Merge branch 'for-3.17' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/percpu

Pull percpu updates from Tejun Heo:

 - Major reorganization of percpu header files which I think makes
   things a lot more readable and logical than before.

 - percpu-refcount is updated so that it requires explicit destruction
   and can be reinitialized if necessary.  This was pulled into the
   block tree to replace the custom percpu refcnting implemented in
   blk-mq.

 - In the process, percpu and percpu-refcount got cleaned up a bit

* 'for-3.17' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/percpu: (21 commits)
  percpu-refcount: implement percpu_ref_reinit() and percpu_ref_is_zero()
  percpu-refcount: require percpu_ref to be exited explicitly
  percpu-refcount: use unsigned long for pcpu_count pointer
  percpu-refcount: add helpers for ->percpu_count accesses
  percpu-refcount: one bit is enough for REF_STATUS
  percpu-refcount, aio: use percpu_ref_cancel_init() in ioctx_alloc()
  workqueue: stronger test in process_one_work()
  workqueue: clear POOL_DISASSOCIATED in rebind_workers()
  percpu: Use ALIGN macro instead of hand coding alignment calculation
  percpu: invoke __verify_pcpu_ptr() from the generic part of accessors and operations
  percpu: preffity percpu header files
  percpu: use raw_cpu_*() to define __this_cpu_*()
  percpu: reorder macros in percpu header files
  percpu: move {raw|this}_cpu_*() definitions to include/linux/percpu-defs.h
  percpu: move generic {raw|this}_cpu_*_N() definitions to include/asm-generic/percpu.h
  percpu: only allow sized arch overrides for {raw|this}_cpu_*() ops
  percpu: reorganize include/linux/percpu-defs.h
  percpu: move accessors from include/linux/percpu.h to percpu-defs.h
  percpu: include/asm-generic/percpu.h should contain only arch-overridable parts
  percpu: introduce arch_raw_cpu_ptr()
  ...

11 files changed, 832 insertions, 811 deletions

diff --git a/arch/x86/include/asm/percpu.h b/arch/x86/include/asm/percpu.h
index 851bcdc5db04..fd472181a1d0 100644
--- a/arch/x86/include/asm/percpu.h
+++ b/arch/x86/include/asm/percpu.h
@@ -52,10 +52,9 @@
  * Compared to the generic __my_cpu_offset version, the following
  * saves one instruction and avoids clobbering a temp register.
  */
-#define raw_cpu_ptr(ptr)				\
+#define arch_raw_cpu_ptr(ptr)				\
 ({							\
 	unsigned long tcp_ptr__;			\
-	__verify_pcpu_ptr(ptr);				\
 	asm volatile("add " __percpu_arg(1) ", %0"	\
 		     : "=r" (tcp_ptr__)			\
 		     : "m" (this_cpu_off), "0" (ptr));	\
diff --git a/drivers/target/target_core_tpg.c b/drivers/target/target_core_tpg.c
index c036595b17cf..fddfae61222f 100644
--- a/drivers/target/target_core_tpg.c
+++ b/drivers/target/target_core_tpg.c
@@ -825,7 +825,7 @@ int core_tpg_add_lun(
 
 	ret = core_dev_export(dev, tpg, lun);
 	if (ret < 0) {
-		percpu_ref_cancel_init(&lun->lun_ref);
+		percpu_ref_exit(&lun->lun_ref);
 		return ret;
 	}
 
@@ -880,5 +880,7 @@ int core_tpg_post_dellun(
 	lun->lun_status = TRANSPORT_LUN_STATUS_FREE;
 	spin_unlock(&tpg->tpg_lun_lock);
 
+	percpu_ref_exit(&lun->lun_ref);
+
 	return 0;
 }
diff --git a/fs/aio.c b/fs/aio.c
index 1c9c5f0a9e2b..bd7ec2cc2674 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -506,6 +506,8 @@ static void free_ioctx(struct work_struct *work)
 
 	aio_free_ring(ctx);
 	free_percpu(ctx->cpu);
+	percpu_ref_exit(&ctx->reqs);
+	percpu_ref_exit(&ctx->users);
 	kmem_cache_free(kioctx_cachep, ctx);
 }
 
@@ -715,8 +717,8 @@ err_ctx:
 err:
 	mutex_unlock(&ctx->ring_lock);
 	free_percpu(ctx->cpu);
-	free_percpu(ctx->reqs.pcpu_count);
-	free_percpu(ctx->users.pcpu_count);
+	percpu_ref_exit(&ctx->reqs);
+	percpu_ref_exit(&ctx->users);
 	kmem_cache_free(kioctx_cachep, ctx);
 	pr_debug("error allocating ioctx %d\n", err);
 	return ERR_PTR(err);
diff --git a/include/asm-generic/percpu.h b/include/asm-generic/percpu.h
index 0703aa75b5e8..4d9f233c4ba8 100644
--- a/include/asm-generic/percpu.h
+++ b/include/asm-generic/percpu.h
@@ -36,93 +36,385 @@ extern unsigned long __per_cpu_offset[NR_CPUS];
 #endif
 
 /*
- * Add a offset to a pointer but keep the pointer as is.
- *
- * Only S390 provides its own means of moving the pointer.
+ * Arch may define arch_raw_cpu_ptr() to provide more efficient address
+ * translations for raw_cpu_ptr().
  */
-#ifndef SHIFT_PERCPU_PTR
-/* Weird cast keeps both GCC and sparse happy. */
-#define SHIFT_PERCPU_PTR(__p, __offset)	({				\
-	__verify_pcpu_ptr((__p));					\
-	RELOC_HIDE((typeof(*(__p)) __kernel __force *)(__p), (__offset)); \
-})
+#ifndef arch_raw_cpu_ptr
+#define arch_raw_cpu_ptr(ptr) SHIFT_PERCPU_PTR(ptr, __my_cpu_offset)
 #endif
 
-/*
- * A percpu variable may point to a discarded regions. The following are
- * established ways to produce a usable pointer from the percpu variable
- * offset.
- */
-#define per_cpu(var, cpu) \
-	(*SHIFT_PERCPU_PTR(&(var), per_cpu_offset(cpu)))
-
-#ifndef raw_cpu_ptr
-#define raw_cpu_ptr(ptr) SHIFT_PERCPU_PTR(ptr, __my_cpu_offset)
+#ifdef CONFIG_HAVE_SETUP_PER_CPU_AREA
+extern void setup_per_cpu_areas(void);
 #endif
-#ifdef CONFIG_DEBUG_PREEMPT
-#define this_cpu_ptr(ptr) SHIFT_PERCPU_PTR(ptr, my_cpu_offset)
+
+#endif	/* SMP */
+
+#ifndef PER_CPU_BASE_SECTION
+#ifdef CONFIG_SMP
+#define PER_CPU_BASE_SECTION ".data..percpu"
 #else
-#define this_cpu_ptr(ptr) raw_cpu_ptr(ptr)
+#define PER_CPU_BASE_SECTION ".data"
+#endif
 #endif
 
-#define __get_cpu_var(var) (*this_cpu_ptr(&(var)))
-#define __raw_get_cpu_var(var) (*raw_cpu_ptr(&(var)))
+#ifndef PER_CPU_ATTRIBUTES
+#define PER_CPU_ATTRIBUTES
+#endif
 
-#ifdef CONFIG_HAVE_SETUP_PER_CPU_AREA
-extern void setup_per_cpu_areas(void);
+#ifndef PER_CPU_DEF_ATTRIBUTES
+#define PER_CPU_DEF_ATTRIBUTES
 #endif
 
-#else /* ! SMP */
+#define raw_cpu_generic_to_op(pcp, val, op)				\
+do {									\
+	*raw_cpu_ptr(&(pcp)) op val;					\
+} while (0)
 
-#define VERIFY_PERCPU_PTR(__p) ({			\
-	__verify_pcpu_ptr((__p));			\
-	(typeof(*(__p)) __kernel __force *)(__p);	\
+#define raw_cpu_generic_add_return(pcp, val)				\
+({									\
+	raw_cpu_add(pcp, val);						\
+	raw_cpu_read(pcp);						\
 })
 
-#define per_cpu(var, cpu)	(*((void)(cpu), VERIFY_PERCPU_PTR(&(var))))
-#define __get_cpu_var(var)	(*VERIFY_PERCPU_PTR(&(var)))
-#define __raw_get_cpu_var(var)	(*VERIFY_PERCPU_PTR(&(var)))
-#define this_cpu_ptr(ptr)	per_cpu_ptr(ptr, 0)
-#define raw_cpu_ptr(ptr)	this_cpu_ptr(ptr)
+#define raw_cpu_generic_xchg(pcp, nval)					\
+({									\
+	typeof(pcp) __ret;						\
+	__ret = raw_cpu_read(pcp);					\
+	raw_cpu_write(pcp, nval);					\
+	__ret;								\
+})
 
-#endif	/* SMP */
+#define raw_cpu_generic_cmpxchg(pcp, oval, nval)			\
+({									\
+	typeof(pcp) __ret;						\
+	__ret = raw_cpu_read(pcp);					\
+	if (__ret == (oval))						\
+		raw_cpu_write(pcp, nval);				\
+	__ret;								\
+})
 
-#ifndef PER_CPU_BASE_SECTION
-#ifdef CONFIG_SMP
-#define PER_CPU_BASE_SECTION ".data..percpu"
-#else
-#define PER_CPU_BASE_SECTION ".data"
+#define raw_cpu_generic_cmpxchg_double(pcp1, pcp2, oval1, oval2, nval1, nval2) \
+({									\
+	int __ret = 0;							\
+	if (raw_cpu_read(pcp1) == (oval1) &&				\
+			 raw_cpu_read(pcp2)  == (oval2)) {		\
+		raw_cpu_write(pcp1, nval1);				\
+		raw_cpu_write(pcp2, nval2);				\
+		__ret = 1;						\
+	}								\
+	(__ret);							\
+})
+
+#define this_cpu_generic_read(pcp)					\
+({									\
+	typeof(pcp) __ret;						\
+	preempt_disable();						\
+	__ret = *this_cpu_ptr(&(pcp));					\
+	preempt_enable();						\
+	__ret;								\
+})
+
+#define this_cpu_generic_to_op(pcp, val, op)				\
+do {									\
+	unsigned long __flags;						\
+	raw_local_irq_save(__flags);					\
+	*raw_cpu_ptr(&(pcp)) op val;					\
+	raw_local_irq_restore(__flags);					\
+} while (0)
+
+#define this_cpu_generic_add_return(pcp, val)				\
+({									\
+	typeof(pcp) __ret;						\
+	unsigned long __flags;						\
+	raw_local_irq_save(__flags);					\
+	raw_cpu_add(pcp, val);						\
+	__ret = raw_cpu_read(pcp);					\
+	raw_local_irq_restore(__flags);					\
+	__ret;								\
+})
+
+#define this_cpu_generic_xchg(pcp, nval)				\
+({									\
+	typeof(pcp) __ret;						\
+	unsigned long __flags;						\
+	raw_local_irq_save(__flags);					\
+	__ret = raw_cpu_read(pcp);					\
+	raw_cpu_write(pcp, nval);					\
+	raw_local_irq_restore(__flags);					\
+	__ret;								\
+})
+
+#define this_cpu_generic_cmpxchg(pcp, oval, nval)			\
+({									\
+	typeof(pcp) __ret;						\
+	unsigned long __flags;						\
+	raw_local_irq_save(__flags);					\
+	__ret = raw_cpu_read(pcp);					\
+	if (__ret == (oval))						\
+		raw_cpu_write(pcp, nval);				\
+	raw_local_irq_restore(__flags);					\
+	__ret;								\
+})
+
+#define this_cpu_generic_cmpxchg_double(pcp1, pcp2, oval1, oval2, nval1, nval2)	\
+({									\
+	int __ret;							\
+	unsigned long __flags;						\
+	raw_local_irq_save(__flags);					\
+	__ret = raw_cpu_generic_cmpxchg_double(pcp1, pcp2,		\
+			oval1, oval2, nval1, nval2);			\
+	raw_local_irq_restore(__flags);					\
+	__ret;								\
+})
+
+#ifndef raw_cpu_read_1
+#define raw_cpu_read_1(pcp)		(*raw_cpu_ptr(&(pcp)))
 #endif
+#ifndef raw_cpu_read_2
+#define raw_cpu_read_2(pcp)		(*raw_cpu_ptr(&(pcp)))
+#endif
+#ifndef raw_cpu_read_4
+#define raw_cpu_read_4(pcp)		(*raw_cpu_ptr(&(pcp)))
+#endif
+#ifndef raw_cpu_read_8
+#define raw_cpu_read_8(pcp)		(*raw_cpu_ptr(&(pcp)))
 #endif
 
-#ifdef CONFIG_SMP
+#ifndef raw_cpu_write_1
+#define raw_cpu_write_1(pcp, val)	raw_cpu_generic_to_op(pcp, val, =)
+#endif
+#ifndef raw_cpu_write_2
+#define raw_cpu_write_2(pcp, val)	raw_cpu_generic_to_op(pcp, val, =)
+#endif
+#ifndef raw_cpu_write_4
+#define raw_cpu_write_4(pcp, val)	raw_cpu_generic_to_op(pcp, val, =)
+#endif
+#ifndef raw_cpu_write_8
+#define raw_cpu_write_8(pcp, val)	raw_cpu_generic_to_op(pcp, val, =)
+#endif
 
-#ifdef MODULE
-#define PER_CPU_SHARED_ALIGNED_SECTION ""
-#define PER_CPU_ALIGNED_SECTION ""
-#else
-#define PER_CPU_SHARED_ALIGNED_SECTION "..shared_aligned"
-#define PER_CPU_ALIGNED_SECTION "..shared_aligned"
+#ifndef raw_cpu_add_1
+#define raw_cpu_add_1(pcp, val)		raw_cpu_generic_to_op(pcp, val, +=)
+#endif
+#ifndef raw_cpu_add_2
+#define raw_cpu_add_2(pcp, val)		raw_cpu_generic_to_op(pcp, val, +=)
+#endif
+#ifndef raw_cpu_add_4
+#define raw_cpu_add_4(pcp, val)		raw_cpu_generic_to_op(pcp, val, +=)
+#endif
+#ifndef raw_cpu_add_8
+#define raw_cpu_add_8(pcp, val)		raw_cpu_generic_to_op(pcp, val, +=)
 #endif
-#define PER_CPU_FIRST_SECTION "..first"
 
-#else
+#ifndef raw_cpu_and_1
+#define raw_cpu_and_1(pcp, val)		raw_cpu_generic_to_op(pcp, val, &=)
+#endif
+#ifndef raw_cpu_and_2
+#define raw_cpu_and_2(pcp, val)		raw_cpu_generic_to_op(pcp, val, &=)
+#endif
+#ifndef raw_cpu_and_4
+#define raw_cpu_and_4(pcp, val)		raw_cpu_generic_to_op(pcp, val, &=)
+#endif
+#ifndef raw_cpu_and_8
+#define raw_cpu_and_8(pcp, val)		raw_cpu_generic_to_op(pcp, val, &=)
+#endif
+
+#ifndef raw_cpu_or_1
+#define raw_cpu_or_1(pcp, val)		raw_cpu_generic_to_op(pcp, val, |=)
+#endif
+#ifndef raw_cpu_or_2
+#define raw_cpu_or_2(pcp, val)		raw_cpu_generic_to_op(pcp, val, |=)
+#endif
+#ifndef raw_cpu_or_4
+#define raw_cpu_or_4(pcp, val)		raw_cpu_generic_to_op(pcp, val, |=)
+#endif
+#ifndef raw_cpu_or_8
+#define raw_cpu_or_8(pcp, val)		raw_cpu_generic_to_op(pcp, val, |=)
+#endif
 
-#define PER_CPU_SHARED_ALIGNED_SECTION ""
-#define PER_CPU_ALIGNED_SECTION "..shared_aligned"
-#define PER_CPU_FIRST_SECTION ""
+#ifndef raw_cpu_add_return_1
+#define raw_cpu_add_return_1(pcp, val)	raw_cpu_generic_add_return(pcp, val)
+#endif
+#ifndef raw_cpu_add_return_2
+#define raw_cpu_add_return_2(pcp, val)	raw_cpu_generic_add_return(pcp, val)
+#endif
+#ifndef raw_cpu_add_return_4
+#define raw_cpu_add_return_4(pcp, val)	raw_cpu_generic_add_return(pcp, val)
+#endif
+#ifndef raw_cpu_add_return_8
+#define raw_cpu_add_return_8(pcp, val)	raw_cpu_generic_add_return(pcp, val)
+#endif
 
+#ifndef raw_cpu_xchg_1
+#define raw_cpu_xchg_1(pcp, nval)	raw_cpu_generic_xchg(pcp, nval)
+#endif
+#ifndef raw_cpu_xchg_2
+#define raw_cpu_xchg_2(pcp, nval)	raw_cpu_generic_xchg(pcp, nval)
+#endif
+#ifndef raw_cpu_xchg_4
+#define raw_cpu_xchg_4(pcp, nval)	raw_cpu_generic_xchg(pcp, nval)
+#endif
+#ifndef raw_cpu_xchg_8
+#define raw_cpu_xchg_8(pcp, nval)	raw_cpu_generic_xchg(pcp, nval)
 #endif
 
-#ifndef PER_CPU_ATTRIBUTES
-#define PER_CPU_ATTRIBUTES
+#ifndef raw_cpu_cmpxchg_1
+#define raw_cpu_cmpxchg_1(pcp, oval, nval) \
+	raw_cpu_generic_cmpxchg(pcp, oval, nval)
+#endif
+#ifndef raw_cpu_cmpxchg_2
+#define raw_cpu_cmpxchg_2(pcp, oval, nval) \
+	raw_cpu_generic_cmpxchg(pcp, oval, nval)
+#endif
+#ifndef raw_cpu_cmpxchg_4
+#define raw_cpu_cmpxchg_4(pcp, oval, nval) \
+	raw_cpu_generic_cmpxchg(pcp, oval, nval)
+#endif
+#ifndef raw_cpu_cmpxchg_8
+#define raw_cpu_cmpxchg_8(pcp, oval, nval) \
+	raw_cpu_generic_cmpxchg(pcp, oval, nval)
 #endif
 
-#ifndef PER_CPU_DEF_ATTRIBUTES
-#define PER_CPU_DEF_ATTRIBUTES
+#ifndef raw_cpu_cmpxchg_double_1
+#define raw_cpu_cmpxchg_double_1(pcp1, pcp2, oval1, oval2, nval1, nval2) \
+	raw_cpu_generic_cmpxchg_double(pcp1, pcp2, oval1, oval2, nval1, nval2)
+#endif
+#ifndef raw_cpu_cmpxchg_double_2
+#define raw_cpu_cmpxchg_double_2(pcp1, pcp2, oval1, oval2, nval1, nval2) \
+	raw_cpu_generic_cmpxchg_double(pcp1, pcp2, oval1, oval2, nval1, nval2)
+#endif
+#ifndef raw_cpu_cmpxchg_double_4
+#define raw_cpu_cmpxchg_double_4(pcp1, pcp2, oval1, oval2, nval1, nval2) \
+	raw_cpu_generic_cmpxchg_double(pcp1, pcp2, oval1, oval2, nval1, nval2)
+#endif
+#ifndef raw_cpu_cmpxchg_double_8
+#define raw_cpu_cmpxchg_double_8(pcp1, pcp2, oval1, oval2, nval1, nval2) \
+	raw_cpu_generic_cmpxchg_double(pcp1, pcp2, oval1, oval2, nval1, nval2)
+#endif
+
+#ifndef this_cpu_read_1
+#define this_cpu_read_1(pcp)		this_cpu_generic_read(pcp)
+#endif
+#ifndef this_cpu_read_2
+#define this_cpu_read_2(pcp)		this_cpu_generic_read(pcp)
+#endif
+#ifndef this_cpu_read_4
+#define this_cpu_read_4(pcp)		this_cpu_generic_read(pcp)
+#endif
+#ifndef this_cpu_read_8
+#define this_cpu_read_8(pcp)		this_cpu_generic_read(pcp)
 #endif
 
-/* Keep until we have removed all uses of __this_cpu_ptr */
-#define __this_cpu_ptr raw_cpu_ptr
+#ifndef this_cpu_write_1
+#define this_cpu_write_1(pcp, val)	this_cpu_generic_to_op(pcp, val, =)
+#endif
+#ifndef this_cpu_write_2
+#define this_cpu_write_2(pcp, val)	this_cpu_generic_to_op(pcp, val, =)
+#endif
+#ifndef this_cpu_write_4
+#define this_cpu_write_4(pcp, val)	this_cpu_generic_to_op(pcp, val, =)
+#endif
+#ifndef this_cpu_write_8
+#define this_cpu_write_8(pcp, val)	this_cpu_generic_to_op(pcp, val, =)
+#endif
+
+#ifndef this_cpu_add_1
+#define this_cpu_add_1(pcp, val)	this_cpu_generic_to_op(pcp, val, +=)
+#endif
+#ifndef this_cpu_add_2
+#define this_cpu_add_2(pcp, val)	this_cpu_generic_to_op(pcp, val, +=)
+#endif
+#ifndef this_cpu_add_4
+#define this_cpu_add_4(pcp, val)	this_cpu_generic_to_op(pcp, val, +=)
+#endif
+#ifndef this_cpu_add_8
+#define this_cpu_add_8(pcp, val)	this_cpu_generic_to_op(pcp, val, +=)
+#endif
+
+#ifndef this_cpu_and_1
+#define this_cpu_and_1(pcp, val)	this_cpu_generic_to_op(pcp, val, &=)
+#endif
+#ifndef this_cpu_and_2
+#define this_cpu_and_2(pcp, val)	this_cpu_generic_to_op(pcp, val, &=)
+#endif
+#ifndef this_cpu_and_4
+#define this_cpu_and_4(pcp, val)	this_cpu_generic_to_op(pcp, val, &=)
+#endif
+#ifndef this_cpu_and_8
+#define this_cpu_and_8(pcp, val)	this_cpu_generic_to_op(pcp, val, &=)
+#endif
+
+#ifndef this_cpu_or_1
+#define this_cpu_or_1(pcp, val)		this_cpu_generic_to_op(pcp, val, |=)
+#endif
+#ifndef this_cpu_or_2
+#define this_cpu_or_2(pcp, val)		this_cpu_generic_to_op(pcp, val, |=)
+#endif
+#ifndef this_cpu_or_4
+#define this_cpu_or_4(pcp, val)		this_cpu_generic_to_op(pcp, val, |=)
+#endif
+#ifndef this_cpu_or_8
+#define this_cpu_or_8(pcp, val)		this_cpu_generic_to_op(pcp, val, |=)
+#endif
+
+#ifndef this_cpu_add_return_1
+#define this_cpu_add_return_1(pcp, val)	this_cpu_generic_add_return(pcp, val)
+#endif
+#ifndef this_cpu_add_return_2
+#define this_cpu_add_return_2(pcp, val)	this_cpu_generic_add_return(pcp, val)
+#endif
+#ifndef this_cpu_add_return_4
+#define this_cpu_add_return_4(pcp, val)	this_cpu_generic_add_return(pcp, val)
+#endif
+#ifndef this_cpu_add_return_8
+#define this_cpu_add_return_8(pcp, val)	this_cpu_generic_add_return(pcp, val)
+#endif
+
+#ifndef this_cpu_xchg_1
+#define this_cpu_xchg_1(pcp, nval)	this_cpu_generic_xchg(pcp, nval)
+#endif
+#ifndef this_cpu_xchg_2
+#define this_cpu_xchg_2(pcp, nval)	this_cpu_generic_xchg(pcp, nval)
+#endif
+#ifndef this_cpu_xchg_4
+#define this_cpu_xchg_4(pcp, nval)	this_cpu_generic_xchg(pcp, nval)
+#endif
+#ifndef this_cpu_xchg_8
+#define this_cpu_xchg_8(pcp, nval)	this_cpu_generic_xchg(pcp, nval)
+#endif
+
+#ifndef this_cpu_cmpxchg_1
+#define this_cpu_cmpxchg_1(pcp, oval, nval) \
+	this_cpu_generic_cmpxchg(pcp, oval, nval)
+#endif
+#ifndef this_cpu_cmpxchg_2
+#define this_cpu_cmpxchg_2(pcp, oval, nval) \
+	this_cpu_generic_cmpxchg(pcp, oval, nval)
+#endif
+#ifndef this_cpu_cmpxchg_4
+#define this_cpu_cmpxchg_4(pcp, oval, nval) \
+	this_cpu_generic_cmpxchg(pcp, oval, nval)
+#endif
+#ifndef this_cpu_cmpxchg_8
+#define this_cpu_cmpxchg_8(pcp, oval, nval) \
+	this_cpu_generic_cmpxchg(pcp, oval, nval)
+#endif
+
+#ifndef this_cpu_cmpxchg_double_1
+#define this_cpu_cmpxchg_double_1(pcp1, pcp2, oval1, oval2, nval1, nval2) \
+	this_cpu_generic_cmpxchg_double(pcp1, pcp2, oval1, oval2, nval1, nval2)
+#endif
+#ifndef this_cpu_cmpxchg_double_2
+#define this_cpu_cmpxchg_double_2(pcp1, pcp2, oval1, oval2, nval1, nval2) \
+	this_cpu_generic_cmpxchg_double(pcp1, pcp2, oval1, oval2, nval1, nval2)
+#endif
+#ifndef this_cpu_cmpxchg_double_4
+#define this_cpu_cmpxchg_double_4(pcp1, pcp2, oval1, oval2, nval1, nval2) \
+	this_cpu_generic_cmpxchg_double(pcp1, pcp2, oval1, oval2, nval1, nval2)
+#endif
+#ifndef this_cpu_cmpxchg_double_8
+#define this_cpu_cmpxchg_double_8(pcp1, pcp2, oval1, oval2, nval1, nval2) \
+	this_cpu_generic_cmpxchg_double(pcp1, pcp2, oval1, oval2, nval1, nval2)
+#endif
 
 #endif /* _ASM_GENERIC_PERCPU_H_ */
diff --git a/include/linux/percpu-defs.h b/include/linux/percpu-defs.h
index dec01d6c3f80..cfd56046ecec 100644
--- a/include/linux/percpu-defs.h
+++ b/include/linux/percpu-defs.h
@@ -1,6 +1,40 @@
+/*
+ * linux/percpu-defs.h - basic definitions for percpu areas
+ *
+ * DO NOT INCLUDE DIRECTLY OUTSIDE PERCPU IMPLEMENTATION PROPER.
+ *
+ * This file is separate from linux/percpu.h to avoid cyclic inclusion
+ * dependency from arch header files.  Only to be included from
+ * asm/percpu.h.
+ *
+ * This file includes macros necessary to declare percpu sections and
+ * variables, and definitions of percpu accessors and operations.  It
+ * should provide enough percpu features to arch header files even when
+ * they can only include asm/percpu.h to avoid cyclic inclusion dependency.
+ */
+
 #ifndef _LINUX_PERCPU_DEFS_H
 #define _LINUX_PERCPU_DEFS_H
 
+#ifdef CONFIG_SMP
+
+#ifdef MODULE
+#define PER_CPU_SHARED_ALIGNED_SECTION ""
+#define PER_CPU_ALIGNED_SECTION ""
+#else
+#define PER_CPU_SHARED_ALIGNED_SECTION "..shared_aligned"
+#define PER_CPU_ALIGNED_SECTION "..shared_aligned"
+#endif
+#define PER_CPU_FIRST_SECTION "..first"
+
+#else
+
+#define PER_CPU_SHARED_ALIGNED_SECTION ""
+#define PER_CPU_ALIGNED_SECTION "..shared_aligned"
+#define PER_CPU_FIRST_SECTION ""
+
+#endif
+
 /*
  * Base implementations of per-CPU variable declarations and definitions, where
  * the section in which the variable is to be placed is provided by the
@@ -19,19 +53,6 @@
 	__attribute__((section(".discard"), unused))
 
 /*
- * Macro which verifies @ptr is a percpu pointer without evaluating
- * @ptr.  This is to be used in percpu accessors to verify that the
- * input parameter is a percpu pointer.
- *
- * + 0 is required in order to convert the pointer type from a
- * potential array type to a pointer to a single item of the array.
- */
-#define __verify_pcpu_ptr(ptr)	do {					\
-	const void __percpu *__vpp_verify = (typeof((ptr) + 0))NULL;	\
-	(void)__vpp_verify;						\
-} while (0)
-
-/*
  * s390 and alpha modules require percpu variables to be defined as
  * weak to force the compiler to generate GOT based external
  * references for them.  This is necessary because percpu sections
@@ -164,4 +185,337 @@
 #define EXPORT_PER_CPU_SYMBOL_GPL(var)
 #endif
 
+/*
+ * Accessors and operations.
+ */
+#ifndef __ASSEMBLY__
+
+/*
+ * __verify_pcpu_ptr() verifies @ptr is a percpu pointer without evaluating
+ * @ptr and is invoked once before a percpu area is accessed by all
+ * accessors and operations.  This is performed in the generic part of
+ * percpu and arch overrides don't need to worry about it; however, if an
+ * arch wants to implement an arch-specific percpu accessor or operation,
+ * it may use __verify_pcpu_ptr() to verify the parameters.
+ *
+ * + 0 is required in order to convert the pointer type from a
+ * potential array type to a pointer to a single item of the array.
+ */
+#define __verify_pcpu_ptr(ptr)						\
+do {									\
+	const void __percpu *__vpp_verify = (typeof((ptr) + 0))NULL;	\
+	(void)__vpp_verify;						\
+} while (0)
+
+#ifdef CONFIG_SMP
+
+/*
+ * Add an offset to a pointer but keep the pointer as-is.  Use RELOC_HIDE()
+ * to prevent the compiler from making incorrect assumptions about the
+ * pointer value.  The weird cast keeps both GCC and sparse happy.
+ */
+#define SHIFT_PERCPU_PTR(__p, __offset)					\
+	RELOC_HIDE((typeof(*(__p)) __kernel __force *)(__p), (__offset))
+
+#define per_cpu_ptr(ptr, cpu)						\
+({									\
+	__verify_pcpu_ptr(ptr);						\
+	SHIFT_PERCPU_PTR((ptr), per_cpu_offset((cpu)));			\
+})
+
+#define raw_cpu_ptr(ptr)						\
+({									\
+	__verify_pcpu_ptr(ptr);						\
+	arch_raw_cpu_ptr(ptr);						\
+})
+
+#ifdef CONFIG_DEBUG_PREEMPT
+#define this_cpu_ptr(ptr)						\
+({									\
+	__verify_pcpu_ptr(ptr);						\
+	SHIFT_PERCPU_PTR(ptr, my_cpu_offset);				\
+})
+#else
+#define this_cpu_ptr(ptr) raw_cpu_ptr(ptr)
+#endif
+
+#else	/* CONFIG_SMP */
+
+#define VERIFY_PERCPU_PTR(__p)						\
+({									\
+	__verify_pcpu_ptr(__p);						\
+	(typeof(*(__p)) __kernel __force *)(__p);			\
+})
+
+#define per_cpu_ptr(ptr, cpu)	({ (void)(cpu); VERIFY_PERCPU_PTR(ptr); })
+#define raw_cpu_ptr(ptr)	per_cpu_ptr(ptr, 0)
+#define this_cpu_ptr(ptr)	raw_cpu_ptr(ptr)
+
+#endif	/* CONFIG_SMP */
+
+#define per_cpu(var, cpu)	(*per_cpu_ptr(&(var), cpu))
+#define __raw_get_cpu_var(var)	(*raw_cpu_ptr(&(var)))
+#define __get_cpu_var(var)	(*this_cpu_ptr(&(var)))
+
+/* keep until we have removed all uses of __this_cpu_ptr */
+#define __this_cpu_ptr(ptr)	raw_cpu_ptr(ptr)
+
+/*
+ * Must be an lvalue. Since @var must be a simple identifier,
+ * we force a syntax error here if it isn't.
+ */
+#define get_cpu_var(var)						\
+(*({									\
+	preempt_disable();						\
+	this_cpu_ptr(&var);						\
+}))
+
+/*
+ * The weird & is necessary because sparse considers (void)(var) to be
+ * a direct dereference of percpu variable (var).
+ */
+#define put_cpu_var(var)						\
+do {									\
+	(void)&(var);							\
+	preempt_enable();						\
+} while (0)
+
+#define get_cpu_ptr(var)						\
+({									\
+	preempt_disable();						\
+	this_cpu_ptr(var);						\
+})
+
+#define put_cpu_ptr(var)						\
+do {									\
+	(void)(var);							\
+	preempt_enable();						\
+} while (0)
+
+/*
+ * Branching function to split up a function into a set of functions that
+ * are called for different scalar sizes of the objects handled.
+ */
+
+extern void __bad_size_call_parameter(void);
+
+#ifdef CONFIG_DEBUG_PREEMPT
+extern void __this_cpu_preempt_check(const char *op);
+#else
+static inline void __this_cpu_preempt_check(const char *op) { }
+#endif
+
+#define __pcpu_size_call_return(stem, variable)				\
+({									\
+	typeof(variable) pscr_ret__;					\
+	__verify_pcpu_ptr(&(variable));					\
+	switch(sizeof(variable)) {					\
+	case 1: pscr_ret__ = stem##1(variable); break;			\
+	case 2: pscr_ret__ = stem##2(variable); break;			\
+	case 4: pscr_ret__ = stem##4(variable); break;			\
+	case 8: pscr_ret__ = stem##8(variable); break;			\
+	default:							\
+		__bad_size_call_parameter(); break;			\
+	}								\
+	pscr_ret__;							\
+})
+
+#define __pcpu_size_call_return2(stem, variable, ...)			\
+({									\
+	typeof(variable) pscr2_ret__;					\
+	__verify_pcpu_ptr(&(variable));					\
+	switch(sizeof(variable)) {					\
+	case 1: pscr2_ret__ = stem##1(variable, __VA_ARGS__); break;	\
+	case 2: pscr2_ret__ = stem##2(variable, __VA_ARGS__); break;	\
+	case 4: pscr2_ret__ = stem##4(variable, __VA_ARGS__); break;	\
+	case 8: pscr2_ret__ = stem##8(variable, __VA_ARGS__); break;	\
+	default:							\
+		__bad_size_call_parameter(); break;			\
+	}								\
+	pscr2_ret__;							\
+})
+
+/*
+ * Special handling for cmpxchg_double.  cmpxchg_double is passed two
+ * percpu variables.  The first has to be aligned to a double word
+ * boundary and the second has to follow directly thereafter.
+ * We enforce this on all architectures even if they don't support
+ * a double cmpxchg instruction, since it's a cheap requirement, and it
+ * avoids breaking the requirement for architectures with the instruction.
+ */
+#define __pcpu_double_call_return_bool(stem, pcp1, pcp2, ...)		\
+({									\
+	bool pdcrb_ret__;						\
+	__verify_pcpu_ptr(&(pcp1));					\
+	BUILD_BUG_ON(sizeof(pcp1) != sizeof(pcp2));			\
+	VM_BUG_ON((unsigned long)(&(pcp1)) % (2 * sizeof(pcp1)));	\
+	VM_BUG_ON((unsigned long)(&(pcp2)) !=				\
+		  (unsigned long)(&(pcp1)) + sizeof(pcp1));		\
+	switch(sizeof(pcp1)) {						\
+	case 1: pdcrb_ret__ = stem##1(pcp1, pcp2, __VA_ARGS__); break;	\
+	case 2: pdcrb_ret__ = stem##2(pcp1, pcp2, __VA_ARGS__); break;	\
+	case 4: pdcrb_ret__ = stem##4(pcp1, pcp2, __VA_ARGS__); break;	\
+	case 8: pdcrb_ret__ = stem##8(pcp1, pcp2, __VA_ARGS__); break;	\
+	default:							\
+		__bad_size_call_parameter(); break;			\
+	}								\
+	pdcrb_ret__;							\
+})
+
+#define __pcpu_size_call(stem, variable, ...)				\
+do {									\
+	__verify_pcpu_ptr(&(variable));					\
+	switch(sizeof(variable)) {					\
+		case 1: stem##1(variable, __VA_ARGS__);break;		\
+		case 2: stem##2(variable, __VA_ARGS__);break;		\
+		case 4: stem##4(variable, __VA_ARGS__);break;		\
+		case 8: stem##8(variable, __VA_ARGS__);break;		\
+		default: 						\
+			__bad_size_call_parameter();break;		\
+	}								\
+} while (0)
+
+/*
+ * this_cpu operations (C) 2008-2013 Christoph Lameter <cl@linux.com>
+ *
+ * Optimized manipulation for memory allocated through the per cpu
+ * allocator or for addresses of per cpu variables.
+ *
+ * These operation guarantee exclusivity of access for other operations
+ * on the *same* processor. The assumption is that per cpu data is only
+ * accessed by a single processor instance (the current one).
+ *
+ * The arch code can provide optimized implementation by defining macros
+ * for certain scalar sizes. F.e. provide this_cpu_add_2() to provide per
+ * cpu atomic operations for 2 byte sized RMW actions. If arch code does
+ * not provide operations for a scalar size then the fallback in the
+ * generic code will be used.
+ *
+ * cmpxchg_double replaces two adjacent scalars at once.  The first two
+ * parameters are per cpu variables which have to be of the same size.  A
+ * truth value is returned to indicate success or failure (since a double
+ * register result is difficult to handle).  There is very limited hardware
+ * support for these operations, so only certain sizes may work.
+ */
+
+/*
+ * Operations for contexts where we do not want to do any checks for
+ * preemptions.  Unless strictly necessary, always use [__]this_cpu_*()
+ * instead.
+ *
+ * If there is no other protection through preempt disable and/or disabling
+ * interupts then one of these RMW operations can show unexpected behavior
+ * because the execution thread was rescheduled on another processor or an
+ * interrupt occurred and the same percpu variable was modified from the
+ * interrupt context.
+ */
+#define raw_cpu_read(pcp)		__pcpu_size_call_return(raw_cpu_read_, pcp)
+#define raw_cpu_write(pcp, val)		__pcpu_size_call(raw_cpu_write_, pcp, val)
+#define raw_cpu_add(pcp, val)		__pcpu_size_call(raw_cpu_add_, pcp, val)
+#define raw_cpu_and(pcp, val)		__pcpu_size_call(raw_cpu_and_, pcp, val)
+#define raw_cpu_or(pcp, val)		__pcpu_size_call(raw_cpu_or_, pcp, val)
+#define raw_cpu_add_return(pcp, val)	__pcpu_size_call_return2(raw_cpu_add_return_, pcp, val)
+#define raw_cpu_xchg(pcp, nval)		__pcpu_size_call_return2(raw_cpu_xchg_, pcp, nval)
+#define raw_cpu_cmpxchg(pcp, oval, nval) \
+	__pcpu_size_call_return2(raw_cpu_cmpxchg_, pcp, oval, nval)
+#define raw_cpu_cmpxchg_double(pcp1, pcp2, oval1, oval2, nval1, nval2) \
+	__pcpu_double_call_return_bool(raw_cpu_cmpxchg_double_, pcp1, pcp2, oval1, oval2, nval1, nval2)
+
+#define raw_cpu_sub(pcp, val)		raw_cpu_add(pcp, -(val))
+#define raw_cpu_inc(pcp)		raw_cpu_add(pcp, 1)
+#define raw_cpu_dec(pcp)		raw_cpu_sub(pcp, 1)
+#define raw_cpu_sub_return(pcp, val)	raw_cpu_add_return(pcp, -(typeof(pcp))(val))
+#define raw_cpu_inc_return(pcp)		raw_cpu_add_return(pcp, 1)
+#define raw_cpu_dec_return(pcp)		raw_cpu_add_return(pcp, -1)
+
+/*
+ * Operations for contexts that are safe from preemption/interrupts.  These
+ * operations verify that preemption is disabled.
+ */
+#define __this_cpu_read(pcp)						\
+({									\
+	__this_cpu_preempt_check("read");				\
+	raw_cpu_read(pcp);						\
+})
+
+#define __this_cpu_write(pcp, val)					\
+({									\
+	__this_cpu_preempt_check("write");				\
+	raw_cpu_write(pcp, val);					\
+})
+
+#define __this_cpu_add(pcp, val)					\
+({									\
+	__this_cpu_preempt_check("add");				\
+	raw_cpu_add(pcp, val);						\
+})
+
+#define __this_cpu_and(pcp, val)					\
+({									\
+	__this_cpu_preempt_check("and");				\
+	raw_cpu_and(pcp, val);						\
+})
+
+#define __this_cpu_or(pcp, val)						\
+({									\
+	__this_cpu_preempt_check("or");					\
+	raw_cpu_or(pcp, val);						\
+})
+
+#define __this_cpu_add_return(pcp, val)					\
+({									\
+	__this_cpu_preempt_check("add_return");				\
+	raw_cpu_add_return(pcp, val);					\
+})
+
+#define __this_cpu_xchg(pcp, nval)					\
+({									\
+	__this_cpu_preempt_check("xchg");				\
+	raw_cpu_xchg(pcp, nval);					\
+})
+
+#define __this_cpu_cmpxchg(pcp, oval, nval)				\
+({									\
+	__this_cpu_preempt_check("cmpxchg");				\
+	raw_cpu_cmpxchg(pcp, oval, nval);				\
+})
+
+#define __this_cpu_cmpxchg_double(pcp1, pcp2, oval1, oval2, nval1, nval2) \
+({	__this_cpu_preempt_check("cmpxchg_double");			\
+	raw_cpu_cmpxchg_double(pcp1, pcp2, oval1, oval2, nval1, nval2);	\
+})
+
+#define __this_cpu_sub(pcp, val)	__this_cpu_add(pcp, -(typeof(pcp))(val))
+#define __this_cpu_inc(pcp)		__this_cpu_add(pcp, 1)
+#define __this_cpu_dec(pcp)		__this_cpu_sub(pcp, 1)
+#define __this_cpu_sub_return(pcp, val)	__this_cpu_add_return(pcp, -(typeof(pcp))(val))
+#define __this_cpu_inc_return(pcp)	__this_cpu_add_return(pcp, 1)
+#define __this_cpu_dec_return(pcp)	__this_cpu_add_return(pcp, -1)
+
+/*
+ * Operations with implied preemption protection.  These operations can be
+ * used without worrying about preemption.  Note that interrupts may still
+ * occur while an operation is in progress and if the interrupt modifies
+ * the variable too then RMW actions may not be reliable.
+ */
+#define this_cpu_read(pcp)		__pcpu_size_call_return(this_cpu_read_, pcp)
+#define this_cpu_write(pcp, val)	__pcpu_size_call(this_cpu_write_, pcp, val)
+#define this_cpu_add(pcp, val)		__pcpu_size_call(this_cpu_add_, pcp, val)
+#define this_cpu_and(pcp, val)		__pcpu_size_call(this_cpu_and_, pcp, val)
+#define this_cpu_or(pcp, val)		__pcpu_size_call(this_cpu_or_, pcp, val)
+#define this_cpu_add_return(pcp, val)	__pcpu_size_call_return2(this_cpu_add_return_, pcp, val)
+#define this_cpu_xchg(pcp, nval)	__pcpu_size_call_return2(this_cpu_xchg_, pcp, nval)
+#define this_cpu_cmpxchg(pcp, oval, nval) \
+	__pcpu_size_call_return2(this_cpu_cmpxchg_, pcp, oval, nval)
+#define this_cpu_cmpxchg_double(pcp1, pcp2, oval1, oval2, nval1, nval2) \
+	__pcpu_double_call_return_bool(this_cpu_cmpxchg_double_, pcp1, pcp2, oval1, oval2, nval1, nval2)
+
+#define this_cpu_sub(pcp, val)		this_cpu_add(pcp, -(typeof(pcp))(val))
+#define this_cpu_inc(pcp)		this_cpu_add(pcp, 1)
+#define this_cpu_dec(pcp)		this_cpu_sub(pcp, 1)
+#define this_cpu_sub_return(pcp, val)	this_cpu_add_return(pcp, -(typeof(pcp))(val))
+#define this_cpu_inc_return(pcp)	this_cpu_add_return(pcp, 1)
+#define this_cpu_dec_return(pcp)	this_cpu_add_return(pcp, -1)
+
+#endif /* __ASSEMBLY__ */
 #endif /* _LINUX_PERCPU_DEFS_H */
diff --git a/include/linux/percpu-refcount.h b/include/linux/percpu-refcount.h
index 5d8920e23073..3dfbf237cd8f 100644
--- a/include/linux/percpu-refcount.h
+++ b/include/linux/percpu-refcount.h
@@ -57,11 +57,9 @@ struct percpu_ref {
 	atomic_t		count;
 	/*
 	 * The low bit of the pointer indicates whether the ref is in percpu
-	 * mode; if set, then get/put will manipulate the atomic_t (this is a
-	 * hack because we need to keep the pointer around for
-	 * percpu_ref_kill_rcu())
+	 * mode; if set, then get/put will manipulate the atomic_t.
 	 */
-	unsigned __percpu	*pcpu_count;
+	unsigned long		pcpu_count_ptr;
 	percpu_ref_func_t	*release;
 	percpu_ref_func_t	*confirm_kill;
 	struct rcu_head		rcu;
@@ -69,7 +67,8 @@ struct percpu_ref {
 
 int __must_check percpu_ref_init(struct percpu_ref *ref,
 				 percpu_ref_func_t *release);
-void percpu_ref_cancel_init(struct percpu_ref *ref);
+void percpu_ref_reinit(struct percpu_ref *ref);
+void percpu_ref_exit(struct percpu_ref *ref);
 void percpu_ref_kill_and_confirm(struct percpu_ref *ref,
 				 percpu_ref_func_t *confirm_kill);
 
@@ -88,12 +87,28 @@ static inline void percpu_ref_kill(struct percpu_ref *ref)
 	return percpu_ref_kill_and_confirm(ref, NULL);
 }
 
-#define PCPU_STATUS_BITS	2<