SYNOPSIS

Fd #include <dispatch/dispatch.h> Ft dispatch_queue_t Fo dispatch_queue_create Fa const char *label dispatch_queue_attr_t attr Fc Ft const char * Fo dispatch_queue_get_label Fa dispatch_queue_t queue Fc Ft dispatch_queue_t Fo dispatch_get_current_queue Fa void Fc Ft dispatch_queue_t Fo dispatch_get_global_queue Fa long priority Fa unsigned long flags Fc Ft dispatch_queue_t Fo dispatch_get_main_queue Fa void Fc Ft void Fo dispatch_main Fa void Fc Ft void Fo dispatch_set_target_queue Fa dispatch_object_t object Fa dispatch_queue_t target Fc

DESCRIPTION

Queues are the fundamental mechanism for scheduling blocks for execution within the dispatch(3) framework.

All blocks submitted to dispatch queues are dequeued in FIFO order. By default, queues created with Fn dispatch_queue_create wait for the previously dequeued block to complete before dequeuing the next block. This FIFO completion behavior is sometimes simply described as a "serial queue." Queues are not bound to any specific thread of execution and blocks submitted to independent queues may execute concurrently. Queues, like all dispatch objects, are reference counted and newly created queues have a reference count of one.

The optional Fa label argument is used to describe the purpose of the queue and is useful during debugging and performance analysis. By convention, clients should pass a reverse DNS style label. If a label is provided, it is copied. If a label is not provided, then Fn dispatch_queue_get_label returns an empty C string. For example:

my_queue = dispatch_queue_create("com.example.subsystem.taskXYZ", NULL);

The Fa attr argument is reserved for future use and must be NULL.

Queues may be temporarily suspended and resumed with the functions Fn dispatch_suspend and Fn dispatch_resume respectively. Suspension is checked prior to block execution and is not preemptive.

MAIN QUEUE

The dispatch framework provides a default serial queue for the application to use. This queue is accessed via Fn dispatch_get_main_queue . Programs must call Fn dispatch_main at the end of Fn main in order to process blocks submitted to the main queue. (See the compatibility section for exceptions.)

GLOBAL CONCURRENT QUEUES

Unlike the main queue or queues allocated with Fn dispatch_queue_create , the global concurrent queues schedule blocks as soon as threads become available (non-FIFO completion order). The global concurrent queues represent three priority bands:

• DISPATCH_QUEUE_PRIORITY_HIGH
• DISPATCH_QUEUE_PRIORITY_DEFAULT
• DISPATCH_QUEUE_PRIORITY_LOW

Blocks submitted to the high priority global queue will be invoked before those submitted to the default or low priority global queues. Blocks submitted to the low priority global queue will only be invoked if no blocks are pending on the default or high priority queues.

RETURN VALUES

The Fn dispatch_queue_create function returns NULL on failure.

The Fn dispatch_queue_get_label function always returns a valid C string. An empty C string is returned if the Fa label was NULL creation time.

The Fn dispatch_get_main_queue function returns the default main queue.

The Fn dispatch_get_current_queue function always returns a valid queue. When called from within a block submitted to a dispatch queue, that queue will be returned. If this function is called from the main thread before Fn dispatch_main is called, then the result of Fn dispatch_get_main_queue is returned. Otherwise, the result of Fo dispatch_get_global_queue Fa DISPATCH_QUEUE_PRIORITY_DEFAULT Fa 0 Fc will be returned in all other cases.

The Fn dispatch_main function never returns.

TARGET QUEUE

The Fn dispatch_set_target_queue function updates the target queue of the given dispatch object. The target queue of an object is responsible for processing the object. Currently only dispatch queues and dispatch sources are supported by this function. The result of using Fn dispatch_set_target_queue with any other dispatch object type is undefined.

The new target queue is retained by the given object before the previous target queue is released. The new target queue will take effect between block executions, but not in the middle of any existing block executions (non-preemptive).

The priority of a dispatch queue is inherited by its target queue. In order to change the priority of a queue created with Fn dispatch_queue_create , use the Fn dispatch_get_global_queue function to obtain a target queue of the desired priority. The Fa flags argument is reserved for future use and must be zero. Passing any value other than zero may result in a Vt NULL return value.

The target queue of a dispatch source specifies where its event handler and cancellation handler blocks will be submitted. See dispatch_source_create3 for more information about dispatch sources.

The result of passing the main queue or a global concurrent queue to the first argument of Fn dispatch_set_target_queue is undefined.

Directly or indirectly setting the target queue of a dispatch queue to itself is undefined.

CAVEATS

Code cannot make any assumptions about the queue returned by Fn dispatch_get_current_queue . The returned queue may have arbitrary policies that may surprise code that tries to schedule work with the queue. The list of policies includes, but is not limited to, queue width (i.e. serial vs. concurrent), scheduling priority, security credential or filesystem configuration. Therefore, Fn dispatch_get_current_queue MUST only be used for identity tests or debugging.

COMPATIBILITY

Cocoa applications need not call Fn dispatch_main . Blocks submitted to the main queue will be executed as part of the "common modes" of the application's main NSRunLoop or CFRunLoop. However, blocks submitted to the main queue in applications using Fn dispatch_main are not guaranteed to execute on the main thread.

The dispatch framework is a pure C level API. As a result, it does not catch exceptions generated by higher level languages such as Objective-C or C++. Applications MUST catch all exceptions before returning from a block submitted to a dispatch queue; otherwise the internal data structures of the dispatch framework will be left in an inconsistent state.

The dispatch framework manages the relationship between dispatch queues and threads of execution. As a result, applications MUST NOT delete or mutate objects that they did not create. The following interfaces MUST NOT be called by blocks submitted to a dispatch queue:

• Fn pthread_cancel
• Fn pthread_detach
• Fn pthread_join
• Fn pthread_kill
• Fn pthread_exit

Applications MAY call the following interfaces from a block submitted to a dispatch queue if and only if they restore the thread to its original state before returning:

• Fn pthread_setcancelstate
• Fn pthread_setcanceltype
• Fn pthread_setschedparam
• Fn pthread_sigmask
• Fn pthread_setugid_np
• Fn pthread_chdir
• Fn pthread_fchdir

Applications MUST NOT rely on the following interfaces returning predictable results between invocations of blocks submitted to a dispatch queue:

• Fn pthread_self
• Fn pthread_getschedparam
• Fn pthread_get_stacksize_np
• Fn pthread_get_stackaddr_np
• Fn pthread_mach_thread_np
• Fn pthread_from_mach_thread_np

While the result of Fn pthread_self may change between invocations of blocks, the value will not change during the execution of any single block. Because the underlying thread may change beteween block invocations on a single queue, using per-thread data as an out-of-band return value is error prone. In other words, the result of calling Fn pthread_setspecific and Fn pthread_getspecific is well defined within a signle block, but not across multiple blocks. Also, one cannot make any assumptions about when the destructor passed to Fn pthread_key_create is called. The destructor may be called between the invocation of blocks on the same queue, or during the idle state of a process.

The following example code correctly handles per-thread return values:

__block int r;
__block int e;
dispatch_sync(queue, ^{
        r = kill(1, 0);
        // Copy the per-thread return value to the callee thread
        e = errno;
});
printf("kill(1,0) returned %d and errno %d, r, e);

Note that in the above example errno is a per-thread variable and must be copied out explicitly as the block may be invoked on different thread of execution than the caller. Another example of per-thread data that would need to be copied is the use of Fn getpwnam instead of Fn getpwnam_r .

As an optimization, Fn dispatch_sync invokes the block on the current thread when possible. In this case, the thread specific data such as errno may persist from the block until back to the caller. Great care should be taken not to accidentally rely on this side-effect.