Merge tag 'mips_4.20' of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux

Pull MIPS updates from Paul Burton:

 - kexec support for the generic MIPS platform when running on a CPU
   including the MIPS Coherence Manager & related hardware.

 - Improvements to the definition of memory barriers used around MMIO
   accesses, and fixes in their use.

 - Switch to CONFIG_NO_BOOTMEM from Mike Rapoport, finally dropping
   reliance on the old bootmem code.

 - A number of fixes & improvements for Loongson 3 systems.

 - DT & config updates for the Microsemi Ocelot platform.

 - Workaround to enable USB power on the Netgear WNDR3400v3.

 - Various cleanups & fixes.

* tag 'mips_4.20' of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux: (51 commits)
  MIPS: Cleanup DSP ASE detection
  MIPS: dts: Change upper case to lower case
  MIPS: generic: Add Network, SPI and I2C to ocelot_defconfig
  MIPS: Loongson-3: Fix BRIDGE irq delivery problem
  MIPS: Loongson-3: Fix CPU UART irq delivery problem
  MIPS: Remove unused PREF, PREFE & PREFX macros
  MIPS: lib: Use kernel_pref & user_pref in memcpy()
  MIPS: Remove unused CAT macro
  MIPS: Add kernel_pref & user_pref helpers
  MIPS: Remove unused TTABLE macro
  MIPS: Remove unused PIC macros
  MIPS: Remove unused MOVN & MOVZ macros
  MIPS: Provide actually relaxed MMIO accessors
  MIPS: Enforce strong ordering for MMIO accessors
  MIPS: Correct `mmiowb' barrier for `wbflush' platforms
  MIPS: Define MMIO ordering barriers
  MIPS: mscc: add PCB120 to the ocelot fitImage
  MIPS: mscc: add DT for Ocelot PCB120
  MIPS: memset: Limit excessive `noreorder' assembly mode use
  MIPS: memset: Fix CPU_DADDI_WORKAROUNDS `small_fixup' regression
  ...

10 files changed, 127 insertions, 265 deletions

diff --git a/arch/mips/include/asm/asm-eva.h b/arch/mips/include/asm/asm-eva.h
index 1e38f0e1ea3e..d80be38c4144 100644
--- a/arch/mips/include/asm/asm-eva.h
+++ b/arch/mips/include/asm/asm-eva.h
@@ -15,6 +15,7 @@
 /* Kernel variants */
 
 #define kernel_cache(op, base)		"cache " op ", " base "\n"
+#define kernel_pref(hint, base)		"pref " hint ", " base "\n"
 #define kernel_ll(reg, addr)		"ll " reg ", " addr "\n"
 #define kernel_sc(reg, addr)		"sc " reg ", " addr "\n"
 #define kernel_lw(reg, addr)		"lw " reg ", " addr "\n"
@@ -51,6 +52,7 @@
 				"	.set	pop\n"
 
 #define user_cache(op, base)		__BUILD_EVA_INSN("cachee", op, base)
+#define user_pref(hint, base)		__BUILD_EVA_INSN("prefe", hint, base)
 #define user_ll(reg, addr)		__BUILD_EVA_INSN("lle", reg, addr)
 #define user_sc(reg, addr)		__BUILD_EVA_INSN("sce", reg, addr)
 #define user_lw(reg, addr)		__BUILD_EVA_INSN("lwe", reg, addr)
@@ -72,6 +74,7 @@
 #else
 
 #define user_cache(op, base)		kernel_cache(op, base)
+#define user_pref(hint, base)		kernel_pref(hint, base)
 #define user_ll(reg, addr)		kernel_ll(reg, addr)
 #define user_sc(reg, addr)		kernel_sc(reg, addr)
 #define user_lw(reg, addr)		kernel_lw(reg, addr)
@@ -99,6 +102,7 @@
 #else /* __ASSEMBLY__ */
 
 #define kernel_cache(op, base)		cache op, base
+#define kernel_pref(hint, base)		pref hint, base
 #define kernel_ll(reg, addr)		ll reg, addr
 #define kernel_sc(reg, addr)		sc reg, addr
 #define kernel_lw(reg, addr)		lw reg, addr
@@ -135,6 +139,7 @@
 				.set	pop;
 
 #define user_cache(op, base)		__BUILD_EVA_INSN(cachee, op, base)
+#define user_pref(hint, base)		__BUILD_EVA_INSN(prefe, hint, base)
 #define user_ll(reg, addr)		__BUILD_EVA_INSN(lle, reg, addr)
 #define user_sc(reg, addr)		__BUILD_EVA_INSN(sce, reg, addr)
 #define user_lw(reg, addr)		__BUILD_EVA_INSN(lwe, reg, addr)
@@ -155,6 +160,7 @@
 #else
 
 #define user_cache(op, base)		kernel_cache(op, base)
+#define user_pref(hint, base)		kernel_pref(hint, base)
 #define user_ll(reg, addr)		kernel_ll(reg, addr)
 #define user_sc(reg, addr)		kernel_sc(reg, addr)
 #define user_lw(reg, addr)		kernel_lw(reg, addr)
diff --git a/arch/mips/include/asm/asm.h b/arch/mips/include/asm/asm.h
index 81fae23ce7cd..c23527ba65d0 100644
--- a/arch/mips/include/asm/asm.h
+++ b/arch/mips/include/asm/asm.h
@@ -20,32 +20,6 @@
 #include <asm/sgidefs.h>
 #include <asm/asm-eva.h>
 
-#ifndef CAT
-#ifdef __STDC__
-#define __CAT(str1, str2) str1##str2
-#else
-#define __CAT(str1, str2) str1/**/str2
-#endif
-#define CAT(str1, str2) __CAT(str1, str2)
-#endif
-
-/*
- * PIC specific declarations
- * Not used for the kernel but here seems to be the right place.
- */
-#ifdef __PIC__
-#define CPRESTORE(register)				\
-		.cprestore register
-#define CPADD(register)					\
-		.cpadd	register
-#define CPLOAD(register)				\
-		.cpload register
-#else
-#define CPRESTORE(register)
-#define CPADD(register)
-#define CPLOAD(register)
-#endif
-
 /*
  * LEAF - declare leaf routine
  */
@@ -130,96 +104,6 @@ symbol		=	value
 		.popsection;
 
 /*
- * Build text tables
- */
-#define TTABLE(string)					\
-		.pushsection .text;			\
-		.word	1f;				\
-		.popsection				\
-		.pushsection .data;			\
-1:		.asciiz string;				\
-		.popsection
-
-/*
- * MIPS IV pref instruction.
- * Use with .set noreorder only!
- *
- * MIPS IV implementations are free to treat this as a nop.  The R5000
- * is one of them.  So we should have an option not to use this instruction.
- */
-#ifdef CONFIG_CPU_HAS_PREFETCH
-
-#define PREF(hint,addr)					\
-		.set	push;				\
-		.set	arch=r5000;			\
-		pref	hint, addr;			\
-		.set	pop
-
-#define PREFE(hint, addr)				\
-		.set	push;				\
-		.set	mips0;				\
-		.set	eva;				\
-		prefe	hint, addr;			\
-		.set	pop
-
-#define PREFX(hint,addr)				\
-		.set	push;				\
-		.set	arch=r5000;			\
-		prefx	hint, addr;			\
-		.set	pop
-
-#else /* !CONFIG_CPU_HAS_PREFETCH */
-
-#define PREF(hint, addr)
-#define PREFE(hint, addr)
-#define PREFX(hint, addr)
-
-#endif /* !CONFIG_CPU_HAS_PREFETCH */
-
-/*
- * MIPS ISA IV/V movn/movz instructions and equivalents for older CPUs.
- */
-#if (_MIPS_ISA == _MIPS_ISA_MIPS1)
-#define MOVN(rd, rs, rt)				\
-		.set	push;				\
-		.set	reorder;			\
-		beqz	rt, 9f;				\
-		move	rd, rs;				\
-		.set	pop;				\
-9:
-#define MOVZ(rd, rs, rt)				\
-		.set	push;				\
-		.set	reorder;			\
-		bnez	rt, 9f;				\
-		move	rd, rs;				\
-		.set	pop;				\
-9:
-#endif /* _MIPS_ISA == _MIPS_ISA_MIPS1 */
-#if (_MIPS_ISA == _MIPS_ISA_MIPS2) || (_MIPS_ISA == _MIPS_ISA_MIPS3)
-#define MOVN(rd, rs, rt)				\
-		.set	push;				\
-		.set	noreorder;			\
-		bnezl	rt, 9f;				\
-		 move	rd, rs;				\
-		.set	pop;				\
-9:
-#define MOVZ(rd, rs, rt)				\
-		.set	push;				\
-		.set	noreorder;			\
-		beqzl	rt, 9f;				\
-		 move	rd, rs;				\
-		.set	pop;				\
-9:
-#endif /* (_MIPS_ISA == _MIPS_ISA_MIPS2) || (_MIPS_ISA == _MIPS_ISA_MIPS3) */
-#if (_MIPS_ISA == _MIPS_ISA_MIPS4 ) || (_MIPS_ISA == _MIPS_ISA_MIPS5) || \
-    (_MIPS_ISA == _MIPS_ISA_MIPS32) || (_MIPS_ISA == _MIPS_ISA_MIPS64)
-#define MOVN(rd, rs, rt)				\
-		movn	rd, rs, rt
-#define MOVZ(rd, rs, rt)				\
-		movz	rd, rs, rt
-#endif /* MIPS IV, MIPS V, MIPS32 or MIPS64 */
-
-/*
  * Stack alignment
  */
 #if (_MIPS_SIM == _MIPS_SIM_ABI32)
diff --git a/arch/mips/include/asm/io.h b/arch/mips/include/asm/io.h
index 54c730aed327..266257d56fb6 100644
--- a/arch/mips/include/asm/io.h
+++ b/arch/mips/include/asm/io.h
@@ -20,6 +20,7 @@
 #include <linux/irqflags.h>
 
 #include <asm/addrspace.h>
+#include <asm/barrier.h>
 #include <asm/bug.h>
 #include <asm/byteorder.h>
 #include <asm/cpu.h>
@@ -34,11 +35,6 @@
 #include <mangle-port.h>
 
 /*
- * Slowdown I/O port space accesses for antique hardware.
- */
-#undef CONF_SLOWDOWN_IO
-
-/*
  * Raw operations are never swapped in software.  OTOH values that raw
  * operations are working on may or may not have been swapped by the bus
  * hardware.  An example use would be for flash memory that's used for
@@ -50,6 +46,11 @@
 # define __raw_ioswabq(a, x)	(x)
 # define ____raw_ioswabq(a, x)	(x)
 
+# define __relaxed_ioswabb ioswabb
+# define __relaxed_ioswabw ioswabw
+# define __relaxed_ioswabl ioswabl
+# define __relaxed_ioswabq ioswabq
+
 /* ioswab[bwlq], __mem_ioswab[bwlq] are defined in mangle-port.h */
 
 #define IO_SPACE_LIMIT 0xffff
@@ -80,31 +81,29 @@ static inline void set_io_port_base(unsigned long base)
 }
 
 /*
- * Thanks to James van Artsdalen for a better timing-fix than
- * the two short jumps: using outb's to a nonexistent port seems
- * to guarantee better timings even on fast machines.
- *
- * On the other hand, I'd like to be sure of a non-existent port:
- * I feel a bit unsafe about using 0x80 (should be safe, though)
- *
- *		Linus
- *
+ * Provide the necessary definitions for generic iomap. We make use of
+ * mips_io_port_base for iomap(), but we don't reserve any low addresses for
+ * use with I/O ports.
  */
 
-#define __SLOW_DOWN_IO \
-	__asm__ __volatile__( \
-		"sb\t$0,0x80(%0)" \
-		: : "r" (mips_io_port_base));
+#define HAVE_ARCH_PIO_SIZE
+#define PIO_OFFSET	mips_io_port_base
+#define PIO_MASK	IO_SPACE_LIMIT
+#define PIO_RESERVED	0x0UL
 
-#ifdef CONF_SLOWDOWN_IO
-#ifdef REALLY_SLOW_IO
-#define SLOW_DOWN_IO { __SLOW_DOWN_IO; __SLOW_DOWN_IO; __SLOW_DOWN_IO; __SLOW_DOWN_IO; }
-#else
-#define SLOW_DOWN_IO __SLOW_DOWN_IO
-#endif
-#else
-#define SLOW_DOWN_IO
-#endif
+/*
+ * Enforce in-order execution of data I/O.  In the MIPS architecture
+ * these are equivalent to corresponding platform-specific memory
+ * barriers defined in <asm/barrier.h>.  API pinched from PowerPC,
+ * with sync additionally defined.
+ */
+#define iobarrier_rw() mb()
+#define iobarrier_r() rmb()
+#define iobarrier_w() wmb()
+#define iobarrier_sync() iob()
+
+/* Some callers use this older API instead.  */
+#define mmiowb() iobarrier_w()
 
 /*
  *     virt_to_phys    -       map virtual addresses to physical
@@ -172,11 +171,6 @@ static inline void *isa_bus_to_virt(unsigned long address)
 extern void __iomem * __ioremap(phys_addr_t offset, phys_addr_t size, unsigned long flags);
 extern void __iounmap(const volatile void __iomem *addr);
 
-#ifndef CONFIG_PCI
-struct pci_dev;
-static inline void pci_iounmap(struct pci_dev *dev, void __iomem *addr) {}
-#endif
-
 static inline void __iomem * __ioremap_mode(phys_addr_t offset, unsigned long size,
 	unsigned long flags)
 {
@@ -316,13 +310,13 @@ static inline void iounmap(const volatile void __iomem *addr)
 #undef __IS_KSEG1
 }
 
-#if defined(CONFIG_CPU_CAVIUM_OCTEON) || defined(CONFIG_LOONGSON3_ENHANCEMENT)
+#if defined(CONFIG_CPU_CAVIUM_OCTEON) || defined(CONFIG_CPU_LOONGSON3)
 #define war_io_reorder_wmb()		wmb()
 #else
 #define war_io_reorder_wmb()		barrier()
 #endif
 
-#define __BUILD_MEMORY_SINGLE(pfx, bwlq, type, irq)			\
+#define __BUILD_MEMORY_SINGLE(pfx, bwlq, type, barrier, relax, irq)	\
 									\
 static inline void pfx##write##bwlq(type val,				\
 				    volatile void __iomem *mem)		\
@@ -330,7 +324,10 @@ static inline void pfx##write##bwlq(type val,				\
 	volatile type *__mem;						\
 	type __val;							\
 									\
-	war_io_reorder_wmb();					\
+	if (barrier)							\
+		iobarrier_rw();						\
+	else								\
+		war_io_reorder_wmb();					\
 									\
 	__mem = (void *)__swizzle_addr_##bwlq((unsigned long)(mem));	\
 									\
@@ -367,6 +364,9 @@ static inline type pfx##read##bwlq(const volatile void __iomem *mem)	\
 									\
 	__mem = (void *)__swizzle_addr_##bwlq((unsigned long)(mem));	\
 									\
+	if (barrier)							\
+		iobarrier_rw();						\
+									\
 	if (sizeof(type) != sizeof(u64) || sizeof(u64) == sizeof(long)) \
 		__val = *__mem;						\
 	else if (cpu_has_64bits) {					\
@@ -390,18 +390,22 @@ static inline type pfx##read##bwlq(const volatile void __iomem *mem)	\
 	}								\
 									\
 	/* prevent prefetching of coherent DMA data prematurely */	\
-	rmb();								\
+	if (!relax)							\
+		rmb();							\
 	return pfx##ioswab##bwlq(__mem, __val);				\
 }
 
-#define __BUILD_IOPORT_SINGLE(pfx, bwlq, type, p, slow)			\
+#define __BUILD_IOPORT_SINGLE(pfx, bwlq, type, barrier, relax, p)	\
 									\
 static inline void pfx##out##bwlq##p(type val, unsigned long port)	\
 {									\
 	volatile type *__addr;						\
 	type __val;							\
 									\
-	war_io_reorder_wmb();					\
+	if (barrier)							\
+		iobarrier_rw();						\
+	else								\
+		war_io_reorder_wmb();					\
 									\
 	__addr = (void *)__swizzle_addr_##bwlq(mips_io_port_base + port); \
 									\
@@ -411,7 +415,6 @@ static inline void pfx##out##bwlq##p(type val, unsigned long port)	\
 	BUILD_BUG_ON(sizeof(type) > sizeof(unsigned long));		\
 									\
 	*__addr = __val;						\
-	slow;								\
 }									\
 									\
 static inline type pfx##in##bwlq##p(unsigned long port)			\
@@ -423,23 +426,27 @@ static inline type pfx##in##bwlq##p(unsigned long port)			\
 									\
 	BUILD_BUG_ON(sizeof(type) > sizeof(unsigned long));		\
 									\
+	if (barrier)							\
+		iobarrier_rw();						\
+									\
 	__val = *__addr;						\
-	slow;								\
 									\
 	/* prevent prefetching of coherent DMA data prematurely */	\
-	rmb();								\
+	if (!relax)							\
+		rmb();							\
 	return pfx##ioswab##bwlq(__addr, __val);			\
 }
 
-#define __BUILD_MEMORY_PFX(bus, bwlq, type)				\
+#define __BUILD_MEMORY_PFX(bus, bwlq, type, relax)			\
 									\
-__BUILD_MEMORY_SINGLE(bus, bwlq, type, 1)
+__BUILD_MEMORY_SINGLE(bus, bwlq, type, 1, relax, 1)
 
 #define BUILDIO_MEM(bwlq, type)						\
 									\
-__BUILD_MEMORY_PFX(__raw_, bwlq, type)					\
-__BUILD_MEMORY_PFX(, bwlq, type)					\
-__BUILD_MEMORY_PFX(__mem_, bwlq, type)					\
+__BUILD_MEMORY_PFX(__raw_, bwlq, type, 0)				\
+__BUILD_MEMORY_PFX(__relaxed_, bwlq, type, 1)				\
+__BUILD_MEMORY_PFX(__mem_, bwlq, type, 0)				\
+__BUILD_MEMORY_PFX(, bwlq, type, 0)
 
 BUILDIO_MEM(b, u8)
 BUILDIO_MEM(w, u16)
@@ -447,8 +454,8 @@ BUILDIO_MEM(l, u32)
 BUILDIO_MEM(q, u64)
 
 #define __BUILD_IOPORT_PFX(bus, bwlq, type)				\
-	__BUILD_IOPORT_SINGLE(bus, bwlq, type, ,)			\
-	__BUILD_IOPORT_SINGLE(bus, bwlq, type, _p, SLOW_DOWN_IO)
+	__BUILD_IOPORT_SINGLE(bus, bwlq, type, 1, 0,)			\
+	__BUILD_IOPORT_SINGLE(bus, bwlq, type, 1, 0, _p)
 
 #define BUILDIO_IOPORT(bwlq, type)					\
 	__BUILD_IOPORT_PFX(, bwlq, type)				\
@@ -463,19 +470,19 @@ BUILDIO_IOPORT(q, u64)
 
 #define __BUILDIO(bwlq, type)						\
 									\
-__BUILD_MEMORY_SINGLE(____raw_, bwlq, type, 0)
+__BUILD_MEMORY_SINGLE(____raw_, bwlq, type, 1, 0, 0)
 
 __BUILDIO(q, u64)
 
-#define readb_relaxed			readb
-#define readw_relaxed			readw
-#define readl_relaxed			readl
-#define readq_relaxed			readq
+#define readb_relaxed			__relaxed_readb
+#define readw_relaxed			__relaxed_readw
+#define readl_relaxed			__relaxed_readl
+#define readq_relaxed			__relaxed_readq
 
-#define writeb_relaxed			writeb
-#define writew_relaxed			writew
-#define writel_relaxed			writel
-#define writeq_relaxed			writeq
+#define writeb_relaxed			__relaxed_writeb
+#define writew_relaxed			__relaxed_writew
+#define writel_relaxed			__relaxed_writel
+#define writeq_relaxed			__relaxed_writeq
 
 #define readb_be(addr)							\
 	__raw_readb((__force unsigned *)(addr))
@@ -561,14 +568,6 @@ BUILDSTRING(l, u32)
 BUILDSTRING(q, u64)
 #endif
 
-
-#ifdef CONFIG_CPU_CAVIUM_OCTEON
-#define mmiowb() wmb()
-#else
-/* Depends on MIPS II instruction set */
-#define mmiowb() asm volatile ("sync" ::: "memory")
-#endif
-
 static inline void memset_io(volatile void __iomem *addr, unsigned char val, int count)
 {
 	memset((void __force *) addr, val, count);
diff --git a/arch/mips/include/asm/kexec.h b/arch/mips/include/asm/kexec.h
index 493a3cc7c39a..40795ca89961 100644
--- a/arch/mips/include/asm/kexec.h
+++ b/arch/mips/include/asm/kexec.h
@@ -12,11 +12,11 @@
 #include <asm/stacktrace.h>
 
 /* Maximum physical address we can use pages from */
-#define KEXEC_SOURCE_MEMORY_LIMIT (0x20000000)
+#define KEXEC_SOURCE_MEMORY_LIMIT (-1UL)
 /* Maximum address we can reach in physical address mode */
-#define KEXEC_DESTINATION_MEMORY_LIMIT (0x20000000)
+#define KEXEC_DESTINATION_MEMORY_LIMIT (-1UL)
  /* Maximum address we can use for the control code buffer */
-#define KEXEC_CONTROL_MEMORY_LIMIT (0x20000000)
+#define KEXEC_CONTROL_MEMORY_LIMIT (-1UL)
 /* Reserve 3*4096 bytes for board-specific info */
 #define KEXEC_CONTROL_PAGE_SIZE (4096 + 3*4096)
 
@@ -39,11 +39,12 @@ extern unsigned long kexec_args[4];
 extern int (*_machine_kexec_prepare)(struct kimage *);
 extern void (*_machine_kexec_shutdown)(void);
 extern void (*_machine_crash_shutdown)(struct pt_regs *regs);
-extern void default_machine_crash_shutdown(struct pt_regs *regs);
+void default_machine_crash_shutdown(struct pt_regs *regs);
+void kexec_nonboot_cpu_jump(void);
+void kexec_reboot(void);
 #ifdef CONFIG_SMP
 extern const unsigned char kexec_smp_wait[];
 extern unsigned long secondary_kexec_args[4];
-extern void (*relocated_kexec_smp_wait) (void *);
 extern atomic_t kexec_ready_to_reboot;
 extern void (*_crash_smp_send_stop)(void);
 #endif
diff --git a/arch/mips/include/asm/mach-loongson64/irq.h b/arch/mips/include/asm/mach-loongson64/irq.h
index 3644b68c0ccc..be9f727a9328 100644
--- a/arch/mips/include/asm/mach-loongson64/irq.h
+++ b/arch/mips/include/asm/mach-loongson64/irq.h
@@ -10,7 +10,7 @@
 #define MIPS_CPU_IRQ_BASE 56
 
 #define LOONGSON_UART_IRQ   (MIPS_CPU_IRQ_BASE + 2) /* UART */
-#define LOONGSON_HT1_IRQ    (MIPS_CPU_IRQ_BASE + 3) /* HT1 */
+#define LOONGSON_BRIDGE_IRQ (MIPS_CPU_IRQ_BASE + 3) /* CASCADE */
 #define LOONGSON_TIMER_IRQ  (MIPS_CPU_IRQ_BASE + 7) /* CPU Timer */
 
 #define LOONGSON_HT1_CFG_BASE		loongson_sysconf.ht_control_base
diff --git a/arch/mips/include/asm/mach-loongson64/kernel-entry-init.h b/arch/mips/include/asm/mach-loongson64/kernel-entry-init.h
index 312739117bb0..cbac603ced19 100644
--- a/arch/mips/include/asm/mach-loongson64/kernel-entry-init.h
+++ b/arch/mips/include/asm/mach-loongson64/kernel-entry-init.h
@@ -11,6 +11,8 @@
 #ifndef __ASM_MACH_LOONGSON64_KERNEL_ENTRY_H
 #define __ASM_MACH_LOONGSON64_KERNEL_ENTRY_H
 
+#include <asm/cpu.h>
+
 /*
  * Override macros used in arch/mips/kernel/head.S.
  */
@@ -26,12 +28,15 @@
 	mfc0	t0, CP0_PAGEGRAIN
 	or	t0, (0x1 << 29)
 	mtc0	t0, CP0_PAGEGRAIN
-#ifdef CONFIG_LOONGSON3_ENHANCEMENT
 	/* Enable STFill Buffer */
+	mfc0	t0, CP0_PRID
+	andi	t0, (PRID_IMP_MASK | PRID_REV_MASK)
+	slti	t0, (PRID_IMP_LOONGSON_64 | PRID_REV_LOONGSON3A_R2)
+	bnez	t0, 1f
 	mfc0	t0, CP0_CONFIG6
 	or	t0, 0x100
 	mtc0	t0, CP0_CONFIG6
-#endif
+1:
 	_ehb
 	.set	pop
 #endif
@@ -52,12 +57,15 @@
 	mfc0	t0, CP0_PAGEGRAIN
 	or	t0, (0x1 << 29)
 	mtc0	t0, CP0_PAGEGRAIN
-#ifdef CONFIG_LOONGSON3_ENHANCEMENT
 	/* Enable STFill Buffer */
+	mfc0	t0, CP0_PRID
+	andi	t0, (PRID_IMP_MASK | PRID_REV_MASK)
+	slti	t0, (PRID_IMP_LOONGSON_64 | PRID_REV_LOONGSON3A_R2)
+	bnez	t0, 1f
 	mfc0	t0, CP0_CONFIG6
 	or	t0, 0x100
 	mtc0	t0, CP0_CONFIG6
-#endif
+1:
 	_ehb
 	.set	pop
 #endif
diff --git a/arch/mips/include/asm/mipsregs.h b/arch/mips/include/asm/mipsregs.h
index 01df9ad62fb8..341a02c92985 100644
--- a/arch/mips/include/asm/mipsregs.h
+++ b/arch/mips/include/asm/mipsregs.h
@@ -2287,13 +2287,14 @@ do {									\
 	_write_32bit_cp1_register(dest, val, )
 #endif
 
-#ifdef HAVE_AS_DSP
+#ifdef TOOLCHAIN_SUPPORTS_DSP
 #define rddsp(mask)							\
 ({									\
 	unsigned int __dspctl;						\
 									\
 	__asm__ __volatile__(						\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	rddsp	%0, %x1					\n"	\
 	"	.set pop					\n"	\
@@ -2306,6 +2307,7 @@ do {									\
 do {									\
 	__asm__ __volatile__(						\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	wrdsp	%0, %x1					\n"	\
 	"	.set pop					\n"	\
@@ -2318,6 +2320,7 @@ do {									\
 	long mflo0;							\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mflo %0, $ac0					\n"	\
 	"	.set pop					\n" 	\
@@ -2330,6 +2333,7 @@ do {									\
 	long mflo1;							\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mflo %0, $ac1					\n"	\
 	"	.set pop					\n" 	\
@@ -2342,6 +2346,7 @@ do {									\
 	long mflo2;							\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mflo %0, $ac2					\n"	\
 	"	.set pop					\n" 	\
@@ -2354,6 +2359,7 @@ do {									\
 	long mflo3;							\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mflo %0, $ac3					\n"	\
 	"	.set pop					\n" 	\
@@ -2366,6 +2372,7 @@ do {									\
 	long mfhi0;							\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mfhi %0, $ac0					\n"	\
 	"	.set pop					\n" 	\
@@ -2378,6 +2385,7 @@ do {									\
 	long mfhi1;							\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mfhi %0, $ac1					\n"	\
 	"	.set pop					\n" 	\
@@ -2390,6 +2398,7 @@ do {									\
 	long mfhi2;							\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mfhi %0, $ac2					\n"	\
 	"	.set pop					\n" 	\
@@ -2402,6 +2411,7 @@ do {									\
 	long mfhi3;							\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mfhi %0, $ac3					\n"	\
 	"	.set pop					\n" 	\
@@ -2414,6 +2424,7 @@ do {									\
 ({									\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mtlo %0, $ac0					\n"	\
 	"	.set pop					\n"	\
@@ -2425,6 +2436,7 @@ do {									\
 ({									\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mtlo %0, $ac1					\n"	\
 	"	.set pop					\n"	\
@@ -2436,6 +2448,7 @@ do {									\
 ({									\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mtlo %0, $ac2					\n"	\
 	"	.set pop					\n"	\
@@ -2447,6 +2460,7 @@ do {									\
 ({									\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mtlo %0, $ac3					\n"	\
 	"	.set pop					\n"	\
@@ -2458,6 +2472,7 @@ do {									\
 ({									\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mthi %0, $ac0					\n"	\
 	"	.set pop					\n"	\
@@ -2469,6 +2484,7 @@ do {									\
 ({									\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mthi %0, $ac1					\n"	\
 	"	.set pop					\n"	\
@@ -2480,6 +2496,7 @@ do {									\
 ({									\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mthi %0, $ac2					\n"	\
 	"	.set pop					\n"	\
@@ -2491,6 +2508,7 @@ do {									\
 ({									\
 	__asm__(							\
 	"	.set push					\n"	\
+	"	.set " MIPS_ISA_LEVEL "				\n"	\
 	"	.set dsp					\n"	\
 	"	mthi %0, $ac3					\n"	\
 	"	.set pop					\n"	\
diff --git a/arch/mips/include/asm/r4kcache.h b/arch/mips/include/asm/r4kcache.h
index 7f12d7e27c94..d19b2d65336b 100644
--- a/arch/mips/include/asm/r4kcache.h
+++ b/arch/mips/include/asm/r4kcache.h
@@ -48,58 +48,14 @@ extern void (*r4k_blast_icache)(void);
 	:								\
 	: "i" (op), "R" (*(unsigned char *)(addr)))
 
-#ifdef CONFIG_MIPS_MT
-
-#define __iflush_prologue						\
-	unsigned long redundance;					\
-	extern int mt_n_iflushes;					\
-	for (redundance = 0; redundance < mt_n_iflushes; redundance++) {
-
-#define __iflush_epilogue						\
-	}
-
-#define __dflush_prologue						\
-	unsigned long redundance;					\
-	extern int mt_n_dflushes;					\
-	for (redundance = 0; redundance < mt_n_dflushes; redundance++) {
-
-#define __dflush_epilogue \
-	}
-
-#define __inv_dflush_prologue __dflush_prologue
-#define __inv_dflush_epilogue __dflush_epilogue
-#define __sflush_prologue {
-#define __sflush_epilogue }
-#define __inv_sflush_prologue __sflush_prologue
-#define __inv_sflush_epilogue __sflush_epilogue
-
-#else /* CONFIG_MIPS_MT */
-
-#define __iflush_prologue {
-#define __iflush_epilogue }
-#define __dflush_prologue {
-#define __dflush_epilogue }
-#define __inv_dflush_prologue {
-#define __inv_dflush_epilogue }
-#define __sflush_prologue {
-#define __sflush_epilogue }
-#define __inv_sflush_prologue {
-#define __inv_sflush_epilogue }
-
-#endif /* CONFIG_MIPS_MT */
-
 static inline void flush_icache_line_indexed(unsigned long addr)
 {
-	__iflush_prologue
 	cache_op(Index_Invalidate_I, addr);
-	__iflush_epilogue
 }
 
 static inline void flush_dcache_line_indexed(unsigned long addr)
 {
-	__dflush_prologue
 	cache_op(Index_Writeback_Inv_D, addr);
-	__dflush_epilogue
 }
 
 static inline void flush_scache_line_indexed(unsigned long addr)
@@ -109,7 +65,6 @@ static inline void flush_scache_line_indexed(unsigned long addr)
 
 static inline void flush_icache_line(unsigned long addr)
 {
-	__iflush_prologue
 	switch (boot_cpu_type()) {
 	case CPU_LOONGSON2:
 		cache_op(Hit_Invalidate_I_Loongson2, addr);
@@ -119,21 +74,16 @@ static inline void flush_icache_line(unsigned long addr)
 		cache_op(Hit_Invalidate_I, addr);
 		break;
 	}
-	__iflush_epilogue
 }
 
 static inline void flush_dcache_line(unsigned long addr)
 {
-	__dflush_prologue
 	cache_op(Hit_Writeback_Inv_D, addr);
-	__dflush_epilogue
 }
 
 static inline void invalidate_dcache_line(unsigned long addr)
 {
-	__dflush_prologue
 	cache_op(Hit_Invalidate_D, addr);
-	__dflush_epilogue
 }
 
 static inline void invalidate_scache_line(unsigned long addr)
@@ -586,13 +536,9 @@ static inline void extra##blast_##pfx##cache##lsize(void)		\
 			       current_cpu_data.desc.waybit;		\
 	unsigned long ws, addr;						\
 									\
-	__##pfx##flush_prologue						\
-									\
 	for (ws = 0; ws < ws_end; ws += ws_inc)				\
 		for (addr = start; addr < end; addr += lsize * 32)	\
 			cache##lsize##_unroll32(addr|ws, indexop);	\
-									\
-	__##pfx##flush_epilogue						\
 }									\
 									\
 static inline void extra##blast_##pfx##cache##lsize##_page(unsigned long page) \
@@ -600,14 +546,10 @@ static inline void extra##blast_##pfx##cache##lsize##_page(unsigned long page) \
 	unsigned long start = page;					\
 	unsigned long end = page + PAGE_SIZE;				\
 									\
-	__##pfx##flush_prologue						\
-									\
 	do {								\
 		cache##lsize##_unroll32(start, hitop);			\
 		start += lsize * 32;					\
 	} while (start < end);						\
-									\
-	__##pfx##flush_epilogue						\
 }									\
 									\
 static inline void extra##blast_##pfx##cache##lsize##_page_indexed(unsigned long page) \
@@ -620,13 +562,9 @@ static inline void extra##blast_##pfx##cache##lsize##_page_indexed(unsigned long
 			       current_cpu_data.desc.waybit;		\
 	unsigned long ws, addr;						\
 									\
-	__##pfx##flush_prologue						\
-									\
 	for (ws = 0; ws < ws_end; ws += ws_inc)				\
 		for (addr = start; addr < end; addr += lsize * 32)	\
 			cache##lsize##_unroll32(addr|ws, indexop);	\
-									\
-	__##pfx##flush_epilogue						\
 }
 
 __BUILD_BLAST_CACHE(d, dcache, Index_Writeback_Inv_D, Hit_Writeback_Inv_D, 16, )
@@ -656,14 +594,10 @@ static inline void blast_##pfx##cache##lsize##_user_page(unsigned long page) \
 	unsigned long start = page;					\
 	unsigned long end = page + PAGE_SIZE;				\
 									\
-	__##pfx##flush_prologue						\
-									\
 	do {								\
 		cache##lsize##_unroll32_user(start, hitop);             \
 		start += lsize * 32;					\
 	} while (start < end);						\
-									\
-	__##pfx##flush_epilogue						\
 }
 
 __BUILD_BLAST_USER_CACHE(d, dcache, Index_Writeback_Inv_D, Hit_Writeback_Inv_D,
@@ -685,16 +619,12 @@ static inline void prot##extra##blast_##pfx##cache##_range(unsigned long start,
 	unsigned long addr = start & ~(lsize - 1);			\
 	unsigned long aend = (end - 1) & ~(lsize - 1);			\
 									\
-	__##pfx##flush_prologue						\
-									\
 	while (1) {							\
 		prot##cache_op(hitop, addr);				\
 		if (addr == aend)					\
 			break;						\
 		addr += lsize;						\
 	}								\
-									\
-	__##pfx##flush_epilogue						\
 }
 
 #ifndef CONFIG_EVA
@@ -712,8 +642,6 @@ static inline void protected_blast_##pfx##cache##_range(unsigned long start,\
 	unsigned long addr = start & ~(lsize - 1);			\
 	unsigned long aend = (end - 1) & ~(lsize - 1);			\
 									\
-	__##pfx##flush_prologue						\
-									\
 	if (!uaccess_kernel()) {					\
 		while (1) {						\
 			protected_cachee_op(hitop, addr);		\
@@ -730,7 +658,6 @@ static inline void protected_blast_##pfx##cache##_range(unsigned long start,\
 		}                                                       \
 									\
 	}								\
-	__##pfx##flush_epilogue						\
 }
 
 __BUILD_PROT_BLAST_CACHE_RANGE(d, dcache, Hit_Writeback_Inv_D)
diff --git a/arch/mips/include/asm/smp-ops.h b/arch/mips/include/asm/smp-ops.h
index 53b2cb8e5966..b7123f9c0785 100644
--- a/arch/mips/include/asm/smp-ops.h
+++ b/arch/mips/include/asm/smp-ops.h
@@ -33,6 +33,9 @@ struct plat_smp_ops {
 	int (*cpu_disable)(void);
 	void (*cpu_die)(unsigned int cpu);
 #endif
+#ifdef CONFIG_KEXEC
+	void (*kexec_nonboot_cpu)(void);
+#endif
 };
 
 extern void register_smp_ops(const struct plat_smp_ops *ops);
diff --git a/arch/mips/include/asm/smp.h b/arch/mips/include/asm/smp.h
index 056a6bf13491..7990c1c70471 100644
--- a/arch/mips/include/asm/smp.h
+++ b/arch/mips/include/asm/smp.h
@@ -91,6 +91,22 @@ static inline void __cpu_die(unsigned int cpu)
 extern void play_dead(void);
 #endif
 
+#ifdef CONFIG_KEXEC
+static inline void kexec_nonboot_cpu(void)
+{
+	extern const struct plat_smp_ops *mp_ops;	/* private */
+
+	return mp_ops->kexec_nonboot_cpu();
+}
+
+static inline void *kexec_nonboot_cpu_func(void)
+{
+	extern const struct plat_smp_ops *mp_ops;	/* private */
+
+	return mp_ops->kexec_nonboot_cpu;
+}
+#endif
+
 /*
  * This function will set up the necessary IPIs for Linux to communicate
  * with the CPUs in mask.