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

Pull MIPS updates from Ralf Baechle:
 "These are the highlists of the main MIPS pull request for 4.4:

   - Add latencytop support
   - Support appended DTBs
   - VDSO support and initially use it for gettimeofday.
   - Drop the .MIPS.abiflags and ELF NOTE sections from vmlinux
   - Support for the 5KE, an internal test core.
   - Switch all MIPS platfroms to libata drivers.
   - Improved support, cleanups for ralink and Lantiq platforms.
   - Support for the new xilfpga platform.
   - A number of DTB improvments for BMIPS.
   - Improved support for CM and CPS.
   - Minor JZ4740 and BCM47xx enhancements"

* 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus: (120 commits)
  MIPS: idle: add case for CPU_5KE
  MIPS: Octeon: Support APPENDED_DTB
  MIPS: vmlinux: create a section for appended DTB
  MIPS: Clean up compat_siginfo_t
  MIPS: Fix PAGE_MASK definition
  MIPS: BMIPS: Enable GZIP ramdisk and timed printks
  MIPS: Add xilfpga defconfig
  MIPS: xilfpga: Add mipsfpga platform code
  MIPS: xilfpga: Add xilfpga device tree files.
  dt-bindings: MIPS: Document xilfpga bindings and boot style
  MIPS: Make MIPS_CMDLINE_DTB default
  MIPS: Make the kernel arguments from dtb available
  MIPS: Use USE_OF as the guard for appended dtb
  MIPS: BCM63XX: Use pr_* instead of printk
  MIPS: Loongson: Cleanup CONFIG_LOONGSON_SUSPEND.
  MIPS: lantiq: Disable xbar fpi burst mode
  MIPS: lantiq: Force the crossbar to big endian
  MIPS: lantiq: Initialize the USB core on boot
  MIPS: lantiq: Return correct value for fpi clock on ar9
  MIPS: ralink: Add missing clock on rt305x
  ...

22 files changed, 618 insertions, 173 deletions

diff --git a/arch/mips/kernel/Makefile b/arch/mips/kernel/Makefile
index d982be1ea1c3..68e2b7db9348 100644
--- a/arch/mips/kernel/Makefile
+++ b/arch/mips/kernel/Makefile
@@ -51,6 +51,7 @@ obj-$(CONFIG_MIPS_MT_FPAFF)	+= mips-mt-fpaff.o
 obj-$(CONFIG_MIPS_MT_SMP)	+= smp-mt.o
 obj-$(CONFIG_MIPS_CMP)		+= smp-cmp.o
 obj-$(CONFIG_MIPS_CPS)		+= smp-cps.o cps-vec.o
+obj-$(CONFIG_MIPS_CPS_NS16550)	+= cps-vec-ns16550.o
 obj-$(CONFIG_MIPS_GIC_IPI)	+= smp-gic.o
 obj-$(CONFIG_MIPS_SPRAM)	+= spram.o
 
diff --git a/arch/mips/kernel/cps-vec-ns16550.S b/arch/mips/kernel/cps-vec-ns16550.S
new file mode 100644
index 000000000000..6d246ad05638
--- /dev/null
+++ b/arch/mips/kernel/cps-vec-ns16550.S
@@ -0,0 +1,202 @@
+/*
+ * Copyright (C) 2015 Imagination Technologies
+ * Author: Paul Burton <paul.burton@imgtec.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the
+ * Free Software Foundation;  either version 2 of the  License, or (at your
+ * option) any later version.
+ */
+
+#include <asm/addrspace.h>
+#include <asm/asm.h>
+#include <asm/asm-offsets.h>
+#include <asm/mipsregs.h>
+#include <asm/regdef.h>
+#include <linux/serial_reg.h>
+
+#define UART_TX_OFS	(UART_TX << CONFIG_MIPS_CPS_NS16550_SHIFT)
+#define UART_LSR_OFS	(UART_LSR << CONFIG_MIPS_CPS_NS16550_SHIFT)
+
+/**
+ * _mips_cps_putc() - write a character to the UART
+ * @a0: ASCII character to write
+ * @t9: UART base address
+ */
+LEAF(_mips_cps_putc)
+1:	lw		t0, UART_LSR_OFS(t9)
+	andi		t0, t0, UART_LSR_TEMT
+	beqz		t0, 1b
+	sb		a0, UART_TX_OFS(t9)
+	jr		ra
+	END(_mips_cps_putc)
+
+/**
+ * _mips_cps_puts() - write a string to the UART
+ * @a0: pointer to NULL-terminated ASCII string
+ * @t9: UART base address
+ *
+ * Write a null-terminated ASCII string to the UART.
+ */
+NESTED(_mips_cps_puts, 0, ra)
+	move		s7, ra
+	move		s6, a0
+
+1:	lb		a0, 0(s6)
+	beqz		a0, 2f
+	jal		_mips_cps_putc
+	PTR_ADDIU	s6, s6, 1
+	b		1b
+
+2:	jr		s7
+	END(_mips_cps_puts)
+
+/**
+ * _mips_cps_putx4 - write a 4b hex value to the UART
+ * @a0: the 4b value to write to the UART
+ * @t9: UART base address
+ *
+ * Write a single hexadecimal character to the UART.
+ */
+NESTED(_mips_cps_putx4, 0, ra)
+	andi		a0, a0, 0xf
+	li		t0, '0'
+	blt		a0, 10, 1f
+	li		t0, 'a'
+	addiu		a0, a0, -10
+1:	addu		a0, a0, t0
+	b		_mips_cps_putc
+	END(_mips_cps_putx4)
+
+/**
+ * _mips_cps_putx8 - write an 8b hex value to the UART
+ * @a0: the 8b value to write to the UART
+ * @t9: UART base address
+ *
+ * Write an 8 bit value (ie. 2 hexadecimal characters) to the UART.
+ */
+NESTED(_mips_cps_putx8, 0, ra)
+	move		s3, ra
+	move		s2, a0
+	srl		a0, a0, 4
+	jal		_mips_cps_putx4
+	move		a0, s2
+	move		ra, s3
+	b		_mips_cps_putx4
+	END(_mips_cps_putx8)
+
+/**
+ * _mips_cps_putx16 - write a 16b hex value to the UART
+ * @a0: the 16b value to write to the UART
+ * @t9: UART base address
+ *
+ * Write a 16 bit value (ie. 4 hexadecimal characters) to the UART.
+ */
+NESTED(_mips_cps_putx16, 0, ra)
+	move		s5, ra
+	move		s4, a0
+	srl		a0, a0, 8
+	jal		_mips_cps_putx8
+	move		a0, s4
+	move		ra, s5
+	b		_mips_cps_putx8
+	END(_mips_cps_putx16)
+
+/**
+ * _mips_cps_putx32 - write a 32b hex value to the UART
+ * @a0: the 32b value to write to the UART
+ * @t9: UART base address
+ *
+ * Write a 32 bit value (ie. 8 hexadecimal characters) to the UART.
+ */
+NESTED(_mips_cps_putx32, 0, ra)
+	move		s7, ra
+	move		s6, a0
+	srl		a0, a0, 16
+	jal		_mips_cps_putx16
+	move		a0, s6
+	move		ra, s7
+	b		_mips_cps_putx16
+	END(_mips_cps_putx32)
+
+#ifdef CONFIG_64BIT
+
+/**
+ * _mips_cps_putx64 - write a 64b hex value to the UART
+ * @a0: the 64b value to write to the UART
+ * @t9: UART base address
+ *
+ * Write a 64 bit value (ie. 16 hexadecimal characters) to the UART.
+ */
+NESTED(_mips_cps_putx64, 0, ra)
+	move		sp, ra
+	move		s8, a0
+	dsrl32		a0, a0, 0
+	jal		_mips_cps_putx32
+	move		a0, s8
+	move		ra, sp
+	b		_mips_cps_putx32
+	END(_mips_cps_putx64)
+
+#define _mips_cps_putxlong _mips_cps_putx64
+
+#else /* !CONFIG_64BIT */
+
+#define _mips_cps_putxlong _mips_cps_putx32
+
+#endif /* !CONFIG_64BIT */
+
+/**
+ * mips_cps_bev_dump() - dump relevant exception state to UART
+ * @a0: pointer to NULL-terminated ASCII string naming the exception
+ *
+ * Write information that may be useful in debugging an exception to the
+ * UART configured by CONFIG_MIPS_CPS_NS16550_*. As this BEV exception
+ * will only be run if something goes horribly wrong very early during
+ * the bringup of a core and it is very likely to be unsafe to perform
+ * memory accesses at that point (cache state indeterminate, EVA may not
+ * be configured, coherence may be disabled) let alone have a stack,
+ * this is all written in assembly using only registers & unmapped
+ * uncached access to the UART registers.
+ */
+LEAF(mips_cps_bev_dump)
+	move		s0, ra
+	move		s1, a0
+
+	li		t9, CKSEG1ADDR(CONFIG_MIPS_CPS_NS16550_BASE)
+
+	PTR_LA		a0, str_newline
+	jal		_mips_cps_puts
+	PTR_LA		a0, str_bev
+	jal		_mips_cps_puts
+	move		a0, s1
+	jal		_mips_cps_puts
+	PTR_LA		a0, str_newline
+	jal		_mips_cps_puts
+	PTR_LA		a0, str_newline
+	jal		_mips_cps_puts
+
+#define DUMP_COP0_REG(reg, name, sz, _mfc0)		\
+	PTR_LA		a0, 8f;				\
+	jal		_mips_cps_puts;			\
+	_mfc0		a0, reg;			\
+	jal		_mips_cps_putx##sz;		\
+	PTR_LA		a0, str_newline;		\
+	jal		_mips_cps_puts;			\
+	TEXT(name)
+
+	DUMP_COP0_REG(CP0_CAUSE,    "Cause:    0x", 32, mfc0)
+	DUMP_COP0_REG(CP0_STATUS,   "Status:   0x", 32, mfc0)
+	DUMP_COP0_REG(CP0_EBASE,    "EBase:    0x", long, MFC0)
+	DUMP_COP0_REG(CP0_BADVADDR, "BadVAddr: 0x", long, MFC0)
+	DUMP_COP0_REG(CP0_BADINSTR, "BadInstr: 0x", 32, mfc0)
+
+	PTR_LA		a0, str_newline
+	jal		_mips_cps_puts
+	jr		s0
+	END(mips_cps_bev_dump)
+
+.pushsection	.data
+str_bev: .asciiz "BEV Exception: "
+str_newline: .asciiz "\r\n"
+.popsection
diff --git a/arch/mips/kernel/cps-vec.S b/arch/mips/kernel/cps-vec.S
index 209ded16806b..8fd5a276cad2 100644
--- a/arch/mips/kernel/cps-vec.S
+++ b/arch/mips/kernel/cps-vec.S
@@ -25,14 +25,32 @@
 
 .set noreorder
 
+#ifdef CONFIG_64BIT
+# define STATUS_BITDEPS		ST0_KX
+#else
+# define STATUS_BITDEPS		0
+#endif
+
+#ifdef CONFIG_MIPS_CPS_NS16550
+
+#define DUMP_EXCEP(name)		\
+	PTR_LA	a0, 8f;			\
+	jal	mips_cps_bev_dump;	\
+	 nop;				\
+	TEXT(name)
+
+#else /* !CONFIG_MIPS_CPS_NS16550 */
+
+#define DUMP_EXCEP(name)
+
+#endif /* !CONFIG_MIPS_CPS_NS16550 */
+
 	/*
 	 * Set dest to non-zero if the core supports the MT ASE, else zero. If
 	 * MT is not supported then branch to nomt.
 	 */
 	.macro	has_mt	dest, nomt
-	mfc0	\dest, CP0_CONFIG
-	bgez	\dest, \nomt
-	 mfc0	\dest, CP0_CONFIG, 1
+	mfc0	\dest, CP0_CONFIG, 1
 	bgez	\dest, \nomt
 	 mfc0	\dest, CP0_CONFIG, 2
 	bgez	\dest, \nomt
@@ -47,11 +65,9 @@
 
 LEAF(mips_cps_core_entry)
 	/*
-	 * These first 12 bytes will be patched by cps_smp_setup to load the
-	 * base address of the CM GCRs into register v1 and the CCA to use into
-	 * register s0.
+	 * These first 4 bytes will be patched by cps_smp_setup to load the
+	 * CCA to use into register s0.
 	 */
-	.quad	0
 	.word	0
 
 	/* Check whether we're here due to an NMI */
@@ -71,7 +87,7 @@ not_nmi:
 	mtc0	t0, CP0_CAUSE
 
 	/* Setup Status */
-	li	t0, ST0_CU1 | ST0_CU0
+	li	t0, ST0_CU1 | ST0_CU0 | ST0_BEV | STATUS_BITDEPS
 	mtc0	t0, CP0_STATUS
 
 	/*
@@ -151,6 +167,12 @@ dcache_done:
 	mtc0	t0, CP0_CONFIG
 	ehb
 
+	/* Calculate an uncached address for the CM GCRs */
+	MFC0	v1, CP0_CMGCRBASE
+	PTR_SLL	v1, v1, 4
+	PTR_LI	t0, UNCAC_BASE
+	PTR_ADDU v1, v1, t0
+
 	/* Enter the coherent domain */
 	li	t0, 0xff
 	sw	t0, GCR_CL_COHERENCE_OFS(v1)
@@ -188,36 +210,42 @@ dcache_done:
 
 .org 0x200
 LEAF(excep_tlbfill)
+	DUMP_EXCEP("TLB Fill")
 	b	.
 	 nop
 	END(excep_tlbfill)
 
 .org 0x280
 LEAF(excep_xtlbfill)
+	DUMP_EXCEP("XTLB Fill")
 	b	.
 	 nop
 	END(excep_xtlbfill)
 
 .org 0x300
 LEAF(excep_cache)
+	DUMP_EXCEP("Cache")
 	b	.
 	 nop
 	END(excep_cache)
 
 .org 0x380
 LEAF(excep_genex)
+	DUMP_EXCEP("General")
 	b	.
 	 nop
 	END(excep_genex)
 
 .org 0x400
 LEAF(excep_intex)
+	DUMP_EXCEP("Interrupt")
 	b	.
 	 nop
 	END(excep_intex)
 
 .org 0x480
 LEAF(excep_ejtag)
+	DUMP_EXCEP("EJTAG")
 	PTR_LA	k0, ejtag_debug_handler
 	jr	k0
 	 nop
diff --git a/arch/mips/kernel/cpu-probe.c b/arch/mips/kernel/cpu-probe.c
index 09a51d091941..6b9064499bd3 100644
--- a/arch/mips/kernel/cpu-probe.c
+++ b/arch/mips/kernel/cpu-probe.c
@@ -536,8 +536,7 @@ static inline unsigned int decode_config3(struct cpuinfo_mips *c)
 		c->options |= MIPS_CPU_SEGMENTS;
 	if (config3 & MIPS_CONF3_MSA)
 		c->ases |= MIPS_ASE_MSA;
-	/* Only tested on 32-bit cores */
-	if ((config3 & MIPS_CONF3_PW) && config_enabled(CONFIG_32BIT)) {
+	if (config3 & MIPS_CONF3_PW) {
 		c->htw_seq = 0;
 		c->options |= MIPS_CPU_HTW;
 	}
diff --git a/arch/mips/kernel/csrc-r4k.c b/arch/mips/kernel/csrc-r4k.c
index e5ed7ada1433..1f910563fdf6 100644
--- a/arch/mips/kernel/csrc-r4k.c
+++ b/arch/mips/kernel/csrc-r4k.c
@@ -28,6 +28,43 @@ static u64 notrace r4k_read_sched_clock(void)
 	return read_c0_count();
 }
 
+static inline unsigned int rdhwr_count(void)
+{
+	unsigned int count;
+
+	__asm__ __volatile__(
+	"	.set push\n"
+	"	.set mips32r2\n"
+	"	rdhwr	%0, $2\n"
+	"	.set pop\n"
+	: "=r" (count));
+
+	return count;
+}
+
+static bool rdhwr_count_usable(void)
+{
+	unsigned int prev, curr, i;
+
+	/*
+	 * Older QEMUs have a broken implementation of RDHWR for the CP0 count
+	 * which always returns a constant value. Try to identify this and don't
+	 * use it in the VDSO if it is broken. This workaround can be removed
+	 * once the fix has been in QEMU stable for a reasonable amount of time.
+	 */
+	for (i = 0, prev = rdhwr_count(); i < 100; i++) {
+		curr = rdhwr_count();
+
+		if (curr != prev)
+			return true;
+
+		prev = curr;
+	}
+
+	pr_warn("Not using R4K clocksource in VDSO due to broken RDHWR\n");
+	return false;
+}
+
 int __init init_r4k_clocksource(void)
 {
 	if (!cpu_has_counter || !mips_hpt_frequency)
@@ -36,6 +73,13 @@ int __init init_r4k_clocksource(void)
 	/* Calculate a somewhat reasonable rating value */
 	clocksource_mips.rating = 200 + mips_hpt_frequency / 10000000;
 
+	/*
+	 * R2 onwards makes the count accessible to user mode so it can be used
+	 * by the VDSO (HWREna is configured by configure_hwrena()).
+	 */
+	if (cpu_has_mips_r2_r6 && rdhwr_count_usable())
+		clocksource_mips.archdata.vdso_clock_mode = VDSO_CLOCK_R4K;
+
 	clocksource_register_hz(&clocksource_mips, mips_hpt_frequency);
 
 	sched_clock_register(r4k_read_sched_clock, 32, mips_hpt_frequency);
diff --git a/arch/mips/kernel/idle.c b/arch/mips/kernel/idle.c
index ab1478d5a4db..46794d64c0bf 100644
--- a/arch/mips/kernel/idle.c
+++ b/arch/mips/kernel/idle.c
@@ -134,6 +134,16 @@ void __init check_wait(void)
 		return;
 	}
 
+	/*
+	 * MIPSr6 specifies that masked interrupts should unblock an executing
+	 * wait instruction, and thus that it is safe for us to use
+	 * r4k_wait_irqoff. Yippee!
+	 */
+	if (cpu_has_mips_r6) {
+		cpu_wait = r4k_wait_irqoff;
+		return;
+	}
+
 	switch (current_cpu_type()) {
 	case CPU_R3081:
 	case CPU_R3081E:
@@ -155,12 +165,12 @@ void __init check_wait(void)
 	case CPU_4KEC:
 	case CPU_4KSC:
 	case CPU_5KC:
+	case CPU_5KE:
 	case CPU_25KF:
 	case CPU_PR4450:
 	case CPU_BMIPS3300:
 	case CPU_BMIPS4350:
 	case CPU_BMIPS4380:
-	case CPU_BMIPS5000:
 	case CPU_CAVIUM_OCTEON:
 	case CPU_CAVIUM_OCTEON_PLUS:
 	case CPU_CAVIUM_OCTEON2:
@@ -171,7 +181,9 @@ void __init check_wait(void)
 	case CPU_XLP:
 		cpu_wait = r4k_wait;
 		break;
-
+	case CPU_BMIPS5000:
+		cpu_wait = r4k_wait_irqoff;
+		break;
 	case CPU_RM7000:
 		cpu_wait = rm7k_wait_irqoff;
 		break;
@@ -196,7 +208,6 @@ 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 b8ceee576cdf..1448c1f43d4e 100644
--- a/arch/mips/kernel/mips-cm.c
+++ b/arch/mips/kernel/mips-cm.c
@@ -9,6 +9,8 @@
  */
 
 #include <linux/errno.h>
+#include <linux/percpu.h>
+#include <linux/spinlock.h>
 
 #include <asm/mips-cm.h>
 #include <asm/mipsregs.h>
@@ -136,6 +138,9 @@ static char *cm3_causes[32] = {
 	"0x19", "0x1a", "0x1b", "0x1c", "0x1d", "0x1e", "0x1f"
 };
 
+static DEFINE_PER_CPU_ALIGNED(spinlock_t, cm_core_lock);
+static DEFINE_PER_CPU_ALIGNED(unsigned long, cm_core_lock_flags);
+
 phys_addr_t __mips_cm_phys_base(void)
 {
 	u32 config3 = read_c0_config3();
@@ -200,6 +205,7 @@ int mips_cm_probe(void)
 {
 	phys_addr_t addr;
 	u32 base_reg;
+	unsigned cpu;
 
 	/*
 	 * No need to probe again if we have already been
@@ -247,38 +253,70 @@ int mips_cm_probe(void)
 	/* determine register width for this CM */
 	mips_cm_is64 = config_enabled(CONFIG_64BIT) && (mips_cm_revision() >= CM_REV_CM3);
 
+	for_each_possible_cpu(cpu)
+		spin_lock_init(&per_cpu(cm_core_lock, cpu));
+
 	return 0;
 }
 
-void mips_cm_error_report(void)
+void mips_cm_lock_other(unsigned int core, unsigned int vp)
 {
-	unsigned long revision = mips_cm_revision();
+	unsigned curr_core;
+	u32 val;
+
+	preempt_disable();
+	curr_core = current_cpu_data.core;
+	spin_lock_irqsave(&per_cpu(cm_core_lock, curr_core),
+			  per_cpu(cm_core_lock_flags, curr_core));
+
+	if (mips_cm_revision() >= CM_REV_CM3) {
+		val = core << CM3_GCR_Cx_OTHER_CORE_SHF;
+		val |= vp << CM3_GCR_Cx_OTHER_VP_SHF;
+	} else {
+		BUG_ON(vp != 0);
+		val = core << CM_GCR_Cx_OTHER_CORENUM_SHF;
+	}
+
+	write_gcr_cl_other(val);
+
 	/*
-	 * 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.
+	 * Ensure the core-other region reflects the appropriate core &
+	 * VP before any accesses to it occur.
 	 */
-	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();
+	mb();
+}
+
+void mips_cm_unlock_other(void)
+{
+	unsigned curr_core = current_cpu_data.core;
+
+	spin_unlock_irqrestore(&per_cpu(cm_core_lock, curr_core),
+			       per_cpu(cm_core_lock_flags, curr_core));
+	preempt_enable();
+}
+
+void mips_cm_error_report(void)
+{
+	u64 cm_error, cm_addr, cm_other;
+	unsigned long revision;
 	int ocause, cause;
 	char buf[256];
 
 	if (!mips_cm_present())
 		return;
 
-	cause = cm_error >> cm_error_cause_sft;
+	revision = mips_cm_revision();
 
-	if (!cause)
-		/* All good */
-		return;
-
-	ocause = cm_other >> CM_GCR_ERROR_MULT_ERR2ND_SHF;
 	if (revision < CM_REV_CM3) { /* CM2 */
+		cm_error = read_gcr_error_cause();
+		cm_addr = read_gcr_error_addr();
+		cm_other = read_gcr_error_mult();
+		cause = cm_error >> CM_GCR_ERROR_CAUSE_ERRTYPE_SHF;
+		ocause = cm_other >> CM_GCR_ERROR_MULT_ERR2ND_SHF;
+
+		if (!cause)
+			return;
+
 		if (cause < 16) {
 			unsigned long cca_bits = (cm_error >> 15) & 7;
 			unsigned long tr_bits = (cm_error >> 12) & 7;
@@ -310,18 +348,30 @@ void mips_cm_error_report(void)
 		}
 			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]);
+		pr_err("CM_ADDR =%08llx\n", cm_addr);
+		pr_err("CM_OTHER=%08llx %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;
+		ulong core_id_bits, vp_id_bits, cmd_bits, cmd_group_bits;
+		ulong cm3_cca_bits, mcp_bits, cm3_tr_bits, sched_bit;
+
+		cm_error = read64_gcr_error_cause();
+		cm_addr = read64_gcr_error_addr();
+		cm_other = read64_gcr_error_mult();
+		cause = cm_error >> CM3_GCR_ERROR_CAUSE_ERRTYPE_SHF;
+		ocause = cm_other >> CM_GCR_ERROR_MULT_ERR2ND_SHF;
+
+		if (!cause)
+			return;
+
+		/* Used by cause == {1,2,3} */
+		core_id_bits = (cm_error >> 22) & 0xf;
+		vp_id_bits = (cm_error >> 18) & 0xf;
+		cmd_bits = (cm_error >> 14) & 0xf;
+		cmd_group_bits = (cm_error >> 11) & 0xf;
+		cm3_cca_bits = (cm_error >> 8) & 7;
+		mcp_bits = (cm_error >> 5) & 0xf;
+		cm3_tr_bits = (cm_error >> 1) & 0xf;
+		sched_bit = cm_error & 0x1;
 
 		if (cause == 1 || cause == 3) { /* Tag ECC */
 			unsigned long tag_ecc = (cm_error >> 57) & 0x1;
@@ -363,12 +413,14 @@ void mips_cm_error_report(void)
 				 cm3_cmd_group[cmd_group_bits],
 				 cm3_cca_bits, 1 << mcp_bits,
 				 cm3_tr[cm3_tr_bits], sched_bit);
+		} else {
+			buf[0] = 0;
 		}
 
 		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]);
+		pr_err("CM_ADDR =%llx\n", cm_addr);
+		pr_err("CM_OTHER=%llx %s\n", cm_other, cm3_causes[ocause]);
 	}
 
 	/* reprime cause register */
diff --git a/arch/mips/kernel/mips-cpc.c b/arch/mips/kernel/mips-cpc.c
index 8af4d627b68b..566b8d2c092c 100644
--- a/arch/mips/kernel/mips-cpc.c
+++ b/arch/mips/kernel/mips-cpc.c
@@ -76,6 +76,12 @@ void mips_cpc_lock_other(unsigned int core)
 	spin_lock_irqsave(&per_cpu(cpc_core_lock, curr_core),
 			  per_cpu(cpc_core_lock_flags, curr_core));
 	write_cpc_cl_other(core << CPC_Cx_OTHER_CORENUM_SHF);
+
+	/*
+	 * Ensure the core-other region reflects the appropriate core &
+	 * VP before any accesses to it occur.
+	 */
+	mb();
 }
 
 void mips_cpc_unlock_other(void)
diff --git a/arch/mips/kernel/mips-r2-to-r6-emul.c b/arch/mips/kernel/mips-r2-to-r6-emul.c
index f2977f00911b..1f5aac7f9ec3 100644
--- a/arch/mips/kernel/mips-r2-to-r6-emul.c
+++ b/arch/mips/kernel/mips-r2-to-r6-emul.c
@@ -22,6 +22,7 @@
 #include <asm/asm.h>
 #include <asm/branch.h>
 #include <asm/break.h>
+#include <asm/debug.h>
 #include <asm/fpu.h>
 #include <asm/fpu_emulator.h>
 #include <asm/inst.h>
@@ -2363,7 +2364,6 @@ static const struct file_operations mipsr2_clear_fops = {
 
 static int __init mipsr2_init_debugfs(void)
 {
-	extern struct dentry	*mips_debugfs_dir;
 	struct dentry		*mipsr2_emul;
 
 	if (!mips_debugfs_dir)
diff --git a/arch/mips/kernel/segment.c b/arch/mips/kernel/segment.c
index 076ead2a9859..87bc74a5a518 100644
--- a/arch/mips/kernel/segment.c
+++ b/arch/mips/kernel/segment.c
@@ -10,6 +10,7 @@
 #include <linux/debugfs.h>
 #include <linux/seq_file.h>
 #include <asm/cpu.h>
+#include <asm/debug.h>
 #include <asm/mipsregs.h>
 
 static void build_segment_config(char *str, unsigned int cfg)
@@ -91,7 +92,6 @@ static const struct file_operations segments_fops = {
 
 static int __init segments_info(void)
 {
-	extern struct dentry *mips_debugfs_dir;
 	struct dentry *segments;
 
 	if (cpu_has_segments) {
diff --git a/arch/mips/kernel/setup.c b/arch/mips/kernel/setup.c
index 479515109e5b..66aac55df349 100644
--- a/arch/mips/kernel/setup.c
+++ b/arch/mips/kernel/setup.c
@@ -33,11 +33,16 @@
 #include <asm/cache.h>
 #include <asm/cdmm.h>
 #include <asm/cpu.h>
+#include <asm/debug.h>
 #include <asm/sections.h>
 #include <asm/setup.h>
 #include <asm/smp-ops.h>
 #include <asm/prom.h>
 
+#ifdef CONFIG_MIPS_ELF_APPENDED_DTB
+const char __section(.appended_dtb) __appended_dtb[0x100000];
+#endif /* CONFIG_MIPS_ELF_APPENDED_DTB */
+
 struct cpuinfo_mips cpu_data[NR_CPUS] __read_mostly;
 
 EXPORT_SYMBOL(cpu_data);
@@ -616,6 +621,10 @@ static void __init request_crashkernel(struct resource *res)
 }
 #endif /* !defined(CONFIG_KEXEC)  */
 
+#define USE_PROM_CMDLINE	IS_ENABLED(CONFIG_MIPS_CMDLINE_FROM_BOOTLOADER)
+#define USE_DTB_CMDLINE		IS_ENABLED(CONFIG_MIPS_CMDLINE_FROM_DTB)
+#define EXTEND_WITH_PROM	IS_ENABLED(CONFIG_MIPS_CMDLINE_EXTEND)
+
 static void __init arch_mem_init(char **cmdline_p)
 {
 	struct memblock_region *reg;
@@ -640,18 +649,24 @@ static void __init arch_mem_init(char **cmdline_p)
 	pr_info("Determined physical RAM map:\n");
 	print_memory_map();
 
-#ifdef CONFIG_CMDLINE_BOOL
-#ifdef CONFIG_CMDLINE_OVERRIDE
+#if defined(CONFIG_CMDLINE_BOOL) && defined(CONFIG_CMDLINE_OVERRIDE)
 	strlcpy(boot_command_line, builtin_cmdline, COMMAND_LINE_SIZE);
 #else
+	if ((USE_PROM_CMDLINE && arcs_cmdline[0]) ||
+	    (USE_DTB_CMDLINE && !boot_command_line[0]))
+		strlcpy(boot_command_line, arcs_cmdline, COMMAND_LINE_SIZE);
+
+	if (EXTEND_WITH_PROM && arcs_cmdline[0]) {
+		strlcat(boot_command_line, " ", COMMAND_LINE_SIZE);
+		strlcat(boot_command_line, arcs_cmdline, COMMAND_LINE_SIZE);
+	}
+
+#if defined(CONFIG_CMDLINE_BOOL)
 	if (builtin_cmdline[0]) {
-		strlcat(arcs_cmdline, " ", COMMAND_LINE_SIZE);
-		strlcat(arcs_cmdline, builtin_cmdline, COMMAND_LINE_SIZE);
+		strlcat(boot_command_line, " ", COMMAND_LINE_SIZE);
+		strlcat(boot_command_line, builtin_cmdline, COMMAND_LINE_SIZE);
 	}
-	strlcpy(boot_command_line, arcs_cmdline, COMMAND_LINE_SIZE);
 #endif
-#else
-	strlcpy(boot_command_line, arcs_cmdline, COMMAND_LINE_SIZE);
 #endif
 	strlcpy(command_line, boot_command_line, COMMAND_LINE_SIZE);
 
diff --git a/arch/mips/kernel/signal.c b/arch/mips/kernel/signal.c
index 2fec67bfc457..bf792e2839a6 100644
--- a/arch/mips/kernel/signal.c
+++ b/arch/mips/kernel/signal.c
@@ -36,7 +36,6 @@
 #include <asm/ucontext.h>
 #include <asm/cpu-features.h>
 #include <asm/war.h>
-#include <asm/vdso.h>
 #include <asm/dsp.h>
 #include <asm/inst.h>
 #include <asm/msa.h>
@@ -752,16 +751,15 @@ static int setup_rt_frame(void *sig_return, struct ksignal *ksig,
 struct mips_abi mips_abi = {
 #ifdef CONFIG_TRAD_SIGNALS
 	.setup_frame	= setup_frame,
-	.signal_return_offset = offsetof(struct mips_vdso, signal_trampoline),
 #endif
 	.setup_rt_frame = setup_rt_frame,
-	.rt_signal_return_offset =
-		offsetof(struct mips_vdso, rt_signal_trampoline),
 	.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),
+
+	.vdso		= &vdso_image,
 };
 
 static void handle_signal(struct ksignal *ksig, struct pt_regs *regs)
@@ -801,11 +799,11 @@ static void handle_signal(struct ksignal *ksig, struct pt_regs *regs)
 	}
 
 	if (sig_uses_siginfo(&ksig->ka))
-		ret = abi->setup_rt_frame(vdso + abi->rt_signal_return_offset,
+		ret = abi->setup_rt_frame(vdso + abi->vdso->off_rt_sigreturn,
 					  ksig, regs, oldset);
 	else
-		ret = abi->setup_frame(vdso + abi->signal_return_offset, ksig,
-				       regs, oldset);
+		ret = abi->setup_frame(vdso + abi->vdso->off_sigreturn,
+				       ksig, regs, oldset);
 
 	signal_setup_done(ret, ksig, 0);
 }
diff --git a/arch/mips/kernel/signal32.c b/arch/mips/kernel/signal32.c
index f7e89524e316..4909639aa35b 100644
--- a/arch/mips/kernel/signal32.c
+++ b/arch/mips/kernel/signal32.c
@@ -31,7 +31,6 @@
 #include <asm/ucontext.h>
 #include <asm/fpu.h>
 #include <asm/war.h>
-#include <asm/vdso.h>
 #include <asm/dsp.h>
 
 #include "signal-common.h"
@@ -406,14 +405,12 @@ static int setup_rt_frame_32(void *sig_return, struct ksignal *ksig,
  */
 struct mips_abi mips_abi_32 = {
 	.setup_frame	= setup_frame_32,
-	.signal_return_offset =
-		offsetof(struct mips_vdso, o32_signal_trampoline),
 	.setup_rt_frame = setup_rt_frame_32,
-	.rt_signal_return_offset =
-		offsetof(struct mips_vdso, o32_rt_signal_trampoline),
 	.restart	= __NR_O32_restart_syscall,
 
 	.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),
+
+	.vdso		= &vdso_image_o32,
 };
diff --git a/arch/mips/kernel/signal_n32.c b/arch/mips/kernel/signal_n32.c
index 0d017fdcaf07..a7bc38430500 100644
--- a/arch/mips/kernel/signal_n32.c
+++ b/arch/mips/kernel/signal_n32.c
@@ -38,7 +38,6 @@
 #include <asm/fpu.h>
 #include <asm/cpu-features.h>
 #include <asm/war.h>
-#include <asm/vdso.h>
 
 #include "signal-common.h"
 
@@ -151,11 +150,11 @@ static int setup_rt_frame_n32(void *sig_return, struct ksignal *ksig,
 
 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,
 
 	.off_sc_fpregs = offsetof(struct sigcontext, sc_fpregs),
 	.off_sc_fpc_csr = offsetof(struct sigcontext, sc_fpc_csr),
 	.off_sc_used_math = offset