Merge branch 'lpc32xx/fixes' of git://git.antcom.de/linux-2.6 into fixes

* 'lpc32xx/fixes' of git://git.antcom.de/linux-2.6: (5 commits)
  ARM: LPC32xx: serial.c: Fixed loop limit
  ARM: LPC32xx: serial.c: HW bug workaround
  ARM: LPC32xx: irq.c: Clear latched event
  ARM: LPC32xx: Fix interrupt controller init
  ARM: LPC32xx: Fix irq on GPI_28

Update to Linux 3.3-rc5

Signed-off-by: Arnd Bergmann <arnd@arndb.de>

57 files changed, 405 insertions, 219 deletions

diff --git a/arch/arm/common/it8152.c b/arch/arm/common/it8152.c
index d1bcd7b13ebc..fb1f1cfce60c 100644
--- a/arch/arm/common/it8152.c
+++ b/arch/arm/common/it8152.c
@@ -320,13 +320,6 @@ err0:
 	return -EBUSY;
 }
 
-/*
- * If we set up a device for bus mastering, we need to check the latency
- * timer as we don't have even crappy BIOSes to set it properly.
- * The implementation is from arch/i386/pci/i386.c
- */
-unsigned int pcibios_max_latency = 255;
-
 /* ITE bridge requires setting latency timer to avoid early bus access
    termination by PCI bus master devices
 */
diff --git a/arch/arm/common/pl330.c b/arch/arm/common/pl330.c
index d8e44a43047c..ff3ad2244824 100644
--- a/arch/arm/common/pl330.c
+++ b/arch/arm/common/pl330.c
@@ -1502,12 +1502,13 @@ int pl330_chan_ctrl(void *ch_id, enum pl330_chan_op op)
 	struct pl330_thread *thrd = ch_id;
 	struct pl330_dmac *pl330;
 	unsigned long flags;
-	int ret = 0, active = thrd->req_running;
+	int ret = 0, active;
 
 	if (!thrd || thrd->free || thrd->dmac->state == DYING)
 		return -EINVAL;
 
 	pl330 = thrd->dmac;
+	active = thrd->req_running;
 
 	spin_lock_irqsave(&pl330->lock, flags);
 
diff --git a/arch/arm/include/asm/assembler.h b/arch/arm/include/asm/assembler.h
index 62f8095d46de..23371b17b23e 100644
--- a/arch/arm/include/asm/assembler.h
+++ b/arch/arm/include/asm/assembler.h
@@ -137,6 +137,11 @@
 	disable_irq
 	.endm
 
+	.macro	save_and_disable_irqs_notrace, oldcpsr
+	mrs	\oldcpsr, cpsr
+	disable_irq_notrace
+	.endm
+
 /*
  * Restore interrupt state previously stored in a register.  We don't
  * guarantee that this will preserve the flags.
diff --git a/arch/arm/include/asm/hardware/pl330.h b/arch/arm/include/asm/hardware/pl330.h
index 575fa8186ca0..c1821385abfa 100644
--- a/arch/arm/include/asm/hardware/pl330.h
+++ b/arch/arm/include/asm/hardware/pl330.h
@@ -41,7 +41,7 @@ enum pl330_dstcachectrl {
 	DCCTRL1, /* Bufferable only */
 	DCCTRL2, /* Cacheable, but do not allocate */
 	DCCTRL3, /* Cacheable and bufferable, but do not allocate */
-	DINVALID1 = 8,
+	DINVALID1,              /* AWCACHE = 0x1000 */
 	DINVALID2,
 	DCCTRL6, /* Cacheable write-through, allocate on writes only */
 	DCCTRL7, /* Cacheable write-back, allocate on writes only */
diff --git a/arch/arm/include/asm/processor.h b/arch/arm/include/asm/processor.h
index ce280b8d613c..cb8d638924fd 100644
--- a/arch/arm/include/asm/processor.h
+++ b/arch/arm/include/asm/processor.h
@@ -22,6 +22,7 @@
 #include <asm/hw_breakpoint.h>
 #include <asm/ptrace.h>
 #include <asm/types.h>
+#include <asm/system.h>
 
 #ifdef __KERNEL__
 #define STACK_TOP	((current->personality & ADDR_LIMIT_32BIT) ? \
diff --git a/arch/arm/kernel/ptrace.c b/arch/arm/kernel/ptrace.c
index e33870ff0ac0..ede6443c34d9 100644
--- a/arch/arm/kernel/ptrace.c
+++ b/arch/arm/kernel/ptrace.c
@@ -23,6 +23,7 @@
 #include <linux/perf_event.h>
 #include <linux/hw_breakpoint.h>
 #include <linux/regset.h>
+#include <linux/audit.h>
 
 #include <asm/pgtable.h>
 #include <asm/system.h>
@@ -904,6 +905,12 @@ long arch_ptrace(struct task_struct *child, long request,
 	return ret;
 }
 
+#ifdef __ARMEB__
+#define AUDIT_ARCH_NR AUDIT_ARCH_ARMEB
+#else
+#define AUDIT_ARCH_NR AUDIT_ARCH_ARM
+#endif
+
 asmlinkage int syscall_trace(int why, struct pt_regs *regs, int scno)
 {
 	unsigned long ip;
@@ -918,7 +925,7 @@ asmlinkage int syscall_trace(int why, struct pt_regs *regs, int scno)
 	if (!ip)
 		audit_syscall_exit(regs);
 	else
-		audit_syscall_entry(AUDIT_ARCH_ARMEB, scno, regs->ARM_r0,
+		audit_syscall_entry(AUDIT_ARCH_NR, scno, regs->ARM_r0,
 				    regs->ARM_r1, regs->ARM_r2, regs->ARM_r3);
 
 	if (!test_thread_flag(TIF_SYSCALL_TRACE))
diff --git a/arch/arm/kernel/smp_twd.c b/arch/arm/kernel/smp_twd.c
index 4285daa077b0..7a79b24597b2 100644
--- a/arch/arm/kernel/smp_twd.c
+++ b/arch/arm/kernel/smp_twd.c
@@ -129,7 +129,7 @@ static struct notifier_block twd_cpufreq_nb = {
 
 static int twd_cpufreq_init(void)
 {
-	if (!IS_ERR(twd_clk))
+	if (twd_evt && *__this_cpu_ptr(twd_evt) && !IS_ERR(twd_clk))
 		return cpufreq_register_notifier(&twd_cpufreq_nb,
 			CPUFREQ_TRANSITION_NOTIFIER);
 
diff --git a/arch/arm/mach-lpc32xx/include/mach/irqs.h b/arch/arm/mach-lpc32xx/include/mach/irqs.h
index 2667f52e3b04..9e3b90df32e1 100644
--- a/arch/arm/mach-lpc32xx/include/mach/irqs.h
+++ b/arch/arm/mach-lpc32xx/include/mach/irqs.h
@@ -61,7 +61,7 @@
  */
 #define IRQ_LPC32XX_JTAG_COMM_TX	LPC32XX_SIC1_IRQ(1)
 #define IRQ_LPC32XX_JTAG_COMM_RX	LPC32XX_SIC1_IRQ(2)
-#define IRQ_LPC32XX_GPI_11		LPC32XX_SIC1_IRQ(4)
+#define IRQ_LPC32XX_GPI_28		LPC32XX_SIC1_IRQ(4)
 #define IRQ_LPC32XX_TS_P		LPC32XX_SIC1_IRQ(6)
 #define IRQ_LPC32XX_TS_IRQ		LPC32XX_SIC1_IRQ(7)
 #define IRQ_LPC32XX_TS_AUX		LPC32XX_SIC1_IRQ(8)
diff --git a/arch/arm/mach-lpc32xx/irq.c b/arch/arm/mach-lpc32xx/irq.c
index 4eae566dfdc7..c74de01ab5b6 100644
--- a/arch/arm/mach-lpc32xx/irq.c
+++ b/arch/arm/mach-lpc32xx/irq.c
@@ -118,6 +118,10 @@ static const struct lpc32xx_event_info lpc32xx_events[NR_IRQS] = {
 		.event_group = &lpc32xx_event_pin_regs,
 		.mask = LPC32XX_CLKPWR_EXTSRC_GPI_06_BIT,
 	},
+	[IRQ_LPC32XX_GPI_28] = {
+		.event_group = &lpc32xx_event_pin_regs,
+		.mask = LPC32XX_CLKPWR_EXTSRC_GPI_28_BIT,
+	},
 	[IRQ_LPC32XX_GPIO_00] = {
 		.event_group = &lpc32xx_event_int_regs,
 		.mask = LPC32XX_CLKPWR_INTSRC_GPIO_00_BIT,
@@ -305,9 +309,18 @@ static int lpc32xx_irq_wake(struct irq_data *d, unsigned int state)
 
 		if (state)
 			eventreg |= lpc32xx_events[d->irq].mask;
-		else
+		else {
 			eventreg &= ~lpc32xx_events[d->irq].mask;
 
+			/*
+			 * When disabling the wakeup, clear the latched
+			 * event
+			 */
+			__raw_writel(lpc32xx_events[d->irq].mask,
+				lpc32xx_events[d->irq].
+				event_group->rawstat_reg);
+		}
+
 		__raw_writel(eventreg,
 			lpc32xx_events[d->irq].event_group->enab_reg);
 
@@ -380,13 +393,15 @@ void __init lpc32xx_init_irq(void)
 
 	/* Setup SIC1 */
 	__raw_writel(0, LPC32XX_INTC_MASK(LPC32XX_SIC1_BASE));
-	__raw_writel(MIC_APR_DEFAULT, LPC32XX_INTC_POLAR(LPC32XX_SIC1_BASE));
-	__raw_writel(MIC_ATR_DEFAULT, LPC32XX_INTC_ACT_TYPE(LPC32XX_SIC1_BASE));
+	__raw_writel(SIC1_APR_DEFAULT, LPC32XX_INTC_POLAR(LPC32XX_SIC1_BASE));
+	__raw_writel(SIC1_ATR_DEFAULT,
+				LPC32XX_INTC_ACT_TYPE(LPC32XX_SIC1_BASE));
 
 	/* Setup SIC2 */
 	__raw_writel(0, LPC32XX_INTC_MASK(LPC32XX_SIC2_BASE));
-	__raw_writel(MIC_APR_DEFAULT, LPC32XX_INTC_POLAR(LPC32XX_SIC2_BASE));
-	__raw_writel(MIC_ATR_DEFAULT, LPC32XX_INTC_ACT_TYPE(LPC32XX_SIC2_BASE));
+	__raw_writel(SIC2_APR_DEFAULT, LPC32XX_INTC_POLAR(LPC32XX_SIC2_BASE));
+	__raw_writel(SIC2_ATR_DEFAULT,
+				LPC32XX_INTC_ACT_TYPE(LPC32XX_SIC2_BASE));
 
 	/* Configure supported IRQ's */
 	for (i = 0; i < NR_IRQS; i++) {
diff --git a/arch/arm/mach-lpc32xx/serial.c b/arch/arm/mach-lpc32xx/serial.c
index 429cfdbb2b3d..f2735281616a 100644
--- a/arch/arm/mach-lpc32xx/serial.c
+++ b/arch/arm/mach-lpc32xx/serial.c
@@ -88,6 +88,7 @@ struct uartinit {
 	char *uart_ck_name;
 	u32 ck_mode_mask;
 	void __iomem *pdiv_clk_reg;
+	resource_size_t mapbase;
 };
 
 static struct uartinit uartinit_data[] __initdata = {
@@ -97,6 +98,7 @@ static struct uartinit uartinit_data[] __initdata = {
 		.ck_mode_mask =
 			LPC32XX_UART_CLKMODE_LOAD(LPC32XX_UART_CLKMODE_ON, 5),
 		.pdiv_clk_reg = LPC32XX_CLKPWR_UART5_CLK_CTRL,
+		.mapbase = LPC32XX_UART5_BASE,
 	},
 #endif
 #ifdef CONFIG_ARCH_LPC32XX_UART3_SELECT
@@ -105,6 +107,7 @@ static struct uartinit uartinit_data[] __initdata = {
 		.ck_mode_mask =
 			LPC32XX_UART_CLKMODE_LOAD(LPC32XX_UART_CLKMODE_ON, 3),
 		.pdiv_clk_reg = LPC32XX_CLKPWR_UART3_CLK_CTRL,
+		.mapbase = LPC32XX_UART3_BASE,
 	},
 #endif
 #ifdef CONFIG_ARCH_LPC32XX_UART4_SELECT
@@ -113,6 +116,7 @@ static struct uartinit uartinit_data[] __initdata = {
 		.ck_mode_mask =
 			LPC32XX_UART_CLKMODE_LOAD(LPC32XX_UART_CLKMODE_ON, 4),
 		.pdiv_clk_reg = LPC32XX_CLKPWR_UART4_CLK_CTRL,
+		.mapbase = LPC32XX_UART4_BASE,
 	},
 #endif
 #ifdef CONFIG_ARCH_LPC32XX_UART6_SELECT
@@ -121,6 +125,7 @@ static struct uartinit uartinit_data[] __initdata = {
 		.ck_mode_mask =
 			LPC32XX_UART_CLKMODE_LOAD(LPC32XX_UART_CLKMODE_ON, 6),
 		.pdiv_clk_reg = LPC32XX_CLKPWR_UART6_CLK_CTRL,
+		.mapbase = LPC32XX_UART6_BASE,
 	},
 #endif
 };
@@ -165,11 +170,24 @@ void __init lpc32xx_serial_init(void)
 
 		/* pre-UART clock divider set to 1 */
 		__raw_writel(0x0101, uartinit_data[i].pdiv_clk_reg);
+
+		/*
+		 * Force a flush of the RX FIFOs to work around a
+		 * HW bug
+		 */
+		puart = uartinit_data[i].mapbase;
+		__raw_writel(0xC1, LPC32XX_UART_IIR_FCR(puart));
+		__raw_writel(0x00, LPC32XX_UART_DLL_FIFO(puart));
+		j = LPC32XX_SUART_FIFO_SIZE;
+		while (j--)
+			tmp = __raw_readl(
+				LPC32XX_UART_DLL_FIFO(puart));
+		__raw_writel(0, LPC32XX_UART_IIR_FCR(puart));
 	}
 
 	/* This needs to be done after all UART clocks are setup */
 	__raw_writel(clkmodes, LPC32XX_UARTCTL_CLKMODE);
-	for (i = 0; i < ARRAY_SIZE(uartinit_data) - 1; i++) {
+	for (i = 0; i < ARRAY_SIZE(uartinit_data); i++) {
 		/* Force a flush of the RX FIFOs to work around a HW bug */
 		puart = serial_std_platform_data[i].mapbase;
 		__raw_writel(0xC1, LPC32XX_UART_IIR_FCR(puart));
diff --git a/arch/arm/mach-omap2/voltagedomains3xxx_data.c b/arch/arm/mach-omap2/voltagedomains3xxx_data.c
index c005e2f5e383..57db2038b23c 100644
--- a/arch/arm/mach-omap2/voltagedomains3xxx_data.c
+++ b/arch/arm/mach-omap2/voltagedomains3xxx_data.c
@@ -108,6 +108,7 @@ void __init omap3xxx_voltagedomains_init(void)
 	 * XXX Will depend on the process, validation, and binning
 	 * for the currently-running IC
 	 */
+#ifdef CONFIG_PM_OPP
 	if (cpu_is_omap3630()) {
 		omap3_voltdm_mpu.volt_data = omap36xx_vddmpu_volt_data;
 		omap3_voltdm_core.volt_data = omap36xx_vddcore_volt_data;
@@ -115,6 +116,7 @@ void __init omap3xxx_voltagedomains_init(void)
 		omap3_voltdm_mpu.volt_data = omap34xx_vddmpu_volt_data;
 		omap3_voltdm_core.volt_data = omap34xx_vddcore_volt_data;
 	}
+#endif
 
 	if (cpu_is_omap3517() || cpu_is_omap3505())
 		voltdms = voltagedomains_am35xx;
diff --git a/arch/arm/mach-omap2/voltagedomains44xx_data.c b/arch/arm/mach-omap2/voltagedomains44xx_data.c
index 4e11d022595d..c3115f6853d4 100644
--- a/arch/arm/mach-omap2/voltagedomains44xx_data.c
+++ b/arch/arm/mach-omap2/voltagedomains44xx_data.c
@@ -100,9 +100,11 @@ void __init omap44xx_voltagedomains_init(void)
 	 * XXX Will depend on the process, validation, and binning
 	 * for the currently-running IC
 	 */
+#ifdef CONFIG_PM_OPP
 	omap4_voltdm_mpu.volt_data = omap44xx_vdd_mpu_volt_data;
 	omap4_voltdm_iva.volt_data = omap44xx_vdd_iva_volt_data;
 	omap4_voltdm_core.volt_data = omap44xx_vdd_core_volt_data;
+#endif
 
 	for (i = 0; voltdm = voltagedomains_omap4[i], voltdm; i++)
 		voltdm->sys_clk.name = sys_clk_name;
diff --git a/arch/arm/mach-shmobile/board-ag5evm.c b/arch/arm/mach-shmobile/board-ag5evm.c
index eff8a96c75ee..068b754bc348 100644
--- a/arch/arm/mach-shmobile/board-ag5evm.c
+++ b/arch/arm/mach-shmobile/board-ag5evm.c
@@ -30,6 +30,7 @@
 #include <linux/serial_sci.h>
 #include <linux/smsc911x.h>
 #include <linux/gpio.h>
+#include <linux/videodev2.h>
 #include <linux/input.h>
 #include <linux/input/sh_keysc.h>
 #include <linux/mmc/host.h>
@@ -37,7 +38,6 @@
 #include <linux/mmc/sh_mobile_sdhi.h>
 #include <linux/mfd/tmio.h>
 #include <linux/sh_clk.h>
-#include <linux/dma-mapping.h>
 #include <video/sh_mobile_lcdc.h>
 #include <video/sh_mipi_dsi.h>
 #include <sound/sh_fsi.h>
@@ -159,19 +159,12 @@ static struct resource sh_mmcif_resources[] = {
 	},
 };
 
-static struct sh_mmcif_dma sh_mmcif_dma = {
-	.chan_priv_rx	= {
-		.slave_id	= SHDMA_SLAVE_MMCIF_RX,
-	},
-	.chan_priv_tx	= {
-		.slave_id	= SHDMA_SLAVE_MMCIF_TX,
-	},
-};
 static struct sh_mmcif_plat_data sh_mmcif_platdata = {
 	.sup_pclk	= 0,
 	.ocr		= MMC_VDD_165_195,
 	.caps		= MMC_CAP_8_BIT_DATA | MMC_CAP_NONREMOVABLE,
-	.dma		= &sh_mmcif_dma,
+	.slave_id_tx	= SHDMA_SLAVE_MMCIF_TX,
+	.slave_id_rx	= SHDMA_SLAVE_MMCIF_RX,
 };
 
 static struct platform_device mmc_device = {
@@ -321,12 +314,11 @@ static struct resource mipidsi0_resources[] = {
 	},
 };
 
-#define DSI0PHYCR	0xe615006c
 static int sh_mipi_set_dot_clock(struct platform_device *pdev,
 				 void __iomem *base,
 				 int enable)
 {
-	struct clk *pck;
+	struct clk *pck, *phy;
 	int ret;
 
 	pck = clk_get(&pdev->dev, "dsip_clk");
@@ -335,18 +327,27 @@ static int sh_mipi_set_dot_clock(struct platform_device *pdev,
 		goto sh_mipi_set_dot_clock_pck_err;
 	}
 
+	phy = clk_get(&pdev->dev, "dsiphy_clk");
+	if (IS_ERR(phy)) {
+		ret = PTR_ERR(phy);
+		goto sh_mipi_set_dot_clock_phy_err;
+	}
+
 	if (enable) {
 		clk_set_rate(pck, clk_round_rate(pck,  24000000));
-		__raw_writel(0x2a809010, DSI0PHYCR);
+		clk_set_rate(phy, clk_round_rate(pck, 510000000));
 		clk_enable(pck);
+		clk_enable(phy);
 	} else {
 		clk_disable(pck);
+		clk_disable(phy);
 	}
 
 	ret = 0;
 
+	clk_put(phy);
+sh_mipi_set_dot_clock_phy_err:
 	clk_put(pck);
-
 sh_mipi_set_dot_clock_pck_err:
 	return ret;
 }
diff --git a/arch/arm/mach-shmobile/board-ap4evb.c b/arch/arm/mach-shmobile/board-ap4evb.c
index aab0a349f759..eeb4d9664584 100644
--- a/arch/arm/mach-shmobile/board-ap4evb.c
+++ b/arch/arm/mach-shmobile/board-ap4evb.c
@@ -295,15 +295,6 @@ static struct resource sh_mmcif_resources[] = {
 	},
 };
 
-static struct sh_mmcif_dma sh_mmcif_dma = {
-	.chan_priv_rx	= {
-		.slave_id	= SHDMA_SLAVE_MMCIF_RX,
-	},
-	.chan_priv_tx	= {
-		.slave_id	= SHDMA_SLAVE_MMCIF_TX,
-	},
-};
-
 static struct sh_mmcif_plat_data sh_mmcif_plat = {
 	.sup_pclk	= 0,
 	.ocr		= MMC_VDD_165_195 | MMC_VDD_32_33 | MMC_VDD_33_34,
@@ -311,7 +302,8 @@ static struct sh_mmcif_plat_data sh_mmcif_plat = {
 			  MMC_CAP_8_BIT_DATA |
 			  MMC_CAP_NEEDS_POLL,
 	.get_cd		= slot_cn7_get_cd,
-	.dma		= &sh_mmcif_dma,
+	.slave_id_tx	= SHDMA_SLAVE_MMCIF_TX,
+	.slave_id_rx	= SHDMA_SLAVE_MMCIF_RX,
 };
 
 static struct platform_device sh_mmcif_device = {
diff --git a/arch/arm/mach-shmobile/board-kota2.c b/arch/arm/mach-shmobile/board-kota2.c
index 857ceeec1bb0..c8e7ca23fc06 100644
--- a/arch/arm/mach-shmobile/board-kota2.c
+++ b/arch/arm/mach-shmobile/board-kota2.c
@@ -143,11 +143,10 @@ static struct gpio_keys_button gpio_buttons[] = {
 static struct gpio_keys_platform_data gpio_key_info = {
 	.buttons        = gpio_buttons,
 	.nbuttons       = ARRAY_SIZE(gpio_buttons),
-	.poll_interval  = 250, /* polled for now */
 };
 
 static struct platform_device gpio_keys_device = {
-	.name   = "gpio-keys-polled", /* polled for now */
+	.name   = "gpio-keys",
 	.id     = -1,
 	.dev    = {
 		.platform_data  = &gpio_key_info,
diff --git a/arch/arm/mach-shmobile/board-mackerel.c b/arch/arm/mach-shmobile/board-mackerel.c
index 9b42fbd10f8e..a2813247b455 100644
--- a/arch/arm/mach-shmobile/board-mackerel.c
+++ b/arch/arm/mach-shmobile/board-mackerel.c
@@ -43,7 +43,6 @@
 #include <linux/smsc911x.h>
 #include <linux/sh_intc.h>
 #include <linux/tca6416_keypad.h>
-#include <linux/usb/r8a66597.h>
 #include <linux/usb/renesas_usbhs.h>
 #include <linux/dma-mapping.h>
 
@@ -145,11 +144,6 @@
  * 1-2 short | VBUS 5V       | Host
  * open      | external VBUS | Function
  *
- * *1
- * CN31 is used as
- * CONFIG_USB_R8A66597_HCD	Host
- * CONFIG_USB_RENESAS_USBHS	Function
- *
  * CAUTION
  *
  * renesas_usbhs driver can use external interrupt mode
@@ -161,15 +155,6 @@
  * mackerel can not use external interrupt (IRQ7-PORT167) mode on "USB0",
  * because Touchscreen is using IRQ7-PORT40.
  * It is impossible to use IRQ7 demux on this board.
- *
- * We can use external interrupt mode USB-Function on "USB1".
- * USB1 can become Host by r8a66597, and become Function by renesas_usbhs.
- * But don't select both drivers in same time.
- * These uses same IRQ number for request_irq(), and aren't supporting
- * IRQF_SHARED / IORESOURCE_IRQ_SHAREABLE.
- *
- * Actually these are old/new version of USB driver.
- * This mean its register will be broken if it supports shared IRQ,
  */
 
 /*
@@ -208,6 +193,16 @@
  */
 
 /*
+ * FSI - AK4642
+ *
+ * it needs amixer settings for playing
+ *
+ * amixer set "Headphone" on
+ * amixer set "HPOUTL Mixer DACH" on
+ * amixer set "HPOUTR Mixer DACH" on
+ */
+
+/*
  * FIXME !!
  *
  * gpio_no_direction
@@ -676,51 +671,16 @@ static struct platform_device usbhs0_device = {
  * Use J30 to select between Host and Function. This setting
  * can however not be detected by software. Hotplug of USBHS1
  * is provided via IRQ8.
+ *
+ * Current USB1 works as "USB Host".
+ *  - set J30 "short"
+ *
+ * If you want to use it as "USB gadget",
+ *  - J30 "open"
+ *  - modify usbhs1_get_id() USBHS_HOST -> USBHS_GADGET
+ *  - add .get_vbus = usbhs_get_vbus in usbhs1_private
  */
 #define IRQ8 evt2irq(0x0300)
-
-/* USBHS1 USB Host support via r8a66597_hcd */
-static void usb1_host_port_power(int port, int power)
-{
-	if (!power) /* only power-on is supported for now */
-		return;
-
-	/* set VBOUT/PWEN and EXTLP1 in DVSTCTR */
-	__raw_writew(__raw_readw(0xE68B0008) | 0x600, 0xE68B0008);
-}
-
-static struct r8a66597_platdata usb1_host_data = {
-	.on_chip	= 1,
-	.port_power	= usb1_host_port_power,
-};
-
-static struct resource usb1_host_resources[] = {
-	[0] = {
-		.name	= "USBHS1",
-		.start	= 0xe68b0000,
-		.end	= 0xe68b00e6 - 1,
-		.flags	= IORESOURCE_MEM,
-	},
-	[1] = {
-		.start	= evt2irq(0x1ce0) /* USB1_USB1I0 */,
-		.flags	= IORESOURCE_IRQ,
-	},
-};
-
-static struct platform_device usb1_host_device = {
-	.name	= "r8a66597_hcd",
-	.id	= 1,
-	.dev = {
-		.dma_mask		= NULL,         /*  not use dma */
-		.coherent_dma_mask	= 0xffffffff,
-		.platform_data		= &usb1_host_data,
-	},
-	.num_resources	= ARRAY_SIZE(usb1_host_resources),
-	.resource	= usb1_host_resources,
-};
-
-/* USBHS1 USB Function support via renesas_usbhs */
-
 #define USB_PHY_MODE		(1 << 4)
 #define USB_PHY_INT_EN		((1 << 3) | (1 << 2))
 #define USB_PHY_ON		(1 << 1)
@@ -776,7 +736,7 @@ static void usbhs1_hardware_exit(struct platform_device *pdev)
 
 static int usbhs1_get_id(struct platform_device *pdev)
 {
-	return USBHS_GADGET;
+	return USBHS_HOST;
 }
 
 static u32 usbhs1_pipe_cfg[] = {
@@ -807,7 +767,6 @@ static struct usbhs_private usbhs1_private = {
 			.hardware_exit	= usbhs1_hardware_exit,
 			.get_id		= usbhs1_get_id,
 			.phy_reset	= usbhs_phy_reset,
-			.get_vbus	= usbhs_get_vbus,
 		},
 		.driver_param = {
 			.buswait_bwait	= 4,
@@ -1184,15 +1143,6 @@ static struct resource sh_mmcif_resources[] = {
 	},
 };
 
-static struct sh_mmcif_dma sh_mmcif_dma = {
-	.chan_priv_rx	= {
-		.slave_id	= SHDMA_SLAVE_MMCIF_RX,
-	},
-	.chan_priv_tx	= {
-		.slave_id	= SHDMA_SLAVE_MMCIF_TX,
-	},
-};
-
 static struct sh_mmcif_plat_data sh_mmcif_plat = {
 	.sup_pclk	= 0,
 	.ocr		= MMC_VDD_165_195 | MMC_VDD_32_33 | MMC_VDD_33_34,
@@ -1200,7 +1150,8 @@ static struct sh_mmcif_plat_data sh_mmcif_plat = {
 			  MMC_CAP_8_BIT_DATA |
 			  MMC_CAP_NEEDS_POLL,
 	.get_cd		= slot_cn7_get_cd,
-	.dma		= &sh_mmcif_dma,
+	.slave_id_tx	= SHDMA_SLAVE_MMCIF_TX,
+	.slave_id_rx	= SHDMA_SLAVE_MMCIF_RX,
 };
 
 static struct platform_device sh_mmcif_device = {
@@ -1311,7 +1262,6 @@ static struct platform_device *mackerel_devices[] __initdata = {
 	&nor_flash_device,
 	&smc911x_device,
 	&lcdc_device,
-	&usb1_host_device,
 	&usbhs1_device,
 	&usbhs0_device,
 	&leds_device,
@@ -1473,9 +1423,6 @@ static void __init mackerel_init(void)
 	gpio_pull_down(GPIO_PORT167CR); /* VBUS0_1 pull down */
 	gpio_request(GPIO_FN_IDIN_1_113, NULL);
 
-	/* USB phy tweak to make the r8a66597_hcd host driver work */
-	__raw_writew(0x8a0a, 0xe6058130);       /* USBCR4 */
-
 	/* enable FSI2 port A (ak4643) */
 	gpio_request(GPIO_FN_FSIAIBT,	NULL);
 	gpio_request(GPIO_FN_FSIAILR,	NULL);
diff --git a/arch/arm/mach-shmobile/clock-sh73a0.c b/arch/arm/mach-shmobile/clock-sh73a0.c
index afbead6a6e17..7727cca6136c 100644
--- a/arch/arm/mach-shmobile/clock-sh73a0.c
+++ b/arch/arm/mach-shmobile/clock-sh73a0.c
@@ -365,6 +365,114 @@ static struct clk div6_clks[DIV6_NR] = {
 			dsi_parent, ARRAY_SIZE(dsi_parent), 12, 3),
 };
 
+/* DSI DIV */
+static unsigned long dsiphy_recalc(struct clk *clk)
+{
+	u32 value;
+
+	value = __raw_readl(clk->mapping->base);
+
+	/* FIXME */
+	if (!(value & 0x000B8000))
+		return clk->parent->rate;
+
+	value &= 0x3f;
+	value += 1;
+
+	if ((value < 12) ||
+	    (value > 33)) {
+		pr_err("DSIPHY has wrong value (%d)", value);
+		return 0;
+	}
+
+	return clk->parent->rate / value;
+}
+
+static long dsiphy_round_rate(struct clk *clk, unsigned long rate)
+{
+	return clk_rate_mult_range_round(clk, 12, 33, rate);
+}
+
+static void dsiphy_disable(struct clk *clk)
+{
+	u32 value;
+
+	value = __raw_readl(clk->mapping->base);
+	value &= ~0x000B8000;
+
+	__raw_writel(value , clk->mapping->base);
+}
+
+static int dsiphy_enable(struct clk *clk)
+{
+	u32 value;
+	int multi;
+
+	value = __raw_readl(clk->mapping->base);
+	multi = (value & 0x3f) + 1;
+
+	if ((multi < 12) || (multi > 33))
+		return -EIO;
+
+	__raw_writel(value | 0x000B8000, clk->mapping->base);
+
+	return 0;
+}
+
+static int dsiphy_set_rate(struct clk *clk, unsigned long rate)
+{
+	u32 value;
+	int idx;
+
+	idx = rate / clk->parent->rate;
+	if ((idx < 12) || (idx > 33))
+		return -EINVAL;
+
+	idx += -1;
+
+	value = __raw_readl(clk->mapping->base);
+	value = (value & ~0x3f) + idx;
+
+	__raw_writel(value, clk->mapping->base);
+
+	return 0;
+}
+
+static struct clk_ops dsiphy_clk_ops = {
+	.recalc		= dsiphy_recalc,
+	.round_rate	= dsiphy_round_rate,
+	.set_rate	= dsiphy_set_rate,
+	.enable		= dsiphy_enable,
+	.disable	= dsiphy_disable,
+};
+
+static struct clk_mapping dsi0phy_clk_mapping = {
+	.phys	= DSI0PHYCR,
+	.len	= 4,
+};
+
+static struct clk_mapping dsi1phy_clk_mapping = {
+	.phys	= DSI1PHYCR,
+	.len	= 4,
+};
+
+static struct clk dsi0phy_clk = {
+	.ops		= &dsiphy_clk_ops,
+	.parent		= &div6_clks[DIV6_DSI0P], /* late install */
+	.mapping	= &dsi0phy_clk_mapping,
+};
+
+static struct clk dsi1phy_clk = {
+	.ops		= &dsiphy_clk_ops,
+	.parent		= &div6_clks[DIV6_DSI1P], /* late install */
+	.mapping	= &dsi1phy_clk_mapping,
+};
+
+static struct clk *late_main_clks[] = {
+	&dsi0phy_clk,
+	&dsi1phy_clk,
+};
+
 enum { MSTP001,
 	MSTP129, MSTP128, MSTP127, MSTP126, MSTP125, MSTP118, MSTP116, MSTP100,
 	MSTP219,
@@ -429,6 +537,8 @@ static struct clk_lookup lookups[] = {
 	CLKDEV_ICK_ID("dsit_clk", "sh-mipi-dsi.1", &div6_clks[DIV6_DSIT]),
 	CLKDEV_ICK_ID("dsip_clk", "sh-mipi-dsi.0", &div6_clks[DIV6_DSI0P]),
 	CLKDEV_ICK_ID("dsip_clk", "sh-mipi-dsi.1", &div6_clks[DIV6_DSI1P]),
+	CLKDEV_ICK_ID("dsiphy_clk", "sh-mipi-dsi.0", &dsi0phy_clk),
+	CLKDEV_ICK_ID("dsiphy_clk", "sh-mipi-dsi.1", &dsi1phy_clk),
 
 	/* MSTP32 clocks */
 	CLKDEV_DEV_ID("i2c-sh_mobile.2", &mstp_clks[MSTP001]), /* I2C2 */
@@ -504,6 +614,9 @@ void __init sh73a0_clock_init(void)
 	if (!ret)
 		ret = sh_clk_mstp32_register(mstp_clks, MSTP_NR);
 
+	for (k = 0; !ret && (k < ARRAY_SIZE(late_main_clks)); k++)
+		ret = clk_register(late_main_clks[k]);
+
 	clkdev_add_table(lookups, ARRAY_SIZE(lookups));
 
 	if (!ret)
diff --git a/arch/arm/mach-shmobile/include/mach/sh73a0.h b/arch/arm/mach-shmobile/include/mach/sh73a0.h
index 881d515a9686..cad57578ceed 100644
--- a/arch/arm/mach-shmobile/include/mach/sh73a0.h
+++ b/arch/arm/mach-shmobile/include/mach/sh73a0.h
@@ -515,8 +515,8 @@ enum {
 	SHDMA_SLAVE_MMCIF_RX,
 };
 
-/* PINT interrupts are located at Linux IRQ 768 and up */
-#define SH73A0_PINT0_IRQ(irq) ((irq) + 768)
-#define SH73A0_PINT1_IRQ(irq) ((irq) + 800)
+/* PINT interrupts are located at Linux IRQ 800 and up */
+#define SH73A0_PINT0_IRQ(irq) ((irq) + 800)
+#define SH73A0_PINT1_IRQ(irq) ((irq) + 832)
 
 #endif /* __ASM_SH73A0_H__ */
diff --git a/arch/arm/mach-shmobile/intc-sh73a0.c b/arch/arm/mach-shmobile/intc-sh73a0.c
index 1eda6b0b69e3..9857595eaa79 100644
--- a/arch/arm/mach-shmobile/intc-sh73a0.c
+++ b/arch/arm/mach-shmobile/intc-sh73a0.c
@@ -19,6 +19,7 @@
 #include <linux/kernel.h>
 #include <linux/init.h>
 #include <linux/interrupt.h>
+#include <linux/module.h>
 #include <linux/irq.h>
 #include <linux/io.h>
 #include <linux/sh_intc.h>
@@ -445,6 +446,7 @@ void __init sh73a0_init_irq(void)
 		setup_irq(gic_spi(1 + k), &sh73a0_irq_pin_cascade[k]);
 
 		n = intcs_evt2irq(to_intc_vect(gic_spi(1 + k)));
+		WARN_ON(irq_alloc_desc_at(n, numa_node_id()) != n);
 		irq_set_chip_and_handler_name(n, &intca_gic_irq_chip,
 					      handle_level_irq, "level");
 		set_irq_flags(n, IRQF_VALID); /* yuck */
diff --git a/arch/arm/mach-shmobile/pfc-r8a7779.c b/arch/arm/mach-shmobile/pfc-r8a7779.c
index 963532f2b2c4..d14c9b048077 100644
--- a/arch/arm/mach-shmobile/pfc-r8a7779.c
+++ b/arch/arm/mach-shmobile/pfc-r8a7779.c
@@ -2120,7 +2120,7 @@ static struct pinmux_cfg_reg pinmux_config_regs[] = {
 	    FN_AUDATA3, 0, 0, 0 }
 	},
 	{ PINMUX_CFG_REG_VAR("IPSR4", 0xfffc0030, 32,
-			     3, 1, 1, 1, 1, 1, 1, 3, 3, 1,
+			     3, 1, 1, 1, 1, 1, 1, 3, 3,
 			     1, 1, 1, 1, 1, 1, 3, 3, 3, 2) {
 	    /* IP4_31_29 [3] */
 	    FN_DU1_DB0, FN_VI2_DATA4_VI2_B4, FN_SCL2_B, FN_SD3_DAT0,
diff --git a/arch/arm/mach-shmobile/pfc-sh7372.c b/arch/arm/mach-shmobile/pfc-sh7372.c
index 1bd6585a6acf..336093f9210a 100644
--- a/arch/arm/mach-shmobile/pfc-sh7372.c
+++ b/arch/arm/mach-shmobile/pfc-sh7372.c
@@ -23,6 +23,7 @@
 #include <linux/init.h>
 #include <linux/kernel.h>
 #include <linux/gpio.h>
+#include <mach/irqs.h>
 #include <mach/sh7372.h>
 
 #define CPU_ALL_PORT(fn, pfx, sfx) \
@@ -1594,6 +1595,43 @@ static struct pinmux_data_reg pinmux_data_regs[] = {
 	{ },
 };
 
+#define EXT_IRQ16L(n) evt2irq(0x200 + ((n) << 5))
+#define EXT_IRQ16H(n) evt2irq(0x3200 + (((n) - 16) << 5))
+static struct pinmux_irq pinmux_irqs[] = {
+	PINMUX_IRQ(EXT_IRQ16L(0), PORT6_FN0, PORT162_FN0),
+	PINMUX_IRQ(EXT_IRQ16L(1), PORT12_FN0),
+	PINMUX_IRQ(EXT_IRQ16L(2), PORT4_FN0, PORT5_FN0),
+	PINMUX_IRQ(EXT_IRQ16L(3), PORT8_FN0, PORT16_FN0),
+	PINMUX_IRQ(EXT_IRQ16L(4), PORT17_FN0, PORT163_FN0),
+	PINMUX_IRQ(EXT_IRQ16L(5), PORT18_FN0),
+	PINMUX_IRQ(EXT_IRQ16L(6), PORT39_FN0, PORT164_FN0),
+	PINMUX_IRQ(EXT_IRQ16L(7), PORT40_FN0, PORT167_FN0),
+	PINMUX_IRQ(EXT_IRQ16L(8), PORT41_FN0, PORT168_FN0),
+	PINMUX_IRQ(EXT_IRQ16L(9), PORT42_FN0, PORT169_FN0),
+	PINMUX_IRQ(EXT_IRQ16L(10), PORT65_FN0),
+	PINMUX_IRQ(EXT_IRQ16L(11), PORT67_FN0),
+	PINMUX_IRQ(EXT_IRQ16L(12), PORT80_FN0, PORT137_FN0),
+	PINMUX_IRQ(EXT_IRQ16L(13), PORT81_FN0, PORT145_FN0),
+	PINMUX_IRQ(EXT_IRQ16L(14), PORT82_FN0, PORT146_FN0),
+	PINMUX_IRQ(EXT_IRQ16L(15), PORT83_FN0, PORT147_FN0),
+	PINMUX_IRQ(EXT_IRQ16H(16), PORT84_FN0, PORT170_FN0),
+	PINMUX_IRQ(EXT_IRQ16H(17), PORT85_FN0),
+	PINMUX_IRQ(EXT_IRQ16H(18), PORT86_FN0),
+	PINMUX_IRQ(EXT_IRQ16H(19), PORT87_FN0),
+	PINMUX_IRQ(EXT_IRQ16H(20), PORT92_FN0),
+	PINMUX_IRQ(EXT_IRQ16H(21), PORT93_FN0),
+	PINMUX_IRQ(EXT_IRQ16H(22), PORT94_FN0),
+	PINMUX_IRQ(EXT_IRQ16H(23), PORT95_FN0),
+	PINMUX_IRQ(EXT_IRQ16H(24), PORT112_FN0),
+	PINMUX_IRQ(EXT_IRQ16H(25), PORT119_FN0),
+	PINMUX_IRQ(EXT_IRQ16H(26), PORT121_FN0, PORT172_FN0),
+	PINMUX_IRQ(EXT_IRQ16H(27), PORT122_FN0, PORT180_FN0),
+	PINMUX_IRQ(EXT_IRQ16H(28), PORT123_FN0, PORT181_FN0),
+	PINMUX_IRQ(EXT_IRQ16H(29), PORT129_FN0, PORT182_FN0),
+	PINMUX_IRQ(EXT_IRQ16H(30), PORT130_FN0, PORT183_FN0),
+	PINMUX_IRQ(EXT_IRQ16H(31), PORT138_FN0, PORT184_FN0),
+};
+
 static struct pinmux_info sh7372_pinmux_info = {
 	.name = "sh7372_pfc",
 	.reserved_id = PINMUX_RESERVED,
@@ -1614,6 +1652,9 @@ static struct pinmux_info sh7372_pinmux_info = {
 
 	.gpio_data = pinmux_data,
 	.gpio_data_size = ARRAY_SIZE(pinmux_data),
+
+	.gpio_irq = pinmux_irqs,
+	.gpio_irq_size = ARRAY_SIZE(pinmux_irqs),
 };
 
 void sh7372_pinmux_init(void)
diff --git a/arch/arm/mach-shmobile/smp-sh73a0.c b/arch/arm/mach-shmobile/smp-sh73a0.c
index 0d159d64a345..2d0d4212be41 100644
--- a/arch/arm/mach-shmobile/smp-sh73a0.c
+++ b/arch/arm/mach-shmobile/smp-sh73a0.c
@@ -80,7 +80,7 @@ int __cpuinit sh73a0_boot_secondary(unsigned int cpu)
 	/* enable cache coherency */
 	modify_scu_cpu_psr(0, 3 << (cpu * 8));
 
-	if (((__raw_readw(__io(PSTR)) >> (4 * cpu)) & 3) == 3)
+	if (((__raw_readl(__io(PSTR)) >> (4 * cpu)) & 3) == 3)
 		__raw_writel(1 << cpu, __io(WUPCR));	/* wake up */
 	else
 		__raw_writel(1 << cpu, __io(SRESCR));	/* reset */
diff --git a/arch/arm/mm/Kconfig b/arch/arm/mm/Kconfig
index 1a3ca2488164..7edef9121632 100644
--- a/arch/arm/mm/Kconfig
+++ b/arch/arm/mm/Kconfig
@@ -631,7 +631,8 @@ comment "Processor Features"
 
 config ARM_LPAE
 	bool "Support for the Large Physical Address Extension"
-	depends on MMU && CPU_V7
+	depends on MMU && CPU_32v7 && !CPU_32v6 && !CPU_32v5 && \
+		!CPU_32v4 && !CPU_32v3
 	help
 	  Say Y if you have an ARMv7 processor supporting the LPAE page
 	  table format and you would like to access memory beyond the
diff --git a/arch/arm/mm/cache-v7.S b/arch/arm/mm/cache-v7.S
index 7a24d39661f0..a655d3da386d 100644
--- a/arch/arm/mm/cache-v7.S
+++ b/arch/arm/mm/cache-v7.S
@@ -55,7 +55,7 @@ loop1:
 	cmp	r1, #2				@ see what cache we have at this level
 	blt	skip				@ skip if no cache, or just i-cache
 #ifdef CONFIG_PREEMPT
-	save_and_disable_irqs r9		@ make cssr&csidr read atomic
+	save_and_disable_irqs_notrace r9	@ make cssr&csidr read atomic
 #endif
 	mcr	p15, 2, r10, c0, c0, 0		@ select current cache level in cssr
 	isb					@ isb to sych the new cssr&csidr
diff --git a/arch/c6x/boot/Makefile b/arch/c6x/boot/Makefile
index ecca820e6041..6891257d514c 100644
--- a/arch/c6x/boot/Makefile
+++ b/arch/c6x/boot/Makefile
@@ -13,7 +13,7 @@ obj-y += linked_dtb.o
 endif
 
 $(obj)/%.dtb: $(src)/dts/%.dts FORCE
-	$(call cmd,dtc)
+	$(call if_changed_dep,dtc)
 
 quiet_cmd_cp = CP      $< $@$2
 	cmd_cp = cat $< >$@$2 || (rm -f $@ && echo false)
diff --git a/arch/m68k/include/asm/mcf_pgtable.h b/arch/m68k/include/asm/mcf_pgtable.h
index 756bde4fb4f8..3c793682e5d9 100644
--- a/arch/m68k/include/asm/mcf_pgtable.h
+++ b/arch/m68k/include/asm/mcf_pgtable.h
@@ -78,7 +78,8 @@
 				 | CF_PAGE_READABLE \
 				 | CF_PAGE_WRITABLE \
 				 | CF_PAGE_EXEC \
-				 | CF_PAGE_SYSTEM)
+				 | CF_PAGE_SYSTEM \
+				 | CF_PAGE_SHARED)
 
 #define PAGE_COPY	__pgprot(CF_PAGE_VALID \
 				 | CF_PAGE_ACCESSED \
diff --git a/arch/m68k/mm/mcfmmu.c b/arch/m68k/mm/mcfmmu.c
index babd5a97cdcb..875b800ef0dd 100644
--- a/arch/m68k/mm/mcfmmu.c
+++ b/arch/m68k/mm/mcfmmu.c
@@ -87,7 +87,7 @@ void __init paging_init(void)
 
 int cf_tlb_miss(struct pt_regs *regs, int write, int dtlb, int extension_word)
 {
-	unsigned long flags, mmuar;
+	unsigned long flags, mmuar, mmutr;
 	struct mm_struct *mm;
 	pgd_t *pgd;
 	pmd_t *pmd;
@@ -137,9 +137,10 @@ int cf_tlb_miss(struct pt_regs *regs, int write, int dtlb, int extension_word)
 	if (!pte_dirty(*pte) && !KMAPAREA(mmuar))
 		set_pte(pte, pte_wrprotect(*pte));
 
-	mmu_write(MMUTR, (mmuar & PAGE_MASK) | (asid << MMUTR_IDN) |
-		(((int)(pte->pte) & (int)CF_PAGE_MMUTR_MASK)
-		>> CF_PAGE_MMUTR_SHIFT) | MMUTR_V);
+	mmutr = (mmuar & PAGE_MASK) | (asid << MMUTR_IDN) | MMUTR_V;
+	if ((mmuar < TASK_UNMAPPED_BASE) || (mmuar >= TASK_SIZE))
+		mmutr |= (pte->pte & CF_PAGE_MMUTR_MASK) >> CF_PAGE_MMUTR_SHIFT;
+	mmu_write(MMUTR, mmutr);
 
 	mmu_write(MMUDR, (pte_val(*pte) & PAGE_MASK) |
 		((pte->pte) & CF_PAGE_MMUDR_MASK) | MMUDR_SZ_8KB | MMUDR_X);
diff --git a/arch/m68k/platform/coldfire/entry.S b/arch/m68k/platform/coldfire/entry.S
index 863889fc31c9..281e38c2b6c7 100644
--- a/arch/m68k/platform/coldfire/entry.S
+++ b/arch/m68k/platform/coldfire/entry.S
@@ -136,7 +136,7 @@ Luser_return:
 	movel	%sp,%d1			/* get thread_info pointer */
 	andl	#-THREAD_SIZE,%d1	/* at base of kernel stack */
 	movel	%d1,%a0
-	movel	%a0@(TINFO_FLAGS),%d1	/* get thread_info->flags */
+	moveb	%a0@(TINFO_FLAGS+3),%d1	/* thread_info->flags (low 8 bits) */
 	jne	Lwork_to_do		/* still work to do */
 
 Lreturn:
@@ -148,8 +148,6 @@ Lwork_to_do:
 	btst	#TIF_NEED_RESCHED,%d1
 	jne	reschedule
 
-	/* GERG: do we need something here for TRACEing?? */
-
 Lsignal_return:
 	subql	#4,%sp			/* dummy return address */
 	SAVE_SWITCH_STACK
diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
index 4f80cf1ce77b..3e57a00b8cba 100644
--- a/arch/powerpc/kernel/entry_32.S
+++ b/arch/powerpc/kernel/entry_32.S
@@ -1213,7 +1213,7 @@ do_user_signal:			/* r10 contains MSR_KERNEL here */
 	stw	r3,_TRAP(r1)
 2:	addi	r3,r1,STACK_FRAME_OVERHEAD
 	mr	r4,r9
-	bl	do_signal
+	bl	do_notify_resume
 	REST_NVGPRS(r1)
 	b	recheck
 
diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S
index d834425186ae..866462cbe2d8 100644
--- a/arch/powerpc/kernel/entry_64.S
+++ b/arch/powerpc/kernel/entry_64.S
@@ -751,12 +751,16 @@ user_work:
 
 	andi.	r0,r4,_TIF_NEED_RESCHED
 	beq	1f
+	li	r5,1
+	TRACE_AND_RESTORE_IRQ(r5);
 	bl	.schedule
 	b	.ret_from_except_lite
 
 1:	bl	.save_nvgprs
+	li	r5,1
+	TRACE_AND_RESTORE_IRQ(r5);
 	addi	r3,r1,STACK_FRAME_OVERHEAD
-	bl	.do_signal
+	bl	.do_notify_resume
 	b	.ret_from_except
 
 unrecov_restore:
diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S
index 3844ca7c5099..15c5a4f6de01 100644
--- a/arch/powerpc/kernel/exceptions-64s.S
+++ b/arch/powerpc/kernel/exceptions-64s.S
@@ -774,8 +774,8 @@ alignment_common:
 program_check_common:
 	EXCEPTION_PROLOG_COMMON(0x700, PACA_EXGEN)
 	bl	.save_nvgprs
-	addi	r3,r1,STACK_FRAME_OVERHEAD
 	DISABLE_INTS
+	addi	r3,r1,STACK_FRAME_OVERHEAD
 	bl	.program_check_exception
 	b	.ret_from_except
 
diff --git a/arch/powerpc/kernel/signal.c b/arch/powerpc/kernel/signal.c
index 2300426e531a..ac6e437b1021 100644
--- a/arch/powerpc/kernel/signal.c
+++ b/arch/powerpc/kernel/signal.c
@@ -11,6 +11,7 @@
 
 #include <linux/tracehook.h>
 #include <linux/signal.h>
+#include <linux/key.h>
 #include <asm/hw_breakpoint.h>
 #include <asm/uaccess.h>
 #include <asm/unistd.h>
@@ -113,8 +114,9 @@ static void check_syscall_restart(struct pt_regs *regs, struct k_sigaction *ka,
 	}
 }
 
-static int do_signal_pending(sigset_t *oldset, struct pt_regs *regs)
+static int do_signal(struct pt_regs *regs)
 {
+	sigset_t *oldset;
 	siginfo_t info;
 	int signr;
 	struct k_sigaction ka;
@@ -123,7 +125,7 @@ static int do_signal_pending(sigset_t *oldset, struct pt_regs *regs)
 
 	if (current_thread_info()->local_flags & _TLF_RESTORE_SIGMASK)
 		oldset = &current->saved_sigmask;
-	else if (!oldset)
+	else
 		oldset = &current->blocked;
 
 	signr = get_signal_to_deliver(&info, &ka, regs, NULL);
@@ -191,14 +193,16 @@ static int do_signal_pending(sigset_t *oldset, struct pt_regs *regs)
 	return ret;
 }
 
-void do_signal(struct pt_regs *regs, unsigned long thread_info_flags)
+void do_notify_resume(struct pt_regs *regs, unsigned long thread_info_flags)
 {
 	if (thread_info_flags & _TIF_SIGPENDING)
-		do_signal_pending(NULL, regs);
+		do_signal(regs);
 
 	if (thread_info_flags & _TIF_NOTIFY_RESUME) {
 		clear_thread_flag(TIF_NOTIFY_RESUME);
 		tracehook_notify_resume(regs);
+		if (current->replacement_session_keyring)
+			key_replace_session_keyring();
 	}
 }
 
diff --git a/arch/powerpc/kernel/signal.h b/arch/powerpc/kernel/signal.h
index 6c0ddfc0603e..8dde973aaaf5 100644
--- a/arch/powerpc/kernel/signal.h
+++ b/arch/powerpc/kernel/signal.h
@@ -12,7 +12,7 @@
 
 #define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP)))
 
-extern void do_signal(struct pt_regs *regs, unsigned long thread_info_flags);
+extern void do_notify_resume(struct pt_regs *regs, unsigned long thread_info_flags);
 
 extern void __user * get_sigframe(struct k_sigaction *ka, struct pt_regs *regs,
 				  size_t frame_size, int is_32);
diff --git a/arch/powerpc/platforms/wsp/smp.c b/arch/powerpc/platforms/wsp/smp.c
index 71bd105f3863..0ba103ae83a5 100644
--- a/arch/powerpc/platforms/wsp/smp.c
+++ b/arch/powerpc/platforms/wsp/smp.c
@@ -71,7 +71,7 @@ int __devinit smp_a2_kick_cpu(int nr)
 
 static int __init smp_a2_probe(void)
 {
-	return cpus_weight(cpu_possible_map);
+	return num_possible_cpus();
 }
 
 static struct smp_ops_t a2_smp_ops = {
diff --git a/arch/s390/kernel/compat_wrapper.S b/arch/s390/kernel/compat_wrapper.S
index 18c51df9fe06..ff605a39cf43 100644
--- a/arch/s390/kernel/compat_wrapper.S
+++ b/arch/s390/kernel/compat_wrapper.S
@@ -662,7 +662,7 @@ ENTRY(sys32_getresuid16_wrapper)
 ENTRY(sys32_poll_wrapper)
 	llgtr	%r2,%r2			# struct pollfd *
 	llgfr	%r3,%r3			# unsigned int
-	lgfr	%r4,%r4			# long
+	lgfr	%r4,%r4			# int
 	jg	sys_poll		# branch to system call
 
 ENTRY(sys32_setresgid16_wrapper)
diff --git a/arch/s390/kernel/process.c b/arch/s390/kernel/process.c
index 3201ae447990..4261aa799774 100644
--- a/arch/s390/kernel/process.c
+++ b/arch/s390/kernel/process.c
@@ -76,7 +76,6 @@ static void default_idle(void)
 	if (test_thread_flag(TIF_MCCK_PENDING)) {
 		local_mcck_enable();
 		local_irq_enable();
-		s390_handle_mcck();
 		return;
 	}
 	trace_hardirqs_on();
@@ -93,10 +92,12 @@ void cpu_idle(void)
 	for (;;) {
 		tick_nohz_idle_enter();
 		rcu_idle_enter();
-		while (!need_resched())
+		while (!need_resched() && !test_thread_flag(TIF_MCCK_PENDING))
 			default_idle();
 		rcu_idle_exit();
 		tick_nohz_idle_exit();
+		if (test_thread_flag(TIF_MCCK_PENDING))
+			s390_handle_mcck();
 		preempt_enable_no_resched();
 		schedule();
 		preempt_disable();
diff --git a/arch/s390/kernel/time.c b/arch/s390/kernel/time.c
index fa02f443f5f6..14da278febbf 100644
--- a/arch/s390/kernel/time.c
+++ b/arch/s390/kernel/time.c
@@ -113,11 +113,14 @@ static void fixup_clock_comparator(unsigned long long delta)
 static int s390_next_ktime(ktime_t expires,
 			   struct clock_event_device *evt)
 {
+	struct timespec ts;
 	u64 nsecs;
 
-	nsecs = ktime_to_ns(ktime_sub(expires, ktime_get_monotonic_offset()));
+	ts.tv_sec = ts.tv_nsec = 0;
+	monotonic_to_bootbased(&ts);
+	nsecs = ktime_to_ns(ktime_add(timespec_to_ktime(ts), expires));
 	do_div(nsecs, 125);
-	S390_lowcore.clock_comparator = TOD_UNIX_EPOCH + (nsecs << 9);
+	S390_lowcore.clock_comparator = sched_clock_base_cc + (nsecs << 9);
 	set_clock_comparator(S390_lowcore.clock_comparator);
 	return 0;
 }
diff --git a/arch/s390/mm/pgtable.c b/arch/s390/mm/pgtable.c
index 9a4d02f64f16..51b0738e13d1 100644
--- a/arch/s390/mm/pgtable.c
+++ b/arch/s390/mm/pgtable.c
@@ -574,7 +574,7 @@ static inline void page_table_free_pgste(unsigned long *table)
 	page = pfn_to_page(__pa(table) >> PAGE_SHIFT);
 	mp = (struct gmap_pgtable *) page->index;
 	BUG_ON(!list_empty(&mp->mapper));
-	pgtable_page_ctor(page);
+	pgtable_page_dtor(page);
 	atomic_set(&page->_mapcount, -1);
 	kfree(mp);
 	__free_page(page);
diff --git a/arch/sh/boards/board-sh7757lcr.c b/arch/sh/boards/board-sh7757lcr.c
index 0838154dd216..24b1ee410daa 100644
--- a/arch/sh/boards/board-sh7757lcr.c
+++ b/arch/sh/boards/board-sh7757lcr.c
@@ -169,6 +169,11 @@ static struct resource sh_eth_giga1_resources[] = {
 		.end    = 0xfee00fff,
 		.flags  = IORESOURCE_MEM,
 	}, {
+		/* TSU */
+		.start  = 0xfee01800,
+		.end    = 0xfee01fff,
+		.flags  = IORESOURCE_MEM,
+	}, {
 		.start  = 316,
 		.end    = 316,
 		.flags  = IORESOURCE_IRQ,
@@ -210,20 +215,13 @@ static struct resource sh_mmcif_resources[] = {
 	},
 };
 
-static struct sh_mmcif_dma sh7757lcr_mmcif_dma = {
-	.chan_priv_tx	= {
-		.slave_id = SHDMA_SLAVE_MMCIF_TX,
-	},
-	.chan_priv_rx	= {
-		.slave_id = SHDMA_SLAVE_MMCIF_RX,
-	}
-};
-
 static struct sh_mmcif_plat_data sh_mmcif_plat = {
-	.dma		= &sh7757lcr_mmcif_dma,
 	.sup_pclk	= 0x0f,
-	.caps		= MMC_CAP_4_BIT_DATA | MMC_CAP_8_BIT_DATA,
+	.caps		= MMC_CAP_4_BIT_DATA | MMC_CAP_8_BIT_DATA |
+			  MMC_CAP_NONREMOVABLE,
 	.ocr		= MMC_VDD_32_33 | MMC_VDD_33_34,
+	.slave_id_tx	= SHDMA_SLAVE_MMCIF_TX,
+	.slave_id_rx	= SHDMA_SLAVE_MMCIF_RX,
 };
 
 static struct platform_device sh_mmcif_device = {
diff --git a/arch/sh/boards/mach-ap325rxa/setup.c b/arch/sh/boards/mach-ap325rxa/setup.c
index 6418e95c2b6b..ebd0f818a25f 100644
--- a/arch/sh/boards/mach-ap325rxa/setup.c
+++ b/arch/sh/boards/mach-ap325rxa/setup.c
@@ -22,6 +22,7 @@
 #include <linux/i2c.h>
 #include <linux/smsc911x.h>
 #include <linux/gpio.h>
+#include <linux/videodev2.h>
 #include <media/ov772x.h>
 #include <media/soc_camera.h>
 #include <media/soc_camera_platform.h>
diff --git a/arch/sh/boards/mach-ecovec24/setup.c b/arch/sh/boards/mach-ecovec24/setup.c
index 033ef2ba621f..cde7c0085ced 100644
--- a/arch/sh/boards/mach-ecovec24/setup.c
+++ b/arch/sh/boards/mach-ecovec24/setup.c
@@ -29,9 +29,11 @@
 #include <linux/input.h>
 #include <linux/input/sh_keysc.h>
 #include <linux/sh_eth.h>
+#include <linux/videodev2.h>
 #include <video/sh_mobile_lcdc.h>
 #include <sound/sh_fsi.h>
 #include <media/sh_mobile_ceu.h>
+#include <media/soc_camera.h>
 #include <media/tw9910.h>
 #include <media/mt9t112.h>
 #include <asm/heartbeat.h>
diff --git a/arch/sh/boards/mach-kfr2r09/setup.c b/arch/sh/boards/mach-kfr2r09/setup.c
index 2a18b06abdaf..5b382e1afaea 100644
--- a/arch/sh/boards/mach-kfr2r09/setup.c
+++ b/arch/sh/boards/mach-kfr2r09/setup.c
@@ -22,6 +22,7 @@
 #include <linux/input/sh_keysc.h>
 #include <linux/i2c.h>
 #include <linux/usb/r8a66597.h>
+#include <linux/videodev2.h>
 #include <media/rj54n1cb0c.h>
 #include <media/soc_camera.h>
 #include <media/sh_mobile_ceu.h>
diff --git a/arch/sh/boards/mach-migor/setup.c b/arch/sh/boards/mach-migor/setup.c
index 68c3d6f42896..d37ba2720527 100644
--- a/arch/sh/boards/mach-migor/setup.c
+++ b/arch/sh/boards/mach-migor/setup.c
@@ -21,9 +21,11 @@
 #include <linux/delay.h>
 #include <linux/clk.h>
 #include <linux/gpio.h>
+#include <linux/videodev2.h>
 #include <video/sh_mobile_lcdc.h>
 #include <media/sh_mobile_ceu.h>
 #include <media/ov772x.h>
+#include <media/soc_camera.h>
 #include <media/tw9910.h>
 #include <asm/clock.h>
 #include <asm/machvec.h>
diff --git a/arch/sh/boards/mach-se/7724/setup.c b/arch/sh/boards/mach-se/7724/setup.c
index 036fe1adaef1..2b07fc016950 100644
--- a/arch/sh/boards/mach-se/7724/setup.c
+++ b/arch/sh/boards/mach-se/7724/setup.c
@@ -24,6 +24,7 @@
 #include <linux/input/sh_keysc.h>
 #include <linux/usb/r8a66597.h>
 #include <linux/sh_eth.h>
+#include <linux/videodev2.h>
 #include <video/sh_mobile_lcdc.h>
 #include <media/sh_mobile_ceu.h>
 #include <sound/sh_fsi.h>
diff --git a/arch/sh/drivers/pci/pci-sh7780.c b/arch/sh/drivers/pci/pci-sh7780.c
index fa7b978cc727..fb8f14990743 100644
--- a/arch/sh/drivers/pci/pci-sh7780.c
+++ b/arch/sh/drivers/pci/pci-sh7780.c
@@ -74,7 +74,7 @@ struct pci_errors {
 	{ SH4_PCIINT_MLCK,	"master lock error" },
 	{ SH4_PCIINT_TABT,	"target-target abort" },
 	{ SH4_PCIINT_TRET,	"target retry time out" },
-	{ SH4_PCIINT_MFDE,	"master function disable erorr" },
+	{ SH4_PCIINT_MFDE,	"master function disable error" },
 	{ SH4_PCIINT_PRTY,	"address parity error" },
 	{ SH4_PCIINT_SERR,	"SERR" },
 	{ SH4_PCIINT_TWDP,	"data parity error for target write" },
diff --git a/arch/sh/include/asm/device.h b/arch/sh/include/asm/device.h
index a1c9c0daec10..071bcb4d4bfd 100644
--- a/arch/sh/include/asm/device.h
+++ b/arch/sh/include/asm/device.h
@@ -3,9 +3,10 @@
  *
  * This file is released under the GPLv2
  */
+#ifndef __ASM_SH_DEVICE_H
+#define __ASM_SH_DEVICE_H
 
-struct dev_archdata {
-};
+#include <asm-generic/device.h>
 
 struct platform_device;
 /* allocate contiguous memory chunk and fill in struct resource */
@@ -14,5 +15,4 @@ int platform_resource_setup_memory(struct platform_device *pdev,
 
 void plat_early_device_setup(void);
 
-struct pdev_archdata {
-};
+#endif /* __ASM_SH_DEVICE_H */
diff --git a/arch/sh/kernel/cpu/sh4a/clock-sh7724.c b/arch/sh/kernel/cpu/sh4a/clock-sh7724.c
index b3c039a5064a..70bd96646f42 100644
--- a/arch/sh/kernel/cpu/sh4a/clock-sh7724.c
+++ b/arch/sh/kernel/cpu/sh4a/clock-sh7724.c
@@ -343,7 +343,7 @@ static struct clk_lookup lookups[] = {
 	CLKDEV_DEV_ID("sh_mobile_ceu.1", &mstp_clks[HWBLK_CEU1]),
 	CLKDEV_CON_ID("beu1", &mstp_clks[HWBLK_BEU1]),
 	CLKDEV_CON_ID("2ddmac0", &mstp_clks[HWBLK_2DDMAC]),
-	CLKDEV_CON_ID("spu0", &mstp_clks[HWBLK_SPU]),
+	CLKDEV_DEV_ID("sh_fsi.0", &mstp_clks[HWBLK_SPU]),
 	CLKDEV_CON_ID("jpu0", &mstp_clks[HWBLK_JPU]),
 	CLKDEV_DEV_ID("sh-vou.0", &mstp_clks[HWBLK_VOU]),
 	CLKDEV_CON_ID("beu0", &mstp_clks[HWBLK_BEU0]),
diff --git a/arch/sh/kernel/cpu/sh4a/setup-sh7757.c b/arch/sh/kernel/cpu/sh4a/setup-sh7757.c
index a7b2da6b3a1a..2875e8be4f72 100644
--- a/arch/sh/kernel/cpu/sh4a/setup-sh7757.c
+++ b/arch/sh/kernel/cpu/sh4a/setup-sh7757.c
@@ -133,7 +133,7 @@ static struct resource spi0_resources[] = {
 	[0] = {
 		.start	= 0xfe002000,
 		.end	= 0xfe0020ff,
-		.flags	= IORESOURCE_MEM,
+		.flags	= IORESOURCE_MEM | IORESOURCE_MEM_32BIT,
 	},
 	[1] = {
 		.start	= 86,
@@ -661,6 +661,25 @@ static struct platform_device spi0_device = {
 	.resource	= spi0_resources,
 };
 
+static struct resource spi1_resources[] = {
+	{
+		.start	= 0xffd8ee70,
+		.end	= 0xffd8eeff,
+		.flags	= IORESOURCE_MEM | IORESOURCE_MEM_8BIT,
+	},
+	{
+		.start	= 54,
+		.flags	= IORESOURCE_IRQ,
+	},
+};
+
+static struct platform_device spi1_device = {
+	.name	= "sh_spi",
+	.id	= 1,
+	.num_resources	= ARRAY_SIZE(spi1_resources),
+	.resource	= spi1_resources,
+};
+
 static struct resource usb_ehci_resources[] = {
 	[0] = {
 		.start	= 0xfe4f1000,
@@ -720,6 +739,7 @@ static struct platform_device *sh7757_devices[] __initdata = {
 	&dma2_device,
 	&dma3_device,
 	&spi0_device,
+	&spi1_device,
 	&usb_ehci_device,
 	&usb_ohci_device,
 };
diff --git a/arch/sh/kernel/smp.c b/arch/sh/kernel/smp.c
index 3147a9a6fb8b..f624174bf239 100644
--- a/arch/sh/kernel/smp.c
+++ b/arch/sh/kernel/smp.c
@@ -63,7 +63,7 @@ void __init smp_prepare_cpus(unsigned int max_cpus)
 	mp_ops->prepare_cpus(max_cpus);
 
 #ifndef CONFIG_HOTPLUG_CPU
-	init_cpu_present(&cpu_possible_map);
+	init_cpu_present(cpu_possible_mask);
 #endif
 }
 
diff --git a/arch/sh/kernel/topology.c b/arch/sh/kernel/topology.c
index 4649a6ff0cfe..772caffba22f 100644
--- a/arch/sh/kernel/topology.c
+++ b/arch/sh/kernel/topology.c
@@ -27,7 +27,7 @@ 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;
+	return *cpu_possible_mask;
 }
 
 const struct cpumask *cpu_coregroup_mask(unsigned int cpu)
diff --git a/arch/sh/mm/cache-sh2a.c b/arch/sh/mm/cache-sh2a.c
index ae08cbbfa569..949e2d3138a0 100644
--- a/arch/sh/mm/cache-sh2a.c
+++ b/arch/sh/mm/cache-sh2a.c
@@ -23,6 +23,7 @@
 #define MAX_OCACHE_PAGES	32
 #define MAX_ICACHE_PAGES	32
 
+#ifdef CONFIG_CACHE_WRITEBACK
 static void sh2a_flush_oc_line(unsigned long v, int way)
 {
 	unsigned long addr = (v & 0x000007f0) | (way << 11);
@@ -34,6 +35,7 @@ static void sh2a_flush_oc_line(unsigned long v, int way)
 		__raw_writel(data, CACHE_OC_ADDRESS_ARRAY | addr);
 	}
 }
+#endif
 
 static void sh2a_invalidate_line(unsigned long cache_addr, unsigned long v)
 {
diff --git a/arch/x86/include/asm/i387.h b/arch/x86/include/asm/i387.h
index a850b4d8d14d..247904945d3f 100644
--- a/arch/x86/include/asm/i387.h
+++ b/arch/x86/include/asm/i387.h
@@ -29,10 +29,11 @@ extern unsigned int sig_xstate_size;
 extern void fpu_init(void);
 extern void mxcsr_feature_mask_init(void);
 extern int init_fpu(struct task_struct *child);
-extern void __math_state_restore(struct task_struct *);
 extern void math_state_restore(void);
 extern int dump_fpu(struct pt_regs *, struct user_i387_struct *);
 
+DECLARE_PER_CPU(struct task_struct *, fpu_owner_task);
+
 extern user_regset_active_fn fpregs_active, xfpregs_active;
 extern user_regset_get_fn fpregs_get, xfpregs_get, fpregs_soft_get,
 				xstateregs_get;
@@ -269,6 +270,16 @@ static inline int fpu_restore_checking(struct fpu *fpu)
 
 static inline int restore_fpu_checking(struct task_struct *tsk)
 {
+	/* AMD K7/K8 CPUs don't save/restore FDP/FIP/FOP unless an exception
+	   is pending.  Clear the x87 state here by setting it to fixed
+	   values. "m" is a random variable that should be in L1 */
+	alternative_input(
+		ASM_NOP8 ASM_NOP2,
+		"emms\n\t"	  	/* clear stack tags */
+		"fildl %P[addr]",	/* set F?P to defined value */
+		X86_FEATURE_FXSAVE_LEAK,
+		[addr] "m" (tsk->thread.fpu.has_fpu));
+
 	return fpu_restore_checking(&tsk->thread.fpu);
 }
 
@@ -279,19 +290,21 @@ static inline int restore_fpu_checking(struct task_struct *tsk)
  */
 static inline int __thread_has_fpu(struct task_struct *tsk)
 {
-	return tsk->thread.has_fpu;
+	return tsk->thread.fpu.has_fpu;
 }
 
 /* Must be paired with an 'stts' after! */
 static inline void __thread_clear_has_fpu(struct task_struct *tsk)
 {
-	tsk->thread.has_fpu = 0;
+	tsk->thread.fpu.has_fpu = 0;
+	percpu_write(fpu_owner_task, NULL);
 }
 
 /* Must be paired with a 'clts' before! */
 static inline void __thread_set_has_fpu(struct task_struct *tsk)
 {
-	tsk->thread.has_fpu = 1;
+	tsk->thread.fpu.has_fpu = 1;
+	percpu_write(fpu_owner_task, tsk);
 }
 
 /*
@@ -336,30 +349,36 @@ typedef struct { int preload; } fpu_switch_t;
  * We don't do that yet, so "fpu_lazy_restore()" always returns
  * false, but some day..
  */
-#define fpu_lazy_restore(tsk) (0)
-#define fpu_lazy_state_intact(tsk) do { } while (0)
+static inline int fpu_lazy_restore(struct task_struct *new, unsigned int cpu)
+{
+	return new == percpu_read_stable(fpu_owner_task) &&
+		cpu == new->thread.fpu.last_cpu;
+}
 
-static inline fpu_switch_t switch_fpu_prepare(struct task_struct *old, struct task_struct *new)
+static inline fpu_switch_t switch_fpu_prepare(struct task_struct *old, struct task_struct *new, int cpu)
 {
 	fpu_switch_t fpu;
 
 	fpu.preload = tsk_used_math(new) && new->fpu_counter > 5;
 	if (__thread_has_fpu(old)) {
-		if (__save_init_fpu(old))
-			fpu_lazy_state_intact(old);
-		__thread_clear_has_fpu(old);
-		old->fpu_counter++;
+		if (!__save_init_fpu(old))
+			cpu = ~0;
+		old->thread.fpu.last_cpu = cpu;
+		old->thread.fpu.has_fpu = 0;	/* But leave fpu_owner_task! */
 
 		/* Don't change CR0.TS if we just switch! */
 		if (fpu.preload) {
+			new->fpu_counter++;
 			__thread_set_has_fpu(new);
 			prefetch(new->thread.fpu.state);
 		} else
 			stts();
 	} else {
 		old->fpu_counter = 0;
+		old->thread.fpu.last_cpu = ~0;
 		if (fpu.preload) {
-			if (fpu_lazy_restore(new))
+			new->fpu_counter++;
+			if (fpu_lazy_restore(new, cpu))
 				fpu.preload = 0;
 			else
 				prefetch(new->thread.fpu.state);
@@ -377,8 +396,10 @@ static inline fpu_switch_t switch_fpu_prepare(struct task_struct *old, struct ta
  */
 static inline void switch_fpu_finish(struct task_struct *new, fpu_switch_t fpu)
 {
-	if (fpu.preload)
-		__math_state_restore(new);
+	if (fpu.preload) {
+		if (unlikely(restore_fpu_checking(new)))
+			__thread_fpu_end(new);
+	}
 }
 
 /*
@@ -451,8 +472,10 @@ static inline void kernel_fpu_begin(void)
 		__save_init_fpu(me);
 		__thread_clear_has_fpu(me);
 		/* We do 'stts()' in kernel_fpu_end() */
-	} else
+	} else {
+		percpu_write(fpu_owner_task, NULL);
 		clts();
+	}
 }
 
 static inline void kernel_fpu_end(void)
diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h
index f7c89e231c6c..58545c97d071 100644
--- a/arch/x86/include/asm/processor.h
+++ b/arch/x86/include/asm/processor.h
@@ -374,6 +374,8 @@ union thread_xstate {
 };
 
 struct fpu {
+	unsigned int last_cpu;
+	unsigned int has_fpu;
 	union thread_xstate *state;
 };
 
@@ -454,7 +456,6 @@ struct thread_struct {
 	unsigned long		trap_no;
 	unsigned long		error_code;
 	/* floating point and extended processor state */
-	unsigned long		has_fpu;
 	struct fpu		fpu;
 #ifdef CONFIG_X86_32
 	/* Virtual 86 mode info */
diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index d43cad74f166..c0f7d68d318f 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -1044,6 +1044,9 @@ DEFINE_PER_CPU(char *, irq_stack_ptr) =
 
 DEFINE_PER_CPU(unsigned int, irq_count) = -1;
 
+DEFINE_PER_CPU(struct task_struct *, fpu_owner_task);
+EXPORT_PER_CPU_SYMBOL(fpu_owner_task);
+
 /*
  * Special IST stacks which the CPU switches to when it calls
  * an IST-marked descriptor entry. Up to 7 stacks (hardware
@@ -1111,6 +1114,8 @@ void debug_stack_reset(void)
 
 DEFINE_PER_CPU(struct task_struct *, current_task) = &init_task;
 EXPORT_PER_CPU_SYMBOL(current_task);
+DEFINE_PER_CPU(struct task_struct *, fpu_owner_task);
+EXPORT_PER_CPU_SYMBOL(fpu_owner_task);
 
 #ifdef CONFIG_CC_STACKPROTECTOR
 DEFINE_PER_CPU_ALIGNED(struct stack_canary, stack_canary);
diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c
index 80bfe1ab0031..c08d1ff12b7c 100644
--- a/arch/x86/kernel/process_32.c
+++ b/arch/x86/kernel/process_32.c
@@ -214,6 +214,7 @@ int copy_thread(unsigned long clone_flags, unsigned long sp,
 
 	task_user_gs(p) = get_user_gs(regs);
 
+	p->fpu_counter = 0;
 	p->thread.io_bitmap_ptr = NULL;
 	tsk = current;
 	err = -ENOMEM;
@@ -303,7 +304,7 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
 
 	/* never put a printk in __switch_to... printk() calls wake_up*() indirectly */
 
-	fpu = switch_fpu_prepare(prev_p, next_p);
+	fpu = switch_fpu_prepare(prev_p, next_p, cpu);
 
 	/*
 	 * Reload esp0.
diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c
index 1fd94bc4279d..cfa5c90c01db 100644
--- a/arch/x86/kernel/process_64.c
+++ b/arch/x86/kernel/process_64.c
@@ -286,6 +286,7 @@ int copy_thread(unsigned long clone_flags, unsigned long sp,
 
 	set_tsk_thread_flag(p, TIF_FORK);
 
+	p->fpu_counter = 0;
 	p->thread.io_bitmap_ptr = NULL;
 
 	savesegment(gs, p->thread.gsindex);
@@ -388,7 +389,7 @@ __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
 	unsigned fsindex, gsindex;
 	fpu_switch_t fpu;
 
-	fpu = switch_fpu_prepare(prev_p, next_p);
+	fpu = switch_fpu_prepare(prev_p, next_p, cpu);
 
 	/*
 	 * Reload esp0, LDT and the page table pointer:
diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
index 77da5b475ad2..4bbe04d96744 100644
--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -571,37 +571,6 @@ asmlinkage void __attribute__((weak)) smp_threshold_interrupt(void)
 }
 
 /*
- * This gets called with the process already owning the
- * FPU state, and with CR0.TS cleared. It just needs to
- * restore the FPU register state.
- */
-void __math_state_restore(struct task_struct *tsk)
-{
-	/* We need a safe address that is cheap to find and that is already
-	   in L1. We've just brought in "tsk->thread.has_fpu", so use that */
-#define safe_address (tsk->thread.has_fpu)
-
-	/* AMD K7/K8 CPUs don't save/restore FDP/FIP/FOP unless an exception
-	   is pending.  Clear the x87 state here by setting it to fixed
-	   values. safe_address is a random variable that should be in L1 */
-	alternative_input(
-		ASM_NOP8 ASM_NOP2,
-		"emms\n\t"	  	/* clear stack tags */
-		"fildl %P[addr]",	/* set F?P to defined value */
-		X86_FEATURE_FXSAVE_LEAK,
-		[addr] "m" (safe_address));
-
-	/*
-	 * Paranoid restore. send a SIGSEGV if we fail to restore the state.
-	 */
-	if (unlikely(restore_fpu_checking(tsk))) {
-		__thread_fpu_end(tsk);
-		force_sig(SIGSEGV, tsk);
-		return;
-	}
-}
-
-/*
  * 'math_state_restore()' saves the current math information in the
  * old math state array, and gets the new ones from the current task
  *
@@ -631,7 +600,14 @@ void math_state_restore(void)
 	}
 
 	__thread_fpu_begin(tsk);
-	__math_state_restore(tsk);
+	/*
+	 * Paranoid restore. send a SIGSEGV if we fail to restore the state.
+	 */
+	if (unlikely(restore_fpu_checking(tsk))) {
+		__thread_fpu_end(tsk);
+		force_sig(SIGSEGV, tsk);
+		return;
+	}
 
 	tsk->fpu_counter++;
 }