diff options
Diffstat (limited to 'net/sunrpc/sched.c')
| -rw-r--r-- | net/sunrpc/sched.c | 817 |
1 files changed, 518 insertions, 299 deletions
diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c index 93a7a4e94d80..016f16ca5779 100644 --- a/net/sunrpc/sched.c +++ b/net/sunrpc/sched.c @@ -1,3 +1,4 @@ +// SPDX-License-Identifier: GPL-2.0-only /* * linux/net/sunrpc/sched.c * @@ -19,15 +20,13 @@ #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" -#ifdef RPC_DEBUG -#define RPCDBG_FACILITY RPCDBG_SCHED -#endif - #define CREATE_TRACE_POINTS #include <trace/events/sunrpc.h> @@ -44,7 +43,7 @@ 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(unsigned long ptr); +static void __rpc_queue_timer_fn(struct work_struct *); /* * RPC tasks sit here while waiting for conditions to improve. @@ -54,7 +53,38 @@ static struct rpc_wait_queue delay_queue; /* * rpciod-related stuff */ -struct workqueue_struct *rpciod_workqueue; +struct workqueue_struct *rpciod_workqueue __read_mostly; +struct workqueue_struct *xprtiod_workqueue __read_mostly; +EXPORT_SYMBOL_GPL(xprtiod_workqueue); + +gfp_t rpc_task_gfp_mask(void) +{ + if (current->flags & PF_WQ_WORKER) + return GFP_KERNEL | __GFP_NORETRY | __GFP_NOWARN; + return GFP_KERNEL; +} +EXPORT_SYMBOL_GPL(rpc_task_gfp_mask); + +bool rpc_task_set_rpc_status(struct rpc_task *task, int rpc_status) +{ + if (cmpxchg(&task->tk_rpc_status, 0, rpc_status) == 0) + return true; + return false; +} + +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 @@ -64,118 +94,121 @@ struct workqueue_struct *rpciod_workqueue; static void __rpc_disable_timer(struct rpc_wait_queue *queue, struct rpc_task *task) { - if (task->tk_timeout == 0) + if (list_empty(&task->u.tk_wait.timer_list)) return; - dprintk("RPC: %5u disabling timer\n", task->tk_pid); task->tk_timeout = 0; list_del(&task->u.tk_wait.timer_list); if (list_empty(&queue->timer_list.list)) - del_timer(&queue->timer_list.timer); + 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; - mod_timer(&queue->timer_list.timer, 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) +__rpc_add_timer(struct rpc_wait_queue *queue, struct rpc_task *task, + unsigned long timeout) { - if (!task->tk_timeout) - return; - - dprintk("RPC: %5u setting alarm for %lu ms\n", - task->tk_pid, task->tk_timeout * 1000 / HZ); - - task->u.tk_wait.expires = jiffies + task->tk_timeout; - if (list_empty(&queue->timer_list.list) || time_before(task->u.tk_wait.expires, queue->timer_list.expires)) - rpc_set_queue_timer(queue, task->u.tk_wait.expires); + 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_rotate_queue_owner(struct rpc_wait_queue *queue) -{ - struct list_head *q = &queue->tasks[queue->priority]; - struct rpc_task *task; - - if (!list_empty(q)) { - task = list_first_entry(q, struct rpc_task, u.tk_wait.list); - if (task->tk_owner == queue->owner) - list_move_tail(&task->u.tk_wait.list, q); - } -} - static void rpc_set_waitqueue_priority(struct rpc_wait_queue *queue, int priority) { if (queue->priority != priority) { - /* Fairness: rotate the list when changing priority */ - rpc_rotate_queue_owner(queue); queue->priority = priority; + queue->nr = 1U << priority; } } -static void rpc_set_waitqueue_owner(struct rpc_wait_queue *queue, pid_t pid) -{ - queue->owner = pid; - queue->nr = RPC_BATCH_COUNT; -} - static void rpc_reset_waitqueue_priority(struct rpc_wait_queue *queue) { rpc_set_waitqueue_priority(queue, queue->maxpriority); - rpc_set_waitqueue_owner(queue, 0); } /* - * Add new request to a priority queue. + * Add a request to a queue list */ -static void __rpc_add_wait_queue_priority(struct rpc_wait_queue *queue, - struct rpc_task *task, - unsigned char queue_priority) +static void +__rpc_list_enqueue_task(struct list_head *q, struct rpc_task *task) { - struct list_head *q; struct rpc_task *t; - INIT_LIST_HEAD(&task->u.tk_wait.links); - if (unlikely(queue_priority > queue->maxpriority)) - queue_priority = queue->maxpriority; - if (queue_priority > queue->priority) - rpc_set_waitqueue_priority(queue, queue_priority); - q = &queue->tasks[queue_priority]; list_for_each_entry(t, q, u.tk_wait.list) { if (t->tk_owner == task->tk_owner) { - list_add_tail(&task->u.tk_wait.list, &t->u.tk_wait.links); + 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) { - WARN_ON_ONCE(RPC_IS_QUEUED(task)); - if (RPC_IS_QUEUED(task)) - return; - + 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; @@ -183,9 +216,6 @@ static void __rpc_add_wait_queue(struct rpc_wait_queue *queue, /* barrier matches the read in rpc_wake_up_task_queue_locked() */ smp_wmb(); rpc_set_queued(task); - - dprintk("RPC: %5u added to queue %p \"%s\"\n", - task->tk_pid, queue, rpc_qname(queue)); } /* @@ -193,13 +223,7 @@ static void __rpc_add_wait_queue(struct rpc_wait_queue *queue, */ static void __rpc_remove_wait_queue_priority(struct rpc_task *task) { - struct rpc_task *t; - - if (!list_empty(&task->u.tk_wait.links)) { - t = list_entry(task->u.tk_wait.links.next, struct rpc_task, u.tk_wait.list); - list_move(&t->u.tk_wait.list, &task->u.tk_wait.list); - list_splice_init(&task->u.tk_wait.links, &t->u.tk_wait.links); - } + __rpc_list_dequeue_task(task); } /* @@ -211,10 +235,9 @@ static void __rpc_remove_wait_queue(struct rpc_wait_queue *queue, struct rpc_tas __rpc_disable_timer(queue, task); if (RPC_IS_PRIORITY(queue)) __rpc_remove_wait_queue_priority(task); - list_del(&task->u.tk_wait.list); + else + list_del(&task->u.tk_wait.list); queue->qlen--; - dprintk("RPC: %5u removed from queue %p \"%s\"\n", - task->tk_pid, queue, rpc_qname(queue)); } static void __rpc_init_priority_wait_queue(struct rpc_wait_queue *queue, const char *qname, unsigned char nr_queues) @@ -227,7 +250,8 @@ static void __rpc_init_priority_wait_queue(struct rpc_wait_queue *queue, const c queue->maxpriority = nr_queues - 1; rpc_reset_waitqueue_priority(queue); queue->qlen = 0; - setup_timer(&queue->timer_list.timer, __rpc_queue_timer_fn, (unsigned long)queue); + 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); } @@ -246,24 +270,32 @@ EXPORT_SYMBOL_GPL(rpc_init_wait_queue); void rpc_destroy_wait_queue(struct rpc_wait_queue *queue) { - del_timer_sync(&queue->timer_list.timer); + cancel_delayed_work_sync(&queue->timer_list.dwork); } EXPORT_SYMBOL_GPL(rpc_destroy_wait_queue); -static int rpc_wait_bit_killable(void *word) +static int rpc_wait_bit_killable(struct wait_bit_key *key, int mode) { - if (fatal_signal_pending(current)) + schedule(); + if (signal_pending_state(mode, current)) return -ERESTARTSYS; - freezable_schedule_unsafe(); return 0; } -#ifdef RPC_DEBUG +#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; + struct rpc_clnt *clnt = task->tk_client; - task->tk_pid = atomic_inc_return(&rpc_pid); + /* Might be a task carrying a reverse-direction operation */ + if (!clnt) { + static atomic_t rpc_pid; + + task->tk_pid = atomic_inc_return(&rpc_pid); + return; + } + + task->tk_pid = atomic_inc_return(&clnt->cl_pid); } #else static inline void rpc_task_set_debuginfo(struct rpc_task *task) @@ -273,10 +305,9 @@ static inline void rpc_task_set_debuginfo(struct rpc_task *task) static void rpc_set_active(struct rpc_task *task) { - trace_rpc_task_begin(task->tk_client, task, NULL); - rpc_task_set_debuginfo(task); set_bit(RPC_TASK_ACTIVE, &task->tk_runstate); + trace_rpc_task_begin(task, NULL); } /* @@ -291,7 +322,7 @@ static int rpc_complete_task(struct rpc_task *task) unsigned long flags; int ret; - trace_rpc_task_complete(task->tk_client, task, NULL); + trace_rpc_task_complete(task, NULL); spin_lock_irqsave(&wq->lock, flags); clear_bit(RPC_TASK_ACTIVE, &task->tk_runstate); @@ -309,14 +340,12 @@ static int rpc_complete_task(struct rpc_task *task) * 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, int (*action)(void *)) +int rpc_wait_for_completion_task(struct rpc_task *task) { - if (action == NULL) - action = rpc_wait_bit_killable; return out_of_line_wait_on_bit(&task->tk_runstate, RPC_TASK_ACTIVE, - action, TASK_KILLABLE); + rpc_wait_bit_killable, TASK_KILLABLE|TASK_FREEZABLE_UNSAFE); } -EXPORT_SYMBOL_GPL(__rpc_wait_for_completion_task); +EXPORT_SYMBOL_GPL(rpc_wait_for_completion_task); /* * Make an RPC task runnable. @@ -329,7 +358,8 @@ EXPORT_SYMBOL_GPL(__rpc_wait_for_completion_task); * lockless RPC_IS_QUEUED() test) before we've had a chance to test * the RPC_TASK_RUNNING flag. */ -static void rpc_make_runnable(struct rpc_task *task) +static void rpc_make_runnable(struct workqueue_struct *wq, + struct rpc_task *task) { bool need_wakeup = !rpc_test_and_set_running(task); @@ -338,9 +368,11 @@ static void rpc_make_runnable(struct rpc_task *task) return; if (RPC_IS_ASYNC(task)) { INIT_WORK(&task->u.tk_work, rpc_async_schedule); - queue_work(rpciod_workqueue, &task->u.tk_work); - } else + queue_work(wq, &task->u.tk_work); + } else { + smp_mb__after_atomic(); wake_up_bit(&task->tk_runstate, RPC_TASK_QUEUED); + } } /* @@ -349,100 +381,175 @@ static void rpc_make_runnable(struct rpc_task *task) * NB: An RPC task will only receive interrupt-driven events as long * as it's on a wait queue. */ -static void __rpc_sleep_on_priority(struct rpc_wait_queue *q, +static void __rpc_do_sleep_on_priority(struct rpc_wait_queue *q, struct rpc_task *task, - rpc_action action, unsigned char queue_priority) { - dprintk("RPC: %5u sleep_on(queue \"%s\" time %lu)\n", - task->tk_pid, rpc_qname(q), jiffies); - - trace_rpc_task_sleep(task->tk_client, task, q); + 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); +} - WARN_ON_ONCE(task->tk_callback != NULL); - task->tk_callback = action; - __rpc_add_timer(q, task); +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; } -void rpc_sleep_on(struct rpc_wait_queue *q, struct rpc_task *task, - rpc_action action) +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 */ - WARN_ON_ONCE(!RPC_IS_ACTIVATED(task)); - if (!RPC_IS_ACTIVATED(task)) { + if (WARN_ON_ONCE(!RPC_IS_ACTIVATED(task))) { task->tk_status = -EIO; rpc_put_task_async(task); - return; + 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_bh(&q->lock); - __rpc_sleep_on_priority(q, task, action, task->tk_priority); - spin_unlock_bh(&q->lock); + 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, - rpc_action action, int priority) + int priority) { - /* We shouldn't ever put an inactive task to sleep */ - WARN_ON_ONCE(!RPC_IS_ACTIVATED(task)); - if (!RPC_IS_ACTIVATED(task)) { - task->tk_status = -EIO; - rpc_put_task_async(task); + if (!rpc_sleep_check_activated(task)) return; - } + WARN_ON_ONCE(task->tk_timeout != 0); + priority -= RPC_PRIORITY_LOW; /* * Protect the queue operations. */ - spin_lock_bh(&q->lock); - __rpc_sleep_on_priority(q, task, action, priority - RPC_PRIORITY_LOW); - spin_unlock_bh(&q->lock); + 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 - wake up a single rpc_task + * __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(struct rpc_wait_queue *queue, struct rpc_task *task) +static void __rpc_do_wake_up_task_on_wq(struct workqueue_struct *wq, + struct rpc_wait_queue *queue, + struct rpc_task *task) { - dprintk("RPC: %5u __rpc_wake_up_task (now %lu)\n", - task->tk_pid, jiffies); - /* 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->tk_client, task, queue); + trace_rpc_task_wakeup(task, queue); __rpc_remove_wait_queue(queue, task); - rpc_make_runnable(task); - - dprintk("RPC: __rpc_wake_up_task done\n"); + rpc_make_runnable(wq, task); } /* * 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) +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) - __rpc_do_wake_up_task(queue, task); + 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); } /* @@ -450,12 +557,48 @@ static void rpc_wake_up_task_queue_locked(struct rpc_wait_queue *queue, struct r */ void rpc_wake_up_queued_task(struct rpc_wait_queue *queue, struct rpc_task *task) { - spin_lock_bh(&queue->lock); + if (!RPC_IS_QUEUED(task)) + return; + spin_lock(&queue->lock); rpc_wake_up_task_queue_locked(queue, task); - spin_unlock_bh(&queue->lock); + 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. */ @@ -465,20 +608,22 @@ static struct rpc_task *__rpc_find_next_queued_priority(struct rpc_wait_queue *q struct rpc_task *task; /* + * Service the privileged queue. + */ + q = &queue->tasks[RPC_NR_PRIORITY - 1]; + if (queue->maxpriority > RPC_PRIORITY_PRIVILEGED && !list_empty(q)) { + task = list_first_entry(q, struct rpc_task, u.tk_wait.list); + goto out; + } + + /* * Service a batch of tasks from a single owner. */ q = &queue->tasks[queue->priority]; - if (!list_empty(q)) { - task = list_entry(q->next, struct rpc_task, u.tk_wait.list); - if (queue->owner == task->tk_owner) { - if (--queue->nr) - goto out; - list_move_tail(&task->u.tk_wait.list, q); - } - /* - * Check if we need to switch queues. - */ - goto new_owner; + if (!list_empty(q) && queue->nr) { + queue->nr--; + task = list_first_entry(q, struct rpc_task, u.tk_wait.list); + goto out; } /* @@ -490,7 +635,7 @@ static struct rpc_task *__rpc_find_next_queued_priority(struct rpc_wait_queue *q else q = q - 1; if (!list_empty(q)) { - task = list_entry(q->next, struct rpc_task, u.tk_wait.list); + task = list_first_entry(q, struct rpc_task, u.tk_wait.list); goto new_queue; } } while (q != &queue->tasks[queue->priority]); @@ -500,8 +645,6 @@ static struct rpc_task *__rpc_find_next_queued_priority(struct rpc_wait_queue *q new_queue: rpc_set_waitqueue_priority(queue, (unsigned int)(q - &queue->tasks[0])); -new_owner: - rpc_set_waitqueue_owner(queue, task->tk_owner); out: return task; } @@ -518,25 +661,30 @@ static struct rpc_task *__rpc_find_next_queued(struct rpc_wait_queue *queue) /* * Wake up the first task on the wait queue. */ -struct rpc_task *rpc_wake_up_first(struct rpc_wait_queue *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; - dprintk("RPC: wake_up_first(%p \"%s\")\n", - queue, rpc_qname(queue)); - spin_lock_bh(&queue->lock); + spin_lock(&queue->lock); task = __rpc_find_next_queued(queue); - if (task != NULL) { - if (func(task, data)) - rpc_wake_up_task_queue_locked(queue, task); - else - task = NULL; - } - spin_unlock_bh(&queue->lock); + 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) @@ -554,6 +702,23 @@ struct rpc_task *rpc_wake_up_next(struct rpc_wait_queue *queue) 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 * @@ -561,25 +726,28 @@ EXPORT_SYMBOL_GPL(rpc_wake_up_next); */ void rpc_wake_up(struct rpc_wait_queue *queue) { - struct list_head *head; + 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; - spin_lock_bh(&queue->lock); - head = &queue->tasks[queue->maxpriority]; for (;;) { - while (!list_empty(head)) { - struct rpc_task *task; - task = list_first_entry(head, - struct rpc_task, - u.tk_wait.list); - rpc_wake_up_task_queue_locked(queue, task); - } - if (head == &queue->tasks[0]) + task = __rpc_find_next_queued(queue); + if (task == NULL) break; - head--; + rpc_wake_up_task_queue_set_status_locked(queue, task, status); } - spin_unlock_bh(&queue->lock); } -EXPORT_SYMBOL_GPL(rpc_wake_up); /** * rpc_wake_up_status - wake up all rpc_tasks and set their status value. @@ -590,39 +758,26 @@ EXPORT_SYMBOL_GPL(rpc_wake_up); */ void rpc_wake_up_status(struct rpc_wait_queue *queue, int status) { - struct list_head *head; - - spin_lock_bh(&queue->lock); - head = &queue->tasks[queue->maxpriority]; - for (;;) { - while (!list_empty(head)) { - struct rpc_task *task; - task = list_first_entry(head, - struct rpc_task, - u.tk_wait.list); - task->tk_status = status; - rpc_wake_up_task_queue_locked(queue, task); - } - if (head == &queue->tasks[0]) - break; - head--; - } - spin_unlock_bh(&queue->lock); + 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(unsigned long ptr) +static void __rpc_queue_timer_fn(struct work_struct *work) { - struct rpc_wait_queue *queue = (struct rpc_wait_queue *)ptr; + 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->u.tk_wait.expires; + timeo = task->tk_timeout; if (time_after_eq(now, timeo)) { - dprintk("RPC: %5u timeout\n", task->tk_pid); + trace_rpc_task_timeout(task, task->tk_action); task->tk_status = -ETIMEDOUT; rpc_wake_up_task_queue_locked(queue, task); continue; @@ -637,7 +792,8 @@ static void __rpc_queue_timer_fn(unsigned long ptr) static void __rpc_atrun(struct rpc_task *task) { - task->tk_status = 0; + if (task->tk_status == -ETIMEDOUT) + task->tk_status = 0; } /* @@ -645,8 +801,7 @@ static void __rpc_atrun(struct rpc_task *task) */ void rpc_delay(struct rpc_task *task, unsigned long delay) { - task->tk_timeout = delay; - rpc_sleep_on(&delay_queue, task, __rpc_atrun); + rpc_sleep_on_timeout(&delay_queue, task, __rpc_atrun, jiffies + delay); } EXPORT_SYMBOL_GPL(rpc_delay); @@ -664,7 +819,6 @@ 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(); @@ -674,8 +828,7 @@ static void rpc_reset_task_statistics(struct rpc_task *task) { task->tk_timeouts = 0; - task->tk_flags &= ~(RPC_CALL_MAJORSEEN|RPC_TASK_KILLED|RPC_TASK_SENT); - + task->tk_flags &= ~(RPC_CALL_MAJORSEEN|RPC_TASK_SENT); rpc_init_task_statistics(task); } @@ -684,11 +837,16 @@ rpc_reset_task_statistics(struct rpc_task *task) */ 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) { + trace_rpc_task_call_done(task, task->tk_ops->rpc_call_done); task->tk_ops->rpc_call_done(task, task->tk_calldata); if (task->tk_action != NULL) { - WARN_ON(RPC_ASSASSINATED(task)); /* Always release the RPC slot and buffer memory */ xprt_release(task); rpc_reset_task_statistics(task); @@ -696,12 +854,37 @@ void rpc_exit_task(struct rpc_task *task) } } +void rpc_signal_task(struct rpc_task *task) +{ + struct rpc_wait_queue *queue; + + if (!RPC_IS_ACTIVATED(task)) + return; + + if (!rpc_task_set_rpc_status(task, -ERESTARTSYS)) + return; + trace_rpc_task_signalled(task, task->tk_action); + queue = READ_ONCE(task->tk_waitqueue); + if (queue) + rpc_wake_up_queued_task(queue, task); +} + +void rpc_task_try_cancel(struct rpc_task *task, int error) +{ + struct rpc_wait_queue *queue; + + if (!rpc_task_set_rpc_status(task, error)) + return; + queue = READ_ONCE(task->tk_waitqueue); + if (queue) + rpc_wake_up_queued_task(queue, task); +} + void rpc_exit(struct rpc_task *task, int status) { task->tk_status = status; task->tk_action = rpc_exit_task; - if (RPC_IS_QUEUED(task)) - rpc_wake_up_queued_task(task->tk_waitqueue, task); + rpc_wake_up_queued_task(task->tk_waitqueue, task); } EXPORT_SYMBOL_GPL(rpc_exit); @@ -711,6 +894,15 @@ void rpc_release_calldata(const struct rpc_call_ops *ops, void *calldata) ops->rpc_release(calldata); } +static bool xprt_needs_memalloc(struct rpc_xprt *xprt, struct rpc_task *tk) +{ + if (!xprt) + return false; + if (!atomic_read(&xprt->swapper)) + return false; + return test_bit(XPRT_LOCKED, &xprt->state) && xprt->snd_task == tk; +} + /* * This is the RPC `scheduler' (or rather, the finite state machine). */ @@ -719,9 +911,7 @@ static void __rpc_execute(struct rpc_task *task) struct rpc_wait_queue *queue; int task_is_async = RPC_IS_ASYNC(task); int status = 0; - - dprintk("RPC: %5u __rpc_execute flags=0x%x\n", - task->tk_pid, task->tk_flags); + unsigned long pflags = current->flags; WARN_ON_ONCE(RPC_IS_QUEUED(task)); if (RPC_IS_QUEUED(task)) @@ -731,29 +921,39 @@ static void __rpc_execute(struct rpc_task *task) void (*do_action)(struct rpc_task *); /* - * Execute any pending callback first. + * Perform the next FSM step or a pending callback. + * + * tk_action may be NULL if the task has been killed. */ - do_action = task->tk_callback; - task->tk_callback = NULL; - if (do_action == NULL) { - /* - * Perform the next FSM step. - * 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 (do_action == NULL) - break; + do_action = task->tk_action; + /* Tasks with an RPC error status should exit */ + if (do_action && do_action != rpc_exit_task && + (status = READ_ONCE(task->tk_rpc_status)) != 0) { + task->tk_status = status; + do_action = rpc_exit_task; } - trace_rpc_task_run_action(task->tk_client, task, task->tk_action); + /* Callbacks override all actions */ + if (task->tk_callback) { + do_action = task->tk_callback; + task->tk_callback = NULL; + } + if (!do_action) + break; + if (RPC_IS_SWAPPER(task) || + xprt_needs_memalloc(task->tk_xprt, task)) + current->flags |= PF_MEMALLOC; + + 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)) + if (!RPC_IS_QUEUED(task)) { + cond_resched(); continue; + } + /* * The queue->lock protects against races with * rpc_make_runnable(). @@ -764,39 +964,43 @@ static void __rpc_execute(struct rpc_task *task) * rpc_task pointer may still be dereferenced. */ queue = task->tk_waitqueue; - spin_lock_bh(&queue->lock); + spin_lock(&queue->lock); if (!RPC_IS_QUEUED(task)) { - spin_unlock_bh(&queue->lock); + spin_unlock(&queue->lock); + continue; + } + /* Wake up any task that has an exit status */ + if (READ_ONCE(task->tk_rpc_status) != 0) { + rpc_wake_up_task_queue_locked(queue, task); + spin_unlock(&queue->lock); continue; } rpc_clear_running(task); - spin_unlock_bh(&queue->lock); + spin_unlock(&queue->lock); if (task_is_async) - return; + goto out; /* sync task: sleep here */ - dprintk("RPC: %5u sync task going to sleep\n", task->tk_pid); + 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 == -ERESTARTSYS) { + TASK_KILLABLE|TASK_FREEZABLE); + 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. */ - dprintk("RPC: %5u got signal\n", task->tk_pid); - task->tk_flags |= RPC_TASK_KILLED; - rpc_exit(task, -ERESTARTSYS); + rpc_signal_task(task); } - dprintk("RPC: %5u sync task resuming\n", task->tk_pid); + trace_rpc_task_sync_wake(task, task->tk_action); } - dprintk("RPC: %5u return %d, status %d\n", task->tk_pid, status, - task->tk_status); /* Release all resources associated with the task */ rpc_release_task(task); +out: + current_restore_flags(pflags, PF_MEMALLOC); } /* @@ -813,83 +1017,83 @@ void rpc_execute(struct rpc_task *task) bool is_async = RPC_IS_ASYNC(task); rpc_set_active(task); - rpc_make_runnable(task); - if (!is_async) + 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) { - current->flags |= PF_FSTRANS; + unsigned int pflags = memalloc_nofs_save(); + __rpc_execute(container_of(work, struct rpc_task, u.tk_work)); - current->flags &= ~PF_FSTRANS; + memalloc_nofs_restore(pflags); } /** - * rpc_malloc - allocate an RPC buffer - * @task: RPC task that will use this buffer - * @size: requested byte size + * 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 NULL if the request cannot be serviced immediately. - * The caller can arrange to sleep in a way that is safe for rpciod. + * 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. - * - * In order to avoid memory starvation triggering more writebacks of - * NFS requests, we avoid using GFP_KERNEL. */ -void *rpc_malloc(struct rpc_task *task, size_t size) +int rpc_malloc(struct rpc_task *task) { + struct rpc_rqst *rqst = task->tk_rqstp; + size_t size = rqst->rq_callsize + rqst->rq_rcvsize; struct rpc_buffer *buf; - gfp_t gfp = GFP_NOWAIT; - - if (RPC_IS_SWAPPER(task)) - gfp |= __GFP_MEMALLOC; + gfp_t gfp = rpc_task_gfp_mask(); size += sizeof(struct rpc_buffer); - if (size <= RPC_BUFFER_MAXSIZE) - buf = mempool_alloc(rpc_buffer_mempool, gfp); - else + if (size <= RPC_BUFFER_MAXSIZE) { + buf = kmem_cache_alloc(rpc_buffer_slabp, gfp); + /* Reach for the mempool if dynamic allocation fails */ + if (!buf && RPC_IS_ASYNC(task)) + buf = mempool_alloc(rpc_buffer_mempool, GFP_NOWAIT); + } else buf = kmalloc(size, gfp); if (!buf) - return NULL; + return -ENOMEM; buf->len = size; - dprintk("RPC: %5u allocated buffer of size %zu at %p\n", - task->tk_pid, size, buf); - return &buf->data; + rqst->rq_buffer = buf->data; + rqst->rq_rbuffer = (char *)rqst->rq_buffer + rqst->rq_callsize; + return 0; } -EXPORT_SYMBOL_GPL(rpc_malloc); /** - * rpc_free - free buffer allocated via rpc_malloc - * @buffer: buffer to free + * rpc_free - free RPC buffer resources allocated via rpc_malloc + * @task: RPC task * */ -void rpc_free(void *buffer) +void rpc_free(struct rpc_task *task) { + void *buffer = task->tk_rqstp->rq_buffer; size_t size; struct rpc_buffer *buf; - if (!buffer) - return; - buf = container_of(buffer, struct rpc_buffer, data); size = buf->len; - dprintk("RPC: freeing buffer of size %zu at %p\n", - size, buf); - if (size <= RPC_BUFFER_MAXSIZE) mempool_free(buf, rpc_buffer_mempool); else kfree(buf); } -EXPORT_SYMBOL_GPL(rpc_free); /* * Creation and deletion of RPC task structures @@ -909,19 +1113,25 @@ static void rpc_init_task(struct rpc_task *task, const struct rpc_task_setup *ta /* Initialize workqueue for async tasks */ task->tk_workqueue = task_setup_data->workqueue; + task->tk_xprt = rpc_task_get_xprt(task_setup_data->rpc_client, + xprt_get(task_setup_data->rpc_xprt)); + + task->tk_op_cred = get_rpccred(task_setup_data->rpc_op_cred); + if (task->tk_ops->rpc_call_prepare != NULL) task->tk_action = rpc_prepare_task; rpc_init_task_statistics(task); - - dprintk("RPC: new task initialized, procpid %u\n", - task_pid_nr(current)); } -static struct rpc_task * -rpc_alloc_task(void) +static struct rpc_task *rpc_alloc_task(void) { - return (struct rpc_task *)mempool_alloc(rpc_task_mempool, GFP_NOIO); + struct rpc_task *task; + + task = kmem_cache_alloc(rpc_task_slabp, rpc_task_gfp_mask()); + if (task) + return task; + return mempool_alloc(rpc_task_mempool, GFP_NOWAIT); } /* @@ -936,7 +1146,7 @@ struct rpc_task *rpc_new_task(const struct rpc_task_setup *setup_data) task = rpc_alloc_task(); if (task == NULL) { rpc_release_calldata(setup_data->callback_ops, - setup_data->callback_data); + setup_data->callback_data); return ERR_PTR(-ENOMEM); } flags = RPC_TASK_DYNAMIC; @@ -944,7 +1154,6 @@ struct rpc_task *rpc_new_task(const struct rpc_task_setup *setup_data) rpc_init_task(task, setup_data); task->tk_flags |= flags; - dprintk("RPC: allocated task %p\n", task); return task; } @@ -971,24 +1180,27 @@ static void rpc_free_task(struct rpc_task *task) { unsigned short tk_flags = task->tk_flags; + put_rpccred(task->tk_op_cred); rpc_release_calldata(task->tk_ops, task->tk_calldata); - if (tk_flags & RPC_TASK_DYNAMIC) { - dprintk("RPC: %5u freeing task\n", task->tk_pid); + if (tk_flags & RPC_TASK_DYNAMIC) mempool_free(task, rpc_task_mempool); - } } static void rpc_async_release(struct work_struct *work) { + unsigned int pflags = memalloc_nofs_save(); + rpc_free_task(container_of(work, struct rpc_task, u.tk_work)); + memalloc_nofs_restore(pflags); } static void rpc_release_resources_task(struct rpc_task *task) { xprt_release(task); if (task->tk_msg.rpc_cred) { - put_rpccred(task->tk_msg.rpc_cred); + if (!(task->tk_flags & RPC_TASK_CRED_NOREF)) + put_cred(task->tk_msg.rpc_cred); task->tk_msg.rpc_cred = NULL; } rpc_task_release_client(task); @@ -1026,8 +1238,6 @@ EXPORT_SYMBOL_GPL(rpc_put_task_async); static void rpc_release_task(struct rpc_task *task) { - dprintk("RPC: %5u release task\n", task->tk_pid); - WARN_ON_ONCE(RPC_IS_QUEUED(task)); rpc_release_resources_task(task); @@ -1068,10 +1278,21 @@ static int rpciod_start(void) /* * Create the rpciod thread and wait for it to start. */ - dprintk("RPC: creating workqueue rpciod\n"); - wq = alloc_workqueue("rpciod", WQ_MEM_RECLAIM, 1); + wq = alloc_workqueue("rpciod", WQ_MEM_RECLAIM | WQ_UNBOUND, 0); + if (!wq) + goto out_failed; rpciod_workqueue = wq; - return rpciod_workqueue != NULL; + wq = alloc_workqueue("xprtiod", WQ_UNBOUND | WQ_MEM_RECLAIM, 0); + if (!wq) + goto free_rpciod; + xprtiod_workqueue = wq; + return 1; +free_rpciod: + wq = rpciod_workqueue; + rpciod_workqueue = NULL; + destroy_workqueue(wq); +out_failed: + return 0; } static void rpciod_stop(void) @@ -1080,25 +1301,23 @@ static void rpciod_stop(void) if (rpciod_workqueue == NULL) return; - dprintk("RPC: destroying workqueue rpciod\n"); wq = rpciod_workqueue; rpciod_workqueue = NULL; destroy_workqueue(wq); + wq = xprtiod_workqueue; + xprtiod_workqueue = NULL; + destroy_workqueue(wq); } void rpc_destroy_mempool(void) { rpciod_stop(); - if (rpc_buffer_mempool) - mempool_destroy(rpc_buffer_mempool); - if (rpc_task_mempool) - mempool_destroy(rpc_task_mempool); - if (rpc_task_slabp) - kmem_cache_destroy(rpc_task_slabp); - if (rpc_buffer_slabp) - kmem_cache_destroy(rpc_buffer_slabp); + mempool_destroy(rpc_buffer_mempool); + mempool_destroy(rpc_task_mempool); + kmem_cache_destroy(rpc_task_slabp); + kmem_cache_destroy(rpc_buffer_slabp); rpc_destroy_wait_queue(&delay_queue); } |