD_INFO_H */
diff --git a/arch/sh/include/asm/topology.h b/arch/sh/include/asm/topology.h
index 65e7bd2f2240..37cdadd975ac 100644
--- a/arch/sh/include/asm/topology.h
+++ b/arch/sh/include/asm/topology.h
@@ -40,6 +40,14 @@
 
 #endif
 
+#define mc_capable()    (1)
+
+const struct cpumask *cpu_coregroup_mask(unsigned int cpu);
+
+extern cpumask_t cpu_core_map[NR_CPUS];
+
+#define topology_core_cpumask(cpu)      (&cpu_core_map[cpu])
+
 #include <asm-generic/topology.h>
 
 #endif /* _ASM_SH_TOPOLOGY_H */
diff --git a/arch/sh/include/asm/ubc.h b/arch/sh/include/asm/ubc.h
index dd7878197b6e..e3c2968b04d3 100644
--- a/arch/sh/include/asm/ubc.h
+++ b/arch/sh/include/asm/ubc.h
@@ -66,16 +66,5 @@
  */
 #define NR_UBC_CHANNELS         2
 
-#ifndef __ASSEMBLY__
-/* arch/sh/kernel/cpu/ubc.S */
-extern void ubc_sleep(void);
-
-#ifdef CONFIG_UBC_WAKEUP
-extern void ubc_wakeup(void);
-#else
-#define ubc_wakeup()    do { } while (0)
-#endif
-#endif
-
 #endif /* __KERNEL__ */
 #endif /* __ASM_SH_UBC_H */
diff --git a/arch/sh/include/asm/watchdog.h b/arch/sh/include/asm/watchdog.h
index 2fe7cee9e43a..19dfff5c8511 100644
--- a/arch/sh/include/asm/watchdog.h
+++ b/arch/sh/include/asm/watchdog.h
@@ -2,6 +2,8 @@
  * include/asm-sh/watchdog.h
  *
  * Copyright (C) 2002, 2003 Paul Mundt
+ * Copyright (C) 2009 Siemens AG
+ * Copyright (C) 2009 Valentin Sitdikov
  *
  * 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
@@ -61,6 +63,61 @@
 #define WTCSR_CKS_2048  0x06
 #define WTCSR_CKS_4096  0x07
 
+#if defined(CONFIG_CPU_SUBTYPE_SH7785) || defined(CONFIG_CPU_SUBTYPE_SH7780)
+/**
+ *      sh_wdt_read_cnt - Read from Counter
+ *      Reads back the WTCNT value.
+ */
+static inline __u32 sh_wdt_read_cnt(void)
+{
+        return ctrl_inl(WTCNT_R);
+}
+
+/**
+ *      sh_wdt_write_cnt - Write to Counter
+ *      @val: Value to write
+ *
+ *      Writes the given value @val to the lower byte of the timer counter.
+ *      The upper byte is set manually on each write.
+ */
+static inline void sh_wdt_write_cnt(__u32 val)
+{
+        ctrl_outl((WTCNT_HIGH << 24) | (__u32)val, WTCNT);
+}
+
+/**
+ *      sh_wdt_write_bst - Write to Counter
+ *      @val: Value to write
+ *
+ *      Writes the given value @val to the lower byte of the timer counter.
+ *      The upper byte is set manually on each write.
+ */
+static inline void sh_wdt_write_bst(__u32 val)
+{
+        ctrl_outl((WTBST_HIGH << 24) | (__u32)val, WTBST);
+}
+/**
+ *      sh_wdt_read_csr - Read from Control/Status Register
+ *
+ *      Reads back the WTCSR value.
+ */
+static inline __u32 sh_wdt_read_csr(void)
+{
+        return ctrl_inl(WTCSR_R);
+}
+
+/**
+ *      sh_wdt_write_csr - Write to Control/Status Register
+ *      @val: Value to write
+ *
+ *      Writes the given value @val to the lower byte of the control/status
+ *      register. The upper byte is set manually on each write.
+ */
+static inline void sh_wdt_write_csr(__u32 val)
+{
+        ctrl_outl((WTCSR_HIGH << 24) | (__u32)val, WTCSR);
+}
+#else
 /**
  *      sh_wdt_read_cnt - Read from Counter
  *      Reads back the WTCNT value.
@@ -103,6 +160,6 @@ static inline void sh_wdt_write_csr(__u8 val)
 {
         ctrl_outw((WTCSR_HIGH << 8) | (__u16)val, WTCSR);
 }
-
+#endif /* CONFIG_CPU_SUBTYPE_SH7785 || CONFIG_CPU_SUBTYPE_SH7780 */
 #endif /* __KERNEL__ */
 #endif /* __ASM_SH_WATCHDOG_H */
diff --git a/arch/sh/include/cpu-sh4/cpu/watchdog.h b/arch/sh/include/cpu-sh4/cpu/watchdog.h
index 259f6a0ce23d..7672301d0c70 100644
--- a/arch/sh/include/cpu-sh4/cpu/watchdog.h
+++ b/arch/sh/include/cpu-sh4/cpu/watchdog.h
@@ -2,6 +2,8 @@
  * include/asm-sh/cpu-sh4/watchdog.h
  *
  * Copyright (C) 2002, 2003 Paul Mundt
+ * Copyright (C) 2009 Siemens AG
+ * Copyright (C) 2009 Sitdikov Valentin
  *
  * 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
@@ -10,9 +12,20 @@
 #ifndef __ASM_CPU_SH4_WATCHDOG_H
 #define __ASM_CPU_SH4_WATCHDOG_H
 
+#if defined(CONFIG_CPU_SUBTYPE_SH7785) || defined(CONFIG_CPU_SUBTYPE_SH7780)
+/* Prefix definition */
+#define WTBST_HIGH      0x55
+/* Register definitions */
+#define WTCNT_R         0xffcc0010 /*WDTCNT*/
+#define WTCSR           0xffcc0004 /*WDTCSR*/
+#define WTCNT           0xffcc0000 /*WDTST*/
+#define WTST            WTCNT
+#define WTBST           0xffcc0008 /*WDTBST*/
+#else
 /* Register definitions */
 #define WTCNT           0xffc00008
 #define WTCSR           0xffc0000c
+#endif
 
 /* Bit definitions */
 #define WTCSR_TME       0x80
diff --git a/arch/sh/include/mach-ecovec24/mach/partner-jet-setup.txt b/arch/sh/include/mach-ecovec24/mach/partner-jet-setup.txt
index 8b8e4fa1fee9..cc737b807334 100644
--- a/arch/sh/include/mach-ecovec24/mach/partner-jet-setup.txt
+++ b/arch/sh/include/mach-ecovec24/mach/partner-jet-setup.txt
@@ -22,13 +22,12 @@ ED 0xff000010, 0x00000004
 LIST "setup clocks"
 ED 0xa4150024, 0x00004000
 ED 0xa4150000, 0x8E003508
-ED 0xa4150004, 0x00000000
 
 WAIT 1
 
 LIST "BSC"
 ED 0xff800020, 0xa5a50000
-ED 0xfec10000, 0x00000013
+ED 0xfec10000, 0x00001013
 ED 0xfec10004, 0x11110400
 ED 0xfec10024, 0x00000440
 
diff --git a/arch/sh/include/mach-se/mach/se7722.h b/arch/sh/include/mach-se/mach/se7722.h
index e971d9a82f4a..16505bfb8a9e 100644
--- a/arch/sh/include/mach-se/mach/se7722.h
+++ b/arch/sh/include/mach-se/mach/se7722.h
@@ -92,18 +92,11 @@
 #define SE7722_FPGA_IRQ_MRSHPC1 3 /* IRQ1 */
 #define SE7722_FPGA_IRQ_MRSHPC2 4 /* IRQ1 */
 #define SE7722_FPGA_IRQ_MRSHPC3 5 /* IRQ1 */
-
 #define SE7722_FPGA_IRQ_NR      6
-#define SE7722_FPGA_IRQ_BASE    110
-
-#define MRSHPC_IRQ3     (SE7722_FPGA_IRQ_BASE + SE7722_FPGA_IRQ_MRSHPC3)
-#define MRSHPC_IRQ2     (SE7722_FPGA_IRQ_BASE + SE7722_FPGA_IRQ_MRSHPC2)
-#define MRSHPC_IRQ1     (SE7722_FPGA_IRQ_BASE + SE7722_FPGA_IRQ_MRSHPC1)
-#define MRSHPC_IRQ0     (SE7722_FPGA_IRQ_BASE + SE7722_FPGA_IRQ_MRSHPC0)
-#define SMC_IRQ         (SE7722_FPGA_IRQ_BASE + SE7722_FPGA_IRQ_SMC)
-#define USB_IRQ         (SE7722_FPGA_IRQ_BASE + SE7722_FPGA_IRQ_USB)
 
 /* arch/sh/boards/se/7722/irq.c */
+extern unsigned int se7722_fpga_irq[];
+
 void init_se7722_IRQ(void);
 
 #define __IO_PREFIX             se7722
diff --git a/arch/sh/kernel/Makefile b/arch/sh/kernel/Makefile
index 649daadd4519..cd4387737ed8 100644
--- a/arch/sh/kernel/Makefile
+++ b/arch/sh/kernel/Makefile
@@ -9,8 +9,12 @@ ifdef CONFIG_FUNCTION_TRACER
 CFLAGS_REMOVE_ftrace.o = -pg
 endif
 
-obj-y   := debugtraps.o dumpstack.o idle.o io.o io_generic.o irq.o      \
-           machvec.o nmi_debug.o process_$(BITS).o ptrace_$(BITS).o     \
+CFLAGS_REMOVE_return_address.o = -pg
+
+obj-y   := debugtraps.o dma-nommu.o dumpstack.o                         \
+           idle.o io.o io_generic.o irq.o                               \
+           irq_$(BITS).o machvec.o nmi_debug.o process_$(BITS).o        \
+           ptrace_$(BITS).o return_address.o                            \
            setup.o signal_$(BITS).o sys_sh.o sys_sh$(BITS).o            \
            syscalls_$(BITS).o time.o topology.o traps.o                 \
            traps_$(BITS).o unwinder.o
@@ -28,13 +32,13 @@ obj-$(CONFIG_CRASH_DUMP)	+= crash_dump.o
 obj-$(CONFIG_STACKTRACE)        += stacktrace.o
 obj-$(CONFIG_IO_TRAPPED)        += io_trapped.o
 obj-$(CONFIG_KPROBES)           += kprobes.o
-obj-$(CONFIG_GENERIC_GPIO)      += gpio.o
 obj-$(CONFIG_DYNAMIC_FTRACE)    += ftrace.o
 obj-$(CONFIG_FTRACE_SYSCALLS)   += ftrace.o
 obj-$(CONFIG_FUNCTION_GRAPH_TRACER) += ftrace.o
 obj-$(CONFIG_DUMP_CODE)         += disassemble.o
 obj-$(CONFIG_HIBERNATION)       += swsusp.o
 obj-$(CONFIG_DWARF_UNWINDER)    += dwarf.o
+obj-$(CONFIG_PERF_EVENTS)       += perf_event.o perf_callchain.o
 
 obj-$(CONFIG_HAVE_HW_BREAKPOINT)                += hw_breakpoint.o
 obj-$(CONFIG_GENERIC_CLOCKEVENTS_BROADCAST)     += localtimer.o
diff --git a/arch/sh/kernel/asm-offsets.c b/arch/sh/kernel/asm-offsets.c
index d218e808294e..08a2be775b6c 100644
--- a/arch/sh/kernel/asm-offsets.c
+++ b/arch/sh/kernel/asm-offsets.c
@@ -34,5 +34,28 @@ int main(void)
         DEFINE(PBE_NEXT, offsetof(struct pbe, next));
         DEFINE(SWSUSP_ARCH_REGS_SIZE, sizeof(struct swsusp_arch_regs));
 #endif
+
+        DEFINE(SH_SLEEP_MODE, offsetof(struct sh_sleep_data, mode));
+        DEFINE(SH_SLEEP_SF_PRE, offsetof(struct sh_sleep_data, sf_pre));
+        DEFINE(SH_SLEEP_SF_POST, offsetof(struct sh_sleep_data, sf_post));
+        DEFINE(SH_SLEEP_RESUME, offsetof(struct sh_sleep_data, resume));
+        DEFINE(SH_SLEEP_VBR, offsetof(struct sh_sleep_data, vbr));
+        DEFINE(SH_SLEEP_SPC, offsetof(struct sh_sleep_data, spc));
+        DEFINE(SH_SLEEP_SR, offsetof(struct sh_sleep_data, sr));
+        DEFINE(SH_SLEEP_SP, offsetof(struct sh_sleep_data, sp));
+        DEFINE(SH_SLEEP_BASE_ADDR, offsetof(struct sh_sleep_data, addr));
+        DEFINE(SH_SLEEP_BASE_DATA, offsetof(struct sh_sleep_data, data));
+        DEFINE(SH_SLEEP_REG_STBCR, offsetof(struct sh_sleep_regs, stbcr));
+        DEFINE(SH_SLEEP_REG_BAR, offsetof(struct sh_sleep_regs, bar));
+        DEFINE(SH_SLEEP_REG_PTEH, offsetof(struct sh_sleep_regs, pteh));
+        DEFINE(SH_SLEEP_REG_PTEL, offsetof(struct sh_sleep_regs, ptel));
+        DEFINE(SH_SLEEP_REG_TTB, offsetof(struct sh_sleep_regs, ttb));
+        DEFINE(SH_SLEEP_REG_TEA, offsetof(struct sh_sleep_regs, tea));
+        DEFINE(SH_SLEEP_REG_MMUCR, offsetof(struct sh_sleep_regs, mmucr));
+        DEFINE(SH_SLEEP_REG_PTEA, offsetof(struct sh_sleep_regs, ptea));
+        DEFINE(SH_SLEEP_REG_PASCR, offsetof(struct sh_sleep_regs, pascr));
+        DEFINE(SH_SLEEP_REG_IRMCR, offsetof(struct sh_sleep_regs, irmcr));
+        DEFINE(SH_SLEEP_REG_CCR, offsetof(struct sh_sleep_regs, ccr));
+        DEFINE(SH_SLEEP_REG_RAMCR, offsetof(struct sh_sleep_regs, ramcr));
         return 0;
 }
diff --git a/arch/sh/kernel/cpu/Makefile b/arch/sh/kernel/cpu/Makefile
index 3d6b9312dc47..d97c803719ec 100644
--- a/arch/sh/kernel/cpu/Makefile
+++ b/arch/sh/kernel/cpu/Makefile
@@ -15,7 +15,6 @@ obj-$(CONFIG_ARCH_SHMOBILE)	+= shmobile/
 
 # Common interfaces.
 
-obj-$(CONFIG_UBC_WAKEUP)        += ubc.o
 obj-$(CONFIG_SH_ADC)            += adc.o
 obj-$(CONFIG_SH_CLK_CPG)        += clock-cpg.o
 
diff --git a/arch/sh/kernel/cpu/init.c b/arch/sh/kernel/cpu/init.c
index e932ebef4738..89b4b76c0d76 100644
--- a/arch/sh/kernel/cpu/init.c
+++ b/arch/sh/kernel/cpu/init.c
@@ -75,16 +75,11 @@ static void __init expmask_init(void)
         /*
          * Future proofing.
          *
-         * Disable support for slottable sleep instruction
-         * and non-nop instructions in the rte delay slot.
+         * Disable support for slottable sleep instruction, non-nop
+         * instructions in the rte delay slot, and associative writes to
+         * the memory-mapped cache array.
          */
-        expmask &= ~(EXPMASK_RTEDS | EXPMASK_BRDSSLP);
-
-        /*
-         * Enable associative writes to the memory-mapped cache array
-         * until the cache flush ops have been rewritten.
-         */
-        expmask |= EXPMASK_MMCAW;
+        expmask &= ~(EXPMASK_RTEDS | EXPMASK_BRDSSLP | EXPMASK_MMCAW);
 
         __raw_writel(expmask, EXPMASK);
         ctrl_barrier();
@@ -311,12 +306,12 @@ asmlinkage void __init sh_cpu_init(void)
         if (fpu_disabled) {
                 printk("FPU Disabled\n");
                 current_cpu_data.flags &= ~CPU_HAS_FPU;
-                disable_fpu();
         }
 
         /* FPU initialization */
+        disable_fpu();
         if ((current_cpu_data.flags & CPU_HAS_FPU)) {
-                clear_thread_flag(TIF_USEDFPU);
+                current_thread_info()->status &= ~TS_USEDFPU;
                 clear_used_math();
         }
 
@@ -338,17 +333,6 @@ asmlinkage void __init sh_cpu_init(void)
         }
 #endif
 
-        /*
-         * Some brain-damaged loaders decided it would be a good idea to put
-         * the UBC to sleep. This causes some issues when it comes to things
-         * like PTRACE_SINGLESTEP or doing hardware watchpoints in GDB.  So ..
-         * we wake it up and hope that all is well.
-         */
-#ifdef CONFIG_SUPERH32
-        if (raw_smp_processor_id() == 0)
-                ubc_wakeup();
-#endif
-
         speculative_execution_init();
         expmask_init();
 }
diff --git a/arch/sh/kernel/cpu/sh2a/fpu.c b/arch/sh/kernel/cpu/sh2a/fpu.c
index 6df2fb98eb30..d395ce5740e7 100644
--- a/arch/sh/kernel/cpu/sh2a/fpu.c
+++ b/arch/sh/kernel/cpu/sh2a/fpu.c
@@ -25,14 +25,12 @@
 
 /*
  * Save FPU registers onto task structure.
- * Assume called with FPU enabled (SR.FD=0).
  */
 void
-save_fpu(struct task_struct *tsk, struct pt_regs *regs)
+save_fpu(struct task_struct *tsk)
 {
         unsigned long dummy;
 
-        clear_tsk_thread_flag(tsk, TIF_USEDFPU);
         enable_fpu();
         asm volatile("sts.l     fpul, @-%0\n\t"
                      "sts.l     fpscr, @-%0\n\t"
@@ -60,7 +58,6 @@ save_fpu(struct task_struct *tsk, struct pt_regs *regs)
                      : "memory");
 
         disable_fpu();
-        release_fpu(regs);
 }
 
 static void
@@ -598,31 +595,31 @@ BUILD_TRAP_HANDLER(fpu_error)
         struct task_struct *tsk = current;
         TRAP_HANDLER_DECL;
 
-        save_fpu(tsk, regs);
+        __unlazy_fpu(tsk, regs);
         if (ieee_fpe_handler(regs)) {
                 tsk->thread.fpu.hard.fpscr &=
                         ~(FPSCR_CAUSE_MASK | FPSCR_FLAG_MASK);
                 grab_fpu(regs);
                 restore_fpu(tsk);
-                set_tsk_thread_flag(tsk, TIF_USEDFPU);
+                task_thread_info(tsk)->status |= TS_USEDFPU;
                 return;
         }
 
         force_sig(SIGFPE, tsk);
 }
 
-BUILD_TRAP_HANDLER(fpu_state_restore)
+void fpu_state_restore(struct pt_regs *regs)
 {
         struct task_struct *tsk = current;
-        TRAP_HANDLER_DECL;
 
         grab_fpu(regs);
-        if (!user_mode(regs)) {
+        if (unlikely(!user_mode(regs))) {
                 printk(KERN_ERR "BUG: FPU is used in kernel mode.\n");
+                BUG();
                 return;
         }
 
-        if (used_math()) {
+        if (likely(used_math())) {
                 /* Using the FPU again.  */
                 restore_fpu(tsk);
         } else  {
@@ -630,5 +627,13 @@ BUILD_TRAP_HANDLER(fpu_state_restore)
                 fpu_init();
                 set_used_math();
         }
-        set_tsk_thread_flag(tsk, TIF_USEDFPU);
+        task_thread_info(tsk)->status |= TS_USEDFPU;
+        tsk->fpu_counter++;
+}
+
+BUILD_TRAP_HANDLER(fpu_state_restore)
+{
+        TRAP_HANDLER_DECL;
+
+        fpu_state_restore(regs);
 }
diff --git a/arch/sh/kernel/cpu/sh3/entry.S b/arch/sh/kernel/cpu/sh3/entry.S
index bb407ef0b91e..3f7e2a22c7c2 100644
--- a/arch/sh/kernel/cpu/sh3/entry.S
+++ b/arch/sh/kernel/cpu/sh3/entry.S
@@ -297,41 +297,8 @@ ENTRY(vbr_base)
 !
         .balign         256,0,256
 general_exception:
-#ifndef CONFIG_CPU_SUBTYPE_SHX3
         bra     handle_exception
          sts    pr, k3          ! save original pr value in k3
-#else
-        mov.l   1f, k4
-        mov.l   @k4, k4
-
-        ! Is EXPEVT larger than 0x800?
-        mov     #0x8, k0
-        shll8   k0
-        cmp/hs  k0, k4
-        bf      0f
-
-        ! then add 0x580 (k2 is 0xd80 or 0xda0)
-        mov     #0x58, k0
-        shll2   k0
-        shll2   k0
-        add     k0, k4
-0:
-        ! Setup stack and save DSP context (k0 contains original r15 on return)
-        bsr     prepare_stack
-         nop
-
-        ! Save registers / Switch to bank 0
-        mov             k4, k2          ! keep vector in k2
-        mov.l   1f, k4          ! SR bits to clear in k4
-        bsr     save_regs       ! needs original pr value in k3
-         nop
-
-        bra     handle_exception_special
-         nop
-
-        .align  2
-1:      .long   EXPEVT
-#endif
 
 ! prepare_stack()
 ! - roll back gRB
diff --git a/arch/sh/kernel/cpu/sh4/Makefile b/arch/sh/kernel/cpu/sh4/Makefile
index 203b18347b83..3a1dbc709831 100644
--- a/arch/sh/kernel/cpu/sh4/Makefile
+++ b/arch/sh/kernel/cpu/sh4/Makefile
@@ -9,6 +9,11 @@ obj-$(CONFIG_HIBERNATION)		+= $(addprefix ../sh3/, swsusp.o)
 obj-$(CONFIG_SH_FPU)                    += fpu.o softfloat.o
 obj-$(CONFIG_SH_STORE_QUEUES)           += sq.o
 
+# Perf events
+perf-$(CONFIG_CPU_SUBTYPE_SH7750)       := perf_event.o
+perf-$(CONFIG_CPU_SUBTYPE_SH7750S)      := perf_event.o
+perf-$(CONFIG_CPU_SUBTYPE_SH7091)       := perf_event.o
+
 # CPU subtype setup
 obj-$(CONFIG_CPU_SUBTYPE_SH7750)        += setup-sh7750.o
 obj-$(CONFIG_CPU_SUBTYPE_SH7750R)       += setup-sh7750.o
@@ -27,4 +32,5 @@ endif
 # Additional clocks by subtype
 clock-$(CONFIG_CPU_SUBTYPE_SH4_202)     += clock-sh4-202.o
 
-obj-y   += $(clock-y)
+obj-y                                   += $(clock-y)
+obj-$(CONFIG_PERF_EVENTS)               += $(perf-y)
diff --git a/arch/sh/kernel/cpu/sh4/fpu.c b/arch/sh/kernel/cpu/sh4/fpu.c
index e3ea5411da6d..e97857aec8a0 100644
--- a/arch/sh/kernel/cpu/sh4/fpu.c
+++ b/arch/sh/kernel/cpu/sh4/fpu.c
@@ -41,13 +41,11 @@ static unsigned int fpu_exception_flags;
 
 /*
  * Save FPU registers onto task structure.
- * Assume called with FPU enabled (SR.FD=0).
  */
-void save_fpu(struct task_struct *tsk, struct pt_regs *regs)
+void save_fpu(struct task_struct *tsk)
 {
         unsigned long dummy;
 
-        clear_tsk_thread_flag(tsk, TIF_USEDFPU);
         enable_fpu();
         asm volatile ("sts.l    fpul, @-%0\n\t"
                       "sts.l    fpscr, @-%0\n\t"
@@ -92,7 +90,6 @@ void save_fpu(struct task_struct *tsk, struct pt_regs *regs)
                       :"memory");
 
         disable_fpu();
-        release_fpu(regs);
 }
 
 static void restore_fpu(struct task_struct *tsk)
@@ -285,7 +282,6 @@ static int ieee_fpe_handler(struct pt_regs *regs)
                 /* fcnvsd */
                 struct task_struct *tsk = current;
 
-                save_fpu(tsk, regs);
                 if ((tsk->thread.fpu.hard.fpscr & FPSCR_CAUSE_ERROR))
                         /* FPU error */
                         denormal_to_double(&tsk->thread.fpu.hard,
@@ -462,7 +458,7 @@ BUILD_TRAP_HANDLER(fpu_error)
         struct task_struct *tsk = current;
         TRAP_HANDLER_DECL;
 
-        save_fpu(tsk, regs);
+        __unlazy_fpu(tsk, regs);
         fpu_exception_flags = 0;
         if (ieee_fpe_handler(regs)) {
                 tsk->thread.fpu.hard.fpscr &=
@@ -473,7 +469,7 @@ BUILD_TRAP_HANDLER(fpu_error)
                 tsk->thread.fpu.hard.fpscr |= (fpu_exception_flags >> 10);
                 grab_fpu(regs);
                 restore_fpu(tsk);
-                set_tsk_thread_flag(tsk, TIF_USEDFPU);
+                task_thread_info(tsk)->status |= TS_USEDFPU;
                 if ((((tsk->thread.fpu.hard.fpscr & FPSCR_ENABLE_MASK) >> 7) &
                      (fpu_exception_flags >> 2)) == 0) {
                         return;
@@ -483,18 +479,18 @@ BUILD_TRAP_HANDLER(fpu_error)
         force_sig(SIGFPE, tsk);
 }
 
-BUILD_TRAP_HANDLER(fpu_state_restore)
+void fpu_state_restore(struct pt_regs *regs)
 {
         struct task_struct *tsk = current;
-        TRAP_HANDLER_DECL;
 
         grab_fpu(regs);
-        if (!user_mode(regs)) {
+        if (unlikely(!user_mode(regs))) {
                 printk(KERN_ERR "BUG: FPU is used in kernel mode.\n");
+                BUG();
                 return;
         }
 
-        if (used_math()) {
+        if (likely(used_math())) {
                 /* Using the FPU again.  */
                 restore_fpu(tsk);
         } else {
@@ -502,5 +498,13 @@ BUILD_TRAP_HANDLER(fpu_state_restore)
                 fpu_init();
                 set_used_math();
         }
-        set_tsk_thread_flag(tsk, TIF_USEDFPU);
+        task_thread_info(tsk)->status |= TS_USEDFPU;
+        tsk->fpu_counter++;
+}
+
+BUILD_TRAP_HANDLER(fpu_state_restore)
+{
+        TRAP_HANDLER_DECL;
+
+        fpu_state_restore(regs);
 }
diff --git a/arch/sh/kernel/cpu/sh4/perf_event.c b/arch/sh/kernel/cpu/sh4/perf_event.c
new file mode 100644
index 000000000000..7f9ecc9c2d02
--- /dev/null
+++ b/arch/sh/kernel/cpu/sh4/perf_event.c
@@ -0,0 +1,253 @@
+/*
+ * Performance events support for SH7750-style performance counters
+ *
+ *  Copyright (C) 2009  Paul Mundt
+ *
+ * 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/io.h>
+#include <linux/irq.h>
+#include <linux/perf_event.h>
+#include <asm/processor.h>
+
+#define PM_CR_BASE      0xff000084      /* 16-bit */
+#define PM_CTR_BASE     0xff100004      /* 32-bit */
+
+#define PMCR(n)         (PM_CR_BASE + ((n) * 0x04))
+#define PMCTRH(n)       (PM_CTR_BASE + 0x00 + ((n) * 0x08))
+#define PMCTRL(n)       (PM_CTR_BASE + 0x04 + ((n) * 0x08))
+
+#define PMCR_PMM_MASK   0x0000003f
+
+#define PMCR_CLKF       0x00000100
+#define PMCR_PMCLR      0x00002000
+#define PMCR_PMST       0x00004000
+#define PMCR_PMEN       0x00008000
+
+static struct sh_pmu sh7750_pmu;
+
+/*
+ * There are a number of events supported by each counter (33 in total).
+ * Since we have 2 counters, each counter will take the event code as it
+ * corresponds to the PMCR PMM setting. Each counter can be configured
+ * independently.
+ *
+ *      Event Code      Description
+ *      ----------      -----------
+ *
+ *      0x01            Operand read access
+ *      0x02            Operand write access
+ *      0x03            UTLB miss
+ *      0x04            Operand cache read miss
+ *      0x05            Operand cache write miss
+ *      0x06            Instruction fetch (w/ cache)
+ *      0x07            Instruction TLB miss
+ *      0x08            Instruction cache miss
+ *      0x09            All operand accesses
+ *      0x0a            All instruction accesses
+ *      0x0b            OC RAM operand access
+ *      0x0d            On-chip I/O space access
+ *      0x0e            Operand access (r/w)
+ *      0x0f            Operand cache miss (r/w)
+ *      0x10            Branch instruction
+ *      0x11            Branch taken
+ *      0x12            BSR/BSRF/JSR
+ *      0x13            Instruction execution
+ *      0x14            Instruction execution in parallel
+ *      0x15            FPU Instruction execution
+ *      0x16            Interrupt
+ *      0x17            NMI
+ *      0x18            trapa instruction execution
+ *      0x19            UBCA match
+ *      0x1a            UBCB match
+ *      0x21            Instruction cache fill
+ *      0x22            Operand cache fill
+ *      0x23            Elapsed time
+ *      0x24            Pipeline freeze by I-cache miss
+ *      0x25            Pipeline freeze by D-cache miss
+ *      0x27            Pipeline freeze by branch instruction
+ *      0x28            Pipeline freeze by CPU register
+ *      0x29            Pipeline freeze by FPU
+ */
+
+static const int sh7750_general_events[] = {
+        [PERF_COUNT_HW_CPU_CYCLES]              = 0x0023,
+        [PERF_COUNT_HW_INSTRUCTIONS]            = 0x000a,
+        [PERF_COUNT_HW_CACHE_REFERENCES]        = 0x0006,       /* I-cache */
+        [PERF_COUNT_HW_CACHE_MISSES]            = 0x0008,       /* I-cache */
+        [PERF_COUNT_HW_BRANCH_INSTRUCTIONS]     = 0x0010,
+        [PERF_COUNT_HW_BRANCH_MISSES]           = -1,
+        [PERF_COUNT_HW_BUS_CYCLES]              = -1,
+};
+
+#define C(x)    PERF_COUNT_HW_CACHE_##x
+
+static const int sh7750_cache_events
+                        [PERF_COUNT_HW_CACHE_MAX]
+                        [PERF_COUNT_HW_CACHE_OP_MAX]
+                        [PERF_COUNT_HW_CACHE_RESULT_MAX] =
+{
+        [ C(L1D) ] = {
+                [ C(OP_READ) ] = {
+                        [ C(RESULT_ACCESS) ] = 0x0001,
+                        [ C(RESULT_MISS)   ] = 0x0004,
+                },
+                [ C(OP_WRITE) ] = {
+                        [ C(RESULT_ACCESS) ] = 0x0002,
+                        [ C(RESULT_MISS)   ] = 0x0005,
+                },
+                [ C(OP_PREFETCH) ] = {
+                        [ C(RESULT_ACCESS) ] = 0,
+                        [ C(RESULT_MISS)   ] = 0,
+                },
+        },
+
+        [ C(L1I) ] = {
+                [ C(OP_READ) ] = {
+                        [ C(RESULT_ACCESS) ] = 0x0006,
+                        [ C(RESULT_MISS)   ] = 0x0008,
+                },
+                [ C(OP_WRITE) ] = {
+                        [ C(RESULT_ACCESS) ] = -1,
+                        [ C(RESULT_MISS)   ] = -1,
+                },
+                [ C(OP_PREFETCH) ] = {
+                        [ C(RESULT_ACCESS) ] = 0,
+                        [ C(RESULT_MISS)   ] = 0,
+                },
+        },
+
+        [ C(LL) ] = {
+                [ C(OP_READ) ] = {
+                        [ C(RESULT_ACCESS) ] = 0,
+                        [ C(RESULT_MISS)   ] = 0,
+                },
+                [ C(OP_WRITE) ] = {
+                        [ C(RESULT_ACCESS) ] = 0,
+                        [ C(RESULT_MISS)   ] = 0,
+                },
+                [ C(OP_PREFETCH) ] = {
+                        [ C(RESULT_ACCESS) ] = 0,
+                        [ C(RESULT_MISS)   ] = 0,
+                },
+        },
+
+        [ C(DTLB) ] = {
+                [ C(OP_READ) ] = {
+                        [ C(RESULT_ACCESS) ] = 0,
+                        [ C(RESULT_MISS)   ] = 0x0003,
+                },
+                [ C(OP_WRITE) ] = {
+                        [ C(RESULT_ACCESS) ] = 0,
+                        [ C(RESULT_MISS)   ] = 0,
+                },
+                [ C(OP_PREFETCH) ] = {
+                        [ C(RESULT_ACCESS) ] = 0,
+                        [ C(RESULT_MISS)   ] = 0,
+                },
+        },
+
+        [ C(ITLB) ] = {
+                [ C(OP_READ) ] = {
+                        [ C(RESULT_ACCESS) ] = 0,
+                        [ C(RESULT_MISS)   ] = 0x0007,
+                },
+                [ C(OP_WRITE) ] = {
+                        [ C(RESULT_ACCESS) ] = -1,
+                        [ C(RESULT_MISS)   ] = -1,
+                },
+                [ C(OP_PREFETCH) ] = {
+                        [ C(RESULT_ACCESS) ] = -1,
+                        [ C(RESULT_MISS)   ] = -1,
+                },
+        },
+
+        [ C(BPU) ] = {
+                [ C(OP_READ) ] = {
+                        [ C(RESULT_ACCESS) ] = -1,
+                        [ C(RESULT_MISS)   ] = -1,
+                },
+                [ C(OP_WRITE) ] = {
+                        [ C(RESULT_ACCESS) ] = -1,
+                        [ C(RESULT_MISS)   ] = -1,
+                },
+                [ C(OP_PREFETCH) ] = {
+                        [ C(RESULT_ACCESS) ] = -1,
+                        [ C(RESULT_MISS)   ] = -1,
+                },
+        },
+};
+
+static int sh7750_event_map(int event)
+{
+        return sh7750_general_events[event];
+}
+
+static u64 sh7750_pmu_read(int idx)
+{
+        return (u64)((u64)(__raw_readl(PMCTRH(idx)) & 0xffff) << 32) |
+                           __raw_readl(PMCTRL(idx));
+}
+
+static void sh7750_pmu_disable(struct hw_perf_event *hwc, int idx)
+{
+        unsigned int tmp;
+
+        tmp = __raw_readw(PMCR(idx));
+        tmp &= ~(PMCR_PMM_MASK | PMCR_PMEN);
+        __raw_writew(tmp, PMCR(idx));
+}
+
+static void sh7750_pmu_enable(struct hw_perf_event *hwc, int idx)
+{
+        __raw_writew(__raw_readw(PMCR(idx)) | PMCR_PMCLR, PMCR(idx));
+        __raw_writew(hwc->config | PMCR_PMEN | PMCR_PMST, PMCR(idx));
+}
+
+static void sh7750_pmu_disable_all(void)
+{
+        int i;
+
+        for (i = 0; i < sh7750_pmu.num_events; i++)
+                __raw_writew(__raw_readw(PMCR(i)) & ~PMCR_PMEN, PMCR(i));
+}
+
+static void sh7750_pmu_enable_all(void)
+{
+        int i;
+
+        for (i = 0; i < sh7750_pmu.num_events; i++)
+                __raw_writew(__raw_readw(PMCR(i)) | PMCR_PMEN, PMCR(i));
+}
+
+static struct sh_pmu sh7750_pmu = {
+        .name           = "SH7750",
+        .num_events     = 2,
+        .event_map      = sh7750_event_map,
+        .max_events     = ARRAY_SIZE(sh7750_general_events),
+        .raw_event_mask = PMCR_PMM_MASK,
+        .cache_events   = &sh7750_cache_events,
+        .read           = sh7750_pmu_read,
+        .disable        = sh7750_pmu_disable,
+        .enable         = sh7750_pmu_enable,
+        .disable_all    = sh7750_pmu_disable_all,
+        .enable_all     = sh7750_pmu_enable_all,
+};
+
+static int __init sh7750_pmu_init(void)
+{
+        /*
+         * Make sure this CPU actually has perf counters.
+         */
+        if (!(boot_cpu_data.flags & CPU_HAS_PERF_COUNTER)) {
+                pr_notice("HW perf events unsupported, software events only.\n");
+                return -ENODEV;
+        }
+
+        return register_sh_pmu(&sh7750_pmu);
+}
+arch_initcall(sh7750_pmu_init);
diff --git a/arch/sh/kernel/cpu/sh4a/Makefile b/arch/sh/kernel/cpu/sh4a/Makefile
index 490d5dc9e372..33bab477d2e2 100644
--- a/arch/sh/kernel/cpu/sh4a/Makefile
+++ b/arch/sh/kernel/cpu/sh4a/Makefile
@@ -44,3 +44,4 @@ pinmux-$(CONFIG_CPU_SUBTYPE_SH7786)	:= pinmux-sh7786.o
 obj-y                           += $(clock-y)
 obj-$(CONFIG_SMP)               += $(smp-y)
 obj-$(CONFIG_GENERIC_GPIO)      += $(pinmux-y)
+obj-$(CONFIG_PERF_EVENTS)       += perf_event.o
diff --git a/arch/sh/kernel/cpu/sh4a/clock-sh7724.c b/arch/sh/kernel/cpu/sh4a/clock-sh7724.c
index dfe9192be63e..9db743802f06 100644
--- a/arch/sh/kernel/cpu/sh4a/clock-sh7724.c
+++ b/arch/sh/kernel/cpu/sh4a/clock-sh7724.c
@@ -152,7 +152,7 @@ struct clk div6_clks[] = {
         SH_CLK_DIV6("fsia_clk", &div3_clk, FCLKACR, 0),
         SH_CLK_DIV6("fsib_clk", &div3_clk, FCLKBCR, 0),
         SH_CLK_DIV6("irda_clk", &div3_clk, IRDACLKCR, 0),
-        SH_CLK_DIV6("spu_clk", &div3_clk, SPUCLKCR, 0),
+        SH_CLK_DIV6("spu_clk", &div3_clk, SPUCLKCR, CLK_ENABLE_ON_INIT),
 };
 
 #define R_CLK (&r_clk)
diff --git a/arch/sh/kernel/cpu/sh4a/perf_event.c b/arch/sh/kernel/cpu/sh4a/perf_event.c
new file mode 100644
index 000000000000..eddc21973fa1
--- /dev/null
+++ b/arch/sh/kernel/cpu/sh4a/perf_event.c
@@ -0,0 +1,269 @@
+/*
+ * Performance events support for SH-4A performance counters
+ *
+ *  Copyright (C) 2009  Paul Mundt
+ *
+ * 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/io.h>
+#include <linux/irq.h>
+#include <linux/perf_event.h>
+#include <asm/processor.h>
+
+#define PPC_CCBR(idx)   (0xff200800 + (sizeof(u32) * idx))
+#define PPC_PMCTR(idx)  (0xfc100000 + (sizeof(u32) * idx))
+
+#define CCBR_CIT_MASK   (0x7ff << 6)
+#define CCBR_DUC        (1 << 3)
+#define CCBR_CMDS       (1 << 1)
+#define CCBR_PPCE       (1 << 0)
+
+#define PPC_PMCAT       0xfc100080
+
+#define PMCAT_OVF3      (1 << 27)
+#define PMCAT_CNN3      (1 << 26)
+#define PMCAT_CLR3      (1 << 25)
+#define PMCAT_OVF2      (1 << 19)
+#define PMCAT_CLR2      (1 << 17)
+#define PMCAT_OVF1      (1 << 11)
+#define PMCAT_CNN1      (1 << 10)
+#define PMCAT_CLR1      (1 << 9)
+#define PMCAT_OVF0      (1 << 3)
+#define PMCAT_CLR0      (1 << 1)
+
+static struct sh_pmu sh4a_pmu;
+
+/*
+ * Supported raw event codes:
+ *
+ *      Event Code      Description
+ *      ----------      -----------
+ *
+ *      0x0000          number of elapsed cycles
+ *      0x0200          number of elapsed cycles in privileged mode
+ *      0x0280          number of elapsed cycles while SR.BL is asserted
+ *      0x0202          instruction execution
+ *      0x0203          instruction execution in parallel
+ *      0x0204          number of unconditional branches
+ *      0x0208          number of exceptions
+ *      0x0209          number of interrupts
+ *      0x0220          UTLB miss caused by instruction fetch
+ *      0x0222          UTLB miss caused by operand access
+ *      0x02a0          number of ITLB misses
+ *      0x0028          number of accesses to instruction memories
+ *      0x0029          number of accesses to instruction cache
+ *      0x002a          instruction cache miss
+ *      0x022e          number of access to instruction X/Y memory
+ *      0x0030          number of reads to operand memories
+ *      0x0038          number of writes to operand memories
+ *      0x0031          number of operand cache read accesses
+ *      0x0039          number of operand cache write accesses
+ *      0x0032          operand cache read miss
+ *      0x003a          operand cache write miss
+ *      0x0236          number of reads to operand X/Y memory
+ *      0x023e          number of writes to operand X/Y memory
+ *      0x0237          number of reads to operand U memory
+ *      0x023f          number of writes to operand U memory
+ *      0x0337          number of U memory read buffer misses
+ *      0x02b4          number of wait cycles due to operand read access
+ *      0x02bc          number of wait cycles due to operand write access
+ *      0x0033          number of wait cycles due to operand cache read miss
+ *      0x003b          number of wait cycles due to operand cache write miss
+ */
+
+/*
+ * Special reserved bits used by hardware emulators, read values will
+ * vary, but writes must always be 0.
+ */
+#define PMCAT_EMU_CLR_MASK      ((1 << 24) | (1 << 16) | (1 << 8) | (1 << 0))
+
+static const int sh4a_general_events[] = {
+        [PERF_COUNT_HW_CPU_CYCLES]              = 0x0000,
+        [PERF_COUNT_HW_INSTRUCTIONS]            = 0x0202,
+        [PERF_COUNT_HW_CACHE_REFERENCES]        = 0x0029,       /* I-cache */
+        [PERF_COUNT_HW_CACHE_MISSES]            = 0x002a,       /* I-cache */
+        [PERF_COUNT_HW_BRANCH_INSTRUCTIONS]     = 0x0204,
+        [PERF_COUNT_HW_BRANCH_MISSES]           = -1,
+        [PERF_COUNT_HW_BUS_CYCLES]              = -1,
+};
+
+#define C(x)    PERF_COUNT_HW_CACHE_##x
+
+static const int sh4a_cache_events
+                        [PERF_COUNT_HW_CACHE_MAX]
+                        [PERF_COUNT_HW_CACHE_OP_MAX]
+                        [PERF_COUNT_HW_CACHE_RESULT_MAX] =
+{
+        [ C(L1D) ] = {
+                [ C(OP_READ) ] = {
+                        [ C(RESULT_ACCESS) ] = 0x0031,
+                        [ C(RESULT_MISS)   ] = 0x0032,
+                },
+                [ C(OP_WRITE) ] = {
+                        [ C(RESULT_ACCESS) ] = 0x0039,
+                        [ C(RESULT_MISS)   ] = 0x003a,
+                },
+                [ C(OP_PREFETCH) ] = {
+                        [ C(RESULT_ACCESS) ] = 0,
+                        [ C(RESULT_MISS)   ] = 0,
+                },
+        },
+
+        [ C(L1I) ] = {
+                [ C(OP_READ) ] = {
+                        [ C(RESULT_ACCESS) ] = 0x0029,
+                        [ C(RESULT_MISS)   ] = 0x002a,
+                },
+                [ C(OP_WRITE) ] = {
+                        [ C(RESULT_ACCESS) ] = -1,
+                        [ C(RESULT_MISS)   ] = -1,
+                },
+                [ C(OP_PREFETCH) ] = {
+                        [ C(RESULT_ACCESS) ] = 0,
+                        [ C(RESULT_MISS)   ] = 0,
+                },
+        },
+
+        [ C(LL) ] = {
+                [ C(OP_READ) ] = {
+                        [ C(RESULT_ACCESS) ] = 0x0030,
+                        [ C(RESULT_MISS)   ] = 0,
+                },
+                [ C(OP_WRITE) ] = {
+                        [ C(RESULT_ACCESS) ] = 0x0038,
+                        [ C(RESULT_MISS)   ] = 0,
+                },
+                [ C(OP_PREFETCH) ] = {
+                        [ C(RESULT_ACCESS) ] = 0,
+                        [ C(RESULT_MISS)   ] = 0,
+                },
+        },
+
+        [ C(DTLB) ] = {
+                [ C(OP_READ) ] = {
+                        [ C(RESULT_ACCESS) ] = 0x0222,
+                        [ C(RESULT_MISS)   ] = 0x0220,
+                },
+                [ C(OP_WRITE) ] = {
+                        [ C(RESULT_ACCESS) ] = 0,
+                        [ C(RESULT_MISS)   ] = 0,
+                },
+                [ C(OP_PREFETCH) ] = {
+                        [ C(RESULT_ACCESS) ] = 0,
+                        [ C(RESULT_MISS)   ] = 0,
+                },
+        },
+
+        [ C(ITLB) ] = {
+                [ C(OP_READ) ] = {
+                        [ C(RESULT_ACCESS) ] = 0,
+                        [ C(RESULT_MISS)   ] = 0x02a0,
+                },
+                [ C(OP_WRITE) ] = {
+                        [ C(RESULT_ACCESS) ] = -1,
+                        [ C(RESULT_MISS)   ] = -1,
+                },
+                [ C(OP_PREFETCH) ] = {
+                        [ C(RESULT_ACCESS) ] = -1,
+                        [ C(RESULT_MISS)   ] = -1,
+                },
+        },
+
+        [ C(BPU) ] = {
+                [ C(OP_READ) ] = {
+                        [ C(RESULT_ACCESS) ] = -1,
+                        [ C(RESULT_MISS)   ] = -1,
+                },
+                [ C(OP_WRITE) ] = {
+                        [ C(RESULT_ACCESS) ] = -1,
+                        [ C(RESULT_MISS)   ] = -1,
+                },
+                [ C(OP_PREFETCH) ] = {
+                        [ C(RESULT_ACCESS) ] = -1,
+                        [ C(RESULT_MISS)   ] = -1,
+                },
+        },
+};
+
+static int sh4a_event_map(int event)
+{
+        return sh4a_general_events[event];
+}
+
+static u64 sh4a_pmu_read(int idx)
+{
+        return __raw_readl(PPC_PMCTR(idx));
+}
+
+static void sh4a_pmu_disable(struct hw_perf_event *hwc, int idx)
+{
+        unsigned int tmp;
+
+        tmp = __raw_readl(PPC_CCBR(idx));
+        tmp &= ~(CCBR_CIT_MASK | CCBR_DUC);
+        __raw_writel(tmp, PPC_CCBR(idx));
+}
+
+static void sh4a_pmu_enable(struct hw_perf_event *hwc, int idx)
+{
+        unsigned int tmp;
+
+        tmp = __raw_readl(PPC_PMCAT);
+        tmp &= ~PMCAT_EMU_CLR_MASK;
+        tmp |= idx ? PMCAT_CLR1 : PMCAT_CLR0;
+        __raw_writel(tmp, PPC_PMCAT);
+
+        tmp = __raw_readl(PPC_CCBR(idx));
+        tmp |= (hwc->config << 6) | CCBR_CMDS | CCBR_PPCE;
+        __raw_writel(tmp, PPC_CCBR(idx));
+
+        __raw_writel(__raw_readl(PPC_CCBR(idx)) | CCBR_DUC, PPC_CCBR(idx));
+}
+
+static void sh4a_pmu_disable_all(void)
+{
+        int i;
+
+        for (i = 0; i < sh4a_pmu.num_events; i++)
+                __raw_writel(__raw_readl(PPC_CCBR(i)) & ~CCBR_DUC, PPC_CCBR(i));
+}
+
+static void sh4a_pmu_enable_all(void)
+{
+        int i;
+
+        for (i = 0; i < sh4a_pmu.num_events; i++)
+                __raw_writel(__raw_readl(PPC_CCBR(i)) | CCBR_DUC, PPC_CCBR(i));
+}
+
+static struct sh_pmu sh4a_pmu = {
+        .name           = "SH-4A",
+        .num_events     = 2,
+        .event_map      = sh4a_event_map,
+        .max_events     = ARRAY_SIZE(sh4a_general_events),
+        .raw_event_mask = 0x3ff,
+        .cache_events   = &sh4a_cache_events,
+        .read           = sh4a_pmu_read,
+        .disable        = sh4a_pmu_disable,
+        .enable         = sh4a_pmu_enable,
+        .disable_all    = sh4a_pmu_disable_all,
+        .enable_all     = sh4a_pmu_enable_all,
+};
+
+static int __init sh4a_pmu_init(void)
+{
+        /*
+         * Make sure this CPU actually has perf counters.
+         */
+        if (!(boot_cpu_data.flags & CPU_HAS_PERF_COUNTER)) {
+                pr_notice("HW perf events unsupported, software events only.\n");
+                return -ENODEV;
+        }
+
+        return register_sh_pmu(&sh4a_pmu);
+}
+arch_initcall(sh4a_pmu_init);
diff --git a/arch/sh/kernel/cpu/sh4a/setup-sh7724.c b/arch/sh/kernel/cpu/sh4a/setup-sh7724.c
index f3851fd757ec..845e89c936e7 100644
--- a/arch/sh/kernel/cpu/sh4a/setup-sh7724.c
+++ b/arch/sh/kernel/cpu/sh4a/setup-sh7724.c
@@ -20,6 +20,8 @@
 #include <linux/uio_driver.h>
 #include <linux/sh_timer.h>
 #include <linux/io.h>
+#include <linux/notifier.h>
+#include <asm/suspend.h>
 #include <asm/clock.h>
 #include <asm/mmzone.h>
 #include <cpu/sh7724.h>
@@ -202,7 +204,7 @@ static struct resource veu0_resources[] = {
         [0] = {
                 .name   = "VEU3F0",
                 .start  = 0xfe920000,
-                .end    = 0xfe9200cb - 1,
+                .end    = 0xfe9200cb,
                 .flags  = IORESOURCE_MEM,
         },
         [1] = {
@@ -234,7 +236,7 @@ static struct resource veu1_resources[] = {
         [0] = {
                 .name   = "VEU3F1",
                 .start  = 0xfe924000,
-                .end    = 0xfe9240cb - 1,
+                .end    = 0xfe9240cb,
                 .flags  = IORESOURCE_MEM,
         },
         [1] = {
@@ -523,6 +525,70 @@ static struct platform_device jpu_device = {
         },
 };
 
+/* SPU2DSP0 */
+static struct uio_info spu0_platform_data = {
+        .name = "SPU2DSP0",
+        .version = "0",
+        .irq = 86,
+};
+
+static struct resource spu0_resources[] = {
+        [0] = {
+                .name   = "SPU2DSP0",
+                .start  = 0xFE200000,
+                .end    = 0xFE2FFFFF,
+                .flags  = IORESOURCE_MEM,
+        },
+        [1] = {
+                /* place holder for contiguous memory */
+        },
+};
+
+static struct platform_device spu0_device = {
+        .name           = "uio_pdrv_genirq",
+        .id             = 4,
+        .dev = {
+                .platform_data  = &spu0_platform_data,
+        },
+        .resource       = spu0_resources,
+        .num_resources  = ARRAY_SIZE(spu0_resources),
+        .archdata = {
+                .hwblk_id = HWBLK_SPU,
+        },
+};
+
+/* SPU2DSP1 */
+static struct uio_info spu1_platform_data = {
+        .name = "SPU2DSP1",
+        .version = "0",
+        .irq = 87,
+};
+
+static struct resource spu1_resources[] = {
+        [0] = {
+                .name   = "SPU2DSP1",
+                .start  = 0xFE300000,
+                .end    = 0xFE3FFFFF,
+                .flags  = IORESOURCE_MEM,
+        },
+        [1] = {
+                /* place holder for contiguous memory */
+        },
+};
+
+static struct platform_device spu1_device = {
+        .name           = "uio_pdrv_genirq",
+        .id             = 5,
+        .dev = {
+                .platform_data  = &spu1_platform_data,
+        },
+        .resource       = spu1_resources,
+        .num_resources  = ARRAY_SIZE(spu1_resources),
+        .archdata = {
+                .hwblk_id = HWBLK_SPU,
+        },
+};
+
 static struct platform_device *sh7724_devices[] __initdata = {
         &cmt_device,
         &tmu0_device,
@@ -539,6 +605,8 @@ static struct platform_device *sh7724_devices[] __initdata = {
         &veu0_device,
         &veu1_device,
         &jpu_device,
+        &spu0_device,
+        &spu1_device,
 };
 
 static int __init sh7724_devices_setup(void)
@@ -547,6 +615,8 @@ static int __init sh7724_devices_setup(void)
         platform_resource_setup_memory(&veu0_device, "veu0", 2 << 20);
         platform_resource_setup_memory(&veu1_device, "veu1", 2 << 20);
         platform_resource_setup_memory(&jpu_device,  "jpu",  2 << 20);
+        platform_resource_setup_memory(&spu0_device, "spu0", 2 << 20);
+        platform_resource_setup_memory(&spu1_device, "spu1", 2 << 20);
 
         return platform_add_devices(sh7724_devices,
                                     ARRAY_SIZE(sh7724_devices));
@@ -827,3 +897,193 @@ void __init plat_irq_setup(void)
 {
         register_intc_controller(&intc_desc);
 }
+
+static struct {
+        /* BSC */
+        unsigned long mmselr;
+        unsigned long cs0bcr;
+        unsigned long cs4bcr;
+        unsigned long cs5abcr;
+        unsigned long cs5bbcr;
+        unsigned long cs6abcr;
+        unsigned long cs6bbcr;
+        unsigned long cs4wcr;
+        unsigned long cs5awcr;
+        unsigned long cs5bwcr;
+        unsigned long cs6awcr;
+        unsigned long cs6bwcr;
+        /* INTC */
+        unsigned short ipra;
+        unsigned short iprb;
+        unsigned short iprc;
+        unsigned short iprd;
+        unsigned short ipre;
+        unsigned short iprf;
+        unsigned short iprg;
+        unsigned short iprh;
+        unsigned short ipri;
+        unsigned short iprj;
+        unsigned short iprk;
+        unsigned short iprl;
+        unsigned char imr0;
+        unsigned char imr1;
+        unsigned char imr2;
+        unsigned char imr3;
+        unsigned char imr4;
+        unsigned char imr5;
+        unsigned char imr6;
+        unsigned char imr7;
+        unsigned char imr8;
+        unsigned char imr9;
+        unsigned char imr10;
+        unsigned char imr11;
+        unsigned char imr12;
+        /* RWDT */
+        unsigned short rwtcnt;
+        unsigned short rwtcsr;
+        /* CPG */
+        unsigned long irdaclk;
+        unsigned long spuclk;
+} sh7724_rstandby_state;
+
+static int sh7724_pre_sleep_notifier_call(struct notifier_block *nb,
+                                          unsigned long flags, void *unused)
+{
+        if (!(flags & SUSP_SH_RSTANDBY))
+                return NOTIFY_DONE;
+
+        /* BCR */
+        sh7724_rstandby_state.mmselr = __raw_readl(0xff800020); /* MMSELR */
+        sh7724_rstandby_state.mmselr |= 0xa5a50000;
+        sh7724_rstandby_state.cs0bcr = __raw_readl(0xfec10004); /* CS0BCR */
+        sh7724_rstandby_state.cs4bcr = __raw_readl(0xfec10010); /* CS4BCR */
+        sh7724_rstandby_state.cs5abcr = __raw_readl(0xfec10014); /* CS5ABCR */
+        sh7724_rstandby_state.cs5bbcr = __raw_readl(0xfec10018); /* CS5BBCR */
+        sh7724_rstandby_state.cs6abcr = __raw_readl(0xfec1001c); /* CS6ABCR */
+        sh7724_rstandby_state.cs6bbcr = __raw_readl(0xfec10020); /* CS6BBCR */
+        sh7724_rstandby_state.cs4wcr = __raw_readl(0xfec10030); /* CS4WCR */
+        sh7724_rstandby_state.cs5awcr = __raw_readl(0xfec10034); /* CS5AWCR */
+        sh7724_rstandby_state.cs5bwcr = __raw_readl(0xfec10038); /* CS5BWCR */
+        sh7724_rstandby_state.cs6awcr = __raw_readl(0xfec1003c); /* CS6AWCR */
+        sh7724_rstandby_state.cs6bwcr = __raw_readl(0xfec10040); /* CS6BWCR */
+
+        /* INTC */
+        sh7724_rstandby_state.ipra = __raw_readw(0xa4080000); /* IPRA */
+        sh7724_rstandby_state.iprb = __raw_readw(0xa4080004); /* IPRB */
+        sh7724_rstandby_state.iprc = __raw_readw(0xa4080008); /* IPRC */
+        sh7724_rstandby_state.iprd = __raw_readw(0xa408000c); /* IPRD */
+        sh7724_rstandby_state.ipre = __raw_readw(0xa4080010); /* IPRE */
+        sh7724_rstandby_state.iprf = __raw_readw(0xa4080014); /* IPRF */
+        sh7724_rstandby_state.iprg = __raw_readw(0xa4080018); /* IPRG */
+        sh7724_rstandby_state.iprh = __raw_readw(0xa408001c); /* IPRH */
+        sh7724_rstandby_state.ipri = __raw_readw(0xa4080020); /* IPRI */
+        sh7724_rstandby_state.iprj = __raw_readw(0xa4080024); /* IPRJ */
+        sh7724_rstandby_state.iprk = __raw_readw(0xa4080028); /* IPRK */
+        sh7724_rstandby_state.iprl = __raw_readw(0xa408002c); /* IPRL */
+        sh7724_rstandby_state.imr0 = __raw_readb(0xa4080080); /* IMR0 */
+        sh7724_rstandby_state.imr1 = __raw_readb(0xa4080084); /* IMR1 */
+        sh7724_rstandby_state.imr2 = __raw_readb(0xa4080088); /* IMR2 */
+        sh7724_rstandby_state.imr3 = __raw_readb(0xa408008c); /* IMR3 */
+        sh7724_rstandby_state.imr4 = __raw_readb(0xa4080090); /* IMR4 */
+        sh7724_rstandby_state.imr5 = __raw_readb(0xa4080094); /* IMR5 */
+        sh7724_rstandby_state.imr6 = __raw_readb(0xa4080098); /* IMR6 */
+        sh7724_rstandby_state.imr7 = __raw_readb(0xa408009c); /* IMR7 */
+        sh7724_rstandby_state.imr8 = __raw_readb(0xa40800a0); /* IMR8 */
+        sh7724_rstandby_state.imr9 = __raw_readb(0xa40800a4); /* IMR9 */
+        sh7724_rstandby_state.imr10 = __raw_readb(0xa40800a8); /* IMR10 */
+        sh7724_rstandby_state.imr11 = __raw_readb(0xa40800ac); /* IMR11 */
+        sh7724_rstandby_state.imr12 = __raw_readb(0xa40800b0); /* IMR12 */
+
+        /* RWDT */
+        sh7724_rstandby_state.rwtcnt = __raw_readb(0xa4520000); /* RWTCNT */
+        sh7724_rstandby_state.rwtcnt |= 0x5a00;
+        sh7724_rstandby_state.rwtcsr = __raw_readb(0xa4520004); /* RWTCSR */
+        sh7724_rstandby_state.rwtcsr |= 0xa500;
+        __raw_writew(sh7724_rstandby_state.rwtcsr & 0x07, 0xa4520004);
+
+        /* CPG */
+        sh7724_rstandby_state.irdaclk = __raw_readl(0xa4150018); /* IRDACLKCR */
+        sh7724_rstandby_state.spuclk = __raw_readl(0xa415003c); /* SPUCLKCR */
+
+        return NOTIFY_DONE;
+}
+
+static int sh7724_post_sleep_notifier_call(struct notifier_block *nb,
+                                           unsigned long flags, void *unused)
+{
+        if (!(flags & SUSP_SH_RSTANDBY))
+                return NOTIFY_DONE;
+
+        /* BCR */
+        __raw_writel(sh7724_rstandby_state.mmselr, 0xff800020); /* MMSELR */
+        __raw_writel(sh7724_rstandby_state.cs0bcr, 0xfec10004); /* CS0BCR */
+        __raw_writel(sh7724_rstandby_state.cs4bcr, 0xfec10010); /* CS4BCR */
+        __raw_writel(sh7724_rstandby_state.cs5abcr, 0xfec10014); /* CS5ABCR */
+        __raw_writel(sh7724_rstandby_state.cs5bbcr, 0xfec10018); /* CS5BBCR */
+        __raw_writel(sh7724_rstandby_state.cs6abcr, 0xfec1001c); /* CS6ABCR */
+        __raw_writel(sh7724_rstandby_state.cs6bbcr, 0xfec10020); /* CS6BBCR */
+        __raw_writel(sh7724_rstandby_state.cs4wcr, 0xfec10030); /* CS4WCR */
+        __raw_writel(sh7724_rstandby_state.cs5awcr, 0xfec10034); /* CS5AWCR */
+        __raw_writel(sh7724_rstandby_state.cs5bwcr, 0xfec10038); /* CS5BWCR */
+        __raw_writel(sh7724_rstandby_state.cs6awcr, 0xfec1003c); /* CS6AWCR */
+        __raw_writel(sh7724_rstandby_state.cs6bwcr, 0xfec10040); /* CS6BWCR */
+
+        /* INTC */
+        __raw_writew(sh7724_rstandby_state.ipra, 0xa4080000); /* IPRA */
+        __raw_writew(sh7724_rstandby_state.iprb, 0xa4080004); /* IPRB */
+        __raw_writew(sh7724_rstandby_state.iprc, 0xa4080008); /* IPRC */
+        __raw_writew(sh7724_rstandby_state.iprd, 0xa408000c); /* IPRD */
+        __raw_writew(sh7724_rstandby_state.ipre, 0xa4080010); /* IPRE */
+        __raw_writew(sh7724_rstandby_state.iprf, 0xa4080014); /* IPRF */
+        __raw_writew(sh7724_rstandby_state.iprg, 0xa4080018); /* IPRG */
+        __raw_writew(sh7724_rstandby_state.iprh, 0xa408001c); /* IPRH */
+        __raw_writew(sh7724_rstandby_state.ipri, 0xa4080020); /* IPRI */
+        __raw_writew(sh7724_rstandby_state.iprj, 0xa4080024); /* IPRJ */
+        __raw_writew(sh7724_rstandby_state.iprk, 0xa4080028); /* IPRK */
+        __raw_writew(sh7724_rstandby_state.iprl, 0xa408002c); /* IPRL */
+        __raw_writeb(sh7724_rstandby_state.imr0, 0xa4080080); /* IMR0 */
+        __raw_writeb(sh7724_rstandby_state.imr1, 0xa4080084); /* IMR1 */
+        __raw_writeb(sh7724_rstandby_state.imr2, 0xa4080088); /* IMR2 */
+        __raw_writeb(sh7724_rstandby_state.imr3, 0xa408008c); /* IMR3 */
+        __raw_writeb(sh7724_rstandby_state.imr4, 0xa4080090); /* IMR4 */
+        __raw_writeb(sh7724_rstandby_state.imr5, 0xa4080094); /* IMR5 */
+        __raw_writeb(sh7724_rstandby_state.imr6, 0xa4080098); /* IMR6 */
+        __raw_writeb(sh7724_rstandby_state.imr7, 0xa408009c); /* IMR7 */
+        __raw_writeb(sh7724_rstandby_state.imr8, 0xa40800a0); /* IMR8 */
+        __raw_writeb(sh7724_rstandby_state.imr9, 0xa40800a4); /* IMR9 */
+        __raw_writeb(sh7724_rstandby_state.imr10, 0xa40800a8); /* IMR10 */
+        __raw_writeb(sh7724_rstandby_state.imr11, 0xa40800ac); /* IMR11 */
+        __raw_writeb(sh7724_rstandby_state.imr12, 0xa40800b0); /* IMR12 */
+
+        /* RWDT */
+        __raw_writew(sh7724_rstandby_state.rwtcnt, 0xa4520000); /* RWTCNT */
+        __raw_writew(sh7724_rstandby_state.rwtcsr, 0xa4520004); /* RWTCSR */
+
+        /* CPG */
+        __raw_writel(sh7724_rstandby_state.irdaclk, 0xa4150018); /* IRDACLKCR */
+        __raw_writel(sh7724_rstandby_state.spuclk, 0xa415003c); /* SPUCLKCR */
+
+        return NOTIFY_DONE;
+}
+
+static struct notifier_block sh7724_pre_sleep_notifier = {
+        .notifier_call = sh7724_pre_sleep_notifier_call,
+        .priority = SH_MOBILE_PRE(SH_MOBILE_SLEEP_CPU),
+};
+
+static struct notifier_block sh7724_post_sleep_notifier = {
+        .notifier_call = sh7724_post_sleep_notifier_call,
+        .priority = SH_MOBILE_POST(SH_MOBILE_SLEEP_CPU),
+};
+
+static int __init sh7724_sleep_setup(void)
+{
+        atomic_notifier_chain_register(&sh_mobile_pre_sleep_notifier_list,
+                                       &sh7724_pre_sleep_notifier);
+
+        atomic_notifier_chain_register(&sh_mobile_post_sleep_notifier_list,
+                                       &sh7724_post_sleep_notifier);
+        return 0;
+}
+arch_initcall(sh7724_sleep_setup);
+
diff --git a/arch/sh/kernel/cpu/sh4a/setup-shx3.c b/arch/sh/kernel/cpu/sh4a/setup-shx3.c
index e848443deeb9..c7ba9166e18a 100644
--- a/arch/sh/kernel/cpu/sh4a/setup-shx3.c
+++ b/arch/sh/kernel/cpu/sh4a/setup-shx3.c
@@ -15,6 +15,15 @@
 #include <linux/sh_timer.h>
 #include <asm/mmzone.h>
 
+/*
+ * This intentionally only registers SCIF ports 0, 1, and 3. SCIF 2
+ * INTEVT values overlap with the FPU EXPEVT ones, requiring special
+ * demuxing in the exception dispatch path.
+ *
+ * As this overlap is something that never should have made it in to
+ * silicon in the first place, we just refuse to deal with the port at
+ * all rather than adding infrastructure to hack around it.
+ */
 static struct plat_sci_port sci_platform_data[] = {
         {
                 .mapbase        = 0xffc30000,
@@ -27,11 +36,6 @@ static struct plat_sci_port sci_platform_data[] = {
                 .type           = PORT_SCIF,
                 .irqs           = { 44, 45, 47, 46 },
         }, {
-                .mapbase        = 0xffc50000,
-                .flags          = UPF_BOOT_AUTOCONF,
-                .type           = PORT_SCIF,
-                .irqs           = { 48, 49, 51, 50 },
-        }, {
                 .mapbase        = 0xffc60000,
                 .flags          = UPF_BOOT_AUTOCONF,
                 .type           = PORT_SCIF,
@@ -268,7 +272,11 @@ enum {
         UNUSED = 0,
 
         /* interrupt sources */
-        IRL, IRQ0, IRQ1, IRQ2, IRQ3,
+        IRL_LLLL, IRL_LLLH, IRL_LLHL, IRL_LLHH,
+        IRL_LHLL, IRL_LHLH, IRL_LHHL, IRL_LHHH,
+        IRL_HLLL, IRL_HLLH, IRL_HLHL, IRL_HLHH,
+        IRL_HHLL, IRL_HHLH, IRL_HHHL,
+        IRQ0, IRQ1, IRQ2, IRQ3,
         HUDII,
         TMU0, TMU1, TMU2, TMU3, TMU4, TMU5,
         PCII0, PCII1, PCII2, PCII3, PCII4,
@@ -291,7 +299,7 @@ enum {
         INTICI4, INTICI5, INTICI6, INTICI7,
 
         /* interrupt groups */
-        PCII56789, SCIF0, SCIF1, SCIF2, SCIF3,
+        IRL, PCII56789, SCIF0, SCIF1, SCIF2, SCIF3,
         DMAC0, DMAC1,
 };
 
@@ -309,8 +317,6 @@ static struct intc_vect vectors[] __initdata = {
         INTC_VECT(SCIF0_BRI, 0x740), INTC_VECT(SCIF0_TXI, 0x760),
         INTC_VECT(SCIF1_ERI, 0x780), INTC_VECT(SCIF1_RXI, 0x7a0),
         INTC_VECT(SCIF1_BRI, 0x7c0), INTC_VECT(SCIF1_TXI, 0x7e0),
-        INTC_VECT(SCIF2_ERI, 0x800), INTC_VECT(SCIF2_RXI, 0x820),
-        INTC_VECT(SCIF2_BRI, 0x840), INTC_VECT(SCIF2_TXI, 0x860),
         INTC_VECT(SCIF3_ERI, 0x880), INTC_VECT(SCIF3_RXI, 0x8a0),
         INTC_VECT(SCIF3_BRI, 0x8c0), INTC_VECT(SCIF3_TXI, 0x8e0),
         INTC_VECT(DMAC0_DMINT0, 0x900), INTC_VECT(DMAC0_DMINT1, 0x920),
@@ -344,10 +350,13 @@ static struct intc_vect vectors[] __initdata = {
 };
 
 static struct intc_group groups[] __initdata = {
+        INTC_GROUP(IRL, IRL_LLLL, IRL_LLLH, IRL_LLHL, IRL_LLHH,
+                   IRL_LHLL, IRL_LHLH, IRL_LHHL, IRL_LHHH,
+                   IRL_HLLL, IRL_HLLH, IRL_HLHL, IRL_HLHH,
+                   IRL_HHLL, IRL_HHLH, IRL_HHHL),
         INTC_GROUP(PCII56789, PCII5, PCII6, PCII7, PCII8, PCII9),
         INTC_GROUP(SCIF0, SCIF0_ERI, SCIF0_RXI, SCIF0_BRI, SCIF0_TXI),
         INTC_GROUP(SCIF1, SCIF1_ERI, SCIF1_RXI, SCIF1_BRI, SCIF1_TXI),
-        INTC_GROUP(SCIF2, SCIF2_ERI, SCIF2_RXI, SCIF2_BRI, SCIF2_TXI),
         INTC_GROUP(SCIF3, SCIF3_ERI, SCIF3_RXI, SCIF3_BRI, SCIF3_TXI),
         INTC_GROUP(DMAC0, DMAC0_DMINT0, DMAC0_DMINT1, DMAC0_DMINT2,
                    DMAC0_DMINT3, DMAC0_DMINT4, DMAC0_DMINT5, DMAC0_DMAE),
@@ -419,14 +428,14 @@ static DECLARE_INTC_DESC(intc_desc_irq, "shx3-irq", vectors_irq, groups,
 
 /* External interrupt pins in IRL mode */
 static struct intc_vect vectors_irl[] __initdata = {
-        INTC_VECT(IRL, 0x200), INTC_VECT(IRL, 0x220),
-        INTC_VECT(IRL, 0x240), INTC_VECT(IRL, 0x260),
-        INTC_VECT(IRL, 0x280), INTC_VECT(IRL, 0x2a0),
-        INTC_VECT(IRL, 0x2c0), INTC_VECT(IRL, 0x2e0),
-        INTC_VECT(IRL, 0x300), INTC_VECT(IRL, 0x320),
-        INTC_VECT(IRL, 0x340), INTC_VECT(IRL, 0x360),
-        INTC_VECT(IRL, 0x380), INTC_VECT(IRL, 0x3a0),
-        INTC_VECT(IRL, 0x3c0),
+        INTC_VECT(IRL_LLLL, 0x200), INTC_VECT(IRL_LLLH, 0x220),
+        INTC_VECT(IRL_LLHL, 0x240), INTC_VECT(IRL_LLHH, 0x260),
+        INTC_VECT(IRL_LHLL, 0x280), INTC_VECT(IRL_LHLH, 0x2a0),
+        INTC_VECT(IRL_LHHL, 0x2c0), INTC_VECT(IRL_LHHH, 0x2e0),
+        INTC_VECT(IRL_HLLL, 0x300), INTC_VECT(IRL_HLLH, 0x320),
+        INTC_VECT(IRL_HLHL, 0x340), INTC_VECT(IRL_HLHH, 0x360),
+        INTC_VECT(IRL_HHLL, 0x380), INTC_VECT(IRL_HHLH, 0x3a0),
+        INTC_VECT(IRL_HHHL, 0x3c0),
 };
 
 static DECLARE_INTC_DESC(intc_desc_irl, "shx3-irl", vectors_irl, groups,
diff --git a/arch/sh/kernel/cpu/sh4a/smp-shx3.c b/arch/sh/kernel/cpu/sh4a/smp-shx3.c
index 185ec3976a25..5863e0c4d02f 100644
--- a/arch/sh/kernel/cpu/sh4a/smp-shx3.c
+++ b/arch/sh/kernel/cpu/sh4a/smp-shx3.c
@@ -14,6 +14,13 @@
 #include <linux/interrupt.h>
 #include <linux/io.h>
 
+#define STBCR_REG(phys_id) (0xfe400004 | (phys_id << 12))
+#define RESET_REG(phys_id) (0xfe400008 | (phys_id << 12))
+
+#define STBCR_MSTP      0x00000001
+#define STBCR_RESET     0x00000002
+#define STBCR_LTSLP     0x80000000
+
 static irqreturn_t ipi_interrupt_handler(int irq, void *arg)
 {
         unsigned int message = (unsigned int)(long)arg;
@@ -21,9 +28,9 @@ static irqreturn_t ipi_interrupt_handler(int irq, void *arg)
         unsigned int offs = 4 * cpu;
         unsigned int x;
 
-        x = ctrl_inl(0xfe410070 + offs); /* C0INITICI..CnINTICI */
+        x = __raw_readl(0xfe410070 + offs); /* C0INITICI..CnINTICI */
         x &= (1 << (message << 2));
-        ctrl_outl(x, 0xfe410080 + offs); /* C0INTICICLR..CnINTICICLR */
+        __raw_writel(x, 0xfe410080 + offs); /* C0INTICICLR..CnINTICICLR */
 
         smp_message_recv(message);
 
@@ -37,6 +44,9 @@ void __init plat_smp_setup(void)
 
         init_cpu_possible(cpumask_of(cpu));
 
+        /* Enable light sleep for the boot CPU */
+        __raw_writel(__raw_readl(STBCR_REG(cpu)) | STBCR_LTSLP, STBCR_REG(cpu));
+
         __cpu_number_map[0] = 0;
         __cpu_logical_map[0] = 0;
 
@@ -66,32 +76,23 @@ void __init plat_prepare_cpus(unsigned int max_cpus)
                             "IPI", (void *)(long)i);
 }
 
-#define STBCR_REG(phys_id) (0xfe400004 | (phys_id << 12))
-#define RESET_REG(phys_id) (0xfe400008 | (phys_id << 12))
-
-#define STBCR_MSTP      0x00000001
-#define STBCR_RESET     0x00000002
-#define STBCR_LTSLP     0x80000000
-
-#define STBCR_AP_VAL    (STBCR_RESET | STBCR_LTSLP)
-
 void plat_start_cpu(unsigned int cpu, unsigned long entry_point)
 {
-        ctrl_outl(entry_point, RESET_REG(cpu));
+        __raw_writel(entry_point, RESET_REG(cpu));
 
-        if (!(ctrl_inl(STBCR_REG(cpu)) & STBCR_MSTP))
-                ctrl_outl(STBCR_MSTP, STBCR_REG(cpu));
+        if (!(__raw_readl(STBCR_REG(cpu)) & STBCR_MSTP))
+                __raw_writel(STBCR_MSTP, STBCR_REG(cpu));
 
-        while (!(ctrl_inl(STBCR_REG(cpu)) & STBCR_MSTP))
+        while (!(__raw_readl(STBCR_REG(cpu)) & STBCR_MSTP))
                 cpu_relax();
 
         /* Start up secondary processor by sending a reset */
-        ctrl_outl(STBCR_AP_VAL, STBCR_REG(cpu));
+        __raw_writel(STBCR_RESET | STBCR_LTSLP, STBCR_REG(cpu));
 }
 
 int plat_smp_processor_id(void)
 {
-        return ctrl_inl(0xff000048); /* CPIDR */
+        return __raw_readl(0xff000048); /* CPIDR */
 }
 
 void plat_send_ipi(unsigned int cpu, unsigned int message)
@@ -100,5 +101,5 @@ void plat_send_ipi(unsigned int cpu, unsigned int message)
 
         BUG_ON(cpu >= 4);
 
-        ctrl_outl(1 << (message << 2), addr); /* C0INTICI..CnINTICI */
+        __raw_writel(1 << (message << 2), addr); /* C0INTICI..CnINTICI */
 }
diff --git a/arch/sh/kernel/cpu/sh5/entry.S b/arch/sh/kernel/cpu/sh5/entry.S
index b0aacf675258..8f13f73cb2cb 100644
--- a/arch/sh/kernel/cpu/sh5/entry.S
+++ b/arch/sh/kernel/cpu/sh5/entry.S
@@ -933,7 +933,7 @@ ret_with_reschedule:
 
         pta     restore_all, tr1
 
-        movi    (_TIF_SIGPENDING | _TIF_RESTORE_SIGMASK), r8
+        movi    _TIF_SIGPENDING, r8
         and     r8, r7, r8
         pta     work_notifysig, tr0
         bne     r8, ZERO, tr0
diff --git a/arch/sh/kernel/cpu/shmobile/cpuidle.c b/arch/sh/kernel/cpu/shmobile/cpuidle.c
index 1c504bd972c3..83972aa319c2 100644
--- a/arch/sh/kernel/cpu/shmobile/cpuidle.c
+++ b/arch/sh/kernel/cpu/shmobile/cpuidle.c
@@ -87,25 +87,31 @@ void sh_mobile_setup_cpuidle(void)
 
         dev->safe_state = state;
 
-        state = &dev->states[i++];
-        snprintf(state->name, CPUIDLE_NAME_LEN, "C1");
-        strncpy(state->desc, "SuperH Sleep Mode [SF]", CPUIDLE_DESC_LEN);
-        state->exit_latency = 100;
-        state->target_residency = 1 * 2;
-        state->power_usage = 1;
-        state->flags = 0;
-        state->flags |= CPUIDLE_FLAG_TIME_VALID;
-        state->enter = cpuidle_sleep_enter;
+        if (sh_mobile_sleep_supported & SUSP_SH_SF) {
+                state = &dev->states[i++];
+                snprintf(state->name, CPUIDLE_NAME_LEN, "C1");
+                strncpy(state->desc, "SuperH Sleep Mode [SF]",
+                        CPUIDLE_DESC_LEN);
+                state->exit_latency = 100;
+                state->target_residency = 1 * 2;
+                state->power_usage = 1;
+                state->flags = 0;
+                state->flags |= CPUIDLE_FLAG_TIME_VALID;
+                state->enter = cpuidle_sleep_enter;
+        }
 
-        state = &dev->states[i++];
-        snprintf(state->name, CPUIDLE_NAME_LEN, "C2");
-        strncpy(state->desc, "SuperH Mobile Standby Mode [SF]", CPUIDLE_DESC_LEN);
-        state->exit_latency = 2300;
-        state->target_residency = 1 * 2;
-        state->power_usage = 1;
-        state->flags = 0;
-        state->flags |= CPUIDLE_FLAG_TIME_VALID;
-        state->enter = cpuidle_sleep_enter;
+        if (sh_mobile_sleep_supported & SUSP_SH_STANDBY) {
+                state = &dev->states[i++];
+                snprintf(state->name, CPUIDLE_NAME_LEN, "C2");
+                strncpy(state->desc, "SuperH Mobile Standby Mode [SF]",
+                        CPUIDLE_DESC_LEN);
+                state->exit_latency = 2300;
+                state->target_residency = 1 * 2;
+                state->power_usage = 1;
+                state->flags = 0;
+                state->flags |= CPUIDLE_FLAG_TIME_VALID;
+                state->enter = cpuidle_sleep_enter;
+        }
 
         dev->state_count = i;
 
diff --git a/arch/sh/kernel/cpu/shmobile/pm.c b/arch/sh/kernel/cpu/shmobile/pm.c
index ee3c2aaf66fb..ca029a44743c 100644
--- a/arch/sh/kernel/cpu/shmobile/pm.c
+++ b/arch/sh/kernel/cpu/shmobile/pm.c
@@ -15,6 +15,13 @@
 #include <linux/suspend.h>
 #include <asm/suspend.h>
 #include <asm/uaccess.h>
+#include <asm/cacheflush.h>
+
+/*
+ * Notifier lists for pre/post sleep notification
+ */
+ATOMIC_NOTIFIER_HEAD(sh_mobile_pre_sleep_notifier_list);
+ATOMIC_NOTIFIER_HEAD(sh_mobile_post_sleep_notifier_list);
 
 /*
  * Sleep modes available on SuperH Mobile:
@@ -26,30 +33,105 @@
 #define SUSP_MODE_SLEEP         (SUSP_SH_SLEEP)
 #define SUSP_MODE_SLEEP_SF      (SUSP_SH_SLEEP | SUSP_SH_SF)
 #define SUSP_MODE_STANDBY_SF    (SUSP_SH_STANDBY | SUSP_SH_SF)
+#define SUSP_MODE_RSTANDBY      (SUSP_SH_RSTANDBY | SUSP_SH_MMU | SUSP_SH_SF)
+ /*
+  * U-standby mode is unsupported since it needs bootloader hacks
+  */
 
-/*
- * The following modes are not there yet:
- *
- * R-standby mode is unsupported, but will be added in the future
- * U-standby mode is low priority since it needs bootloader hacks
- */
-
-#define ILRAM_BASE 0xe5200000
-
-extern const unsigned char sh_mobile_standby[];
-extern const unsigned int sh_mobile_standby_size;
+#ifdef CONFIG_CPU_SUBTYPE_SH7724
+#define RAM_BASE 0xfd800000 /* RSMEM */
+#else
+#define RAM_BASE 0xe5200000 /* ILRAM */
+#endif
 
 void sh_mobile_call_standby(unsigned long mode)
 {
-        void *onchip_mem = (void *)ILRAM_BASE;
-        void (*standby_onchip_mem)(unsigned long, unsigned long) = onchip_mem;
+        void *onchip_mem = (void *)RAM_BASE;
+        struct sh_sleep_data *sdp = onchip_mem;
+        void (*standby_onchip_mem)(unsigned long, unsigned long);
+
+        /* code located directly after data structure */
+        standby_onchip_mem = (void *)(sdp + 1);
+
+        atomic_notifier_call_chain(&sh_mobile_pre_sleep_notifier_list,
+                                   mode, NULL);
+
+        /* flush the caches if MMU flag is set */
+        if (mode & SUSP_SH_MMU)
+                flush_cache_all();
 
         /* Let assembly snippet in on-chip memory handle the rest */
-        standby_onchip_mem(mode, ILRAM_BASE);
+        standby_onchip_mem(mode, RAM_BASE);
+
+        atomic_notifier_call_chain(&sh_mobile_post_sleep_notifier_list,
+                                   mode, NULL);
+}
+
+extern char sh_mobile_sleep_enter_start;
+extern char sh_mobile_sleep_enter_end;
+
+extern char sh_mobile_sleep_resume_start;
+extern char sh_mobile_sleep_resume_end;
+
+unsigned long sh_mobile_sleep_supported = SUSP_SH_SLEEP;
+
+void sh_mobile_register_self_refresh(unsigned long flags,
+                                     void *pre_start, void *pre_end,
+                                     void *post_start, void *post_end)
+{
+        void *onchip_mem = (void *)RAM_BASE;
+        void *vp;
+        struct sh_sleep_data *sdp;
+        int n;
+
+        /* part 0: data area */
+        sdp = onchip_mem;
+        sdp->addr.stbcr = 0xa4150020; /* STBCR */
+        sdp->addr.bar = 0xa4150040; /* BAR */
+        sdp->addr.pteh = 0xff000000; /* PTEH */
+        sdp->addr.ptel = 0xff000004; /* PTEL */
+        sdp->addr.ttb = 0xff000008; /* TTB */
+        sdp->addr.tea = 0xff00000c; /* TEA */
+        sdp->addr.mmucr = 0xff000010; /* MMUCR */
+        sdp->addr.ptea = 0xff000034; /* PTEA */
+        sdp->addr.pascr = 0xff000070; /* PASCR */
+        sdp->addr.irmcr = 0xff000078; /* IRMCR */
+        sdp->addr.ccr = 0xff00001c; /* CCR */
+        sdp->addr.ramcr = 0xff000074; /* RAMCR */
+        vp = sdp + 1;
+
+        /* part 1: common code to enter sleep mode */
+        n = &sh_mobile_sleep_enter_end - &sh_mobile_sleep_enter_start;
+        memcpy(vp, &sh_mobile_sleep_enter_start, n);
+        vp += roundup(n, 4);
+
+        /* part 2: board specific code to enter self-refresh mode */
+        n = pre_end - pre_start;
+        memcpy(vp, pre_start, n);
+        sdp->sf_pre = (unsigned long)vp;
+        vp += roundup(n, 4);
+
+        /* part 3: board specific code to resume from self-refresh mode */
+        n = post_end - post_start;
+        memcpy(vp, post_start, n);
+        sdp->sf_post = (unsigned long)vp;
+        vp += roundup(n, 4);
+
+        /* part 4: common code to resume from sleep mode */
+        WARN_ON(vp > (onchip_mem + 0x600));
+        vp = onchip_mem + 0x600; /* located at interrupt vector */
+        n = &sh_mobile_sleep_resume_end - &sh_mobile_sleep_resume_start;
+        memcpy(vp, &sh_mobile_sleep_resume_start, n);
+        sdp->resume = (unsigned long)vp;
+
+        sh_mobile_sleep_supported |= flags;
 }
 
 static int sh_pm_enter(suspend_state_t state)
 {
+        if (!(sh_mobile_sleep_supported & SUSP_MODE_STANDBY_SF))
+                return -ENXIO;
+
         local_irq_disable();
         set_bl_bit();
         sh_mobile_call_standby(SUSP_MODE_STANDBY_SF);
@@ -65,13 +147,6 @@ static struct platform_suspend_ops sh_pm_ops = {
 
 static int __init sh_pm_init(void)
 {
-        void *onchip_mem = (void *)ILRAM_BASE;
-
-        /* Copy the assembly snippet to the otherwise ununsed ILRAM */
-        memcpy(onchip_mem, sh_mobile_standby, sh_mobile_standby_size);
-        wmb();
-        ctrl_barrier();
-
         suspend_set_ops(&sh_pm_ops);
         sh_mobile_setup_cpuidle();
         return 0;
diff --git a/arch/sh/kernel/cpu/shmobile/pm_runtime.c b/arch/sh/kernel/cpu/shmobile/pm_runtime.c
index 7c615b17e209..6dcb8166a64d 100644
--- a/arch/sh/kernel/cpu/shmobile/pm_runtime.c
+++ b/arch/sh/kernel/cpu/shmobile/pm_runtime.c
@@ -45,12 +45,14 @@ static int __platform_pm_runtime_resume(struct platform_device *pdev)
 
         dev_dbg(d, "__platform_pm_runtime_resume() [%d]\n", hwblk);
 
-        if (d->driver && d->driver->pm && d->driver->pm->runtime_resume) {
+        if (d->driver) {
                 hwblk_enable(hwblk_info, hwblk);
                 ret = 0;
 
                 if (test_bit(PDEV_ARCHDATA_FLAG_SUSP, &ad->flags)) {
-                        ret = d->driver->pm->runtime_resume(d);
+                        if (d->driver->pm && d->driver->pm->runtime_resume)
+                                ret = d->driver->pm->runtime_resume(d);
+
                         if (!ret)
                                 clear_bit(PDEV_ARCHDATA_FLAG_SUSP, &ad->flags);
                         else
@@ -73,12 +75,15 @@ static int __platform_pm_runtime_suspend(struct platform_device *pdev)
 
         dev_dbg(d, "__platform_pm_runtime_suspend() [%d]\n", hwblk);
 
-        if (d->driver && d->driver->pm && d->driver->pm->runtime_suspend) {
+        if (d->driver) {
                 BUG_ON(!test_bit(PDEV_ARCHDATA_FLAG_IDLE, &ad->flags));
+                ret = 0;
 
-                hwblk_enable(hwblk_info, hwblk);
-                ret = d->driver->pm->runtime_suspend(d);
-                hwblk_disable(hwblk_info, hwblk);
+                if (d->driver->pm && d->driver->pm->runtime_suspend) {
+                        hwblk_enable(hwblk_info, hwblk);
+                        ret = d->driver->pm->runtime_suspend(d);
+                        hwblk_disable(hwblk_info, hwblk);
+                }
 
                 if (!ret) {
                         set_bit(PDEV_ARCHDATA_FLAG_SUSP, &ad->flags);
diff --git a/arch/sh/kernel/cpu/shmobile/sleep.S b/arch/sh/kernel/cpu/shmobile/sleep.S
index a439e6c7824f..e9dd7fa0abd2 100644
--- a/arch/sh/kernel/cpu/shmobile/sleep.S
+++ b/arch/sh/kernel/cpu/shmobile/sleep.S
@@ -20,79 +20,103 @@
  * Kernel mode register usage, see entry.S:
  *      k0      scratch
  *      k1      scratch
- *      k4      scratch
  */
 #define k0      r0
 #define k1      r1
-#define k4      r4
 
-/* manage self-refresh and enter standby mode.
+/* manage self-refresh and enter standby mode. must be self-contained.
  * this code will be copied to on-chip memory and executed from there.
  */
+        .balign 4
+ENTRY(sh_mobile_sleep_enter_start)
 
-        .balign         4096,0,4096
-ENTRY(sh_mobile_standby)
+        /* save mode flags */
+        mov.l   r4, @(SH_SLEEP_MODE, r5)
 
         /* save original vbr */
-        stc     vbr, r1
-        mova    saved_vbr, r0
-        mov.l   r1, @r0
+        stc     vbr, r0
+        mov.l   r0, @(SH_SLEEP_VBR, r5)
 
         /* point vbr to our on-chip memory page */
         ldc     r5, vbr
 
         /* save return address */
-        mova    saved_spc, r0
-        sts     pr, r5
-        mov.l   r5, @r0
+        sts     pr, r0
+        mov.l   r0, @(SH_SLEEP_SPC, r5)
 
         /* save sr */
-        mova    saved_sr, r0
-        stc     sr, r5
-        mov.l   r5, @r0
+        stc     sr, r0
+        mov.l   r0, @(SH_SLEEP_SR, r5)
 
-        /* save mode flags */
-        mova    saved_mode, r0
-        mov.l   r4, @r0
+        /* save sp */
+        mov.l   r15, @(SH_SLEEP_SP, r5)
+
+        /* save stbcr */
+        bsr     save_register
+         mov    #SH_SLEEP_REG_STBCR, r0
+
+        /* save mmu and cache context if needed */
+        mov.l   @(SH_SLEEP_MODE, r5), r0
+        tst     #SUSP_SH_MMU, r0
+        bt      skip_mmu_save_disable
+
+       /* save mmu state */
+        bsr     save_register
+         mov    #SH_SLEEP_REG_PTEH, r0
+
+        bsr     save_register
+         mov    #SH_SLEEP_REG_PTEL, r0
+
+        bsr     save_register
+         mov    #SH_SLEEP_REG_TTB, r0
+
+        bsr     save_register
+         mov    #SH_SLEEP_REG_TEA, r0
+
+        bsr     save_register
+         mov    #SH_SLEEP_REG_MMUCR, r0
+
+        bsr     save_register
+         mov    #SH_SLEEP_REG_PTEA, r0
+
+        bsr     save_register
+         mov    #SH_SLEEP_REG_PASCR, r0
 
-        /* put mode flags in r0 */
-        mov     r4, r0
+        bsr     save_register
+         mov    #SH_SLEEP_REG_IRMCR, r0
 
+        /* invalidate TLBs and disable the MMU */
+        bsr     get_register
+         mov    #SH_SLEEP_REG_MMUCR, r0
+        mov     #4, r1
+        mov.l   r1, @r0
+        icbi    @r0
+
+        /* save cache registers and disable caches */
+        bsr     save_register
+         mov    #SH_SLEEP_REG_CCR, r0
+
+        bsr     save_register
+         mov    #SH_SLEEP_REG_RAMCR, r0
+
+        bsr     get_register
+         mov    #SH_SLEEP_REG_CCR, r0
+        mov     #0, r1
+        mov.l   r1, @r0
+        icbi    @r0
+
+skip_mmu_save_disable:
+        /* call self-refresh entering code if needed */
+        mov.l   @(SH_SLEEP_MODE, r5), r0
         tst     #SUSP_SH_SF, r0
         bt      skip_set_sf
-#ifdef CONFIG_CPU_SUBTYPE_SH7724
-        /* DBSC: put memory in self-refresh mode */
-        mov.l   dben_reg, r4
-        mov.l   dben_data0, r1
-        mov.l   r1, @r4
-
-        mov.l   dbrfpdn0_reg, r4
-        mov.l   dbrfpdn0_data0, r1
-        mov.l   r1, @r4
-
-        mov.l   dbcmdcnt_reg, r4
-        mov.l   dbcmdcnt_data0, r1
-        mov.l   r1, @r4
-
-        mov.l   dbcmdcnt_reg, r4
-        mov.l   dbcmdcnt_data1, r1
-        mov.l   r1, @r4
-
-        mov.l   dbrfpdn0_reg, r4
-        mov.l   dbrfpdn0_data1, r1
-        mov.l   r1, @r4
-#else
-        /* SBSC: disable power down and put in self-refresh mode */
-        mov.l   1f, r4
-        mov.l   2f, r1
-        mov.l   @r4, r2
-        or      r1, r2
-        mov.l   3f, r3
-        and     r3, r2
-        mov.l   r2, @r4
-#endif
+
+        mov.l   @(SH_SLEEP_SF_PRE, r5), r0
+        jsr     @r0
+         nop
 
 skip_set_sf:
+        mov.l   @(SH_SLEEP_MODE, r5), r0
         tst     #SUSP_SH_STANDBY, r0
         bt      test_rstandby
 
@@ -104,6 +128,12 @@ test_rstandby:
         tst     #SUSP_SH_RSTANDBY, r0
         bt      test_ustandby
 
+        /* setup BAR register */
+        bsr     get_register
+         mov    #SH_SLEEP_REG_BAR, r0
+        mov.l   @(SH_SLEEP_RESUME, r5), r1
+        mov.l   r1, @r0
+
         /* set mode to "r-standby mode" */
         bra     do_sleep
          mov    #0x20, r1
@@ -123,124 +153,136 @@ force_sleep:
 
 do_sleep:
         /* setup and enter selected standby mode */
-        mov.l   5f, r4
-        mov.l   r1, @r4
+        bsr     get_register
+         mov    #SH_SLEEP_REG_STBCR, r0
+        mov.l   r1, @r0
 again:
         sleep
         bra     again
          nop
 
-restore_jump_vbr:
+save_register:
+        add     #SH_SLEEP_BASE_ADDR, r0
+        mov.l   @(r0, r5), r1
+        add     #-SH_SLEEP_BASE_ADDR, r0
+        mov.l   @r1, r1
+        add     #SH_SLEEP_BASE_DATA, r0
+        mov.l   r1, @(r0, r5)
+        add     #-SH_SLEEP_BASE_DATA, r0
+        rts
+         nop
+
+get_register:
+        add     #SH_SLEEP_BASE_ADDR, r0
+        mov.l   @(r0, r5), r0
+        rts
+         nop
+ENTRY(sh_mobile_sleep_enter_end)
+
+        .balign 4
+ENTRY(sh_mobile_sleep_resume_start)
+
+        /* figure out start address */
+        bsr     0f
+         nop
+0:
+        sts     pr, k1
+        mov.l   1f, k0
+        and     k0, k1
+
+        /* store pointer to data area in VBR */
+        ldc     k1, vbr
+
+        /* setup sr with saved sr */
+        mov.l   @(SH_SLEEP_SR, k1), k0
+        ldc     k0, sr
+
+        /* now: user register set! */
+        stc     vbr, r5
+
         /* setup spc with return address to c code */
-        mov.l   saved_spc, k0
-        ldc     k0, spc
+        mov.l   @(SH_SLEEP_SPC, r5), r0
+        ldc     r0, spc
 
         /* restore vbr */
-        mov.l   saved_vbr, k0
-        ldc     k0, vbr
+        mov.l   @(SH_SLEEP_VBR, r5), r0
+        ldc     r0, vbr
 
         /* setup ssr with saved sr */
-        mov.l   saved_sr, k0
-        ldc     k0, ssr
+        mov.l   @(SH_SLEEP_SR, r5), r0
+        ldc     r0, ssr
 
-        /* get mode flags */
-        mov.l   saved_mode, k0
+        /* restore sp */
+        mov.l   @(SH_SLEEP_SP, r5), r15
 
-done_sleep:
-        /* reset standby mode to sleep mode */
-        mov.l   5f, k4
-        mov     #0x00, k1
-        mov.l   k1, @k4
+        /* restore sleep mode register */
+        bsr     restore_register
+         mov    #SH_SLEEP_REG_STBCR, r0
 
-        tst     #SUSP_SH_SF, k0
+        /* call self-refresh resume code if needed */
+        mov.l   @(SH_SLEEP_MODE, r5), r0
+        tst     #SUSP_SH_SF, r0
         bt      skip_restore_sf
 
-#ifdef CONFIG_CPU_SUBTYPE_SH7724
-        /* DBSC: put memory in auto-refresh mode */
-        mov.l   dbrfpdn0_reg, k4
-        mov.l   dbrfpdn0_data0, k1
-        mov.l   k1, @k4
-
-        nop /* sleep 140 ns */
-        nop
-        nop
-        nop
-
-        mov.l   dbcmdcnt_reg, k4
-        mov.l   dbcmdcnt_data0, k1
-        mov.l   k1, @k4
-
-        mov.l   dbcmdcnt_reg, k4
-        mov.l   dbcmdcnt_data1, k1
-        mov.l   k1, @k4
-
-        mov.l   dben_reg, k4
-        mov.l   dben_data1, k1
-        mov.l   k1, @k4
-
-        mov.l   dbrfpdn0_reg, k4
-        mov.l   dbrfpdn0_data2, k1
-        mov.l   k1, @k4
-#else
-        /* SBSC: set auto-refresh mode */
-        mov.l   1f, k4
-        mov.l   @k4, k0
-        mov.l   4f, k1
-        and     k1, k0
-        mov.l   k0, @k4
-        mov.l   6f, k4
-        mov.l   8f, k0
-        mov.l   @k4, k1
-        mov     #-1, k4
-        add     k4, k1
-        or      k1, k0
-        mov.l   7f, k1
-        mov.l   k0, @k1
-#endif
+        mov.l   @(SH_SLEEP_SF_POST, r5), r0
+        jsr     @r0
+         nop
+
 skip_restore_sf:
-        /* jump to vbr vector */
-        mov.l   saved_vbr, k0
-        mov.l   offset_vbr, k4
-        add     k4, k0
-        jmp     @k0
+        /* restore mmu and cache state if needed */
+        mov.l   @(SH_SLEEP_MODE, r5), r0
+        tst     #SUSP_SH_MMU, r0
+        bt      skip_restore_mmu
+
+        /* restore mmu state */
+        bsr     restore_register
+         mov    #SH_SLEEP_REG_PTEH, r0
+
+        bsr     restore_register
+         mov    #SH_SLEEP_REG_PTEL, r0
+
+        bsr     restore_register
+         mov    #SH_SLEEP_REG_TTB, r0
+
+        bsr     restore_register
+         mov    #SH_SLEEP_REG_TEA, r0
+
+        bsr     restore_register
+         mov    #SH_SLEEP_REG_PTEA, r0
+
+        bsr     restore_register
+         mov    #SH_SLEEP_REG_PASCR, r0
+
+        bsr     restore_register
+         mov    #SH_SLEEP_REG_IRMCR, r0
+
+        bsr     restore_register
+         mov    #SH_SLEEP_REG_MMUCR, r0
+        icbi    @r0
+
+        /* restore cache settings */
+        bsr     restore_register
+         mov    #SH_SLEEP_REG_RAMCR, r0
+        icbi    @r0
+
+        bsr     restore_register
+         mov    #SH_SLEEP_REG_CCR, r0
+        icbi    @r0
+
+skip_restore_mmu:
+        rte
          nop
 
-        .balign 4
-saved_mode:     .long   0
-saved_spc:      .long   0
-saved_sr:       .long   0
-saved_vbr:      .long   0
-offset_vbr:     .long   0x600
-#ifdef CONFIG_CPU_SUBTYPE_SH7724
-dben_reg:       .long   0xfd000010 /* DBEN */
-dben_data0:     .long   0
-dben_data1:     .long   1
-dbrfpdn0_reg:   .long   0xfd000040 /* DBRFPDN0 */
-dbrfpdn0_data0: .long   0
-dbrfpdn0_data1: .long   1
-dbrfpdn0_data2: .long   0x00010000
-dbcmdcnt_reg:   .long   0xfd000014 /* DBCMDCNT */
-dbcmdcnt_data0: .long   2
-dbcmdcnt_data1: .long   4
-#else
-1:      .long   0xfe400008 /* SDCR0 */
-2:      .long   0x00000400
-3:      .long   0xffff7fff
-4:      .long   0xfffffbff
-#endif
-5:      .long   0xa4150020 /* STBCR */
-6:      .long   0xfe40001c /* RTCOR */
-7:      .long   0xfe400018 /* RTCNT */
-8:      .long   0xa55a0000
-
-
-/* interrupt vector @ 0x600 */
-        .balign         0x400,0,0x400
-        .long   0xdeadbeef
-        .balign         0x200,0,0x200
-        bra     restore_jump_vbr
+restore_register:
+        add     #SH_SLEEP_BASE_DATA, r0
+        mov.l   @(r0, r5), r1
+        add     #-SH_SLEEP_BASE_DATA, r0
+        add     #SH_SLEEP_BASE_ADDR, r0
+        mov.l   @(r0, r5), r0
+        mov.l   r1, @r0
+        rts
          nop
-sh_mobile_standby_end:
 
-ENTRY(sh_mobile_standby_size)
-        .long sh_mobile_standby_end - sh_mobile_standby
+        .balign 4
+1:      .long   ~0x7ff
+ENTRY(sh_mobile_sleep_resume_end)
diff --git a/arch/sh/kernel/cpu/ubc.S b/arch/sh/kernel/cpu/ubc.S
deleted file mode 100644
index 81923079fa12..000000000000
--- a/arch/sh/kernel/cpu/ubc.S
+++ /dev/null
@@ -1,59 +0,0 @@
-/*
- * arch/sh/kernel/cpu/ubc.S
- *
- * Set of management routines for the User Break Controller (UBC)
- *
- * Copyright (C) 2002 Paul Mundt
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms of the GNU General Public License as published by the
- * Free Software Foundation; either version 2 of the License, or (at your
- * option) any later version.
- */
-#include <linux/linkage.h>
-#include <asm/ubc.h>
-
-#define STBCR2          0xffc00010
-
-ENTRY(ubc_sleep)
-        mov     #0, r0
-
-        mov.l   1f, r1          ! Zero out UBC_BBRA ..
-        mov.w   r0, @r1
-
-        mov.l   2f, r1          ! .. same for BBRB ..
-        mov.w   r0, @r1
-
-        mov.l   3f, r1          ! .. and again for BRCR.
-        mov.w   r0, @r1
-
-        mov.w   @r1, r0         ! Dummy read BRCR
-
-        mov.l   4f, r1          ! Set MSTP5 in STBCR2
-        mov.b   @r1, r0
-        or      #0x01, r0
-        mov.b   r0, @r1
-
-        mov.b   @r1, r0         ! Two dummy reads ..
-        mov.b   @r1, r0
-
-        rts
-        nop
-
-ENTRY(ubc_wakeup)
-        mov.l   4f, r1          ! Clear MSTP5
-        mov.b   @r1, r0
-        and     #0xfe, r0
-        mov.b   r0, @r1
-
-        mov.b   @r1, r0         ! Two more dummy reads ..
-        mov.b   @r1, r0
-
-        rts
-        nop
-
-1:      .long   UBC_BBRA
-2:      .long   UBC_BBRB
-3:      .long   UBC_BRCR
-4:      .long   STBCR2
-
diff --git a/arch/sh/kernel/dma-nommu.c b/arch/sh/kernel/dma-nommu.c
new file mode 100644
index 000000000000..3c55b87f8b63
--- /dev/null
+++ b/arch/sh/kernel/dma-nommu.c
@@ -0,0 +1,82 @@
+/*
+ * DMA mapping support for platforms lacking IOMMUs.
+ *
+ * Copyright (C) 2009  Paul Mundt
+ *
+ * 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/dma-mapping.h>
+#include <linux/io.h>
+
+static dma_addr_t nommu_map_page(struct device *dev, struct page *page,
+                                 unsigned long offset, size_t size,
+                                 enum dma_data_direction dir,
+                                 struct dma_attrs *attrs)
+{
+        dma_addr_t addr = page_to_phys(page) + offset;
+
+        WARN_ON(size == 0);
+        dma_cache_sync(dev, page_address(page) + offset, size, dir);
+
+        return addr;
+}
+
+static int nommu_map_sg(struct device *dev, struct scatterlist *sg,
+                        int nents, enum dma_data_direction dir,
+                        struct dma_attrs *attrs)
+{
+        struct scatterlist *s;
+        int i;
+
+        WARN_ON(nents == 0 || sg[0].length == 0);
+
+        for_each_sg(sg, s, nents, i) {
+                BUG_ON(!sg_page(s));
+
+                dma_cache_sync(dev, sg_virt(s), s->length, dir);
+
+                s->dma_address = sg_phys(s);
+                s->dma_length = s->length;
+        }
+
+        return nents;
+}
+
+#ifdef CONFIG_DMA_NONCOHERENT
+static void nommu_sync_single(struct device *dev, dma_addr_t addr,
+                              size_t size, enum dma_data_direction dir)
+{
+        dma_cache_sync(dev, phys_to_virt(addr), size, dir);
+}
+
+static void nommu_sync_sg(struct device *dev, struct scatterlist *sg,
+                          int nelems, enum dma_data_direction dir)
+{
+        struct scatterlist *s;
+        int i;
+
+        for_each_sg(sg, s, nelems, i)
+                dma_cache_sync(dev, sg_virt(s), s->length, dir);
+}
+#endif
+
+struct dma_map_ops nommu_dma_ops = {
+        .alloc_coherent         = dma_generic_alloc_coherent,
+        .free_coherent          = dma_generic_free_coherent,
+        .map_page               = nommu_map_page,
+        .map_sg                 = nommu_map_sg,
+#ifdef CONFIG_DMA_NONCOHERENT
+        .sync_single_for_device = nommu_sync_single,
+        .sync_sg_for_device     = nommu_sync_sg,
+#endif
+        .is_phys                = 1,
+};
+
+void __init no_iommu_init(void)
+{
+        if (dma_ops)
+                return;
+        dma_ops = &nommu_dma_ops;
+}
diff --git a/arch/sh/kernel/dwarf.c b/arch/sh/kernel/dwarf.c
index d76a23170dbb..3576b709f052 100644
--- a/arch/sh/kernel/dwarf.c
+++ b/arch/sh/kernel/dwarf.c
@@ -20,6 +20,7 @@
 #include <linux/list.h>
 #include <linux/mempool.h>
 #include <linux/mm.h>
+#include <linux/elf.h>
 #include <linux/ftrace.h>
 #include <asm/dwarf.h>
 #include <asm/unwinder.h>
@@ -530,7 +531,18 @@ static int dwarf_cfa_execute_insns(unsigned char *insn_start,
 }
 
 /**
- *      dwarf_unwind_stack - recursively unwind the stack
+ *      dwarf_free_frame - free the memory allocated for @frame
+ *      @frame: the frame to free
+ */
+void dwarf_free_frame(struct dwarf_frame *frame)
+{
+        dwarf_frame_free_regs(frame);
+        mempool_free(frame, dwarf_frame_pool);
+}
+
+/**
+ *      dwarf_unwind_stack - unwind the stack
+ *
  *      @pc: address of the function to unwind
  *      @prev: struct dwarf_frame of the previous stackframe on the callstack
  *
@@ -548,9 +560,9 @@ struct dwarf_frame * dwarf_unwind_stack(unsigned long pc,
         unsigned long addr;
 
         /*
-         * If this is the first invocation of this recursive function we
-         * need get the contents of a physical register to get the CFA
-         * in order to begin the virtual unwinding of the stack.
+         * If we're starting at the top of the stack we need get the
+         * contents of a physical register to get the CFA in order to
+         * begin the virtual unwinding of the stack.
          *
          * NOTE: the return address is guaranteed to be setup by the
          * time this function makes its first function call.
@@ -593,9 +605,8 @@ struct dwarf_frame * dwarf_unwind_stack(unsigned long pc,
         fde = dwarf_lookup_fde(pc);
         if (!fde) {
                 /*
-                 * This is our normal exit path - the one that stops the
-                 * recursion. There's two reasons why we might exit
-                 * here,
+                 * This is our normal exit path. There are two reasons
+                 * why we might exit here,
                  *
                  *      a) pc has no asscociated DWARF frame info and so
                  *      we don't know how to unwind this frame. This is
@@ -637,10 +648,10 @@ struct dwarf_frame * dwarf_unwind_stack(unsigned long pc,
 
                 } else {
                         /*
-                         * Again, this is the first invocation of this
-                         * recurisve function. We need to physically
-                         * read the contents of a register in order to
-                         * get the Canonical Frame Address for this
+                         * Again, we're starting from the top of the
+                         * stack. We need to physically read
+                         * the contents of a register in order to get
+                         * the Canonical Frame Address for this
                          * function.
                          */
                         frame->cfa = dwarf_read_arch_reg(frame->cfa_register);
@@ -670,13 +681,12 @@ struct dwarf_frame * dwarf_unwind_stack(unsigned long pc,
         return frame;
 
 bail:
-        dwarf_frame_free_regs(frame);
-        mempool_free(frame, dwarf_frame_pool);
+        dwarf_free_frame(frame);
         return NULL;
 }
 
 static int dwarf_parse_cie(void *entry, void *p, unsigned long len,
-                           unsigned char *end)
+                           unsigned char *end, struct module *mod)
 {
         struct dwarf_cie *cie;
         unsigned long flags;
@@ -772,6 +782,8 @@ static int dwarf_parse_cie(void *entry, void *p, unsigned long len,
         cie->initial_instructions = p;
         cie->instructions_end = end;
 
+        cie->mod = mod;
+
         /* Add to list */
         spin_lock_irqsave(&dwarf_cie_lock, flags);
         list_add_tail(&cie->link, &dwarf_cie_list);
@@ -782,7 +794,7 @@ static int dwarf_parse_cie(void *entry, void *p, unsigned long len,
 
 static int dwarf_parse_fde(void *entry, u32 entry_type,
                            void *start, unsigned long len,
-                           unsigned char *end)
+                           unsigned char *end, struct module *mod)
 {
         struct dwarf_fde *fde;
         struct dwarf_cie *cie;
@@ -831,6 +843,8 @@ static int dwarf_parse_fde(void *entry, u32 entry_type,
         fde->instructions = p;
         fde->end = end;
 
+        fde->mod = mod;
+
         /* Add to list. */
         spin_lock_irqsave(&dwarf_fde_lock, flags);
         list_add_tail(&fde->link, &dwarf_fde_list);
@@ -854,10 +868,8 @@ static void dwarf_unwinder_dump(struct task_struct *task,
         while (1) {
                 frame = dwarf_unwind_stack(return_addr, _frame);
 
-                if (_frame) {
-                        dwarf_frame_free_regs(_frame);
-                        mempool_free(_frame, dwarf_frame_pool);
-                }
+                if (_frame)
+                        dwarf_free_frame(_frame);
 
                 _frame = frame;
 
@@ -867,6 +879,9 @@ static void dwarf_unwinder_dump(struct task_struct *task,
                 return_addr = frame->return_addr;
                 ops->address(data, return_addr, 1);
         }
+
+        if (frame)
+                dwarf_free_frame(frame);
 }
 
 static struct unwinder dwarf_unwinder = {
@@ -896,48 +911,28 @@ static void dwarf_unwinder_cleanup(void)
 }
 
 /**
- *      dwarf_unwinder_init - initialise the dwarf unwinder
+ *      dwarf_parse_section - parse DWARF section
+ *      @eh_frame_start: start address of the .eh_frame section
+ *      @eh_frame_end: end address of the .eh_frame section
+ *      @mod: the kernel module containing the .eh_frame section
  *
- *      Build the data structures describing the .dwarf_frame section to
- *      make it easier to lookup CIE and FDE entries. Because the
- *      .eh_frame section is packed as tightly as possible it is not
- *      easy to lookup the FDE for a given PC, so we build a list of FDE
- *      and CIE entries that make it easier.
+ *      Parse the information in a .eh_frame section.
  */
-static int __init dwarf_unwinder_init(void)
+static int dwarf_parse_section(char *eh_frame_start, char *eh_frame_end,
+                               struct module *mod)
 {
         u32 entry_type;
         void *p, *entry;
         int count, err = 0;
-        unsigned long len;
+        unsigned long len = 0;
         unsigned int c_entries, f_entries;
         unsigned char *end;
-        INIT_LIST_HEAD(&dwarf_cie_list);
-        INIT_LIST_HEAD(&dwarf_fde_list);
 
         c_entries = 0;
         f_entries = 0;
-        entry = &__start_eh_frame;
-
-        dwarf_frame_cachep = kmem_cache_create("dwarf_frames",
-                        sizeof(struct dwarf_frame), 0,
-                        SLAB_PANIC | SLAB_HWCACHE_ALIGN | SLAB_NOTRACK, NULL);
-
-        dwarf_reg_cachep = kmem_cache_create("dwarf_regs",
-                        sizeof(struct dwarf_reg), 0,
-                        SLAB_PANIC | SLAB_HWCACHE_ALIGN | SLAB_NOTRACK, NULL);
+        entry = eh_frame_start;
 
-        dwarf_frame_pool = mempool_create(DWARF_FRAME_MIN_REQ,
-                                          mempool_alloc_slab,
-                                          mempool_free_slab,
-                                          dwarf_frame_cachep);
-
-        dwarf_reg_pool = mempool_create(DWARF_REG_MIN_REQ,
-                                         mempool_alloc_slab,
-                                         mempool_free_slab,
-                                         dwarf_reg_cachep);
-
-        while ((char *)entry < __stop_eh_frame) {
+        while ((char *)entry < eh_frame_end) {
                 p = entry;
 
                 count = dwarf_entry_len(p, &len);
@@ -949,6 +944,7 @@ static int __init dwarf_unwinder_init(void)
                          * entry and move to the next one because 'len'
                          * tells us where our next entry is.
                          */
+                        err = -EINVAL;
                         goto out;
                 } else
                         p += count;
@@ -960,13 +956,14 @@ static int __init dwarf_unwinder_init(void)
                 p += 4;
 
                 if (entry_type == DW_EH_FRAME_CIE) {
-                        err = dwarf_parse_cie(entry, p, len, end);
+                        err = dwarf_parse_cie(entry, p, len, end, mod);
                         if (err < 0)
                                 goto out;
                         else
                                 c_entries++;
                 } else {
-                        err = dwarf_parse_fde(entry, entry_type, p, len, end);
+                        err = dwarf_parse_fde(entry, entry_type, p, len,
+                                              end, mod);
                         if (err < 0)
                                 goto out;
                         else
@@ -979,6 +976,129 @@ static int __init dwarf_unwinder_init(void)
         printk(KERN_INFO "DWARF unwinder initialised: read %u CIEs, %u FDEs\n",
                c_entries, f_entries);
 
+        return 0;
+
+out:
+        return err;
+}
+
+#ifdef CONFIG_MODULES
+int module_dwarf_finalize(const Elf_Ehdr *hdr, const Elf_Shdr *sechdrs,
+                          struct module *me)
+{
+        unsigned int i, err;
+        unsigned long start, end;
+        char *secstrings = (void *)hdr + sechdrs[hdr->e_shstrndx].sh_offset;
+
+        start = end = 0;
+
+        for (i = 1; i < hdr->e_shnum; i++) {
+                /* Alloc bit cleared means "ignore it." */
+                if ((sechdrs[i].sh_flags & SHF_ALLOC)
+                    && !strcmp(secstrings+sechdrs[i].sh_name, ".eh_frame")) {
+                        start = sechdrs[i].sh_addr;
+                        end = start + sechdrs[i].sh_size;
+                        break;
+                }
+        }
+
+        /* Did we find the .eh_frame section? */
+        if (i != hdr->e_shnum) {
+                err = dwarf_parse_section((char *)start, (char *)end, me);
+                if (err) {
+                        printk(KERN_WARNING "%s: failed to parse DWARF info\n",
+                               me->name);
+                        return err;
+                }
+        }
+
+        return 0;
+}
+
+/**
+ *      module_dwarf_cleanup - remove FDE/CIEs associated with @mod
+ *      @mod: the module that is being unloaded
+ *
+ *      Remove any FDEs and CIEs from the global lists that came from
+ *      @mod's .eh_frame section because @mod is being unloaded.
+ */
+void module_dwarf_cleanup(struct module *mod)
+{
+        struct dwarf_fde *fde;
+        struct dwarf_cie *cie;
+        unsigned long flags;
+
+        spin_lock_irqsave(&dwarf_cie_lock, flags);
+
+again_cie:
+        list_for_each_entry(cie, &dwarf_cie_list, link) {
+                if (cie->mod == mod)
+                        break;
+        }
+
+        if (&cie->link != &dwarf_cie_list) {
+                list_del(&cie->link);
+                kfree(cie);
+                goto again_cie;
+        }
+
+        spin_unlock_irqrestore(&dwarf_cie_lock, flags);
+
+        spin_lock_irqsave(&dwarf_fde_lock, flags);
+
+again_fde:
+        list_for_each_entry(fde, &dwarf_fde_list, link) {
+                if (fde->mod == mod)
+                        break;
+        }
+
+        if (&fde->link != &dwarf_fde_list) {
+                list_del(&fde->link);
+                kfree(fde);
+                goto again_fde;
+        }
+
+        spin_unlock_irqrestore(&dwarf_fde_lock, flags);
+}
+#endif /* CONFIG_MODULES */
+
+/**
+ *      dwarf_unwinder_init - initialise the dwarf unwinder
+ *
+ *      Build the data structures describing the .dwarf_frame section to
+ *      make it easier to lookup CIE and FDE entries. Because the
+ *      .eh_frame section is packed as tightly as possible it is not
+ *      easy to lookup the FDE for a given PC, so we build a list of FDE
+ *      and CIE entries that make it easier.
+ */
+static int __init dwarf_unwinder_init(void)
+{
+        int err;
+        INIT_LIST_HEAD(&dwarf_cie_list);
+        INIT_LIST_HEAD(&dwarf_fde_list);
+
+        dwarf_frame_cachep = kmem_cache_create("dwarf_frames",
+                        sizeof(struct dwarf_frame), 0,
+                        SLAB_PANIC | SLAB_HWCACHE_ALIGN | SLAB_NOTRACK, NULL);
+
+        dwarf_reg_cachep = kmem_cache_create("dwarf_regs",
+                        sizeof(struct dwarf_reg), 0,
+                        SLAB_PANIC | SLAB_HWCACHE_ALIGN | SLAB_NOTRACK, NULL);
+
+        dwarf_frame_pool = mempool_create(DWARF_FRAME_MIN_REQ,
+                                          mempool_alloc_slab,
+                                          mempool_free_slab,
+                                          dwarf_frame_cachep);
+
+        dwarf_reg_pool = mempool_create(DWARF_REG_MIN_REQ,
+                                         mempool_alloc_slab,
+                                         mempool_free_slab,
+                                         dwarf_reg_cachep);
+
+        err = dwarf_parse_section(__start_eh_frame, __stop_eh_frame, NULL);
+        if (err)
+                goto out;
+
         err = unwinder_register(&dwarf_unwinder);
         if (err)
                 goto out;
diff --git a/arch/sh/kernel/entry-common.S b/arch/sh/kernel/entry-common.S
index 3eb84931d2aa..f0abd58c3a69 100644
--- a/arch/sh/kernel/entry-common.S
+++ b/arch/sh/kernel/entry-common.S
@@ -133,7 +133,7 @@ work_pending:
         ! r8: current_thread_info
         ! t:  result of "tst    #_TIF_NEED_RESCHED, r0"
         bf/s    work_resched
-         tst    #(_TIF_SIGPENDING | _TIF_RESTORE_SIGMASK), r0
+         tst    #_TIF_SIGPENDING, r0
 work_notifysig:
         bt/s    __restore_all
          mov    r15, r4
diff --git a/arch/sh/kernel/ftrace.c b/arch/sh/kernel/ftrace.c
index 2c48e267256e..b6f41c109beb 100644
--- a/arch/sh/kernel/ftrace.c
+++ b/arch/sh/kernel/ftrace.c
@@ -62,6 +62,150 @@ static unsigned char *ftrace_call_replace(unsigned long ip, unsigned long addr)
         return ftrace_replaced_code;
 }
 
+/*
+ * Modifying code must take extra care. On an SMP machine, if
+ * the code being modified is also being executed on another CPU
+ * that CPU will have undefined results and possibly take a GPF.
+ * We use kstop_machine to stop other CPUS from exectuing code.
+ * But this does not stop NMIs from happening. We still need
+ * to protect against that. We separate out the modification of
+ * the code to take care of this.
+ *
+ * Two buffers are added: An IP buffer and a "code" buffer.
+ *
+ * 1) Put the instruction pointer into the IP buffer
+ *    and the new code into the "code" buffer.
+ * 2) Wait for any running NMIs to finish and set a flag that says
+ *    we are modifying code, it is done in an atomic operation.
+ * 3) Write the code
+ * 4) clear the flag.
+ * 5) Wait for any running NMIs to finish.
+ *
+ * If an NMI is executed, the first thing it does is to call
+ * "ftrace_nmi_enter". This will check if the flag is set to write
+ * and if it is, it will write what is in the IP and "code" buffers.
+ *
+ * The trick is, it does not matter if everyone is writing the same
+ * content to the code location. Also, if a CPU is executing code
+ * it is OK to write to that code location if the contents being written
+ * are the same as what exists.
+ */
+#define MOD_CODE_WRITE_FLAG (1 << 31)   /* set when NMI should do the write */
+static atomic_t nmi_running = ATOMIC_INIT(0);
+static int mod_code_status;             /* holds return value of text write */
+static void *mod_code_ip;               /* holds the IP to write to */
+static void *mod_code_newcode;          /* holds the text to write to the IP */
+
+static unsigned nmi_wait_count;
+static atomic_t nmi_update_count = ATOMIC_INIT(0);
+
+int ftrace_arch_read_dyn_info(char *buf, int size)
+{
+        int r;
+
+        r = snprintf(buf, size, "%u %u",
+                     nmi_wait_count,
+                     atomic_read(&nmi_update_count));
+        return r;
+}
+
+static void clear_mod_flag(void)
+{
+        int old = atomic_read(&nmi_running);
+
+        for (;;) {
+                int new = old & ~MOD_CODE_WRITE_FLAG;
+
+                if (old == new)
+                        break;
+
+                old = atomic_cmpxchg(&nmi_running, old, new);
+        }
+}
+
+static void ftrace_mod_code(void)
+{
+        /*
+         * Yes, more than one CPU process can be writing to mod_code_status.
+         *    (and the code itself)
+         * But if one were to fail, then they all should, and if one were
+         * to succeed, then they all should.
+         */
+        mod_code_status = probe_kernel_write(mod_code_ip, mod_code_newcode,
+                                             MCOUNT_INSN_SIZE);
+
+        /* if we fail, then kill any new writers */
+        if (mod_code_status)
+                clear_mod_flag();
+}
+
+void ftrace_nmi_enter(void)
+{
+        if (atomic_inc_return(&nmi_running) & MOD_CODE_WRITE_FLAG) {
+                smp_rmb();
+                ftrace_mod_code();
+                atomic_inc(&nmi_update_count);
+        }
+        /* Must have previous changes seen before executions */
+        smp_mb();
+}
+
+void ftrace_nmi_exit(void)
+{
+        /* Finish all executions before clearing nmi_running */
+        smp_mb();
+        atomic_dec(&nmi_running);
+}
+
+static void wait_for_nmi_and_set_mod_flag(void)
+{
+        if (!atomic_cmpxchg(&nmi_running, 0, MOD_CODE_WRITE_FLAG))
+                return;
+
+        do {
+                cpu_relax();
+        } while (atomic_cmpxchg(&nmi_running, 0, MOD_CODE_WRITE_FLAG));
+
+        nmi_wait_count++;
+}
+
+static void wait_for_nmi(void)
+{
+        if (!atomic_read(&nmi_running))
+                return;
+
+        do {
+                cpu_relax();
+        } while (atomic_read(&nmi_running));
+
+        nmi_wait_count++;
+}
+
+static int
+do_ftrace_mod_code(unsigned long ip, void *new_code)
+{
+        mod_code_ip = (void *)ip;
+        mod_code_newcode = new_code;
+
+        /* The buffers need to be visible before we let NMIs write them */
+        smp_mb();
+
+        wait_for_nmi_and_set_mod_flag();
+
+        /* Make sure all running NMIs have finished before we write the code */
+        smp_mb();
+
+        ftrace_mod_code();
+
+        /* Make sure the write happens before clearing the bit */
+        smp_mb();
+
+        clear_mod_flag();
+        wait_for_nmi();
+
+        return mod_code_status;
+}
+
 static int ftrace_modify_code(unsigned long ip, unsigned char *old_code,
                        unsigned char *new_code)
 {
@@ -86,7 +230,7 @@ static int ftrace_modify_code(unsigned long ip, unsigned char *old_code,
                 return -EINVAL;
 
         /* replace the text with the new text */
-        if (probe_kernel_write((void *)ip, new_code, MCOUNT_INSN_SIZE))
+        if (do_ftrace_mod_code(ip, new_code))
                 return -EPERM;
 
         flush_icache_range(ip, ip + MCOUNT_INSN_SIZE);
diff --git a/arch/sh/kernel/gpio.c b/arch/sh/kernel/gpio.c
deleted file mode 100644
index d22e5af699f9..000000000000
--- a/arch/sh/kernel/gpio.c
+++ /dev/null
@@ -1,584 +0,0 @@
-/*
- * Pinmuxed GPIO support for SuperH.
- *
- * Copyright (C) 2008 Magnus Damm
- *
- * 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/errno.h>
-#include <linux/kernel.h>
-#include <linux/list.h>
-#include <linux/module.h>
-#include <linux/clk.h>
-#include <linux/err.h>
-#include <linux/io.h>
-#include <linux/irq.h>
-#include <linux/bitops.h>
-#include <linux/gpio.h>
-
-static int enum_in_range(pinmux_enum_t enum_id, struct pinmux_range *r)
-{
-        if (enum_id < r->begin)
-                return 0;
-
-        if (enum_id > r->end)
-                return 0;
-
-        return 1;
-}
-
-static unsigned long gpio_read_raw_reg(unsigned long reg,
-                                       unsigned long reg_width)
-{
-        switch (reg_width) {
-        case 8:
-                return ctrl_inb(reg);
-        case 16:
-                return ctrl_inw(reg);
-        case 32:
-                return ctrl_inl(reg);
-        }
-
-        BUG();
-        return 0;
-}
-
-static void gpio_write_raw_reg(unsigned long reg,
-                               unsigned long reg_width,
-                               unsigned long data)
-{
-        switch (reg_width) {
-        case 8:
-                ctrl_outb(data, reg);
-                return;
-        case 16:
-                ctrl_outw(data, reg);
-                return;
-        case 32:
-                ctrl_outl(data, reg);
-                return;
-        }
-
-        BUG();
-}
-
-static void gpio_write_bit(struct pinmux_data_reg *dr,
-                           unsigned long in_pos, unsigned long value)
-{
-        unsigned long pos;
-
-        pos = dr->reg_width - (in_pos + 1);
-
-#ifdef DEBUG
-        pr_info("write_bit addr = %lx, value = %ld, pos = %ld, "
-                "r_width = %ld\n",
-                dr->reg, !!value, pos, dr->reg_width);
-#endif
-
-        if (value)
-                set_bit(pos, &dr->reg_shadow);
-        else
-                clear_bit(pos, &dr->reg_shadow);
-
-        gpio_write_raw_reg(dr->reg, dr->reg_width, dr->reg_shadow);
-}
-
-static int gpio_read_reg(unsigned long reg, unsigned long reg_width,
-                         unsigned long field_width, unsigned long in_pos)
-{
-        unsigned long data, mask, pos;
-
-        data = 0;
-        mask = (1 << field_width) - 1;
-        pos = reg_width - ((in_pos + 1) * field_width);
-
-#ifdef DEBUG
-        pr_info("read_reg: addr = %lx, pos = %ld, "
-                "r_width = %ld, f_width = %ld\n",
-                reg, pos, reg_width, field_width);
-#endif
-
-        data = gpio_read_raw_reg(reg, reg_width);
-        return (data >> pos) & mask;
-}
-
-static void gpio_write_reg(unsigned long reg, unsigned long reg_width,
-                           unsigned long field_width, unsigned long in_pos,
-                           unsigned long value)
-{
-        unsigned long mask, pos;
-
-        mask = (1 << field_width) - 1;
-        pos = reg_width - ((in_pos + 1) * field_width);
-
-#ifdef DEBUG
-        pr_info("write_reg addr = %lx, value = %ld, pos = %ld, "
-                "r_width = %ld, f_width = %ld\n",
-                reg, value, pos, reg_width, field_width);
-#endif
-
-        mask = ~(mask << pos);
-        value = value << pos;
-
-        switch (reg_width) {
-        case 8:
-                ctrl_outb((ctrl_inb(reg) & mask) | value, reg);
-                break;
-        case 16:
-                ctrl_outw((ctrl_inw(reg) & mask) | value, reg);
-                break;
-        case 32:
-                ctrl_outl((ctrl_inl(reg) & mask) | value, reg);
-                break;
-        }
-}
-
-static int setup_data_reg(struct pinmux_info *gpioc, unsigned gpio)
-{
-        struct pinmux_gpio *gpiop = &gpioc->gpios[gpio];
-        struct pinmux_data_reg *data_reg;
-        int k, n;
-
-        if (!enum_in_range(gpiop->enum_id, &gpioc->data))
-                return -1;
-
-        k = 0;
-        while (1) {
-                data_reg = gpioc->data_regs + k;
-
-                if (!data_reg->reg_width)
-                        break;
-
-                for (n = 0; n < data_reg->reg_width; n++) {
-                        if (data_reg->enum_ids[n] == gpiop->enum_id) {
-                                gpiop->flags &= ~PINMUX_FLAG_DREG;
-                                gpiop->flags |= (k << PINMUX_FLAG_DREG_SHIFT);
-                                gpiop->flags &= ~PINMUX_FLAG_DBIT;
-                                gpiop->flags |= (n << PINMUX_FLAG_DBIT_SHIFT);
-                                return 0;
-                        }
-                }
-                k++;
-        }
-
-        BUG();
-
-        return -1;
-}
-
-static void setup_data_regs(struct pinmux_info *gpioc)
-{
-        struct pinmux_data_reg *drp;
-        int k;
-
-        for (k = gpioc->first_gpio; k <= gpioc->last_gpio; k++)
-                setup_data_reg(gpioc, k);
-
-        k = 0;
-        while (1) {
-                drp = gpioc->data_regs + k;
-
-                if (!drp->reg_width)
-                        break;
-
-                drp->reg_shadow = gpio_read_raw_reg(drp->reg, drp->reg_width);
-                k++;
-        }
-}
-
-static int get_data_reg(struct pinmux_info *gpioc, unsigned gpio,
-                        struct pinmux_data_reg **drp, int *bitp)
-{
-        struct pinmux_gpio *gpiop = &gpioc->gpios[gpio];
-        int k, n;
-
-        if (!enum_in_range(gpiop->enum_id, &gpioc->data))
-                return -1;
-
-        k = (gpiop->flags & PINMUX_FLAG_DREG) >> PINMUX_FLAG_DREG_SHIFT;
-        n = (gpiop->flags & PINMUX_FLAG_DBIT) >> PINMUX_FLAG_DBIT_SHIFT;
-        *drp = gpioc->data_regs + k;
-        *bitp = n;
-        return 0;
-}
-
-static int get_config_reg(struct pinmux_info *gpioc, pinmux_enum_t enum_id,
-                          struct pinmux_cfg_reg **crp, int *indexp,
-                          unsigned long **cntp)
-{
-        struct pinmux_cfg_reg *config_reg;
-        unsigned long r_width, f_width;
-        int k, n;
-
-        k = 0;
-        while (1) {
-                config_reg = gpioc->cfg_regs + k;
-
-                r_width = config_reg->reg_width;
-                f_width = config_reg->field_width;
-
-                if (!r_width)
-                        break;
-                for (n = 0; n < (r_width / f_width) * 1 << f_width; n++) {
-                        if (config_reg->enum_ids[n] == enum_id) {
-                                *crp = config_reg;
-                                *indexp = n;
-                                *cntp = &config_reg->cnt[n / (1 << f_width)];
-                                return 0;
-                        }
-                }
-                k++;
-        }
-
-        return -1;
-}
-
-static int get_gpio_enum_id(struct pinmux_info *gpioc, unsigned gpio,
-                            int pos, pinmux_enum_t *enum_idp)
-{
-        pinmux_enum_t enum_id = gpioc->gpios[gpio].enum_id;
-        pinmux_enum_t *data = gpioc->gpio_data;
-        int k;
-
-        if (!enum_in_range(enum_id, &gpioc->data)) {
-                if (!enum_in_range(enum_id, &gpioc->mark)) {
-                        pr_err("non data/mark enum_id for gpio %d\n", gpio);
-                        return -1;
-                }
-        }
-
-        if (pos) {
-                *enum_idp = data[pos + 1];
-                return pos + 1;
-        }
-
-        for (k = 0; k < gpioc->gpio_data_size; k++) {
-                if (data[k] == enum_id) {
-                        *enum_idp = data[k + 1];
-                        return k + 1;
-                }
-        }
-
-        pr_err("cannot locate data/mark enum_id for gpio %d\n", gpio);
-        return -1;
-}
-
-static void write_config_reg(struct pinmux_info *gpioc,
-                             struct pinmux_cfg_reg *crp,
-                             int index)
-{
-        unsigned long ncomb, pos, value;
-
-        ncomb = 1 << crp->field_width;
-        pos = index / ncomb;
-        value = index % ncomb;
-
-        gpio_write_reg(crp->reg, crp->reg_width, crp->field_width, pos, value);
-}
-
-static int check_config_reg(struct pinmux_info *gpioc,
-                            struct pinmux_cfg_reg *crp,
-                            int index)
-{
-        unsigned long ncomb, pos, value;
-
-        ncomb = 1 << crp->field_width;
-        pos = index / ncomb;
-        value = index % ncomb;
-
-        if (gpio_read_reg(crp->reg, crp->reg_width,
-                          crp->field_width, pos) == value)
-                return 0;
-
-        return -1;
-}
-
-enum { GPIO_CFG_DRYRUN, GPIO_CFG_REQ, GPIO_CFG_FREE };
-
-static int pinmux_config_gpio(struct pinmux_info *gpioc, unsigned gpio,
-                              int pinmux_type, int cfg_mode)
-{
-        struct pinmux_cfg_reg *cr = NULL;
-        pinmux_enum_t enum_id;
-        struct pinmux_range *range;
-        int in_range, pos, index;
-        unsigned long *cntp;
-
-        switch (pinmux_type) {
-
-        case PINMUX_TYPE_FUNCTION:
-                range = NULL;
-                break;
-
-        case PINMUX_TYPE_OUTPUT:
-                range = &gpioc->output;
-                break;
-
-        case PINMUX_TYPE_INPUT:
-                range = &gpioc->input;
-                break;
-
-        case PINMUX_TYPE_INPUT_PULLUP:
-                range = &gpioc->input_pu;
-                break;
-
-        case PINMUX_TYPE_INPUT_PULLDOWN:
-                range = &gpioc->input_pd;
-                break;
-
-        default:
-                goto out_err;
-        }
-
-        pos = 0;
-        enum_id = 0;
-        index = 0;
-        while (1) {
-                pos = get_gpio_enum_id(gpioc, gpio, pos, &enum_id);
-                if (pos <= 0)
-                        goto out_err;
-
-                if (!enum_id)
-                        break;
-
-                in_range = enum_in_range(enum_id, &gpioc->function);
-                if (!in_range && range) {
-                        in_range = enum_in_range(enum_id, range);
-
-                        if (in_range && enum_id == range->force)
-                                continue;
-                }
-
-                if (!in_range)
-                        continue;
-
-                if (get_config_reg(gpioc, enum_id, &cr, &index, &cntp) != 0)
-                        goto out_err;
-
-                switch (cfg_mode) {
-                case GPIO_CFG_DRYRUN:
-                        if (!*cntp || !check_config_reg(gpioc, cr, index))
-                                continue;
-                        break;
-
-                case GPIO_CFG_REQ:
-                        write_config_reg(gpioc, cr, index);
-                        *cntp = *cntp + 1;
-                        break;
-
-                case GPIO_CFG_FREE:
-                        *cntp = *cntp - 1;
-                        break;
-                }
-        }
-
-        return 0;
- out_err:
-        return -1;
-}
-
-static DEFINE_SPINLOCK(gpio_lock);
-
-static struct pinmux_info *chip_to_pinmux(struct gpio_chip *chip)
-{
-        return container_of(chip, struct pinmux_info, chip);
-}
-
-static int sh_gpio_request(struct gpio_chip *chip, unsigned offset)
-{
-        struct pinmux_info *gpioc = chip_to_pinmux(chip);
-        struct pinmux_data_reg *dummy;
-        unsigned long flags;
-        int i, ret, pinmux_type;
-
-        ret = -EINVAL;
-
-        if (!gpioc)
-                goto err_out;
-
-        spin_lock_irqsave(&gpio_lock, flags);
-
-        if ((gpioc->gpios[offset].flags & PINMUX_FLAG_TYPE) != PINMUX_TYPE_NONE)
-                goto err_unlock;
-
-        /* setup pin function here if no data is associated with pin */
-
-        if (get_data_reg(gpioc, offset, &dummy, &i) != 0)
-                pinmux_type = PINMUX_TYPE_FUNCTION;
-        else
-                pinmux_type = PINMUX_TYPE_GPIO;
-
-        if (pinmux_type == PINMUX_TYPE_FUNCTION) {
-                if (pinmux_config_gpio(gpioc, offset,
-                                       pinmux_type,
-                                       GPIO_CFG_DRYRUN) != 0)
-                        goto err_unlock;
-
-                if (pinmux_config_gpio(gpioc, offset,
-                                       pinmux_type,
-                                       GPIO_CFG_REQ) != 0)
-                        BUG();
-        }
-
-        gpioc->gpios[offset].flags &= ~PINMUX_FLAG_TYPE;
-        gpioc->gpios[offset].flags |= pinmux_type;
-
-        ret = 0;
- err_unlock:
-        spin_unlock_irqrestore(&gpio_lock, flags);
- err_out:
-        return ret;
-}
-
-static void sh_gpio_free(struct gpio_chip *chip, unsigned offset)
-{
-        struct pinmux_info *gpioc = chip_to_pinmux(chip);
-        unsigned long flags;
-        int pinmux_type;
-
-        if (!gpioc)
-                return;
-
-        spin_lock_irqsave(&gpio_lock, flags);
-
-        pinmux_type = gpioc->gpios[offset].flags & PINMUX_FLAG_TYPE;
-        pinmux_config_gpio(gpioc, offset, pinmux_type, GPIO_CFG_FREE);
-        gpioc->gpios[offset].flags &= ~PINMUX_FLAG_TYPE;
-        gpioc->gpios[offset].flags |= PINMUX_TYPE_NONE;
-
-        spin_unlock_irqrestore(&gpio_lock, flags);
-}
-
-static int pinmux_direction(struct pinmux_info *gpioc,
-                            unsigned gpio, int new_pinmux_type)
-{
-        int pinmux_type;
-        int ret = -EINVAL;
-
-        if (!gpioc)
-                goto err_out;
-
-        pinmux_type = gpioc->gpios[gpio].flags & PINMUX_FLAG_TYPE;
-
-        switch (pinmux_type) {
-        case PINMUX_TYPE_GPIO:
-                break;
-        case PINMUX_TYPE_OUTPUT:
-        case PINMUX_TYPE_INPUT:
-        case PINMUX_TYPE_INPUT_PULLUP:
-        case PINMUX_TYPE_INPUT_PULLDOWN:
-                pinmux_config_gpio(gpioc, gpio, pinmux_type, GPIO_CFG_FREE);
-                break;
-        default:
-                goto err_out;
-        }
-
-        if (pinmux_config_gpio(gpioc, gpio,
-                               new_pinmux_type,
-                               GPIO_CFG_DRYRUN) != 0)
-                goto err_out;
-
-        if (pinmux_config_gpio(gpioc, gpio,
-                               new_pinmux_type,
-                               GPIO_CFG_REQ) != 0)
-                BUG();
-
-        gpioc->gpios[gpio].flags &= ~PINMUX_FLAG_TYPE;
-        gpioc->gpios[gpio].flags |= new_pinmux_type;
-
-        ret = 0;
- err_out:
-        return ret;
-}
-
-static int sh_gpio_direction_input(struct gpio_chip *chip, unsigned offset)
-{
-        struct pinmux_info *gpioc = chip_to_pinmux(chip);
-        unsigned long flags;
-        int ret;
-
-        spin_lock_irqsave(&gpio_lock, flags);
-        ret = pinmux_direction(gpioc, offset, PINMUX_TYPE_INPUT);
-        spin_unlock_irqrestore(&gpio_lock, flags);
-
-        return ret;
-}
-
-static void sh_gpio_set_value(struct pinmux_info *gpioc,
-                             unsigned gpio, int value)
-{
-        struct pinmux_data_reg *dr = NULL;
-        int bit = 0;
-
-        if (!gpioc || get_data_reg(gpioc, gpio, &dr, &bit) != 0)
-                BUG();
-        else
-                gpio_write_bit(dr, bit, value);
-}
-
-static int sh_gpio_direction_output(struct gpio_chip *chip, unsigned offset,
-                                    int value)
-{
-        struct pinmux_info *gpioc = chip_to_pinmux(chip);
-        unsigned long flags;
-        int ret;
-
-        sh_gpio_set_value(gpioc, offset, value);
-        spin_lock_irqsave(&gpio_lock, flags);
-        ret = pinmux_direction(gpioc, offset, PINMUX_TYPE_OUTPUT);
-        spin_unlock_irqrestore(&gpio_lock, flags);
-
-        return ret;
-}
-
-static int sh_gpio_get_value(struct pinmux_info *gpioc, unsigned gpio)
-{
-        struct pinmux_data_reg *dr = NULL;
-        int bit = 0;
-
-        if (!gpioc || get_data_reg(gpioc, gpio, &dr, &bit) != 0) {
-                BUG();
-                return 0;
-        }
-
-        return gpio_read_reg(dr->reg, dr->reg_width, 1, bit);
-}
-
-static int sh_gpio_get(struct gpio_chip *chip, unsigned offset)
-{
-        return sh_gpio_get_value(chip_to_pinmux(chip), offset);
-}
-
-static void sh_gpio_set(struct gpio_chip *chip, unsigned offset, int value)
-{
-        sh_gpio_set_value(chip_to_pinmux(chip), offset, value);
-}
-
-int register_pinmux(struct pinmux_info *pip)
-{
-        struct gpio_chip *chip = &pip->chip;
-
-        pr_info("sh pinmux: %s handling gpio %d -> %d\n",
-                pip->name, pip->first_gpio, pip->last_gpio);
-
-        setup_data_regs(pip);
-
-        chip->request = sh_gpio_request;
-        chip->free = sh_gpio_free;
-        chip->direction_input = sh_gpio_direction_input;
-        chip->get = sh_gpio_get;
-        chip->direction_output = sh_gpio_direction_output;
-        chip->set = sh_gpio_set;
-
-        WARN_ON(pip->first_gpio != 0); /* needs testing */
-
-        chip->label = pip->name;
-        chip->owner = THIS_MODULE;
-        chip->base = pip->first_gpio;
-        chip->ngpio = (pip->last_gpio - pip->first_gpio) + 1;
-
-        return gpiochip_add(chip);
-}
diff --git a/arch/sh/kernel/head_32.S b/arch/sh/kernel/head_32.S
index a78be74b8d3e..1151ecdffa71 100644
--- a/arch/sh/kernel/head_32.S
+++ b/arch/sh/kernel/head_32.S
@@ -33,7 +33,7 @@ ENTRY(empty_zero_page)
         .long   1               /* LOADER_TYPE */
         .long   0x00000000      /* INITRD_START */
         .long   0x00000000      /* INITRD_SIZE */
-#ifdef CONFIG_32BIT
+#if defined(CONFIG_32BIT) && defined(CONFIG_PMB_FIXED)
         .long   0x53453f00 + 32 /* "SE?" = 32 bit */
 #else
         .long   0x53453f00 + 29 /* "SE?" = 29 bit */
diff --git a/arch/sh/kernel/idle.c b/arch/sh/kernel/idle.c
index 27ff2dc093c7..aaff0037fcd7 100644
--- a/arch/sh/kernel/idle.c
+++ b/arch/sh/kernel/idle.c
@@ -21,7 +21,7 @@
 #include <asm/atomic.h>
 
 static int hlt_counter;
-void (*pm_idle)(void);
+void (*pm_idle)(void) = NULL;
 void (*pm_power_off)(void);
 EXPORT_SYMBOL(pm_power_off);
 
@@ -39,48 +39,92 @@ static int __init hlt_setup(char *__unused)
 }
 __setup("hlt", hlt_setup);
 
+static inline int hlt_works(void)
+{
+        return !hlt_counter;
+}
+
+/*
+ * On SMP it's slightly faster (but much more power-consuming!)
+ * to poll the ->work.need_resched flag instead of waiting for the
+ * cross-CPU IPI to arrive. Use this option with caution.
+ */
+static void poll_idle(void)
+{
+        local_irq_enable();
+        while (!need_resched())
+                cpu_relax();
+}
+
 void default_idle(void)
 {
-        if (!hlt_counter) {
+        if (hlt_works()) {
                 clear_thread_flag(TIF_POLLING_NRFLAG);
                 smp_mb__after_clear_bit();
-                set_bl_bit();
-                stop_critical_timings();
 
-                while (!need_resched())
+                if (!need_resched()) {
+                        local_irq_enable();
                         cpu_sleep();
+                } else
+                        local_irq_enable();
 
-                start_critical_timings();
-                clear_bl_bit();
                 set_thread_flag(TIF_POLLING_NRFLAG);
         } else
-                while (!need_resched())
-                        cpu_relax();
+                poll_idle();
 }
 
+/*
+ * The idle thread. There's no useful work to be done, so just try to conserve
+ * power and have a low exit latency (ie sit in a loop waiting for somebody to
+ * say that they'd like to reschedule)
+ */
 void cpu_idle(void)
 {
+        unsigned int cpu = smp_processor_id();
+
         set_thread_flag(TIF_POLLING_NRFLAG);
 
         /* endless idle loop with no priority at all */
         while (1) {
-                void (*idle)(void) = pm_idle;
+                tick_nohz_stop_sched_tick(1);
 
-                if (!idle)
-                        idle = default_idle;
+                while (!need_resched() && cpu_online(cpu)) {
+                        check_pgt_cache();
+                        rmb();
 
-                tick_nohz_stop_sched_tick(1);
-                while (!need_resched())
-                        idle();
-                tick_nohz_restart_sched_tick();
+                        local_irq_disable();
+                        /* Don't trace irqs off for idle */
+                        stop_critical_timings();
+                        pm_idle();
+                        /*
+                         * Sanity check to ensure that pm_idle() returns
+                         * with IRQs enabled
+                         */
+                        WARN_ON(irqs_disabled());
+                        start_critical_timings();
+                }
 
+                tick_nohz_restart_sched_tick();
                 preempt_enable_no_resched();
                 schedule();
                 preempt_disable();
-                check_pgt_cache();
         }
 }
 
+void __cpuinit select_idle_routine(void)
+{
+        /*
+         * If a platform has set its own idle routine, leave it alone.
+         */
+        if (pm_idle)
+                return;
+
+        if (hlt_works())
+                pm_idle = default_idle;
+        else
+                pm_idle = poll_idle;
+}
+
 static void do_nothing(void *unused)
 {
 }
diff --git a/arch/sh/kernel/io_generic.c b/arch/sh/kernel/io_generic.c
index b8fa6524760a..e1e1dbd19557 100644
--- a/arch/sh/kernel/io_generic.c
+++ b/arch/sh/kernel/io_generic.c
@@ -24,7 +24,7 @@
 #define dummy_read()
 #endif
 
-unsigned long generic_io_base;
+unsigned long generic_io_base = 0;
 
 u8 generic_inb(unsigned long port)
 {
@@ -147,8 +147,10 @@ void generic_outsl(unsigned long port, const void *src, unsigned long count)
 
 void __iomem *generic_ioport_map(unsigned long addr, unsigned int size)
 {
+#ifdef P1SEG
         if (PXSEG(addr) >= P1SEG)
                 return (void __iomem *)addr;
+#endif
 
         return (void __iomem *)(addr + generic_io_base);
 }
diff --git a/arch/sh/kernel/irq.c b/arch/sh/kernel/irq.c
index eac7da772fc2..e1913f28f418 100644
--- a/arch/sh/kernel/irq.c
+++ b/arch/sh/kernel/irq.c
@@ -37,7 +37,15 @@ void ack_bad_irq(unsigned int irq)
  */
 static int show_other_interrupts(struct seq_file *p, int prec)
 {
+        int j;
+
+        seq_printf(p, "%*s: ", prec, "NMI");
+        for_each_online_cpu(j)
+                seq_printf(p, "%10u ", irq_stat[j].__nmi_count);
+        seq_printf(p, "  Non-maskable interrupts\n");
+
         seq_printf(p, "%*s: %10u\n", prec, "ERR", atomic_read(&irq_err_count));
+
         return 0;
 }
 
@@ -255,6 +263,12 @@ void __init init_IRQ(void)
 {
         plat_irq_setup();
 
+        /*
+         * Pin any of the legacy IRQ vectors that haven't already been
+         * grabbed by the platform
+         */
+        reserve_irq_legacy();
+
         /* Perform the machine specific initialisation */
         if (sh_mv.mv_init_irq)
                 sh_mv.mv_init_irq();
diff --git a/arch/sh/kernel/irq_32.c b/arch/sh/kernel/irq_32.c
new file mode 100644
index 000000000000..e33ab15831f9
--- /dev/null
+++ b/arch/sh/kernel/irq_32.c
@@ -0,0 +1,57 @@
+/*
+ * SHcompact irqflags support
+ *
+ * Copyright (C) 2006 - 2009 Paul Mundt
+ *
+ * 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/irqflags.h>
+#include <linux/module.h>
+
+void notrace raw_local_irq_restore(unsigned long flags)
+{
+        unsigned long __dummy0, __dummy1;
+
+        if (flags == RAW_IRQ_DISABLED) {
+                __asm__ __volatile__ (
+                        "stc    sr, %0\n\t"
+                        "or     #0xf0, %0\n\t"
+                        "ldc    %0, sr\n\t"
+                        : "=&z" (__dummy0)
+                        : /* no inputs */
+                        : "memory"
+                );
+        } else {
+                __asm__ __volatile__ (
+                        "stc    sr, %0\n\t"
+                        "and    %1, %0\n\t"
+#ifdef CONFIG_CPU_HAS_SR_RB
+                        "stc    r6_bank, %1\n\t"
+                        "or     %1, %0\n\t"
+#endif
+                        "ldc    %0, sr\n\t"
+                        : "=&r" (__dummy0), "=r" (__dummy1)
+                        : "1" (~RAW_IRQ_DISABLED)
+                        : "memory"
+                );
+        }
+}
+EXPORT_SYMBOL(raw_local_irq_restore);
+
+unsigned long notrace __raw_local_save_flags(void)
+{
+        unsigned long flags;
+
+        __asm__ __volatile__ (
+                "stc    sr, %0\n\t"
+                "and    #0xf0, %0\n\t"
+                : "=&z" (flags)
+                : /* no inputs */
+                : "memory"
+        );
+
+        return flags;
+}
+EXPORT_SYMBOL(__raw_local_save_flags);
diff --git a/arch/sh/kernel/irq_64.c b/arch/sh/kernel/irq_64.c
new file mode 100644
index 000000000000..32365ba0e039
--- /dev/null
+++ b/arch/sh/kernel/irq_64.c
@@ -0,0 +1,51 @@
+/*
+ * SHmedia irqflags support
+ *
+ * Copyright (C) 2006 - 2009 Paul Mundt
+ *
+ * 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/irqflags.h>
+#include <linux/module.h>
+#include <cpu/registers.h>
+
+void notrace raw_local_irq_restore(unsigned long flags)
+{
+        unsigned long long __dummy;
+
+        if (flags == RAW_IRQ_DISABLED) {
+                __asm__ __volatile__ (
+                        "getcon " __SR ", %0\n\t"
+                        "or     %0, %1, %0\n\t"
+                        "putcon %0, " __SR "\n\t"
+                        : "=&r" (__dummy)
+                        : "r" (RAW_IRQ_DISABLED)
+                );
+        } else {
+                __asm__ __volatile__ (
+                        "getcon " __SR ", %0\n\t"
+                        "and    %0, %1, %0\n\t"
+                        "putcon %0, " __SR "\n\t"
+                        : "=&r" (__dummy)
+                        : "r" (~RAW_IRQ_DISABLED)
+                );
+        }
+}
+EXPORT_SYMBOL(raw_local_irq_restore);
+
+unsigned long notrace __raw_local_save_flags(void)
+{
+        unsigned long flags;
+
+        __asm__ __volatile__ (
+                "getcon " __SR ", %0\n\t"
+                "and    %0, %1, %0"
+                : "=&r" (flags)
+                : "r" (RAW_IRQ_DISABLED)
+        );
+
+        return flags;
+}
+EXPORT_SYMBOL(__raw_local_save_flags);
diff --git a/arch/sh/kernel/machine_kexec.c b/arch/sh/kernel/machine_kexec.c
index 7ea2704ea033..76f280223ebd 100644
--- a/arch/sh/kernel/machine_kexec.c
+++ b/arch/sh/kernel/machine_kexec.c
@@ -46,12 +46,6 @@ void machine_crash_shutdown(struct pt_regs *regs)
  */
 int machine_kexec_prepare(struct kimage *image)
 {
-        /* older versions of kexec-tools are passing
-         * the zImage entry point as a virtual address.
-         */
-        if (image->start != PHYSADDR(image->start))
-                return -EINVAL; /* upgrade your kexec-tools */
-
         return 0;
 }
 
diff --git a/arch/sh/kernel/machvec.c b/arch/sh/kernel/machvec.c
index cbce639b108a..1652340ba3f2 100644
--- a/arch/sh/kernel/machvec.c
+++ b/arch/sh/kernel/machvec.c
@@ -135,5 +135,9 @@ void __init sh_mv_setup(void)
         if (!sh_mv.mv_nr_irqs)
                 sh_mv.mv_nr_irqs = NR_IRQS;
 
+#ifdef P2SEG
         __set_io_port_base(P2SEG);
+#else
+        __set_io_port_base(0);
+#endif
 }
diff --git a/arch/sh/kernel/module.c b/arch/sh/kernel/module.c
index c2efdcde266f..43adddfe4c04 100644
--- a/arch/sh/kernel/module.c
+++ b/arch/sh/kernel/module.c
@@ -32,6 +32,7 @@
 #include <linux/string.h>
 #include <linux/kernel.h>
 #include <asm/unaligned.h>
+#include <asm/dwarf.h>
 
 void *module_alloc(unsigned long size)
 {
@@ -145,10 +146,16 @@ int module_finalize(const Elf_Ehdr *hdr,
                     const Elf_Shdr *sechdrs,
                     struct module *me)
 {
-        return module_bug_finalize(hdr, sechdrs, me);
+        int ret = 0;
+
+        ret |= module_dwarf_finalize(hdr, sechdrs, me);
+        ret |= module_bug_finalize(hdr, sechdrs, me);
+
+        return ret;
 }
 
 void module_arch_cleanup(struct module *mod)
 {
         module_bug_cleanup(mod);
+        module_dwarf_cleanup(mod);
 }
diff --git a/arch/sh/kernel/perf_callchain.c b/arch/sh/kernel/perf_callchain.c
new file mode 100644
index 000000000000..24ea837eac5b
--- /dev/null
+++ b/arch/sh/kernel/perf_callchain.c
@@ -0,0 +1,98 @@
+/*
+ * Performance event callchain support - SuperH architecture code
+ *
+ * Copyright (C) 2009  Paul Mundt
+ *
+ * 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/sched.h>
+#include <linux/perf_event.h>
+#include <linux/percpu.h>
+#include <asm/unwinder.h>
+#include <asm/ptrace.h>
+
+static inline void callchain_store(struct perf_callchain_entry *entry, u64 ip)
+{
+        if (entry->nr < PERF_MAX_STACK_DEPTH)
+                entry->ip[entry->nr++] = ip;
+}
+
+static void callchain_warning(void *data, char *msg)
+{
+}
+
+static void
+callchain_warning_symbol(void *data, char *msg, unsigned long symbol)
+{
+}
+
+static int callchain_stack(void *data, char *name)
+{
+        return 0;
+}
+
+static void callchain_address(void *data, unsigned long addr, int reliable)
+{
+        struct perf_callchain_entry *entry = data;
+
+        if (reliable)
+                callchain_store(entry, addr);
+}
+
+static const struct stacktrace_ops callchain_ops = {
+        .warning        = callchain_warning,
+        .warning_symbol = callchain_warning_symbol,
+        .stack          = callchain_stack,
+        .address        = callchain_address,
+};
+
+static void
+perf_callchain_kernel(struct pt_regs *regs, struct perf_callchain_entry *entry)
+{
+        callchain_store(entry, PERF_CONTEXT_KERNEL);
+        callchain_store(entry, regs->pc);
+
+        unwind_stack(NULL, regs, NULL, &callchain_ops, entry);
+}
+
+static void
+perf_do_callchain(struct pt_regs *regs, struct perf_callchain_entry *entry)
+{
+        int is_user;
+
+        if (!regs)
+                return;
+
+        is_user = user_mode(regs);
+
+        if (!current || current->pid == 0)
+                return;
+
+        if (is_user && current->state != TASK_RUNNING)
+                return;
+
+        /*
+         * Only the kernel side is implemented for now.
+         */
+        if (!is_user)
+                perf_callchain_kernel(regs, entry);
+}
+
+/*
+ * No need for separate IRQ and NMI entries.
+ */
+static DEFINE_PER_CPU(struct perf_callchain_entry, callchain);
+
+struct perf_callchain_entry *perf_callchain(struct pt_regs *regs)
+{
+        struct perf_callchain_entry *entry = &__get_cpu_var(callchain);
+
+        entry->nr = 0;
+
+        perf_do_callchain(regs, entry);
+
+        return entry;
+}
diff --git a/arch/sh/kernel/perf_event.c b/arch/sh/kernel/perf_event.c
new file mode 100644
index 000000000000..7ff0943e7a08
--- /dev/null
+++ b/arch/sh/kernel/perf_event.c
@@ -0,0 +1,312 @@
+/*
+ * Performance event support framework for SuperH hardware counters.
+ *
+ *  Copyright (C) 2009  Paul Mundt
+ *
+ * Heavily based on the x86 and PowerPC implementations.
+ *
+ * x86:
+ *  Copyright (C) 2008 Thomas Gleixner <tglx@linutronix.de>
+ *  Copyright (C) 2008-2009 Red Hat, Inc., Ingo Molnar
+ *  Copyright (C) 2009 Jaswinder Singh Rajput
+ *  Copyright (C) 2009 Advanced Micro Devices, Inc., Robert Richter
+ *  Copyright (C) 2008-2009 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com>
+ *  Copyright (C) 2009 Intel Corporation, <markus.t.metzger@intel.com>
+ *
+ * ppc:
+ *  Copyright 2008-2009 Paul Mackerras, IBM Corporation.
+ *
+ * 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/io.h>
+#include <linux/irq.h>
+#include <linux/perf_event.h>
+#include <asm/processor.h>
+
+struct cpu_hw_events {
+        struct perf_event       *events[MAX_HWEVENTS];
+        unsigned long           used_mask[BITS_TO_LONGS(MAX_HWEVENTS)];
+        unsigned long           active_mask[BITS_TO_LONGS(MAX_HWEVENTS)];
+};
+
+DEFINE_PER_CPU(struct cpu_hw_events, cpu_hw_events);
+
+static struct sh_pmu *sh_pmu __read_mostly;
+
+/* Number of perf_events counting hardware events */
+static atomic_t num_events;
+/* Used to avoid races in calling reserve/release_pmc_hardware */
+static DEFINE_MUTEX(pmc_reserve_mutex);
+
+/*
+ * Stub these out for now, do something more profound later.
+ */
+int reserve_pmc_hardware(void)
+{
+        return 0;
+}
+
+void release_pmc_hardware(void)
+{
+}
+
+static inline int sh_pmu_initialized(void)
+{
+        return !!sh_pmu;
+}
+
+/*
+ * Release the PMU if this is the last perf_event.
+ */
+static void hw_perf_event_destroy(struct perf_event *event)
+{
+        if (!atomic_add_unless(&num_events, -1, 1)) {
+                mutex_lock(&pmc_reserve_mutex);
+                if (atomic_dec_return(&num_events) == 0)
+                        release_pmc_hardware();
+                mutex_unlock(&pmc_reserve_mutex);
+        }
+}
+
+static int hw_perf_cache_event(int config, int *evp)
+{
+        unsigned long type, op, result;
+        int ev;
+
+        if (!sh_pmu->cache_events)
+                return -EINVAL;
+
+        /* unpack config */
+        type = config & 0xff;
+        op = (config >> 8) & 0xff;
+        result = (config >> 16) & 0xff;
+
+        if (type >= PERF_COUNT_HW_CACHE_MAX ||
+            op >= PERF_COUNT_HW_CACHE_OP_MAX ||
+            result >= PERF_COUNT_HW_CACHE_RESULT_MAX)
+                return -EINVAL;
+
+        ev = (*sh_pmu->cache_events)[type][op][result];
+        if (ev == 0)
+                return -EOPNOTSUPP;
+        if (ev == -1)
+                return -EINVAL;
+        *evp = ev;
+        return 0;
+}
+
+static int __hw_perf_event_init(struct perf_event *event)
+{
+        struct perf_event_attr *attr = &event->attr;
+        struct hw_perf_event *hwc = &event->hw;
+        int config = -1;
+        int err;
+
+        if (!sh_pmu_initialized())
+                return -ENODEV;
+
+        /*
+         * All of the on-chip counters are "limited", in that they have
+         * no interrupts, and are therefore unable to do sampling without
+         * further work and timer assistance.
+         */
+        if (hwc->sample_period)
+                return -EINVAL;
+
+        /*
+         * See if we need to reserve the counter.
+         *
+         * If no events are currently in use, then we have to take a
+         * mutex to ensure that we don't race with another task doing
+         * reserve_pmc_hardware or release_pmc_hardware.
+         */
+        err = 0;
+        if (!atomic_inc_not_zero(&num_events)) {
+                mutex_lock(&pmc_reserve_mutex);
+                if (atomic_read(&num_events) == 0 &&
+                    reserve_pmc_hardware())
+                        err = -EBUSY;
+                else
+                        atomic_inc(&num_events);
+                mutex_unlock(&pmc_reserve_mutex);
+        }
+
+        if (err)
+                return err;
+
+        event->destroy = hw_perf_event_destroy;
+
+        switch (attr->type) {
+        case PERF_TYPE_RAW:
+                config = attr->config & sh_pmu->raw_event_mask;
+                break;
+        case PERF_TYPE_HW_CACHE:
+                err = hw_perf_cache_event(attr->config, &config);
+                if (err)
+                        return err;
+                break;
+        case PERF_TYPE_HARDWARE:
+                if (attr->config >= sh_pmu->max_events)
+                        return -EINVAL;
+
+                config = sh_pmu->event_map(attr->config);
+                break;
+        }
+
+        if (config == -1)
+                return -EINVAL;
+
+        hwc->config |= config;
+
+        return 0;
+}
+
+static void sh_perf_event_update(struct perf_event *event,
+                                   struct hw_perf_event *hwc, int idx)
+{
+        u64 prev_raw_count, new_raw_count;
+        s64 delta;
+        int shift = 0;
+
+        /*
+         * Depending on the counter configuration, they may or may not
+         * be chained, in which case the previous counter value can be
+         * updated underneath us if the lower-half overflows.
+         *
+         * Our tactic to handle this is to first atomically read and
+         * exchange a new raw count - then add that new-prev delta
+         * count to the generic counter atomically.
+         *
+         * As there is no interrupt associated with the overflow events,
+         * this is the simplest approach for maintaining consistency.
+         */
+again:
+        prev_raw_count = atomic64_read(&hwc->prev_count);
+        new_raw_count = sh_pmu->read(idx);
+
+        if (atomic64_cmpxchg(&hwc->prev_count, prev_raw_count,
+                             new_raw_count) != prev_raw_count)
+                goto again;
+
+        /*
+         * Now we have the new raw value and have updated the prev
+         * timestamp already. We can now calculate the elapsed delta
+         * (counter-)time and add that to the generic counter.
+         *
+         * Careful, not all hw sign-extends above the physical width
+         * of the count.
+         */
+        delta = (new_raw_count << shift) - (prev_raw_count << shift);
+        delta >>= shift;
+
+        atomic64_add(delta, &event->count);
+}
+
+static void sh_pmu_disable(struct perf_event *event)
+{
+        struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+        struct hw_perf_event *hwc = &event->hw;
+        int idx = hwc->idx;
+
+        clear_bit(idx, cpuc->active_mask);
+        sh_pmu->disable(hwc, idx);
+
+        barrier();
+
+        sh_perf_event_update(event, &event->hw, idx);
+
+        cpuc->events[idx] = NULL;
+        clear_bit(idx, cpuc->used_mask);
+
+        perf_event_update_userpage(event);
+}
+
+static int sh_pmu_enable(struct perf_event *event)
+{
+        struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events);
+        struct hw_perf_event *hwc = &event->hw;
+        int idx = hwc->idx;
+
+        if (test_and_set_bit(idx, cpuc->used_mask)) {
+                idx = find_first_zero_bit(cpuc->used_mask, sh_pmu->num_events);
+                if (idx == sh_pmu->num_events)
+                        return -EAGAIN;
+
+                set_bit(idx, cpuc->used_mask);
+                hwc->idx = idx;
+        }
+
+        sh_pmu->disable(hwc, idx);
+
+        cpuc->events[idx] = event;
+        set_bit(idx, cpuc->active_mask);
+
+        sh_pmu->enable(hwc, idx);
+
+        perf_event_update_userpage(event);
+
+        return 0;
+}
+
+static void sh_pmu_read(struct perf_event *event)
+{
+        sh_perf_event_update(event, &event->hw, event->hw.idx);
+}
+
+static const struct pmu pmu = {
+        .enable         = sh_pmu_enable,
+        .disable        = sh_pmu_disable,
+        .read           = sh_pmu_read,
+};
+
+const struct pmu *hw_perf_event_init(struct perf_event *event)
+{
+        int err = __hw_perf_event_init(event);
+        if (unlikely(err)) {
+                if (event->destroy)
+                        event->destroy(event);
+                return ERR_PTR(err);
+        }
+
+        return &pmu;
+}
+
+void hw_perf_event_setup(int cpu)
+{
+        struct cpu_hw_events *cpuhw = &per_cpu(cpu_hw_events, cpu);
+
+        memset(cpuhw, 0, sizeof(struct cpu_hw_events));
+}
+
+void hw_perf_enable(void)
+{
+        if (!sh_pmu_initialized())
+                return;
+
+        sh_pmu->enable_all();
+}
+
+void hw_perf_disable(void)
+{
+        if (!sh_pmu_initialized())
+                return;
+
+        sh_pmu->disable_all();
+}
+
+int register_sh_pmu(struct sh_pmu *pmu)
+{
+        if (sh_pmu)
+                return -EBUSY;
+        sh_pmu = pmu;
+
+        pr_info("Performance Events: %s support registered\n", pmu->name);
+
+        WARN_ON(pmu->num_events > MAX_HWEVENTS);
+
+        return 0;
+}
diff --git a/arch/sh/kernel/process_32.c b/arch/sh/kernel/process_32.c
index 4a2c866f9773..7399d78fc8ed 100644
--- a/arch/sh/kernel/process_32.c
+++ b/arch/sh/kernel/process_32.c
@@ -133,7 +133,10 @@ int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)
         regs.regs[5] = (unsigned long)fn;
 
         regs.pc = (unsigned long)kernel_thread_helper;
-        regs.sr = (1 << 30);
+        regs.sr = SR_MD;
+#if defined(CONFIG_SH_FPU)
+        regs.sr |= SR_FD;
+#endif
 
         /* Ok, create the new process.. */
         pid = do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0,
@@ -141,6 +144,7 @@ int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)
 
         return pid;
 }
+EXPORT_SYMBOL(kernel_thread);
 
 /*
  * Free current thread data structures etc..
@@ -184,6 +188,16 @@ int dump_fpu(struct pt_regs *regs, elf_fpregset_t *fpu)
 
         return fpvalid;
 }
+EXPORT_SYMBOL(dump_fpu);
+
+/*
+ * This gets called before we allocate a new thread and copy
+ * the current task into it.
+ */
+void prepare_to_copy(struct task_struct *tsk)
+{
+        unlazy_fpu(tsk, task_pt_regs(tsk));
+}
 
 asmlinkage void ret_from_fork(void);
 
@@ -193,15 +207,10 @@ int copy_thread(unsigned long clone_flags, unsigned long usp,
 {
         struct thread_info *ti = task_thread_info(p);
         struct pt_regs *childregs;
-        struct task_struct *tsk = current;
-
-#if defined(CONFIG_SH_FPU)
-        unlazy_fpu(tsk, regs);
-        p->thread.fpu = tsk->thread.fpu;
-        copy_to_stopped_child_used_math(p);
-#endif
 
 #if defined(CONFIG_SH_DSP)
+        struct task_struct *tsk = current;
+
         if (is_dsp_enabled(tsk)) {
                 /* We can use the __save_dsp or just copy the struct:
                  * __save_dsp(p);
@@ -220,6 +229,8 @@ int copy_thread(unsigned long clone_flags, unsigned long usp,
         } else {
                 childregs->regs[15] = (unsigned long)childregs;
                 ti->addr_limit = KERNEL_DS;
+                ti->status &= ~TS_USEDFPU;
+                p->fpu_counter = 0;
         }
 
         if (clone_flags & CLONE_SETTLS)
@@ -242,9 +253,13 @@ int copy_thread(unsigned long clone_flags, unsigned long usp,
 __notrace_funcgraph struct task_struct *
 __switch_to(struct task_struct *prev, struct task_struct *next)
 {
-#if defined(CONFIG_SH_FPU)
+        struct thread_struct *next_t = &next->thread;
+
         unlazy_fpu(prev, task_pt_regs(prev));
-#endif
+
+        /* we're going to use this soon, after a few expensive things */
+        if (next->fpu_counter > 5)
+                prefetch(&next_t->fpu.hard);
 
 #ifdef CONFIG_MMU
         /*
@@ -256,6 +271,14 @@ __switch_to(struct task_struct *prev, struct task_struct *next)
                      : "r" (task_thread_info(next)));
 #endif
 
+        /*
+         * If the task has used fpu the last 5 timeslices, just do a full
+         * restore of the math state immediately to avoid the trap; the
+         * chances of needing FPU soon are obviously high now
+         */
+        if (next->fpu_counter > 5)
+                fpu_state_restore(task_pt_regs(next));
+
         return prev;
 }
 
diff --git a/arch/sh/kernel/process_64.c b/arch/sh/kernel/process_64.c
index 1192398ef582..359b8a2f4d2e 100644
--- a/arch/sh/kernel/process_64.c
+++ b/arch/sh/kernel/process_64.c
@@ -335,6 +335,7 @@ int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)
         return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0,
                       &regs, 0, NULL, NULL);
 }
+EXPORT_SYMBOL(kernel_thread);
 
 /*
  * Free current thread data structures etc..
@@ -417,6 +418,7 @@ int dump_fpu(struct pt_regs *regs, elf_fpregset_t *fpu)
         return 0; /* Task didn't use the fpu at all. */
 #endif
 }
+EXPORT_SYMBOL(dump_fpu);
 
 asmlinkage void ret_from_fork(void);
 
diff --git a/arch/sh/kernel/return_address.c b/arch/sh/kernel/return_address.c
new file mode 100644
index 000000000000..df3ab5811074
--- /dev/null
+++ b/arch/sh/kernel/return_address.c
@@ -0,0 +1,54 @@
+/*
+ * arch/sh/kernel/return_address.c
+ *
+ * Copyright (C) 2009  Matt Fleming
+ * Copyright (C) 2009  Paul Mundt
+ *
+ * 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 <asm/dwarf.h>
+
+#ifdef CONFIG_DWARF_UNWINDER
+
+void *return_address(unsigned int depth)
+{
+        struct dwarf_frame *frame;
+        unsigned long ra;
+        int i;
+
+        for (i = 0, frame = NULL, ra = 0; i <= depth; i++) {
+                struct dwarf_frame *tmp;
+
+                tmp = dwarf_unwind_stack(ra, frame);
+
+                if (frame)
+                        dwarf_free_frame(frame);
+
+                frame = tmp;
+
+                if (!frame || !frame->return_addr)
+                        break;
+
+                ra = frame->return_addr;
+        }
+
+        /* Failed to unwind the stack to the specified depth. */
+        WARN_ON(i != depth + 1);
+
+        if (frame)
+                dwarf_free_frame(frame);
+
+        return (void *)ra;
+}
+
+#else
+
+void *return_address(unsigned int depth)
+{
+        return NULL;
+}
+
+#endif
diff --git a/arch/sh/kernel/setup.c b/arch/sh/kernel/setup.c
index 99b4fb553bf1..5a947a2567e4 100644
--- a/arch/sh/kernel/setup.c
+++ b/arch/sh/kernel/setup.c
@@ -453,6 +453,10 @@ void __init setup_arch(char **cmdline_p)
 
         paging_init();
 
+#ifdef CONFIG_PMB_ENABLE
+        pmb_init();
+#endif
+
 #ifdef CONFIG_SMP
         plat_smp_setup();
 #endif
diff --git a/arch/sh/kernel/sh_ksyms_32.c b/arch/sh/kernel/sh_ksyms_32.c
index 444cce3ae921..3896f26efa4a 100644
--- a/arch/sh/kernel/sh_ksyms_32.c
+++ b/arch/sh/kernel/sh_ksyms_32.c
@@ -1,37 +1,11 @@
 #include <linux/module.h>
-#include <linux/smp.h>
-#include <linux/user.h>
-#include <linux/elfcore.h>
-#include <linux/sched.h>
-#include <linux/in6.h>
-#include <linux/interrupt.h>
-#include <linux/vmalloc.h>
-#include <linux/pci.h>
-#include <linux/irq.h>
-#include <asm/sections.h>
-#include <asm/processor.h>
-#include <asm/uaccess.h>
+#include <linux/string.h>
+#include <linux/uaccess.h>
+#include <linux/delay.h>
+#include <linux/mm.h>
 #include <asm/checksum.h>
-#include <asm/io.h>
-#include <asm/delay.h>
-#include <asm/tlbflush.h>
-#include <asm/cacheflush.h>
-#include <asm/ftrace.h>
-
-extern int dump_fpu(struct pt_regs *, elf_fpregset_t *);
-
-/* platform dependent support */
-EXPORT_SYMBOL(dump_fpu);
-EXPORT_SYMBOL(kernel_thread);
-EXPORT_SYMBOL(strlen);
-
-/* PCI exports */
-#ifdef CONFIG_PCI
-EXPORT_SYMBOL(pci_alloc_consistent);
-EXPORT_SYMBOL(pci_free_consistent);
-#endif
+#include <asm/sections.h>
 
-/* mem exports */
 EXPORT_SYMBOL(memchr);
 EXPORT_SYMBOL(memcpy);
 EXPORT_SYMBOL(memset);
@@ -40,6 +14,13 @@ EXPORT_SYMBOL(__copy_user);
 EXPORT_SYMBOL(__udelay);
 EXPORT_SYMBOL(__ndelay);
 EXPORT_SYMBOL(__const_udelay);
+EXPORT_SYMBOL(strlen);
+EXPORT_SYMBOL(csum_partial);
+EXPORT_SYMBOL(csum_partial_copy_generic);
+EXPORT_SYMBOL(copy_page);
+EXPORT_SYMBOL(__clear_user);
+EXPORT_SYMBOL(_ebss);
+EXPORT_SYMBOL(empty_zero_page);
 
 #define DECLARE_EXPORT(name)            \
         extern void name(void);EXPORT_SYMBOL(name)
@@ -107,30 +88,6 @@ DECLARE_EXPORT(__sdivsi3_i4);
 DECLARE_EXPORT(__udivsi3_i4);
 DECLARE_EXPORT(__sdivsi3_i4i);
 DECLARE_EXPORT(__udivsi3_i4i);
-
-#if !defined(CONFIG_CACHE_OFF) && (defined(CONFIG_CPU_SH4) || \
-        defined(CONFIG_SH7705_CACHE_32KB))
-/* needed by some modules */
-EXPORT_SYMBOL(flush_cache_all);
-EXPORT_SYMBOL(flush_cache_range);
-EXPORT_SYMBOL(flush_dcache_page);
-#endif
-
 #ifdef CONFIG_MCOUNT
 DECLARE_EXPORT(mcount);
 #endif
-EXPORT_SYMBOL(csum_partial);
-EXPORT_SYMBOL(csum_partial_copy_generic);
-#ifdef CONFIG_IPV6
-EXPORT_SYMBOL(csum_ipv6_magic);
-#endif
-EXPORT_SYMBOL(copy_page);
-EXPORT_SYMBOL(__clear_user);
-EXPORT_SYMBOL(_ebss);
-EXPORT_SYMBOL(empty_zero_page);
-
-#ifndef CONFIG_CACHE_OFF
-EXPORT_SYMBOL(__flush_purge_region);
-EXPORT_SYMBOL(__flush_wback_region);
-EXPORT_SYMBOL(__flush_invalidate_region);
-#endif
diff --git a/arch/sh/kernel/sh_ksyms_64.c b/arch/sh/kernel/sh_ksyms_64.c
index d008e17eb257..45afa5c51f67 100644
--- a/arch/sh/kernel/sh_ksyms_64.c
+++ b/arch/sh/kernel/sh_ksyms_64.c
@@ -24,16 +24,6 @@
 #include <asm/delay.h>
 #include <asm/irq.h>
 
-extern int dump_fpu(struct pt_regs *, elf_fpregset_t *);
-
-/* platform dependent support */
-EXPORT_SYMBOL(dump_fpu);
-EXPORT_SYMBOL(kernel_thread);
-
-#ifdef CONFIG_VT
-EXPORT_SYMBOL(screen_info);
-#endif
-
 EXPORT_SYMBOL(__put_user_asm_b);
 EXPORT_SYMBOL(__put_user_asm_w);
 EXPORT_SYMBOL(__put_user_asm_l);
diff --git a/arch/sh/kernel/signal_32.c b/arch/sh/kernel/signal_32.c
index 3db37425210d..12815ce01ecd 100644
--- a/arch/sh/kernel/signal_32.c
+++ b/arch/sh/kernel/signal_32.c
@@ -67,7 +67,8 @@ sys_sigsuspend(old_sigset_t mask,
 
         current->state = TASK_INTERRUPTIBLE;
         schedule();
-        set_thread_flag(TIF_RESTORE_SIGMASK);
+        set_restore_sigmask();
+
         return -ERESTARTNOHAND;
 }
 
@@ -590,7 +591,7 @@ static void do_signal(struct pt_regs *regs, unsigned int save_r0)
         if (try_to_freeze())
                 goto no_signal;
 
-        if (test_thread_flag(TIF_RESTORE_SIGMASK))
+        if (current_thread_info()->status & TS_RESTORE_SIGMASK)
                 oldset = &current->saved_sigmask;
         else
                 oldset = &current->blocked;
@@ -602,12 +603,13 @@ static void do_signal(struct pt_regs *regs, unsigned int save_r0)
                 /* Whee!  Actually deliver the signal.  */
                 if (handle_signal(signr, &ka, &info, oldset,
                                   regs, save_r0) == 0) {
-                        /* a signal was successfully delivered; the saved
+                        /*
+                         * A signal was successfully delivered; the saved
                          * sigmask will have been stored in the signal frame,
                          * and will be restored by sigreturn, so we can simply
-                         * clear the TIF_RESTORE_SIGMASK flag */
-                        if (test_thread_flag(TIF_RESTORE_SIGMASK))
-                                clear_thread_flag(TIF_RESTORE_SIGMASK);
+                         * clear the TS_RESTORE_SIGMASK flag
+                         */
+                        current_thread_info()->status &= ~TS_RESTORE_SIGMASK;
 
                         tracehook_signal_handler(signr, &info, &ka, regs,
                                         test_thread_flag(TIF_SINGLESTEP));
@@ -631,10 +633,12 @@ no_signal:
                 }
         }
 
-        /* if there's no signal to deliver, we just put the saved sigmask
-         * back */
-        if (test_thread_flag(TIF_RESTORE_SIGMASK)) {
-                clear_thread_flag(TIF_RESTORE_SIGMASK);
+        /*
+         * If there's no signal to deliver, we just put the saved sigmask
+         * back.
+         */
+        if (current_thread_info()->status & TS_RESTORE_SIGMASK) {
+                current_thread_info()->status &= ~TS_RESTORE_SIGMASK;
                 sigprocmask(SIG_SETMASK, &current->saved_sigmask, NULL);
         }
 }
diff --git a/arch/sh/kernel/signal_64.c b/arch/sh/kernel/signal_64.c
index 74793c80a57a..feb3dddd3192 100644
--- a/arch/sh/kernel/signal_64.c
+++ b/arch/sh/kernel/signal_64.c
@@ -101,7 +101,7 @@ static int do_signal(struct pt_regs *regs, sigset_t *oldset)
         if (try_to_freeze())
                 goto no_signal;
 
-        if (test_thread_flag(TIF_RESTORE_SIGMASK))
+        if (current_thread_info()->status & TS_RESTORE_SIGMASK)
                 oldset = &current->saved_sigmask;
         else if (!oldset)
                 oldset = &current->blocked;
@@ -115,11 +115,9 @@ static int do_signal(struct pt_regs *regs, sigset_t *oldset)
                         /*
                          * If a signal was successfully delivered, the
                          * saved sigmask is in its frame, and we can
-                         * clear the TIF_RESTORE_SIGMASK flag.
+                         * clear the TS_RESTORE_SIGMASK flag.
                          */
-                        if (test_thread_flag(TIF_RESTORE_SIGMASK))
-                                clear_thread_flag(TIF_RESTORE_SIGMASK);
-
+                        current_thread_info()->status &= ~TS_RESTORE_SIGMASK;
                         tracehook_signal_handler(signr, &info, &ka, regs, 0);
                         return 1;
                 }
@@ -146,8 +144,8 @@ no_signal:
         }
 
         /* No signal to deliver -- put the saved sigmask back */
-        if (test_thread_flag(TIF_RESTORE_SIGMASK)) {
-                clear_thread_flag(TIF_RESTORE_SIGMASK);
+        if (current_thread_info()->status & TS_RESTORE_SIGMASK) {
+                current_thread_info()->status &= ~TS_RESTORE_SIGMASK;
                 sigprocmask(SIG_SETMASK, &current->saved_sigmask, NULL);
         }
 
@@ -176,6 +174,7 @@ sys_sigsuspend(old_sigset_t mask,
         while (1) {
                 current->state = TASK_INTERRUPTIBLE;
                 schedule();
+                set_restore_sigmask();
                 regs->pc += 4;    /* because sys_sigreturn decrements the pc */
                 if (do_signal(regs, &saveset)) {
                         /* pc now points at signal handler. Need to decrement
diff --git a/arch/sh/kernel/smp.c b/arch/sh/kernel/smp.c
index 160db1003cfb..983e0792d5f3 100644
--- a/arch/sh/kernel/smp.c
+++ b/arch/sh/kernel/smp.c
@@ -122,7 +122,9 @@ int __cpuinit __cpu_up(unsigned int cpu)
         stack_start.bss_start = 0; /* don't clear bss for secondary cpus */
         stack_start.start_kernel_fn = start_secondary;
 
-        flush_cache_all();
+        flush_icache_range((unsigned long)&stack_start,
+                           (unsigned long)&stack_start + sizeof(stack_start));
+        wmb();
 
         plat_start_cpu(cpu, (unsigned long)_stext);
 
diff --git a/arch/sh/kernel/topology.c b/arch/sh/kernel/topology.c
index 0838942b7083..9b0b633b6c92 100644
--- a/arch/sh/kernel/topology.c
+++ b/arch/sh/kernel/topology.c
@@ -16,6 +16,32 @@
 
 static DEFINE_PER_CPU(struct cpu, cpu_devices);
 
+cpumask_t cpu_core_map[NR_CPUS];
+
+static cpumask_t cpu_coregroup_map(unsigned int cpu)
+{
+        /*
+         * Presently all SH-X3 SMP cores are multi-cores, so just keep it
+         * simple until we have a method for determining topology..
+         */
+        return cpu_possible_map;
+}
+
+const struct cpumask *cpu_coregroup_mask(unsigned int cpu)
+{
+        return &cpu_core_map[cpu];
+}
+
+int arch_update_cpu_topology(void)
+{
+        unsigned int cpu;
+
+        for_each_possible_cpu(cpu)
+                cpu_core_map[cpu] = cpu_coregroup_map(cpu);
+
+        return 0;
+}
+
 static int __init topology_init(void)
 {
         int i, ret;
diff --git a/arch/sh/kernel/traps.c b/arch/sh/kernel/traps.c
index a8396f36bd14..7b036339dc92 100644
--- a/arch/sh/kernel/traps.c
+++ b/arch/sh/kernel/traps.c
@@ -9,8 +9,8 @@
 #include <asm/unwinder.h>
 #include <asm/system.h>
 
-#ifdef CONFIG_BUG
-void handle_BUG(struct pt_regs *regs)
+#ifdef CONFIG_GENERIC_BUG
+static void handle_BUG(struct pt_regs *regs)
 {
         const struct bug_entry *bug;
         unsigned long bugaddr = regs->pc;
@@ -81,7 +81,7 @@ BUILD_TRAP_HANDLER(bug)
                        SIGTRAP) == NOTIFY_STOP)
                 return;
 
-#ifdef CONFIG_BUG
+#ifdef CONFIG_GENERIC_BUG
         if (__kernel_text_address(instruction_pointer(regs))) {
                 insn_size_t insn = *(insn_size_t *)instruction_pointer(regs);
                 if (insn == TRAPA_BUG_OPCODE)
@@ -95,9 +95,11 @@ BUILD_TRAP_HANDLER(bug)
 
 BUILD_TRAP_HANDLER(nmi)
 {
+        unsigned int cpu = smp_processor_id();
         TRAP_HANDLER_DECL;
 
         nmi_enter();
+        nmi_count(cpu)++;
 
         switch (notify_die(DIE_NMI, "NMI", regs, 0, vec & 0xff, SIGINT)) {
         case NOTIFY_OK:
diff --git a/arch/sh/kernel/traps_32.c b/arch/sh/kernel/traps_32.c
index 7a2ee3a6b8e7..3da5a125d884 100644
--- a/arch/sh/kernel/traps_32.c
+++ b/arch/sh/kernel/traps_32.c
@@ -25,6 +25,7 @@
 #include <linux/kexec.h>
 #include <linux/limits.h>
 #include <linux/proc_fs.h>
+#include <linux/seq_file.h>
 #include <linux/sysfs.h>
 #include <asm/system.h>
 #include <asm/uaccess.h>
@@ -68,61 +69,49 @@ static const char *se_usermode_action[] = {
         "signal+warn"
 };
 
-static int
-proc_alignment_read(char *page, char **start, off_t off, int count, int *eof,
-                    void *data)
+static int alignment_proc_show(struct seq_file *m, void *v)
 {
-        char *p = page;
-        int len;
-
-        p += sprintf(p, "User:\t\t%lu\n", se_user);
-        p += sprintf(p, "System:\t\t%lu\n", se_sys);
-        p += sprintf(p, "Half:\t\t%lu\n", se_half);
-        p += sprintf(p, "Word:\t\t%lu\n", se_word);
-        p += sprintf(p, "DWord:\t\t%lu\n", se_dword);
-        p += sprintf(p, "Multi:\t\t%lu\n", se_multi);
-        p += sprintf(p, "User faults:\t%i (%s)\n", se_usermode,
+        seq_printf(m, "User:\t\t%lu\n", se_user);
+        seq_printf(m, "System:\t\t%lu\n", se_sys);
+        seq_printf(m, "Half:\t\t%lu\n", se_half);
+        seq_printf(m, "Word:\t\t%lu\n", se_word);
+        seq_printf(m, "DWord:\t\t%lu\n", se_dword);
+        seq_printf(m, "Multi:\t\t%lu\n", se_multi);
+        seq_printf(m, "User faults:\t%i (%s)\n", se_usermode,
                         se_usermode_action[se_usermode]);
-        p += sprintf(p, "Kernel faults:\t%i (fixup%s)\n", se_kernmode_warn,
+        seq_printf(m, "Kernel faults:\t%i (fixup%s)\n", se_kernmode_warn,
                         se_kernmode_warn ? "+warn" : "");
-
-        len = (p - page) - off;
-        if (len < 0)
-                len = 0;
-
-        *eof = (len <= count) ? 1 : 0;
-        *start = page + off;
-
-        return len;
+        return 0;
 }
 
-static int proc_alignment_write(struct file *file, const char __user *buffer,
-                                unsigned long count, void *data)
+static int alignment_proc_open(struct inode *inode, struct file *file)
 {
-        char mode;
-
-        if (count > 0) {
-                if (get_user(mode, buffer))
-                        return -EFAULT;
-                if (mode >= '0' && mode <= '5')
-                        se_usermode = mode - '0';
-        }
-        return count;
+        return single_open(file, alignment_proc_show, NULL);
 }
 
-static int proc_alignment_kern_write(struct file *file, const char __user *buffer,
-                                     unsigned long count, void *data)
+static ssize_t alignment_proc_write(struct file *file,
+                const char __user *buffer, size_t count, loff_t *pos)
 {
+        int *data = PDE(file->f_path.dentry->d_inode)->data;
         char mode;
 
         if (count > 0) {
                 if (get_user(mode, buffer))
                         return -EFAULT;
-                if (mode >= '0' && mode <= '1')
-                        se_kernmode_warn = mode - '0';
+                if (mode >= '0' && mode <= '5')
+                        *data = mode - '0';
         }
         return count;
 }
+
+static const struct file_operations alignment_proc_fops = {
+        .owner          = THIS_MODULE,
+        .open           = alignment_proc_open,
+        .read           = seq_read,
+        .llseek         = seq_lseek,
+        .release        = single_release,
+        .write          = alignment_proc_write,
+};
 #endif
 
 static void dump_mem(const char *str, unsigned long bottom, unsigned long top)
@@ -945,14 +934,9 @@ void __init trap_init(void)
         set_exception_table_evt(0x800, do_reserved_inst);
         set_exception_table_evt(0x820, do_illegal_slot_inst);
 #elif defined(CONFIG_SH_FPU)
-#ifdef CONFIG_CPU_SUBTYPE_SHX3
-        set_exception_table_evt(0xd80, fpu_state_restore_trap_handler);
-        set_exception_table_evt(0xda0, fpu_state_restore_trap_handler);
-#else
         set_exception_table_evt(0x800, fpu_state_restore_trap_handler);
         set_exception_table_evt(0x820, fpu_state_restore_trap_handler);
 #endif
-#endif
 
 #ifdef CONFIG_CPU_SH2
         set_exception_table_vec(TRAP_ADDRESS_ERROR, address_error_trap_handler);
@@ -1011,20 +995,16 @@ static int __init alignment_init(void)
         if (!dir)
                 return -ENOMEM;
 
-        res = create_proc_entry("alignment", S_IWUSR | S_IRUGO, dir);
+        res = proc_create_data("alignment", S_IWUSR | S_IRUGO, dir,
+                               &alignment_proc_fops, &se_usermode);
         if (!res)
                 return -ENOMEM;
 
-        res->read_proc = proc_alignment_read;
-        res->write_proc = proc_alignment_write;
-
-        res = create_proc_entry("kernel_alignment", S_IWUSR | S_IRUGO, dir);
+        res = proc_create_data("kernel_alignment", S_IWUSR | S_IRUGO, dir,
+                               &alignment_proc_fops, &se_kernmode_warn);
         if (!res)
                 return -ENOMEM;
 
-        res->read_proc = proc_alignment_read;
-        res->write_proc = proc_alignment_kern_write;
-
         return 0;
 }
 
diff --git a/arch/sh/lib/Makefile b/arch/sh/lib/Makefile
index a969b47c5463..dab4d2129812 100644
--- a/arch/sh/lib/Makefile
+++ b/arch/sh/lib/Makefile
@@ -2,7 +2,7 @@
 # Makefile for SuperH-specific library files..
 #
 
-lib-y  = delay.o memset.o memmove.o memchr.o \
+lib-y  = delay.o memmove.o memchr.o \
          checksum.o strlen.o div64.o div64-generic.o
 
 # Extracted from libgcc
@@ -23,8 +23,11 @@ obj-y				+= io.o
 memcpy-y                        := memcpy.o
 memcpy-$(CONFIG_CPU_SH4)        := memcpy-sh4.o
 
+memset-y                        := memset.o
+memset-$(CONFIG_CPU_SH4)        := memset-sh4.o
+
 lib-$(CONFIG_MMU)               += copy_page.o __clear_user.o
 lib-$(CONFIG_MCOUNT)            += mcount.o
-lib-y                           += $(memcpy-y) $(udivsi3-y)
+lib-y                           += $(memcpy-y) $(memset-y) $(udivsi3-y)
 
 EXTRA_CFLAGS += -Werror
diff --git a/arch/sh/lib/memset-sh4.S b/arch/sh/lib/memset-sh4.S
new file mode 100644
index 000000000000..1a6e32cc4e4d
--- /dev/null
+++ b/arch/sh/lib/memset-sh4.S
@@ -0,0 +1,107 @@
+/*
+ * "memset" implementation for SH4
+ *
+ * Copyright (C) 1999  Niibe Yutaka
+ * Copyright (c) 2009  STMicroelectronics Limited
+ * Author: Stuart Menefy <stuart.menefy:st.com>
+ */
+
+/*
+ *            void *memset(void *s, int c, size_t n);
+ */
+
+#include <linux/linkage.h>
+
+ENTRY(memset)
+        mov     #12,r0
+        add     r6,r4
+        cmp/gt  r6,r0
+        bt/s    40f             ! if it's too small, set a byte at once
+         mov    r4,r0
+        and     #3,r0
+        cmp/eq  #0,r0
+        bt/s    2f              ! It's aligned
+         sub    r0,r6
+1:
+        dt      r0
+        bf/s    1b
+         mov.b  r5,@-r4
+2:                              ! make VVVV
+        extu.b  r5,r5
+        swap.b  r5,r0           !   V0
+        or      r0,r5           !   VV
+        swap.w  r5,r0           ! VV00
+        or      r0,r5           ! VVVV
+
+        ! Check if enough bytes need to be copied to be worth the big loop
+        mov     #0x40, r0       ! (MT)
+        cmp/gt  r6,r0           ! (MT)  64 > len => slow loop
+
+        bt/s    22f
+         mov    r6,r0
+
+        ! align the dst to the cache block size if necessary
+        mov     r4, r3
+        mov     #~(0x1f), r1
+
+        and     r3, r1
+        cmp/eq  r3, r1
+
+        bt/s    11f             ! dst is already aligned
+         sub    r1, r3          ! r3-r1 -> r3
+        shlr2   r3              ! number of loops
+
+10:     mov.l   r5,@-r4
+        dt      r3
+        bf/s    10b
+         add    #-4, r6
+
+11:     ! dst is 32byte aligned
+        mov     r6,r2
+        mov     #-5,r0
+        shld    r0,r2           ! number of loops
+
+        add     #-32, r4
+        mov     r5, r0
+12:
+        movca.l r0,@r4
+        mov.l   r5,@(4, r4)
+        mov.l   r5,@(8, r4)
+        mov.l   r5,@(12,r4)
+        mov.l   r5,@(16,r4)
+        mov.l   r5,@(20,r4)
+        add     #-0x20, r6
+        mov.l   r5,@(24,r4)
+        dt      r2
+        mov.l   r5,@(28,r4)
+        bf/s    12b
+         add    #-32, r4
+
+        add     #32, r4
+        mov     #8, r0
+        cmp/ge  r0, r6
+        bf      40f
+
+        mov     r6,r0
+22:
+        shlr2   r0
+        shlr    r0              ! r0 = r6 >> 3
+3:
+        dt      r0
+        mov.l   r5,@-r4         ! set 8-byte at once
+        bf/s    3b
+         mov.l  r5,@-r4
+        !
+        mov     #7,r0
+        and     r0,r6
+
+        ! fill bytes (length may be zero)
+40:     tst     r6,r6
+        bt      5f
+4:
+        dt      r6
+        bf/s    4b
+         mov.b  r5,@-r4
+5:
+        rts
+         mov    r4,r0
diff --git a/arch/sh/math-emu/math.c b/arch/sh/math-emu/math.c
index ac2d7abd2567..d6c15cae0912 100644
--- a/arch/sh/math-emu/math.c
+++ b/arch/sh/math-emu/math.c
@@ -558,7 +558,7 @@ static int ieee_fpe_handler(struct pt_regs *regs)
                                             (finsn >> 8) & 0xf);
                         tsk->thread.fpu.hard.fpscr &=
                                 ~(FPSCR_CAUSE_MASK | FPSCR_FLAG_MASK);
-                        set_tsk_thread_flag(tsk, TIF_USEDFPU);
+                        task_thread_info(tsk)->status |= TS_USEDFPU;
                 } else {
                         info.si_signo = SIGFPE;
                         info.si_errno = 0;
@@ -619,10 +619,10 @@ int do_fpu_inst(unsigned short inst, struct pt_regs *regs)
         struct task_struct *tsk = current;
         struct sh_fpu_soft_struct *fpu = &(tsk->thread.fpu.soft);
 
-        if (!test_tsk_thread_flag(tsk, TIF_USEDFPU)) {
+        if (!(task_thread_info(tsk)->status & TS_USEDFPU)) {
                 /* initialize once. */
                 fpu_init(fpu);
-                set_tsk_thread_flag(tsk, TIF_USEDFPU);
+                task_thread_info(tsk)->status |= TS_USEDFPU;
         }
 
         return fpu_emulate(inst, fpu, regs);
diff --git a/arch/sh/mm/Kconfig b/arch/sh/mm/Kconfig
index 7f7b52f9beba..0e7ba8e891cf 100644
--- a/arch/sh/mm/Kconfig
+++ b/arch/sh/mm/Kconfig
@@ -82,8 +82,7 @@ config 32BIT
 
 config PMB_ENABLE
         bool "Support 32-bit physical addressing through PMB"
-        depends on MMU && EXPERIMENTAL && (CPU_SUBTYPE_SH7757 || CPU_SUBTYPE_SH7780 || CPU_SUBTYPE_SH7785)
-        select 32BIT
+        depends on MMU && EXPERIMENTAL && CPU_SH4A
         default y
         help
           If you say Y here, physical addressing will be extended to
@@ -97,8 +96,7 @@ choice
 
 config PMB
         bool "PMB"
-        depends on MMU && EXPERIMENTAL && (CPU_SUBTYPE_SH7757 || CPU_SUBTYPE_SH7780 || CPU_SUBTYPE_SH7785)
-        select 32BIT
+        depends on MMU && EXPERIMENTAL && CPU_SH4A
         help
           If you say Y here, physical addressing will be extended to
           32-bits through the SH-4A PMB. If this is not set, legacy
@@ -106,9 +104,7 @@ config PMB
 
 config PMB_FIXED
         bool "fixed PMB"
-        depends on MMU && EXPERIMENTAL && (CPU_SUBTYPE_SH7757 || \
-                                           CPU_SUBTYPE_SH7780 || \
-                                           CPU_SUBTYPE_SH7785)
+        depends on MMU && EXPERIMENTAL && CPU_SH4A
         select 32BIT
         help
           If this option is enabled, fixed PMB mappings are inherited
@@ -258,6 +254,15 @@ endchoice
 
 source "mm/Kconfig"
 
+config SCHED_MC
+        bool "Multi-core scheduler support"
+        depends on SMP
+        default y
+        help
+          Multi-core scheduler support improves the CPU scheduler's decision
+          making when dealing with multi-core CPU chips at a cost of slightly
+          increased overhead in some places. If unsure say N here.
+
 endmenu
 
 menu "Cache configuration"
diff --git a/arch/sh/mm/Makefile b/arch/sh/mm/Makefile
index 3759bf853293..8a70535fa7ce 100644
--- a/arch/sh/mm/Makefile
+++ b/arch/sh/mm/Makefile
@@ -33,8 +33,7 @@ obj-y				+= $(tlb-y)
 endif
 
 obj-$(CONFIG_HUGETLB_PAGE)      += hugetlbpage.o
-obj-$(CONFIG_PMB)               += pmb.o
-obj-$(CONFIG_PMB_FIXED)         += pmb-fixed.o
+obj-$(CONFIG_PMB_ENABLE)        += pmb.o
 obj-$(CONFIG_NUMA)              += numa.o
 
 # Special flags for fault_64.o.  This puts restrictions on the number of
diff --git a/arch/sh/mm/cache-sh4.c b/arch/sh/mm/cache-sh4.c
index b7f235c74d66..f36a08bf3d5c 100644
--- a/arch/sh/mm/cache-sh4.c
+++ b/arch/sh/mm/cache-sh4.c
@@ -2,7 +2,7 @@
  * arch/sh/mm/cache-sh4.c
  *
  * Copyright (C) 1999, 2000, 2002  Niibe Yutaka
- * Copyright (C) 2001 - 2007  Paul Mundt
+ * Copyright (C) 2001 - 2009  Paul Mundt
  * Copyright (C) 2003  Richard Curnow
  * Copyright (c) 2007 STMicroelectronics (R&D) Ltd.
  *
@@ -15,6 +15,8 @@
 #include <linux/io.h>
 #include <linux/mutex.h>
 #include <linux/fs.h>
+#include <linux/highmem.h>
+#include <asm/pgtable.h>
 #include <asm/mmu_context.h>
 #include <asm/cacheflush.h>
 
@@ -23,21 +25,12 @@
  * flushing. Anything exceeding this will simply flush the dcache in its
  * entirety.
  */
-#define MAX_DCACHE_PAGES        64      /* XXX: Tune for ways */
 #define MAX_ICACHE_PAGES        32
 
 static void __flush_cache_one(unsigned long addr, unsigned long phys,
                                unsigned long exec_offset);
 
 /*
- * This is initialised here to ensure that it is not placed in the BSS.  If
- * that were to happen, note that cache_init gets called before the BSS is
- * cleared, so this would get nulled out which would be hopeless.
- */
-static void (*__flush_dcache_segment_fn)(unsigned long, unsigned long) =
-        (void (*)(unsigned long, unsigned long))0xdeadbeef;
-
-/*
  * Write back the range of D-cache, and purge the I-cache.
  *
  * Called from kernel/module.c:sys_init_module and routine for a.out format,
@@ -97,15 +90,15 @@ static inline void flush_cache_one(unsigned long start, unsigned long phys)
         unsigned long flags, exec_offset = 0;
 
         /*
-         * All types of SH-4 require PC to be in P2 to operate on the I-cache.
-         * Some types of SH-4 require PC to be in P2 to operate on the D-cache.
+         * All types of SH-4 require PC to be uncached to operate on the I-cache.
+         * Some types of SH-4 require PC to be uncached to operate on the D-cache.
          */
         if ((boot_cpu_data.flags & CPU_HAS_P2_FLUSH_BUG) ||
             (start < CACHE_OC_ADDRESS_ARRAY))
-                exec_offset = 0x20000000;
+                exec_offset = cached_to_uncached;
 
         local_irq_save(flags);
-        __flush_cache_one(start | SH_CACHE_ASSOC, P1SEGADDR(phys), exec_offset);
+        __flush_cache_one(start, phys, exec_offset);
         local_irq_restore(flags);
 }
 
@@ -124,7 +117,7 @@ static void sh4_flush_dcache_page(void *arg)
         else
 #endif
         {
-                unsigned long phys = PHYSADDR(page_address(page));
+                unsigned long phys = page_to_phys(page);
                 unsigned long addr = CACHE_OC_ADDRESS_ARRAY;
                 int i, n;
 
@@ -159,10 +152,27 @@ static void __uses_jump_to_uncached flush_icache_all(void)
         local_irq_restore(flags);
 }
 
-static inline void flush_dcache_all(void)
+static void flush_dcache_all(void)
 {
-        (*__flush_dcache_segment_fn)(0UL, boot_cpu_data.dcache.way_size);
-        wmb();
+        unsigned long addr, end_addr, entry_offset;
+
+        end_addr = CACHE_OC_ADDRESS_ARRAY +
+                (current_cpu_data.dcache.sets <<
+                 current_cpu_data.dcache.entry_shift) *
+                        current_cpu_data.dcache.ways;
+
+        entry_offset = 1 << current_cpu_data.dcache.entry_shift;
+
+        for (addr = CACHE_OC_ADDRESS_ARRAY; addr < end_addr; ) {
+                __raw_writel(0, addr); addr += entry_offset;
+                __raw_writel(0, addr); addr += entry_offset;
+                __raw_writel(0, addr); addr += entry_offset;
+                __raw_writel(0, addr); addr += entry_offset;
+                __raw_writel(0, addr); addr += entry_offset;
+                __raw_writel(0, addr); addr += entry_offset;
+                __raw_writel(0, addr); addr += entry_offset;
+                __raw_writel(0, addr); addr += entry_offset;
+        }
 }
 
 static void sh4_flush_cache_all(void *unused)
@@ -171,89 +181,13 @@ static void sh4_flush_cache_all(void *unused)
         flush_icache_all();
 }
 
-static void __flush_cache_mm(struct mm_struct *mm, unsigned long start,
-                             unsigned long end)
-{
-        unsigned long d = 0, p = start & PAGE_MASK;
-        unsigned long alias_mask = boot_cpu_data.dcache.alias_mask;
-        unsigned long n_aliases = boot_cpu_data.dcache.n_aliases;
-        unsigned long select_bit;
-        unsigned long all_aliases_mask;
-        unsigned long addr_offset;
-        pgd_t *dir;
-        pmd_t *pmd;
-        pud_t *pud;
-        pte_t *pte;
-        int i;
-
-        dir = pgd_offset(mm, p);
-        pud = pud_offset(dir, p);
-        pmd = pmd_offset(pud, p);
-        end = PAGE_ALIGN(end);
-
-        all_aliases_mask = (1 << n_aliases) - 1;
-
-        do {
-                if (pmd_none(*pmd) || unlikely(pmd_bad(*pmd))) {
-                        p &= PMD_MASK;
-                        p += PMD_SIZE;
-                        pmd++;
-
-                        continue;
-                }
-
-                pte = pte_offset_kernel(pmd, p);
-
-                do {
-                        unsigned long phys;
-                        pte_t entry = *pte;
-
-                        if (!(pte_val(entry) & _PAGE_PRESENT)) {
-                                pte++;
-                                p += PAGE_SIZE;
-                                continue;
-                        }
-
-                        phys = pte_val(entry) & PTE_PHYS_MASK;
-
-                        if ((p ^ phys) & alias_mask) {
-                                d |= 1 << ((p & alias_mask) >> PAGE_SHIFT);
-                                d |= 1 << ((phys & alias_mask) >> PAGE_SHIFT);
-
-                                if (d == all_aliases_mask)
-                                        goto loop_exit;
-                        }
-
-                        pte++;
-                        p += PAGE_SIZE;
-                } while (p < end && ((unsigned long)pte & ~PAGE_MASK));
-                pmd++;
-        } while (p < end);
-
-loop_exit:
-        addr_offset = 0;
-        select_bit = 1;
-
-        for (i = 0; i < n_aliases; i++) {
-                if (d & select_bit) {
-                        (*__flush_dcache_segment_fn)(addr_offset, PAGE_SIZE);
-                        wmb();
-                }
-
-                select_bit <<= 1;
-                addr_offset += PAGE_SIZE;
-        }
-}
-
 /*
  * Note : (RPC) since the caches are physically tagged, the only point
  * of flush_cache_mm for SH-4 is to get rid of aliases from the
  * D-cache.  The assumption elsewhere, e.g. flush_cache_range, is that
  * lines can stay resident so long as the virtual address they were
  * accessed with (hence cache set) is in accord with the physical
- * address (i.e. tag).  It's no different here.  So I reckon we don't
- * need to flush the I-cache, since aliases don't matter for that.  We
- * should try that.
+ * address (i.e. tag).  It's no different here.
  *
  * Caller takes mm->mmap_sem.
  */
@@ -264,33 +198,7 @@ static void sh4_flush_cache_mm(void *arg)
         if (cpu_context(smp_processor_id(), mm) == NO_CONTEXT)
                 return;
 
-        /*
-         * If cache is only 4k-per-way, there are never any 'aliases'.  Since
-         * the cache is physically tagged, the data can just be left in there.
-         */
-        if (boot_cpu_data.dcache.n_aliases == 0)
-                return;
-
-        /*
-         * Don't bother groveling around the dcache for the VMA ranges
-         * if there are too many PTEs to make it worthwhile.
-         */
-        if (mm->nr_ptes >= MAX_DCACHE_PAGES)
-                flush_dcache_all();
-        else {
-                struct vm_area_struct *vma;
-
-                /*
-                 * In this case there are reasonably sized ranges to flush,
-                 * iterate through the VMA list and take care of any aliases.
-                 */
-                for (vma = mm->mmap; vma; vma = vma->vm_next)
-                        __flush_cache_mm(mm, vma->vm_start, vma->vm_end);
-        }
-
-        /* Only touch the icache if one of the VMAs has VM_EXEC set. */
-        if (mm->exec_vm)
-                flush_icache_all();
+        flush_dcache_all();
 }
 
 /*
@@ -303,44 +211,63 @@ static void sh4_flush_cache_page(void *args)
 {
         struct flusher_data *data = args;
         struct vm_area_struct *vma;
+        struct page *page;
         unsigned long address, pfn, phys;
-        unsigned int alias_mask;
+        int map_coherent = 0;
+        pgd_t *pgd;
+        pud_t *pud;
+        pmd_t *pmd;
+        pte_t *pte;
+        void *vaddr;
 
         vma = data->vma;
-        address = data->addr1;
+        address = data->addr1 & PAGE_MASK;
         pfn = data->addr2;
         phys = pfn << PAGE_SHIFT;
+        page = pfn_to_page(pfn);
 
         if (cpu_context(smp_processor_id(), vma->vm_mm) == NO_CONTEXT)
                 return;
 
-        alias_mask = boot_cpu_data.dcache.alias_mask;
-
-        /* We only need to flush D-cache when we have alias */
-        if ((address^phys) & alias_mask) {
-                /* Loop 4K of the D-cache */
-                flush_cache_one(
-                        CACHE_OC_ADDRESS_ARRAY | (address & alias_mask),
-                        phys);
-                /* Loop another 4K of the D-cache */
-                flush_cache_one(
-                        CACHE_OC_ADDRESS_ARRAY | (phys & alias_mask),
-                        phys);
-        }
+        pgd = pgd_offset(vma->vm_mm, address);
+        pud = pud_offset(pgd, address);
+        pmd = pmd_offset(pud, address);
+        pte = pte_offset_kernel(pmd, address);
+
+        /* If the page isn't present, there is nothing to do here. */
+        if (!(pte_val(*pte) & _PAGE_PRESENT))
+                return;
 
-        alias_mask = boot_cpu_data.icache.alias_mask;
-        if (vma->vm_flags & VM_EXEC) {
+        if ((vma->vm_mm == current->active_mm))
+                vaddr = NULL;
+        else {
                 /*
-                 * Evict entries from the portion of the cache from which code
-                 * may have been executed at this address (virtual).  There's
-                 * no need to evict from the portion corresponding to the
-                 * physical address as for the D-cache, because we know the
-                 * kernel has never executed the code through its identity
-                 * translation.
+                 * Use kmap_coherent or kmap_atomic to do flushes for
+                 * another ASID than the current one.
                  */
-                flush_cache_one(
-                        CACHE_IC_ADDRESS_ARRAY | (address & alias_mask),
-                        phys);
+                map_coherent = (current_cpu_data.dcache.n_aliases &&
+                        !test_bit(PG_dcache_dirty, &page->flags) &&
+                        page_mapped(page));
+                if (map_coherent)
+                        vaddr = kmap_coherent(page, address);
+                else
+                        vaddr = kmap_atomic(page, KM_USER0);
+
+                address = (unsigned long)vaddr;
+        }
+
+        if (pages_do_alias(address, phys))
+                flush_cache_one(CACHE_OC_ADDRESS_ARRAY |
+                        (address & shm_align_mask), phys);
+
+        if (vma->vm_flags & VM_EXEC)
+                flush_icache_all();
+
+        if (vaddr) {
+                if (map_coherent)
+                        kunmap_coherent(vaddr);
+                else
+                        kunmap_atomic(vaddr, KM_USER0);
         }
 }
 
@@ -373,24 +300,10 @@ static void sh4_flush_cache_range(void *args)
         if (boot_cpu_data.dcache.n_aliases == 0)
                 return;
 
-        /*
-         * Don't bother with the lookup and alias check if we have a
-         * wide range to cover, just blow away the dcache in its
-         * entirety instead. -- PFM.
-         */
-        if (((end - start) >> PAGE_SHIFT) >= MAX_DCACHE_PAGES)
-                flush_dcache_all();
-        else
-                __flush_cache_mm(vma->vm_mm, start, end);
+        flush_dcache_all();
 
-        if (vma->vm_flags & VM_EXEC) {
-                /*
-                 * TODO: Is this required???  Need to look at how I-cache
-                 * coherency is assured when new programs are loaded to see if
-                 * this matters.
-                 */
+        if (vma->vm_flags & VM_EXEC)
                 flush_icache_all();
-        }
 }
 
 /**
@@ -464,245 +377,6 @@ static void __flush_cache_one(unsigned long addr, unsigned long phys,
         } while (--way_count != 0);
 }
 
-/*
- * Break the 1, 2 and 4 way variants of this out into separate functions to
- * avoid nearly all the overhead of having the conditional stuff in the function
- * bodies (+ the 1 and 2 way cases avoid saving any registers too).
- *
- * We want to eliminate unnecessary bus transactions, so this code uses
- * a non-obvious technique.
- *
- * Loop over a cache way sized block of, one cache line at a time. For each
- * line, use movca.a to cause the current cache line contents to be written
- * back, but without reading anything from main memory. However this has the
- * side effect that the cache is now caching that memory location. So follow
- * this with a cache invalidate to mark the cache line invalid. And do all
- * this with interrupts disabled, to avoid the cache line being accidently
- * evicted while it is holding garbage.
- *
- * This also breaks in a number of circumstances:
- * - if there are modifications to the region of memory just above
- *   empty_zero_page (for example because a breakpoint has been placed
- *   there), then these can be lost.
- *
- *   This is because the the memory address which the cache temporarily
- *   caches in the above description is empty_zero_page. So the
- *   movca.l hits the cache (it is assumed that it misses, or at least
- *   isn't dirty), modifies the line and then invalidates it, losing the
- *   required change.
- *
- * - If caches are disabled or configured in write-through mode, then
- *   the movca.l writes garbage directly into memory.
- */
-static void __flush_dcache_segment_writethrough(unsigned long start,
-                                                unsigned long extent_per_way)
-{
-        unsigned long addr;
-        int i;
-
-        addr = CACHE_OC_ADDRESS_ARRAY | (start & cpu_data->dcache.entry_mask);
-
-        while (extent_per_way) {
-                for (i = 0; i < cpu_data->dcache.ways; i++)
-                        __raw_writel(0, addr + cpu_data->dcache.way_incr * i);
-
-                addr += cpu_data->dcache.linesz;
-                extent_per_way -= cpu_data->dcache.linesz;
-        }
-}
-
-static void __flush_dcache_segment_1way(unsigned long start,
-                                        unsigned long extent_per_way)
-{
-        unsigned long orig_sr, sr_with_bl;
-        unsigned long base_addr;
-        unsigned long way_incr, linesz, way_size;
-        struct cache_info *dcache;
-        register unsigned long a0, a0e;
-
-        asm volatile("stc sr, %0" : "=r" (orig_sr));
-        sr_with_bl = orig_sr | (1<<28);
-        base_addr = ((unsigned long)&empty_zero_page[0]);
-
-        /*
-         * The previous code aligned base_addr to 16k, i.e. the way_size of all
-         * existing SH-4 D-caches.  Whilst I don't see a need to have this
-         * aligned to any better than the cache line size (which it will be
-         * anyway by construction), let's align it to at least the way_size of
-         * any existing or conceivable SH-4 D-cache.  -- RPC
-         */
-        base_addr = ((base_addr >> 16) << 16);
-        base_addr |= start;
-
-        dcache = &boot_cpu_data.dcache;
-        linesz = dcache->linesz;
-        way_incr = dcache->way_incr;
-        way_size = dcache->way_size;
-
-        a0 = base_addr;
-        a0e = base_addr + extent_per_way;
-        do {
-                asm volatile("ldc %0, sr" : : "r" (sr_with_bl));
-                asm volatile("movca.l r0, @%0\n\t"
-                             "ocbi @%0" : : "r" (a0));
-                a0 += linesz;
-                asm volatile("movca.l r0, @%0\n\t"
-                             "ocbi @%0" : : "r" (a0));
-                a0 += linesz;
-                asm volatile("movca.l r0, @%0\n\t"
-                             "ocbi @%0" : : "r" (a0));
-                a0 += linesz;
-                asm volatile("movca.l r0, @%0\n\t"
-                             "ocbi @%0" : : "r" (a0));
-                asm volatile("ldc %0, sr" : : "r" (orig_sr));
-                a0 += linesz;
-        } while (a0 < a0e);
-}
-
-static void __flush_dcache_segment_2way(unsigned long start,
-                                        unsigned long extent_per_way)
-{
-        unsigned long orig_sr, sr_with_bl;
-        unsigned long base_addr;
-        unsigned long way_incr, linesz, way_size;
-        struct cache_info *dcache;
-        register unsigned long a0, a1, a0e;
-
-        asm volatile("stc sr, %0" : "=r" (orig_sr));
-        sr_with_bl = orig_sr | (1<<28);
-        base_addr = ((unsigned long)&empty_zero_page[0]);
-
-        /* See comment under 1-way above */
-        base_addr = ((base_addr >> 16) << 16);
-        base_addr |= start;
-
-        dcache = &boot_cpu_data.dcache;
-        linesz = dcache->linesz;
-        way_incr = dcache->way_incr;
-        way_size = dcache->way_size;
-
-        a0 = base_addr;
-        a1 = a0 + way_incr;
-        a0e = base_addr + extent_per_way;
-        do {
-                asm volatile("ldc %0, sr" : : "r" (sr_with_bl));
-                asm volatile("movca.l r0, @%0\n\t"
-                             "movca.l r0, @%1\n\t"
-                             "ocbi @%0\n\t"
-                             "ocbi @%1" : :
-                             "r" (a0), "r" (a1));
-                a0 += linesz;
-                a1 += linesz;
-                asm volatile("movca.l r0, @%0\n\t"
-                             "movca.l r0, @%1\n\t"
-                             "ocbi @%0\n\t"
-                             "ocbi @%1" : :
-                             "r" (a0), "r" (a1));
-                a0 += linesz;
-                a1 += linesz;
-                asm volatile("movca.l r0, @%0\n\t"
-                             "movca.l r0, @%1\n\t"
-                             "ocbi @%0\n\t"
-                             "ocbi @%1" : :
-                             "r" (a0), "r" (a1));
-                a0 += linesz;
-                a1 += linesz;
-                asm volatile("movca.l r0, @%0\n\t"
-                             "movca.l r0, @%1\n\t"
-                             "ocbi @%0\n\t"
-                             "ocbi @%1" : :
-                             "r" (a0), "r" (a1));
-                asm volatile("ldc %0, sr" : : "r" (orig_sr));
-                a0 += linesz;
-                a1 += linesz;
-        } while (a0 < a0e);
-}
-
-static void __flush_dcache_segment_4way(unsigned long start,
-                                        unsigned long extent_per_way)
-{
-        unsigned long orig_sr, sr_with_bl;
-        unsigned long base_addr;
-        unsigned long way_incr, linesz, way_size;
-        struct cache_info *dcache;
-        register unsigned long a0, a1, a2, a3, a0e;
-
-        asm volatile("stc sr, %0" : "=r" (orig_sr));
-        sr_with_bl = orig_sr | (1<<28);
-        base_addr = ((unsigned long)&empty_zero_page[0]);
-
-        /* See comment under 1-way above */
-        base_addr = ((base_addr >> 16) << 16);
-        base_addr |= start;
-
-        dcache = &boot_cpu_data.dcache;
-        linesz = dcache->linesz;
-        way_incr = dcache->way_incr;
-        way_size = dcache->way_size;
-
-        a0 = base_addr;
-        a1 = a0 + way_incr;
-        a2 = a1 + way_incr;
-        a3 = a2 + way_incr;
-        a0e = base_addr + extent_per_way;
-        do {
-                asm volatile("ldc %0, sr" : : "r" (sr_with_bl));
-                asm volatile("movca.l r0, @%0\n\t"
-                             "movca.l r0, @%1\n\t"
-                             "movca.l r0, @%2\n\t"
-                             "movca.l r0, @%3\n\t"
-                             "ocbi @%0\n\t"
-                             "ocbi @%1\n\t"
-                             "ocbi @%2\n\t"
-                             "ocbi @%3\n\t" : :
-                             "r" (a0), "r" (a1), "r" (a2), "r" (a3));
-                a0 += linesz;
-                a1 += linesz;
-                a2 += linesz;
-                a3 += linesz;
-                asm volatile("movca.l r0, @%0\n\t"
-                             "movca.l r0, @%1\n\t"
-                             "movca.l r0, @%2\n\t"
-                             "movca.l r0, @%3\n\t"
-                             "ocbi @%0\n\t"
-                             "ocbi @%1\n\t"
-                             "ocbi @%2\n\t"
-                             "ocbi @%3\n\t" : :
-                             "r" (a0), "r" (a1), "r" (a2), "r" (a3));
-                a0 += linesz;
-                a1 += linesz;
-                a2 += linesz;
-                a3 += linesz;
-                asm volatile("movca.l r0, @%0\n\t"
-                             "movca.l r0, @%1\n\t"
-                             "movca.l r0, @%2\n\t"
-                             "movca.l r0, @%3\n\t"
-                             "ocbi @%0\n\t"
-                             "ocbi @%1\n\t"
-                             "ocbi @%2\n\t"
-                             "ocbi @%3\n\t" : :
-                             "r" (a0), "r" (a1), "r" (a2), "r" (a3));
-                a0 += linesz;
-                a1 += linesz;
-                a2 += linesz;
-                a3 += linesz;
-                asm volatile("movca.l r0, @%0\n\t"
-                             "movca.l r0, @%1\n\t"
-                             "movca.l r0, @%2\n\t"
-                             "movca.l r0, @%3\n\t"
-                             "ocbi @%0\n\t"
-                             "ocbi @%1\n\t"
-                             "ocbi @%2\n\t"
-                             "ocbi @%3\n\t" : :
-                             "r" (a0), "r" (a1), "r" (a2), "r" (a3));
-                asm volatile("ldc %0, sr" : : "r" (orig_sr));
-                a0 += linesz;
-                a1 += linesz;
-                a2 += linesz;
-                a3 += linesz;
-        } while (a0 < a0e);
-}
-
 extern void __weak sh4__flush_region_init(void);
 
 /*
@@ -710,32 +384,11 @@ extern void __weak sh4__flush_region_init(void);
  */
 void __init sh4_cache_init(void)
 {
-        unsigned int wt_enabled = !!(__raw_readl(CCR) & CCR_CACHE_WT);
-
         printk("PVR=%08x CVR=%08x PRR=%08x\n",
                 ctrl_inl(CCN_PVR),
                 ctrl_inl(CCN_CVR),
                 ctrl_inl(CCN_PRR));
 
-        if (wt_enabled)
-                __flush_dcache_segment_fn = __flush_dcache_segment_writethrough;
-        else {
-                switch (boot_cpu_data.dcache.ways) {
-                case 1:
-                        __flush_dcache_segment_fn = __flush_dcache_segment_1way;
-                        break;
-                case 2:
-                        __flush_dcache_segment_fn = __flush_dcache_segment_2way;
-                        break;
-                case 4:
-                        __flush_dcache_segment_fn = __flush_dcache_segment_4way;
-                        break;
-                default:
-                        panic("unknown number of cache ways\n");
-                        break;
-                }
-        }
-
         local_flush_icache_range        = sh4_flush_icache_range;
         local_flush_dcache_page         = sh4_flush_dcache_page;
         local_flush_cache_all           = sh4_flush_cache_all;
diff --git a/arch/sh/mm/cache-sh5.c b/arch/sh/mm/cache-sh5.c
index 467ff8e260f7..eb4cc4ec7952 100644
--- a/arch/sh/mm/cache-sh5.c
+++ b/arch/sh/mm/cache-sh5.c
@@ -563,7 +563,7 @@ static void sh5_flush_cache_page(void *args)
 
 static void sh5_flush_dcache_page(void *page)
 {
-        sh64_dcache_purge_phy_page(page_to_phys(page));
+        sh64_dcache_purge_phy_page(page_to_phys((struct page *)page));
         wmb();
 }
 
diff --git a/arch/sh/mm/cache-sh7705.c b/arch/sh/mm/cache-sh7705.c
index 2601935eb589..f527fb70fce6 100644
--- a/arch/sh/mm/cache-sh7705.c
+++ b/arch/sh/mm/cache-sh7705.c
@@ -141,7 +141,7 @@ static void sh7705_flush_dcache_page(void *arg)
         if (mapping && !mapping_mapped(mapping))
                 set_bit(PG_dcache_dirty, &page->flags);
         else
-                __flush_dcache_page(PHYSADDR(page_address(page)));
+                __flush_dcache_page(__pa(page_address(page)));
 }
 
 static void __uses_jump_to_uncached sh7705_flush_cache_all(void *args)
diff --git a/arch/sh/mm/cache.c b/arch/sh/mm/cache.c
index a2dc7f9ecc51..e9415d3ea94a 100644
--- a/arch/sh/mm/cache.c
+++ b/arch/sh/mm/cache.c
@@ -27,8 +27,11 @@ void (*local_flush_icache_page)(void *args) = cache_noop;
 void (*local_flush_cache_sigtramp)(void *args) = cache_noop;
 
 void (*__flush_wback_region)(void *start, int size);
+EXPORT_SYMBOL(__flush_wback_region);
 void (*__flush_purge_region)(void *start, int size);
+EXPORT_SYMBOL(__flush_purge_region);
 void (*__flush_invalidate_region)(void *start, int size);
+EXPORT_SYMBOL(__flush_invalidate_region);
 
 static inline void noop__flush_region(void *start, int size)
 {
@@ -161,14 +164,21 @@ void flush_cache_all(void)
 {
         cacheop_on_each_cpu(local_flush_cache_all, NULL, 1);
 }
+EXPORT_SYMBOL(flush_cache_all);
 
 void flush_cache_mm(struct mm_struct *mm)
 {
+        if (boot_cpu_data.dcache.n_aliases == 0)
+                return;
+
         cacheop_on_each_cpu(local_flush_cache_mm, mm, 1);
 }
 
 void flush_cache_dup_mm(struct mm_struct *mm)
 {
+        if (boot_cpu_data.dcache.n_aliases == 0)
+                return;
+
         cacheop_on_each_cpu(local_flush_cache_dup_mm, mm, 1);
 }
 
@@ -195,11 +205,13 @@ void flush_cache_range(struct vm_area_struct *vma, unsigned long start,
 
         cacheop_on_each_cpu(local_flush_cache_range, (void *)&data, 1);
 }
+EXPORT_SYMBOL(flush_cache_range);
 
 void flush_dcache_page(struct page *page)
 {
         cacheop_on_each_cpu(local_flush_dcache_page, page, 1);
 }
+EXPORT_SYMBOL(flush_dcache_page);
 
 void flush_icache_range(unsigned long start, unsigned long end)
 {
@@ -265,7 +277,11 @@ static void __init emit_cache_params(void)
 
 void __init cpu_cache_init(void)
 {
-        unsigned int cache_disabled = !(__raw_readl(CCR) & CCR_CACHE_ENABLE);
+        unsigned int cache_disabled = 0;
+
+#ifdef CCR
+        cache_disabled = !(__raw_readl(CCR) & CCR_CACHE_ENABLE);
+#endif
 
         compute_alias(&boot_cpu_data.icache);
         compute_alias(&boot_cpu_data.dcache);
diff --git a/arch/sh/mm/consistent.c b/arch/sh/mm/consistent.c
index e098ec158ddb..902967e3f841 100644
--- a/arch/sh/mm/consistent.c
+++ b/arch/sh/mm/consistent.c
@@ -15,11 +15,15 @@
 #include <linux/dma-mapping.h>
 #include <linux/dma-debug.h>
 #include <linux/io.h>
+#include <linux/module.h>
 #include <asm/cacheflush.h>
 #include <asm/addrspace.h>
 
 #define PREALLOC_DMA_DEBUG_ENTRIES      4096
 
+struct dma_map_ops *dma_ops;
+EXPORT_SYMBOL(dma_ops);
+
 static int __init dma_init(void)
 {
         dma_debug_init(PREALLOC_DMA_DEBUG_ENTRIES);
@@ -27,15 +31,12 @@ static int __init dma_init(void)
 }
 fs_initcall(dma_init);
 
-void *dma_alloc_coherent(struct device *dev, size_t size,
-                           dma_addr_t *dma_handle, gfp_t gfp)
+void *dma_generic_alloc_coherent(struct device *dev, size_t size,
+                                 dma_addr_t *dma_handle, gfp_t gfp)
 {
         void *ret, *ret_nocache;
         int order = get_order(size);
 
-        if (dma_alloc_from_coherent(dev, size, dma_handle, &ret))
-                return ret;
-
         ret = (void *)__get_free_pages(gfp, order);
         if (!ret)
                 return NULL;
@@ -57,35 +58,26 @@ void *dma_alloc_coherent(struct device *dev, size_t size,
 
         *dma_handle = virt_to_phys(ret);
 
-        debug_dma_alloc_coherent(dev, size, *dma_handle, ret_nocache);
-
         return ret_nocache;
 }
-EXPORT_SYMBOL(dma_alloc_coherent);
 
-void dma_free_coherent(struct device *dev, size_t size,
-                         void *vaddr, dma_addr_t dma_handle)
+void dma_generic_free_coherent(struct device *dev, size_t size,
+                               void *vaddr, dma_addr_t dma_handle)
 {
         int order = get_order(size);
         unsigned long pfn = dma_handle >> PAGE_SHIFT;
         int k;
 
-        WARN_ON(irqs_disabled());       /* for portability */
-
-        if (dma_release_from_coherent(dev, order, vaddr))
-                return;
-
-        debug_dma_free_coherent(dev, size, vaddr, dma_handle);
         for (k = 0; k < (1 << order); k++)
                 __free_pages(pfn_to_page(pfn + k), 0);
+
         iounmap(vaddr);
 }
-EXPORT_SYMBOL(dma_free_coherent);
 
 void dma_cache_sync(struct device *dev, void *vaddr, size_t size,
                     enum dma_data_direction direction)
 {
-#ifdef CONFIG_CPU_SH5
+#if defined(CONFIG_CPU_SH5) || defined(CONFIG_PMB)
         void *p1addr = vaddr;
 #else
         void *p1addr = (void*) P1SEGADDR((unsigned long)vaddr);
diff --git a/arch/sh/mm/init.c b/arch/sh/mm/init.c
index 8173e38afd38..432acd07e76a 100644
--- a/arch/sh/mm/init.c
+++ b/arch/sh/mm/init.c
@@ -15,6 +15,7 @@
 #include <linux/pagemap.h>
 #include <linux/percpu.h>
 #include <linux/io.h>
+#include <linux/dma-mapping.h>
 #include <asm/mmu_context.h>
 #include <asm/tlb.h>
 #include <asm/cacheflush.h>
@@ -186,11 +187,21 @@ void __init paging_init(void)
         set_fixmap_nocache(FIX_UNCACHED, __pa(&__uncached_start));
 }
 
+/*
+ * Early initialization for any I/O MMUs we might have.
+ */
+static void __init iommu_init(void)
+{
+        no_iommu_init();
+}
+
 void __init mem_init(void)
 {
         int codesize, datasize, initsize;
         int nid;
 
+        iommu_init();
+
         num_physpages = 0;
         high_memory = NULL;
 
@@ -323,4 +334,12 @@ int memory_add_physaddr_to_nid(u64 addr)
 }
 EXPORT_SYMBOL_GPL(memory_add_physaddr_to_nid);
 #endif
+
 #endif /* CONFIG_MEMORY_HOTPLUG */
+
+#ifdef CONFIG_PMB
+int __in_29bit_mode(void)
+{
+        return !(ctrl_inl(PMB_PASCR) & PASCR_SE);
+}
+#endif /* CONFIG_PMB */
diff --git a/arch/sh/mm/kmap.c b/arch/sh/mm/kmap.c
index 16e01b5fed04..15d74ea42094 100644
--- a/arch/sh/mm/kmap.c
+++ b/arch/sh/mm/kmap.c
@@ -39,7 +39,9 @@ void *kmap_coherent(struct page *page, unsigned long addr)
         pagefault_disable();
 
         idx = FIX_CMAP_END -
-                ((addr & current_cpu_data.dcache.alias_mask) >> PAGE_SHIFT);
+                (((addr >> PAGE_SHIFT) & (FIX_N_COLOURS - 1)) +
+                 (FIX_N_COLOURS * smp_processor_id()));
+
         vaddr = __fix_to_virt(idx);
 
         BUG_ON(!pte_none(*(kmap_coherent_pte - idx)));
diff --git a/arch/sh/mm/numa.c b/arch/sh/mm/numa.c
index 9b784fdb947c..6c524446c0f6 100644
--- a/arch/sh/mm/numa.c
+++ b/arch/sh/mm/numa.c
@@ -60,7 +60,7 @@ void __init setup_bootmem_node(int nid, unsigned long start, unsigned long end)
         unsigned long bootmem_paddr;
 
         /* Don't allow bogus node assignment */
-        BUG_ON(nid > MAX_NUMNODES || nid == 0);
+        BUG_ON(nid > MAX_NUMNODES || nid <= 0);
 
         start_pfn = start >> PAGE_SHIFT;
         end_pfn = end >> PAGE_SHIFT;
diff --git a/arch/sh/mm/pmb-fixed.c b/arch/sh/mm/pmb-fixed.c
deleted file mode 100644
index 43c8eac4d8a1..000000000000
--- a/arch/sh/mm/pmb-fixed.c
+++ /dev/null
@@ -1,45 +0,0 @@
-/*
- * arch/sh/mm/fixed_pmb.c
- *
- * Copyright (C) 2009  Renesas Solutions Corp.
- *
- * 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/init.h>
-#include <linux/mm.h>
-#include <linux/io.h>
-#include <asm/mmu.h>
-#include <asm/mmu_context.h>
-
-static int __uses_jump_to_uncached fixed_pmb_init(void)
-{
-        int i;
-        unsigned long addr, data;
-
-        jump_to_uncached();
-
-        for (i = 0; i < PMB_ENTRY_MAX; i++) {
-                addr = PMB_DATA + (i << PMB_E_SHIFT);
-                data = ctrl_inl(addr);
-                if (!(data & PMB_V))
-                        continue;
-
-                if (data & PMB_C) {
-#if defined(CONFIG_CACHE_WRITETHROUGH)
-                        data |= PMB_WT;
-#elif defined(CONFIG_CACHE_WRITEBACK)
-                        data &= ~PMB_WT;
-#else
-                        data &= ~(PMB_C | PMB_WT);
-#endif
-                }
-                ctrl_outl(data, addr);
-        }
-
-        back_to_cached();
-
-        return 0;
-}
-arch_initcall(fixed_pmb_init);
diff --git a/arch/sh/mm/pmb.c b/arch/sh/mm/pmb.c
index aade31102112..280f6a166035 100644
--- a/arch/sh/mm/pmb.c
+++ b/arch/sh/mm/pmb.c
@@ -35,29 +35,9 @@
 
 static void __pmb_unmap(struct pmb_entry *);
 
-static struct kmem_cache *pmb_cache;
+static struct pmb_entry pmb_entry_list[NR_PMB_ENTRIES];
 static unsigned long pmb_map;
 
-static struct pmb_entry pmb_init_map[] = {
-        /* vpn         ppn         flags (ub/sz/c/wt) */
-
-        /* P1 Section Mappings */
-        { 0x80000000, 0x00000000, PMB_SZ_64M  | PMB_C, },
-        { 0x84000000, 0x04000000, PMB_SZ_64M  | PMB_C, },
-        { 0x88000000, 0x08000000, PMB_SZ_128M | PMB_C, },
-        { 0x90000000, 0x10000000, PMB_SZ_64M  | PMB_C, },
-        { 0x94000000, 0x14000000, PMB_SZ_64M  | PMB_C, },
-        { 0x98000000, 0x18000000, PMB_SZ_64M  | PMB_C, },
-
-        /* P2 Section Mappings */
-        { 0xa0000000, 0x00000000, PMB_UB | PMB_SZ_64M  | PMB_WT, },
-        { 0xa4000000, 0x04000000, PMB_UB | PMB_SZ_64M  | PMB_WT, },
-        { 0xa8000000, 0x08000000, PMB_UB | PMB_SZ_128M | PMB_WT, },
-        { 0xb0000000, 0x10000000, PMB_UB | PMB_SZ_64M  | PMB_WT, },
-        { 0xb4000000, 0x14000000, PMB_UB | PMB_SZ_64M  | PMB_WT, },
-        { 0xb8000000, 0x18000000, PMB_UB | PMB_SZ_64M  | PMB_WT, },
-};
-
 static inline unsigned long mk_pmb_entry(unsigned int entry)
 {
         return (entry & PMB_E_MASK) << PMB_E_SHIFT;
@@ -73,81 +53,68 @@ static inline unsigned long mk_pmb_data(unsigned int entry)
         return mk_pmb_entry(entry) | PMB_DATA;
 }
 
-static DEFINE_SPINLOCK(pmb_list_lock);
-static struct pmb_entry *pmb_list;
-
-static inline void pmb_list_add(struct pmb_entry *pmbe)
+static int pmb_alloc_entry(void)
 {
-        struct pmb_entry **p, *tmp;
+        unsigned int pos;
 
-        p = &pmb_list;
-        while ((tmp = *p) != NULL)
-                p = &tmp->next;
+repeat:
+        pos = find_first_zero_bit(&pmb_map, NR_PMB_ENTRIES);
 
-        pmbe->next = tmp;
-        *p = pmbe;
-}
+        if (unlikely(pos > NR_PMB_ENTRIES))
+                return -ENOSPC;
 
-static inline void pmb_list_del(struct pmb_entry *pmbe)
-{
-        struct pmb_entry **p, *tmp;
+        if (test_and_set_bit(pos, &pmb_map))
+                goto repeat;
 
-        for (p = &pmb_list; (tmp = *p); p = &tmp->next)
-                if (tmp == pmbe) {
-                        *p = tmp->next;
-                        return;
-                }
+        return pos;
 }
 
-struct pmb_entry *pmb_alloc(unsigned long vpn, unsigned long ppn,
-                            unsigned long flags)
+static struct pmb_entry *pmb_alloc(unsigned long vpn, unsigned long ppn,
+                                   unsigned long flags, int entry)
 {
         struct pmb_entry *pmbe;
+        int pos;
+
+        if (entry == PMB_NO_ENTRY) {
+                pos = pmb_alloc_entry();
+                if (pos < 0)
+                        return ERR_PTR(pos);
+        } else {
+                if (test_bit(entry, &pmb_map))
+                        return ERR_PTR(-ENOSPC);
+                pos = entry;
+        }
 
-        pmbe = kmem_cache_alloc(pmb_cache, GFP_KERNEL);
+        pmbe = &pmb_entry_list[pos];
         if (!pmbe)
                 return ERR_PTR(-ENOMEM);
 
         pmbe->vpn       = vpn;
         pmbe->ppn       = ppn;
         pmbe->flags     = flags;
-
-        spin_lock_irq(&pmb_list_lock);
-        pmb_list_add(pmbe);
-        spin_unlock_irq(&pmb_list_lock);
+        pmbe->entry     = pos;
 
         return pmbe;
 }
 
-void pmb_free(struct pmb_entry *pmbe)
+static void pmb_free(struct pmb_entry *pmbe)
 {
-        spin_lock_irq(&pmb_list_lock);
-        pmb_list_del(pmbe);
-        spin_unlock_irq(&pmb_list_lock);
+        int pos = pmbe->entry;
 
-        kmem_cache_free(pmb_cache, pmbe);
+        pmbe->vpn       = 0;
+        pmbe->ppn       = 0;
+        pmbe->flags     = 0;
+        pmbe->entry     = 0;
+
+        clear_bit(pos, &pmb_map);
 }
 
 /*
  * Must be in P2 for __set_pmb_entry()
  */
-int __set_pmb_entry(unsigned long vpn, unsigned long ppn,
-                    unsigned long flags, int *entry)
+static void __set_pmb_entry(unsigned long vpn, unsigned long ppn,
+                            unsigned long flags, int pos)
 {
-        unsigned int pos = *entry;
-
-        if (unlikely(pos == PMB_NO_ENTRY))
-                pos = find_first_zero_bit(&pmb_map, NR_PMB_ENTRIES);
-
-repeat:
-        if (unlikely(pos > NR_PMB_ENTRIES))
-                return -ENOSPC;
-
-        if (test_and_set_bit(pos, &pmb_map)) {
-                pos = find_first_zero_bit(&pmb_map, NR_PMB_ENTRIES);
-                goto repeat;
-        }
-
         ctrl_outl(vpn | PMB_V, mk_pmb_addr(pos));
 
 #ifdef CONFIG_CACHE_WRITETHROUGH
@@ -161,35 +128,21 @@ repeat:
 #endif
 
         ctrl_outl(ppn | flags | PMB_V, mk_pmb_data(pos));
-
-        *entry = pos;
-
-        return 0;
 }
 
-int __uses_jump_to_uncached set_pmb_entry(struct pmb_entry *pmbe)
+static void __uses_jump_to_uncached set_pmb_entry(struct pmb_entry *pmbe)
 {
-        int ret;
-
         jump_to_uncached();
-        ret = __set_pmb_entry(pmbe->vpn, pmbe->ppn, pmbe->flags, &pmbe->entry);
+        __set_pmb_entry(pmbe->vpn, pmbe->ppn, pmbe->flags, pmbe->entry);
         back_to_cached();
-
-        return ret;
 }
 
-void __uses_jump_to_uncached clear_pmb_entry(struct pmb_entry *pmbe)
+static void __uses_jump_to_uncached clear_pmb_entry(struct pmb_entry *pmbe)
 {
         unsigned int entry = pmbe->entry;
         unsigned long addr;
 
-        /*
-         * Don't allow clearing of wired init entries, P1 or P2 access
-         * without a corresponding mapping in the PMB will lead to reset
-         * by the TLB.
-         */
-        if (unlikely(entry < ARRAY_SIZE(pmb_init_map) ||
-                     entry >= NR_PMB_ENTRIES))
+        if (unlikely(entry >= NR_PMB_ENTRIES))
                 return;
 
         jump_to_uncached();
@@ -202,8 +155,6 @@ void __uses_jump_to_uncached clear_pmb_entry(struct pmb_entry *pmbe)
         ctrl_outl(ctrl_inl(addr) & ~PMB_V, addr);
 
         back_to_cached();
-
-        clear_bit(entry, &pmb_map);
 }
 
 
@@ -239,23 +190,17 @@ long pmb_remap(unsigned long vaddr, unsigned long phys,
 
 again:
         for (i = 0; i < ARRAY_SIZE(pmb_sizes); i++) {
-                int ret;
-
                 if (size < pmb_sizes[i].size)
                         continue;
 
-                pmbe = pmb_alloc(vaddr, phys, pmb_flags | pmb_sizes[i].flag);
+                pmbe = pmb_alloc(vaddr, phys, pmb_flags | pmb_sizes[i].flag,
+                                 PMB_NO_ENTRY);
                 if (IS_ERR(pmbe)) {
                         err = PTR_ERR(pmbe);
                         goto out;
                 }
 
-                ret = set_pmb_entry(pmbe);
-                if (ret != 0) {
-                        pmb_free(pmbe);
-                        err = -EBUSY;
-                        goto out;
-                }
+                set_pmb_entry(pmbe);
 
                 phys    += pmb_sizes[i].size;
                 vaddr   += pmb_sizes[i].size;
@@ -292,11 +237,16 @@ out:
 
 void pmb_unmap(unsigned long addr)
 {
-        struct pmb_entry **p, *pmbe;
+        struct pmb_entry *pmbe = NULL;
+        int i;
 
-        for (p = &pmb_list; (pmbe = *p); p = &pmbe->next)
-                if (pmbe->vpn == addr)
-                        break;
+        for (i = 0; i < ARRAY_SIZE(pmb_entry_list); i++) {
+                if (test_bit(i, &pmb_map)) {
+                        pmbe = &pmb_entry_list[i];
+                        if (pmbe->vpn == addr)
+                                break;
+                }
+        }
 
         if (unlikely(!pmbe))
                 return;
@@ -306,13 +256,22 @@ void pmb_unmap(unsigned long addr)
 
 static void __pmb_unmap(struct pmb_entry *pmbe)
 {
-        WARN_ON(!test_bit(pmbe->entry, &pmb_map));
+        BUG_ON(!test_bit(pmbe->entry, &pmb_map));
 
         do {
                 struct pmb_entry *pmblink = pmbe;
 
-                if (pmbe->entry != PMB_NO_ENTRY)
-                        clear_pmb_entry(pmbe);
+                /*
+                 * We may be called before this pmb_entry has been
+                 * entered into the PMB table via set_pmb_entry(), but
+                 * that's OK because we've allocated a unique slot for
+                 * this entry in pmb_alloc() (even if we haven't filled
+                 * it yet).
+                 *
+                 * Therefore, calling clear_pmb_entry() is safe as no
+                 * other mapping can be using that slot.
+                 */
+                clear_pmb_entry(pmbe);
 
                 pmbe = pmblink->link;
 
@@ -320,42 +279,34 @@ static void __pmb_unmap(struct pmb_entry *pmbe)
         } while (pmbe);
 }
 
-static void pmb_cache_ctor(void *pmb)
+#ifdef CONFIG_PMB
+int __uses_jump_to_uncached pmb_init(void)
 {
-        struct pmb_entry *pmbe = pmb;
-
-        memset(pmb, 0, sizeof(struct pmb_entry));
-
-        pmbe->entry = PMB_NO_ENTRY;
-}
-
-static int __uses_jump_to_uncached pmb_init(void)
-{
-        unsigned int nr_entries = ARRAY_SIZE(pmb_init_map);
-        unsigned int entry, i;
-
-        BUG_ON(unlikely(nr_entries >= NR_PMB_ENTRIES));
-
-        pmb_cache = kmem_cache_create("pmb", sizeof(struct pmb_entry), 0,
-                                      SLAB_PANIC, pmb_cache_ctor);
+        unsigned int i;
+        long size, ret;
 
         jump_to_uncached();
 
         /*
-         * Ordering is important, P2 must be mapped in the PMB before we
-         * can set PMB.SE, and P1 must be mapped before we jump back to
-         * P1 space.
+         * Insert PMB entries for the P1 and P2 areas so that, after
+         * we've switched the MMU to 32-bit mode, the semantics of P1
+         * and P2 are the same as in 29-bit mode, e.g.
+         *
+         *      P1 - provides a cached window onto physical memory
+         *      P2 - provides an uncached window onto physical memory
          */
-        for (entry = 0; entry < nr_entries; entry++) {
-                struct pmb_entry *pmbe = pmb_init_map + entry;
+        size = __MEMORY_START + __MEMORY_SIZE;
 
-                __set_pmb_entry(pmbe->vpn, pmbe->ppn, pmbe->flags, &entry);
-        }
+        ret = pmb_remap(P1SEG, 0x00000000, size, PMB_C);
+        BUG_ON(ret != size);
+
+        ret = pmb_remap(P2SEG, 0x00000000, size, PMB_WT | PMB_UB);
+        BUG_ON(ret != size);
 
         ctrl_outl(0, PMB_IRMCR);
 
         /* PMB.SE and UB[7] */
-        ctrl_outl((1 << 31) | (1 << 7), PMB_PASCR);
+        ctrl_outl(PASCR_SE | (1 << 7), PMB_PASCR);
 
         /* Flush out the TLB */
         i =  ctrl_inl(MMUCR);
@@ -366,7 +317,53 @@ static int __uses_jump_to_uncached pmb_init(void)
 
         return 0;
 }
-arch_initcall(pmb_init);
+#else
+int __uses_jump_to_uncached pmb_init(void)
+{
+        int i;
+        unsigned long addr, data;
+
+        jump_to_uncached();
+
+        for (i = 0; i < PMB_ENTRY_MAX; i++) {
+                struct pmb_entry *pmbe;
+                unsigned long vpn, ppn, flags;
+
+                addr = PMB_DATA + (i << PMB_E_SHIFT);
+                data = ctrl_inl(addr);
+                if (!(data & PMB_V))
+                        continue;
+
+                if (data & PMB_C) {
+#if defined(CONFIG_CACHE_WRITETHROUGH)
+                        data |= PMB_WT;
+#elif defined(CONFIG_CACHE_WRITEBACK)
+                        data &= ~PMB_WT;
+#else
+                        data &= ~(PMB_C | PMB_WT);
+#endif
+                }
+                ctrl_outl(data, addr);
+
+                ppn = data & PMB_PFN_MASK;
+
+                flags = data & (PMB_C | PMB_WT | PMB_UB);
+                flags |= data & PMB_SZ_MASK;
+
+                addr = PMB_ADDR + (i << PMB_E_SHIFT);
+                data = ctrl_inl(addr);
+
+                vpn = data & PMB_PFN_MASK;
+
+                pmbe = pmb_alloc(vpn, ppn, flags, i);
+                WARN_ON(IS_ERR(pmbe));
+        }
+
+        back_to_cached();
+
+        return 0;
+}
+#endif /* CONFIG_PMB */
 
 static int pmb_seq_show(struct seq_file *file, void *iter)
 {
@@ -434,15 +431,18 @@ postcore_initcall(pmb_debugfs_init);
 static int pmb_sysdev_suspend(struct sys_device *dev, pm_message_t state)
 {
         static pm_message_t prev_state;
+        int i;
 
         /* Restore the PMB after a resume from hibernation */
         if (state.event == PM_EVENT_ON &&
             prev_state.event == PM_EVENT_FREEZE) {
                 struct pmb_entry *pmbe;
-                spin_lock_irq(&pmb_list_lock);
-                for (pmbe = pmb_list; pmbe; pmbe = pmbe->next)
-                        set_pmb_entry(pmbe);
-                spin_unlock_irq(&pmb_list_lock);
+                for (i = 0; i < ARRAY_SIZE(pmb_entry_list); i++) {
+                        if (test_bit(i, &pmb_map)) {
+                                pmbe = &pmb_entry_list[i];
+                                set_pmb_entry(pmbe);
+                        }
+                }
         }
         prev_state = state;
         return 0;
diff --git a/arch/sh/oprofile/Makefile b/arch/sh/oprofile/Makefile
index 8e6eec91c14c..4886c5c1786c 100644
--- a/arch/sh/oprofile/Makefile
+++ b/arch/sh/oprofile/Makefile
@@ -7,7 +7,3 @@ DRIVER_OBJS = $(addprefix ../../../drivers/oprofile/, \
                 timer_int.o )
 
 oprofile-y      := $(DRIVER_OBJS) common.o backtrace.o
-
-oprofile-$(CONFIG_CPU_SUBTYPE_SH7750S)  += op_model_sh7750.o
-oprofile-$(CONFIG_CPU_SUBTYPE_SH7750)   += op_model_sh7750.o
-oprofile-$(CONFIG_CPU_SUBTYPE_SH7091)   += op_model_sh7750.o
diff --git a/arch/sh/oprofile/common.c b/arch/sh/oprofile/common.c
index 44f4e31c6d63..ac604937f3ee 100644
--- a/arch/sh/oprofile/common.c
+++ b/arch/sh/oprofile/common.c
@@ -20,9 +20,6 @@
 #include <asm/processor.h>
 #include "op_impl.h"
 
-extern struct op_sh_model op_model_sh7750_ops __weak;
-extern struct op_sh_model op_model_sh4a_ops __weak;
-
 static struct op_sh_model *model;
 
 static struct op_counter_config ctr[20];
@@ -94,33 +91,14 @@ int __init oprofile_arch_init(struct oprofile_operations *ops)
          */
         ops->backtrace = sh_backtrace;
 
-        switch (current_cpu_data.type) {
-        /* SH-4 types */
-        case CPU_SH7750:
-        case CPU_SH7750S:
-                lmodel = &op_model_sh7750_ops;
-                break;
-
-        /* SH-4A types */
-        case CPU_SH7763:
-        case CPU_SH7770:
-        case CPU_SH7780:
-        case CPU_SH7781:
-        case CPU_SH7785:
-        case CPU_SH7786:
-        case CPU_SH7723:
-        case CPU_SH7724:
-        case CPU_SHX3:
-                lmodel = &op_model_sh4a_ops;
-                break;
-
-        /* SH4AL-DSP types */
-        case CPU_SH7343:
-        case CPU_SH7722:
-        case CPU_SH7366:
-                lmodel = &op_model_sh4a_ops;
-                break;
-        }
+        /*
+         * XXX
+         *
+         * All of the SH7750/SH-4A counters have been converted to perf,
+         * this infrastructure hook is left for other users until they've
+         * had a chance to convert over, at which point all of this
+         * will be deleted.
+         */
 
         if (!lmodel)
                 return -ENODEV;
diff --git a/arch/sh/oprofile/op_impl.h b/arch/sh/oprofile/op_impl.h
index 4d509975eba6..1244479ceb29 100644
--- a/arch/sh/oprofile/op_impl.h
+++ b/arch/sh/oprofile/op_impl.h
@@ -6,7 +6,7 @@ struct op_counter_config {
         unsigned long enabled;
         unsigned long event;
 
-        unsigned long long count;
+        unsigned long count;
 
         /* Dummy values for userspace tool compliance */
         unsigned long kernel;
diff --git a/arch/sh/oprofile/op_model_sh7750.c b/arch/sh/oprofile/op_model_sh7750.c
deleted file mode 100644
index c892c7c30c2f..000000000000
--- a/arch/sh/oprofile/op_model_sh7750.c
+++ /dev/null
@@ -1,255 +0,0 @@
-/*
- * arch/sh/oprofile/op_model_sh7750.c
- *
- * OProfile support for SH7750/SH7750S Performance Counters
- *
- * Copyright (C) 2003 - 2008  Paul Mundt
- *
- * 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/oprofile.h>
-#include <linux/profile.h>
-#include <linux/init.h>
-#include <linux/errno.h>
-#include <linux/interrupt.h>
-#include <linux/io.h>
-#include <linux/fs.h>
-#include "op_impl.h"
-
-#define PM_CR_BASE      0xff000084      /* 16-bit */
-#define PM_CTR_BASE     0xff100004      /* 32-bit */
-
-#define PMCR(n)         (PM_CR_BASE + ((n) * 0x04))
-#define PMCTRH(n)       (PM_CTR_BASE + 0x00 + ((n) * 0x08))
-#define PMCTRL(n)       (PM_CTR_BASE + 0x04 + ((n) * 0x08))
-
-#define PMCR_PMM_MASK   0x0000003f
-
-#define PMCR_CLKF       0x00000100
-#define PMCR_PMCLR      0x00002000
-#define PMCR_PMST       0x00004000
-#define PMCR_PMEN       0x00008000
-
-struct op_sh_model op_model_sh7750_ops;
-
-#define NR_CNTRS        2
-
-static struct sh7750_ppc_register_config {
-        unsigned int ctrl;
-        unsigned long cnt_hi;
-        unsigned long cnt_lo;
-} regcache[NR_CNTRS];
-
-/*
- * There are a number of events supported by each counter (33 in total).
- * Since we have 2 counters, each counter will take the event code as it
- * corresponds to the PMCR PMM setting. Each counter can be configured
- * independently.
- *
- *      Event Code      Description
- *      ----------      -----------
- *
- *      0x01            Operand read access
- *      0x02            Operand write access
- *      0x03            UTLB miss
- *      0x04            Operand cache read miss
- *      0x05            Operand cache write miss
- *      0x06            Instruction fetch (w/ cache)
- *      0x07            Instruction TLB miss
- *      0x08            Instruction cache miss
- *      0x09            All operand accesses
- *      0x0a            All instruction accesses
- *      0x0b            OC RAM operand access
- *      0x0d            On-chip I/O space access
- *      0x0e            Operand access (r/w)
- *      0x0f            Operand cache miss (r/w)
- *      0x10            Branch instruction
- *      0x11            Branch taken
- *      0x12            BSR/BSRF/JSR
- *      0x13            Instruction execution
- *      0x14            Instruction execution in parallel
- *      0x15            FPU Instruction execution
- *      0x16            Interrupt
- *      0x17            NMI
- *      0x18            trapa instruction execution
- *      0x19            UBCA match
- *      0x1a            UBCB match
- *      0x21            Instruction cache fill
- *      0x22            Operand cache fill
- *      0x23            Elapsed time
- *      0x24            Pipeline freeze by I-cache miss
- *      0x25            Pipeline freeze by D-cache miss
- *      0x27            Pipeline freeze by branch instruction
- *      0x28            Pipeline freeze by CPU register
- *      0x29            Pipeline freeze by FPU
- *
- * Unfortunately we don't have a native exception or interrupt for counter
- * overflow (although since these counters can run for 16.3 days without
- * overflowing, it's not really necessary).
- *
- * OProfile on the other hand likes to have samples taken periodically, so
- * for now we just piggyback the timer interrupt to get the expected
- * behavior.
- */
-
-static int sh7750_timer_notify(struct pt_regs *regs)
-{
-        oprofile_add_sample(regs, 0);
-        return 0;
-}
-
-static u64 sh7750_read_counter(int counter)
-{
-        return (u64)((u64)(__raw_readl(PMCTRH(counter)) & 0xffff) << 32) |
-                           __raw_readl(PMCTRL(counter));
-}
-
-/*
- * Files will be in a path like:
- *
- *  /<oprofilefs mount point>/<counter number>/<file>
- *
- * So when dealing with <file>, we look to the parent dentry for the counter
- * number.
- */
-static inline int to_counter(struct file *file)
-{
-        const unsigned char *name = file->f_path.dentry->d_parent->d_name.name;
-
-        return (int)simple_strtol(name, NULL, 10);
-}
-
-/*
- * XXX: We have 48-bit counters, so we're probably going to want something
- * more along the lines of oprofilefs_ullong_to_user().. Truncating to
- * unsigned long works fine for now though, as long as we don't attempt to
- * profile for too horribly long.
- */
-static ssize_t sh7750_read_count(struct file *file, char __user *buf,
-                                 size_t count, loff_t *ppos)
-{
-        int counter = to_counter(file);
-        u64 val = sh7750_read_counter(counter);
-
-        return oprofilefs_ulong_to_user((unsigned long)val, buf, count, ppos);
-}
-
-static ssize_t sh7750_write_count(struct file *file, const char __user *buf,
-                                  size_t count, loff_t *ppos)
-{
-        int counter = to_counter(file);
-        unsigned long val;
-
-        if (oprofilefs_ulong_from_user(&val, buf, count))
-                return -EFAULT;
-
-        /*
-         * Any write will clear the counter, although only 0 should be
-         * written for this purpose, as we do not support setting the
-         * counter to an arbitrary value.
-         */
-        WARN_ON(val != 0);
-
-        __raw_writew(__raw_readw(PMCR(counter)) | PMCR_PMCLR, PMCR(counter));
-
-        return count;
-}
-
-static const struct file_operations count_fops = {
-        .read           = sh7750_read_count,
-        .write          = sh7750_write_count,
-};
-
-static int sh7750_ppc_create_files(struct super_block *sb, struct dentry *dir)
-{
-        return oprofilefs_create_file(sb, dir, "count", &count_fops);
-}
-
-static void sh7750_ppc_reg_setup(struct op_counter_config *ctr)
-{
-        unsigned int counters = op_model_sh7750_ops.num_counters;
-        int i;
-
-        for (i = 0; i < counters; i++) {
-                regcache[i].ctrl        = 0;
-                regcache[i].cnt_hi      = 0;
-                regcache[i].cnt_lo      = 0;
-
-                if (!ctr[i].enabled)
-                        continue;
-
-                regcache[i].ctrl |= ctr[i].event | PMCR_PMEN | PMCR_PMST;
-                regcache[i].cnt_hi = (unsigned long)((ctr->count >> 32) & 0xffff);
-                regcache[i].cnt_lo = (unsigned long)(ctr->count & 0xffffffff);
-        }
-}
-
-static void sh7750_ppc_cpu_setup(void *args)
-{
-        unsigned int counters = op_model_sh7750_ops.num_counters;
-        int i;
-
-        for (i = 0; i < counters; i++) {
-                __raw_writew(0, PMCR(i));
-                __raw_writel(regcache[i].cnt_hi, PMCTRH(i));
-                __raw_writel(regcache[i].cnt_lo, PMCTRL(i));
-        }
-}
-
-static void sh7750_ppc_cpu_start(void *args)
-{
-        unsigned int counters = op_model_sh7750_ops.num_counters;
-        int i;
-
-        for (i = 0; i < counters; i++)
-                __raw_writew(regcache[i].ctrl, PMCR(i));
-}
-
-static void sh7750_ppc_cpu_stop(void *args)
-{
-        unsigned int counters = op_model_sh7750_ops.num_counters;
-        int i;
-
-        /* Disable the counters */
-        for (i = 0; i < counters; i++)
-                __raw_writew(__raw_readw(PMCR(i)) & ~PMCR_PMEN, PMCR(i));
-}
-
-static inline void sh7750_ppc_reset(void)
-{
-        unsigned int counters = op_model_sh7750_ops.num_counters;
-        int i;
-
-        /* Clear the counters */
-        for (i = 0; i < counters; i++)
-                __raw_writew(__raw_readw(PMCR(i)) | PMCR_PMCLR, PMCR(i));
-}
-
-static int sh7750_ppc_init(void)
-{
-        sh7750_ppc_reset();
-
-        return register_timer_hook(sh7750_timer_notify);
-}
-
-static void sh7750_ppc_exit(void)
-{
-        unregister_timer_hook(sh7750_timer_notify);
-
-        sh7750_ppc_reset();
-}
-
-struct op_sh_model op_model_sh7750_ops = {
-        .cpu_type       = "sh/sh7750",
-        .num_counters   = NR_CNTRS,
-        .reg_setup      = sh7750_ppc_reg_setup,
-        .cpu_setup      = sh7750_ppc_cpu_setup,
-        .cpu_start      = sh7750_ppc_cpu_start,
-        .cpu_stop       = sh7750_ppc_cpu_stop,
-        .init           = sh7750_ppc_init,
-        .exit           = sh7750_ppc_exit,
-        .create_files   = sh7750_ppc_create_files,
-};