TASKQUEUE(9FREEBSD) - man page online | system kernel interfaces

TASKQUEUE(9)                      BSD Kernel Developer's Manual                      TASKQUEUE(9)

NAME
     taskqueue — asynchronous task execution

SYNOPSIS
     #include <sys/param.h>
     #include <sys/kernel.h>
     #include <sys/malloc.h>
     #include <sys/queue.h>
     #include <sys/taskqueue.h>

     typedef void (*task_fn_t)(void *context, int pending);

     typedef void (*taskqueue_enqueue_fn)(void *context);

     struct task {
             STAILQ_ENTRY(task)      ta_link;        /* link for queue */
             u_short                 ta_pending;     /* count times queued */
             u_short                 ta_priority;    /* priority of task in queue */
             task_fn_t               ta_func;        /* task handler */
             void                    *ta_context;    /* argument for handler */
     };

     enum taskqueue_callback_type {
             TASKQUEUE_CALLBACK_TYPE_INIT,
             TASKQUEUE_CALLBACK_TYPE_SHUTDOWN,
     };

     typedef void (*taskqueue_callback_fn)(void *context);

     struct timeout_task;

     struct taskqueue *
     taskqueue_create(const char *name, int mflags, taskqueue_enqueue_fn enqueue, void *context);

     struct taskqueue *
     taskqueue_create_fast(const char *name, int mflags, taskqueue_enqueue_fn enqueue,
         void *context);

     int
     taskqueue_start_threads(struct taskqueue **tqp, int count, int pri, const char *name, ...);

     void
     taskqueue_set_callback(struct taskqueue *queue, enum taskqueue_callback_type cb_type,
         taskqueue_callback_fn callback, void *context);

     void
     taskqueue_free(struct taskqueue *queue);

     int
     taskqueue_enqueue(struct taskqueue *queue, struct task *task);

     int
     taskqueue_enqueue_fast(struct taskqueue *queue, struct task *task);

     int
     taskqueue_enqueue_timeout(struct taskqueue *queue, struct timeout_task *timeout_task,
         int ticks);

     int
     taskqueue_cancel(struct taskqueue *queue, struct task *task, u_int *pendp);

     int
     taskqueue_cancel_timeout(struct taskqueue *queue, struct timeout_task *timeout_task,
         u_int *pendp);

     void
     taskqueue_drain(struct taskqueue *queue, struct task *task);

     void
     taskqueue_drain_timeout(struct taskqueue *queue, struct timeout_task *timeout_task);

     void
     taskqueue_drain_all(struct taskqueue *queue);

     void
     taskqueue_block(struct taskqueue *queue);

     void
     taskqueue_unblock(struct taskqueue *queue);

     int
     taskqueue_member(struct taskqueue *queue, struct thread *td);

     void
     taskqueue_run(struct taskqueue *queue);

     TASK_INIT(struct task *task, int priority, task_fn_t func, void *context);

     TASK_INITIALIZER(int priority, task_fn_t func, void *context);

     TASKQUEUE_DECLARE(name);

     TASKQUEUE_DEFINE(name, taskqueue_enqueue_fn enqueue, void *context, init);

     TASKQUEUE_FAST_DEFINE(name, taskqueue_enqueue_fn enqueue, void *context, init);

     TASKQUEUE_DEFINE_THREAD(name);

     TASKQUEUE_FAST_DEFINE_THREAD(name);

     TIMEOUT_TASK_INIT(struct taskqueue *queue, struct timeout_task *timeout_task, int priority,
         task_fn_t func, void *context);

DESCRIPTION
     These functions provide a simple interface for asynchronous execution of code.

     The function taskqueue_create() is used to create new queues.  The arguments to
     taskqueue_create() include a name that should be unique, a set of malloc(9) flags that spec‐
     ify whether the call to malloc() is allowed to sleep, a function that is called from
     taskqueue_enqueue() when a task is added to the queue, and a pointer to the memory location
     where the identity of the thread that services the queue is recorded.  The function called
     from taskqueue_enqueue() must arrange for the queue to be processed (for instance by sched‐
     uling a software interrupt or waking a kernel thread).  The memory location where the thread
     identity is recorded is used to signal the service thread(s) to terminate--when this value
     is set to zero and the thread is signaled it will terminate.  If the queue is intended for
     use in fast interrupt handlers taskqueue_create_fast() should be used in place of
     taskqueue_create().

     The function taskqueue_free() should be used to free the memory used by the queue.  Any
     tasks that are on the queue will be executed at this time after which the thread servicing
     the queue will be signaled that it should exit.

     Once a taskqueue has been created, its threads should be started using
     taskqueue_start_threads().  Callbacks may optionally be registered using
     taskqueue_set_callback().  Currently, callbacks may be registered for the following pur‐
     poses:

     TASKQUEUE_CALLBACK_TYPE_INIT      This callback is called by every thread in the taskqueue,
                                       before it executes any tasks.  This callback must be set
                                       before the taskqueue's threads are started.

     TASKQUEUE_CALLBACK_TYPE_SHUTDOWN  This callback is called by every thread in the taskqueue,
                                       after it executes its last task.  This callback will
                                       always be called before the taskqueue structure is
                                       reclaimed.

     To add a task to the list of tasks queued on a taskqueue, call taskqueue_enqueue() with
     pointers to the queue and task.  If the task's ta_pending field is non-zero, then it is sim‐
     ply incremented to reflect the number of times the task was enqueued, up to a cap of
     USHRT_MAX.  Otherwise, the task is added to the list before the first task which has a lower
     ta_priority value or at the end of the list if no tasks have a lower priority.  Enqueueing a
     task does not perform any memory allocation which makes it suitable for calling from an
     interrupt handler.  This function will return EPIPE if the queue is being freed.

     The function taskqueue_enqueue_fast() should be used in place of taskqueue_enqueue() when
     the enqueuing must happen from a fast interrupt handler.  This method uses spin locks to
     avoid the possibility of sleeping in the fast interrupt context.

     When a task is executed, first it is removed from the queue, the value of ta_pending is
     recorded and then the field is zeroed.  The function ta_func from the task structure is
     called with the value of the field ta_context as its first argument and the value of
     ta_pending as its second argument.  After the function ta_func returns, wakeup(9) is called
     on the task pointer passed to taskqueue_enqueue().

     The taskqueue_enqueue_timeout() is used to schedule the enqueue after the specified amount
     of ticks.  Only non-fast task queues can be used for timeout_task scheduling.  If the ticks
     argument is negative, the already scheduled enqueueing is not re-scheduled.  Otherwise, the
     task is scheduled for enqueueing in the future, after the absolute value of ticks is passed.

     The taskqueue_cancel() function is used to cancel a task.  The ta_pending count is cleared,
     and the old value returned in the reference parameter pendp, if it is non-NULL.  If the task
     is currently running, EBUSY is returned, otherwise 0.  To implement a blocking
     taskqueue_cancel() that waits for a running task to finish, it could look like:

           while (taskqueue_cancel(tq, task, NULL) != 0)
                   taskqueue_drain(tq, task);

     Note that, as with taskqueue_drain(), the caller is responsible for ensuring that the task
     is not re-enqueued after being canceled.

     Similarly, the taskqueue_cancel_timeout() function is used to cancel the scheduled task exe‐
     cution.

     The taskqueue_drain() function is used to wait for the task to finish, and the
     taskqueue_drain_timeout() function is used to wait for the scheduled task to finish.  There
     is no guarantee that the task will not be enqueued after call to taskqueue_drain().  If the
     caller wants to put the task into a known state, then before calling taskqueue_drain() the
     caller should use out-of-band means to ensure that the task would not be enqueued.  For
     example, if the task is enqueued by an interrupt filter, then the interrupt could be dis‐
     abled.

     The taskqueue_drain_all() function is used to wait for all pending and running tasks that
     are enqueued on the taskqueue to finish.  The caller must arrange that the tasks are not re-
     enqueued.  Note that taskqueue_drain_all() currently does not handle tasks with delayed
     enqueueing.

     The taskqueue_block() function blocks the taskqueue.  It prevents any enqueued but not run‐
     ning tasks from being executed.  Future calls to taskqueue_enqueue() will enqueue tasks, but
     the tasks will not be run until taskqueue_unblock() is called.  Please note that
     taskqueue_block() does not wait for any currently running tasks to finish.  Thus, the
     taskqueue_block() does not provide a guarantee that taskqueue_run() is not running after
     taskqueue_block() returns, but it does provide a guarantee that taskqueue_run() will not be
     called again until taskqueue_unblock() is called.  If the caller requires a guarantee that
     taskqueue_run() is not running, then this must be arranged by the caller.  Note that if
     taskqueue_drain() is called on a task that is enqueued on a taskqueue that is blocked by
     taskqueue_block(), then taskqueue_drain() can not return until the taskqueue is unblocked.
     This can result in a deadlock if the thread blocked in taskqueue_drain() is the thread that
     is supposed to call taskqueue_unblock().  Thus, use of taskqueue_drain() after
     taskqueue_block() is discouraged, because the state of the task can not be known in advance.
     The same caveat applies to taskqueue_drain_all().

     The taskqueue_unblock() function unblocks the previously blocked taskqueue.  All enqueued
     tasks can be run after this call.

     The taskqueue_member() function returns 1 if the given thread td is part of the given
     taskqueue queue and 0 otherwise.

     The taskqueue_run() function will run all pending tasks in the specified queue.  Normally
     this function is only used internally.

     A convenience macro, TASK_INIT(task, priority, func, context) is provided to initialise a
     task structure.  The TASK_INITIALIZER() macro generates an initializer for a task structure.
     A macro TIMEOUT_TASK_INIT(queue, timeout_task, priority, func, context) initializes the
     timeout_task structure.  The values of priority, func, and context are simply copied into
     the task structure fields and the ta_pending field is cleared.

     Five macros TASKQUEUE_DECLARE(name), TASKQUEUE_DEFINE(name, enqueue, context, init),
     TASKQUEUE_FAST_DEFINE(name, enqueue, context, init), and TASKQUEUE_DEFINE_THREAD(name)
     TASKQUEUE_FAST_DEFINE_THREAD(name) are used to declare a reference to a global queue, to
     define the implementation of the queue, and declare a queue that uses its own thread.  The
     TASKQUEUE_DEFINE() macro arranges to call taskqueue_create() with the values of its name,
     enqueue and context arguments during system initialisation.  After calling
     taskqueue_create(), the init argument to the macro is executed as a C statement, allowing
     any further initialisation to be performed (such as registering an interrupt handler etc.)

     The TASKQUEUE_DEFINE_THREAD() macro defines a new taskqueue with its own kernel thread to
     serve tasks.  The variable struct taskqueue *taskqueue_name is used to enqueue tasks onto
     the queue.

     TASKQUEUE_FAST_DEFINE() and TASKQUEUE_FAST_DEFINE_THREAD() act just like TASKQUEUE_DEFINE()
     and TASKQUEUE_DEFINE_THREAD() respectively but taskqueue is created with
     taskqueue_create_fast().

   Predefined Task Queues
     The system provides four global taskqueues, taskqueue_fast, taskqueue_swi,
     taskqueue_swi_giant, and taskqueue_thread.  The taskqueue_fast queue is for swi handlers
     dispatched from fast interrupt handlers, where sleep mutexes cannot be used.  The swi
     taskqueues are run via a software interrupt mechanism.  The taskqueue_swi queue runs without
     the protection of the Giant kernel lock, and the taskqueue_swi_giant queue runs with the
     protection of the Giant kernel lock.  The thread taskqueue taskqueue_thread runs in a kernel
     thread context, and tasks run from this thread do not run under the Giant kernel lock.  If
     the caller wants to run under Giant, he should explicitly acquire and release Giant in his
     taskqueue handler routine.

     To use these queues, call taskqueue_enqueue() with the value of the global taskqueue vari‐
     able for the queue you wish to use (taskqueue_swi, taskqueue_swi_giant, or
     taskqueue_thread).  Use taskqueue_enqueue_fast() for the global taskqueue variable
     taskqueue_fast.

     The software interrupt queues can be used, for instance, for implementing interrupt handlers
     which must perform a significant amount of processing in the handler.  The hardware inter‐
     rupt handler would perform minimal processing of the interrupt and then enqueue a task to
     finish the work.  This reduces to a minimum the amount of time spent with interrupts dis‐
     abled.

     The thread queue can be used, for instance, by interrupt level routines that need to call
     kernel functions that do things that can only be done from a thread context.  (e.g., call
     malloc with the M_WAITOK flag.)

     Note that tasks queued on shared taskqueues such as taskqueue_swi may be delayed an indeter‐
     minate amount of time before execution.  If queueing delays cannot be tolerated then a pri‐
     vate taskqueue should be created with a dedicated processing thread.

SEE ALSO
     ithread(9), kthread(9), swi(9)

HISTORY
     This interface first appeared in FreeBSD 5.0.  There is a similar facility called work_queue
     in the Linux kernel.

AUTHORS
     This manual page was written by Doug Rabson.

BSD                                      January 24, 2014                                     BSD