A thread of execution (also known as a thread) is a single flow of control within a program, including the initial invocation of a specific top-level function, and recursively including every function invocation subsequently executed by the thread. [ Note: When one thread creates another, the initial call to the top-level function of the new thread is executed by the new thread, not by the creating thread. — end note ] Every thread in a program can potentially access every object and function in a program.11 Under a hosted implementation, a C++ program can have more than one thread running concurrently. The execution of each thread proceeds as defined by the remainder of this International Standard. The execution of the entire program consists of an execution of all of its threads. [ Note: Usually the execution can be viewed as an interleaving of all its threads. However, some kinds of atomic operations, for example, allow executions inconsistent with a simple interleaving, as described below. — end note ] Under a freestanding implementation, it is implementation-defined whether a program can have more than one thread of execution.
For a signal handler that is not executed as a result of a call to the std::raise function, it is unspecified which thread of execution contains the signal handler invocation.
The value of an object visible to a thread T at a particular point is the initial value of the object, a value assigned to the object by T, or a value assigned to the object by another thread, according to the rules below. [ Note: In some cases, there may instead be undefined behavior. Much of this section is motivated by the desire to support atomic operations with explicit and detailed visibility constraints. However, it also implicitly supports a simpler view for more restricted programs. — end note ]
The library defines a number of atomic operations and operations on mutexes that are specially identified as synchronization operations. These operations play a special role in making assignments in one thread visible to another. A synchronization operation on one or more memory locations is either a consume operation, an acquire operation, a release operation, or both an acquire and release operation. A synchronization operation without an associated memory location is a fence and can be either an acquire fence, a release fence, or both an acquire and release fence. In addition, there are relaxed atomic operations, which are not synchronization operations, and atomic read-modify-write operations, which have special characteristics. [ Note: For example, a call that acquires a mutex will perform an acquire operation on the locations comprising the mutex. Correspondingly, a call that releases the same mutex will perform a release operation on those same locations. Informally, performing a release operation on A forces prior side effects on other memory locations to become visible to other threads that later perform a consume or an acquire operation on A. “Relaxed” atomic operations are not synchronization operations even though, like synchronization operations, they cannot contribute to data races. — end note ]
All modifications to a particular atomic object M occur in some particular total order, called the modification order of M. [ Note: There is a separate order for each atomic object. There is no requirement that these can be combined into a single total order for all objects. In general this will be impossible since different threads may observe modifications to different objects in inconsistent orders. — end note ]
A release sequence headed by a release operation A on an atomic object M is a maximal contiguous sub-sequence of side effects in the modification order of M, where the first operation is A, and every subsequent operation
is performed by the same thread that performed A, or
is an atomic read-modify-write operation.
Certain library calls synchronize with other library calls performed by another thread. For example, an atomic store-release synchronizes with a load-acquire that takes its value from the store ([atomics.order]). [ Note: Except in the specified cases, reading a later value does not necessarily ensure visibility as described below. Such a requirement would sometimes interfere with efficient implementation. — end note ] [ Note: The specifications of the synchronization operations define when one reads the value written by another. For atomic objects, the definition is clear. All operations on a given mutex occur in a single total order. Each mutex acquisition “reads the value written” by the last mutex release. — end note ]
An evaluation A carries a dependency to an evaluation B if
the value of A is used as an operand of B, unless:
B is an invocation of any specialization of std::kill_dependency, or
A is the left operand of a conditional (?:, see [expr.cond]) operator, or
A is the left operand of the built-in comma (,) operator;
A writes a scalar object or bit-field M, B reads the value written by A from M, and A is sequenced before B, or
for some evaluation X, A carries a dependency to X, and X carries a dependency to B.
[ Note: “Carries a dependency to” is a subset of “is sequenced before”, and is similarly strictly intra-thread. — end note ]
An evaluation A is dependency-ordered before an evaluation B if
for some evaluation X, A is dependency-ordered before X and X carries a dependency to B.
[ Note: The relation “is dependency-ordered before” is analogous to “synchronizes with”, but uses release/consume in place of release/acquire. — end note ]
An evaluation A inter-thread happens before an evaluation B if
A synchronizes with B, or
A is dependency-ordered before B, or
for some evaluation X
A synchronizes with X and X is sequenced before B, or
A is sequenced before X and X inter-thread happens before B, or
A inter-thread happens before X and X inter-thread happens before B.
[ Note: The “inter-thread happens before” relation describes arbitrary concatenations of “sequenced before”, “synchronizes with” and “dependency-ordered before” relationships, with two exceptions. The first exception is that a concatenation is not permitted to end with “dependency-ordered before” followed by “sequenced before”. The reason for this limitation is that a consume operation participating in a “dependency-ordered before” relationship provides ordering only with respect to operations to which this consume operation actually carries a dependency. The reason that this limitation applies only to the end of such a concatenation is that any subsequent release operation will provide the required ordering for a prior consume operation. The second exception is that a concatenation is not permitted to consist entirely of “sequenced before”. The reasons for this limitation are (1) to permit “inter-thread happens before” to be transitively closed and (2) the “happens before” relation, defined below, provides for relationships consisting entirely of “sequenced before”. — end note ]
A is sequenced before B, or
A inter-thread happens before B.
The implementation shall ensure that no program execution demonstrates a cycle in the “happens before” relation. [ Note: This cycle would otherwise be possible only through the use of consume operations. — end note ]
An evaluation A strongly happens before an evaluation B if either
A is sequenced before B, or
A synchronizes with B, or
A strongly happens before X and X strongly happens before B.
[ Note: In the absence of consume operations, the happens before and strongly happens before relations are identical. Strongly happens before essentially excludes consume operations. — end note ]
A visible side effect A on a scalar object or bit-field M with respect to a value computation B of M satisfies the conditions:
A happens before B and
The value of a non-atomic scalar object or bit-field M, as determined by evaluation B, shall be the value stored by the visible side effect A. [ Note: If there is ambiguity about which side effect to a non-atomic object or bit-field is visible, then the behavior is either unspecified or undefined. — end note ] [ Note: This states that operations on ordinary objects are not visibly reordered. This is not actually detectable without data races, but it is necessary to ensure that data races, as defined below, and with suitable restrictions on the use of atomics, correspond to data races in a simple interleaved (sequentially consistent) execution. — end note ]
The value of an atomic object M, as determined by evaluation B, shall be the value stored by some side effect A that modifies M, where B does not happen before A. [ Note: The set of such side effects is also restricted by the rest of the rules described here, and in particular, by the coherence requirements below. — end note ]
If an operation A that modifies an atomic object M happens before an operation B that modifies M, then A shall be earlier than B in the modification order of M. [ Note: This requirement is known as write-write coherence. — end note ]
If a value computation A of an atomic object M happens before a value computation B of M, and A takes its value from a side effect X on M, then the value computed by B shall either be the value stored by X or the value stored by a side effect Y on M, where Y follows X in the modification order of M. [ Note: This requirement is known as read-read coherence. — end note ]
If a value computation A of an atomic object M happens before an operation B that modifies M, then A shall take its value from a side effect X on M, where X precedes B in the modification order of M. [ Note: This requirement is known as read-write coherence. — end note ]
If a side effect X on an atomic object M happens before a value computation B of M, then the evaluation B shall take its value from X or from a side effect Y that follows X in the modification order of M. [ Note: This requirement is known as write-read coherence. — end note ]
[ Note: The four preceding coherence requirements effectively disallow compiler reordering of atomic operations to a single object, even if both operations are relaxed loads. This effectively makes the cache coherence guarantee provided by most hardware available to C++ atomic operations. — end note ]
[ Note: The value observed by a load of an atomic depends on the “happens before” relation, which depends on the values observed by loads of atomics. The intended reading is that there must exist an association of atomic loads with modifications they observe that, together with suitably chosen modification orders and the “happens before” relation derived as described above, satisfy the resulting constraints as imposed here. — end note ]
Two actions are potentially concurrent if
they are performed by different threads, or
they are unsequenced, at least one is performed by a signal handler, and they are not both performed by the same signal handler invocation.
The execution of a program contains a data race if it contains two potentially concurrent conflicting actions, at least one of which is not atomic, and neither happens before the other, except for the special case for signal handlers described below. Any such data race results in undefined behavior. [ Note: It can be shown that programs that correctly use mutexes and memory_order_seq_cst operations to prevent all data races and use no other synchronization operations behave as if the operations executed by their constituent threads were simply interleaved, with each value computation of an object being taken from the last side effect on that object in that interleaving. This is normally referred to as “sequential consistency”. However, this applies only to data-race-free programs, and data-race-free programs cannot observe most program transformations that do not change single-threaded program semantics. In fact, most single-threaded program transformations continue to be allowed, since any program that behaves differently as a result must perform an undefined operation. — end note ]
Two accesses to the same object of type volatile std::sig_atomic_t do not result in a data race if both occur in the same thread, even if one or more occurs in a signal handler. For each signal handler invocation, evaluations performed by the thread invoking a signal handler can be divided into two groups A and B, such that no evaluations in B happen before evaluations in A, and the evaluations of such volatile std::sig_atomic_t objects take values as though all evaluations in A happened before the execution of the signal handler and the execution of the signal handler happened before all evaluations in B.
[ Note: Compiler transformations that introduce assignments to a potentially shared memory location that would not be modified by the abstract machine are generally precluded by this International Standard, since such an assignment might overwrite another assignment by a different thread in cases in which an abstract machine execution would not have encountered a data race. This includes implementations of data member assignment that overwrite adjacent members in separate memory locations. Reordering of atomic loads in cases in which the atomics in question may alias is also generally precluded, since this may violate the coherence rules. — end note ]
[ Note: Transformations that introduce a speculative read of a potentially shared memory location may not preserve the semantics of the C++ program as defined in this International Standard, since they potentially introduce a data race. However, they are typically valid in the context of an optimizing compiler that targets a specific machine with well-defined semantics for data races. They would be invalid for a hypothetical machine that is not tolerant of races or provides hardware race detection. — end note ]
The implementation may assume that any thread will eventually do one of the following:
make a call to a library I/O function,
perform an access through a volatile glvalue, or
perform a synchronization operation or an atomic operation.
[ Note: This is intended to allow compiler transformations such as removal of empty loops, even when termination cannot be proven. — end note ]
If there is only one thread that is not blocked in a standard library function, a lock-free execution in that thread shall complete. [ Note: Concurrently executing threads may prevent progress of a lock-free execution. For example, this situation can occur with load-locked store-conditional implementations. This property is sometimes termed obstruction-free. — end note ]
When one or more lock-free executions run concurrently, at least one should complete. [ Note: It is difficult for some implementations to provide absolute guarantees to this effect, since repeated and particularly inopportune interference from other threads may prevent forward progress, e.g., by repeatedly stealing a cache line for unrelated purposes between load-locked and store-conditional instructions. Implementations should ensure that such effects cannot indefinitely delay progress under expected operating conditions, and that such anomalies can therefore safely be ignored by programmers. Outside this document, this property is sometimes termed lock-free. — end note ]
During the execution of a thread of execution, each of the following is termed an execution step:
termination of the thread of execution,
performing an access through a volatile glvalue, or
completion of a call to a library I/O function, a synchronization operation, or an atomic operation.
An invocation of a standard library function that blocks is considered to continuously execute execution steps while waiting for the condition that it blocks on to be satisfied. [ Example: A library I/O function that blocks until the I/O operation is complete can be considered to continuously check whether the operation is complete. Each such check might consist of one or more execution steps, for example using observable behavior of the abstract machine. — end example ]
[ Note: Because of this and the preceding requirement regarding what threads of execution have to perform eventually, it follows that no thread of execution can execute forever without an execution step occurring. — end note ]
A thread of execution makes progress when an execution step occurs or a lock-free execution does not complete because there are other concurrent threads that are not blocked in a standard library function (see above).
For a thread of execution providing concurrent forward progress guarantees, the implementation ensures that the thread will eventually make progress for as long as it has not terminated. [ Note: This is required regardless of whether or not other threads of executions (if any) have been or are making progress. To eventually fulfill this requirement means that this will happen in an unspecified but finite amount of time. — end note ]
It is implementation-defined whether the implementation-created thread of execution that executes main and the threads of execution created by std::thread provide concurrent forward progress guarantees. [ Note: General-purpose implementations are encouraged to provide these guarantees. — end note ]
For a thread of execution providing parallel forward progress guarantees, the implementation is not required to ensure that the thread will eventually make progress if it has not yet executed any execution step; once this thread has executed a step, it provides concurrent forward progress guarantees.
[ Note: This does not specify a requirement for when to start this thread of execution, which will typically be specified by the entity that creates this thread of execution. For example, a thread of execution that provides concurrent forward progress guarantees and executes tasks from a set of tasks in an arbitrary order, one after the other, satisfies the requirements of parallel forward progress for these tasks. — end note ]
For a thread of execution providing weakly parallel forward progress guarantees, the implementation does not ensure that the thread will eventually make progress.
[ Note: Threads of execution providing weakly parallel forward progress guarantees cannot be expected to make progress regardless of whether other threads make progress or not; however, blocking with forward progress guarantee delegation, as defined below, can be used to ensure that such threads of execution make progress eventually. — end note ]
Concurrent forward progress guarantees are stronger than parallel forward progress guarantees, which in turn are stronger than weakly parallel forward progress guarantees. [ Note: For example, some kinds of synchronization between threads of execution may only make progress if the respective threads of execution provide parallel forward progress guarantees, but will fail to make progress under weakly parallel guarantees. — end note ]
When a thread of execution P is specified to block with forward progress guarantee delegation on the completion of a set S of threads of execution, then throughout the whole time of P being blocked on S, the implementation shall ensure that the forward progress guarantees provided by at least one thread of execution in S is at least as strong as P's forward progress guarantees. [ Note: It is unspecified which thread or threads of execution in S are chosen and for which number of execution steps. The strengthening is not permanent and not necessarily in place for the rest of the lifetime of the affected thread of execution. As long as P is blocked, the implementation has to eventually select and potentially strengthen a thread of execution in S. — end note ] Once a thread of execution in S terminates, it is removed from S. Once S is empty, P is unblocked.
[ Note: A thread of execution B thus can temporarily provide an effectively stronger forward progress guarantee for a certain amount of time, due to a second thread of execution A being blocked on it with forward progress guarantee delegation. In turn, if B then blocks with forward progress guarantee delegation on C, this may also temporarily provide a stronger forward progress guarantee to C. — end note ]
[ Note: If all threads of execution in S finish executing (e.g., they terminate and do not use blocking synchronization incorrectly), then P's execution of the operation that blocks with forward progress guarantee delegation will not result in P's progress guarantee being effectively weakened. — end note ]
[ Note: This does not remove any constraints regarding blocking synchronization for threads of execution providing parallel or weakly parallel forward progress guarantees because the implementation is not required to strengthen a particular thread of execution whose too-weak progress guarantee is preventing overall progress. — end note ]