freebsd-skq/sys/amd64/include/atomic.h
Konstantin Belousov 8954a9a4e6 Add the atomic_thread_fence() family of functions with intent to
provide a semantic defined by the C11 fences with corresponding
memory_order.

atomic_thread_fence_acq() gives r | r, w, where r and w are read and
write accesses, and | denotes the fence itself.

atomic_thread_fence_rel() is r, w | w.

atomic_thread_fence_acq_rel() is the combination of the acquire and
release in single operation.  Note that reads after the acq+rel fence
could be made visible before writes preceeding the fence.

atomic_thread_fence_seq_cst() orders all accesses before/after the
fence, and the fence itself is globally ordered against other
sequentially consistent atomic operations.

Reviewed by:	alc
Discussed with:	bde
Sponsored by:	The FreeBSD Foundation
MFC after:	3 weeks
2015-07-08 18:12:24 +00:00

578 lines
18 KiB
C

/*-
* Copyright (c) 1998 Doug Rabson
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* $FreeBSD$
*/
#ifndef _MACHINE_ATOMIC_H_
#define _MACHINE_ATOMIC_H_
#ifndef _SYS_CDEFS_H_
#error this file needs sys/cdefs.h as a prerequisite
#endif
#define mb() __asm __volatile("mfence;" : : : "memory")
#define wmb() __asm __volatile("sfence;" : : : "memory")
#define rmb() __asm __volatile("lfence;" : : : "memory")
/*
* Various simple operations on memory, each of which is atomic in the
* presence of interrupts and multiple processors.
*
* atomic_set_char(P, V) (*(u_char *)(P) |= (V))
* atomic_clear_char(P, V) (*(u_char *)(P) &= ~(V))
* atomic_add_char(P, V) (*(u_char *)(P) += (V))
* atomic_subtract_char(P, V) (*(u_char *)(P) -= (V))
*
* atomic_set_short(P, V) (*(u_short *)(P) |= (V))
* atomic_clear_short(P, V) (*(u_short *)(P) &= ~(V))
* atomic_add_short(P, V) (*(u_short *)(P) += (V))
* atomic_subtract_short(P, V) (*(u_short *)(P) -= (V))
*
* atomic_set_int(P, V) (*(u_int *)(P) |= (V))
* atomic_clear_int(P, V) (*(u_int *)(P) &= ~(V))
* atomic_add_int(P, V) (*(u_int *)(P) += (V))
* atomic_subtract_int(P, V) (*(u_int *)(P) -= (V))
* atomic_swap_int(P, V) (return (*(u_int *)(P)); *(u_int *)(P) = (V);)
* atomic_readandclear_int(P) (return (*(u_int *)(P)); *(u_int *)(P) = 0;)
*
* atomic_set_long(P, V) (*(u_long *)(P) |= (V))
* atomic_clear_long(P, V) (*(u_long *)(P) &= ~(V))
* atomic_add_long(P, V) (*(u_long *)(P) += (V))
* atomic_subtract_long(P, V) (*(u_long *)(P) -= (V))
* atomic_swap_long(P, V) (return (*(u_long *)(P)); *(u_long *)(P) = (V);)
* atomic_readandclear_long(P) (return (*(u_long *)(P)); *(u_long *)(P) = 0;)
*/
/*
* The above functions are expanded inline in the statically-linked
* kernel. Lock prefixes are generated if an SMP kernel is being
* built.
*
* Kernel modules call real functions which are built into the kernel.
* This allows kernel modules to be portable between UP and SMP systems.
*/
#if defined(KLD_MODULE) || !defined(__GNUCLIKE_ASM)
#define ATOMIC_ASM(NAME, TYPE, OP, CONS, V) \
void atomic_##NAME##_##TYPE(volatile u_##TYPE *p, u_##TYPE v); \
void atomic_##NAME##_barr_##TYPE(volatile u_##TYPE *p, u_##TYPE v)
int atomic_cmpset_int(volatile u_int *dst, u_int expect, u_int src);
int atomic_cmpset_long(volatile u_long *dst, u_long expect, u_long src);
u_int atomic_fetchadd_int(volatile u_int *p, u_int v);
u_long atomic_fetchadd_long(volatile u_long *p, u_long v);
int atomic_testandset_int(volatile u_int *p, u_int v);
int atomic_testandset_long(volatile u_long *p, u_int v);
void atomic_thread_fence_acq(void);
void atomic_thread_fence_acq_rel(void);
void atomic_thread_fence_rel(void);
void atomic_thread_fence_seq_cst(void);
#define ATOMIC_LOAD(TYPE) \
u_##TYPE atomic_load_acq_##TYPE(volatile u_##TYPE *p)
#define ATOMIC_STORE(TYPE) \
void atomic_store_rel_##TYPE(volatile u_##TYPE *p, u_##TYPE v)
#else /* !KLD_MODULE && __GNUCLIKE_ASM */
/*
* For userland, always use lock prefixes so that the binaries will run
* on both SMP and !SMP systems.
*/
#if defined(SMP) || !defined(_KERNEL)
#define MPLOCKED "lock ; "
#else
#define MPLOCKED
#endif
/*
* The assembly is volatilized to avoid code chunk removal by the compiler.
* GCC aggressively reorders operations and memory clobbering is necessary
* in order to avoid that for memory barriers.
*/
#define ATOMIC_ASM(NAME, TYPE, OP, CONS, V) \
static __inline void \
atomic_##NAME##_##TYPE(volatile u_##TYPE *p, u_##TYPE v)\
{ \
__asm __volatile(MPLOCKED OP \
: "+m" (*p) \
: CONS (V) \
: "cc"); \
} \
\
static __inline void \
atomic_##NAME##_barr_##TYPE(volatile u_##TYPE *p, u_##TYPE v)\
{ \
__asm __volatile(MPLOCKED OP \
: "+m" (*p) \
: CONS (V) \
: "memory", "cc"); \
} \
struct __hack
/*
* Atomic compare and set, used by the mutex functions
*
* if (*dst == expect) *dst = src (all 32 bit words)
*
* Returns 0 on failure, non-zero on success
*/
static __inline int
atomic_cmpset_int(volatile u_int *dst, u_int expect, u_int src)
{
u_char res;
__asm __volatile(
" " MPLOCKED " "
" cmpxchgl %3,%1 ; "
" sete %0 ; "
"# atomic_cmpset_int"
: "=q" (res), /* 0 */
"+m" (*dst), /* 1 */
"+a" (expect) /* 2 */
: "r" (src) /* 3 */
: "memory", "cc");
return (res);
}
static __inline int
atomic_cmpset_long(volatile u_long *dst, u_long expect, u_long src)
{
u_char res;
__asm __volatile(
" " MPLOCKED " "
" cmpxchgq %3,%1 ; "
" sete %0 ; "
"# atomic_cmpset_long"
: "=q" (res), /* 0 */
"+m" (*dst), /* 1 */
"+a" (expect) /* 2 */
: "r" (src) /* 3 */
: "memory", "cc");
return (res);
}
/*
* Atomically add the value of v to the integer pointed to by p and return
* the previous value of *p.
*/
static __inline u_int
atomic_fetchadd_int(volatile u_int *p, u_int v)
{
__asm __volatile(
" " MPLOCKED " "
" xaddl %0,%1 ; "
"# atomic_fetchadd_int"
: "+r" (v), /* 0 */
"+m" (*p) /* 1 */
: : "cc");
return (v);
}
/*
* Atomically add the value of v to the long integer pointed to by p and return
* the previous value of *p.
*/
static __inline u_long
atomic_fetchadd_long(volatile u_long *p, u_long v)
{
__asm __volatile(
" " MPLOCKED " "
" xaddq %0,%1 ; "
"# atomic_fetchadd_long"
: "+r" (v), /* 0 */
"+m" (*p) /* 1 */
: : "cc");
return (v);
}
static __inline int
atomic_testandset_int(volatile u_int *p, u_int v)
{
u_char res;
__asm __volatile(
" " MPLOCKED " "
" btsl %2,%1 ; "
" setc %0 ; "
"# atomic_testandset_int"
: "=q" (res), /* 0 */
"+m" (*p) /* 1 */
: "Ir" (v & 0x1f) /* 2 */
: "cc");
return (res);
}
static __inline int
atomic_testandset_long(volatile u_long *p, u_int v)
{
u_char res;
__asm __volatile(
" " MPLOCKED " "
" btsq %2,%1 ; "
" setc %0 ; "
"# atomic_testandset_long"
: "=q" (res), /* 0 */
"+m" (*p) /* 1 */
: "Jr" ((u_long)(v & 0x3f)) /* 2 */
: "cc");
return (res);
}
/*
* We assume that a = b will do atomic loads and stores. Due to the
* IA32 memory model, a simple store guarantees release semantics.
*
* However, a load may pass a store if they are performed on distinct
* addresses, so for atomic_load_acq we introduce a Store/Load barrier
* before the load in SMP kernels. We use "lock addl $0,mem", as
* recommended by the AMD Software Optimization Guide, and not mfence.
* In the kernel, we use a private per-cpu cache line as the target
* for the locked addition, to avoid introducing false data
* dependencies. In userspace, a word in the red zone on the stack
* (-8(%rsp)) is utilized.
*
* For UP kernels, however, the memory of the single processor is
* always consistent, so we only need to stop the compiler from
* reordering accesses in a way that violates the semantics of acquire
* and release.
*/
#if defined(_KERNEL)
/*
* OFFSETOF_MONITORBUF == __pcpu_offset(pc_monitorbuf).
*
* The open-coded number is used instead of the symbolic expression to
* avoid a dependency on sys/pcpu.h in machine/atomic.h consumers.
* An assertion in amd64/vm_machdep.c ensures that the value is correct.
*/
#define OFFSETOF_MONITORBUF 0x180
#if defined(SMP)
static __inline void
__storeload_barrier(void)
{
__asm __volatile("lock; addl $0,%%gs:%0"
: "+m" (*(u_int *)OFFSETOF_MONITORBUF) : : "memory", "cc");
}
#else /* _KERNEL && UP */
static __inline void
__storeload_barrier(void)
{
__compiler_membar();
}
#endif /* SMP */
#else /* !_KERNEL */
static __inline void
__storeload_barrier(void)
{
__asm __volatile("lock; addl $0,-8(%%rsp)" : : : "memory", "cc");
}
#endif /* _KERNEL*/
/*
* C11-standard acq/rel semantics only apply when the variable in the
* call is the same for acq as it is for rel. However, our previous
* (x86) implementations provided much stronger ordering than required
* (essentially what is called seq_cst order in C11). This
* implementation provides the historical strong ordering since some
* callers depend on it.
*/
#define ATOMIC_LOAD(TYPE) \
static __inline u_##TYPE \
atomic_load_acq_##TYPE(volatile u_##TYPE *p) \
{ \
u_##TYPE res; \
\
__storeload_barrier(); \
res = *p; \
__compiler_membar(); \
return (res); \
} \
struct __hack
#define ATOMIC_STORE(TYPE) \
static __inline void \
atomic_store_rel_##TYPE(volatile u_##TYPE *p, u_##TYPE v) \
{ \
\
__compiler_membar(); \
*p = v; \
} \
struct __hack
static __inline void
atomic_thread_fence_acq(void)
{
__compiler_membar();
}
static __inline void
atomic_thread_fence_rel(void)
{
__compiler_membar();
}
static __inline void
atomic_thread_fence_acq_rel(void)
{
__compiler_membar();
}
static __inline void
atomic_thread_fence_seq_cst(void)
{
__storeload_barrier();
}
#endif /* KLD_MODULE || !__GNUCLIKE_ASM */
ATOMIC_ASM(set, char, "orb %b1,%0", "iq", v);
ATOMIC_ASM(clear, char, "andb %b1,%0", "iq", ~v);
ATOMIC_ASM(add, char, "addb %b1,%0", "iq", v);
ATOMIC_ASM(subtract, char, "subb %b1,%0", "iq", v);
ATOMIC_ASM(set, short, "orw %w1,%0", "ir", v);
ATOMIC_ASM(clear, short, "andw %w1,%0", "ir", ~v);
ATOMIC_ASM(add, short, "addw %w1,%0", "ir", v);
ATOMIC_ASM(subtract, short, "subw %w1,%0", "ir", v);
ATOMIC_ASM(set, int, "orl %1,%0", "ir", v);
ATOMIC_ASM(clear, int, "andl %1,%0", "ir", ~v);
ATOMIC_ASM(add, int, "addl %1,%0", "ir", v);
ATOMIC_ASM(subtract, int, "subl %1,%0", "ir", v);
ATOMIC_ASM(set, long, "orq %1,%0", "ir", v);
ATOMIC_ASM(clear, long, "andq %1,%0", "ir", ~v);
ATOMIC_ASM(add, long, "addq %1,%0", "ir", v);
ATOMIC_ASM(subtract, long, "subq %1,%0", "ir", v);
#define ATOMIC_LOADSTORE(TYPE) \
ATOMIC_LOAD(TYPE); \
ATOMIC_STORE(TYPE)
ATOMIC_LOADSTORE(char);
ATOMIC_LOADSTORE(short);
ATOMIC_LOADSTORE(int);
ATOMIC_LOADSTORE(long);
#undef ATOMIC_ASM
#undef ATOMIC_LOAD
#undef ATOMIC_STORE
#undef ATOMIC_LOADSTORE
#ifndef WANT_FUNCTIONS
/* Read the current value and store a new value in the destination. */
#ifdef __GNUCLIKE_ASM
static __inline u_int
atomic_swap_int(volatile u_int *p, u_int v)
{
__asm __volatile(
" xchgl %1,%0 ; "
"# atomic_swap_int"
: "+r" (v), /* 0 */
"+m" (*p)); /* 1 */
return (v);
}
static __inline u_long
atomic_swap_long(volatile u_long *p, u_long v)
{
__asm __volatile(
" xchgq %1,%0 ; "
"# atomic_swap_long"
: "+r" (v), /* 0 */
"+m" (*p)); /* 1 */
return (v);
}
#else /* !__GNUCLIKE_ASM */
u_int atomic_swap_int(volatile u_int *p, u_int v);
u_long atomic_swap_long(volatile u_long *p, u_long v);
#endif /* __GNUCLIKE_ASM */
#define atomic_set_acq_char atomic_set_barr_char
#define atomic_set_rel_char atomic_set_barr_char
#define atomic_clear_acq_char atomic_clear_barr_char
#define atomic_clear_rel_char atomic_clear_barr_char
#define atomic_add_acq_char atomic_add_barr_char
#define atomic_add_rel_char atomic_add_barr_char
#define atomic_subtract_acq_char atomic_subtract_barr_char
#define atomic_subtract_rel_char atomic_subtract_barr_char
#define atomic_set_acq_short atomic_set_barr_short
#define atomic_set_rel_short atomic_set_barr_short
#define atomic_clear_acq_short atomic_clear_barr_short
#define atomic_clear_rel_short atomic_clear_barr_short
#define atomic_add_acq_short atomic_add_barr_short
#define atomic_add_rel_short atomic_add_barr_short
#define atomic_subtract_acq_short atomic_subtract_barr_short
#define atomic_subtract_rel_short atomic_subtract_barr_short
#define atomic_set_acq_int atomic_set_barr_int
#define atomic_set_rel_int atomic_set_barr_int
#define atomic_clear_acq_int atomic_clear_barr_int
#define atomic_clear_rel_int atomic_clear_barr_int
#define atomic_add_acq_int atomic_add_barr_int
#define atomic_add_rel_int atomic_add_barr_int
#define atomic_subtract_acq_int atomic_subtract_barr_int
#define atomic_subtract_rel_int atomic_subtract_barr_int
#define atomic_cmpset_acq_int atomic_cmpset_int
#define atomic_cmpset_rel_int atomic_cmpset_int
#define atomic_set_acq_long atomic_set_barr_long
#define atomic_set_rel_long atomic_set_barr_long
#define atomic_clear_acq_long atomic_clear_barr_long
#define atomic_clear_rel_long atomic_clear_barr_long
#define atomic_add_acq_long atomic_add_barr_long
#define atomic_add_rel_long atomic_add_barr_long
#define atomic_subtract_acq_long atomic_subtract_barr_long
#define atomic_subtract_rel_long atomic_subtract_barr_long
#define atomic_cmpset_acq_long atomic_cmpset_long
#define atomic_cmpset_rel_long atomic_cmpset_long
#define atomic_readandclear_int(p) atomic_swap_int(p, 0)
#define atomic_readandclear_long(p) atomic_swap_long(p, 0)
/* Operations on 8-bit bytes. */
#define atomic_set_8 atomic_set_char
#define atomic_set_acq_8 atomic_set_acq_char
#define atomic_set_rel_8 atomic_set_rel_char
#define atomic_clear_8 atomic_clear_char
#define atomic_clear_acq_8 atomic_clear_acq_char
#define atomic_clear_rel_8 atomic_clear_rel_char
#define atomic_add_8 atomic_add_char
#define atomic_add_acq_8 atomic_add_acq_char
#define atomic_add_rel_8 atomic_add_rel_char
#define atomic_subtract_8 atomic_subtract_char
#define atomic_subtract_acq_8 atomic_subtract_acq_char
#define atomic_subtract_rel_8 atomic_subtract_rel_char
#define atomic_load_acq_8 atomic_load_acq_char
#define atomic_store_rel_8 atomic_store_rel_char
/* Operations on 16-bit words. */
#define atomic_set_16 atomic_set_short
#define atomic_set_acq_16 atomic_set_acq_short
#define atomic_set_rel_16 atomic_set_rel_short
#define atomic_clear_16 atomic_clear_short
#define atomic_clear_acq_16 atomic_clear_acq_short
#define atomic_clear_rel_16 atomic_clear_rel_short
#define atomic_add_16 atomic_add_short
#define atomic_add_acq_16 atomic_add_acq_short
#define atomic_add_rel_16 atomic_add_rel_short
#define atomic_subtract_16 atomic_subtract_short
#define atomic_subtract_acq_16 atomic_subtract_acq_short
#define atomic_subtract_rel_16 atomic_subtract_rel_short
#define atomic_load_acq_16 atomic_load_acq_short
#define atomic_store_rel_16 atomic_store_rel_short
/* Operations on 32-bit double words. */
#define atomic_set_32 atomic_set_int
#define atomic_set_acq_32 atomic_set_acq_int
#define atomic_set_rel_32 atomic_set_rel_int
#define atomic_clear_32 atomic_clear_int
#define atomic_clear_acq_32 atomic_clear_acq_int
#define atomic_clear_rel_32 atomic_clear_rel_int
#define atomic_add_32 atomic_add_int
#define atomic_add_acq_32 atomic_add_acq_int
#define atomic_add_rel_32 atomic_add_rel_int
#define atomic_subtract_32 atomic_subtract_int
#define atomic_subtract_acq_32 atomic_subtract_acq_int
#define atomic_subtract_rel_32 atomic_subtract_rel_int
#define atomic_load_acq_32 atomic_load_acq_int
#define atomic_store_rel_32 atomic_store_rel_int
#define atomic_cmpset_32 atomic_cmpset_int
#define atomic_cmpset_acq_32 atomic_cmpset_acq_int
#define atomic_cmpset_rel_32 atomic_cmpset_rel_int
#define atomic_swap_32 atomic_swap_int
#define atomic_readandclear_32 atomic_readandclear_int
#define atomic_fetchadd_32 atomic_fetchadd_int
#define atomic_testandset_32 atomic_testandset_int
/* Operations on 64-bit quad words. */
#define atomic_set_64 atomic_set_long
#define atomic_set_acq_64 atomic_set_acq_long
#define atomic_set_rel_64 atomic_set_rel_long
#define atomic_clear_64 atomic_clear_long
#define atomic_clear_acq_64 atomic_clear_acq_long
#define atomic_clear_rel_64 atomic_clear_rel_long
#define atomic_add_64 atomic_add_long
#define atomic_add_acq_64 atomic_add_acq_long
#define atomic_add_rel_64 atomic_add_rel_long
#define atomic_subtract_64 atomic_subtract_long
#define atomic_subtract_acq_64 atomic_subtract_acq_long
#define atomic_subtract_rel_64 atomic_subtract_rel_long
#define atomic_load_acq_64 atomic_load_acq_long
#define atomic_store_rel_64 atomic_store_rel_long
#define atomic_cmpset_64 atomic_cmpset_long
#define atomic_cmpset_acq_64 atomic_cmpset_acq_long
#define atomic_cmpset_rel_64 atomic_cmpset_rel_long
#define atomic_swap_64 atomic_swap_long
#define atomic_readandclear_64 atomic_readandclear_long
#define atomic_testandset_64 atomic_testandset_long
/* Operations on pointers. */
#define atomic_set_ptr atomic_set_long
#define atomic_set_acq_ptr atomic_set_acq_long
#define atomic_set_rel_ptr atomic_set_rel_long
#define atomic_clear_ptr atomic_clear_long
#define atomic_clear_acq_ptr atomic_clear_acq_long
#define atomic_clear_rel_ptr atomic_clear_rel_long
#define atomic_add_ptr atomic_add_long
#define atomic_add_acq_ptr atomic_add_acq_long
#define atomic_add_rel_ptr atomic_add_rel_long
#define atomic_subtract_ptr atomic_subtract_long
#define atomic_subtract_acq_ptr atomic_subtract_acq_long
#define atomic_subtract_rel_ptr atomic_subtract_rel_long
#define atomic_load_acq_ptr atomic_load_acq_long
#define atomic_store_rel_ptr atomic_store_rel_long
#define atomic_cmpset_ptr atomic_cmpset_long
#define atomic_cmpset_acq_ptr atomic_cmpset_acq_long
#define atomic_cmpset_rel_ptr atomic_cmpset_rel_long
#define atomic_swap_ptr atomic_swap_long
#define atomic_readandclear_ptr atomic_readandclear_long
#endif /* !WANT_FUNCTIONS */
#endif /* !_MACHINE_ATOMIC_H_ */