Merge remote-tracking branch 'linus/master' into staging/for_v3.8

* linus/master: (1428 commits)
  futex: avoid wake_futex() for a PI futex_q
  watchdog: using u64 in get_sample_period()
  writeback: put unused inodes to LRU after writeback completion
  mm: vmscan: check for fatal signals iff the process was throttled
  Revert "mm: remove __GFP_NO_KSWAPD"
  proc: check vma->vm_file before dereferencing
  UAPI: strip the _UAPI prefix from header guards during header installation
  include/linux/bug.h: fix sparse warning related to BUILD_BUG_ON_INVALID
  Linux 3.7-rc7
  powerpc/eeh: Do not invalidate PE properly
  ALSA: hda - Fix build without CONFIG_PM
  of/address: sparc: Declare of_iomap as an extern function for sparc again
  PM / QoS: fix wrong error-checking condition
  bnx2x: remove redundant warning log
  vxlan: fix command usage in its doc
  8139cp: revert "set ring address before enabling receiver"
  MPI: Fix compilation on MIPS with GCC 4.4 and newer
  MIPS: Fix crash that occurs when function tracing is enabled
  MIPS: Merge overlapping bootmem ranges
  jbd: Fix lock ordering bug in journal_unmap_buffer()
  ...

40 files changed, 484 insertions, 116 deletions

diff --git a/arch/sparc/Kconfig b/arch/sparc/Kconfig
index b6b442b0d793..9f2edb5c5551 100644
--- a/arch/sparc/Kconfig
+++ b/arch/sparc/Kconfig
@@ -20,6 +20,7 @@ config SPARC
 	select HAVE_ARCH_TRACEHOOK
 	select SYSCTL_EXCEPTION_TRACE
 	select ARCH_WANT_OPTIONAL_GPIOLIB
+	select ARCH_HAS_ATOMIC64_DEC_IF_POSITIVE
 	select RTC_CLASS
 	select RTC_DRV_M48T59
 	select HAVE_IRQ_WORK
diff --git a/arch/sparc/crypto/Makefile b/arch/sparc/crypto/Makefile
index 6ae1ad5e502b..5d469d81761f 100644
--- a/arch/sparc/crypto/Makefile
+++ b/arch/sparc/crypto/Makefile
@@ -13,13 +13,13 @@ obj-$(CONFIG_CRYPTO_DES_SPARC64) += camellia-sparc64.o
 
 obj-$(CONFIG_CRYPTO_CRC32C_SPARC64) += crc32c-sparc64.o
 
-sha1-sparc64-y := sha1_asm.o sha1_glue.o crop_devid.o
-sha256-sparc64-y := sha256_asm.o sha256_glue.o crop_devid.o
-sha512-sparc64-y := sha512_asm.o sha512_glue.o crop_devid.o
-md5-sparc64-y := md5_asm.o md5_glue.o crop_devid.o
+sha1-sparc64-y := sha1_asm.o sha1_glue.o
+sha256-sparc64-y := sha256_asm.o sha256_glue.o
+sha512-sparc64-y := sha512_asm.o sha512_glue.o
+md5-sparc64-y := md5_asm.o md5_glue.o
 
-aes-sparc64-y := aes_asm.o aes_glue.o crop_devid.o
-des-sparc64-y := des_asm.o des_glue.o crop_devid.o
-camellia-sparc64-y := camellia_asm.o camellia_glue.o crop_devid.o
+aes-sparc64-y := aes_asm.o aes_glue.o
+des-sparc64-y := des_asm.o des_glue.o
+camellia-sparc64-y := camellia_asm.o camellia_glue.o
 
-crc32c-sparc64-y := crc32c_asm.o crc32c_glue.o crop_devid.o
+crc32c-sparc64-y := crc32c_asm.o crc32c_glue.o
diff --git a/arch/sparc/crypto/aes_glue.c b/arch/sparc/crypto/aes_glue.c
index 8f1c9980f637..3965d1d36dfa 100644
--- a/arch/sparc/crypto/aes_glue.c
+++ b/arch/sparc/crypto/aes_glue.c
@@ -475,3 +475,5 @@ MODULE_LICENSE("GPL");
 MODULE_DESCRIPTION("AES Secure Hash Algorithm, sparc64 aes opcode accelerated");
 
 MODULE_ALIAS("aes");
+
+#include "crop_devid.c"
diff --git a/arch/sparc/crypto/camellia_glue.c b/arch/sparc/crypto/camellia_glue.c
index 42905c084299..62c89af3fd3f 100644
--- a/arch/sparc/crypto/camellia_glue.c
+++ b/arch/sparc/crypto/camellia_glue.c
@@ -320,3 +320,5 @@ MODULE_LICENSE("GPL");
 MODULE_DESCRIPTION("Camellia Cipher Algorithm, sparc64 camellia opcode accelerated");
 
 MODULE_ALIAS("aes");
+
+#include "crop_devid.c"
diff --git a/arch/sparc/crypto/crc32c_glue.c b/arch/sparc/crypto/crc32c_glue.c
index 0bd89cea8d8e..5162fad912ce 100644
--- a/arch/sparc/crypto/crc32c_glue.c
+++ b/arch/sparc/crypto/crc32c_glue.c
@@ -177,3 +177,5 @@ MODULE_LICENSE("GPL");
 MODULE_DESCRIPTION("CRC32c (Castagnoli), sparc64 crc32c opcode accelerated");
 
 MODULE_ALIAS("crc32c");
+
+#include "crop_devid.c"
diff --git a/arch/sparc/crypto/des_glue.c b/arch/sparc/crypto/des_glue.c
index c4940c2d3073..41524cebcc49 100644
--- a/arch/sparc/crypto/des_glue.c
+++ b/arch/sparc/crypto/des_glue.c
@@ -527,3 +527,5 @@ MODULE_LICENSE("GPL");
 MODULE_DESCRIPTION("DES & Triple DES EDE Cipher Algorithms, sparc64 des opcode accelerated");
 
 MODULE_ALIAS("des");
+
+#include "crop_devid.c"
diff --git a/arch/sparc/crypto/md5_glue.c b/arch/sparc/crypto/md5_glue.c
index 603d723038ce..09a9ea1dfb69 100644
--- a/arch/sparc/crypto/md5_glue.c
+++ b/arch/sparc/crypto/md5_glue.c
@@ -186,3 +186,5 @@ MODULE_LICENSE("GPL");
 MODULE_DESCRIPTION("MD5 Secure Hash Algorithm, sparc64 md5 opcode accelerated");
 
 MODULE_ALIAS("md5");
+
+#include "crop_devid.c"
diff --git a/arch/sparc/crypto/sha1_glue.c b/arch/sparc/crypto/sha1_glue.c
index 2bbb20bee9f1..6cd5f29e1e0d 100644
--- a/arch/sparc/crypto/sha1_glue.c
+++ b/arch/sparc/crypto/sha1_glue.c
@@ -181,3 +181,5 @@ MODULE_LICENSE("GPL");
 MODULE_DESCRIPTION("SHA1 Secure Hash Algorithm, sparc64 sha1 opcode accelerated");
 
 MODULE_ALIAS("sha1");
+
+#include "crop_devid.c"
diff --git a/arch/sparc/crypto/sha256_glue.c b/arch/sparc/crypto/sha256_glue.c
index 591e656bd891..04f555ab2680 100644
--- a/arch/sparc/crypto/sha256_glue.c
+++ b/arch/sparc/crypto/sha256_glue.c
@@ -239,3 +239,5 @@ MODULE_DESCRIPTION("SHA-224 and SHA-256 Secure Hash Algorithm, sparc64 sha256 op
 
 MODULE_ALIAS("sha224");
 MODULE_ALIAS("sha256");
+
+#include "crop_devid.c"
diff --git a/arch/sparc/crypto/sha512_glue.c b/arch/sparc/crypto/sha512_glue.c
index 486f0a2b7001..f04d1994d19a 100644
--- a/arch/sparc/crypto/sha512_glue.c
+++ b/arch/sparc/crypto/sha512_glue.c
@@ -224,3 +224,5 @@ MODULE_DESCRIPTION("SHA-384 and SHA-512 Secure Hash Algorithm, sparc64 sha512 op
 
 MODULE_ALIAS("sha384");
 MODULE_ALIAS("sha512");
+
+#include "crop_devid.c"
diff --git a/arch/sparc/include/asm/atomic_64.h b/arch/sparc/include/asm/atomic_64.h
index ce35a1cf1a20..be56a244c9cf 100644
--- a/arch/sparc/include/asm/atomic_64.h
+++ b/arch/sparc/include/asm/atomic_64.h
@@ -1,7 +1,7 @@
 /* atomic.h: Thankfully the V9 is at least reasonable for this
  *           stuff.
  *
- * Copyright (C) 1996, 1997, 2000 David S. Miller (davem@redhat.com)
+ * Copyright (C) 1996, 1997, 2000, 2012 David S. Miller (davem@redhat.com)
  */
 
 #ifndef __ARCH_SPARC64_ATOMIC__
@@ -106,6 +106,8 @@ static inline long atomic64_add_unless(atomic64_t *v, long a, long u)
 
 #define atomic64_inc_not_zero(v) atomic64_add_unless((v), 1, 0)
 
+extern long atomic64_dec_if_positive(atomic64_t *v);
+
 /* Atomic operations are already serializing */
 #define smp_mb__before_atomic_dec()	barrier()
 #define smp_mb__after_atomic_dec()	barrier()
diff --git a/arch/sparc/include/asm/backoff.h b/arch/sparc/include/asm/backoff.h
index db3af0d30fb1..4e02086b839c 100644
--- a/arch/sparc/include/asm/backoff.h
+++ b/arch/sparc/include/asm/backoff.h
@@ -1,6 +1,46 @@
 #ifndef _SPARC64_BACKOFF_H
 #define _SPARC64_BACKOFF_H
 
+/* The macros in this file implement an exponential backoff facility
+ * for atomic operations.
+ *
+ * When multiple threads compete on an atomic operation, it is
+ * possible for one thread to be continually denied a successful
+ * completion of the compare-and-swap instruction.  Heavily
+ * threaded cpu implementations like Niagara can compound this
+ * problem even further.
+ *
+ * When an atomic operation fails and needs to be retried, we spin a
+ * certain number of times.  At each subsequent failure of the same
+ * operation we double the spin count, realizing an exponential
+ * backoff.
+ *
+ * When we spin, we try to use an operation that will cause the
+ * current cpu strand to block, and therefore make the core fully
+ * available to any other other runnable strands.  There are two
+ * options, based upon cpu capabilities.
+ *
+ * On all cpus prior to SPARC-T4 we do three dummy reads of the
+ * condition code register.  Each read blocks the strand for something
+ * between 40 and 50 cpu cycles.
+ *
+ * For SPARC-T4 and later we have a special "pause" instruction
+ * available.  This is implemented using writes to register %asr27.
+ * The cpu will block the number of cycles written into the register,
+ * unless a disrupting trap happens first.  SPARC-T4 specifically
+ * implements pause with a granularity of 8 cycles.  Each strand has
+ * an internal pause counter which decrements every 8 cycles.  So the
+ * chip shifts the %asr27 value down by 3 bits, and writes the result
+ * into the pause counter.  If a value smaller than 8 is written, the
+ * chip blocks for 1 cycle.
+ *
+ * To achieve the same amount of backoff as the three %ccr reads give
+ * on earlier chips, we shift the backoff value up by 7 bits.  (Three
+ * %ccr reads block for about 128 cycles, 1 << 7 == 128) We write the
+ * whole amount we want to block into the pause register, rather than
+ * loop writing 128 each time.
+ */
+
 #define BACKOFF_LIMIT	(4 * 1024)
 
 #ifdef CONFIG_SMP
@@ -11,16 +51,25 @@
 #define BACKOFF_LABEL(spin_label, continue_label) \
 	spin_label
 
-#define BACKOFF_SPIN(reg, tmp, label)	\
-	mov	reg, tmp; \
-88:	brnz,pt	tmp, 88b; \
-	 sub	tmp, 1, tmp; \
-	set	BACKOFF_LIMIT, tmp; \
-	cmp	reg, tmp; \
-	bg,pn	%xcc, label; \
-	 nop; \
-	ba,pt	%xcc, label; \
-	 sllx	reg, 1, reg;
+#define BACKOFF_SPIN(reg, tmp, label)		\
+	mov		reg, tmp;		\
+88:	rd		%ccr, %g0;		\
+	rd		%ccr, %g0;		\
+	rd		%ccr, %g0;		\
+	.section	.pause_3insn_patch,"ax";\
+	.word		88b;			\
+	sllx		tmp, 7, tmp;		\
+	wr		tmp, 0, %asr27;		\
+	clr		tmp;			\
+	.previous;				\
+	brnz,pt		tmp, 88b;		\
+	 sub		tmp, 1, tmp;		\
+	set		BACKOFF_LIMIT, tmp;	\
+	cmp		reg, tmp;		\
+	bg,pn		%xcc, label;		\
+	 nop;					\
+	ba,pt		%xcc, label;		\
+	 sllx		reg, 1, reg;
 
 #else
 
diff --git a/arch/sparc/include/asm/compat.h b/arch/sparc/include/asm/compat.h
index cef99fbc0a21..830502fe62b4 100644
--- a/arch/sparc/include/asm/compat.h
+++ b/arch/sparc/include/asm/compat.h
@@ -232,9 +232,10 @@ static inline void __user *arch_compat_alloc_user_space(long len)
 	struct pt_regs *regs = current_thread_info()->kregs;
 	unsigned long usp = regs->u_regs[UREG_I6];
 
-	if (!(test_thread_flag(TIF_32BIT)))
+	if (test_thread_64bit_stack(usp))
 		usp += STACK_BIAS;
-	else
+
+	if (test_thread_flag(TIF_32BIT))
 		usp &= 0xffffffffUL;
 
 	usp -= len;
diff --git a/arch/sparc/include/asm/processor_64.h b/arch/sparc/include/asm/processor_64.h
index 4e5a483122a0..721e25f0e2ea 100644
--- a/arch/sparc/include/asm/processor_64.h
+++ b/arch/sparc/include/asm/processor_64.h
@@ -196,7 +196,22 @@ extern unsigned long get_wchan(struct task_struct *task);
 #define KSTK_EIP(tsk)  (task_pt_regs(tsk)->tpc)
 #define KSTK_ESP(tsk)  (task_pt_regs(tsk)->u_regs[UREG_FP])
 
-#define cpu_relax()	barrier()
+/* Please see the commentary in asm/backoff.h for a description of
+ * what these instructions are doing and how they have been choosen.
+ * To make a long story short, we are trying to yield the current cpu
+ * strand during busy loops.
+ */
+#define cpu_relax()	asm volatile("\n99:\n\t"			\
+				     "rd	%%ccr, %%g0\n\t"	\
+				     "rd	%%ccr, %%g0\n\t"	\
+				     "rd	%%ccr, %%g0\n\t"	\
+				     ".section	.pause_3insn_patch,\"ax\"\n\t"\
+				     ".word	99b\n\t"		\
+				     "wr	%%g0, 128, %%asr27\n\t"	\
+				     "nop\n\t"				\
+				     "nop\n\t"				\
+				     ".previous"			\
+				     ::: "memory")
 
 /* Prefetch support.  This is tuned for UltraSPARC-III and later.
  * UltraSPARC-I will treat these as nops, and UltraSPARC-II has
diff --git a/arch/sparc/include/asm/prom.h b/arch/sparc/include/asm/prom.h
index c28765110706..67c62578d170 100644
--- a/arch/sparc/include/asm/prom.h
+++ b/arch/sparc/include/asm/prom.h
@@ -63,5 +63,13 @@ extern char *of_console_options;
 extern void irq_trans_init(struct device_node *dp);
 extern char *build_path_component(struct device_node *dp);
 
+/* SPARC has local implementations */
+extern int of_address_to_resource(struct device_node *dev, int index,
+				  struct resource *r);
+#define of_address_to_resource of_address_to_resource
+
+void __iomem *of_iomap(struct device_node *node, int index);
+#define of_iomap of_iomap
+
 #endif /* __KERNEL__ */
 #endif /* _SPARC_PROM_H */
diff --git a/arch/sparc/include/asm/ptrace.h b/arch/sparc/include/asm/ptrace.h
index 0c6f6b068289..da43bdc62294 100644
--- a/arch/sparc/include/asm/ptrace.h
+++ b/arch/sparc/include/asm/ptrace.h
@@ -42,7 +42,18 @@ struct global_reg_snapshot {
 	struct thread_info	*thread;
 	unsigned long		pad1;
 };
-extern struct global_reg_snapshot global_reg_snapshot[NR_CPUS];
+
+struct global_pmu_snapshot {
+	unsigned long		pcr[4];
+	unsigned long		pic[4];
+};
+
+union global_cpu_snapshot {
+	struct global_reg_snapshot	reg;
+	struct global_pmu_snapshot	pmu;
+};
+
+extern union global_cpu_snapshot global_cpu_snapshot[NR_CPUS];
 
 #define force_successful_syscall_return()	    \
 do {	current_thread_info()->syscall_noerror = 1; \
diff --git a/arch/sparc/include/asm/smp_64.h b/arch/sparc/include/asm/smp_64.h
index 29862a9e9065..dd3bef4b9896 100644
--- a/arch/sparc/include/asm/smp_64.h
+++ b/arch/sparc/include/asm/smp_64.h
@@ -48,6 +48,7 @@ extern void smp_fill_in_sib_core_maps(void);
 extern void cpu_play_dead(void);
 
 extern void smp_fetch_global_regs(void);
+extern void smp_fetch_global_pmu(void);
 
 struct seq_file;
 void smp_bogo(struct seq_file *);
@@ -65,6 +66,7 @@ extern void __cpu_die(unsigned int cpu);
 #define hard_smp_processor_id()		0
 #define smp_fill_in_sib_core_maps() do { } while (0)
 #define smp_fetch_global_regs() do { } while (0)
+#define smp_fetch_global_pmu() do { } while (0)
 
 #endif /* !(CONFIG_SMP) */
 
diff --git a/arch/sparc/include/asm/thread_info_64.h b/arch/sparc/include/asm/thread_info_64.h
index 4e2276631081..a3fe4dcc0aa6 100644
--- a/arch/sparc/include/asm/thread_info_64.h
+++ b/arch/sparc/include/asm/thread_info_64.h
@@ -259,6 +259,11 @@ static inline bool test_and_clear_restore_sigmask(void)
 
 #define tsk_is_polling(t) test_tsk_thread_flag(t, TIF_POLLING_NRFLAG)
 
+#define thread32_stack_is_64bit(__SP) (((__SP) & 0x1) != 0)
+#define test_thread_64bit_stack(__SP) \
+	((test_thread_flag(TIF_32BIT) && !thread32_stack_is_64bit(__SP)) ? \
+	 false : true)
+
 #endif	/* !__ASSEMBLY__ */
 
 #endif /* __KERNEL__ */
diff --git a/arch/sparc/include/asm/ttable.h b/arch/sparc/include/asm/ttable.h
index 48f2807d3265..71b5a67522ab 100644
--- a/arch/sparc/include/asm/ttable.h
+++ b/arch/sparc/include/asm/ttable.h
@@ -372,7 +372,9 @@ etrap_spill_fixup_64bit:				\
 
 /* Normal 32bit spill */
 #define SPILL_2_GENERIC(ASI)				\
-	srl	%sp, 0, %sp;				\
+	and	%sp, 1, %g3;				\
+	brnz,pn	%g3, (. - (128 + 4));			\
+	 srl	%sp, 0, %sp;				\
 	stwa	%l0, [%sp + %g0] ASI;			\
 	mov	0x04, %g3;				\
 	stwa	%l1, [%sp + %g3] ASI;			\
@@ -398,14 +400,16 @@ etrap_spill_fixup_64bit:				\
 	stwa	%i6, [%g1 + %g0] ASI;			\
 	stwa	%i7, [%g1 + %g3] ASI;			\
 	saved;						\
-        retry; nop; nop;				\
+        retry;						\
 	b,a,pt	%xcc, spill_fixup_dax;			\
 	b,a,pt	%xcc, spill_fixup_mna;			\
 	b,a,pt	%xcc, spill_fixup;
 
 #define SPILL_2_GENERIC_ETRAP		\
 etrap_user_spill_32bit:			\
-	srl	%sp, 0, %sp;		\
+	and	%sp, 1, %g3;		\
+	brnz,pn	%g3, etrap_user_spill_64bit;	\
+	 srl	%sp, 0, %sp;		\
 	stwa	%l0, [%sp + 0x00] %asi;	\
 	stwa	%l1, [%sp + 0x04] %asi;	\
 	stwa	%l2, [%sp + 0x08] %asi;	\
@@ -427,7 +431,7 @@ etrap_user_spill_32bit:			\
 	ba,pt	%xcc, etrap_save;	\
 	 wrpr	%g1, %cwp;		\
 	nop; nop; nop; nop;		\
-	nop; nop; nop; nop;		\
+	nop; nop;			\
 	ba,a,pt	%xcc, etrap_spill_fixup_32bit; \
 	ba,a,pt	%xcc, etrap_spill_fixup_32bit; \
 	ba,a,pt	%xcc, etrap_spill_fixup_32bit;
@@ -592,7 +596,9 @@ user_rtt_fill_64bit:					\
 
 /* Normal 32bit fill */
 #define FILL_2_GENERIC(ASI)				\
-	srl	%sp, 0, %sp;				\
+	and	%sp, 1, %g3;				\
+	brnz,pn	%g3, (. - (128 + 4));			\
+	 srl	%sp, 0, %sp;				\
 	lduwa	[%sp + %g0] ASI, %l0;			\
 	mov	0x04, %g2;				\
 	mov	0x08, %g3;				\
@@ -616,14 +622,16 @@ user_rtt_fill_64bit:					\
 	lduwa	[%g1 + %g3] ASI, %i6;			\
 	lduwa	[%g1 + %g5] ASI, %i7;			\
 	restored;					\
-	retry; nop; nop; nop; nop;			\
+	retry; nop; nop;				\
 	b,a,pt	%xcc, fill_fixup_dax;			\
 	b,a,pt	%xcc, fill_fixup_mna;			\
 	b,a,pt	%xcc, fill_fixup;
 
 #define FILL_2_GENERIC_RTRAP				\
 user_rtt_fill_32bit:					\
-	srl	%sp, 0, %sp;				\
+	and	%sp, 1, %g3;				\
+	brnz,pn	%g3, user_rtt_fill_64bit;		\
+	 srl	%sp, 0, %sp;				\
 	lduwa	[%sp + 0x00] %asi, %l0;			\
 	lduwa	[%sp + 0x04] %asi, %l1;			\
 	lduwa	[%sp + 0x08] %asi, %l2;			\
@@ -643,7 +651,7 @@ user_rtt_fill_32bit:					\
 	ba,pt	%xcc, user_rtt_pre_restore;		\
 	 restored;					\
 	nop; nop; nop; nop; nop;			\
-	nop; nop; nop; nop; nop;			\
+	nop; nop; nop;					\
 	ba,a,pt	%xcc, user_rtt_fill_fixup;		\
 	ba,a,pt	%xcc, user_rtt_fill_fixup;		\
 	ba,a,pt	%xcc, user_rtt_fill_fixup;
diff --git a/arch/sparc/include/uapi/asm/sigcontext.h b/arch/sparc/include/uapi/asm/sigcontext.h
index e69de29bb2d1..ae5704fa77ad 100644
--- a/arch/sparc/include/uapi/asm/sigcontext.h
+++ b/arch/sparc/include/uapi/asm/sigcontext.h
@@ -0,0 +1,4 @@
+/*
+ * There isn't anything here anymore, but the file must not be empty or patch
+ * will delete it.
+ */
diff --git a/arch/sparc/include/uapi/asm/unistd.h b/arch/sparc/include/uapi/asm/unistd.h
index 8974ef7ae920..cac719d1bc5c 100644
--- a/arch/sparc/include/uapi/asm/unistd.h
+++ b/arch/sparc/include/uapi/asm/unistd.h
@@ -405,8 +405,13 @@
 #define __NR_setns		337
 #define __NR_process_vm_readv	338
 #define __NR_process_vm_writev	339
+#define __NR_kern_features	340
+#define __NR_kcmp		341
 
-#define NR_syscalls		340
+#define NR_syscalls		342
+
+/* Bitmask values returned from kern_features system call.  */
+#define KERN_FEATURE_MIXED_MODE_STACK	0x00000001
 
 #ifdef __32bit_syscall_numbers__
 /* Sparc 32-bit only has the "setresuid32", "getresuid32" variants,
diff --git a/arch/sparc/kernel/entry.h b/arch/sparc/kernel/entry.h
index 0c218e4c0881..cc3c5cb47cda 100644
--- a/arch/sparc/kernel/entry.h
+++ b/arch/sparc/kernel/entry.h
@@ -59,6 +59,13 @@ struct popc_6insn_patch_entry {
 extern struct popc_6insn_patch_entry __popc_6insn_patch,
 	__popc_6insn_patch_end;
 
+struct pause_patch_entry {
+	unsigned int	addr;
+	unsigned int	insns[3];
+};
+extern struct pause_patch_entry __pause_3insn_patch,
+	__pause_3insn_patch_end;
+
 extern void __init per_cpu_patch(void);
 extern void sun4v_patch_1insn_range(struct sun4v_1insn_patch_entry *,
 				    struct sun4v_1insn_patch_entry *);
diff --git a/arch/sparc/kernel/leon_kernel.c b/arch/sparc/kernel/leon_kernel.c
index f8b6eee40bde..87f60ee65433 100644
--- a/arch/sparc/kernel/leon_kernel.c
+++ b/arch/sparc/kernel/leon_kernel.c
@@ -56,11 +56,13 @@ static inline unsigned int leon_eirq_get(int cpu)
 static void leon_handle_ext_irq(unsigned int irq, struct irq_desc *desc)
 {
 	unsigned int eirq;
+	struct irq_bucket *p;
 	int cpu = sparc_leon3_cpuid();
 
 	eirq = leon_eirq_get(cpu);
-	if ((eirq & 0x10) && irq_map[eirq]->irq) /* bit4 tells if IRQ happened */
-		generic_handle_irq(irq_map[eirq]->irq);
+	p = irq_map[eirq];
+	if ((eirq & 0x10) && p && p->irq) /* bit4 tells if IRQ happened */
+		generic_handle_irq(p->irq);
 }
 
 /* The extended IRQ controller has been found, this function registers it */
diff --git a/arch/sparc/kernel/perf_event.c b/arch/sparc/kernel/perf_event.c
index e48651dace1b..b5c38faa4ead 100644
--- a/arch/sparc/kernel/perf_event.c
+++ b/arch/sparc/kernel/perf_event.c
@@ -817,15 +817,17 @@ static u64 nop_for_index(int idx)
 
 static inline void sparc_pmu_enable_event(struct cpu_hw_events *cpuc, struct hw_perf_event *hwc, int idx)
 {
-	u64 val, mask = mask_for_index(idx);
+	u64 enc, val, mask = mask_for_index(idx);
 	int pcr_index = 0;
 
 	if (sparc_pmu->num_pcrs > 1)
 		pcr_index = idx;
 
+	enc = perf_event_get_enc(cpuc->events[idx]);
+
 	val = cpuc->pcr[pcr_index];
 	val &= ~mask;
-	val |= hwc->config;
+	val |= event_encoding(enc, idx);
 	cpuc->pcr[pcr_index] = val;
 
 	pcr_ops->write_pcr(pcr_index, cpuc->pcr[pcr_index]);
@@ -1738,8 +1740,6 @@ static void perf_callchain_user_64(struct perf_callchain_entry *entry,
 {
 	unsigned long ufp;
 
-	perf_callchain_store(entry, regs->tpc);
-
 	ufp = regs->u_regs[UREG_I6] + STACK_BIAS;
 	do {
 		struct sparc_stackf *usf, sf;
@@ -1760,19 +1760,27 @@ static void perf_callchain_user_32(struct perf_callchain_entry *entry,
 {
 	unsigned long ufp;
 
-	perf_callchain_store(entry, regs->tpc);
-
 	ufp = regs->u_regs[UREG_I6] & 0xffffffffUL;
 	do {
-		struct sparc_stackf32 *usf, sf;
 		unsigned long pc;
 
-		usf = (struct sparc_stackf32 *) ufp;
-		if (__copy_from_user_inatomic(&sf, usf, sizeof(sf)))
-			break;
+		if (thread32_stack_is_64bit(ufp)) {
+			struct sparc_stackf *usf, sf;
 
-		pc = sf.callers_pc;
-		ufp = (unsigned long)sf.fp;
+			ufp += STACK_BIAS;
+			usf = (struct sparc_stackf *) ufp;
+			if (__copy_from_user_inatomic(&sf, usf, sizeof(sf)))
+				break;
+			pc = sf.callers_pc & 0xffffffff;
+			ufp = ((unsigned long) sf.fp) & 0xffffffff;
+		} else {
+			struct sparc_stackf32 *usf, sf;
+			usf = (struct sparc_stackf32 *) ufp;
+			if (__copy_from_user_inatomic(&sf, usf, sizeof(sf)))
+				break;
+			pc = sf.callers_pc;
+			ufp = (unsigned long)sf.fp;
+		}
 		perf_callchain_store(entry, pc);
 	} while (entry->nr < PERF_MAX_STACK_DEPTH);
 }
@@ -1780,6 +1788,11 @@ static void perf_callchain_user_32(struct perf_callchain_entry *entry,
 void
 perf_callchain_user(struct perf_callchain_entry *entry, struct pt_regs *regs)
 {
+	perf_callchain_store(entry, regs->tpc);
+
+	if (!current->mm)
+		return;
+
 	flushw_user();
 	if (test_thread_flag(TIF_32BIT))
 		perf_callchain_user_32(entry, regs);
diff --git a/arch/sparc/kernel/process_64.c b/arch/sparc/kernel/process_64.c
index fcaa59421126..c6e0c2910043 100644
--- a/arch/sparc/kernel/process_64.c
+++ b/arch/sparc/kernel/process_64.c
@@ -27,6 +27,7 @@
 #include <linux/tick.h>
 #include <linux/init.h>
 #include <linux/cpu.h>
+#include <linux/perf_event.h>
 #include <linux/elfcore.h>
 #include <linux/sysrq.h>
 #include <linux/nmi.h>
@@ -47,6 +48,7 @@
 #include <asm/syscalls.h>
 #include <asm/irq_regs.h>
 #include <asm/smp.h>
+#include <asm/pcr.h>
 
 #include "kstack.h"
 
@@ -204,18 +206,22 @@ void show_regs(struct pt_regs *regs)
 	show_stack(current, (unsigned long *) regs->u_regs[UREG_FP]);
 }
 
-struct global_reg_snapshot global_reg_snapshot[NR_CPUS];
-static DEFINE_SPINLOCK(global_reg_snapshot_lock);
+union global_cpu_snapshot global_cpu_snapshot[NR_CPUS];
+static DEFINE_SPINLOCK(global_cpu_snapshot_lock);
 
 static void __global_reg_self(struct thread_info *tp, struct pt_regs *regs,
 			      int this_cpu)
 {
+	struct global_reg_snapshot *rp;
+
 	flushw_all();
 
-	global_reg_snapshot[this_cpu].tstate = regs->tstate;
-	global_reg_snapshot[this_cpu].tpc = regs->tpc;
-	global_reg_snapshot[this_cpu].tnpc = regs->tnpc;
-	global_reg_snapshot[this_cpu].o7 = regs->u_regs[UREG_I7];
+	rp = &global_cpu_snapshot[this_cpu].reg;
+
+	rp->tstate = regs->tstate;
+	rp->tpc = regs->tpc;
+	rp->tnpc = regs->tnpc;
+	rp->o7 = regs->u_regs[UREG_I7];
 
 	if (regs->tstate & TSTATE_PRIV) {
 		struct reg_window *rw;
@@ -223,17 +229,17 @@ static void __global_reg_self(struct thread_info *tp, struct pt_regs *regs,
 		rw = (struct reg_window *)
 			(regs->u_regs[UREG_FP] + STACK_BIAS);
 		if (kstack_valid(tp, (unsigned long) rw)) {
-			global_reg_snapshot[this_cpu].i7 = rw->ins[7];
+			rp->i7 = rw->ins[7];
 			rw = (struct reg_window *)
 				(rw->ins[6] + STACK_BIAS);
 			if (kstack_valid(tp, (unsigned long) rw))
-				global_reg_snapshot[this_cpu].rpc = rw->ins[7];
+				rp->rpc = rw->ins[7];
 		}
 	} else {
-		global_reg_snapshot[this_cpu].i7 = 0;
-		global_reg_snapshot[this_cpu].rpc = 0;
+		rp->i7 = 0;
+		rp->rpc = 0;
 	}
-	global_reg_snapshot[this_cpu].thread = tp;
+	rp->thread = tp;
 }
 
 /* In order to avoid hangs we do not try to synchronize with the
@@ -261,9 +267,9 @@ void arch_trigger_all_cpu_backtrace(void)
 	if (!regs)
 		regs = tp->kregs;
 
-	spin_lock_irqsave(&global_reg_snapshot_lock, flags);
+	spin_lock_irqsave(&global_cpu_snapshot_lock, flags);
 
-	memset(global_reg_snapshot, 0, sizeof(global_reg_snapshot));
+	memset(global_cpu_snapshot, 0, sizeof(global_cpu_snapshot));
 
 	this_cpu = raw_smp_processor_id();
 
@@ -272,7 +278,7 @@ void arch_trigger_all_cpu_backtrace(void)
 	smp_fetch_global_regs();
 
 	for_each_online_cpu(cpu) {
-		struct global_reg_snapshot *gp = &global_reg_snapshot[cpu];
+		struct global_reg_snapshot *gp = &global_cpu_snapshot[cpu].reg;
 
 		__global_reg_poll(gp);
 
@@ -295,9 +301,9 @@ void arch_trigger_all_cpu_backtrace(void)
 		}
 	}
 
-	memset(global_reg_snapshot, 0, sizeof(global_reg_snapshot));
+	memset(global_cpu_snapshot, 0, sizeof(global_cpu_snapshot));
 
-	spin_unlock_irqrestore(&global_reg_snapshot_lock, flags);
+	spin_unlock_irqrestore(&global_cpu_snapshot_lock, flags);
 }
 
 #ifdef CONFIG_MAGIC_SYSRQ
@@ -309,16 +315,90 @@ static void sysrq_handle_globreg(int key)
 
 static struct sysrq_key_op sparc_globalreg_op = {
 	.handler	= sysrq_handle_globreg,
-	.help_msg	= "Globalregs",
+	.help_msg	= "global-regs(Y)",
 	.action_msg	= "Show Global CPU Regs",
 };
 
-static int __init sparc_globreg_init(void)
+static void __global_pmu_self(int this_cpu)
 {
-	return register_sysrq_key('y', &sparc_globalreg_op);
+	struct global_pmu_snapshot *pp;
+	int i, num;
+
+	pp = &global_cpu_snapshot[this_cpu].pmu;
+
+	num = 1;
+	if (tlb_type == hypervisor &&
+	    sun4v_chip_type >= SUN4V_CHIP_NIAGARA4)
+		num = 4;
+
+	for (i = 0; i < num; i++) {
+		pp->pcr[i] = pcr_ops->read_pcr(i);
+		pp->pic[i] = pcr_ops->read_pic(i);
+	}
 }
 
-core_initcall(sparc_globreg_init);
+static void __global_pmu_poll(struct global_pmu_snapshot *pp)
+{
+	int limit = 0;
+
+	while (!pp->pcr[0] && ++limit < 100) {
+		barrier();
+		udelay(1);
+	}
+}
+
+static void pmu_snapshot_all_cpus(void)
+{
+	unsigned long flags;
+	int this_cpu, cpu;
+
+	spin_lock_irqsave(&global_cpu_snapshot_lock, flags);
+
+	memset(global_cpu_snapshot, 0, sizeof(global_cpu_snapshot));
+
+	this_cpu = raw_smp_processor_id();
+
+	__global_pmu_self(this_cpu);
+
+	smp_fetch_global_pmu();
+
+	for_each_online_cpu(cpu) {
+		struct global_pmu_snapshot *pp = &global_cpu_snapshot[cpu].pmu;
+
+		__global_pmu_poll(pp);
+
+		printk("%c CPU[%3d]: PCR[%08lx:%08lx:%08lx:%08lx] PIC[%08lx:%08lx:%08lx:%08lx]\n",
+		       (cpu == this_cpu ? '*' : ' '), cpu,
+		       pp->pcr[0], pp->pcr[1], pp->pcr[2], pp->pcr[3],
+		       pp->pic[0], pp->pic[1], pp->pic[2], pp->pic[3]);
+	}
+
+	memset(global_cpu_snapshot, 0, sizeof(global_cpu_snapshot));
+
+	spin_unlock_irqrestore(&global_cpu_snapshot_lock, flags);
+}
+
+static void sysrq_handle_globpmu(int key)
+{
+	pmu_snapshot_all_cpus();
+}
+
+static struct sysrq_key_op sparc_globalpmu_op = {
+	.handler	= sysrq_handle_globpmu,
+	.help_msg	= "global-pmu(X)",
+	.action_msg	= "Show Global PMU Regs",
+};
+
+static int __init sparc_sysrq_init(void)
+{
+	int ret = register_sysrq_key('y', &sparc_globalreg_op);
+
+	if (!ret)
+		ret = register_sysrq_key('x', &sparc_globalpmu_op);
+	return ret;
+}
+
+core_initcall(sparc_sysrq_init);
 
 #endif
 
@@ -372,13 +452,16 @@ void flush_thread(void)
 /* It's a bit more tricky when 64-bit tasks are involved... */
 static unsigned long clone_stackframe(unsigned long csp, unsigned long psp)
 {
+	bool stack_64bit = test_thread_64bit_stack(psp);
 	unsigned long fp, distance, rval;
 
-	if (!(test_thread_flag(TIF_32BIT))) {
+	if (stack_64bit) {
 		csp += STACK_BIAS;
 		psp += STACK_BIAS;
 		__get_user(fp, &(((struct reg_window __user *)psp)->ins[6]));
 		fp += STACK_BIAS;
+		if (test_thread_flag(TIF_32BIT))
+			fp &= 0xffffffff;
 	} else
 		__get_user(fp, &(((struct reg_window32 __user *)psp)->ins[6]));
 
@@ -392,7 +475,7 @@ static unsigned long clone_stackframe(unsigned long csp, unsigned long psp)
 	rval = (csp - distance);
 	if (copy_in_user((void __user *) rval, (void __user *) psp, distance))
 		rval = 0;
-	else if (test_thread_flag(TIF_32BIT)) {
+	else if (!stack_64bit) {
 		if (put_user(((u32)csp),
 			     &(((struct reg_window32 __user *)rval)->ins[6])))
 			rval = 0;
@@ -427,18 +510,18 @@ void synchronize_user_stack(void)
 
 	flush_user_windows();
 	if ((window = get_thread_wsaved()) != 0) {
-		int winsize = sizeof(struct reg_window);
-		int bias = 0;
-
-		if (test_thread_flag(TIF_32BIT))
-			winsize = sizeof(struct reg_window32);
-		else
-			bias = STACK_BIAS;
-
 		window -= 1;
 		do {
-			unsigned long sp = (t->rwbuf_stkptrs[window] + bias);
 			struct reg_window *rwin = &t->reg_window[window];
+			int winsize = sizeof(struct reg_window);
+			unsigned long sp;
+
+			sp = t->rwbuf_stkptrs[window];
+
+			if (test_thread_64bit_stack(sp))
+				sp += STACK_BIAS;
+			else
+				winsize = sizeof(struct reg_window32);
 
 			if (!copy_to_user((char __user *)sp, rwin, winsize)) {
 				shift_window_buffer(window, get_thread_wsaved() - 1, t);
@@ -464,13 +547,6 @@ void fault_in_user_windows(void)
 {
 	struct thread_info *t = current_thread_info();
 	unsigned long window;
-	int winsize = sizeof(struct reg_window);
-	int bias = 0;
-
-	if (test_thread_flag(TIF_32BIT))
-		winsize = sizeof(struct reg_window32);
-	else
-		bias = STACK_BIAS;
 
 	flush_user_windows();
 	window = get_thread_wsaved();
@@ -478,8 +554,16 @@ void fault_in_user_windows(void)
 	if (likely(window != 0)) {
 		window -= 1;
 		do {
-			unsigned long sp = (t->rwbuf_stkptrs[window] + bias);
 			struct reg_window *rwin = &t->reg_window[window];
+			int winsize = sizeof(struct reg_window);
+			unsigned long sp;
+
+			sp = t->rwbuf_stkptrs[window];
+
+			if (test_thread_64bit_stack(sp))
+				sp += STACK_BIAS;
+			else
+				winsize = sizeof(struct reg_window32);
 
 			if (unlikely(sp & 0x7UL))
 				stack_unaligned(sp);
diff --git a/arch/sparc/kernel/ptrace_64.c b/arch/sparc/kernel/ptrace_64.c
index 484dabac7045..7ff45e4ba681 100644
--- a/arch/sparc/kernel/ptrace_64.c
+++ b/arch/sparc/kernel/ptrace_64.c
@@ -151,7 +151,7 @@ static int regwindow64_get(struct task_struct *target,
 {
 	unsigned long rw_addr = regs->u_regs[UREG_I6];
 
-	if (test_tsk_thread_flag(current, TIF_32BIT)) {
+	if (!test_thread_64bit_stack(rw_addr)) {
 		struct reg_window32 win32;
 		int i;
 
@@ -176,7 +176,7 @@ static int regwindow64_set(struct task_struct *target,
 {
 	unsigned long rw_addr = regs->u_regs[UREG_I6];
 
-	if (test_tsk_thread_flag(current, TIF_32BIT)) {
+	if (!test_thread_64bit_stack(rw_addr)) {
 		struct reg_window32 win32;
 		int i;
 
diff --git a/arch/sparc/kernel/setup_64.c b/arch/sparc/kernel/setup_64.c
index 0800e71d8a88..0eaf0059aaef 100644
--- a/arch/sparc/kernel/setup_64.c
+++ b/arch/sparc/kernel/setup_64.c
@@ -316,6 +316,25 @@ static void __init popc_patch(void)
 	}
 }
 
+static void __init pause_patch(void)
+{
+	struct pause_patch_entry *p;
+
+	p = &__pause_3insn_patch;
+	while (p < &__pause_3insn_patch_end) {
+		unsigned long i, addr = p->addr;
+
+		for (i = 0; i < 3; i++) {
+			*(unsigned int *) (addr +  (i * 4)) = p->insns[i];
+			wmb();
+			__asm__ __volatile__("flush	%0"
+					     : : "r" (addr +  (i * 4)));
+		}
+
+		p++;
+	}
+}
+
 #ifdef CONFIG_SMP
 void __init boot_cpu_id_too_large(int cpu)
 {
@@ -528,6 +547,8 @@ static void __init init_sparc64_elf_hwcap(void)
 
 	if (sparc64_elf_hwcap & AV_SPARC_POPC)
 		popc_patch();
+	if (sparc64_elf_hwcap & AV_SPARC_PAUSE)
+		pause_patch();
 }
 
 void __init setup_arch(char **cmdline_p)
diff --git a/arch/sparc/kernel/signal_64.c b/arch/sparc/kernel/signal_64.c
index 867de2f8189c..689e1ba62809 100644
--- a/arch/sparc/kernel/signal_64.c
+++ b/arch/sparc/kernel/signal_64.c
@@ -295,9 +295,7 @@ void do_rt_sigreturn(struct pt_regs *regs)
 		err |= restore_fpu_state(regs, fpu_save);
 
 	err |= __copy_from_user(&set, &sf->mask, sizeof(sigset_t));
-	err |= do_sigaltstack(&sf->stack, NULL, (unsigned long)sf);
-
-	if (err)
+	if (err || do_sigaltstack(&sf->stack, NULL, (unsigned long)sf) == -EFAULT)
 		goto segv;
 
 	err |= __get_user(rwin_save, &sf->rwin_save);
diff --git a/arch/sparc/kernel/smp_64.c b/arch/sparc/kernel/smp_64.c
index 781bcb10b8bd..d94b878577b7 100644
--- a/arch/sparc/kernel/smp_64.c
+++ b/arch/sparc/kernel/smp_64.c
@@ -852,6 +852,8 @@ extern unsigned long xcall_flush_tlb_mm;
 extern unsigned long xcall_flush_tlb_pending;
 extern unsigned long xcall_flush_tlb_kernel_range;
 extern unsigned long xcall_fetch_glob_regs;
+extern unsigned long xcall_fetch_glob_pmu;
+extern unsigned long xcall_fetch_glob_pmu_n4;
 extern unsigned long xcall_receive_signal;
 extern unsigned long xcall_new_mmu_context_version;
 #ifdef CONFIG_KGDB
@@ -1000,6 +1002,15 @@ void smp_fetch_global_regs(void)
 	smp_cross_call(&xcall_fetch_glob_regs, 0, 0, 0);
 }
 
+void smp_fetch_global_pmu(void)
+{
+	if (tlb_type == hypervisor &&
+	    sun4v_chip_type >= SUN4V_CHIP_NIAGARA4)
+		smp_cross_call(&xcall_fetch_glob_pmu_n4, 0, 0, 0);
+	else
+		smp_cross_call(&xcall_fetch_glob_pmu, 0, 0, 0);
+}
+
 /* We know that the window frames of the user have been flushed
  * to the stack before we get here because all callers of us
  * are flush_tlb_*() routines, and these run after flush_cache_*()
diff --git a/arch/sparc/kernel/sys_sparc_64.c b/arch/sparc/kernel/sys_sparc_64.c
index 11c6c9603e71..878ef3d5fec5 100644
--- a/arch/sparc/kernel/sys_sparc_64.c
+++ b/arch/sparc/kernel/sys_sparc_64.c
@@ -751,3 +751,8 @@ int kernel_execve(const char *filename,
 		      : "cc");
 	return __res;
 }
+
+asmlinkage long sys_kern_features(void)
+{
+	return KERN_FEATURE_MIXED_MODE_STACK;
+}
diff --git a/arch/sparc/kernel/systbls_32.S b/arch/sparc/kernel/systbls_32.S
index 63402f9e9f51..5147f574f125 100644
--- a/arch/sparc/kernel/systbls_32.S
+++ b/arch/sparc/kernel/systbls_32.S
@@ -85,3 +85,4 @@ sys_call_table:
 /*325*/	.long sys_pwritev, sys_rt_tgsigqueueinfo, sys_perf_event_open, sys_recvmmsg, sys_fanotify_init
 /*330*/	.long sys_fanotify_mark, sys_prlimit64, sys_name_to_handle_at, sys_open_by_handle_at, sys_clock_adjtime
 /*335*/	.long sys_syncfs, sys_sendmmsg, sys_setns, sys_process_vm_readv, sys_process_vm_writev
+/*340*/	.long sys_ni_syscall, sys_kcmp
diff --git a/arch/sparc/kernel/systbls_64.S b/arch/sparc/kernel/systbls_64.S
index 3a58e0d66f51..1c9af9fa38e9 100644
--- a/arch/sparc/kernel/systbls_64.S
+++ b/arch/sparc/kernel/systbls_64.S
@@ -86,6 +86,7 @@ sys_call_table32:
 	.word compat_sys_pwritev, compat_sys_rt_tgsigqueueinfo, sys_perf_event_open, compat_sys_recvmmsg, sys_fanotify_init
 /*330*/	.word sys32_fanotify_mark, sys_prlimit64, sys_name_to_handle_at, compat_sys_open_by_handle_at, compat_sys_clock_adjtime
 	.word sys_syncfs, compat_sys_sendmmsg, sys_setns, compat_sys_process_vm_readv, compat_sys_process_vm_writev
+/*340*/	.word sys_kern_features, sys_kcmp
 
 #endif /* CONFIG_COMPAT */
 
@@ -163,3 +164,4 @@ sys_call_table:
 	.word sys_pwritev, sys_rt_tgsigqueueinfo, sys_perf_event_open, sys_recvmmsg, sys_fanotify_init
 /*330*/	.word sys_fanotify_mark, sys_prlimit64, sys_name_to_handle_at, sys_open_by_handle_at, sys_clock_adjtime
 	.word sys_syncfs, sys_sendmmsg, sys_setns, sys_process_vm_readv, sys_process_vm_writev
+/*340*/	.word sys_kern_features, sys_kcmp
diff --git a/arch/sparc/kernel/unaligned_64.c b/arch/sparc/kernel/unaligned_64.c
index f81d038f7340..8201c25e7669 100644
--- a/arch/sparc/kernel/unaligned_64.c
+++ b/arch/sparc/kernel/unaligned_64.c
@@ -113,21 +113,24 @@ static inline long sign_extend_imm13(long imm)
 
 static unsigned long fetch_reg(unsigned int reg, struct pt_regs *regs)
 {
-	unsigned long value;
+	unsigned long value, fp;
 	
 	if (reg < 16)
 		return (!reg ? 0 : regs->u_regs[reg]);
+
+	fp = regs->u_regs[UREG_FP];
+
 	if (regs->tstate & TSTATE_PRIV) {
 		struct reg_window *win;
-		win = (struct reg_window *)(regs->u_regs[UREG_FP] + STACK_BIAS);
+		win = (struct reg_window *)(fp + STACK_BIAS);
 		value = win->locals[reg - 16];
-	} else if (test_thread_flag(TIF_32BIT)) {
+	} else if (!test_thread_64bit_stack(fp)) {
 		struct reg_window32 __user *win32;
-		win32 = (struct reg_window32 __user *)((unsigned long)((u32)regs->u_regs[UREG_FP]));
+		win32 = (struct reg_window32 __user *)((unsigned long)((u32)fp));
 		get_user(value, &win32->locals[reg - 16]);
 	} else {
 		struct reg_window __user *win;
-		win = (struct reg_window __user *)(regs->u_regs[UREG_FP] + STACK_BIAS);
+		win = (struct reg_window __user *)(fp + STACK_BIAS);
 		get_user(value, &win->locals[reg - 16]);
 	}
 	return value;
@@ -135,19 +138,24 @@ static unsigned long fetch_reg(unsigned int reg, struct pt_regs *regs)
 
 static unsigned long *fetch_reg_addr(unsigned int reg, struct pt_regs *regs)
 {
+	unsigned long fp;
+
 	if (reg < 16)
 		return &regs->u_regs[reg];
+
+	fp = regs->u_regs[UREG_FP];
+
 	if (regs->tstate & TSTATE_PRIV) {
 		struct reg_window *win;
-		win = (struct reg_window *)(regs->u_regs[UREG_FP] + STACK_BIAS);
+		win = (struct reg_window *)(fp + STACK_BIAS);
 		return &win->locals[reg - 16];
-	} else if (test_thread_flag(TIF_32BIT)) {
+	} else if (!test_thread_64bit_stack(fp)) {
 		struct reg_window32 *win32;
-		win32 = (struct reg_window32 *)((unsigned long)((u32)regs->u_regs[UREG_FP]));
+		win32 = (struct reg_window32 *)((unsigned long)((u32)fp));
 		return (unsigned long *)&win32->locals[reg - 16];
 	} else {
 		struct reg_window *win;
-		win = (struct reg_window *)(regs->u_regs[UREG_FP] + STACK_BIAS);
+		win = (struct reg_window *)(fp + STACK_BIAS);
 		return &win->locals[reg - 16];
 	}
 }
@@ -392,13 +400,15 @@ int handle_popc(u32 insn, struct pt_regs *regs)
 		if (rd)
 			regs->u_regs[rd] = ret;
 	} else {
-		if (test_thread_flag(TIF_32BIT)) {
+		unsigned long fp = regs->u_regs[UREG_FP];
+
+		if (!test_thread_64bit_stack(fp)) {
 			struct reg_window32 __user *win32;
-			win32 = (struct reg_window32 __user *)((unsigned long)((u32)regs->u_regs[UREG_FP]));
+			win32 = (struct reg_window32 __user *)((unsigned long)((u32)fp));
 			put_user(ret, &win32->locals[rd - 16]);
 		} else {
 			struct reg_window __user *win;
-			win = (struct reg_window __user *)(regs->u_regs[UREG_FP] + STACK_BIAS);
+			win = (struct reg_window __user *)(fp + STACK_BIAS);
 			put_user(ret, &win->locals[rd - 16]);
 		}
 	}
@@ -554,7 +564,7 @@ void handle_ld_nf(u32 insn, struct pt_regs *regs)
 		reg[0] = 0;
 		if ((insn & 0x780000) == 0x180000)
 			reg[1] = 0;
-	} else if (test_thread_flag(TIF_32BIT)) {
+	} else if (!test_thread_64bit_stack(regs->u_regs[UREG_FP])) {
 		put_user(0, (int __user *) reg);
 		if ((insn & 0x780000) == 0x180000)
 			put_user(0, ((int __user *) reg) + 1);
diff --git a/arch/sparc/kernel/visemul.c b/arch/sparc/kernel/visemul.c
index 08e074b7eb6a..c096c624ac4d 100644
--- a/arch/sparc/kernel/visemul.c
+++ b/arch/sparc/kernel/visemul.c
@@ -149,21 +149,24 @@ static inline void maybe_flush_windows(unsigned int rs1, unsigned int rs2,
 
 static unsigned long fetch_reg(unsigned int reg, struct pt_regs *regs)
 {
-	unsigned long value;
+	unsigned long value, fp;
 	
 	if (reg < 16)
 		return (!reg ? 0 : regs->u_regs[reg]);
+
+	fp = regs->u_regs[UREG_FP];
+
 	if (regs->tstate & TSTATE_PRIV) {
 		struct reg_window *win;
-		win = (struct reg_window *)(regs->u_regs[UREG_FP] + STACK_BIAS);
+		win = (struct reg_window *)(fp + STACK_BIAS);
 		value = win->locals[reg - 16];
-	} else if (test_thread_flag(TIF_32BIT)) {
+	} else if (!test_thread_64bit_stack(fp)) {
 		struct reg_window32 __user *win32;
-		win32 = (struct reg_window32 __user *)((unsigned long)((u32)regs->u_regs[UREG_FP]));
+		win32 = (struct reg_window32 __user *)((unsigned long)((u32)fp));
 		get_user(value, &win32->locals[reg - 16]);
 	} else {
 		struct reg_window __user *win;
-		win = (struct reg_window __user *)(regs->u_regs[UREG_FP] + STACK_BIAS);
+		win = (struct reg_window __user *)(fp + STACK_BIAS);
 		get_user(value, &win->locals[reg - 16]);
 	}
 	return value;
@@ -172,16 +175,18 @@ static unsigned long fetch_reg(unsigned int reg, struct pt_regs *regs)
 static inline unsigned long __user *__fetch_reg_addr_user(unsigned int reg,
 							  struct pt_regs *regs)
 {
+	unsigned long fp = regs->u_regs[UREG_FP];
+
 	BUG_ON(reg < 16);
 	BUG_ON(regs->tstate & TSTATE_PRIV);
 
-	if (test_thread_flag(TIF_32BIT)) {
+	if (!test_thread_64bit_stack(fp)) {
 		struct reg_window32 __user *win32;
-		win32 = (struct reg_window32 __user *)((unsigned long)((u32)regs->u_regs[UREG_FP]));
+		win32 = (struct reg_window32 __user *)((unsigned long)((u32)fp));
 		return (unsigned long __user *)&win32->locals[reg - 16];
 	} else {
 		struct reg_window __user *win;
-		win = (struct reg_window __user *)(regs->u_regs[UREG_FP] + STACK_BIAS);
+		win = (struct reg_window __user *)(fp + STACK_BIAS);
 		return &win->locals[reg - 16];
 	}
 }
@@ -204,7 +209,7 @@ static void store_reg(struct pt_regs *regs, unsigned long val, unsigned long rd)
 	} else {
 		unsigned long __user *rd_user = __fetch_reg_addr_user(rd, regs);
 
-		if (test_thread_flag(TIF_32BIT))
+		if (!test_thread_64bit_stack(regs->u_regs[UREG_FP]))
 			__put_user((u32)val, (u32 __user *)rd_user);
 		else
 			__put_user(val, rd_user);
diff --git a/arch/sparc/kernel/vmlinux.lds.S b/arch/sparc/kernel/vmlinux.lds.S
index 89c2c29f154b..0bacceb19150 100644
--- a/arch/sparc/kernel/vmlinux.lds.S
+++ b/arch/sparc/kernel/vmlinux.lds.S
@@ -132,6 +132,11 @@ SECTIONS
 		*(.popc_6insn_patch)
 		__popc_6insn_patch_end = .;
 	}
+	.pause_3insn_patch : {
+		__pause_3insn_patch = .;
+		*(.pause_3insn_patch)
+		__pause_3insn_patch_end = .;
+	}
 	PERCPU_SECTION(SMP_CACHE_BYTES)
 
 	. = ALIGN(PAGE_SIZE);
diff --git a/arch/sparc/kernel/winfixup.S b/arch/sparc/kernel/winfixup.S
index a6b0863c27df..1e67ce958369 100644
--- a/arch/sparc/kernel/winfixup.S
+++ b/arch/sparc/kernel/winfixup.S
@@ -43,6 +43,8 @@ spill_fixup_mna:
 spill_fixup_dax:
 	TRAP_LOAD_THREAD_REG(%g6, %g1)
 	ldx	[%g6 + TI_FLAGS], %g1
+	andcc	%sp, 0x1, %g0
+	movne	%icc, 0, %g1
 	andcc	%g1, _TIF_32BIT, %g0
 	ldub	[%g6 + TI_WSAVED], %g1
 	sll	%g1, 3, %g3
diff --git a/arch/sparc/lib/atomic_64.S b/arch/sparc/lib/atomic_64.S
index 4d502da3de78..85c233d0a340 100644
--- a/arch/sparc/lib/atomic_64.S
+++ b/arch/sparc/lib/atomic_64.S
@@ -1,6 +1,6 @@
 /* atomic.S: These things are too big to do inline.
  *
- * Copyright (C) 1999, 2007 David S. Miller (davem@davemloft.net)
+ * Copyright (C) 1999, 2007 2012 David S. Miller (davem@davemloft.net)
  */
 
 #include <linux/linkage.h>
@@ -117,3 +117,17 @@ ENTRY(atomic64_sub_ret) /* %o0 = decrement, %o1 = atomic_ptr */
 	 sub	%g1, %o0, %o0
 2:	BACKOFF_SPIN(%o2, %o3, 1b)
 ENDPROC(atomic64_sub_ret)
+
+ENTRY(atomic64_dec_if_positive) /* %o0 = atomic_ptr */
+	BACKOFF_SETUP(%o2)
+1:	ldx	[%o0], %g1
+	brlez,pn %g1, 3f
+	 sub	%g1, 1, %g7
+	casx	[%o0], %g1, %g7
+	cmp	%g1, %g7
+	bne,pn	%xcc, BACKOFF_LABEL(2f, 1b)
+	 nop
+3:	retl
+	 sub	%g1, 1, %o0
+2:	BACKOFF_SPIN(%o2, %o3, 1b)
+ENDPROC(atomic64_dec_if_positive)
diff --git a/arch/sparc/lib/ksyms.c b/arch/sparc/lib/ksyms.c
index ee31b884c61b..0c4e35e522fa 100644
--- a/arch/sparc/lib/ksyms.c
+++ b/arch/sparc/lib/ksyms.c
@@ -116,6 +116,7 @@ EXPORT_SYMBOL(atomic64_add);
 EXPORT_SYMBOL(atomic64_add_ret);
 EXPORT_SYMBOL(atomic64_sub);
 EXPORT_SYMBOL(atomic64_sub_ret);
+EXPORT_SYMBOL(atomic64_dec_if_positive);
 
 /* Atomic bit operations. */
 EXPORT_SYMBOL(test_and_set_bit);
diff --git a/arch/sparc/math-emu/math_64.c b/arch/sparc/math-emu/math_64.c
index 1704068da928..034aadbff036 100644
--- a/arch/sparc/math-emu/math_64.c
+++ b/arch/sparc/math-emu/math_64.c
@@ -320,7 +320,7 @@ int do_mathemu(struct pt_regs *regs, struct fpustate *f, bool illegal_insn_trap)
 					XR = 0;
 				else if (freg < 16)
 					XR = regs->u_regs[freg];
-				else if (test_thread_flag(TIF_32BIT)) {
+				else if (!test_thread_64bit_stack(regs->u_regs[UREG_FP])) {
 					struct reg_window32 __user *win32;
 					flushw_user ();
 					win32 = (struct reg_window32 __user *)((unsigned long)((u32)regs->u_regs[UREG_FP]));
diff --git a/arch/sparc/mm/ultra.S b/arch/sparc/mm/ultra.S
index 874162a11ceb..f8e13d421fcb 100644
--- a/arch/sparc/mm/ultra.S
+++ b/arch/sparc/mm/ultra.S
@@ -481,8 +481,8 @@ xcall_sync_tick:
 
 	.globl		xcall_fetch_glob_regs
 xcall_fetch_glob_regs:
-	sethi		%hi(global_reg_snapshot), %g1
-	or		%g1, %lo(global_reg_snapshot), %g1
+	sethi		%hi(global_cpu_snapshot), %g1
+	or		%g1, %lo(global_cpu_snapshot), %g1
 	__GET_CPUID(%g2)
 	sllx		%g2, 6, %g3
 	add		%g1, %g3, %g1
@@ -509,6 +509,66 @@ xcall_fetch_glob_regs:
 	stx		%g3, [%g1 + GR_SNAP_THREAD]
 	retry
 
+	.globl		xcall_fetch_glob_pmu
+xcall_fetch_glob_pmu:
+	sethi		%hi(global_cpu_snapshot), %g1
+	or		%g1, %lo(global_cpu_snapshot), %g1
+	__GET_CPUID(%g2)
+	sllx		%g2, 6, %g3
+	add		%g1, %g3, %g1
+	rd		%pic, %g7
+	stx		%g7, [%g1 + (4 * 8)]
+	rd		%pcr, %g7
+	stx		%g7, [%g1 + (0 * 8)]
+	retry
+
+	.globl		xcall_fetch_glob_pmu_n4
+xcall_fetch_glob_pmu_n4:
+	sethi		%hi(global_cpu_snapshot), %g1
+	or		%g1, %lo(global_cpu_snapshot), %g1
+	__GET_CPUID(%g2)
+	sllx		%g2, 6, %g3
+	add		%g1, %g3, %g1
+
+	ldxa		[%g0] ASI_PIC, %g7
+	stx		%g7, [%g1 + (4 * 8)]
+	mov		0x08, %g3
+	ldxa		[%g3] ASI_PIC, %g7
+	stx		%g7, [%g1 + (5 * 8)]
+	mov		0x10, %g3
+	ldxa		[%g3] ASI_PIC, %g7
+	stx		%g7, [%g1 + (6 * 8)]
+	mov		0x18, %g3
+	ldxa		[%g3] ASI_PIC, %g7
+	stx		%g7, [%g1 + (7 * 8)]
+
+	mov		%o0, %g2
+	mov		%o1, %g3
+	mov		%o5, %g7
+
+	mov		HV_FAST_VT_GET_PERFREG, %o5
+	mov		3, %o0
+	ta		HV_FAST_TRAP
+	stx		%o1, [%g1 + (3 * 8)]
+	mov		HV_FAST_VT_GET_PERFREG, %o5
+	mov		2, %o0
+	ta		HV_FAST_TRAP
+	stx		%o1, [%g1 + (2 * 8)]
+	mov		HV_FAST_VT_GET_PERFREG, %o5
+	mov		1, %o0
+	ta		HV_FAST_TRAP
+	stx		%o1, [%g1 + (1 * 8)]
+	mov		HV_FAST_VT_GET_PERFREG, %o5
+	mov		0, %o0
+	ta		HV_FAST_TRAP
+	stx		%o1, [%g1 + (0 * 8)]
+
+	mov		%g2, %o0
+	mov		%g3, %o1
+	mov		%g7, %o5
+
+	retry
+
 #ifdef DCACHE_ALIASING_POSSIBLE
 	.align		32
 	.globl		xcall_flush_dcache_page_cheetah