Document the semantics of atomic_thread_fence operations.

[FreeBSD/FreeBSD.git] / share / man / man9 / atomic.9

.\" Copyright (c) 2000-2001 John H. Baldwin <jhb@FreeBSD.org>
.\" 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 DEVELOPERS ``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 DEVELOPERS 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.
.\"
.\" $FreeBSD$
.\"
.Dd December 19, 2017
.Dt ATOMIC 9
.Os
.Sh NAME
.Nm atomic_add ,
.Nm atomic_clear ,
.Nm atomic_cmpset ,
.Nm atomic_fcmpset ,
.Nm atomic_fetchadd ,
.Nm atomic_load ,
.Nm atomic_readandclear ,
.Nm atomic_set ,
.Nm atomic_subtract ,
.Nm atomic_store ,
.Nm atomic_thread_fence
.Nd atomic operations
.Sh SYNOPSIS
.In sys/types.h
.In machine/atomic.h
.Ft void
.Fn atomic_add_[acq_|rel_]<type> "volatile <type> *p" "<type> v"
.Ft void
.Fn atomic_clear_[acq_|rel_]<type> "volatile <type> *p" "<type> v"
.Ft int
.Fo atomic_cmpset_[acq_|rel_]<type>
.Fa "volatile <type> *dst"
.Fa "<type> old"
.Fa "<type> new"
.Fc
.Ft int
.Fo atomic_fcmpset_[acq_|rel_]<type>
.Fa "volatile <type> *dst"
.Fa "<type> *old"
.Fa "<type> new"
.Fc
.Ft <type>
.Fn atomic_fetchadd_<type> "volatile <type> *p" "<type> v"
.Ft <type>
.Fn atomic_load_[acq_]<type> "volatile <type> *p"
.Ft <type>
.Fn atomic_readandclear_<type> "volatile <type> *p"
.Ft void
.Fn atomic_set_[acq_|rel_]<type> "volatile <type> *p" "<type> v"
.Ft void
.Fn atomic_subtract_[acq_|rel_]<type> "volatile <type> *p" "<type> v"
.Ft void
.Fn atomic_store_[rel_]<type> "volatile <type> *p" "<type> v"
.Ft <type>
.Fn atomic_swap_<type> "volatile <type> *p" "<type> v"
.Ft int
.Fn atomic_testandclear_<type> "volatile <type> *p" "u_int v"
.Ft int
.Fn atomic_testandset_<type> "volatile <type> *p" "u_int v"
.Ft void
.Fn atomic_thread_fence_[acq|acq_rel|rel|seq_cst] "void"
.Sh DESCRIPTION
Atomic operations are commonly used to implement reference counts and as
building blocks for synchronization primitives, such as mutexes.
.Pp
All of these operations are performed
.Em atomically
across multiple threads and in the presence of interrupts, meaning that they
are performed in an indivisible manner from the perspective of concurrently
running threads and interrupt handlers.
.Pp
On all architectures supported by
.Fx ,
ordinary loads and stores of integers in cache-coherent memory are
inherently atomic if the integer is naturally aligned and its size does not
exceed the processor's word size.
However, such loads and stores may be elided from the program by
the compiler, whereas atomic operations are always performed.
.Pp
When atomic operations are performed on cache-coherent memory, all
operations on the same location are totally ordered.
.Pp
When an atomic load is performed on a location in cache-coherent memory,
it reads the entire value that was defined by the last atomic store to
each byte of the location.
An atomic load will never return a value out of thin air.
When an atomic store is performed on a location, no other thread or
interrupt handler will observe a
.Em torn write ,
or partial modification of the location.
.Pp
Except as noted below, the semantics of these operations are almost
identical to the semantics of similarly named C11 atomic operations.
.Ss Types
Most atomic operations act upon a specific
.Fa type .
That type is indicated in the function name.
In contrast to C11 atomic operations,
.Fx Ns 's
atomic operations are performed on ordinary integer types.
The available types are:
.Pp
.Bl -tag -offset indent -width short -compact
.It Li int
unsigned integer
.It Li long
unsigned long integer
.It Li ptr
unsigned integer the size of a pointer
.It Li 32
unsigned 32-bit integer
.It Li 64
unsigned 64-bit integer
.El
.Pp
For example, the function to atomically add two integers is called
.Fn atomic_add_int .
.Pp
Certain architectures also provide operations for types smaller than
.Dq Li int .
.Pp
.Bl -tag -offset indent -width short -compact
.It Li char
unsigned character
.It Li short
unsigned short integer
.It Li 8
unsigned 8-bit integer
.It Li 16
unsigned 16-bit integer
.El
.Pp
These types must not be used in machine-independent code.
.Ss Acquire and Release Operations
By default, a thread's accesses to different memory locations might not be
performed in
.Em program order ,
that is, the order in which the accesses appear in the source code.
To optimize the program's execution, both the compiler and processor might
reorder the thread's accesses.
However, both ensure that their reordering of the accesses is not visible to
the thread.
Otherwise, the traditional memory model that is expected by single-threaded
programs would be violated.
Nonetheless, other threads in a multithreaded program, such as the
.Fx
kernel, might observe the reordering.
Moreover, in some cases, such as the implementation of synchronization between
threads, arbitrary reordering might result in the incorrect execution of the
program.
To constrain the reordering that both the compiler and processor might perform
on a thread's accesses, a programmer can use atomic operations with
.Em acquire
and
.Em release
semantics.
.Pp
Atomic operations on memory have up to three variants.
The first variant performs the operation without imposing any ordering
constraints on memory accesses to other locations.
The second variant has acquire semantics, and the third variant has release
semantics.
.Pp
When an atomic operation has acquire semantics, the operation must have
completed before any subsequent load or store (by program order) is
performed.
Conversely, acquire semantics do not require that prior loads or stores have
completed before the atomic operation is performed.
An atomic operation can only have acquire semantics if it performs a load
from memory.
To denote acquire semantics, the suffix
.Dq Li _acq
is inserted into the function name immediately prior to the
.Dq Li _ Ns Aq Fa type
suffix.
For example, to subtract two integers ensuring that the subtraction is
completed before any subsequent loads and stores are performed, use
.Fn atomic_subtract_acq_int .
.Pp
When an atomic operation has release semantics, all prior loads or stores
(by program order) must have completed before the operation is performed.
Conversely, release semantics do not require that the atomic operation must
have completed before any subsequent load or store is performed.
An atomic operation can only have release semantics if it performs a store
to memory.
To denote release semantics, the suffix
.Dq Li _rel
is inserted into the function name immediately prior to the
.Dq Li _ Ns Aq Fa type
suffix.
For example, to add two long integers ensuring that all prior loads and
stores are completed before the addition is performed, use
.Fn atomic_add_rel_long .
.Pp
When a release operation by one thread
.Em synchronizes with
an acquire operation by another thread, usually meaning that the acquire
operation reads the value written by the release operation, then the effects
of all prior stores by the releasing thread must become visible to
subsequent loads by the acquiring thread.
Moreover, the effects of all stores (by other threads) that were visible to
the releasing thread must also become visible to the acquiring thread.
These rules only apply to the synchronizing threads.
Other threads might observe these stores in a different order.
.Pp
In effect, atomic operations with acquire and release semantics establish
one-way barriers to reordering that enable the implementations of
synchronization primitives to express their ordering requirements without
also imposing unnecessary ordering.
For example, for a critical section guarded by a mutex, an acquire operation
when the mutex is locked and a release operation when the mutex is unlocked
will prevent any loads or stores from moving outside of the critical
section.
However, they will not prevent the compiler or processor from moving loads
or stores into the critical section, which does not violate the semantics of
a mutex.
.Ss Thread Fence Operations
Alternatively, a programmer can use atomic thread fence operations to
constrain the reordering of accesses.
In contrast to other atomic operations, fences do not, themselves, access
memory.
.Pp
When a fence has acquire semantics, all prior loads (by program order) must
have completed before any subsequent load or store is performed.
Thus, an acquire fence is a two-way barrier for load operations.
To denote acquire semantics, the suffix
.Dq Li _acq
is appended to the function name, for example,
.Fn atomic_thread_fence_acq .
.Pp
When a fence has release semantics, all prior loads or stores (by program
order) must have completed before any subsequent store operation is
performed.
Thus, a release fence is a two-way barrier for store operations.
To denote release semantics, the suffix
.Dq Li _rel
is appended to the function name, for example,
.Fn atomic_thread_fence_rel .
.Pp
Although
.Fn atomic_thread_fence_acq_rel
implements both acquire and release semantics, it is not a full barrier.
For example, a store prior to the fence (in program order) may be completed
after a load subsequent to the fence.
In contrast,
.Fn atomic_thread_fence_seq_cst
implements a full barrier.
Neither loads nor stores may cross this barrier in either direction.
.Pp
In C11, a release fence by one thread synchronizes with an acquire fence by
another thread when an atomic load that is prior to the acquire fence (by
program order) reads the value written by an atomic store that is subsequent
to the release fence.
In constrast, in FreeBSD, because of the atomicity of ordinary, naturally
aligned loads and stores, fences can also be synchronized by ordinary loads
and stores.
This simplifies the implementation and use of some synchronization
primitives in
.Fx .
.Pp
Since neither a compiler nor a processor can foresee which (atomic) load
will read the value written by an (atomic) store, the ordering constraints
imposed by fences must be more restrictive than acquire loads and release
stores.
Essentially, this is why fences are two-way barriers.
.Pp
Although fences impose more restrictive ordering than acquire loads and
release stores, by separating access from ordering, they can sometimes
facilitate more efficient implementations of synchronization primitives.
For example, they can be used to avoid executing a memory barrier until a
memory access shows that some condition is satisfied.
.Ss Multiple Processors
In multiprocessor systems, the atomicity of the atomic operations on memory
depends on support for cache coherence in the underlying architecture.
In general, cache coherence on the default memory type,
.Dv VM_MEMATTR_DEFAULT ,
is guaranteed by all architectures that are supported by
.Fx .
For example, cache coherence is guaranteed on write-back memory by the
.Tn amd64
and
.Tn i386
architectures.
However, on some architectures, cache coherence might not be enabled on all
memory types.
To determine if cache coherence is enabled for a non-default memory type,
consult the architecture's documentation.
.Ss Semantics
This section describes the semantics of each operation using a C like notation.
.Bl -hang
.It Fn atomic_add p v
.Bd -literal -compact
*p += v;
.Ed
.It Fn atomic_clear p v
.Bd -literal -compact
*p &= ~v;
.Ed
.It Fn atomic_cmpset dst old new
.Bd -literal -compact
if (*dst == old) {
        *dst = new;
        return (1);
} else
        return (0);
.Ed
.El
.Pp
Some architectures do not implement the
.Fn atomic_cmpset
functions for the types
.Dq Li char ,
.Dq Li short ,
.Dq Li 8 ,
and
.Dq Li 16 .
.Bl -hang
.It Fn atomic_fcmpset dst *old new
.El
.Pp
On architectures implementing
.Em Compare And Swap
operation in hardware, the functionality can be described as
.Bd -literal -offset indent -compact
if (*dst == *old) {
        *dst = new;
        return (1);
} else {
        *old = *dst;
        return (0);
}
.Ed
On architectures which provide
.Em Load Linked/Store Conditional
primitive, the write to
.Dv *dst
might also fail for several reasons, most important of which
is a parallel write to
.Dv *dst
cache line by other CPU.
In this case
.Fn atomic_fcmpset
function also returns
.Dv false ,
despite
.Dl *old == *dst .
.Pp
Some architectures do not implement the
.Fn atomic_fcmpset
functions for the types
.Dq Li char ,
.Dq Li short ,
.Dq Li 8 ,
and
.Dq Li 16 .
.Bl -hang
.It Fn atomic_fetchadd p v
.Bd -literal -compact
tmp = *p;
*p += v;
return (tmp);
.Ed
.El
.Pp
The
.Fn atomic_fetchadd
functions are only implemented for the types
.Dq Li int ,
.Dq Li long
and
.Dq Li 32
and do not have any variants with memory barriers at this time.
.Bl -hang
.It Fn atomic_load p
.Bd -literal -compact
return (*p);
.Ed
.It Fn atomic_readandclear p
.Bd -literal -compact
tmp = *p;
*p = 0;
return (tmp);
.Ed
.El
.Pp
The
.Fn atomic_readandclear
functions are not implemented for the types
.Dq Li char ,
.Dq Li short ,
.Dq Li ptr ,
.Dq Li 8 ,
and
.Dq Li 16
and do not have any variants with memory barriers at this time.
.Bl -hang
.It Fn atomic_set p v
.Bd -literal -compact
*p |= v;
.Ed
.It Fn atomic_subtract p v
.Bd -literal -compact
*p -= v;
.Ed
.It Fn atomic_store p v
.Bd -literal -compact
*p = v;
.Ed
.It Fn atomic_swap p v
.Bd -literal -compact
tmp = *p;
*p = v;
return (tmp);
.Ed
.El
.Pp
The
.Fn atomic_swap
functions are not implemented for the types
.Dq Li char ,
.Dq Li short ,
.Dq Li ptr ,
.Dq Li 8 ,
and
.Dq Li 16
and do not have any variants with memory barriers at this time.
.Bl -hang
.It Fn atomic_testandclear p v
.Bd -literal -compact
bit = 1 << (v % (sizeof(*p) * NBBY));
tmp = (*p & bit) != 0;
*p &= ~bit;
return (tmp);
.Ed
.El
.Bl -hang
.It Fn atomic_testandset p v
.Bd -literal -compact
bit = 1 << (v % (sizeof(*p) * NBBY));
tmp = (*p & bit) != 0;
*p |= bit;
return (tmp);
.Ed
.El
.Pp
The
.Fn atomic_testandset
and
.Fn atomic_testandclear
functions are only implemented for the types
.Dq Li int ,
.Dq Li long
and
.Dq Li 32
and do not have any variants with memory barriers at this time.
.Pp
The type
.Dq Li 64
is currently not implemented for any of the atomic operations on the
.Tn arm ,
.Tn i386 ,
and
.Tn powerpc
architectures.
.Sh RETURN VALUES
The
.Fn atomic_cmpset
function returns the result of the compare operation.
The
.Fn atomic_fcmpset
function returns
.Dv true
if the operation succeeded.
Otherwise it returns
.Dv false
and sets
.Va *old
to the found value.
The
.Fn atomic_fetchadd ,
.Fn atomic_load ,
.Fn atomic_readandclear ,
and
.Fn atomic_swap
functions return the value at the specified address.
The
.Fn atomic_testandset
and
.Fn atomic_testandclear
function returns the result of the test operation.
.Sh EXAMPLES
This example uses the
.Fn atomic_cmpset_acq_ptr
and
.Fn atomic_set_ptr
functions to obtain a sleep mutex and handle recursion.
Since the
.Va mtx_lock
member of a
.Vt "struct mtx"
is a pointer, the
.Dq Li ptr
type is used.
.Bd -literal
/* Try to obtain mtx_lock once. */
#define _obtain_lock(mp, tid)                                           \\
        atomic_cmpset_acq_ptr(&(mp)->mtx_lock, MTX_UNOWNED, (tid))

/* Get a sleep lock, deal with recursion inline. */
#define _get_sleep_lock(mp, tid, opts, file, line) do {                 \\
        uintptr_t _tid = (uintptr_t)(tid);                              \\
                                                                        \\
        if (!_obtain_lock(mp, tid)) {                                   \\
                if (((mp)->mtx_lock & MTX_FLAGMASK) != _tid)            \\
                        _mtx_lock_sleep((mp), _tid, (opts), (file), (line));\\
                else {                                                  \\
                        atomic_set_ptr(&(mp)->mtx_lock, MTX_RECURSE);   \\
                        (mp)->mtx_recurse++;                            \\
                }                                                       \\
        }                                                               \\
} while (0)
.Ed
.Sh HISTORY
The
.Fn atomic_add ,
.Fn atomic_clear ,
.Fn atomic_set ,
and
.Fn atomic_subtract
operations were first introduced in
.Fx 3.0 .
This first set only supported the types
.Dq Li char ,
.Dq Li short ,
.Dq Li int ,
and
.Dq Li long .
The
.Fn atomic_cmpset ,
.Fn atomic_load ,
.Fn atomic_readandclear ,
and
.Fn atomic_store
operations were added in
.Fx 5.0 .
The types
.Dq Li 8 ,
.Dq Li 16 ,
.Dq Li 32 ,
.Dq Li 64 ,
and
.Dq Li ptr
and all of the acquire and release variants
were added in
.Fx 5.0
as well.
The
.Fn atomic_fetchadd
operations were added in
.Fx 6.0 .
The
.Fn atomic_swap
and
.Fn atomic_testandset
operations were added in
.Fx 10.0 .
.Fn atomic_testandclear
operation was added in
.Fx 11.0 .