root/sys/kern/subr_taskqueue.c
/*-
 * Copyright (c) 2000 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: src/sys/kern/subr_taskqueue.c,v 1.69 2012/08/28 13:35:37 jhb Exp $"
 */

#include <sys/param.h>
#include <sys/queue.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/taskqueue.h>
#include <sys/interrupt.h>
#include <sys/lock.h>
#include <sys/malloc.h>
#include <sys/kthread.h>
#include <sys/spinlock.h>
#include <sys/spinlock2.h>
#include <sys/serialize.h>
#include <sys/proc.h>

MALLOC_DEFINE(M_TASKQUEUE, "taskqueue", "Task Queues");

static STAILQ_HEAD(taskqueue_list, taskqueue) taskqueue_queues;
static struct lock      taskqueue_queues_lock;
static struct spinlock  taskqueue_queues_spin;

struct taskqueue {
        STAILQ_ENTRY(taskqueue) tq_link;
        STAILQ_HEAD(, task)     tq_queue;
        const char              *tq_name;
        /* NOTE: tq must be locked before calling tq_enqueue */
        taskqueue_enqueue_fn    tq_enqueue;
        void                    *tq_context;

        struct task             *tq_running;
        struct spinlock         tq_lock;
        struct thread           **tq_threads;
        int                     tq_tcount;
        int                     tq_flags;
        int                     tq_callouts;
};

#define TQ_FLAGS_ACTIVE         (1 << 0)
#define TQ_FLAGS_BLOCKED        (1 << 1)
#define TQ_FLAGS_PENDING        (1 << 2)

#define DT_CALLOUT_ARMED        (1 << 0)

void
_timeout_task_init(struct taskqueue *queue, struct timeout_task *timeout_task,
    int priority, task_fn_t func, void *context)
{

        TASK_INIT(&timeout_task->t, priority, func, context);
        callout_init_mp(&timeout_task->c);
        timeout_task->t.ta_queue = queue;
        timeout_task->f = 0;
}

static void taskqueue_run(struct taskqueue *queue, int lock_held);

static __inline void
TQ_LOCK_INIT(struct taskqueue *tq)
{
        spin_init(&tq->tq_lock, "tqlock");
}

static __inline void
TQ_LOCK_UNINIT(struct taskqueue *tq)
{
        spin_uninit(&tq->tq_lock);
}

static __inline void
TQ_LOCK(struct taskqueue *tq)
{
        spin_lock(&tq->tq_lock);
}

static __inline void
TQ_UNLOCK(struct taskqueue *tq)
{
        spin_unlock(&tq->tq_lock);
}

static __inline void
TQ_SLEEP(struct taskqueue *tq, void *ident, const char *wmesg)
{
        ssleep(ident, &tq->tq_lock, 0, wmesg, 0);
}

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

        queue = kmalloc(sizeof(*queue), M_TASKQUEUE, mflags | M_ZERO);
        if (!queue)
                return NULL;
        STAILQ_INIT(&queue->tq_queue);
        queue->tq_name = name;
        queue->tq_enqueue = enqueue;
        queue->tq_context = context;
        queue->tq_flags |= TQ_FLAGS_ACTIVE;
        TQ_LOCK_INIT(queue);

        lockmgr(&taskqueue_queues_lock, LK_EXCLUSIVE);
        STAILQ_INSERT_TAIL(&taskqueue_queues, queue, tq_link);
        lockmgr(&taskqueue_queues_lock, LK_RELEASE);

        return queue;
}

/* NOTE: tq must be locked */
static void
taskqueue_terminate(struct thread **pp, struct taskqueue *tq)
{
        while (tq->tq_tcount > 0) {
                /* Unlock spinlock before wakeup() */
                TQ_UNLOCK(tq);
                wakeup(tq);
                TQ_LOCK(tq);
                TQ_SLEEP(tq, pp, "taskqueue_terminate");
        }
}

void
taskqueue_free(struct taskqueue *queue)
{
        TQ_LOCK(queue);
        queue->tq_flags &= ~TQ_FLAGS_ACTIVE;
        taskqueue_run(queue, 1);
        taskqueue_terminate(queue->tq_threads, queue);
        TQ_UNLOCK(queue);

        lockmgr(&taskqueue_queues_lock, LK_EXCLUSIVE);
        STAILQ_REMOVE(&taskqueue_queues, queue, taskqueue, tq_link);
        lockmgr(&taskqueue_queues_lock, LK_RELEASE);

        TQ_LOCK_UNINIT(queue);

        kfree(queue, M_TASKQUEUE);
}

struct taskqueue *
taskqueue_find(const char *name)
{
        struct taskqueue *queue;

        lockmgr(&taskqueue_queues_lock, LK_EXCLUSIVE);
        STAILQ_FOREACH(queue, &taskqueue_queues, tq_link) {
                if (!strcmp(queue->tq_name, name)) {
                        lockmgr(&taskqueue_queues_lock, LK_RELEASE);
                        return queue;
                }
        }
        lockmgr(&taskqueue_queues_lock, LK_RELEASE);
        return NULL;
}

/*
 * NOTE!  If using the per-cpu taskqueues ``taskqueue_thread[mycpuid]'',
 * be sure NOT TO SHARE the ``task'' between CPUs.  TASKS ARE NOT LOCKED.
 * So either use a throwaway task which will only be enqueued once, or
 * use one task per CPU!
 */
static int
taskqueue_enqueue_locked(struct taskqueue *queue, struct task *task)
{
        struct task *ins;
        struct task *prev;

        /*
         * Don't allow new tasks on a queue which is being freed.
         */
        if ((queue->tq_flags & TQ_FLAGS_ACTIVE) == 0)
                return EPIPE;

        /*
         * Count multiple enqueues.
         */
        if (task->ta_pending) {
                KKASSERT(queue == task->ta_queue);
                task->ta_pending++;
                return 0;
        }
        task->ta_queue = queue;

        /*
         * Optimise the case when all tasks have the same priority.
         */
        prev = STAILQ_LAST(&queue->tq_queue, task, ta_link);
        if (!prev || prev->ta_priority >= task->ta_priority) {
                STAILQ_INSERT_TAIL(&queue->tq_queue, task, ta_link);
        } else {
                prev = NULL;
                for (ins = STAILQ_FIRST(&queue->tq_queue); ins;
                     prev = ins, ins = STAILQ_NEXT(ins, ta_link))
                        if (ins->ta_priority < task->ta_priority)
                                break;

                if (prev)
                        STAILQ_INSERT_AFTER(&queue->tq_queue, prev, task, ta_link);
                else
                        STAILQ_INSERT_HEAD(&queue->tq_queue, task, ta_link);
        }

        task->ta_pending = 1;
        if ((queue->tq_flags & TQ_FLAGS_BLOCKED) == 0) {
                if (queue->tq_enqueue)
                        queue->tq_enqueue(queue->tq_context);
        } else {
                queue->tq_flags |= TQ_FLAGS_PENDING;
        }

        return 0;
}

/*
 * This version requires that the task not be moved between queues
 * in an uncontrolled fashion.
 */
int
taskqueue_enqueue(struct taskqueue *queue, struct task *task)
{
        int res;

        TQ_LOCK(queue);
        res = taskqueue_enqueue_locked(queue, task);
        TQ_UNLOCK(queue);

        return (res);
}

/*
 * This version allows a task to be moved between queues in an uncontrolled
 * fashion.  (*qpp) is set to the queue the task is (possibly already)
 * enqueued on, or the specified queue if it is possible to move the task.
 */
int
taskqueue_enqueue_optq(struct taskqueue *queue, struct taskqueue **qpp,
                       struct task *task)
{
        struct taskqueue *qtmp;
        int res;

        /*
         * Interlock for task structure check, handle the case where we
         * are unable to safely shift the task to the specified queue.
         */
        for (;;) {
                qtmp = task->ta_queue;
                cpu_ccfence();

                if (qtmp == NULL) {
                        spin_lock(&taskqueue_queues_spin);
                        if (task->ta_queue == NULL)
                                task->ta_queue = queue;
                        spin_unlock(&taskqueue_queues_spin);
                } else {
                        TQ_LOCK(qtmp);
                        if (task->ta_queue == qtmp) {
                                if (qtmp == queue)
                                        break;

                                /*
                                 * If qtmp is pending on a different queue
                                 * it must stay on that queue.
                                 *
                                 * WARNING: Once ta_queue is reassigned
                                 *          our qtmp lock is no longer
                                 *          sufficient and we lose control
                                 *          of the task.
                                 */
                                if (task->ta_pending) {
                                        task->ta_pending++;
                                        *qpp = qtmp;
                                        TQ_UNLOCK(qtmp);
                                        return 0;
                                }
                                cpu_sfence();
                                task->ta_queue = queue;
                                cpu_ccfence();
                        }
                        TQ_UNLOCK(qtmp);
                }
                /* retry */
        }

        /*
         * The task is assigned to (queue), enqueue it there.
         */
        *qpp = queue;
        res = taskqueue_enqueue_locked(queue, task);
        TQ_UNLOCK(queue);

        return (res);
}

static void
taskqueue_timeout_func(void *arg)
{
        struct taskqueue *queue;
        struct timeout_task *timeout_task;

        timeout_task = arg;
        queue = timeout_task->t.ta_queue;

        TQ_LOCK(queue);
        KASSERT((timeout_task->f & DT_CALLOUT_ARMED) != 0, ("Stray timeout"));
        timeout_task->f &= ~DT_CALLOUT_ARMED;
        queue->tq_callouts--;
        taskqueue_enqueue_locked(queue, &timeout_task->t);
        TQ_UNLOCK(queue);
}

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

        TQ_LOCK(queue);
        KASSERT(timeout_task->t.ta_queue == NULL ||
                timeout_task->t.ta_queue == queue,
                ("Migrated queue"));
        timeout_task->t.ta_queue = queue;
        res = timeout_task->t.ta_pending;
        if (ticks == 0) {
                taskqueue_enqueue_locked(queue, &timeout_task->t);
                TQ_UNLOCK(queue);
        } else {
                if ((timeout_task->f & DT_CALLOUT_ARMED) != 0) {
                        res++;
                } else {
                        queue->tq_callouts++;
                        timeout_task->f |= DT_CALLOUT_ARMED;
                }
                TQ_UNLOCK(queue);
                callout_reset(&timeout_task->c, ticks, taskqueue_timeout_func,
                              timeout_task);
        }
        return (res);
}

void
taskqueue_block(struct taskqueue *queue)
{
        TQ_LOCK(queue);
        queue->tq_flags |= TQ_FLAGS_BLOCKED;
        TQ_UNLOCK(queue);
}

void
taskqueue_unblock(struct taskqueue *queue)
{
        TQ_LOCK(queue);
        queue->tq_flags &= ~TQ_FLAGS_BLOCKED;
        if (queue->tq_flags & TQ_FLAGS_PENDING) {
                queue->tq_flags &= ~TQ_FLAGS_PENDING;
                if (queue->tq_enqueue)
                        queue->tq_enqueue(queue->tq_context);
        }
        TQ_UNLOCK(queue);
}

static void
taskqueue_run(struct taskqueue *queue, int lock_held)
{
        struct task *task;
        int pending;

        if (lock_held == 0)
                TQ_LOCK(queue);
        while (STAILQ_FIRST(&queue->tq_queue)) {
                /*
                 * Carefully remove the first task from the queue and
                 * zero its pending count.
                 */
                task = STAILQ_FIRST(&queue->tq_queue);
                STAILQ_REMOVE_HEAD(&queue->tq_queue, ta_link);
                pending = task->ta_pending;
                task->ta_pending = 0;
                queue->tq_running = task;

                TQ_UNLOCK(queue);
                task->ta_func(task->ta_context, pending);
                queue->tq_running = NULL;
                wakeup(task);
                TQ_LOCK(queue);
        }
        if (lock_held == 0)
                TQ_UNLOCK(queue);
}

static int
taskqueue_cancel_locked(struct taskqueue *queue, struct task *task,
    u_int *pendp)
{

        if (task->ta_pending > 0)
                STAILQ_REMOVE(&queue->tq_queue, task, task, ta_link);
        if (pendp != NULL)
                *pendp = task->ta_pending;
        task->ta_pending = 0;
        return (task == queue->tq_running ? EBUSY : 0);
}

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

        TQ_LOCK(queue);
        error = taskqueue_cancel_locked(queue, task, pendp);
        TQ_UNLOCK(queue);

        return (error);
}

int
taskqueue_cancel_simple(struct task *task)
{
        struct taskqueue *queue;
        int error;

        for (;;) {
                queue = task->ta_queue;
                cpu_ccfence();
                if (queue == NULL) {
                        error = 0;
                        break;
                }
                TQ_LOCK(queue);
                if (queue == task->ta_queue) {
                        error = taskqueue_cancel_locked(queue, task, NULL);
                        TQ_UNLOCK(queue);
                        break;
                }
                TQ_UNLOCK(queue);
        }
        return error;
}

int
taskqueue_cancel_timeout(struct taskqueue *queue,
                         struct timeout_task *timeout_task, u_int *pendp)
{
        u_int pending, pending1;
        int error;

        pending = !!callout_stop(&timeout_task->c);
        TQ_LOCK(queue);
        error = taskqueue_cancel_locked(queue, &timeout_task->t, &pending1);
        if ((timeout_task->f & DT_CALLOUT_ARMED) != 0) {
                timeout_task->f &= ~DT_CALLOUT_ARMED;
                queue->tq_callouts--;
        }
        TQ_UNLOCK(queue);

        if (pendp != NULL)
                *pendp = pending + pending1;
        return (error);
}

void
taskqueue_drain(struct taskqueue *queue, struct task *task)
{
        TQ_LOCK(queue);
        while (task->ta_pending != 0 || task == queue->tq_running)
                TQ_SLEEP(queue, task, "-");
        TQ_UNLOCK(queue);
}

/*
 * Wait for the task to drain and return
 */
void
taskqueue_drain_simple(struct task *task)
{
        struct taskqueue *queue;

        for (;;) {
                queue = task->ta_queue;
                cpu_ccfence();
                if (queue == NULL)
                        return;
                TQ_LOCK(queue);
                if (task->ta_pending == 0 && task != queue->tq_running) {
                        TQ_UNLOCK(queue);
                        return;
                }
                TQ_SLEEP(queue, task, "-");
                TQ_UNLOCK(queue);
        }
}

void
taskqueue_drain_timeout(struct taskqueue *queue,
    struct timeout_task *timeout_task)
{
        callout_cancel(&timeout_task->c);
        taskqueue_drain(queue, &timeout_task->t);
}

static void
taskqueue_swi_enqueue(void *context)
{
        setsofttq();
}

static void
taskqueue_swi_run(void *arg, void *frame)
{
        taskqueue_run(taskqueue_swi, 0);
}

static void
taskqueue_swi_mp_run(void *arg, void *frame)
{
        taskqueue_run(taskqueue_swi_mp, 0);
}

int
taskqueue_start_threads(struct taskqueue **tqp, int count, int pri, int ncpu,
                        const char *fmt, ...)
{
        __va_list ap;
        struct thread *td;
        struct taskqueue *tq;
        int i, error, cpu;
        char ktname[MAXCOMLEN];

        tq = *tqp;
        cpu = ncpu;

        /* catch call argument mistakes */
        KKASSERT(pri > 0 && pri < TDPRI_MAX);
        KKASSERT(tq->tq_enqueue == taskqueue_thread_enqueue);

        if (count <= 0)
                return EINVAL;

        __va_start(ap, fmt);
        kvsnprintf(ktname, MAXCOMLEN, fmt, ap);
        __va_end(ap);

        tq->tq_threads = kmalloc(sizeof(struct thread *) * count, M_TASKQUEUE,
            M_WAITOK | M_ZERO);

        for (i = 0; i < count; i++) {
                /*
                 * If no specific cpu was specified and more than one thread
                 * is to be created, we distribute the threads amongst all
                 * cpus.
                 */
                if ((ncpu <= -1) && (count > 1))
                        cpu = i % ncpus;

                if (count == 1) {
                        error = lwkt_create(taskqueue_thread_loop, tqp,
                                            &tq->tq_threads[i], NULL,
                                            TDF_NOSTART, cpu,
                                            "%s", ktname);
                } else {
                        error = lwkt_create(taskqueue_thread_loop, tqp,
                                            &tq->tq_threads[i], NULL,
                                            TDF_NOSTART, cpu,
                                            "%s_%d", ktname, i);
                }
                if (error) {
                        kprintf("%s: lwkt_create(%s): error %d", __func__,
                            ktname, error);
                        tq->tq_threads[i] = NULL;
                } else {
                        td = tq->tq_threads[i];
                        lwkt_setpri_initial(td, pri);
                        lwkt_schedule(td);
                        tq->tq_tcount++;
                }
        }

        return 0;
}

void
taskqueue_thread_loop(void *arg)
{
        struct taskqueue **tqp, *tq;

        tqp = arg;
        tq = *tqp;
        TQ_LOCK(tq);
        while ((tq->tq_flags & TQ_FLAGS_ACTIVE) != 0) {
                taskqueue_run(tq, 1);
                TQ_SLEEP(tq, tq, "tqthr");
        }

        /* rendezvous with thread that asked us to terminate */
        tq->tq_tcount--;
        TQ_UNLOCK(tq);
        wakeup_one(tq->tq_threads);
        lwkt_exit();
}

/* NOTE: tq must be locked */
void
taskqueue_thread_enqueue(void *context)
{
        struct taskqueue **tqp, *tq;

        tqp = context;
        tq = *tqp;

        /* Unlock spinlock before wakeup_one() */
        TQ_UNLOCK(tq);
        wakeup_one(tq);
        TQ_LOCK(tq);
}

TASKQUEUE_DEFINE(swi, taskqueue_swi_enqueue, 0,
         register_swi(SWI_TQ, taskqueue_swi_run, NULL, "swi_taskq", NULL, -1));
/*
 * XXX: possibly use a different SWI_TQ_MP or so.
 * related: sys/interrupt.h
 * related: platform/XXX/isa/ipl_funcs.c
 */
TASKQUEUE_DEFINE(swi_mp, taskqueue_swi_enqueue, 0,
    register_swi_mp(SWI_TQ, taskqueue_swi_mp_run, NULL, "swi_mp_taskq", NULL,
                    -1));

struct taskqueue *taskqueue_thread[MAXCPU];

static void
taskqueue_init(void)
{
        int cpu;

        lockinit(&taskqueue_queues_lock, "tqqueues", 0, 0);
        spin_init(&taskqueue_queues_spin, "tqspin");
        STAILQ_INIT(&taskqueue_queues);

        for (cpu = 0; cpu < ncpus; cpu++) {
                taskqueue_thread[cpu] = taskqueue_create("thread", M_INTWAIT,
                    taskqueue_thread_enqueue, &taskqueue_thread[cpu]);
                taskqueue_start_threads(&taskqueue_thread[cpu], 1,
                    TDPRI_KERN_DAEMON, cpu, "taskq_cpu %d", cpu);
        }
}

SYSINIT(taskqueueinit, SI_SUB_PRE_DRIVERS, SI_ORDER_FIRST, taskqueue_init, NULL);