sched.c

// SPDX-License-Identifier: GPL-2.0-only
/*
 * linux/net/sunrpc/sched.c
 *
 * Scheduling for synchronous and asynchronous RPC requests.
 *
 * Copyright (C) 1996 Olaf Kirch, <okir@monad.swb.de>
 *
 * TCP NFS related read + write fixes
 * (C) 1999 Dave Airlie, University of Limerick, Ireland <airlied@linux.ie>
 */

#include <linux/module.h>

#include <linux/sched.h>
#include <linux/interrupt.h>
#include <linux/slab.h>
#include <linux/mempool.h>
#include <linux/smp.h>
#include <linux/spinlock.h>
#include <linux/mutex.h>
#include <linux/freezer.h>
#include <linux/sched/mm.h>

#include <linux/sunrpc/clnt.h>
#include <linux/sunrpc/metrics.h>

#include "sunrpc.h"

#define CREATE_TRACE_POINTS
#include <trace/events/sunrpc.h>

/*
 * RPC slabs and memory pools
 */
#define RPC_BUFFER_MAXSIZE	(2048)
#define RPC_BUFFER_POOLSIZE	(8)
#define RPC_TASK_POOLSIZE	(8)
static struct kmem_cache	*rpc_task_slabp __read_mostly;
static struct kmem_cache	*rpc_buffer_slabp __read_mostly;
static mempool_t	*rpc_task_mempool __read_mostly;
static mempool_t	*rpc_buffer_mempool __read_mostly;

static void			rpc_async_schedule(struct work_struct *);
static void			 rpc_release_task(struct rpc_task *task);
static void __rpc_queue_timer_fn(struct work_struct *);

/*
 * RPC tasks sit here while waiting for conditions to improve.
 */
static struct rpc_wait_queue delay_queue;

/*
 * rpciod-related stuff
 */
struct workqueue_struct *rpciod_workqueue __read_mostly;
struct workqueue_struct *xprtiod_workqueue __read_mostly;
EXPORT_SYMBOL_GPL(xprtiod_workqueue);

unsigned long
rpc_task_timeout(const struct rpc_task *task)
{
	unsigned long timeout = READ_ONCE(task->tk_timeout);

	if (timeout != 0) {
		unsigned long now = jiffies;
		if (time_before(now, timeout))
			return timeout - now;
	}
	return 0;
}
EXPORT_SYMBOL_GPL(rpc_task_timeout);

/*
 * Disable the timer for a given RPC task. Should be called with
 * queue->lock and bh_disabled in order to avoid races within
 * rpc_run_timer().
 */
static void
__rpc_disable_timer(struct rpc_wait_queue *queue, struct rpc_task *task)
{
	if (list_empty(&task->u.tk_wait.timer_list))
		return;
	task->tk_timeout = 0;
	list_del(&task->u.tk_wait.timer_list);
	if (list_empty(&queue->timer_list.list))
		cancel_delayed_work(&queue->timer_list.dwork);
}

static void
rpc_set_queue_timer(struct rpc_wait_queue *queue, unsigned long expires)
{
	unsigned long now = jiffies;
	queue->timer_list.expires = expires;
	if (time_before_eq(expires, now))
		expires = 0;
	else
		expires -= now;
	mod_delayed_work(rpciod_workqueue, &queue->timer_list.dwork, expires);
}

/*
 * Set up a timer for the current task.
 */
static void
__rpc_add_timer(struct rpc_wait_queue *queue, struct rpc_task *task,
		unsigned long timeout)
{
	task->tk_timeout = timeout;
	if (list_empty(&queue->timer_list.list) || time_before(timeout, queue->timer_list.expires))
		rpc_set_queue_timer(queue, timeout);
	list_add(&task->u.tk_wait.timer_list, &queue->timer_list.list);
}

static void rpc_set_waitqueue_priority(struct rpc_wait_queue *queue, int priority)
{
	if (queue->priority != priority) {
		queue->priority = priority;
		queue->nr = 1U << priority;
	}
}

static void rpc_reset_waitqueue_priority(struct rpc_wait_queue *queue)
{
	rpc_set_waitqueue_priority(queue, queue->maxpriority);
}

/*
 * Add a request to a queue list
 */
static void
__rpc_list_enqueue_task(struct list_head *q, struct rpc_task *task)
{
	struct rpc_task *t;

	list_for_each_entry(t, q, u.tk_wait.list) {
		if (t->tk_owner == task->tk_owner) {
			list_add_tail(&task->u.tk_wait.links,
					&t->u.tk_wait.links);
			/* Cache the queue head in task->u.tk_wait.list */
			task->u.tk_wait.list.next = q;
			task->u.tk_wait.list.prev = NULL;
			return;
		}
	}
	INIT_LIST_HEAD(&task->u.tk_wait.links);
	list_add_tail(&task->u.tk_wait.list, q);
}

/*
 * Remove request from a queue list
 */
static void
__rpc_list_dequeue_task(struct rpc_task *task)
{
	struct list_head *q;
	struct rpc_task *t;

	if (task->u.tk_wait.list.prev == NULL) {
		list_del(&task->u.tk_wait.links);
		return;
	}
	if (!list_empty(&task->u.tk_wait.links)) {
		t = list_first_entry(&task->u.tk_wait.links,
				struct rpc_task,
				u.tk_wait.links);
		/* Assume __rpc_list_enqueue_task() cached the queue head */
		q = t->u.tk_wait.list.next;
		list_add_tail(&t->u.tk_wait.list, q);
		list_del(&task->u.tk_wait.links);
	}
	list_del(&task->u.tk_wait.list);
}

/*
 * Add new request to a priority queue.
 */
static void __rpc_add_wait_queue_priority(struct rpc_wait_queue *queue,
		struct rpc_task *task,
		unsigned char queue_priority)
{
	if (unlikely(queue_priority > queue->maxpriority))
		queue_priority = queue->maxpriority;
	__rpc_list_enqueue_task(&queue->tasks[queue_priority], task);
}

/*
 * Add new request to wait queue.
 *
 * Swapper tasks always get inserted at the head of the queue.
 * This should avoid many nasty memory deadlocks and hopefully
 * improve overall performance.
 * Everyone else gets appended to the queue to ensure proper FIFO behavior.
 */
static void __rpc_add_wait_queue(struct rpc_wait_queue *queue,
		struct rpc_task *task,
		unsigned char queue_priority)
{
	INIT_LIST_HEAD(&task->u.tk_wait.timer_list);
	if (RPC_IS_PRIORITY(queue))
		__rpc_add_wait_queue_priority(queue, task, queue_priority);
	else if (RPC_IS_SWAPPER(task))
		list_add(&task->u.tk_wait.list, &queue->tasks[0]);
	else
		list_add_tail(&task->u.tk_wait.list, &queue->tasks[0]);
	task->tk_waitqueue = queue;
	queue->qlen++;
	/* barrier matches the read in rpc_wake_up_task_queue_locked() */
	smp_wmb();
	rpc_set_queued(task);
}

/*
 * Remove request from a priority queue.
 */
static void __rpc_remove_wait_queue_priority(struct rpc_task *task)
{
	__rpc_list_dequeue_task(task);
}

/*
 * Remove request from queue.
 * Note: must be called with spin lock held.
 */
static void __rpc_remove_wait_queue(struct rpc_wait_queue *queue, struct rpc_task *task)
{
	__rpc_disable_timer(queue, task);
	if (RPC_IS_PRIORITY(queue))
		__rpc_remove_wait_queue_priority(task);
	else
		list_del(&task->u.tk_wait.list);
	queue->qlen--;
}

static void __rpc_init_priority_wait_queue(struct rpc_wait_queue *queue, const char *qname, unsigned char nr_queues)
{
	int i;

	spin_lock_init(&queue->lock);
	for (i = 0; i < ARRAY_SIZE(queue->tasks); i++)
		INIT_LIST_HEAD(&queue->tasks[i]);
	queue->maxpriority = nr_queues - 1;
	rpc_reset_waitqueue_priority(queue);
	queue->qlen = 0;
	queue->timer_list.expires = 0;
	INIT_DELAYED_WORK(&queue->timer_list.dwork, __rpc_queue_timer_fn);
	INIT_LIST_HEAD(&queue->timer_list.list);
	rpc_assign_waitqueue_name(queue, qname);
}

void rpc_init_priority_wait_queue(struct rpc_wait_queue *queue, const char *qname)
{
	__rpc_init_priority_wait_queue(queue, qname, RPC_NR_PRIORITY);
}
EXPORT_SYMBOL_GPL(rpc_init_priority_wait_queue);

void rpc_init_wait_queue(struct rpc_wait_queue *queue, const char *qname)
{
	__rpc_init_priority_wait_queue(queue, qname, 1);
}
EXPORT_SYMBOL_GPL(rpc_init_wait_queue);

void rpc_destroy_wait_queue(struct rpc_wait_queue *queue)
{
	cancel_delayed_work_sync(&queue->timer_list.dwork);
}
EXPORT_SYMBOL_GPL(rpc_destroy_wait_queue);

static int rpc_wait_bit_killable(struct wait_bit_key *key, int mode)
{
	freezable_schedule_unsafe();
	if (signal_pending_state(mode, current))
		return -ERESTARTSYS;
	return 0;
}

#if IS_ENABLED(CONFIG_SUNRPC_DEBUG) || IS_ENABLED(CONFIG_TRACEPOINTS)
static void rpc_task_set_debuginfo(struct rpc_task *task)
{
	static atomic_t rpc_pid;

	task->tk_pid = atomic_inc_return(&rpc_pid);
}
#else
static inline void rpc_task_set_debuginfo(struct rpc_task *task)
{
}
#endif

static void rpc_set_active(struct rpc_task *task)
{
	rpc_task_set_debuginfo(task);
	set_bit(RPC_TASK_ACTIVE, &task->tk_runstate);
	trace_rpc_task_begin(task, NULL);
}

/*
 * Mark an RPC call as having completed by clearing the 'active' bit
 * and then waking up all tasks that were sleeping.
 */
static int rpc_complete_task(struct rpc_task *task)
{
	void *m = &task->tk_runstate;
	wait_queue_head_t *wq = bit_waitqueue(m, RPC_TASK_ACTIVE);
	struct wait_bit_key k = __WAIT_BIT_KEY_INITIALIZER(m, RPC_TASK_ACTIVE);
	unsigned long flags;
	int ret;

	trace_rpc_task_complete(task, NULL);

	spin_lock_irqsave(&wq->lock, flags);
	clear_bit(RPC_TASK_ACTIVE, &task->tk_runstate);
	ret = atomic_dec_and_test(&task->tk_count);
	if (waitqueue_active(wq))
		__wake_up_locked_key(wq, TASK_NORMAL, &k);
	spin_unlock_irqrestore(&wq->lock, flags);
	return ret;
}

/*
 * Allow callers to wait for completion of an RPC call
 *
 * Note the use of out_of_line_wait_on_bit() rather than wait_on_bit()
 * to enforce taking of the wq->lock and hence avoid races with
 * rpc_complete_task().
 */
int __rpc_wait_for_completion_task(struct rpc_task *task, wait_bit_action_f *action)
{
	if (action == NULL)
		action = rpc_wait_bit_killable;
	return out_of_line_wait_on_bit(&task->tk_runstate, RPC_TASK_ACTIVE,
			action, TASK_KILLABLE);
}
EXPORT_SYMBOL_GPL(__rpc_wait_for_completion_task);

/*
 * Make an RPC task runnable.
 *
 * Note: If the task is ASYNC, and is being made runnable after sitting on an
 * rpc_wait_queue, this must be called with the queue spinlock held to protect
 * the wait queue operation.
 * Note the ordering of rpc_test_and_set_running() and rpc_clear_queued(),
 * which is needed to ensure that __rpc_execute() doesn't loop (due to the
 * lockless RPC_IS_QUEUED() test) before we've had a chance to test
 * the RPC_TASK_RUNNING flag.
 */
static void rpc_make_runnable(struct workqueue_struct *wq,
		struct rpc_task *task)
{
	bool need_wakeup = !rpc_test_and_set_running(task);

	rpc_clear_queued(task);
	if (!need_wakeup)
		return;
	if (RPC_IS_ASYNC(task)) {
		INIT_WORK(&task->u.tk_work, rpc_async_schedule);
		queue_work(wq, &task->u.tk_work);
	} else
		wake_up_bit(&task->tk_runstate, RPC_TASK_QUEUED);
}

/*
 * Prepare for sleeping on a wait queue.
 * By always appending tasks to the list we ensure FIFO behavior.
 * NB: An RPC task will only receive interrupt-driven events as long
 * as it's on a wait queue.
 */
static void __rpc_do_sleep_on_priority(struct rpc_wait_queue *q,
		struct rpc_task *task,
		unsigned char queue_priority)
{
	trace_rpc_task_sleep(task, q);

	__rpc_add_wait_queue(q, task, queue_priority);
}

static void __rpc_sleep_on_priority(struct rpc_wait_queue *q,
		struct rpc_task *task,
		unsigned char queue_priority)
{
	if (WARN_ON_ONCE(RPC_IS_QUEUED(task)))
		return;
	__rpc_do_sleep_on_priority(q, task, queue_priority);
}

static void __rpc_sleep_on_priority_timeout(struct rpc_wait_queue *q,
		struct rpc_task *task, unsigned long timeout,
		unsigned char queue_priority)
{
	if (WARN_ON_ONCE(RPC_IS_QUEUED(task)))
		return;
	if (time_is_after_jiffies(timeout)) {
		__rpc_do_sleep_on_priority(q, task, queue_priority);
		__rpc_add_timer(q, task, timeout);
	} else
		task->tk_status = -ETIMEDOUT;
}

static void rpc_set_tk_callback(struct rpc_task *task, rpc_action action)
{
	if (action && !WARN_ON_ONCE(task->tk_callback != NULL))
		task->tk_callback = action;
}

static bool rpc_sleep_check_activated(struct rpc_task *task)
{
	/* We shouldn't ever put an inactive task to sleep */
	if (WARN_ON_ONCE(!RPC_IS_ACTIVATED(task))) {
		task->tk_status = -EIO;
		rpc_put_task_async(task);
		return false;
	}
	return true;
}

void rpc_sleep_on_timeout(struct rpc_wait_queue *q, struct rpc_task *task,
				rpc_action action, unsigned long timeout)
{
	if (!rpc_sleep_check_activated(task))
		return;

	rpc_set_tk_callback(task, action);

	/*
	 * Protect the queue operations.
	 */
	spin_lock(&q->lock);
	__rpc_sleep_on_priority_timeout(q, task, timeout, task->tk_priority);
	spin_unlock(&q->lock);
}
EXPORT_SYMBOL_GPL(rpc_sleep_on_timeout);

void rpc_sleep_on(struct rpc_wait_queue *q, struct rpc_task *task,
				rpc_action action)
{
	if (!rpc_sleep_check_activated(task))
		return;

	rpc_set_tk_callback(task, action);

	WARN_ON_ONCE(task->tk_timeout != 0);
	/*
	 * Protect the queue operations.
	 */
	spin_lock(&q->lock);
	__rpc_sleep_on_priority(q, task, task->tk_priority);
	spin_unlock(&q->lock);
}
EXPORT_SYMBOL_GPL(rpc_sleep_on);

void rpc_sleep_on_priority_timeout(struct rpc_wait_queue *q,
		struct rpc_task *task, unsigned long timeout, int priority)
{
	if (!rpc_sleep_check_activated(task))
		return;

	priority -= RPC_PRIORITY_LOW;
	/*
	 * Protect the queue operations.
	 */
	spin_lock(&q->lock);
	__rpc_sleep_on_priority_timeout(q, task, timeout, priority);
	spin_unlock(&q->lock);
}
EXPORT_SYMBOL_GPL(rpc_sleep_on_priority_timeout);

void rpc_sleep_on_priority(struct rpc_wait_queue *q, struct rpc_task *task,
		int priority)
{
	if (!rpc_sleep_check_activated(task))
		return;

	WARN_ON_ONCE(task->tk_timeout != 0);
	priority -= RPC_PRIORITY_LOW;
	/*
	 * Protect the queue operations.
	 */
	spin_lock(&q->lock);
	__rpc_sleep_on_priority(q, task, priority);
	spin_unlock(&q->lock);
}
EXPORT_SYMBOL_GPL(rpc_sleep_on_priority);

/**
 * __rpc_do_wake_up_task_on_wq - wake up a single rpc_task
 * @wq: workqueue on which to run task
 * @queue: wait queue
 * @task: task to be woken up
 *
 * Caller must hold queue->lock, and have cleared the task queued flag.
 */
static void __rpc_do_wake_up_task_on_wq(struct workqueue_struct *wq,
		struct rpc_wait_queue *queue,
		struct rpc_task *task)
{
	/* Has the task been executed yet? If not, we cannot wake it up! */
	if (!RPC_IS_ACTIVATED(task)) {
		printk(KERN_ERR "RPC: Inactive task (%p) being woken up!\n", task);
		return;
	}

	trace_rpc_task_wakeup(task, queue);

	__rpc_remove_wait_queue(queue, task);

	rpc_make_runnable(wq, task);
}

/*
 * Wake up a queued task while the queue lock is being held
 */
static struct rpc_task *
rpc_wake_up_task_on_wq_queue_action_locked(struct workqueue_struct *wq,
		struct rpc_wait_queue *queue, struct rpc_task *task,
		bool (*action)(struct rpc_task *, void *), void *data)
{
	if (RPC_IS_QUEUED(task)) {
		smp_rmb();
		if (task->tk_waitqueue == queue) {
			if (action == NULL || action(task, data)) {
				__rpc_do_wake_up_task_on_wq(wq, queue, task);
				return task;
			}
		}
	}
	return NULL;
}

/*
 * Wake up a queued task while the queue lock is being held
 */
static void rpc_wake_up_task_queue_locked(struct rpc_wait_queue *queue,
					  struct rpc_task *task)
{
	rpc_wake_up_task_on_wq_queue_action_locked(rpciod_workqueue, queue,
						   task, NULL, NULL);
}

/*
 * Wake up a task on a specific queue
 */
void rpc_wake_up_queued_task(struct rpc_wait_queue *queue, struct rpc_task *task)
{
	if (!RPC_IS_QUEUED(task))
		return;
	spin_lock(&queue->lock);
	rpc_wake_up_task_queue_locked(queue, task);
	spin_unlock(&queue->lock);
}
EXPORT_SYMBOL_GPL(rpc_wake_up_queued_task);

static bool rpc_task_action_set_status(struct rpc_task *task, void *status)
{
	task->tk_status = *(int *)status;
	return true;
}

static void
rpc_wake_up_task_queue_set_status_locked(struct rpc_wait_queue *queue,
		struct rpc_task *task, int status)
{
	rpc_wake_up_task_on_wq_queue_action_locked(rpciod_workqueue, queue,
			task, rpc_task_action_set_status, &status);
}

/**
 * rpc_wake_up_queued_task_set_status - wake up a task and set task->tk_status
 * @queue: pointer to rpc_wait_queue
 * @task: pointer to rpc_task
 * @status: integer error value
 *
 * If @task is queued on @queue, then it is woken up, and @task->tk_status is
 * set to the value of @status.
 */
void
rpc_wake_up_queued_task_set_status(struct rpc_wait_queue *queue,
		struct rpc_task *task, int status)
{
	if (!RPC_IS_QUEUED(task))
		return;
	spin_lock(&queue->lock);
	rpc_wake_up_task_queue_set_status_locked(queue, task, status);
	spin_unlock(&queue->lock);
}

/*
 * Wake up the next task on a priority queue.
 */
static struct rpc_task *__rpc_find_next_queued_priority(struct rpc_wait_queue *queue)
{
	struct list_head *q;
	struct rpc_task *task;

	/*
	 * Service a batch of tasks from a single owner.
	 */
	q = &queue->tasks[queue->priority];
	if (!list_empty(q) && --queue->nr) {
		task = list_first_entry(q, struct rpc_task, u.tk_wait.list);
		goto out;
	}

	/*
	 * Service the next queue.
	 */
	do {
		if (q == &queue->tasks[0])
			q = &queue->tasks[queue->maxpriority];
		else
			q = q - 1;
		if (!list_empty(q)) {
			task = list_first_entry(q, struct rpc_task, u.tk_wait.list);
			goto new_queue;
		}
	} while (q != &queue->tasks[queue->priority]);

	rpc_reset_waitqueue_priority(queue);
	return NULL;

new_queue:
	rpc_set_waitqueue_priority(queue, (unsigned int)(q - &queue->tasks[0]));
out:
	return task;
}

static struct rpc_task *__rpc_find_next_queued(struct rpc_wait_queue *queue)
{
	if (RPC_IS_PRIORITY(queue))
		return __rpc_find_next_queued_priority(queue);
	if (!list_empty(&queue->tasks[0]))
		return list_first_entry(&queue->tasks[0], struct rpc_task, u.tk_wait.list);
	return NULL;
}

/*
 * Wake up the first task on the wait queue.
 */
struct rpc_task *rpc_wake_up_first_on_wq(struct workqueue_struct *wq,
		struct rpc_wait_queue *queue,
		bool (*func)(struct rpc_task *, void *), void *data)
{
	struct rpc_task	*task = NULL;

	spin_lock(&queue->lock);
	task = __rpc_find_next_queued(queue);
	if (task != NULL)
		task = rpc_wake_up_task_on_wq_queue_action_locked(wq, queue,
				task, func, data);
	spin_unlock(&queue->lock);

	return task;
}

/*
 * Wake up the first task on the wait queue.
 */
struct rpc_task *rpc_wake_up_first(struct rpc_wait_queue *queue,
		bool (*func)(struct rpc_task *, void *), void *data)
{
	return rpc_wake_up_first_on_wq(rpciod_workqueue, queue, func, data);
}
EXPORT_SYMBOL_GPL(rpc_wake_up_first);

static bool rpc_wake_up_next_func(struct rpc_task *task, void *data)
{
	return true;
}

/*
 * Wake up the next task on the wait queue.
*/
struct rpc_task *rpc_wake_up_next(struct rpc_wait_queue *queue)
{
	return rpc_wake_up_first(queue, rpc_wake_up_next_func, NULL);
}
EXPORT_SYMBOL_GPL(rpc_wake_up_next);

/**
 * rpc_wake_up_locked - wake up all rpc_tasks
 * @queue: rpc_wait_queue on which the tasks are sleeping
 *
 */
static void rpc_wake_up_locked(struct rpc_wait_queue *queue)
{
	struct rpc_task *task;

	for (;;) {
		task = __rpc_find_next_queued(queue);
		if (task == NULL)
			break;
		rpc_wake_up_task_queue_locked(queue, task);
	}
}

/**
 * rpc_wake_up - wake up all rpc_tasks
 * @queue: rpc_wait_queue on which the tasks are sleeping
 *
 * Grabs queue->lock
 */
void rpc_wake_up(struct rpc_wait_queue *queue)
{
	spin_lock(&queue->lock);
	rpc_wake_up_locked(queue);
	spin_unlock(&queue->lock);
}
EXPORT_SYMBOL_GPL(rpc_wake_up);

/**
 * rpc_wake_up_status_locked - wake up all rpc_tasks and set their status value.
 * @queue: rpc_wait_queue on which the tasks are sleeping
 * @status: status value to set
 */
static void rpc_wake_up_status_locked(struct rpc_wait_queue *queue, int status)
{
	struct rpc_task *task;

	for (;;) {
		task = __rpc_find_next_queued(queue);
		if (task == NULL)
			break;
		rpc_wake_up_task_queue_set_status_locked(queue, task, status);
	}
}

/**
 * rpc_wake_up_status - wake up all rpc_tasks and set their status value.
 * @queue: rpc_wait_queue on which the tasks are sleeping
 * @status: status value to set
 *
 * Grabs queue->lock
 */
void rpc_wake_up_status(struct rpc_wait_queue *queue, int status)
{
	spin_lock(&queue->lock);
	rpc_wake_up_status_locked(queue, status);
	spin_unlock(&queue->lock);
}
EXPORT_SYMBOL_GPL(rpc_wake_up_status);

static void __rpc_queue_timer_fn(struct work_struct *work)
{
	struct rpc_wait_queue *queue = container_of(work,
			struct rpc_wait_queue,
			timer_list.dwork.work);
	struct rpc_task *task, *n;
	unsigned long expires, now, timeo;

	spin_lock(&queue->lock);
	expires = now = jiffies;
	list_for_each_entry_safe(task, n, &queue->timer_list.list, u.tk_wait.timer_list) {
		timeo = task->tk_timeout;
		if (time_after_eq(now, timeo)) {
			trace_rpc_task_timeout(task, task->tk_action);
			task->tk_status = -ETIMEDOUT;
			rpc_wake_up_task_queue_locked(queue, task);
			continue;
		}
		if (expires == now || time_after(expires, timeo))
			expires = timeo;
	}
	if (!list_empty(&queue->timer_list.list))
		rpc_set_queue_timer(queue, expires);
	spin_unlock(&queue->lock);
}

static void __rpc_atrun(struct rpc_task *task)
{
	if (task->tk_status == -ETIMEDOUT)
		task->tk_status = 0;
}

/*
 * Run a task at a later time
 */
void rpc_delay(struct rpc_task *task, unsigned long delay)
{
	rpc_sleep_on_timeout(&delay_queue, task, __rpc_atrun, jiffies + delay);
}
EXPORT_SYMBOL_GPL(rpc_delay);

/*
 * Helper to call task->tk_ops->rpc_call_prepare
 */
void rpc_prepare_task(struct rpc_task *task)
{
	task->tk_ops->rpc_call_prepare(task, task->tk_calldata);
}

static void
rpc_init_task_statistics(struct rpc_task *task)
{
	/* Initialize retry counters */
	task->tk_garb_retry = 2;
	task->tk_cred_retry = 2;
	task->tk_rebind_retry = 2;

	/* starting timestamp */
	task->tk_start = ktime_get();
}

static void
rpc_reset_task_statistics(struct rpc_task *task)
{
	task->tk_timeouts = 0;
	task->tk_flags &= ~(RPC_CALL_MAJORSEEN|RPC_TASK_SENT);
	rpc_init_task_statistics(task);
}

/*
 * Helper that calls task->tk_ops->rpc_call_done if it exists
 */
void rpc_exit_task(struct rpc_task *task)
{
	trace_rpc_task_end(task, task->tk_action);
	task->tk_action = NULL;
	if (task->tk_ops->rpc_count_stats)
		task->tk_ops->rpc_count_stats(task, task->tk_calldata);
	else if (task->tk_client)
		rpc_count_iostats(task, task->tk_client->cl_metrics);
	if (task->tk_ops->rpc_call_done != NULL) {
		task->tk_ops->rpc_call_done(task, task->tk_calldata);
		if (task->tk_action != NULL) {
			/* Always release the RPC slot and buffer memory */
			xprt_release(task);
			rpc_reset_task_statistics(task);
		}
	}
}

void rpc_signal_task(struct rpc_task *task)
{
	struct rpc_wait_queue *queue;

	if (!RPC_IS_ACTIVATED(task))
		return;

	trace_rpc_task_signalled(task, task->tk_action);
	set_bit(RPC_TASK_SIGNALLED, &task->tk_runstate);
	smp_mb__after_atomic();
	queue = READ_ONCE(task->tk_waitqueue);
	if (queue)
		rpc_wake_up_queued_task_set_status(queue, task, -ERESTARTSYS);
}

void rpc_exit(struct rpc_task *task, int status)
{
	task->tk_status = status;
	task->tk_action = rpc_exit_task;
	rpc_wake_up_queued_task(task->tk_waitqueue, task);
}
EXPORT_SYMBOL_GPL(rpc_exit);

void rpc_release_calldata(const struct rpc_call_ops *ops, void *calldata)
{
	if (ops->rpc_release != NULL)
		ops->rpc_release(calldata);
}

/*
 * This is the RPC `scheduler' (or rather, the finite state machine).
 */
static void __rpc_execute(struct rpc_task *task)
{
	struct rpc_wait_queue *queue;
	int task_is_async = RPC_IS_ASYNC(task);
	int status = 0;

	WARN_ON_ONCE(RPC_IS_QUEUED(task));
	if (RPC_IS_QUEUED(task))
		return;

	for (;;) {
		void (*do_action)(struct rpc_task *);

		/*
		 * Perform the next FSM step or a pending callback.
		 *
		 * tk_action may be NULL if the task has been killed.
		 * In particular, note that rpc_killall_tasks may
		 * do this at any time, so beware when dereferencing.
		 */
		do_action = task->tk_action;
		if (task->tk_callback) {
			do_action = task->tk_callback;
			task->tk_callback = NULL;
		}
		if (!do_action)
			break;
		trace_rpc_task_run_action(task, do_action);
		do_action(task);

		/*
		 * Lockless check for whether task is sleeping or not.
		 */
		if (!RPC_IS_QUEUED(task))
			continue;

		/*
		 * Signalled tasks should exit rather than sleep.
		 */
		if (RPC_SIGNALLED(task)) {
			task->tk_rpc_status = -ERESTARTSYS;
			rpc_exit(task, -ERESTARTSYS);
		}

		/*
		 * The queue->lock protects against races with
		 * rpc_make_runnable().
		 *
		 * Note that once we clear RPC_TASK_RUNNING on an asynchronous
		 * rpc_task, rpc_make_runnable() can assign it to a
		 * different workqueue. We therefore cannot assume that the
		 * rpc_task pointer may still be dereferenced.
		 */
		queue = task->tk_waitqueue;
		spin_lock(&queue->lock);
		if (!RPC_IS_QUEUED(task)) {
			spin_unlock(&queue->lock);
			continue;
		}
		rpc_clear_running(task);
		spin_unlock(&queue->lock);
		if (task_is_async)
			return;

		/* sync task: sleep here */
		trace_rpc_task_sync_sleep(task, task->tk_action);
		status = out_of_line_wait_on_bit(&task->tk_runstate,
				RPC_TASK_QUEUED, rpc_wait_bit_killable,
				TASK_KILLABLE);
		if (status < 0) {
			/*
			 * When a sync task receives a signal, it exits with
			 * -ERESTARTSYS. In order to catch any callbacks that
			 * clean up after sleeping on some queue, we don't
			 * break the loop here, but go around once more.
			 */
			trace_rpc_task_signalled(task, task->tk_action);
			set_bit(RPC_TASK_SIGNALLED, &task->tk_runstate);
			task->tk_rpc_status = -ERESTARTSYS;
			rpc_exit(task, -ERESTARTSYS);
		}
		trace_rpc_task_sync_wake(task, task->tk_action);
	}

	/* Release all resources associated with the task */
	rpc_release_task(task);
}

/*
 * User-visible entry point to the scheduler.
 *
 * This may be called recursively if e.g. an async NFS task updates
 * the attributes and finds that dirty pages must be flushed.
 * NOTE: Upon exit of this function the task is guaranteed to be
 *	 released. In particular note that tk_release() will have
 *	 been called, so your task memory may have been freed.
 */
void rpc_execute(struct rpc_task *task)
{
	bool is_async = RPC_IS_ASYNC(task);

	rpc_set_active(task);
	rpc_make_runnable(rpciod_workqueue, task);
	if (!is_async) {
		unsigned int pflags = memalloc_nofs_save();
		__rpc_execute(task);
		memalloc_nofs_restore(pflags);
	}
}

static void rpc_async_schedule(struct work_struct *work)
{
	unsigned int pflags = memalloc_nofs_save();

	__rpc_execute(container_of(work, struct rpc_task, u.tk_work));
	memalloc_nofs_restore(pflags);
}

/**
 * rpc_malloc - allocate RPC buffer resources
 * @task: RPC task
 *
 * A single memory region is allocated, which is split between the
 * RPC call and RPC reply that this task is being used for. When
 * this RPC is retired, the memory is released by calling rpc_free.
 *
 * To prevent rpciod from hanging, this allocator never sleeps,
 * returning -ENOMEM and suppressing warning if the request cannot
 * be serviced immediately. The caller can arrange to sleep in a
 * way that is safe for rpciod.
 *
 * Most requests are 'small' (under 2KiB) and can be serviced from a
 * mempool, ensuring that NFS reads and writes can always proceed,
 * and that there is good locality of reference for these buffers.
 */
int rpc_malloc(struct rpc_task *task)
{
	struct rpc_rqst *rqst = task->tk_rqstp;