Merge branch 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus

Pull MIPS updates from Ralf Baechle:
 "This is the main pull request for 4.3 for MIPS.  Here's the summary:

  Three fixes that didn't make 4.2-stable:

   - a -Os build might compile the kernel using the MIPS16 instruction
     set but the R2 optimized inline functions in <uapi/asm/swab.h> are
     implemented using 32-bit wide instructions which is invalid.

   - a build error in pgtable-bits.h for a particular kernel
     configuration.

   - accessing registers of the CM GCR might have been compiled to use
     64 bit accesses but these registers are onl 32 bit wide.

  And also a few new bits:

   - move the ATH79 GPIO driver to drivers/gpio

   - the definition of IRQCHIP_DECLARE has moved to linux/irqchip.h,
     change ATH79 accordingly.

   - fix definition of pgprot_writecombine

   - add an implementation of dma_map_ops.mmap

   - fix alignment of quiet build output for vmlinuz link

   - BCM47xx: Use kmemdup rather than duplicating its implementation

   - Netlogic: Fix 0x0x prefixes of constants.

   - merge Bjorn Helgaas' series to remove most of the weak keywords
     from function declarations.

   - CP0 and CP1 registers are best considered treated as unsigned
     values to avoid large values from becoming negative values.

   - improve support for the MIPS GIC timer.

   - enable common clock framework for Malta and SEAD3.

   - a number of improvments and fixes to dump_tlb().

   - document the MIPS TLB dump functionality in Magic SysRq.

   - Cavium Octeon CN68XX improvments.

   - NetLogic improvments.

   - irq: Use access helper irq_data_get_affinity_mask.

   - handle MSA unaligned accesses.

   - a number of R6-related math-emu fixes.

   - support for I6400.

   - improvments to MSA support.

   - add uprobes support.

   - move from deprecated __initcall to arch_initcall.

   - remove finish_arch_switch().

   - IRQ cleanups by Thomas Gleixner.

   - migrate to new 'set-state' interface.

   - random small cleanups"

* 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus: (148 commits)
  MIPS: UAPI: Fix unrecognized opcode WSBH/DSBH/DSHD when using MIPS16.
  MIPS: Fix alignment of quiet build output for vmlinuz link
  MIPS: math-emu: Remove unused handle_dsemul function declaration
  MIPS: math-emu: Add support for the MIPS R6 MAX{, A} FPU instruction
  MIPS: math-emu: Add support for the MIPS R6 MIN{, A} FPU instruction
  MIPS: math-emu: Add support for the MIPS R6 CLASS FPU instruction
  MIPS: math-emu: Add support for the MIPS R6 RINT FPU instruction
  MIPS: math-emu: Add support for the MIPS R6 MSUBF FPU instruction
  MIPS: math-emu: Add support for the MIPS R6 MADDF FPU instruction
  MIPS: math-emu: Add support for the MIPS R6 SELNEZ FPU instruction
  MIPS: math-emu: Add support for the MIPS R6 SELEQZ FPU instruction
  MIPS: math-emu: Add support for the CMP.condn.fmt R6 instruction
  MIPS: inst.h: Add new MIPS R6 FPU opcodes
  MIPS: Octeon: Fix management port MII address on Kontron S1901
  MIPS: BCM47xx: Use kmemdup rather than duplicating its implementation
  STAGING: Octeon: Use common helpers for determining interface and port
  MIPS: Octeon: Support interfaces 4 and 5
  MIPS: Octeon: Set up 1:1 mapping between CN68XX PKO queues and ports
  MIPS: Octeon: Initialize CN68XX PKO
  STAGING: Octeon: Support CN68XX style WQE
  ...

28 files changed, 1652 insertions, 694 deletions

diff --git a/arch/mips/kernel/Makefile b/arch/mips/kernel/Makefile
index 3156c8d..d982be1 100644
--- a/arch/mips/kernel/Makefile
+++ b/arch/mips/kernel/Makefile
@@ -99,6 +99,7 @@ obj-$(CONFIG_PERF_EVENTS)	+= perf_event.o
 obj-$(CONFIG_HW_PERF_EVENTS)	+= perf_event_mipsxx.o
 
 obj-$(CONFIG_JUMP_LABEL)	+= jump_label.o
+obj-$(CONFIG_UPROBES)		+= uprobes.o
 
 obj-$(CONFIG_MIPS_CM)		+= mips-cm.o
 obj-$(CONFIG_MIPS_CPC)		+= mips-cpc.o
diff --git a/arch/mips/kernel/asm-offsets.c b/arch/mips/kernel/asm-offsets.c
index 072fab1..154e203 100644
--- a/arch/mips/kernel/asm-offsets.c
+++ b/arch/mips/kernel/asm-offsets.c
@@ -128,6 +128,7 @@ void output_thread_defines(void)
 	       thread.cp0_baduaddr);
 	OFFSET(THREAD_ECODE, task_struct, \
 	       thread.error_code);
+	OFFSET(THREAD_TRAPNO, task_struct, thread.trap_nr);
 	BLANK();
 }
 
@@ -245,17 +246,6 @@ void output_sc_defines(void)
 }
 #endif
 
-#ifdef CONFIG_MIPS32_COMPAT
-void output_sc32_defines(void)
-{
-	COMMENT("Linux 32-bit sigcontext offsets.");
-	OFFSET(SC32_FPREGS, sigcontext32, sc_fpregs);
-	OFFSET(SC32_FPC_CSR, sigcontext32, sc_fpc_csr);
-	OFFSET(SC32_FPC_EIR, sigcontext32, sc_fpc_eir);
-	BLANK();
-}
-#endif
-
 void output_signal_defined(void)
 {
 	COMMENT("Linux signal numbers.");
diff --git a/arch/mips/kernel/cevt-bcm1480.c b/arch/mips/kernel/cevt-bcm1480.c
index 7976457..940ac00 100644
--- a/arch/mips/kernel/cevt-bcm1480.c
+++ b/arch/mips/kernel/cevt-bcm1480.c
@@ -40,8 +40,8 @@
  * The general purpose timer ticks at 1MHz independent if
  * the rest of the system
  */
-static void sibyte_set_mode(enum clock_event_mode mode,
-			   struct clock_event_device *evt)
+
+static int sibyte_set_periodic(struct clock_event_device *evt)
 {
 	unsigned int cpu = smp_processor_id();
 	void __iomem *cfg, *init;
@@ -49,24 +49,22 @@ static void sibyte_set_mode(enum clock_event_mode mode,
 	cfg = IOADDR(A_SCD_TIMER_REGISTER(cpu, R_SCD_TIMER_CFG));
 	init = IOADDR(A_SCD_TIMER_REGISTER(cpu, R_SCD_TIMER_INIT));
 
-	switch (mode) {
-	case CLOCK_EVT_MODE_PERIODIC:
-		__raw_writeq(0, cfg);
-		__raw_writeq((V_SCD_TIMER_FREQ / HZ) - 1, init);
-		__raw_writeq(M_SCD_TIMER_ENABLE | M_SCD_TIMER_MODE_CONTINUOUS,
-			     cfg);
-		break;
-
-	case CLOCK_EVT_MODE_ONESHOT:
-		/* Stop the timer until we actually program a shot */
-	case CLOCK_EVT_MODE_SHUTDOWN:
-		__raw_writeq(0, cfg);
-		break;
-
-	case CLOCK_EVT_MODE_UNUSED:	/* shuddup gcc */
-	case CLOCK_EVT_MODE_RESUME:
-		;
-	}
+	__raw_writeq(0, cfg);
+	__raw_writeq((V_SCD_TIMER_FREQ / HZ) - 1, init);
+	__raw_writeq(M_SCD_TIMER_ENABLE | M_SCD_TIMER_MODE_CONTINUOUS, cfg);
+	return 0;
+}
+
+static int sibyte_shutdown(struct clock_event_device *evt)
+{
+	unsigned int cpu = smp_processor_id();
+	void __iomem *cfg;
+
+	cfg = IOADDR(A_SCD_TIMER_REGISTER(cpu, R_SCD_TIMER_CFG));
+
+	/* Stop the timer until we actually program a shot */
+	__raw_writeq(0, cfg);
+	return 0;
 }
 
 static int sibyte_next_event(unsigned long delta, struct clock_event_device *cd)
@@ -91,7 +89,7 @@ static irqreturn_t sibyte_counter_handler(int irq, void *dev_id)
 	void __iomem *cfg;
 	unsigned long tmode;
 
-	if (cd->mode == CLOCK_EVT_MODE_PERIODIC)
+	if (clockevent_state_periodic(cd))
 		tmode = M_SCD_TIMER_ENABLE | M_SCD_TIMER_MODE_CONTINUOUS;
 	else
 		tmode = 0;
@@ -130,7 +128,9 @@ void sb1480_clockevent_init(void)
 	cd->irq			= irq;
 	cd->cpumask		= cpumask_of(cpu);
 	cd->set_next_event	= sibyte_next_event;
-	cd->set_mode		= sibyte_set_mode;
+	cd->set_state_shutdown	= sibyte_shutdown;
+	cd->set_state_periodic	= sibyte_set_periodic;
+	cd->set_state_oneshot	= sibyte_shutdown;
 	clockevents_register_device(cd);
 
 	bcm1480_mask_irq(cpu, irq);
diff --git a/arch/mips/kernel/cevt-ds1287.c b/arch/mips/kernel/cevt-ds1287.c
index ff1f01b..77a5ddf 100644
--- a/arch/mips/kernel/cevt-ds1287.c
+++ b/arch/mips/kernel/cevt-ds1287.c
@@ -59,27 +59,32 @@ static int ds1287_set_next_event(unsigned long delta,
 	return -EINVAL;
 }
 
-static void ds1287_set_mode(enum clock_event_mode mode,
-			    struct clock_event_device *evt)
+static int ds1287_shutdown(struct clock_event_device *evt)
 {
 	u8 val;
 
 	spin_lock(&rtc_lock);
 
 	val = CMOS_READ(RTC_REG_B);
+	val &= ~RTC_PIE;
+	CMOS_WRITE(val, RTC_REG_B);
 
-	switch (mode) {
-	case CLOCK_EVT_MODE_PERIODIC:
-		val |= RTC_PIE;
-		break;
-	default:
-		val &= ~RTC_PIE;
-		break;
-	}
+	spin_unlock(&rtc_lock);
+	return 0;
+}
 
+static int ds1287_set_periodic(struct clock_event_device *evt)
+{
+	u8 val;
+
+	spin_lock(&rtc_lock);
+
+	val = CMOS_READ(RTC_REG_B);
+	val |= RTC_PIE;
 	CMOS_WRITE(val, RTC_REG_B);
 
 	spin_unlock(&rtc_lock);
+	return 0;
 }
 
 static void ds1287_event_handler(struct clock_event_device *dev)
@@ -87,11 +92,13 @@ static void ds1287_event_handler(struct clock_event_device *dev)
 }
 
 static struct clock_event_device ds1287_clockevent = {
-	.name		= "ds1287",
-	.features	= CLOCK_EVT_FEAT_PERIODIC,
-	.set_next_event = ds1287_set_next_event,
-	.set_mode	= ds1287_set_mode,
-	.event_handler	= ds1287_event_handler,
+	.name			= "ds1287",
+	.features		= CLOCK_EVT_FEAT_PERIODIC,
+	.set_next_event		= ds1287_set_next_event,
+	.set_state_shutdown	= ds1287_shutdown,
+	.set_state_periodic	= ds1287_set_periodic,
+	.tick_resume		= ds1287_shutdown,
+	.event_handler		= ds1287_event_handler,
 };
 
 static irqreturn_t ds1287_interrupt(int irq, void *dev_id)
diff --git a/arch/mips/kernel/cevt-gt641xx.c b/arch/mips/kernel/cevt-gt641xx.c
index f069460..6604005 100644
--- a/arch/mips/kernel/cevt-gt641xx.c
+++ b/arch/mips/kernel/cevt-gt641xx.c
@@ -64,8 +64,7 @@ static int gt641xx_timer0_set_next_event(unsigned long delta,
 	return 0;
 }
 
-static void gt641xx_timer0_set_mode(enum clock_event_mode mode,
-				    struct clock_event_device *evt)
+static int gt641xx_timer0_shutdown(struct clock_event_device *evt)
 {
 	u32 ctrl;
 
@@ -73,21 +72,39 @@ static void gt641xx_timer0_set_mode(enum clock_event_mode mode,
 
 	ctrl = GT_READ(GT_TC_CONTROL_OFS);
 	ctrl &= ~(GT_TC_CONTROL_ENTC0_MSK | GT_TC_CONTROL_SELTC0_MSK);
+	GT_WRITE(GT_TC_CONTROL_OFS, ctrl);
+
+	raw_spin_unlock(&gt641xx_timer_lock);
+	return 0;
+}
+
+static int gt641xx_timer0_set_oneshot(struct clock_event_device *evt)
+{
+	u32 ctrl;
+
+	raw_spin_lock(&gt641xx_timer_lock);
+
+	ctrl = GT_READ(GT_TC_CONTROL_OFS);
+	ctrl &= ~GT_TC_CONTROL_SELTC0_MSK;
+	ctrl |= GT_TC_CONTROL_ENTC0_MSK;
+	GT_WRITE(GT_TC_CONTROL_OFS, ctrl);
+
+	raw_spin_unlock(&gt641xx_timer_lock);
+	return 0;
+}
 
-	switch (mode) {
-	case CLOCK_EVT_MODE_PERIODIC:
-		ctrl |= GT_TC_CONTROL_ENTC0_MSK | GT_TC_CONTROL_SELTC0_MSK;
-		break;
-	case CLOCK_EVT_MODE_ONESHOT:
-		ctrl |= GT_TC_CONTROL_ENTC0_MSK;
-		break;
-	default:
-		break;
-	}
+static int gt641xx_timer0_set_periodic(struct clock_event_device *evt)
+{
+	u32 ctrl;
+
+	raw_spin_lock(&gt641xx_timer_lock);
 
+	ctrl = GT_READ(GT_TC_CONTROL_OFS);
+	ctrl |= GT_TC_CONTROL_ENTC0_MSK | GT_TC_CONTROL_SELTC0_MSK;
 	GT_WRITE(GT_TC_CONTROL_OFS, ctrl);
 
 	raw_spin_unlock(&gt641xx_timer_lock);
+	return 0;
 }
 
 static void gt641xx_timer0_event_handler(struct clock_event_device *dev)
@@ -95,12 +112,16 @@ static void gt641xx_timer0_event_handler(struct clock_event_device *dev)
 }
 
 static struct clock_event_device gt641xx_timer0_clockevent = {
-	.name		= "gt641xx-timer0",
-	.features	= CLOCK_EVT_FEAT_PERIODIC | CLOCK_EVT_FEAT_ONESHOT,
-	.irq		= GT641XX_TIMER0_IRQ,
-	.set_next_event = gt641xx_timer0_set_next_event,
-	.set_mode	= gt641xx_timer0_set_mode,
-	.event_handler	= gt641xx_timer0_event_handler,
+	.name			= "gt641xx-timer0",
+	.features		= CLOCK_EVT_FEAT_PERIODIC |
+				  CLOCK_EVT_FEAT_ONESHOT,
+	.irq			= GT641XX_TIMER0_IRQ,
+	.set_next_event		= gt641xx_timer0_set_next_event,
+	.set_state_shutdown	= gt641xx_timer0_shutdown,
+	.set_state_periodic	= gt641xx_timer0_set_periodic,
+	.set_state_oneshot	= gt641xx_timer0_set_oneshot,
+	.tick_resume		= gt641xx_timer0_shutdown,
+	.event_handler		= gt641xx_timer0_event_handler,
 };
 
 static irqreturn_t gt641xx_timer0_interrupt(int irq, void *dev_id)
diff --git a/arch/mips/kernel/cevt-r4k.c b/arch/mips/kernel/cevt-r4k.c
index d70c4d8..8dfe6a6 100644
--- a/arch/mips/kernel/cevt-r4k.c
+++ b/arch/mips/kernel/cevt-r4k.c
@@ -28,12 +28,6 @@ static int mips_next_event(unsigned long delta,
 	return res;
 }
 
-void mips_set_clock_mode(enum clock_event_mode mode,
-				struct clock_event_device *evt)
-{
-	/* Nothing to do ...  */
-}
-
 DEFINE_PER_CPU(struct clock_event_device, mips_clockevent_device);
 int cp0_timer_irq_installed;
 
@@ -174,6 +168,11 @@ int c0_compare_int_usable(void)
 	return 1;
 }
 
+unsigned int __weak get_c0_compare_int(void)
+{
+	return MIPS_CPU_IRQ_BASE + cp0_compare_irq;
+}
+
 int r4k_clockevent_init(void)
 {
 	unsigned int cpu = smp_processor_id();
@@ -189,11 +188,9 @@ int r4k_clockevent_init(void)
 	/*
 	 * With vectored interrupts things are getting platform specific.
 	 * get_c0_compare_int is a hook to allow a platform to return the
-	 * interrupt number of it's liking.
+	 * interrupt number of its liking.
 	 */
-	irq = MIPS_CPU_IRQ_BASE + cp0_compare_irq;
-	if (get_c0_compare_int)
-		irq = get_c0_compare_int();
+	irq = get_c0_compare_int();
 
 	cd = &per_cpu(mips_clockevent_device, cpu);
 
@@ -212,7 +209,6 @@ int r4k_clockevent_init(void)
 	cd->irq			= irq;
 	cd->cpumask		= cpumask_of(cpu);
 	cd->set_next_event	= mips_next_event;
-	cd->set_mode		= mips_set_clock_mode;
 	cd->event_handler	= mips_event_handler;
 
 	clockevents_register_device(cd);
diff --git a/arch/mips/kernel/cevt-sb1250.c b/arch/mips/kernel/cevt-sb1250.c
index 5ea6d6b..3d860ef 100644
--- a/arch/mips/kernel/cevt-sb1250.c
+++ b/arch/mips/kernel/cevt-sb1250.c
@@ -38,8 +38,20 @@
  * The general purpose timer ticks at 1MHz independent if
  * the rest of the system
  */
-static void sibyte_set_mode(enum clock_event_mode mode,
-			   struct clock_event_device *evt)
+
+static int sibyte_shutdown(struct clock_event_device *evt)
+{
+	void __iomem *cfg;
+
+	cfg = IOADDR(A_SCD_TIMER_REGISTER(smp_processor_id(), R_SCD_TIMER_CFG));
+
+	/* Stop the timer until we actually program a shot */
+	__raw_writeq(0, cfg);
+
+	return 0;
+}
+
+static int sibyte_set_periodic(struct clock_event_device *evt)
 {
 	unsigned int cpu = smp_processor_id();
 	void __iomem *cfg, *init;
@@ -47,24 +59,11 @@ static void sibyte_set_mode(enum clock_event_mode mode,
 	cfg = IOADDR(A_SCD_TIMER_REGISTER(cpu, R_SCD_TIMER_CFG));
 	init = IOADDR(A_SCD_TIMER_REGISTER(cpu, R_SCD_TIMER_INIT));
 
-	switch (mode) {
-	case CLOCK_EVT_MODE_PERIODIC:
-		__raw_writeq(0, cfg);
-		__raw_writeq((V_SCD_TIMER_FREQ / HZ) - 1, init);
-		__raw_writeq(M_SCD_TIMER_ENABLE | M_SCD_TIMER_MODE_CONTINUOUS,
-			     cfg);
-		break;
-
-	case CLOCK_EVT_MODE_ONESHOT:
-		/* Stop the timer until we actually program a shot */
-	case CLOCK_EVT_MODE_SHUTDOWN:
-		__raw_writeq(0, cfg);
-		break;
-
-	case CLOCK_EVT_MODE_UNUSED:	/* shuddup gcc */
-	case CLOCK_EVT_MODE_RESUME:
-		;
-	}
+	__raw_writeq(0, cfg);
+	__raw_writeq((V_SCD_TIMER_FREQ / HZ) - 1, init);
+	__raw_writeq(M_SCD_TIMER_ENABLE | M_SCD_TIMER_MODE_CONTINUOUS, cfg);
+
+	return 0;
 }
 
 static int sibyte_next_event(unsigned long delta, struct clock_event_device *cd)
@@ -89,7 +88,7 @@ static irqreturn_t sibyte_counter_handler(int irq, void *dev_id)
 	void __iomem *cfg;
 	unsigned long tmode;
 
-	if (cd->mode == CLOCK_EVT_MODE_PERIODIC)
+	if (clockevent_state_periodic(cd))
 		tmode = M_SCD_TIMER_ENABLE | M_SCD_TIMER_MODE_CONTINUOUS;
 	else
 		tmode = 0;
@@ -129,7 +128,9 @@ void sb1250_clockevent_init(void)
 	cd->irq			= irq;
 	cd->cpumask		= cpumask_of(cpu);
 	cd->set_next_event	= sibyte_next_event;
-	cd->set_mode		= sibyte_set_mode;
+	cd->set_state_shutdown	= sibyte_shutdown;
+	cd->set_state_periodic	= sibyte_set_periodic;
+	cd->set_state_oneshot	= sibyte_shutdown;
 	clockevents_register_device(cd);
 
 	sb1250_mask_irq(cpu, irq);
diff --git a/arch/mips/kernel/cevt-txx9.c b/arch/mips/kernel/cevt-txx9.c
index 7239324..537eefd 100644
--- a/arch/mips/kernel/cevt-txx9.c
+++ b/arch/mips/kernel/cevt-txx9.c
@@ -85,36 +85,54 @@ static void txx9tmr_stop_and_clear(struct txx9_tmr_reg __iomem *tmrptr)
 	__raw_writel(0, &tmrptr->tisr);
 }
 
-static void txx9tmr_set_mode(enum clock_event_mode mode,
-			     struct clock_event_device *evt)
+static int txx9tmr_set_state_periodic(struct clock_event_device *evt)
 {
 	struct txx9_clock_event_device *txx9_cd =
 		container_of(evt, struct txx9_clock_event_device, cd);
 	struct txx9_tmr_reg __iomem *tmrptr = txx9_cd->tmrptr;
 
 	txx9tmr_stop_and_clear(tmrptr);
-	switch (mode) {
-	case CLOCK_EVT_MODE_PERIODIC:
-		__raw_writel(TXx9_TMITMR_TIIE | TXx9_TMITMR_TZCE,
-			     &tmrptr->itmr);
-		/* start timer */
-		__raw_writel(((u64)(NSEC_PER_SEC / HZ) * evt->mult) >>
-			     evt->shift,
-			     &tmrptr->cpra);
-		__raw_writel(TCR_BASE | TXx9_TMTCR_TCE, &tmrptr->tcr);
-		break;
-	case CLOCK_EVT_MODE_SHUTDOWN:
-	case CLOCK_EVT_MODE_UNUSED:
-		__raw_writel(0, &tmrptr->itmr);
-		break;
-	case CLOCK_EVT_MODE_ONESHOT:
-		__raw_writel(TXx9_TMITMR_TIIE, &tmrptr->itmr);
-		break;
-	case CLOCK_EVT_MODE_RESUME:
-		__raw_writel(TIMER_CCD, &tmrptr->ccdr);
-		__raw_writel(0, &tmrptr->itmr);
-		break;
-	}
+
+	__raw_writel(TXx9_TMITMR_TIIE | TXx9_TMITMR_TZCE, &tmrptr->itmr);
+	/* start timer */
+	__raw_writel(((u64)(NSEC_PER_SEC / HZ) * evt->mult) >> evt->shift,
+		     &tmrptr->cpra);
+	__raw_writel(TCR_BASE | TXx9_TMTCR_TCE, &tmrptr->tcr);
+	return 0;
+}
+
+static int txx9tmr_set_state_oneshot(struct clock_event_device *evt)
+{
+	struct txx9_clock_event_device *txx9_cd =
+		container_of(evt, struct txx9_clock_event_device, cd);
+	struct txx9_tmr_reg __iomem *tmrptr = txx9_cd->tmrptr;
+
+	txx9tmr_stop_and_clear(tmrptr);
+	__raw_writel(TXx9_TMITMR_TIIE, &tmrptr->itmr);
+	return 0;
+}
+
+static int txx9tmr_set_state_shutdown(struct clock_event_device *evt)
+{
+	struct txx9_clock_event_device *txx9_cd =
+		container_of(evt, struct txx9_clock_event_device, cd);
+	struct txx9_tmr_reg __iomem *tmrptr = txx9_cd->tmrptr;
+
+	txx9tmr_stop_and_clear(tmrptr);
+	__raw_writel(0, &tmrptr->itmr);
+	return 0;
+}
+
+static int txx9tmr_tick_resume(struct clock_event_device *evt)
+{
+	struct txx9_clock_event_device *txx9_cd =
+		container_of(evt, struct txx9_clock_event_device, cd);
+	struct txx9_tmr_reg __iomem *tmrptr = txx9_cd->tmrptr;
+
+	txx9tmr_stop_and_clear(tmrptr);
+	__raw_writel(TIMER_CCD, &tmrptr->ccdr);
+	__raw_writel(0, &tmrptr->itmr);
+	return 0;
 }
 
 static int txx9tmr_set_next_event(unsigned long delta,
@@ -133,12 +151,15 @@ static int txx9tmr_set_next_event(unsigned long delta,
 
 static struct txx9_clock_event_device txx9_clock_event_device = {
 	.cd = {
-		.name		= "TXx9",
-		.features	= CLOCK_EVT_FEAT_PERIODIC |
-				  CLOCK_EVT_FEAT_ONESHOT,
-		.rating		= 200,
-		.set_mode	= txx9tmr_set_mode,
-		.set_next_event = txx9tmr_set_next_event,
+		.name			= "TXx9",
+		.features		= CLOCK_EVT_FEAT_PERIODIC |
+					  CLOCK_EVT_FEAT_ONESHOT,
+		.rating			= 200,
+		.set_state_shutdown	= txx9tmr_set_state_shutdown,
+		.set_state_periodic	= txx9tmr_set_state_periodic,
+		.set_state_oneshot	= txx9tmr_set_state_oneshot,
+		.tick_resume		= txx9tmr_tick_resume,
+		.set_next_event		= txx9tmr_set_next_event,
 	},
 };
 
diff --git a/arch/mips/kernel/cps-vec.S b/arch/mips/kernel/cps-vec.S
index 1b6ca63..9f71c06 100644
--- a/arch/mips/kernel/cps-vec.S
+++ b/arch/mips/kernel/cps-vec.S
@@ -152,7 +152,7 @@ dcache_done:
 
 	/* Enter the coherent domain */
 	li	t0, 0xff
-	PTR_S	t0, GCR_CL_COHERENCE_OFS(v1)
+	sw	t0, GCR_CL_COHERENCE_OFS(v1)
 	ehb
 
 	/* Jump to kseg0 */
@@ -302,7 +302,7 @@ LEAF(mips_cps_boot_vpes)
 	PTR_L	t0, 0(t0)
 
 	/* Calculate a pointer to this cores struct core_boot_config */
-	PTR_L	t0, GCR_CL_ID_OFS(t0)
+	lw	t0, GCR_CL_ID_OFS(t0)
 	li	t1, COREBOOTCFG_SIZE
 	mul	t0, t0, t1
 	PTR_LA	t1, mips_cps_core_bootcfg
diff --git a/arch/mips/kernel/cpu-probe.c b/arch/mips/kernel/cpu-probe.c
index dbe0792..571a8e6 100644
--- a/arch/mips/kernel/cpu-probe.c
+++ b/arch/mips/kernel/cpu-probe.c
@@ -32,6 +32,9 @@
 #include <asm/spram.h>
 #include <asm/uaccess.h>
 
+/* Hardware capabilities */
+unsigned int elf_hwcap __read_mostly;
+
 /*
  * Get the FPU Implementation/Revision.
  */
@@ -188,7 +191,7 @@ __setup("nohtw", htw_disable);
 static int mips_ftlb_disabled;
 static int mips_has_ftlb_configured;
 
-static void set_ftlb_enable(struct cpuinfo_mips *c, int enable);
+static int set_ftlb_enable(struct cpuinfo_mips *c, int enable);
 
 static int __init ftlb_disable(char *s)
 {
@@ -202,7 +205,10 @@ static int __init ftlb_disable(char *s)
 		return 1;
 
 	/* Disable it in the boot cpu */
-	set_ftlb_enable(&cpu_data[0], 0);
+	if (set_ftlb_enable(&cpu_data[0], 0)) {
+		pr_warn("Can't turn FTLB off\n");
+		return 1;
+	}
 
 	back_to_back_c0_hazard();
 
@@ -364,30 +370,41 @@ static unsigned int calculate_ftlb_probability(struct cpuinfo_mips *c)
 		return 3;
 }
 
-static void set_ftlb_enable(struct cpuinfo_mips *c, int enable)
+static int set_ftlb_enable(struct cpuinfo_mips *c, int enable)
 {
-	unsigned int config6;
+	unsigned int config;
 
 	/* It's implementation dependent how the FTLB can be enabled */
 	switch (c->cputype) {
 	case CPU_PROAPTIV:
 	case CPU_P5600:
 		/* proAptiv & related cores use Config6 to enable the FTLB */
-		config6 = read_c0_config6();
+		config = read_c0_config6();
 		/* Clear the old probability value */
-		config6 &= ~(3 << MIPS_CONF6_FTLBP_SHIFT);
+		config &= ~(3 << MIPS_CONF6_FTLBP_SHIFT);
 		if (enable)
 			/* Enable FTLB */
-			write_c0_config6(config6 |
+			write_c0_config6(config |
 					 (calculate_ftlb_probability(c)
 					  << MIPS_CONF6_FTLBP_SHIFT)
 					 | MIPS_CONF6_FTLBEN);
 		else
 			/* Disable FTLB */
-			write_c0_config6(config6 &  ~MIPS_CONF6_FTLBEN);
-		back_to_back_c0_hazard();
+			write_c0_config6(config &  ~MIPS_CONF6_FTLBEN);
 		break;
+	case CPU_I6400:
+		/* I6400 & related cores use Config7 to configure FTLB */
+		config = read_c0_config7();
+		/* Clear the old probability value */
+		config &= ~(3 << MIPS_CONF7_FTLBP_SHIFT);
+		write_c0_config7(config | (calculate_ftlb_probability(c)
+					   << MIPS_CONF7_FTLBP_SHIFT));
+		break;
+	default:
+		return 1;
 	}
+
+	return 0;
 }
 
 static inline unsigned int decode_config0(struct cpuinfo_mips *c)
@@ -524,6 +541,8 @@ static inline unsigned int decode_config3(struct cpuinfo_mips *c)
 	}
 	if (config3 & MIPS_CONF3_CDMM)
 		c->options |= MIPS_CPU_CDMM;
+	if (config3 & MIPS_CONF3_SP)
+		c->options |= MIPS_CPU_SP;
 
 	return config3 & MIPS_CONF_M;
 }
@@ -540,7 +559,16 @@ static inline unsigned int decode_config4(struct cpuinfo_mips *c)
 	if (cpu_has_tlb) {
 		if (((config4 & MIPS_CONF4_IE) >> 29) == 2)
 			c->options |= MIPS_CPU_TLBINV;
-		mmuextdef = config4 & MIPS_CONF4_MMUEXTDEF;
+		/*
+		 * This is a bit ugly. R6 has dropped that field from
+		 * config4 and the only valid configuration is VTLB+FTLB so
+		 * set a good value for mmuextdef for that case.
+		 */
+		if (cpu_has_mips_r6)
+			mmuextdef = MIPS_CONF4_MMUEXTDEF_VTLBSIZEEXT;
+		else
+			mmuextdef = config4 & MIPS_CONF4_MMUEXTDEF;
+
 		switch (mmuextdef) {
 		case MIPS_CONF4_MMUEXTDEF_MMUSIZEEXT:
 			c->tlbsize += (config4 & MIPS_CONF4_MMUSIZEEXT) * 0x40;
@@ -1121,6 +1149,10 @@ static inline void cpu_probe_mips(struct cpuinfo_mips *c, unsigned int cpu)
 		c->cputype = CPU_P5600;
 		__cpu_name[cpu] = "MIPS P5600";
 		break;
+	case PRID_IMP_I6400:
+		c->cputype = CPU_I6400;
+		__cpu_name[cpu] = "MIPS I6400";
+		break;
 	case PRID_IMP_M5150:
 		c->cputype = CPU_M5150;
 		__cpu_name[cpu] = "MIPS M5150";
@@ -1492,10 +1524,14 @@ void cpu_probe(void)
 	else
 		c->srsets = 1;
 
+	if (cpu_has_mips_r6)
+		elf_hwcap |= HWCAP_MIPS_R6;
+
 	if (cpu_has_msa) {
 		c->msa_id = cpu_get_msa_id();
 		WARN(c->msa_id & MSA_IR_WRPF,
 		     "Vector register partitioning unimplemented!");
+		elf_hwcap |= HWCAP_MIPS_MSA;
 	}
 
 	cpu_probe_vmbits(c);
diff --git a/arch/mips/kernel/idle.c b/arch/mips/kernel/idle.c
index e4f62b7..ab1478d 100644
--- a/arch/mips/kernel/idle.c
+++ b/arch/mips/kernel/idle.c
@@ -196,6 +196,7 @@ void __init check_wait(void)
 	case CPU_INTERAPTIV:
 	case CPU_M5150:
 	case CPU_QEMU_GENERIC:
+	case CPU_I6400:
 		cpu_wait = r4k_wait;
 		if (read_c0_config7() & MIPS_CONF7_WII)
 			cpu_wait = r4k_wait_irqoff;
diff --git a/arch/mips/kernel/mips-cm.c b/arch/mips/kernel/mips-cm.c
index 85bbe9b..b8ceee5 100644
--- a/arch/mips/kernel/mips-cm.c
+++ b/arch/mips/kernel/mips-cm.c
@@ -15,11 +15,131 @@
 
 void __iomem *mips_cm_base;
 void __iomem *mips_cm_l2sync_base;
+int mips_cm_is64;
+
+static char *cm2_tr[8] = {
+	"mem",	"gcr",	"gic",	"mmio",
+	"0x04", "cpc", "0x06", "0x07"
+};
+
+/* CM3 Tag ECC transation type */
+static char *cm3_tr[16] = {
+	[0x0] = "ReqNoData",
+	[0x1] = "0x1",
+	[0x2] = "ReqWData",
+	[0x3] = "0x3",
+	[0x4] = "IReqNoResp",
+	[0x5] = "IReqWResp",
+	[0x6] = "IReqNoRespDat",
+	[0x7] = "IReqWRespDat",
+	[0x8] = "RespNoData",
+	[0x9] = "RespDataFol",
+	[0xa] = "RespWData",
+	[0xb] = "RespDataOnly",
+	[0xc] = "IRespNoData",
+	[0xd] = "IRespDataFol",
+	[0xe] = "IRespWData",
+	[0xf] = "IRespDataOnly"
+};
+
+static char *cm2_cmd[32] = {
+	[0x00] = "0x00",
+	[0x01] = "Legacy Write",
+	[0x02] = "Legacy Read",
+	[0x03] = "0x03",
+	[0x04] = "0x04",
+	[0x05] = "0x05",
+	[0x06] = "0x06",
+	[0x07] = "0x07",
+	[0x08] = "Coherent Read Own",
+	[0x09] = "Coherent Read Share",
+	[0x0a] = "Coherent Read Discard",
+	[0x0b] = "Coherent Ready Share Always",
+	[0x0c] = "Coherent Upgrade",
+	[0x0d] = "Coherent Writeback",
+	[0x0e] = "0x0e",
+	[0x0f] = "0x0f",
+	[0x10] = "Coherent Copyback",
+	[0x11] = "Coherent Copyback Invalidate",
+	[0x12] = "Coherent Invalidate",
+	[0x13] = "Coherent Write Invalidate",
+	[0x14] = "Coherent Completion Sync",
+	[0x15] = "0x15",
+	[0x16] = "0x16",
+	[0x17] = "0x17",
+	[0x18] = "0x18",
+	[0x19] = "0x19",
+	[0x1a] = "0x1a",
+	[0x1b] = "0x1b",
+	[0x1c] = "0x1c",
+	[0x1d] = "0x1d",
+	[0x1e] = "0x1e",
+	[0x1f] = "0x1f"
+};
+
+/* CM3 Tag ECC command type */
+static char *cm3_cmd[16] = {
+	[0x0] = "Legacy Read",
+	[0x1] = "Legacy Write",
+	[0x2] = "Coherent Read Own",
+	[0x3] = "Coherent Read Share",
+	[0x4] = "Coherent Read Discard",
+	[0x5] = "Coherent Evicted",
+	[0x6] = "Coherent Upgrade",
+	[0x7] = "Coherent Upgrade for Store Conditional",
+	[0x8] = "Coherent Writeback",
+	[0x9] = "Coherent Write Invalidate",
+	[0xa] = "0xa",
+	[0xb] = "0xb",
+	[0xc] = "0xc",
+	[0xd] = "0xd",
+	[0xe] = "0xe",
+	[0xf] = "0xf"
+};
+
+/* CM3 Tag ECC command group */
+static char *cm3_cmd_group[8] = {
+	[0x0] = "Normal",
+	[0x1] = "Registers",
+	[0x2] = "TLB",
+	[0x3] = "0x3",
+	[0x4] = "L1I",
+	[0x5] = "L1D",
+	[0x6] = "L3",
+	[0x7] = "L2"
+};
+
+static char *cm2_core[8] = {
+	"Invalid/OK",	"Invalid/Data",
+	"Shared/OK",	"Shared/Data",
+	"Modified/OK",	"Modified/Data",
+	"Exclusive/OK", "Exclusive/Data"
+};
+
+static char *cm2_causes[32] = {
+	"None", "GC_WR_ERR", "GC_RD_ERR", "COH_WR_ERR",
+	"COH_RD_ERR", "MMIO_WR_ERR", "MMIO_RD_ERR", "0x07",
+	"0x08", "0x09", "0x0a", "0x0b",
+	"0x0c", "0x0d", "0x0e", "0x0f",
+	"0x10", "0x11", "0x12", "0x13",
+	"0x14", "0x15", "0x16", "INTVN_WR_ERR",
+	"INTVN_RD_ERR", "0x19", "0x1a", "0x1b",
+	"0x1c", "0x1d", "0x1e", "0x1f"
+};
+
+static char *cm3_causes[32] = {
+	"0x0", "MP_CORRECTABLE_ECC_ERR", "MP_REQUEST_DECODE_ERR",
+	"MP_UNCORRECTABLE_ECC_ERR", "MP_PARITY_ERR", "MP_COHERENCE_ERR",
+	"CMBIU_REQUEST_DECODE_ERR", "CMBIU_PARITY_ERR", "CMBIU_AXI_RESP_ERR",
+	"0x9", "RBI_BUS_ERR", "0xb", "0xc", "0xd", "0xe", "0xf", "0x10",
+	"0x11", "0x12", "0x13", "0x14", "0x15", "0x16", "0x17", "0x18",
+	"0x19", "0x1a", "0x1b", "0x1c", "0x1d", "0x1e", "0x1f"
+};
 
 phys_addr_t __mips_cm_phys_base(void)
 {
 	u32 config3 = read_c0_config3();
-	u32 cmgcr;
+	unsigned long cmgcr;
 
 	/* Check the CMGCRBase register is implemented */
 	if (!(config3 & MIPS_CONF3_CMGCR))
@@ -81,6 +201,13 @@ int mips_cm_probe(void)
 	phys_addr_t addr;
 	u32 base_reg;
 
+	/*
+	 * No need to probe again if we have already been
+	 * here before.
+	 */
+	if (mips_cm_base)
+		return 0;
+
 	addr = mips_cm_phys_base();
 	BUG_ON((addr & CM_GCR_BASE_GCRBASE_MSK) != addr);
 	if (!addr)
@@ -117,5 +244,133 @@ int mips_cm_probe(void)
 	/* probe for an L2-only sync region */
 	mips_cm_probe_l2sync();
 
+	/* determine register width for this CM */
+	mips_cm_is64 = config_enabled(CONFIG_64BIT) && (mips_cm_revision() >= CM_REV_CM3);
+
 	return 0;
 }
+
+void mips_cm_error_report(void)
+{
+	unsigned long revision = mips_cm_revision();
+	/*
+	 * CM3 has a 64-bit Error cause register with 0:57 containing the error
+	 * info and 63:58 the error type. For old CMs, everything is contained
+	 * in a single 32-bit register (0:26 and 31:27 respectively). Even
+	 * though the cm_error is u64, we will simply ignore the upper word
+	 * for CM2.
+	 */
+	u64 cm_error = read_gcr_error_cause();
+	int cm_error_cause_sft = CM_GCR_ERROR_CAUSE_ERRTYPE_SHF +
+				 ((revision >= CM_REV_CM3) ? 31 : 0);
+	unsigned long cm_addr = read_gcr_error_addr();
+	unsigned long cm_other = read_gcr_error_mult();
+	int ocause, cause;
+	char buf[256];
+
+	if (!mips_cm_present())
+		return;
+
+	cause = cm_error >> cm_error_cause_sft;
+
+	if (!cause)
+		/* All good */
+		return;
+
+	ocause = cm_other >> CM_GCR_ERROR_MULT_ERR2ND_SHF;
+	if (revision < CM_REV_CM3) { /* CM2 */
+		if (cause < 16) {
+			unsigned long cca_bits = (cm_error >> 15) & 7;
+			unsigned long tr_bits = (cm_error >> 12) & 7;
+			unsigned long cmd_bits = (cm_error >> 7) & 0x1f;
+			unsigned long stag_bits = (cm_error >> 3) & 15;
+			unsigned long sport_bits = (cm_error >> 0) & 7;
+
+			snprintf(buf, sizeof(buf),
+				 "CCA=%lu TR=%s MCmd=%s STag=%lu "
+				 "SPort=%lu\n", cca_bits, cm2_tr[tr_bits],
+				 cm2_cmd[cmd_bits], stag_bits, sport_bits);
+		} else {
+			/* glob state & sresp together */
+			unsigned long c3_bits = (cm_error >> 18) & 7;
+			unsigned long c2_bits = (cm_error >> 15) & 7;
+			unsigned long c1_bits = (cm_error >> 12) & 7;
+			unsigned long c0_bits = (cm_error >> 9) & 7;
+			unsigned long sc_bit = (cm_error >> 8) & 1;
+			unsigned long cmd_bits = (cm_error >> 3) & 0x1f;
+			unsigned long sport_bits = (cm_error >> 0) & 7;
+
+			snprintf(buf, sizeof(buf),
+				 "C3=%s C2=%s C1=%s C0=%s SC=%s "
+				 "MCmd=%s SPort=%lu\n",
+				 cm2_core[c3_bits], cm2_core[c2_bits],
+				 cm2_core[c1_bits], cm2_core[c0_bits],
+				 sc_bit ? "True" : "False",
+				 cm2_cmd[cmd_bits], sport_bits);
+		}
+			pr_err("CM_ERROR=%08llx %s <%s>\n", cm_error,
+			       cm2_causes[cause], buf);
+		pr_err("CM_ADDR =%08lx\n", cm_addr);
+		pr_err("CM_OTHER=%08lx %s\n", cm_other, cm2_causes[ocause]);
+	} else { /* CM3 */
+	/* Used by cause == {1,2,3} */
+		unsigned long core_id_bits = (cm_error >> 22) & 0xf;
+		unsigned long vp_id_bits = (cm_error >> 18) & 0xf;
+		unsigned long cmd_bits = (cm_error >> 14) & 0xf;
+		unsigned long cmd_group_bits = (cm_error >> 11) & 0xf;
+		unsigned long cm3_cca_bits = (cm_error >> 8) & 7;
+		unsigned long mcp_bits = (cm_error >> 5) & 0xf;
+		unsigned long cm3_tr_bits = (cm_error >> 1) & 0xf;
+		unsigned long sched_bit = cm_error & 0x1;
+
+		if (cause == 1 || cause == 3) { /* Tag ECC */
+			unsigned long tag_ecc = (cm_error >> 57) & 0x1;
+			unsigned long tag_way_bits = (cm_error >> 29) & 0xffff;
+			unsigned long dword_bits = (cm_error >> 49) & 0xff;
+			unsigned long data_way_bits = (cm_error >> 45) & 0xf;
+			unsigned long data_sets_bits = (cm_error >> 29) & 0xfff;
+			unsigned long bank_bit = (cm_error >> 28) & 0x1;
+			snprintf(buf, sizeof(buf),
+				 "%s ECC Error: Way=%lu (DWORD=%lu, Sets=%lu)"
+				 "Bank=%lu CoreID=%lu VPID=%lu Command=%s"
+				 "Command Group=%s CCA=%lu MCP=%d"
+				 "Transaction type=%s Scheduler=%lu\n",
+				 tag_ecc ? "TAG" : "DATA",
+				 tag_ecc ? (unsigned long)ffs(tag_way_bits) - 1 :
+				 data_way_bits, bank_bit, dword_bits,
+				 data_sets_bits,
+				 core_id_bits, vp_id_bits,
+				 cm3_cmd[cmd_bits],
+				 cm3_cmd_group[cmd_group_bits],
+				 cm3_cca_bits, 1 << mcp_bits,
+				 cm3_tr[cm3_tr_bits], sched_bit);
+		} else if (cause == 2) {
+			unsigned long data_error_type = (cm_error >> 41) & 0xfff;
+			unsigned long data_decode_cmd = (cm_error >> 37) & 0xf;
+			unsigned long data_decode_group = (cm_error >> 34) & 0x7;
+			unsigned long data_decode_destination_id = (cm_error >> 28) & 0x3f;
+
+			snprintf(buf, sizeof(buf),
+				 "Decode Request Error: Type=%lu, Command=%lu"
+				 "Command Group=%lu Destination ID=%lu"
+				 "CoreID=%lu VPID=%lu Command=%s"
+				 "Command Group=%s CCA=%lu MCP=%d"
+				 "Transaction type=%s Scheduler=%lu\n",
+				 data_error_type, data_decode_cmd,
+				 data_decode_group, data_decode_destination_id,
+				 core_id_bits, vp_id_bits,
+				 cm3_cmd[cmd_bits],
+				 cm3_cmd_group[cmd_group_bits],
+				 cm3_cca_bits, 1 << mcp_bits,
+				 cm3_tr[cm3_tr_bits], sched_bit);
+		}
+
+		pr_err("CM_ERROR=%llx %s <%s>\n", cm_error,
+		       cm3_causes[cause], buf);
+		pr_err("CM_ADDR =%lx\n", cm_addr);
+		pr_err("CM_OTHER=%lx %s\n", cm_other, cm3_causes[ocause]);
+	}
+
+	/* reprime cause register */
+	write_gcr_error_cause(0);
+}
diff --git a/arch/mips/kernel/mips-cpc.c b/arch/mips/kernel/mips-cpc.c
index 1196450..8af4d62 100644
--- a/arch/mips/kernel/mips-cpc.c
+++ b/arch/mips/kernel/mips-cpc.c
@@ -21,9 +21,16 @@ static DEFINE_PER_CPU_ALIGNED(spinlock_t, cpc_core_lock);
 
 static DEFINE_PER_CPU_ALIGNED(unsigned long, cpc_core_lock_flags);
 
-phys_addr_t __weak mips_cpc_phys_base(void)
+/**
+ * mips_cpc_phys_base - retrieve the physical base address of the CPC
+ *
+ * This function returns the physical base address of the Cluster Power
+ * Controller memory mapped registers, or 0 if no Cluster Power Controller
+ * is present.
+ */
+static phys_addr_t mips_cpc_phys_base(void)
 {
-	u32 cpc_base;
+	unsigned long cpc_base;
 
 	if (!mips_cm_present())
 		return 0;
diff --git a/arch/mips/kernel/perf_event_mipsxx.c b/arch/mips/kernel/perf_event_mipsxx.c
index cc1b6fa..d7b8dd4 100644
--- a/arch/mips/kernel/perf_event_mipsxx.c
+++ b/arch/mips/kernel/perf_event_mipsxx.c
@@ -1556,6 +1556,7 @@ static const struct mips_perf_event *mipsxx_pmu_map_raw_event(u64 config)
 #endif
 		break;
 	case CPU_P5600:
+	case CPU_I6400:
 		/* 8-bit event numbers */
 		raw_id = config & 0x1ff;
 		base_id = raw_id & 0xff;
@@ -1717,6 +1718,11 @@ init_hw_perf_events(void)
 		mipspmu.general_event_map = &mipsxxcore_event_map2;
 		mipspmu.cache_event_map = &mipsxxcore_cache_map2;
 		break;
+	case CPU_I6400:
+		mipspmu.name = "mips/I6400";
+		mipspmu.general_event_map = &mipsxxcore_event_map2;
+		mipspmu.cache_event_map = &mipsxxcore_cache_map2;
+		break;
 	case CPU_1004K:
 		mipspmu.name = "mips/1004K";
 		mipspmu.general_event_map = &mipsxxcore_event_map;
diff --git a/arch/mips/kernel/pm-cps.c b/arch/mips/kernel/pm-cps.c
index 0614717..f63a289 100644
--- a/arch/mips/kernel/pm-cps.c
+++ b/arch/mips/kernel/pm-cps.c
@@ -267,6 +267,7 @@ static int __init cps_gen_flush_fsb(u32 **pp, struct uasm_label **pl,
 
 	/* CPUs which do not require the workaround */
 	case CPU_P5600:
+	case CPU_I6400:
 		return 0;
 
 	default:
@@ -671,6 +672,7 @@ static int __init cps_pm_init(void)
 	case CPU_PROAPTIV:
 	case CPU_M5150:
 	case CPU_P5600:
+	case CPU_I6400:
 		stype_intervention = 0x2;
 		stype_memory = 0x3;
 		stype_ordering = 0x10;
diff --git a/arch/mips/kernel/ptrace.c b/arch/mips/kernel/ptrace.c
index e933a30..4f0ac78 100644
--- a/arch/mips/kernel/ptrace.c
+++ b/arch/mips/kernel/ptrace.c
@@ -25,6 +25,7 @@
 #include <linux/regset.h>
 #include <linux/smp.h>
 #include <linux/security.h>
+#include <linux/stddef.h>
 #include <linux/tracehook.h>
 #include <linux/audit.h>
 #include <linux/seccomp.h>
@@ -490,6 +491,93 @@ enum mips_regset {
 	REGSET_FPR,
 };
 
+struct pt_regs_offset {
+	const char *name;
+	int offset;
+};
+
+#define REG_OFFSET_NAME(reg, r) {					\
+	.name = #reg,							\
+	.offset = offsetof(struct pt_regs, r)				\
+}
+
+#define REG_OFFSET_END {						\
+	.name = NULL,							\
+	.offset = 0							\
+}
+
+static const struct pt_regs_offset regoffset_table[] = {
+	REG_OFFSET_NAME(r0, regs[0]),
+	REG_OFFSET_NAME(r1, regs[1]),
+	REG_OFFSET_NAME(r2, regs[2]),
+	REG_OFFSET_NAME(r3, regs[3]),
+	REG_OFFSET_NAME(r4, regs[4]),
+	REG_OFFSET_NAME(r5, regs[5]),
+	REG_OFFSET_NAME(r6, regs[6]),
+	REG_OFFSET_NAME(r7, regs[7]),
+	REG_OFFSET_NAME(r8, regs[8]),
+	REG_OFFSET_NAME(r9, regs[9]),
+	REG_OFFSET_NAME(r10, regs[10]),
+	REG_OFFSET_NAME(r11, regs[11]),
+	REG_OFFSET_NAME(r12, regs[12]),
+	REG_OFFSET_NAME(r13, regs[13]),
+	REG_OFFSET_NAME(r14, regs[14]),
+	REG_OFFSET_NAME(r15, regs[15]),
+	REG_OFFSET_NAME(r16, regs[16]),
+	REG_OFFSET_NAME(r17, regs[17]),
+	REG_OFFSET_NAME(r18, regs[18]),
+	REG_OFFSET_NAME(r19, regs[19]),
+	REG_OFFSET_NAME(r20, regs[20]),
+	REG_OFFSET_NAME(r21, regs[21]),
+	REG_OFFSET_NAME(r22, regs[22]),
+	REG_OFFSET_NAME(r23, regs[23]),
+	REG_OFFSET_NAME(r24, regs[24]),
+	REG_OFFSET_NAME(r25, regs[25]),
+	REG_OFFSET_NAME(r26, regs[26]),
+	REG_OFFSET_NAME(r27, regs[27]),
+	REG_OFFSET_NAME(r28, regs[28]),
+	REG_OFFSET_NAME(r29, regs[29]),
+	REG_OFFSET_NAME(r30, regs[30]),
+	REG_OFFSET_NAME(r31, regs[31]),
+	REG_OFFSET_NAME(c0_status, cp0_status),
+	REG_OFFSET_NAME(hi, hi),
+	REG_OFFSET_NAME(lo, lo),
+#ifdef CONFIG_CPU_HAS_SMARTMIPS
+	REG_OFFSET_NAME(acx, acx),
+#endif
+	REG_OFFSET_NAME(c0_badvaddr, cp0_badvaddr),
+	REG_OFFSET_NAME(c0_cause, cp0_cause),
+	REG_OFFSET_NAME(c0_epc, cp0_epc),
+#ifdef CONFIG_MIPS_MT_SMTC
+	REG_OFFSET_NAME(c0_tcstatus, cp0_tcstatus),
+#endif
+#ifdef CONFIG_CPU_CAVIUM_OCTEON
+	REG_OFFSET_NAME(mpl0, mpl[0]),
+	REG_OFFSET_NAME(mpl1, mpl[1]),
+	REG_OFFSET_NAME(mpl2, mpl[2]),
+	REG_OFFSET_NAME(mtp0, mtp[0]),
+	REG_OFFSET_NAME(mtp1, mtp[1]),
+	REG_OFFSET_NAME(mtp2, mtp[2]),
+#endif
+	REG_OFFSET_END,
+};
+
+/**
+ * regs_query_register_offset() - query register offset from its name
+ * @name:       the name of a register
+ *
+ * regs_query_register_offset() returns the offset of a register in struct
+ * pt_regs from its name. If the name is invalid, this returns -EINVAL;
+ */
+int regs_query_register_offset(const char *name)
+{
+        const struct pt_regs_offset *roff;
+        for (roff = regoffset_table; roff->name != NULL; roff++)
+                if (!strcmp(roff->name, name))
+                        return roff->offset;
+        return -EINVAL;
+}
+
 #if defined(CONFIG_32BIT) || defined(CONFIG_MIPS32_O32)
 
 static const struct user_regset mips_regsets[] = {
diff --git a/arch/mips/kernel/r4k_fpu.S b/arch/mips/kernel/r4k_fpu.S
index 1d88af2..f09546e 100644
--- a/arch/mips/kernel/r4k_fpu.S
+++ b/arch/mips/kernel/r4k_fpu.S
@@ -13,6 +13,7 @@
  * Copyright (C) 1999, 2001 Silicon Graphics, Inc.
  */
 #include <asm/asm.h>
+#include <asm/asmmacro.h>
 #include <asm/errno.h>
 #include <asm/fpregdef.h>
 #include <asm/mipsregs.h>
@@ -35,6 +36,14 @@
 
 	.set	noreorder
 
+/**
+ * _save_fp_context() - save FP context from the FPU
+ * @a0 - pointer to fpregs field of sigcontext
+ * @a1 - pointer to fpc_csr field of sigcontext
+ *
+ * Save FP context, including the 32 FP data registers and the FP
+ * control & status register, from the FPU to signal context.
+ */
 LEAF(_save_fp_context)
 	.set	push
 	SET_HARDFLOAT
@@ -54,117 +63,60 @@ LEAF(_save_fp_context)
 	 nop
 #endif
 	/* Store the 16 odd double precision registers */
-	EX	sdc1 $f1, SC_FPREGS+8(a0)
-	EX	sdc1 $f3, SC_FPREGS+24(a0)
-	EX	sdc1 $f5, SC_FPREGS+40(a0)
-	EX	sdc1 $f7, SC_FPREGS+56(a0)
-	EX	sdc1 $f9, SC_FPREGS+72(a0)
-	EX	sdc1 $f11, SC_FPREGS+88(a0)
-	EX	sdc1 $f13, SC_FPREGS+104(a0)
-	EX	sdc1 $f15, SC_FPREGS+120(a0)
-	EX	sdc1 $f17, SC_FPREGS+136(a0)
-	EX	sdc1 $f19, SC_FPREGS+152(a0)
-	EX	sdc1 $f21, SC_FPREGS+168(a0)
-	EX	sdc1 $f23, SC_FPREGS+184(a0)
-	EX	sdc1 $f25, SC_FPREGS+200(a0)
-	EX	sdc1 $f27, SC_FPREGS+216(a0)
-	EX	sdc1 $f29, SC_FPREGS+232(a0)
-	EX	sdc1 $f31, SC_FPREGS+248(a0)
+	EX	sdc1 $f1, 8(a0)
+	EX	sdc1 $f3, 24(a0)
+	EX	sdc1 $f5, 40(a0)
+	EX	sdc1 $f7, 56(a0)
+	EX	sdc1 $f9, 72(a0)
+	EX	sdc1 $f11, 88(a0)
+	EX	sdc1 $f13, 104(a0)
+	EX	sdc1 $f15, 120(a0)
+	EX	sdc1 $f17, 136(a0)
+	EX	sdc1 $f19, 152(a0)
+	EX	sdc1 $f21, 168(a0)
+	EX	sdc1 $f23, 184(a0)
+	EX	sdc1 $f25, 200(a0)
+	EX	sdc1 $f27, 216(a0)
+	EX	sdc1 $f29, 232(a0)
+	EX	sdc1 $f31, 248(a0)
 1:	.set	pop
 #endif
 
 	.set push
 	SET_HARDFLOAT
 	/* Store the 16 even double precision registers */
-	EX	sdc1 $f0, SC_FPREGS+0(a0)
-	EX	sdc1 $f2, SC_FPREGS+16(a0)
-	EX	sdc1 $f4, SC_FPREGS+32(a0)
-	EX	sdc1 $f6, SC_FPREGS+48(a0)
-	EX	sdc1 $f8, SC_FPREGS+64(a0)
-	EX	sdc1 $f10, SC_FPREGS+80(a0)
-	EX	sdc1 $f12, SC_FPREGS+96(a0)
-	EX	sdc1 $f14, SC_FPREGS+112(a0)
-	EX	sdc1 $f16, SC_FPREGS+128(a0)
-	EX	sdc1 $f18, SC_FPREGS+144(a0)
-	EX	sdc1 $f20, SC_FPREGS+160(a0)
-	EX	sdc1 $f22, SC_FPREGS+176(a0)
-	EX	sdc1 $f24, SC_FPREGS+192(a0)
-	EX	sdc1 $f26, SC_FPREGS+208(a0)
-	EX	sdc1 $f28, SC_FPREGS+224(a0)
-	EX	sdc1 $f30, SC_FPREGS+240(a0)
-	EX	sw t1, SC_FPC_CSR(a0)
+	EX	sdc1 $f0, 0(a0)
+	EX	sdc1 $f2, 16(a0)
+	EX	sdc1 $f4, 32(a0)
+	EX	sdc1 $f6, 48(a0)
+	EX	sdc1 $f8, 64(a0)
+	EX	sdc1 $f10, 80(a0)
+	EX	sdc1 $f12, 96(a0)
+	EX	sdc1 $f14, 112(a0)
+	EX	sdc1 $f16, 128(a0)
+	EX	sdc1 $f18, 144(a0)
+	EX	sdc1 $f20, 160(a0)
+	EX	sdc1 $f22, 176(a0)
+	EX	sdc1 $f24, 192(a0)
+	EX	sdc1 $f26, 208(a0)
+	EX	sdc1 $f28, 224(a0)
+	EX	sdc1 $f30, 240(a0)
+	EX	sw t1, 0(a1)
 	jr	ra
 	 li	v0, 0					# success
 	.set pop
 	END(_save_fp_context)
 
-#ifdef CONFIG_MIPS32_COMPAT
-	/* Save 32-bit process floating point context */
-LEAF(_save_fp_context32)
-	.set push
-	.set MIPS_ISA_ARCH_LEVEL_RAW
-	SET_HARDFLOAT
-	cfc1	t1, fcr31
-
-#ifndef CONFIG_CPU_MIPS64_R6
-	mfc0	t0, CP0_STATUS
-	sll	t0, t0, 5
-	bgez	t0, 1f			# skip storing odd if FR=0
-	 nop
-#endif
-
-	/* Store the 16 odd double precision registers */
-	EX      sdc1 $f1, SC32_FPREGS+8(a0)
-	EX      sdc1 $f3, SC32_FPREGS+24(a0)
-	EX      sdc1 $f5, SC32_FPREGS+40(a0)
-	EX      sdc1 $f7, SC32_FPREGS+56(a0)
-	EX      sdc1 $f9, SC32_FPREGS+72(a0)
-	EX      sdc1 $f11, SC32_FPREGS+88(a0)
-	EX      sdc1 $f13, SC32_FPREGS+104(a0)
-	EX      sdc1 $f15, SC32_FPREGS+120(a0)
-	EX      sdc1 $f17, SC32_FPREGS+136(a0)
-	EX      sdc1 $f19, SC32_FPREGS+152(a0)
-	EX      sdc1 $f21, SC32_FPREGS+168(a0)
-	EX      sdc1 $f23, SC32_FPREGS+184(a0)
-	EX      sdc1 $f25, SC32_FPREGS+200(a0)
-	EX      sdc1 $f27, SC32_FPREGS+216(a0)
-	EX      sdc1 $f29, SC32_FPREGS+232(a0)
-	EX      sdc1 $f31, SC32_FPREGS+248(a0)
-
-	/* Store the 16 even double precision registers */
-1:	EX	sdc1 $f0, SC32_FPREGS+0(a0)
-	EX	sdc1 $f2, SC32_FPREGS+16(a0)
-	EX	sdc1 $f4, SC32_FPREGS+32(a0)
-	EX	sdc1 $f6, SC32_FPREGS+48(a0)
-	EX	sdc1 $f8, SC32_FPREGS+64(a0)
-	EX	sdc1 $f10, SC32_FPREGS+80(a0)
-	EX	sdc1 $f12, SC32_FPREGS+96(a0)
-	EX	sdc1 $f14, SC32_FPREGS+112(a0)
-	EX	sdc1 $f16, SC32_FPREGS+128(a0)
-	EX	sdc1 $f18, SC32_FPREGS+144(a0)
-	EX	sdc1 $f20, SC32_FPREGS+160(a0)
-	EX	sdc1 $f22, SC32_FPREGS+176(a0)
-	EX	sdc1 $f24, SC32_FPREGS+192(a0)
-	EX	sdc1 $f26, SC32_FPREGS+208(a0)
-	EX	sdc1 $f28, SC32_FPREGS+224(a0)
-	EX	sdc1 $f30, SC32_FPREGS+240(a0)
-	EX	sw t1, SC32_FPC_CSR(a0)
-	cfc1	t0, $0				# implementation/version
-	EX	sw t0, SC32_FPC_EIR(a0)
-	.set pop
-
-	jr	ra
-	 li	v0, 0					# success
-	END(_save_fp_context32)
-#endif
-
-/*
- * Restore FPU state:
- *  - fp gp registers
- *  - cp1 status/control register
+/**
+ * _restore_fp_context() - restore FP context to the FPU
+ * @a0 - pointer to fpregs field of sigcontext
+ * @a1 - pointer to fpc_csr field of sigcontext
+ *
+ * Restore FP context, including the 32 FP data registers and the FP
+ * control & status register, from signal context to the FPU.
  */
 LEAF(_restore_fp_context)
-	EX	lw t1, SC_FPC_CSR(a0)
+	EX	lw t1, 0(a1)
 
 #if defined(CONFIG_64BIT) || defined(CONFIG_CPU_MIPS32_R2)  || \
 		defined(CONFIG_CPU_MIPS32_R6)
@@ -178,101 +130,231 @@ LEAF(_restore_fp_context)
 	bgez	t0, 1f			# skip loading odd if FR=0
 	 nop
 #endif
-	EX	ldc1 $f1, SC_FPREGS+8(a0)
-	EX	ldc1 $f3, SC_FPREGS+24(a0)
-	EX	ldc1 $f5, SC_FPREGS+40(a0)
-	EX	ldc1 $f7, SC_FPREGS+56(a0)
-	EX	ldc1 $f9, SC_FPREGS+72(a0)
-	EX	ldc1 $f11, SC_FPREGS+88(a0)
-	EX	ldc1 $f13, SC_FPREGS+104(a0)
-	EX	ldc1 $f15, SC_FPREGS+120(a0)
-	EX	ldc1 $f17, SC_FPREGS+136(a0)
-	EX	ldc1 $f19, SC_FPREGS+152(a0)
-	EX	ldc1 $f21, SC_FPREGS+168(a0)
-	EX	ldc1 $f23, SC_FPREGS+184(a0)
-	EX	ldc1 $f25, SC_FPREGS+200(a0)
-	EX	ldc1 $f27, SC_FPREGS+216(a0)
-	EX	ldc1 $f29, SC_FPREGS+232(a0)
-	EX	ldc1 $f31, SC_FPREGS+248(a0)
+	EX	ldc1 $f1, 8(a0)
+	EX	ldc1 $f3, 24(a0)
+	EX	ldc1 $f5, 40(a0)
+	EX	ldc1 $f7, 56(a0)
+	EX	ldc1 $f9, 72(a0)
+	EX	ldc1 $f11, 88(a0)
+	EX	ldc1 $f13, 104(a0)
+	EX	ldc1 $f15, 120(a0)
+	EX	ldc1 $f17, 136(a0)
+	EX	ldc1 $f19, 152(a0)
+	EX	ldc1 $f21, 168(a0)
+	EX	ldc1 $f23, 184(a0)
+	EX	ldc1 $f25, 200(a0)
+	EX	ldc1 $f27, 216(a0)
+	EX	ldc1 $f29, 232(a0)
+	EX	ldc1 $f31, 248(a0)
 1:	.set pop
 #endif
 	.set push
 	SET_HARDFLOAT
-	EX	ldc1 $f0, SC_FPREGS+0(a0)
-	EX	ldc1 $f2, SC_FPREGS+16(a0)
-	EX	ldc1 $f4, SC_FPREGS+32(a0)
-	EX	ldc1 $f6, SC_FPREGS+48(a0)
-	EX	ldc1 $f8, SC_FPREGS+64(a0)
-	EX	ldc1 $f10, SC_FPREGS+80(a0)
-	EX	ldc1 $f12, SC_FPREGS+96(a0)
-	EX	ldc1 $f14, SC_FPREGS+112(a0)
-	EX	ldc1 $f16, SC_FPREGS+128(a0)
-	EX	ldc1 $f18, SC_FPREGS+144(a0)
-	EX	ldc1 $f20, SC_FPREGS+160(a0)
-	EX	ldc1 $f22, SC_FPREGS+176(a0)
-	EX	ldc1 $f24, SC_FPREGS+192(a0)
-	EX	ldc1 $f26, SC_FPREGS+208(a0)
-	EX	ldc1 $f28, SC_FPREGS+224(a0)
-	EX	ldc1 $f30, SC_FPREGS+240(a0)
+	EX	ldc1 $f0, 0(a0)
+	EX	ldc1 $f2, 16(a0)
+	EX	ldc1 $f4, 32(a0)
+	EX	ldc1 $f6, 48(a0)
+	EX	ldc1 $f8, 64(a0)
+	EX	ldc1 $f10, 80(a0)
+	EX	ldc1 $f12, 96(a0)
+	EX	ldc1 $f14, 112(a0)
+	EX	ldc1 $f16, 128(a0)
+	EX	ldc1 $f18, 144(a0)
+	EX	ldc1 $f20, 160(a0)
+	EX	ldc1 $f22, 176(a0)
+	EX	ldc1 $f24, 192(a0)
+	EX	ldc1 $f26, 208(a0)
+	EX	ldc1 $f28, 224(a0)
+	EX	ldc1 $f30, 240(a0)
 	ctc1	t1, fcr31
 	.set pop
 	jr	ra
 	 li	v0, 0					# success
 	END(_restore_fp_context)
 
-#ifdef CONFIG_MIPS32_COMPAT
-LEAF(_restore_fp_context32)
-	/* Restore an o32 sigcontext.  */
-	.set push
-	SET_HARDFLOAT
-	EX	lw t1, SC32_FPC_CSR(a0)
+#ifdef CONFIG_CPU_HAS_MSA
 
-#ifndef CONFIG_CPU_MIPS64_R6
-	mfc0	t0, CP0_STATUS
-	sll	t0, t0, 5
-	bgez	t0, 1f			# skip loading odd if FR=0
+	.macro	op_one_wr	op, idx, base
+	.align	4
+\idx:	\op	\idx, 0, \base
+	jr	ra
 	 nop
-#endif
+	.endm
 
-	EX      ldc1 $f1, SC32_FPREGS+8(a0)
-	EX      ldc1 $f3, SC32_FPREGS+24(a0)
-	EX      ldc1 $f5, SC32_FPREGS+40(a0)
-	EX      ldc1 $f7, SC32_FPREGS+56(a0)
-	EX      ldc1 $f9, SC32_FPREGS+72(a0)
-	EX      ldc1 $f11, SC32_FPREGS+88(a0)
-	EX      ldc1 $f13, SC32_FPREGS+104(a0)
-	EX      ldc1 $f15, SC32_FPREGS+120(a0)
-	EX      ldc1 $f17, SC32_FPREGS+136(a0)
-	EX      ldc1 $f19, SC32_FPREGS+152(a0)
-	EX      ldc1 $f21, SC32_FPREGS+168(a0)
-	EX      ldc1 $f23, SC32_FPREGS+184(a0)
-	EX      ldc1 $f25, SC32_FPREGS+200(a0)
-	EX      ldc1 $f27, SC32_FPREGS+216(a0)
-	EX      ldc1 $f29, SC32_FPREGS+232(a0)
-	EX      ldc1 $f31, SC32_FPREGS+248(a0)
+	.macro	op_msa_wr	name, op
+LEAF(\name)
+	.set		push
+	.set		noreorder
+	sll		t0, a0, 4
+	PTR_LA		t1, 0f
+	PTR_ADDU	t0, t0, t1
+	jr		t0
+	  nop
+	op_one_wr	\op, 0, a1
+	op_one_wr	\op, 1, a1
+	op_one_wr	\op, 2, a1
+	op_one_wr	\op, 3, a1
+	op_one_wr	\op, 4, a1
+	op_one_wr	\op, 5, a1
+	op_one_wr	\op, 6, a1
+	op_one_wr	\op, 7, a1
+	op_one_wr	\op, 8, a1
+	op_one_wr	\op, 9, a1
+	op_one_wr	\op, 10, a1
+	op_one_wr	\op, 11, a1
+	op_one_wr	\op, 12, a1
+	op_one_wr	\op, 13, a1
+	op_one_wr	\op, 14, a1
+	op_one_wr	\op, 15, a1
+	op_one_wr	\op, 16, a1
+	op_one_wr	\op, 17, a1
+	op_one_wr	\op, 18, a1
+	op_one_wr	\op, 19, a1
+	op_one_wr	\op, 20, a1
+	op_one_wr	\op, 21, a1
+	op_one_wr	\op, 22, a1
+	op_one_wr	\op, 23, a1
+	op_one_wr	\op, 24, a1
+	op_one_wr	\op, 25, a1
+	op_one_wr	\op, 26, a1
+	op_one_wr	\op, 27, a1
+	op_one_wr	\op, 28, a1
+	op_one_wr	\op, 29, a1
+	op_one_wr	\op, 30, a1
+	op_one_wr	\op, 31, a1
+	.set		pop
+	END(\name)
+	.endm
 
-1:	EX	ldc1 $f0, SC32_FPREGS+0(a0)
-	EX	ldc1 $f2, SC32_FPREGS+16(a0)
-	EX	ldc1 $f4, SC32_FPREGS+32(a0)
-	EX	ldc1 $f6, SC32_FPREGS+48(a0)
-	EX	ldc1 $f8, SC32_FPREGS+64(a0)
-	EX	ldc1 $f10, SC32_FPREGS+80(a0)
-	EX	ldc1 $f12, SC32_FPREGS+96(a0)
-	EX	ldc1 $f14, SC32_FPREGS+112(a0)
-	EX	ldc1 $f16, SC32_FPREGS+128(a0)
-	EX	ldc1 $f18, SC32_FPREGS+144(a0)
-	EX	ldc1 $f20, SC32_FPREGS+160(a0)
-	EX	ldc1 $f22, SC32_FPREGS+176(a0)
-	EX	ldc1 $f24, SC32_FPREGS+192(a0)
-	EX	ldc1 $f26, SC32_FPREGS+208(a0)
-	EX	ldc1 $f28, SC32_FPREGS+224(a0)
-	EX	ldc1 $f30, SC32_FPREGS+240(a0)
-	ctc1	t1, fcr31
+	op_msa_wr	read_msa_wr_b, st_b
+	op_msa_wr	read_msa_wr_h, st_h
+	op_msa_wr	read_msa_wr_w, st_w
+	op_msa_wr	read_msa_wr_d, st_d
+
+	op_msa_wr	write_msa_wr_b, ld_b
+	op_msa_wr	write_msa_wr_h, ld_h
+	op_msa_wr	write_msa_wr_w, ld_w
+	op_msa_wr	write_msa_wr_d, ld_d
+
+#endif /* CONFIG_CPU_HAS_MSA */
+
+#ifdef CONFIG_CPU_HAS_MSA
+
+	.macro	save_msa_upper	wr, off, base
+	.set	push
+	.set	noat
+#ifdef CONFIG_64BIT
+	copy_u_d \wr, 1
+	EX sd	$1, \off(\base)
+#elif defined(CONFIG_CPU_LITTLE_ENDIAN)
+	copy_u_w \wr, 2
+	EX sw	$1, \off(\base)
+	copy_u_w \wr, 3
+	EX sw	$1, (\off+4)(\base)
+#else /* CONFIG_CPU_BIG_ENDIAN */
+	copy_u_w \wr, 2
+	EX sw	$1, (\off+4)(\base)
+	copy_u_w \wr, 3
+	EX sw	$1, \off(\base)
+#endif
+	.set	pop
+	.endm
+
+LEAF(_save_msa_all_upper)
+	save_msa_upper	0, 0x00, a0
+	save_msa_upper	1, 0x08, a0
+	save_msa_upper	2, 0x10, a0
+	save_msa_upper	3, 0x18, a0
+	save_msa_upper	4, 0x20, a0
+	save_msa_upper	5, 0x28, a0
+	save_msa_upper	6, 0x30, a0
+	save_msa_upper	7, 0x38, a0
+	save_msa_upper	8, 0x40, a0
+	save_msa_upper	9, 0x48, a0
+	save_msa_upper	10, 0x50, a0
+	save_msa_upper	11, 0x58, a0
+	save_msa_upper	12, 0x60, a0
+	save_msa_upper	13, 0x68, a0
+	save_msa_upper	14, 0x70, a0
+	save_msa_upper	15, 0x78, a0
+	save_msa_upper	16, 0x80, a0
+	save_msa_upper	17, 0x88, a0
+	save_msa_upper	18, 0x90, a0
+	save_msa_upper	19, 0x98, a0
+	save_msa_upper	20, 0xa0, a0
+	save_msa_upper	21, 0xa8, a0
+	save_msa_upper	22, 0xb0, a0
+	save_msa_upper	23, 0xb8, a0
+	save_msa_upper	24, 0xc0, a0
+	save_msa_upper	25, 0xc8, a0
+	save_msa_upper	26, 0xd0, a0
+	save_msa_upper	27, 0xd8, a0
+	save_msa_upper	28, 0xe0, a0
+	save_msa_upper	29, 0xe8, a0
+	save_msa_upper	30, 0xf0, a0
+	save_msa_upper	31, 0xf8, a0
 	jr	ra
-	 li	v0, 0					# success
-	.set pop
-	END(_restore_fp_context32)
+	 li	v0, 0
+	END(_save_msa_all_upper)
+
+	.macro	restore_msa_upper	wr, off, base
+	.set	push
+	.set	noat
+#ifdef CONFIG_64BIT
+	EX ld	$1, \off(\base)
+	insert_d \wr, 1
+#elif defined(CONFIG_CPU_LITTLE_ENDIAN)
+	EX lw	$1, \off(\base)
+	insert_w \wr, 2
+	EX lw	$1, (\off+4)(\base)
+	insert_w \wr, 3
+#else /* CONFIG_CPU_BIG_ENDIAN */
+	EX lw	$1, (\off+4)(\base)
+	insert_w \wr, 2
+	EX lw	$1, \off(\base)
+	insert_w \wr, 3
 #endif
+	.set	pop
+	.endm
+
+LEAF(_restore_msa_all_upper)
+	restore_msa_upper	0, 0x00, a0
+	restore_msa_upper	1, 0x08, a0
+	restore_msa_upper	2, 0x10, a0
+	restore_msa_upper	3, 0x18, a0
+	restore_msa_upper	4, 0x20, a0
+	restore_msa_upper	5, 0x28, a0
+	restore_msa_upper	6, 0x30, a0
+	restore_msa_upper	7, 0x38, a0
+	restore_msa_upper	8, 0x40, a0
+	restore_msa_upper	9, 0x48, a0
+	restore_msa_upper	10, 0x50, a0
+	restore_msa_upper	11, 0x58, a0
+	restore_msa_upper	12, 0x60, a0
+	restore_msa_upper	13, 0x68, a0
+	restore_msa_upper	14, 0x70, a0
+	restore_msa_upper	15, 0x78, a0
+	restore_msa_upper	16, 0x80, a0
+	restore_msa_upper	17, 0x88, a0
+	restore_msa_upper	18, 0x90, a0
+	restore_msa_upper	19, 0x98, a0
+	restore_msa_upper	20, 0xa0, a0
+	restore_msa_upper	21, 0xa8, a0
+	restore_msa_upper	22, 0xb0, a0
+	restore_msa_upper	23, 0xb8, a0
+	restore_msa_upper	24, 0xc0, a0
+	restore_msa_upper	25, 0xc8, a0
+	restore_msa_upper	26, 0xd0, a0
+	restore_msa_upper	27, 0xd8, a0
+	restore_msa_upper	28, 0xe0, a0
+	restore_msa_upper	29, 0xe8, a0
+	restore_msa_upper	30, 0xf0, a0
+	restore_msa_upper	31, 0xf8, a0
+	jr	ra
+	 li	v0, 0
+	END(_restore_msa_all_upper)
+
+#endif /* CONFIG_CPU_HAS_MSA */
 
 	.set	reorder
 
diff --git a/arch/mips/kernel/r4k_switch.S b/arch/mips/kernel/r4k_switch.S
index 04cbbde3..92cd051 100644
--- a/arch/mips/kernel/r4k_switch.S
+++ b/arch/mips/kernel/r4k_switch.S
@@ -34,7 +34,7 @@
 #ifndef USE_ALTERNATE_RESUME_IMPL
 /*
  * task_struct *resume(task_struct *prev, task_struct *next,
- *		       struct thread_info *next_ti, s32 fp_save)
+ *		       struct thread_info *next_ti)
  */
 	.align	5
 	LEAF(resume)
@@ -43,45 +43,6 @@
 	cpu_save_nonscratch a0
 	LONG_S	ra, THREAD_REG31(a0)
 
-	/*
-	 * Check whether we need to save any FP context. FP context is saved
-	 * iff the process has used the context with the scalar FPU or the MSA
-	 * ASE in the current time slice, as indicated by _TIF_USEDFPU and
-	 * _TIF_USEDMSA respectively. switch_to will have set fp_save
-	 * accordingly to an FP_SAVE_ enum value.
-	 */
-	beqz	a3, 2f
-
-	/*
-	 * We do. Clear the saved CU1 bit for prev, such that next time it is
-	 * scheduled it will start in userland with the FPU disabled. If the
-	 * task uses the FPU then it will be enabled again via the do_cpu trap.
-	 * This allows us to lazily restore the FP context.
-	 */
-	PTR_L	t3, TASK_THREAD_INFO(a0)
-	LONG_L	t0, ST_OFF(t3)
-	li	t1, ~ST0_CU1
-	and	t0, t0, t1
-	LONG_S	t0, ST_OFF(t3)
-
-	/* Check whether we're saving scalar or vector context. */
-	bgtz	a3, 1f
-
-	/* Save 128b MSA vector context + scalar FP control & status. */
-	.set push
-	SET_HARDFLOAT
-	cfc1	t1, fcr31
-	msa_save_all	a0
-	.set pop	/* SET_HARDFLOAT */
-
-	sw	t1, THREAD_FCR31(a0)
-	b	2f
-
-1:	/* Save 32b/64b scalar FP context. */
-	fpu_save_double a0 t0 t1		# c0_status passed in t0
-						# clobbers t1
-2:
-
 #if defined(CONFIG_CC_STACKPROTECTOR) && !defined(CONFIG_SMP)
 	PTR_LA	t8, __stack_chk_guard
 	LONG_L	t9, TASK_STACK_CANARY(a1)
diff --git a/arch/mips/kernel/signal-common.h b/arch/mips/kernel/signal-common.h
index 0b85f82..f50d484 100644
--- a/arch/mips/kernel/signal-common.h
+++ b/arch/mips/kernel/signal-common.h
@@ -31,4 +31,13 @@ extern int fpcsr_pending(unsigned int __user *fpcsr);
 #define lock_fpu_owner()	({ preempt_disable(); pagefault_disable(); })
 #define unlock_fpu_owner()	({ pagefault_enable(); preempt_enable(); })
 
+/* Assembly functions to move context to/from the FPU */
+extern asmlinkage int
+_save_fp_context(void __user *fpregs, void __user *csr);
+extern asmlinkage int
+_restore_fp_context(void __user *fpregs, void __user *csr);
+
+extern asmlinkage int _save_msa_all_upper(void __user *buf);
+extern asmlinkage int _restore_msa_all_upper(void __user *buf);
+
 #endif	/* __SIGNAL_COMMON_H */
diff --git a/arch/mips/kernel/signal.c b/arch/mips/kernel/signal.c
index 6a28c79..2fec67b 100644
--- a/arch/mips/kernel/signal.c
+++ b/arch/mips/kernel/signal.c
@@ -21,6 +21,7 @@
 #include <linux/wait.h>
 #include <linux/ptrace.h>
 #include <linux/unistd.h>
+#include <linux/uprobes.h>
 #include <linux/compiler.h>
 #include <linux/syscalls.h>
 #include <linux/uaccess.h>
@@ -38,20 +39,21 @@
 #include <asm/vdso.h>
 #include <asm/dsp.h>
 #include <asm/inst.h>
+#include <asm/msa.h>
 
 #include "signal-common.h"
 
-static int (*save_fp_context)(struct sigcontext __user *sc);
-static int (*restore_fp_context)(struct sigcontext __user *sc);
-
-extern asmlinkage int _save_fp_context(struct sigcontext __user *sc);
-extern asmlinkage int _restore_fp_context(struct sigcontext __user *sc);
+static int (*save_fp_context)(void __user *sc);
+static int (*restore_fp_context)(void __user *sc);
 
 struct sigframe {
 	u32 sf_ass[4];		/* argument save space for o32 */
 	u32 sf_pad[2];		/* Was: signal trampoline */
+
+	/* Matches struct ucontext from its uc_mcontext field onwards */
 	struct sigcontext sf_sc;
 	sigset_t sf_mask;
+	unsigned long long sf_extcontext[0];
 };
 
 struct rt_sigframe {
@@ -65,43 +67,255 @@ struct rt_sigframe {
  * Thread saved context copy to/from a signal context presumed to be on the
  * user stack, and therefore accessed with appropriate macros from uaccess.h.
  */
-static int copy_fp_to_sigcontext(struct sigcontext __user *sc)
+static int copy_fp_to_sigcontext(void __user *sc)
 {
+	struct mips_abi *abi = current->thread.abi;
+	uint64_t __user *fpregs = sc + abi->off_sc_fpregs;
+	uint32_t __user *csr = sc + abi->off_sc_fpc_csr;
 	int i;
 	int err = 0;
+	int inc = test_thread_flag(TIF_32BIT_FPREGS) ? 2 : 1;
 
-	for (i = 0; i < NUM_FPU_REGS; i++) {
+	for (i = 0; i < NUM_FPU_REGS; i += inc) {
 		err |=
 		    __put_user(get_fpr64(&current->thread.fpu.fpr[i], 0),
-			       &sc->sc_fpregs[i]);
+			       &fpregs[i]);
 	}
-	err |= __put_user(current->thread.fpu.fcr31, &sc->sc_fpc_csr);
+	err |= __put_user(current->thread.fpu.fcr31, csr);
 
 	return err;
 }
 
-static int copy_fp_from_sigcontext(struct sigcontext __user *sc)
+static int copy_fp_from_sigcontext(void __user *sc)
 {
+	struct mips_abi *abi = current->thread.abi;
+	uint64_t __user *fpregs = sc + abi->off_sc_fpregs;
+	uint32_t __user *csr = sc + abi->off_sc_fpc_csr;
 	int i;
 	int err = 0;
+	int inc = test_thread_flag(TIF_32BIT_FPREGS) ? 2 : 1;
 	u64 fpr_val;
 
-	for (i = 0; i < NUM_FPU_REGS; i++) {
-		err |= __get_user(fpr_val, &sc->sc_fpregs[i]);
+	for (i = 0; i < NUM_FPU_REGS; i += inc) {
+		err |= __get_user(fpr_val, &fpregs[i]);
 		set_fpr64(&current->thread.fpu.fpr[i], 0, fpr_val);
 	}
-	err |= __get_user(current->thread.fpu.fcr31, &sc->sc_fpc_csr);
+	err |= __get_user(current->thread.fpu.fcr31, csr);
 
 	return err;
 }
 
 /*
+ * Wrappers for the assembly _{save,restore}_fp_context functions.
+ */
+static int save_hw_fp_context(void __user *sc)
+{
+	struct mips_abi *abi = current->thread.abi;
+	uint64_t __user *fpregs = sc + abi->off_sc_fpregs;
+	uint32_t __user *csr = sc + abi->off_sc_fpc_csr;
+
+	return _save_fp_context(fpregs, csr);
+}
+
+static int restore_hw_fp_context(void __user *sc)
+{
+	struct mips_abi *abi = current->thread.abi;
+	uint64_t __user *fpregs = sc + abi->off_sc_fpregs;
+	uint32_t __user *csr = sc + abi->off_sc_fpc_csr;
+
+	return _restore_fp_context(fpregs, csr);
+}
+
+/*
+ * Extended context handling.
+ */
+
+static inline void __user *sc_to_extcontext(void __user *sc)
+{
+	struct ucontext __user *uc;
+
+	/*
+	 * We can just pretend the sigcontext is always embedded in a struct
+	 * ucontext here, because the offset from sigcontext to extended
+	 * context is the same in the struct sigframe case.
+	 */
+	uc = container_of(sc, struct ucontext, uc_mcontext);
+	return &uc->uc_extcontext;
+}
+
+static int save_msa_extcontext(void __user *buf)
+{
+	struct msa_extcontext __user *msa = buf;
+	uint64_t val;
+	int i, err;
+
+	if (!thread_msa_context_live())
+		return 0;
+
+	/*
+	 * Ensure that we can't lose the live MSA context between checking
+	 * for it & writing it to memory.
+	 */
+	preempt_disable();
+
+	if (is_msa_enabled()) {
+		/*
+		 * There are no EVA versions of the vector register load/store
+		 * instructions, so MSA context has to be saved to kernel memory
+		 * and then copied to user memory. The save to kernel memory
+		 * should already have been done when handling scalar FP
+		 * context.
+		 */
+		BUG_ON(config_enabled(CONFIG_EVA));
+
+		err = __put_user(read_msa_csr(), &msa->csr);
+		err |= _save_msa_all_upper(&msa->wr);
+
+		preempt_enable();
+	} else {
+		preempt_enable();
+
+		err = __put_user(current->thread.fpu.msacsr, &msa->csr);
+
+		for (i = 0; i < NUM_FPU_REGS; i++) {
+			val = get_fpr64(&current->thread.fpu.fpr[i], 1);
+			err |= __put_user(val, &msa->wr[i]);
+		}
+	}
+
+	err |= __put_user(MSA_EXTCONTEXT_MAGIC, &msa->ext.magic);
+	err |= __put_user(sizeof(*msa), &msa->ext.size);
+
+	return err ? -EFAULT : sizeof(*msa);
+}
+
+static int restore_msa_extcontext(void __user *buf, unsigned int size)
+{
+	struct msa_extcontext __user *msa = buf;
+	unsigned long long val;
+	unsigned int csr;
+	int i, err;
+
+	if (size != sizeof(*msa))
+		return -EINVAL;
+
+	err = get_user(csr, &msa->csr);
+	if (err)
+		return err;
+
+	preempt_disable();
+
+	if (is_msa_enabled()) {
+		/*
+		 * There are no EVA versions of the vector register load/store
+		 * instructions, so MSA context has to be copied to kernel
+		 * memory and later loaded to registers. The same is true of
+		 * scalar FP context, so FPU & MSA should have already been
+		 * disabled whilst handling scalar FP context.
+		 */
+		BUG_ON(config_enabled(CONFIG_EVA));
+
+		write_msa_csr(csr);
+		err |= _restore_msa_all_upper(&msa->wr);
+		preempt_enable();
+	} else {
+		preempt_enable();
+
+		current->thread.fpu.msacsr = csr;
+
+		for (i = 0; i < NUM_FPU_REGS; i++) {
+			err |= __get_user(val, &msa->wr[i]);
+			set_fpr64(&current->thread.fpu.fpr[i], 1, val);
+		}
+	}
+
+	return err;
+}
+
+static int save_extcontext(void __user *buf)
+{
+	int sz;
+
+	sz = save_msa_extcontext(buf);
+	if (sz < 0)
+		return sz;
+	buf += sz;
+
+	/* If no context was saved then trivially return */
+	if (!sz)
+		return 0;
+
+	/* Write the end marker */
+	if (__put_user(END_EXTCONTEXT_MAGIC, (u32 *)buf))
+		return -EFAULT;
+
+	sz += sizeof(((struct extcontext *)NULL)->magic);
+	return sz;
+}
+
+static int restore_extcontext(void __user *buf)
+{
+	struct extcontext ext;
+	int err;
+
+	while (1) {
+		err = __get_user(ext.magic, (unsigned int *)buf);
+		if (err)
+			return err;
+
+		if (ext.magic == END_EXTCONTEXT_MAGIC)
+			return 0;
+
+		err = __get_user(ext.size, (unsigned int *)(buf
+			+ offsetof(struct extcontext, size)));
+		if (err)
+			return err;
+
+		switch (ext.magic) {
+		case MSA_EXTCONTEXT_MAGIC:
+			err = restore_msa_extcontext(buf, ext.size);
+			break;
+
+		default:
+			err = -EINVAL;
+			break;
+		}
+
+		if (err)
+			return err;
+
+		buf += ext.size;
+	}
+}
+
+/*
  * Helper routines
  */
-static int protected_save_fp_context(struct sigcontext __user *sc)
+int protected_save_fp_context(void __user *sc)
 {
+	struct mips_abi *abi = current->thread.abi;
+	uint64_t __user *fpregs = sc + abi->off_sc_fpregs;
+	uint32_t __user *csr = sc + abi->off_sc_fpc_csr;
+	uint32_t __user *used_math = sc + abi->off_sc_used_math;
+	unsigned int used, ext_sz;
 	int err;
-#ifndef CONFIG_EVA
+
+	used = used_math() ? USED_FP : 0;
+	if (!used)
+		goto fp_done;
+
+	if (!test_thread_flag(TIF_32BIT_FPREGS))
+		used |= USED_FR1;
+	if (test_thread_flag(TIF_HYBRID_FPREGS))
+		used |= USED_HYBRID_FPRS;
+
+	/*
+	 * EVA does not have userland equivalents of ldc1 or sdc1, so
+	 * save to the kernel FP context & copy that to userland below.
+	 */
+	if (config_enabled(CONFIG_EVA))
+		lose_fpu(1);
+
 	while (1) {
 		lock_fpu_owner();
 		if (is_fpu_owner()) {
@@ -114,27 +328,57 @@ static int protected_save_fp_context(struct sigcontext __user *sc)
 		if (likely(!err))
 			break;
 		/* touch the sigcontext and try again */
-		err = __put_user(0, &sc->sc_fpregs[0]) |
-			__put_user(0, &sc->sc_fpregs[31]) |
-			__put_user(0, &sc->sc_fpc_csr);
+		err = __put_user(0, &fpregs[0]) |
+			__put_user(0, &fpregs[31]) |
+			__put_user(0, csr);
 		if (err)
-			break;	/* really bad sigcontext */
+			return err;	/* really bad sigcontext */
 	}
-#else
-	/*
-	 * EVA does not have FPU EVA instructions so saving fpu context directly
-	 * does not work.
-	 */
-	lose_fpu(1);
-	err = save_fp_context(sc); /* this might fail */
-#endif
-	return err;
+
+fp_done:
+	ext_sz = err = save_extcontext(sc_to_extcontext(sc));
+	if (err < 0)
+		return err;
+	used |= ext_sz ? USED_EXTCONTEXT : 0;
+
+	return __put_user(used, used_math);
 }
 
-static int protected_restore_fp_context(struct sigcontext __user *sc)
+int protected_restore_fp_context(void __user *sc)
 {
-	int err, tmp __maybe_unused;
-#ifndef CONFIG_EVA
+	struct mips_abi *abi = current->thread.abi;
+	uint64_t __user *fpregs = sc + abi->off_sc_fpregs;
+	uint32_t __user *csr = sc + abi->off_sc_fpc_csr;
+	uint32_t __user *used_math = sc + abi->off_sc_used_math;
+	unsigned int used;
+	int err, sig = 0, tmp __maybe_unused;
+
+	err = __get_user(used, used_math);
+	conditional_used_math(used & USED_FP);
+
+	/*
+	 * The signal handler may have used FPU; give it up if the program
+	 * doesn't want it following sigreturn.
+	 */
+	if (err || !(used & USED_FP))
+		lose_fpu(0);
+	if (err)
+		return err;
+	if (!(used & USED_FP))
+		goto fp_done;
+
+	err = sig = fpcsr_pending(csr);
+	if (err < 0)
+		return err;
+
+	/*
+	 * EVA does not have userland equivalents of ldc1 or sdc1, so we
+	 * disable the FPU here such that the code below simply copies to
+	 * the kernel FP context.
+	 */
+	if (config_enabled(CONFIG_EVA))
+		lose_fpu(0);
+
 	while (1) {
 		lock_fpu_owner();
 		if (is_fpu_owner()) {
@@ -147,28 +391,24 @@ static int protected_restore_fp_context(struct sigcontext __user *sc)
 		if (likely(!err))
 			break;
 		/* touch the sigcontext and try again */
-		err = __get_user(tmp, &sc->sc_fpregs[0]) |
-			__get_user(tmp, &sc->sc_fpregs[31]) |
-			__get_user(tmp, &sc->sc_fpc_csr);
+		err = __get_user(tmp, &fpregs[0]) |
+			__get_user(tmp, &fpregs[31]) |
+			__get_user(tmp, csr);
 		if (err)
 			break;	/* really bad sigcontext */
 	}
-#else
-	/*
-	 * EVA does not have FPU EVA instructions so restoring fpu context
-	 * directly does not work.
-	 */
-	lose_fpu(0);
-	err = restore_fp_context(sc); /* this might fail */
-#endif
-	return err;
+
+fp_done:
+	if (used & USED_EXTCONTEXT)
+		err |= restore_extcontext(sc_to_extcontext(sc));
+
+	return err ?: sig;
 }
 
 int setup_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc)
 {
 	int err = 0;
 	int i;
-	unsigned int used_math;
 
 	err |= __put_user(regs->cp0_epc, &sc->sc_pc);
 
@@ -191,19 +431,38 @@ int setup_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc)
 		err |= __put_user(rddsp(DSP_MASK), &sc->sc_dsp);
 	}
 
-	used_math = !!used_math();
-	err |= __put_user(used_math, &sc->sc_used_math);
 
-	if (used_math) {
-		/*
-		 * Save FPU state to signal context. Signal handler
-		 * will "inherit" current FPU state.
-		 */
-		err |= protected_save_fp_context(sc);
-	}
+	/*
+	 * Save FPU state to signal context. Signal handler
+	 * will "inherit" current FPU state.
+	 */
+	err |= protected_save_fp_context(sc);
+
 	return err;
 }
 
+static size_t extcontext_max_size(void)
+{
+	size_t sz = 0;
+
+	/*
+	 * The assumption here is that between this point & the point at which
+	 * the extended context is saved the size of the context should only
+	 * ever be able to shrink (if the task is preempted), but never grow.
+	 * That is, what this function returns is an upper bound on the size of
+	 * the extended context for the current task at the current time.
+	 */
+
+	if (thread_msa_context_live())
+		sz += sizeof(struct msa_extcontext);
+
+	/* If any context is saved then we'll append the end marker */
+	if (sz)
+		sz += sizeof(((struct extcontext *)NULL)->magic);
+
+	return sz;
+}
+
 int fpcsr_pending(unsigned int __user *fpcsr)
 {
 	int err, sig = 0;
@@ -223,21 +482,8 @@ int fpcsr_pending(unsigned int __user *fpcsr)
 	return err ?: sig;
 }
 
-static int
-check_and_restore_fp_context(struct sigcontext __user *sc)
-{
-	int err, sig;
-
-	err = sig = fpcsr_pending(&sc->sc_fpc_csr);
-	if (err > 0)
-		err = 0;
-	err |= protected_restore_fp_context(sc);
-	return err ?: sig;
-}
-
 int restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc)
 {
-	unsigned int used_math;
 	unsigned long treg;
 	int err = 0;
 	int i;
@@ -265,19 +511,7 @@ int restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc)
 	for (i = 1; i < 32; i++)
 		err |= __get_user(regs->regs[i], &sc->sc_regs[i]);
 
-	err |= __get_user(used_math, &sc->sc_used_math);
-	conditional_used_math(used_math);
-
-	if (used_math) {
-		/* restore fpu context if we have used it before */
-		if (!err)
-			err = check_and_restore_fp_context(sc);
-	} else {
-		/* signal handler may have used FPU.  Give it up. */
-		lose_fpu(0);
-	}
-
-	return err;
+	return err ?: protected_restore_fp_context(sc);
 }
 
 void __user *get_sigframe(struct ksignal *ksig, struct pt_regs *regs,
@@ -285,6 +519,9 @@ void __user *get_sigframe(struct ksignal *ksig, struct pt_regs *regs,
 {
 	unsigned long sp;
 
+	/* Leave space for potential extended context */
+	frame_size += extcontext_max_size();
+
 	/* Default to using normal stack */
 	sp = regs->regs[29];
 
@@ -520,7 +757,11 @@ struct mips_abi mips_abi = {
 	.setup_rt_frame = setup_rt_frame,
 	.rt_signal_return_offset =
 		offsetof(struct mips_vdso, rt_signal_trampoline),
-	.restart	= __NR_restart_syscall
+	.restart	= __NR_restart_syscall,
+
+	.off_sc_fpregs = offsetof(struct sigcontext, sc_fpregs),
+	.off_sc_fpc_csr = offsetof(struct sigcontext, sc_fpc_csr),
+	.off_sc_used_math = offsetof(struct sigcontext, sc_used_math),
 };
 
 static void handle_signal(struct ksignal *ksig, struct pt_regs *regs)
@@ -616,6 +857,9 @@ asmlinkage void do_notify_resume(struct pt_regs *regs, void *unused,
 
 	user_exit();
 
+	if (thread_info_flags & _TIF_UPROBE)
+		uprobe_notify_resume(regs);
+
 	/* deal with pending signal delivery */
 	if (thread_info_flags & _TIF_SIGPENDING)
 		do_signal(regs);
@@ -629,43 +873,46 @@ asmlinkage void do_notify_resume(struct pt_regs *regs, void *unused,
 }
 
 #ifdef CONFIG_SMP
-#ifndef CONFIG_EVA
-static int smp_save_fp_context(struct sigcontext __user *sc)
+static int smp_save_fp_context(void __user *sc)
 {
 	return raw_cpu_has_fpu
-	       ? _save_fp_context(sc)
+	       ? save_hw_fp_context(sc)
 	       : copy_fp_to_sigcontext(sc);
 }
 
-static int smp_restore_fp_context(struct sigcontext __user *sc)
+static int smp_restore_fp_context(void __user *sc)
 {
 	return raw_cpu_has_fpu
-	       ? _restore_fp_context(sc)
+	       ? restore_hw_fp_context(sc)
 	       : copy_fp_from_sigcontext(sc);
 }
-#endif /* CONFIG_EVA */
 #endif
 
 static int signal_setup(void)
 {
-#ifndef CONFIG_EVA
+	/*
+	 * The offset from sigcontext to extended context should be the same
+	 * regardless of the type of signal, such that userland can always know
+	 * where to look if it wishes to find the extended context structures.
+	 */
+	BUILD_BUG_ON((offsetof(struct sigframe, sf_extcontext) -
+		      offsetof(struct sigframe, sf_sc)) !=
+		     (offsetof(struct rt_sigframe, rs_uc.uc_extcontext) -
+		      offsetof(struct rt_sigframe, rs_uc.uc_mcontext)));
+
 #ifdef CONFIG_SMP
 	/* For now just do the cpu_has_fpu check when the functions are invoked */
 	save_fp_context = smp_save_fp_context;
 	restore_fp_context = smp_restore_fp_context;
 #else
 	if (cpu_has_fpu) {
-		save_fp_context = _save_fp_context;
-		restore_fp_context = _restore_fp_context;
+		save_fp_context = save_hw_fp_context;
+		restore_fp_context = restore_hw_fp_context;
 	} else {
 		save_fp_context = copy_fp_to_sigcontext;
 		restore_fp_context = copy_fp_from_sigcontext;
 	}
 #endif /* CONFIG_SMP */
-#else
-	save_fp_context = copy_fp_to_sigcontext;
-	restore_fp_context = copy_fp_from_sigcontext;
-#endif
 
 	return 0;
 }
diff --git a/arch/mips/kernel/signal32.c b/arch/mips/kernel/signal32.c
index 5d7f263..f7e89524 100644
--- a/arch/mips/kernel/signal32.c
+++ b/arch/mips/kernel/signal32.c
@@ -36,12 +36,6 @@
 
 #include "signal-common.h"
 
-static int (*save_fp_context32)(struct sigcontext32 __user *sc);
-static int (*restore_fp_context32)(struct sigcontext32 __user *sc);
-
-extern asmlinkage int _save_fp_context32(struct sigcontext32 __user *sc);
-extern asmlinkage int _restore_fp_context32(struct sigcontext32 __user *sc);
-
 /*
  * Including <asm/unistd.h> would give use the 64-bit syscall numbers ...
  */
@@ -74,99 +68,11 @@ struct rt_sigframe32 {
 	struct ucontext32 rs_uc;
 };
 
-/*
- * Thread saved context copy to/from a signal context presumed to be on the
- * user stack, and therefore accessed with appropriate macros from uaccess.h.
- */
-static int copy_fp_to_sigcontext32(struct sigcontext32 __user *sc)
-{
-	int i;
-	int err = 0;
-	int inc = test_thread_flag(TIF_32BIT_FPREGS) ? 2 : 1;
-
-	for (i = 0; i < NUM_FPU_REGS; i += inc) {
-		err |=
-		    __put_user(get_fpr64(&current->thread.fpu.fpr[i], 0),
-			       &sc->sc_fpregs[i]);
-	}
-	err |= __put_user(current->thread.fpu.fcr31, &sc->sc_fpc_csr);
-
-	return err;
-}
-
-static int copy_fp_from_sigcontext32(struct sigcontext32 __user *sc)
-{
-	int i;
-	int err = 0;
-	int inc = test_thread_flag(TIF_32BIT_FPREGS) ? 2 : 1;
-	u64 fpr_val;
-
-	for (i = 0; i < NUM_FPU_REGS; i += inc) {
-		err |= __get_user(fpr_val, &sc->sc_fpregs[i]);
-		set_fpr64(&current->thread.fpu.fpr[i], 0, fpr_val);
-	}
-	err |= __get_user(current->thread.fpu.fcr31, &sc->sc_fpc_csr);
-
-	return err;
-}
-
-/*
- * sigcontext handlers
- */
-static int protected_save_fp_context32(struct sigcontext32 __user *sc)
-{
-	int err;
-	while (1) {
-		lock_fpu_owner();
-		if (is_fpu_owner()) {
-			err = save_fp_context32(sc);
-			unlock_fpu_owner();
-		} else {
-			unlock_fpu_owner();
-			err = copy_fp_to_sigcontext32(sc);
-		}
-		if (likely(!err))
-			break;
-		/* touch the sigcontext and try again */
-		err = __put_user(0, &sc->sc_fpregs[0]) |
-			__put_user(0, &sc->sc_fpregs[31]) |
-			__put_user(0, &sc->sc_fpc_csr);
-		if (err)
-			break;	/* really bad sigcontext */
-	}
-	return err;
-}
-
-static int protected_restore_fp_context32(struct sigcontext32 __user *sc)
-{
-	int err, tmp __maybe_unused;
-	while (1) {
-		lock_fpu_owner();
-		if (is_fpu_owner()) {
-			err = restore_fp_context32(sc);
-			unlock_fpu_owner();
-		} else {
-			unlock_fpu_owner();
-			err = copy_fp_from_sigcontext32(sc);
-		}
-		if (likely(!err))
-			break;
-		/* touch the sigcontext and try again */
-		err = __get_user(tmp, &sc->sc_fpregs[0]) |
-			__get_user(tmp, &sc->sc_fpregs[31]) |
-			__get_user(tmp, &sc->sc_fpc_csr);
-		if (err)
-			break;	/* really bad sigcontext */
-	}
-	return err;
-}
-
 static int setup_sigcontext32(struct pt_regs *regs,
 			      struct sigcontext32 __user *sc)
 {
 	int err = 0;
 	int i;
-	u32 used_math;
 
 	err |= __put_user(regs->cp0_epc, &sc->sc_pc);
 
@@ -186,35 +92,18 @@ static int setup_sigcontext32(struct pt_regs *regs,
 		err |= __put_user(mflo3(), &sc->sc_lo3);
 	}
 
-	used_math = !!used_math();
-	err |= __put_user(used_math, &sc->sc_used_math);
+	/*
+	 * Save FPU state to signal context.  Signal handler
+	 * will "inherit" current FPU state.
+	 */
+	err |= protected_save_fp_context(sc);
 
-	if (used_math) {
-		/*
-		 * Save FPU state to signal context.  Signal handler
-		 * will "inherit" current FPU state.
-		 */
-		err |= protected_save_fp_context32(sc);
-	}
 	return err;
 }
 
-static int
-check_and_restore_fp_context32(struct sigcontext32 __user *sc)
-{
-	int err, sig;
-
-	err = sig = fpcsr_pending(&sc->sc_fpc_csr);
-	if (err > 0)
-		err = 0;
-	err |= protected_restore_fp_context32(sc);
-	return err ?: sig;
-}
-
 static int restore_sigcontext32(struct pt_regs *regs,
 				struct sigcontext32 __user *sc)
 {
-	u32 used_math;
 	int err = 0;
 	s32 treg;
 	int i;
@@ -238,70 +127,7 @@ static int restore_sigcontext32(struct pt_regs *regs,
 	for (i = 1; i < 32; i++)
 		err |= __get_user(regs->regs[i], &sc->sc_regs[i]);
 
-	err |= __get_user(used_math, &sc->sc_used_math);
-	conditional_used_math(used_math);
-
-	if (used_math) {
-		/* restore fpu context if we have used it before */
-		if (!err)
-			err = check_and_restore_fp_context32(sc);
-	} else {
-		/* signal handler may have used FPU.  Give it up. */
-		lose_fpu(0);
-	}
-
-	return err;
-}
-
-/*
- *
- */
-extern void __put_sigset_unknown_nsig(void);
-extern void __get_sigset_unknown_nsig(void);
-
-static inline int put_sigset(const sigset_t *kbuf, compat_sigset_t __user *ubuf)
-{
-	int err = 0;
-
-	if (!access_ok(VERIFY_WRITE, ubuf, sizeof(*ubuf)))
-		return -EFAULT;
-
-	switch (_NSIG_WORDS) {
-	default:
-		__put_sigset_unknown_nsig();
-	case 2:
-		err |= __put_user(kbuf->sig[1] >> 32, &ubuf->sig[3]);
-		err |= __put_user(kbuf->sig[1] & 0xffffffff, &ubuf->sig[2]);
-	case 1:
-		err |= __put_user(kbuf->sig[0] >> 32, &ubuf->sig[1]);
-		err |= __put_user(kbuf->sig[0] & 0xffffffff, &ubuf->sig[0]);
-	}
-
-	return err;
-}
-
-static inline int get_sigset(sigset_t *kbuf, const compat_sigset_t __user *ubuf)
-{
-	int err = 0;
-	unsigned long sig[4];
-
-	if (!access_ok(VERIFY_READ, ubuf, sizeof(*ubuf)))
-		return -EFAULT;
-
-	switch (_NSIG_WORDS) {
-	default:
-		__get_sigset_unknown_nsig();
-	case 2:
-		err |= __get_user(sig[3], &ubuf->sig[3]);
-		err |= __get_user(sig[2], &ubuf->sig[2]);
-		kbuf->sig[1] = sig[2] | (sig[3] << 32);
-	case 1:
-		err |= __get_user(sig[1], &ubuf->sig[1]);
-		err |= __get_user(sig[0], &ubuf->sig[0]);
-		kbuf->sig[0] = sig[0] | (sig[1] << 32);
-	}
-
-	return err;
+	return err ?: protected_restore_fp_context(sc);
 }
 
 /*
@@ -585,20 +411,9 @@ struct mips_abi mips_abi_32 = {
 	.setup_rt_frame = setup_rt_frame_32,
 	.rt_signal_return_offset =
 		offsetof(struct mips_vdso, o32_rt_signal_trampoline),
-	.restart	= __NR_O32_restart_syscall
-};
-
-static int signal32_init(void)
-{
-	if (cpu_has_fpu) {
-		save_fp_context32 = _save_fp_context32;
-		restore_fp_context32 = _restore_fp_context32;
-	} else {
-		save_fp_context32 = copy_fp_to_sigcontext32;
-		restore_fp_context32 = copy_fp_from_sigcontext32;
-	}
+	.restart	= __NR_O32_restart_syscall,
 
-	return 0;
-}
-
-arch_initcall(signal32_init);
+	.off_sc_fpregs = offsetof(struct sigcontext32, sc_fpregs),
+	.off_sc_fpc_csr = offsetof(struct sigcontext32, sc_fpc_csr),
+	.off_sc_used_math = offsetof(struct sigcontext32, sc_used_math),
+};
diff --git a/arch/mips/kernel/signal_n32.c b/arch/mips/kernel/signal_n32.c
index f1d4751..0d017fd 100644
--- a/arch/mips/kernel/signal_n32.c
+++ b/arch/mips/kernel/signal_n32.c
@@ -153,5 +153,9 @@ struct mips_abi mips_abi_n32 = {
 	.setup_rt_frame = setup_rt_frame_n32,
 	.rt_signal_return_offset =
 		offsetof(struct mips_vdso, n32_rt_signal_trampoline),
-	.restart	= __NR_N32_restart_syscall
+	.restart	= __NR_N32_restart_syscall,
+
+	.off_sc_fpregs = offsetof(struct sigcontext, sc_fpregs),
+	.off_sc_fpc_csr = offsetof(struct sigcontext, sc_fpc_csr),
+	.off_sc_used_math = offsetof(struct sigcontext, sc_used_math),
 };
diff --git a/arch/mips/kernel/spram.c b/arch/mips/kernel/spram.c
index d1168d7..8489c88 100644
--- a/arch/mips/kernel/spram.c
+++ b/arch/mips/kernel/spram.c
@@ -209,6 +209,7 @@ void spram_config(void)
 	case CPU_PROAPTIV:
 	case CPU_P5600:
 	case CPU_QEMU_GENERIC:
+	case CPU_I6400:
 		config0 = read_c0_config();
 		/* FIXME: addresses are Malta specific */
 		if (config0 & (1<<24)) {
diff --git a/arch/mips/kernel/sysrq.c b/arch/mips/kernel/sysrq.c
index 5b539f5..5f05539 100644
--- a/arch/mips/kernel/sysrq.c
+++ b/arch/mips/kernel/sysrq.c
@@ -21,24 +21,12 @@ static DEFINE_SPINLOCK(show_lock);
 
 static void sysrq_tlbdump_single(void *dummy)
 {
-	const int field = 2 * sizeof(unsigned long);
 	unsigned long flags;
 
 	spin_lock_irqsave(&show_lock, flags);
 
 	pr_info("CPU%d:\n", smp_processor_id());
-	pr_info("Index	: %0x\n", read_c0_index());
-	pr_info("Pagemask: %0x\n", read_c0_pagemask());
-	pr_info("EntryHi : %0*lx\n", field, read_c0_entryhi());
-	pr_info("EntryLo0: %0*lx\n", field, read_c0_entrylo0());
-	pr_info("EntryLo1: %0*lx\n", field, read_c0_entrylo1());
-	pr_info("Wired   : %0x\n", read_c0_wired());
-	pr_info("Pagegrain: %0x\n", read_c0_pagegrain());
-	if (cpu_has_htw) {
-		pr_info("PWField : %0*lx\n", field, read_c0_pwfield());
-		pr_info("PWSize  : %0*lx\n", field, read_c0_pwsize());
-		pr_info("PWCtl   : %0x\n", read_c0_pwctl());
-	}
+	dump_tlb_regs();
 	pr_info("\n");
 	dump_tlb_all();
 	pr_info("\n");
diff --git a/arch/mips/kernel/traps.c b/arch/mips/kernel/traps.c
index 8ea28e6..fdb392b 100644
--- a/arch/mips/kernel/traps.c
+++ b/arch/mips/kernel/traps.c
@@ -370,11 +370,6 @@ void show_registers(struct pt_regs *regs)
 	set_fs(old_fs);
 }
 
-static int regs_to_trapnr(struct pt_regs *regs)
-{
-	return (regs->cp0_cause >> 2) & 0x1f;
-}
-
 static DEFINE_RAW_SPINLOCK(die_lock);
 
 void __noreturn die(const char *str, struct pt_regs *regs)
@@ -384,7 +379,7 @@ void __noreturn die(const char *str, struct pt_regs *regs)
 
 	oops_enter();
 
-	if (notify_die(DIE_OOPS, str, regs, 0, regs_to_trapnr(regs),
+	if (notify_die(DIE_OOPS, str, regs, 0, current->thread.trap_nr,
 		       SIGSEGV) == NOTIFY_STOP)
 		sig = 0;
 
@@ -470,7 +465,7 @@ asmlinkage void do_be(struct pt_regs *regs)
 	printk(KERN_ALERT "%s bus error, epc == %0*lx, ra == %0*lx\n",
 	       data ? "Data" : "Instruction",
 	       field, regs->cp0_epc, field, regs->regs[31]);
-	if (notify_die(DIE_OOPS, "bus error", regs, 0, regs_to_trapnr(regs),
+	if (notify_die(DIE_OOPS, "bus error", regs, 0, current->thread.trap_nr,
 		       SIGBUS) == NOTIFY_STOP)
 		goto out;
 
@@ -826,7 +821,7 @@ asmlinkage void do_fpe(struct pt_regs *regs, unsigned long fcr31)
 	int sig;
 
 	prev_state = exception_enter();
-	if (notify_die(DIE_FP, "FP exception", regs, 0, regs_to_trapnr(regs),
+	if (notify_die(DIE_FP, "FP exception", regs, 0, current->thread.trap_nr,
 		       SIGFPE) == NOTIFY_STOP)
 		goto out;
 
@@ -882,11 +877,12 @@ void do_trap_or_bp(struct pt_regs *regs, unsigned int code,
 	char b[40];
 
 #ifdef CONFIG_KGDB_LOW_LEVEL_TRAP
-	if (kgdb_ll_trap(DIE_TRAP, str, regs, code, regs_to_trapnr(regs), SIGTRAP) == NOTIFY_STOP)
+	if (kgdb_ll_trap(DIE_TRAP, str, regs, code, current->thread.trap_nr,
+			 SIGTRAP) == NOTIFY_STOP)
 		return;
 #endif /* CONFIG_KGDB_LOW_LEVEL_TRAP */
 
-	if (notify_die(DIE_TRAP, str, regs, code, regs_to_trapnr(regs),
+	if (notify_die(DIE_TRAP, str, regs, code, current->thread.trap_nr,
 		       SIGTRAP) == NOTIFY_STOP)
 		return;
 
@@ -948,6 +944,7 @@ asmlinkage void do_bp(struct pt_regs *regs)
 		set_fs(KERNEL_DS);
 
 	prev_state = exception_enter();
+	current->thread.trap_nr = (regs->cp0_cause >> 2) & 0x1f;
 	if (get_isa16_mode(regs->cp0_epc)) {
 		u16 instr[2];
 
@@ -987,15 +984,27 @@ asmlinkage void do_bp(struct pt_regs *regs)
 	 * pertain to them.
 	 */
 	switch (bcode) {
+	case BRK_UPROBE:
+		if (notify_die(DIE_UPROBE, "uprobe", regs, bcode,
+			       current->thread.trap_nr, SIGTRAP) == NOTIFY_STOP)
+			goto out;
+		else
+			break;
+	case BRK_UPROBE_XOL:
+		if (notify_die(DIE_UPROBE_XOL, "uprobe_xol", regs, bcode,
+			       current->thread.trap_nr, SIGTRAP) == NOTIFY_STOP)
+			goto out;
+		else
+			break;
 	case BRK_KPROBE_BP:
 		if (notify_die(DIE_BREAK, "debug", regs, bcode,
-			       regs_to_trapnr(regs), SIGTRAP) == NOTIFY_STOP)
+			       current->thread.trap_nr, SIGTRAP) == NOTIFY_STOP)
 			goto out;
 		else
 			break;
 	case BRK_KPROBE_SSTEPBP:
 		if (notify_die(DIE_SSTEPBP, "single_step", regs, bcode,
-			       regs_to_trapnr(regs), SIGTRAP) == NOTIFY_STOP)
+			       current->thread.trap_nr, SIGTRAP) == NOTIFY_STOP)
 			goto out;
 		else
 			break;
@@ -1028,6 +1037,7 @@ asmlinkage void do_tr(struct pt_regs *regs)
 		set_fs(get_ds());
 
 	prev_state = exception_enter();
+	current->thread.trap_nr = (regs->cp0_cause >> 2) & 0x1f;
 	if (get_isa16_mode(regs->cp0_epc)) {
 		if (__get_user(instr[0], (u16 __user *)(epc + 0)) ||
 		    __get_user(instr[1], (u16 __user *)(epc + 2)))
@@ -1094,8 +1104,9 @@ asmlinkage void do_ri(struct pt_regs *regs)
 no_r2_instr:
 
 	prev_state = exception_enter();
+	current->thread.trap_nr = (regs->cp0_cause >> 2) & 0x1f;
 
-	if (notify_die(DIE_RI, "RI Fault", regs, 0, regs_to_trapnr(regs),
+	if (notify_die(DIE_RI, "RI Fault", regs, 0, current->thread.trap_nr,
 		       SIGILL) == NOTIFY_STOP)
 		goto out;
 
@@ -1444,8 +1455,9 @@ asmlinkage void do_msa_fpe(struct pt_regs *regs, unsigned int msacsr)
 	enum ctx_state prev_state;
 
 	prev_state = exception_enter();
+	current->thread.trap_nr = (regs->cp0_cause >> 2) & 0x1f;
 	if (notify_die(DIE_MSAFP, "MSA FP exception", regs, 0,
-		       regs_to_trapnr(regs), SIGFPE) == NOTIFY_STOP)
+		       current->thread.trap_nr, SIGFPE) == NOTIFY_STOP)
 		goto out;
 
 	/* Clear MSACSR.Cause before enabling interrupts */
@@ -1523,7 +1535,6 @@ asmlinkage void do_watch(struct pt_regs *regs)
 
 asmlinkage void do_mcheck(struct pt_regs *regs)
 {
-	const int field = 2 * sizeof(unsigned long);
 	int multi_match = regs->cp0_status & ST0_TS;
 	enum ctx_state prev_state;
 	mm_segment_t old_fs = get_fs();
@@ -1532,19 +1543,8 @@ asmlinkage void do_mcheck(struct pt_regs *regs)
 	show_regs(regs);
 
 	if (multi_match) {
-		pr_err("Index	: %0x\n", read_c0_index());
-		pr_err("Pagemask: %0x\n", read_c0_pagemask());
-		pr_err("EntryHi : %0*lx\n", field, read_c0_entryhi());
-		pr_err("EntryLo0: %0*lx\n", field, read_c0_entrylo0());
-		pr_err("EntryLo1: %0*lx\n", field, read_c0_entrylo1());
-		pr_err("Wired   : %0x\n", read_c0_wired());
-		pr_err("Pagegrain: %0x\n", read_c0_pagegrain());
-		if (cpu_has_htw) {
-			pr_err("PWField : %0*lx\n", field, read_c0_pwfield());
-			pr_err("PWSize  : %0*lx\n", field, read_c0_pwsize());
-			pr_err("PWCtl   : %0x\n", read_c0_pwctl());
-		}
-		pr_err("\n");
+		dump_tlb_regs();
+		pr_info("\n");
 		dump_tlb_all();
 	}
 
@@ -1651,6 +1651,7 @@ static inline void parity_protection_init(void)
 	case CPU_PROAPTIV:
 	case CPU_P5600:
 	case CPU_QEMU_GENERIC:
+	case CPU_I6400:
 		{
 #define ERRCTL_PE	0x80000000
 #define ERRCTL_L2P	0x00800000
diff --git a/arch/mips/kernel/unaligned.c b/arch/mips/kernel/unaligned.c
index eb3efd1..990354d 100644
--- a/arch/mips/kernel/unaligned.c
+++ b/arch/mips/kernel/unaligned.c
@@ -891,6 +891,9 @@ static void emulate_load_store_insn(struct pt_regs *regs,
 #ifdef	CONFIG_EVA
 	mm_segment_t seg;
 #endif
+	union fpureg *fpr;
+	enum msa_2b_fmt df;
+	unsigned int wd;
 	origpc = (unsigned long)pc;
 	orig31 = regs->regs[31];
 
@@ -1202,6 +1205,75 @@ static void emulate_load_store_insn(struct pt_regs *regs,
 			break;
 		return;
 
+	case msa_op:
+		if (!cpu_has_msa)
+			goto sigill;
+
+		/*
+		 * If we've reached this point then userland should have taken
+		 * the MSA disabled exception & initialised vector context at
+		 * some point in the past.
+		 */
+		BUG_ON(!thread_msa_context_live());
+
+		df = insn.msa_mi10_format.df;
+		wd = insn.msa_mi10_format.wd;
+		fpr = &current->thread.fpu.fpr[wd];
+
+		switch (insn.msa_mi10_format.func) {
+		case msa_ld_op:
+			if (!access_ok(VERIFY_READ, addr, sizeof(*fpr)))
+				goto sigbus;
+
+			/*
+			 * Disable preemption to avoid a race between copying
+			 * state from userland, migrating to another CPU and
+			 * updating the hardware vector register below.
+			 */
+			preempt_disable();
+
+			res = __copy_from_user_inatomic(fpr, addr,
+							sizeof(*fpr));
+			if (res)
+				goto fault;
+
+			/*
+			 * Update the hardware register if it is in use by the
+			 * task in this quantum, in order to avoid having to
+			 * save & restore the whole vector context.
+			 */
+			if (test_thread_flag(TIF_USEDMSA))
+				write_msa_wr(wd, fpr, df);
+
+			preempt_enable();
+			break;
+
+		case msa_st_op:
+			if (!access_ok(VERIFY_WRITE, addr, sizeof(*fpr)))
+				goto sigbus;
+
+			/*
+			 * Update from the hardware register if it is in use by
+			 * the task in this quantum, in order to avoid having to
+			 * save & restore the whole vector context.
+			 */
+			preempt_disable();
+			if (test_thread_flag(TIF_USEDMSA))
+				read_msa_wr(wd, fpr, df);
+			preempt_enable();
+
+			res = __copy_to_user_inatomic(addr, fpr, sizeof(*fpr));
+			if (res)
+				goto fault;
+			break;
+
+		default:
+			goto sigbus;
+		}
+
+		compute_return_epc(regs);
+		break;
+
 #ifndef CONFIG_CPU_MIPSR6
 	/*
 	 * COP2 is available to implementor for application specific use.
@@ -2240,5 +2312,5 @@ static int __init debugfs_unaligned(void)
 		return -ENOMEM;
 	return 0;
 }
-__initcall(debugfs_unaligned);
+arch_initcall(debugfs_unaligned);
 #endif
diff --git a/arch/mips/kernel/uprobes.c b/arch/mips/kernel/uprobes.c
new file mode 100644
index 0000000..8452d93
--- /dev/null
+++ b/arch/mips/kernel/uprobes.c
@@ -0,0 +1,341 @@
+#include <linux/highmem.h>
+#include <linux/kdebug.h>
+#include <linux/types.h>
+#include <linux/notifier.h>
+#include <linux/sched.h>
+#include <linux/uprobes.h>
+
+#include <asm/branch.h>
+#include <asm/cpu-features.h>
+#include <asm/ptrace.h>
+#include <asm/inst.h>
+
+static inline int insn_has_delay_slot(const union mips_instruction insn)
+{
+	switch (insn.i_format.opcode) {
+	/*
+	 * jr and jalr are in r_format format.
+	 */
+	case spec_op:
+		switch (insn.r_format.func) {
+		case jalr_op:
+		case jr_op:
+			return 1;
+		}
+		break;
+
+	/*
+	 * This group contains:
+	 * bltz_op, bgez_op, bltzl_op, bgezl_op,
+	 * bltzal_op, bgezal_op, bltzall_op, bgezall_op.
+	 */
+	case bcond_op:
+		switch (insn.i_format.rt) {
+		case bltz_op:
+		case bltzl_op:
+		case bgez_op:
+		case bgezl_op:
+		case bltzal_op:
+		case bltzall_op:
+		case bgezal_op:
+		case bgezall_op:
+		case bposge32_op:
+			return 1;
+		}
+		break;
+
+	/*
+	 * These are unconditional and in j_format.
+	 */
+	case jal_op:
+	case j_op:
+	case beq_op:
+	case beql_op:
+	case bne_op:
+	case bnel_op:
+	case blez_op: /* not really i_format */
+	case blezl_op:
+	case bgtz_op:
+	case bgtzl_op:
+		return 1;
+
+	/*
+	 * And now the FPA/cp1 branch instructions.
+	 */
+	case cop1_op:
+#ifdef CONFIG_CPU_CAVIUM_OCTEON
+	case lwc2_op: /* This is bbit0 on Octeon */
+	case ldc2_op: /* This is bbit032 on Octeon */
+	case swc2_op: /* This is bbit1 on Octeon */
+	case sdc2_op: /* This is bbit132 on Octeon */
+#endif
+		return 1;
+	}
+
+	return 0;
+}
+
+/**
+ * arch_uprobe_analyze_insn - instruction analysis including validity and fixups.
+ * @mm: the probed address space.
+ * @arch_uprobe: the probepoint information.
+ * @addr: virtual address at which to install the probepoint
+ * Return 0 on success or a -ve number on error.
+ */
+int arch_uprobe_analyze_insn(struct arch_uprobe *aup,
+	struct mm_struct *mm, unsigned long addr)
+{
+	union mips_instruction inst;
+
+	/*
+	 * For the time being this also blocks attempts to use uprobes with
+	 * MIPS16 and microMIPS.
+	 */
+	if (addr & 0x03)
+		return -EINVAL;
+
+	inst.word = aup->insn[0];
+	aup->ixol[0] = aup->insn[insn_has_delay_slot(inst)];
+	aup->ixol[1] = UPROBE_BRK_UPROBE_XOL;		/* NOP  */
+
+	return 0;
+}
+
+/**
+ * is_trap_insn - check if the instruction is a trap variant
+ * @insn: instruction to be checked.
+ * Returns true if @insn is a trap variant.
+ *
+ * This definition overrides the weak definition in kernel/events/uprobes.c.
+ * and is needed for the case where an architecture has multiple trap
+ * instructions (like PowerPC or MIPS).  We treat BREAK just like the more
+ * modern conditional trap instructions.
+ */
+bool is_trap_insn(uprobe_opcode_t *insn)
+{
+	union mips_instruction inst;
+
+	inst.word = *insn;
+
+	switch (inst.i_format.opcode) {
+	case spec_op:
+		switch (inst.r_format.func) {
+		case break_op:
+		case teq_op:
+		case tge_op:
+		case tgeu_op:
+		case tlt_op:
+		case tltu_op:
+		case tne_op:
+			return 1;
+		}
+		break;
+
+	case bcond_op:	/* Yes, really ...  */
+		switch (inst.u_format.rt) {
+		case teqi_op:
+		case tgei_op:
+		case tgeiu_op:
+		case tlti_op:
+		case tltiu_op:
+		case tnei_op:
+			return 1;
+		}
+		break;
+	}
+
+	return 0;
+}
+
+#define UPROBE_TRAP_NR	ULONG_MAX
+
+/*
+ * arch_uprobe_pre_xol - prepare to execute out of line.
+ * @auprobe: the probepoint information.
+ * @regs: reflects the saved user state of current task.
+ */
+int arch_uprobe_pre_xol(struct arch_uprobe *aup, struct pt_regs *regs)
+{
+	struct uprobe_task *utask = current->utask;
+	union mips_instruction insn;
+
+	/*
+	 * Now find the EPC where to resume after the breakpoint has been
+	 * dealt with.  This may require emulation of a branch.
+	 */
+	aup->resume_epc = regs->cp0_epc + 4;
+	if (insn_has_delay_slot((union mips_instruction) aup->insn[0])) {
+		unsigned long epc;
+
+		epc = regs->cp0_epc;
+		__compute_return_epc_for_insn(regs, insn);
+		aup->resume_epc = regs->cp0_epc;
+	}
+
+	utask->autask.saved_trap_nr = current->thread.trap_nr;
+	current->thread.trap_nr = UPROBE_TRAP_NR;
+	regs->cp0_epc = current->utask->xol_vaddr;
+
+	return 0;
+}
+
+int arch_uprobe_post_xol(struct arch_uprobe *aup, struct pt_regs *regs)
+{
+	struct uprobe_task *utask = current->utask;
+
+	current->thread.trap_nr = utask->autask.saved_trap_nr;
+	regs->cp0_epc = aup->resume_epc;
+
+	return 0;
+}
+
+/*
+ * If xol insn itself traps and generates a signal(Say,
+ * SIGILL/SIGSEGV/etc), then detect the case where a singlestepped
+ * instruction jumps back to its own address. It is assumed that anything
+ * like do_page_fault/do_trap/etc sets thread.trap_nr != -1.
+ *
+ * arch_uprobe_pre_xol/arch_uprobe_post_xol save/restore thread.trap_nr,
+ * arch_uprobe_xol_was_trapped() simply checks that ->trap_nr is not equal to
+ * UPROBE_TRAP_NR == -1 set by arch_uprobe_pre_xol().
+ */
+bool arch_uprobe_xol_was_trapped(struct task_struct *tsk)
+{
+	if (tsk->thread.trap_nr != UPROBE_TRAP_NR)
+		return true;
+
+	return false;
+}
+
+int arch_uprobe_exception_notify(struct notifier_block *self,
+	unsigned long val, void *data)
+{
+	struct die_args *args = data;
+	struct pt_regs *regs = args->regs;
+
+	/* regs == NULL is a kernel bug */
+	if (WARN_ON(!regs))
+		return NOTIFY_DONE;
+
+	/* We are only interested in userspace traps */
+	if (!user_mode(regs))
+		return NOTIFY_DONE;
+
+	switch (val) {
+	case DIE_BREAK:
+		if (uprobe_pre_sstep_notifier(regs))
+			return NOTIFY_STOP;
+		break;
+	case DIE_UPROBE_XOL:
+		if (uprobe_post_sstep_notifier(regs))
+			return NOTIFY_STOP;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+/*
+ * This function gets called when XOL instruction either gets trapped or
+ * the thread has a fatal signal. Reset the instruction pointer to its
+ * probed address for the potential restart or for post mortem analysis.
+ */
+void arch_uprobe_abort_xol(struct arch_uprobe *aup,
+	struct pt_regs *regs)
+{
+	struct uprobe_task *utask = current->utask;
+
+	instruction_pointer_set(regs, utask->vaddr);
+}
+
+unsigned long arch_uretprobe_hijack_return_addr(
+	unsigned long trampoline_vaddr, struct pt_regs *regs)
+{
+	unsigned long ra;
+
+	ra = regs->regs[31];
+
+	/* Replace the return address with the trampoline address */
+	regs->regs[31] = ra;
+
+	return ra;
+}
+
+/**
+ * set_swbp - store breakpoint at a given address.
+ * @auprobe: arch specific probepoint information.
+ * @mm: the probed process address space.
+ * @vaddr: the virtual address to insert the opcode.
+ *
+ * For mm @mm, store the breakpoint instruction at @vaddr.
+ * Return 0 (success) or a negative errno.
+ *
+ * This version overrides the weak version in kernel/events/uprobes.c.
+ * It is required to handle MIPS16 and microMIPS.
+ */
+int __weak set_swbp(struct arch_uprobe *auprobe, struct mm_struct *mm,
+	unsigned long vaddr)
+{
+	return uprobe_write_opcode(mm, vaddr, UPROBE_SWBP_INSN);
+}
+
+/**
+ * set_orig_insn - Restore the original instruction.
+ * @mm: the probed process address space.
+ * @auprobe: arch specific probepoint information.
+ * @vaddr: the virtual address to insert the opcode.
+ *
+ * For mm @mm, restore the original opcode (opcode) at @vaddr.
+ * Return 0 (success) or a negative errno.
+ *
+ * This overrides the weak version in kernel/events/uprobes.c.
+ */
+int set_orig_insn(struct arch_uprobe *auprobe, struct mm_struct *mm,
+		 unsigned long vaddr)
+{
+	return uprobe_write_opcode(mm, vaddr,
+			*(uprobe_opcode_t *)&auprobe->orig_inst[0].word);
+}
+
+void __weak arch_uprobe_copy_ixol(struct page *page, unsigned long vaddr,
+				  void *src, unsigned long len)
+{
+	void *kaddr;
+
+	/* Initialize the slot */
+	kaddr = kmap_atomic(page);
+	memcpy(kaddr + (vaddr & ~PAGE_MASK), src, len);
+	kunmap_atomic(kaddr);
+
+	/*
+	 * The MIPS version of flush_icache_range will operate safely on
+	 * user space addresses and more importantly, it doesn't require a
+	 * VMA argument.
+	 */
+	flush_icache_range(vaddr, vaddr + len);
+}
+
+/**
+ * uprobe_get_swbp_addr - compute address of swbp given post-swbp regs
+ * @regs: Reflects the saved state of the task after it has hit a breakpoint
+ * instruction.
+ * Return the address of the breakpoint instruction.
+ *
+ * This overrides the weak version in kernel/events/uprobes.c.
+ */
+unsigned long uprobe_get_swbp_addr(struct pt_regs *regs)
+{
+	return instruction_pointer(regs);
+}
+
+/*
+ * See if the instruction can be emulated.
+ * Returns true if instruction was emulated, false otherwise.
+ *
+ * For now we always emulate so this function just returns 0.
+ */
+bool arch_uprobe_skip_sstep(struct arch_uprobe *auprobe, struct pt_regs *regs)
+{
+	return 0;
+}
diff --git a/arch/mips/kernel/vpe.c b/arch/mips/kernel/vpe.c
index 11da314..9067b65 100644
--- a/arch/mips/kernel/vpe.c
+++ b/arch/mips/kernel/vpe.c
@@ -817,6 +817,7 @@ static int vpe_open(struct inode *inode, struct file *filp)
 
 static int vpe_release(struct inode *inode, struct file *filp)
 {
+#if defined(CONFIG_MIPS_VPE_LOADER_MT) || defined(CONFIG_MIPS_VPE_LOADER_CMP)
 	struct vpe *v;
 	Elf_Ehdr *hdr;
 	int ret = 0;
@@ -827,7 +828,7 @@ static int vpe_release(struct inode *inode, struct file *filp)
 
 	hdr = (Elf_Ehdr *) v->pbuffer;
 	if (memcmp(hdr->e_ident, ELFMAG, SELFMAG) == 0) {
-		if ((vpe_elfload(v) >= 0) && vpe_run) {
+		if (vpe_elfload(v) >= 0) {
 			vpe_run(v);
 		} else {
 			pr_warn("VPE loader: ELF load failed.\n");
@@ -850,6 +851,10 @@ static int vpe_release(struct inode *inode, struct file *filp)
 	v->plen = 0;
 
 	return ret;
+#else
+	pr_warn("VPE loader: ELF load failed.\n");
+	return -ENOEXEC;
+#endif
 }
 
 static ssize_t vpe_write(struct file *file, const char __user *buffer,