Merge branch 'locking-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull locking and atomic updates from Ingo Molnar:
 "Main changes in this cycle are:

   - Extend atomic primitives with coherent logic op primitives
     (atomic_{or,and,xor}()) and deprecate the old partial APIs
     (atomic_{set,clear}_mask())

     The old ops were incoherent with incompatible signatures across
     architectures and with incomplete support.  Now every architecture
     supports the primitives consistently (by Peter Zijlstra)

   - Generic support for 'relaxed atomics':

       - _acquire/release/relaxed() flavours of xchg(), cmpxchg() and {add,sub}_return()
       - atomic_read_acquire()
       - atomic_set_release()

     This came out of porting qwrlock code to arm64 (by Will Deacon)

   - Clean up the fragile static_key APIs that were causing repeat bugs,
     by introducing a new one:

       DEFINE_STATIC_KEY_TRUE(name);
       DEFINE_STATIC_KEY_FALSE(name);

     which define a key of different types with an initial true/false
     value.

     Then allow:

       static_branch_likely()
       static_branch_unlikely()

     to take a key of either type and emit the right instruction for the
     case.  To be able to know the 'type' of the static key we encode it
     in the jump entry (by Peter Zijlstra)

   - Static key self-tests (by Jason Baron)

   - qrwlock optimizations (by Waiman Long)

   - small futex enhancements (by Davidlohr Bueso)

   - ... and misc other changes"

* 'locking-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (63 commits)
  jump_label/x86: Work around asm build bug on older/backported GCCs
  locking, ARM, atomics: Define our SMP atomics in terms of _relaxed() operations
  locking, include/llist: Use linux/atomic.h instead of asm/cmpxchg.h
  locking/qrwlock: Make use of _{acquire|release|relaxed}() atomics
  locking/qrwlock: Implement queue_write_unlock() using smp_store_release()
  locking/lockref: Remove homebrew cmpxchg64_relaxed() macro definition
  locking, asm-generic: Add _{relaxed|acquire|release}() variants for 'atomic_long_t'
  locking, asm-generic: Rework atomic-long.h to avoid bulk code duplication
  locking/atomics: Add _{acquire|release|relaxed}() variants of some atomic operations
  locking, compiler.h: Cast away attributes in the WRITE_ONCE() magic
  locking/static_keys: Make verify_keys() static
  jump label, locking/static_keys: Update docs
  locking/static_keys: Provide a selftest
  jump_label: Provide a self-test
  s390/uaccess, locking/static_keys: employ static_branch_likely()
  x86, tsc, locking/static_keys: Employ static_branch_likely()
  locking/static_keys: Add selftest
  locking/static_keys: Add a new static_key interface
  locking/static_keys: Rework update logic
  locking/static_keys: Add static_key_{en,dis}able() helpers
  ...

3 files changed, 236 insertions, 142 deletions

diff --git a/arch/frv/include/asm/atomic.h b/arch/frv/include/asm/atomic.h
index 102190a61d65..0da689def4cc 100644
--- a/arch/frv/include/asm/atomic.h
+++ b/arch/frv/include/asm/atomic.h
@@ -15,7 +15,6 @@
 #define _ASM_ATOMIC_H
 
 #include <linux/types.h>
-#include <asm/spr-regs.h>
 #include <asm/cmpxchg.h>
 #include <asm/barrier.h>
 
@@ -23,6 +22,8 @@
 #error not SMP safe
 #endif
 
+#include <asm/atomic_defs.h>
+
 /*
  * Atomic operations that C can't guarantee us.  Useful for
  * resource counting etc..
@@ -34,56 +35,26 @@
 #define atomic_read(v)		ACCESS_ONCE((v)->counter)
 #define atomic_set(v, i)	(((v)->counter) = (i))
 
-#ifndef CONFIG_FRV_OUTOFLINE_ATOMIC_OPS
-static inline int atomic_add_return(int i, atomic_t *v)
+static inline int atomic_inc_return(atomic_t *v)
 {
-	unsigned long val;
+	return __atomic_add_return(1, &v->counter);
+}
 
-	asm("0:						\n"
-	    "	orcc		gr0,gr0,gr0,icc3	\n"	/* set ICC3.Z */
-	    "	ckeq		icc3,cc7		\n"
-	    "	ld.p		%M0,%1			\n"	/* LD.P/ORCR must be atomic */
-	    "	orcr		cc7,cc7,cc3		\n"	/* set CC3 to true */
-	    "	add%I2		%1,%2,%1		\n"
-	    "	cst.p		%1,%M0		,cc3,#1	\n"
-	    "	corcc		gr29,gr29,gr0	,cc3,#1	\n"	/* clear ICC3.Z if store happens */
-	    "	beq		icc3,#0,0b		\n"
-	    : "+U"(v->counter), "=&r"(val)
-	    : "NPr"(i)
-	    : "memory", "cc7", "cc3", "icc3"
-	    );
+static inline int atomic_dec_return(atomic_t *v)
+{
+	return __atomic_sub_return(1, &v->counter);
+}
 
-	return val;
+static inline int atomic_add_return(int i, atomic_t *v)
+{
+	return __atomic_add_return(i, &v->counter);
 }
 
 static inline int atomic_sub_return(int i, atomic_t *v)
 {
-	unsigned long val;
-
-	asm("0:						\n"
-	    "	orcc		gr0,gr0,gr0,icc3	\n"	/* set ICC3.Z */
-	    "	ckeq		icc3,cc7		\n"
-	    "	ld.p		%M0,%1			\n"	/* LD.P/ORCR must be atomic */
-	    "	orcr		cc7,cc7,cc3		\n"	/* set CC3 to true */
-	    "	sub%I2		%1,%2,%1		\n"
-	    "	cst.p		%1,%M0		,cc3,#1	\n"
-	    "	corcc		gr29,gr29,gr0	,cc3,#1	\n"	/* clear ICC3.Z if store happens */
-	    "	beq		icc3,#0,0b		\n"
-	    : "+U"(v->counter), "=&r"(val)
-	    : "NPr"(i)
-	    : "memory", "cc7", "cc3", "icc3"
-	    );
-
-	return val;
+	return __atomic_sub_return(i, &v->counter);
 }
 
-#else
-
-extern int atomic_add_return(int i, atomic_t *v);
-extern int atomic_sub_return(int i, atomic_t *v);
-
-#endif
-
 static inline int atomic_add_negative(int i, atomic_t *v)
 {
 	return atomic_add_return(i, v) < 0;
@@ -101,17 +72,14 @@ static inline void atomic_sub(int i, atomic_t *v)
 
 static inline void atomic_inc(atomic_t *v)
 {
-	atomic_add_return(1, v);
+	atomic_inc_return(v);
 }
 
 static inline void atomic_dec(atomic_t *v)
 {
-	atomic_sub_return(1, v);
+	atomic_dec_return(v);
 }
 
-#define atomic_dec_return(v)		atomic_sub_return(1, (v))
-#define atomic_inc_return(v)		atomic_add_return(1, (v))
-
 #define atomic_sub_and_test(i,v)	(atomic_sub_return((i), (v)) == 0)
 #define atomic_dec_and_test(v)		(atomic_sub_return(1, (v)) == 0)
 #define atomic_inc_and_test(v)		(atomic_add_return(1, (v)) == 0)
@@ -120,18 +88,19 @@ static inline void atomic_dec(atomic_t *v)
  * 64-bit atomic ops
  */
 typedef struct {
-	volatile long long counter;
+	long long counter;
 } atomic64_t;
 
 #define ATOMIC64_INIT(i)	{ (i) }
 
-static inline long long atomic64_read(atomic64_t *v)
+static inline long long atomic64_read(const atomic64_t *v)
 {
 	long long counter;
 
 	asm("ldd%I1 %M1,%0"
 	    : "=e"(counter)
 	    : "m"(v->counter));
+
 	return counter;
 }
 
@@ -142,10 +111,25 @@ static inline void atomic64_set(atomic64_t *v, long long i)
 		     : "e"(i));
 }
 
-extern long long atomic64_inc_return(atomic64_t *v);
-extern long long atomic64_dec_return(atomic64_t *v);
-extern long long atomic64_add_return(long long i, atomic64_t *v);
-extern long long atomic64_sub_return(long long i, atomic64_t *v);
+static inline long long atomic64_inc_return(atomic64_t *v)
+{
+	return __atomic64_add_return(1, &v->counter);
+}
+
+static inline long long atomic64_dec_return(atomic64_t *v)
+{
+	return __atomic64_sub_return(1, &v->counter);
+}
+
+static inline long long atomic64_add_return(long long i, atomic64_t *v)
+{
+	return __atomic64_add_return(i, &v->counter);
+}
+
+static inline long long atomic64_sub_return(long long i, atomic64_t *v)
+{
+	return __atomic64_sub_return(i, &v->counter);
+}
 
 static inline long long atomic64_add_negative(long long i, atomic64_t *v)
 {
@@ -176,6 +160,7 @@ static inline void atomic64_dec(atomic64_t *v)
 #define atomic64_dec_and_test(v)	(atomic64_dec_return((v)) == 0)
 #define atomic64_inc_and_test(v)	(atomic64_inc_return((v)) == 0)
 
+
 #define atomic_cmpxchg(v, old, new)	(cmpxchg(&(v)->counter, old, new))
 #define atomic_xchg(v, new)		(xchg(&(v)->counter, new))
 #define atomic64_cmpxchg(v, old, new)	(__cmpxchg_64(old, new, &(v)->counter))
@@ -196,5 +181,21 @@ static __inline__ int __atomic_add_unless(atomic_t *v, int a, int u)
 	return c;
 }
 
+#define ATOMIC_OP(op)							\
+static inline void atomic_##op(int i, atomic_t *v)			\
+{									\
+	(void)__atomic32_fetch_##op(i, &v->counter);			\
+}									\
+									\
+static inline void atomic64_##op(long long i, atomic64_t *v)		\
+{									\
+	(void)__atomic64_fetch_##op(i, &v->counter);			\
+}
+
+ATOMIC_OP(or)
+ATOMIC_OP(and)
+ATOMIC_OP(xor)
+
+#undef ATOMIC_OP
 
 #endif /* _ASM_ATOMIC_H */
diff --git a/arch/frv/include/asm/atomic_defs.h b/arch/frv/include/asm/atomic_defs.h
new file mode 100644
index 000000000000..36e126d2f801
--- /dev/null
+++ b/arch/frv/include/asm/atomic_defs.h
@@ -0,0 +1,172 @@
+
+#include <asm/spr-regs.h>
+
+#ifdef __ATOMIC_LIB__
+
+#ifdef CONFIG_FRV_OUTOFLINE_ATOMIC_OPS
+
+#define ATOMIC_QUALS
+#define ATOMIC_EXPORT(x)	EXPORT_SYMBOL(x)
+
+#else /* !OUTOFLINE && LIB */
+
+#define ATOMIC_OP_RETURN(op)
+#define ATOMIC_FETCH_OP(op)
+
+#endif /* OUTOFLINE */
+
+#else /* !__ATOMIC_LIB__ */
+
+#define ATOMIC_EXPORT(x)
+
+#ifdef CONFIG_FRV_OUTOFLINE_ATOMIC_OPS
+
+#define ATOMIC_OP_RETURN(op)						\
+extern int __atomic_##op##_return(int i, int *v);			\
+extern long long __atomic64_##op##_return(long long i, long long *v);
+
+#define ATOMIC_FETCH_OP(op)						\
+extern int __atomic32_fetch_##op(int i, int *v);			\
+extern long long __atomic64_fetch_##op(long long i, long long *v);
+
+#else /* !OUTOFLINE && !LIB */
+
+#define ATOMIC_QUALS	static inline
+
+#endif /* OUTOFLINE */
+#endif /* __ATOMIC_LIB__ */
+
+
+/*
+ * Note on the 64 bit inline asm variants...
+ *
+ * CSTD is a conditional instruction and needs a constrained memory reference.
+ * Normally 'U' provides the correct constraints for conditional instructions
+ * and this is used for the 32 bit version, however 'U' does not appear to work
+ * for 64 bit values (gcc-4.9)
+ *
+ * The exact constraint is that conditional instructions cannot deal with an
+ * immediate displacement in the memory reference, so what we do is we read the
+ * address through a volatile cast into a local variable in order to insure we
+ * _have_ to compute the correct address without displacement. This allows us
+ * to use the regular 'm' for the memory address.
+ *
+ * Furthermore, the %Ln operand, which prints the low word register (r+1),
+ * really only works for registers, this means we cannot allow immediate values
+ * for the 64 bit versions -- like we do for the 32 bit ones.
+ *
+ */
+
+#ifndef ATOMIC_OP_RETURN
+#define ATOMIC_OP_RETURN(op)						\
+ATOMIC_QUALS int __atomic_##op##_return(int i, int *v)			\
+{									\
+	int val;							\
+									\
+	asm volatile(							\
+	    "0:						\n"		\
+	    "	orcc		gr0,gr0,gr0,icc3	\n"		\
+	    "	ckeq		icc3,cc7		\n"		\
+	    "	ld.p		%M0,%1			\n"		\
+	    "	orcr		cc7,cc7,cc3		\n"		\
+	    "   "#op"%I2	%1,%2,%1		\n"		\
+	    "	cst.p		%1,%M0		,cc3,#1	\n"		\
+	    "	corcc		gr29,gr29,gr0	,cc3,#1	\n"		\
+	    "	beq		icc3,#0,0b		\n"		\
+	    : "+U"(*v), "=&r"(val)					\
+	    : "NPr"(i)							\
+	    : "memory", "cc7", "cc3", "icc3"				\
+	    );								\
+									\
+	return val;							\
+}									\
+ATOMIC_EXPORT(__atomic_##op##_return);					\
+									\
+ATOMIC_QUALS long long __atomic64_##op##_return(long long i, long long *v)	\
+{									\
+	long long *__v = READ_ONCE(v);					\
+	long long val;							\
+									\
+	asm volatile(							\
+	    "0:						\n"		\
+	    "	orcc		gr0,gr0,gr0,icc3	\n"		\
+	    "	ckeq		icc3,cc7		\n"		\
+	    "	ldd.p		%M0,%1			\n"		\
+	    "	orcr		cc7,cc7,cc3		\n"		\
+	    "   "#op"cc		%L1,%L2,%L1,icc0	\n"		\
+	    "   "#op"x		%1,%2,%1,icc0		\n"		\
+	    "	cstd.p		%1,%M0		,cc3,#1	\n"		\
+	    "	corcc		gr29,gr29,gr0	,cc3,#1	\n"		\
+	    "	beq		icc3,#0,0b		\n"		\
+	    : "+m"(*__v), "=&e"(val)					\
+	    : "e"(i)							\
+	    : "memory", "cc7", "cc3", "icc0", "icc3"			\
+	    );								\
+									\
+	return val;							\
+}									\
+ATOMIC_EXPORT(__atomic64_##op##_return);
+#endif
+
+#ifndef ATOMIC_FETCH_OP
+#define ATOMIC_FETCH_OP(op)						\
+ATOMIC_QUALS int __atomic32_fetch_##op(int i, int *v)			\
+{									\
+	int old, tmp;							\
+									\
+	asm volatile(							\
+		"0:						\n"	\
+		"	orcc		gr0,gr0,gr0,icc3	\n"	\
+		"	ckeq		icc3,cc7		\n"	\
+		"	ld.p		%M0,%1			\n"	\
+		"	orcr		cc7,cc7,cc3		\n"	\
+		"	"#op"%I3	%1,%3,%2		\n"	\
+		"	cst.p		%2,%M0		,cc3,#1	\n"	\
+		"	corcc		gr29,gr29,gr0	,cc3,#1	\n"	\
+		"	beq		icc3,#0,0b		\n"	\
+		: "+U"(*v), "=&r"(old), "=r"(tmp)			\
+		: "NPr"(i)						\
+		: "memory", "cc7", "cc3", "icc3"			\
+		);							\
+									\
+	return old;							\
+}									\
+ATOMIC_EXPORT(__atomic32_fetch_##op);					\
+									\
+ATOMIC_QUALS long long __atomic64_fetch_##op(long long i, long long *v)	\
+{									\
+	long long *__v = READ_ONCE(v);					\
+	long long old, tmp;						\
+									\
+	asm volatile(							\
+		"0:						\n"	\
+		"	orcc		gr0,gr0,gr0,icc3	\n"	\
+		"	ckeq		icc3,cc7		\n"	\
+		"	ldd.p		%M0,%1			\n"	\
+		"	orcr		cc7,cc7,cc3		\n"	\
+		"	"#op"		%L1,%L3,%L2		\n"	\
+		"	"#op"		%1,%3,%2		\n"	\
+		"	cstd.p		%2,%M0		,cc3,#1	\n"	\
+		"	corcc		gr29,gr29,gr0	,cc3,#1	\n"	\
+		"	beq		icc3,#0,0b		\n"	\
+		: "+m"(*__v), "=&e"(old), "=e"(tmp)			\
+		: "e"(i)						\
+		: "memory", "cc7", "cc3", "icc3"			\
+		);							\
+									\
+	return old;							\
+}									\
+ATOMIC_EXPORT(__atomic64_fetch_##op);
+#endif
+
+ATOMIC_FETCH_OP(or)
+ATOMIC_FETCH_OP(and)
+ATOMIC_FETCH_OP(xor)
+
+ATOMIC_OP_RETURN(add)
+ATOMIC_OP_RETURN(sub)
+
+#undef ATOMIC_FETCH_OP
+#undef ATOMIC_OP_RETURN
+#undef ATOMIC_QUALS
+#undef ATOMIC_EXPORT
diff --git a/arch/frv/include/asm/bitops.h b/arch/frv/include/asm/bitops.h
index 96de220ef131..0df8e95e3715 100644
--- a/arch/frv/include/asm/bitops.h
+++ b/arch/frv/include/asm/bitops.h
@@ -25,109 +25,30 @@
 
 #include <asm-generic/bitops/ffz.h>
 
-#ifndef CONFIG_FRV_OUTOFLINE_ATOMIC_OPS
-static inline
-unsigned long atomic_test_and_ANDNOT_mask(unsigned long mask, volatile unsigned long *v)
-{
-	unsigned long old, tmp;
-
-	asm volatile(
-		"0:						\n"
-		"	orcc		gr0,gr0,gr0,icc3	\n"	/* set ICC3.Z */
-		"	ckeq		icc3,cc7		\n"
-		"	ld.p		%M0,%1			\n"	/* LD.P/ORCR are atomic */
-		"	orcr		cc7,cc7,cc3		\n"	/* set CC3 to true */
-		"	and%I3		%1,%3,%2		\n"
-		"	cst.p		%2,%M0		,cc3,#1	\n"	/* if store happens... */
-		"	corcc		gr29,gr29,gr0	,cc3,#1	\n"	/* ... clear ICC3.Z */
-		"	beq		icc3,#0,0b		\n"
-		: "+U"(*v), "=&r"(old), "=r"(tmp)
-		: "NPr"(~mask)
-		: "memory", "cc7", "cc3", "icc3"
-		);
-
-	return old;
-}
-
-static inline
-unsigned long atomic_test_and_OR_mask(unsigned long mask, volatile unsigned long *v)
-{
-	unsigned long old, tmp;
-
-	asm volatile(
-		"0:						\n"
-		"	orcc		gr0,gr0,gr0,icc3	\n"	/* set ICC3.Z */
-		"	ckeq		icc3,cc7		\n"
-		"	ld.p		%M0,%1			\n"	/* LD.P/ORCR are atomic */
-		"	orcr		cc7,cc7,cc3		\n"	/* set CC3 to true */
-		"	or%I3		%1,%3,%2		\n"
-		"	cst.p		%2,%M0		,cc3,#1	\n"	/* if store happens... */
-		"	corcc		gr29,gr29,gr0	,cc3,#1	\n"	/* ... clear ICC3.Z */
-		"	beq		icc3,#0,0b		\n"
-		: "+U"(*v), "=&r"(old), "=r"(tmp)
-		: "NPr"(mask)
-		: "memory", "cc7", "cc3", "icc3"
-		);
-
-	return old;
-}
-
-static inline
-unsigned long atomic_test_and_XOR_mask(unsigned long mask, volatile unsigned long *v)
-{
-	unsigned long old, tmp;
-
-	asm volatile(
-		"0:						\n"
-		"	orcc		gr0,gr0,gr0,icc3	\n"	/* set ICC3.Z */
-		"	ckeq		icc3,cc7		\n"
-		"	ld.p		%M0,%1			\n"	/* LD.P/ORCR are atomic */
-		"	orcr		cc7,cc7,cc3		\n"	/* set CC3 to true */
-		"	xor%I3		%1,%3,%2		\n"
-		"	cst.p		%2,%M0		,cc3,#1	\n"	/* if store happens... */
-		"	corcc		gr29,gr29,gr0	,cc3,#1	\n"	/* ... clear ICC3.Z */
-		"	beq		icc3,#0,0b		\n"
-		: "+U"(*v), "=&r"(old), "=r"(tmp)
-		: "NPr"(mask)
-		: "memory", "cc7", "cc3", "icc3"
-		);
-
-	return old;
-}
-
-#else
-
-extern unsigned long atomic_test_and_ANDNOT_mask(unsigned long mask, volatile unsigned long *v);
-extern unsigned long atomic_test_and_OR_mask(unsigned long mask, volatile unsigned long *v);
-extern unsigned long atomic_test_and_XOR_mask(unsigned long mask, volatile unsigned long *v);
-
-#endif
-
-#define atomic_clear_mask(mask, v)	atomic_test_and_ANDNOT_mask((mask), (v))
-#define atomic_set_mask(mask, v)	atomic_test_and_OR_mask((mask), (v))
+#include <asm/atomic.h>
 
 static inline int test_and_clear_bit(unsigned long nr, volatile void *addr)
 {
-	volatile unsigned long *ptr = addr;
-	unsigned long mask = 1UL << (nr & 31);
+	unsigned int *ptr = (void *)addr;
+	unsigned int mask = 1UL << (nr & 31);
 	ptr += nr >> 5;
-	return (atomic_test_and_ANDNOT_mask(mask, ptr) & mask) != 0;
+	return (__atomic32_fetch_and(~mask, ptr) & mask) != 0;
 }
 
 static inline int test_and_set_bit(unsigned long nr, volatile void *addr)
 {
-	volatile unsigned long *ptr = addr;
-	unsigned long mask = 1UL << (nr & 31);
+	unsigned int *ptr = (void *)addr;
+	unsigned int mask = 1UL << (nr & 31);
 	ptr += nr >> 5;
-	return (atomic_test_and_OR_mask(mask, ptr) & mask) != 0;
+	return (__atomic32_fetch_or(mask, ptr) & mask) != 0;
 }
 
 static inline int test_and_change_bit(unsigned long nr, volatile void *addr)
 {
-	volatile unsigned long *ptr = addr;
-	unsigned long mask = 1UL << (nr & 31);
+	unsigned int *ptr = (void *)addr;
+	unsigned int mask = 1UL << (nr & 31);
 	ptr += nr >> 5;
-	return (atomic_test_and_XOR_mask(mask, ptr) & mask) != 0;
+	return (__atomic32_fetch_xor(mask, ptr) & mask) != 0;
 }
 
 static inline void clear_bit(unsigned long nr, volatile void *addr)