elevator.c

Linux块设备驱动源码

		BUG_ON(q->ordered == QUEUE_ORDERED_NONE);

		if (q->ordered == QUEUE_ORDERED_FLUSH &&
		    !blk_barrier_preflush(rq))
			rq = blk_start_pre_flush(q, rq);
	}

	return rq;
}

struct request *elv_next_request(request_queue_t *q)
{
	struct request *rq;
	int ret;

	while ((rq = __elv_next_request(q)) != NULL) {
		/*
		 * just mark as started even if we don't start it, a request
		 * that has been delayed should not be passed by new incoming
		 * requests
		 */
		rq->flags |= REQ_STARTED;

		if (rq == q->last_merge)
			q->last_merge = NULL;

		if ((rq->flags & REQ_DONTPREP) || !q->prep_rq_fn)
			break;

		ret = q->prep_rq_fn(q, rq);
		if (ret == BLKPREP_OK) {
			break;
		} else if (ret == BLKPREP_DEFER) {
			/*
			 * the request may have been (partially) prepped.
			 * we need to keep this request in the front to
			 * avoid resource deadlock.  turn on softbarrier.
			 */
			rq->flags |= REQ_SOFTBARRIER;
			rq = NULL;
			break;
		} else if (ret == BLKPREP_KILL) {
			int nr_bytes = rq->hard_nr_sectors << 9;

			if (!nr_bytes)
				nr_bytes = rq->data_len;

			blkdev_dequeue_request(rq);
			rq->flags |= REQ_QUIET;
			end_that_request_chunk(rq, 0, nr_bytes);
			end_that_request_last(rq);
		} else {
			printk(KERN_ERR "%s: bad return=%d\n", __FUNCTION__,
								ret);
			break;
		}
	}

	return rq;
}

void elv_remove_request(request_queue_t *q, struct request *rq)
{
	elevator_t *e = q->elevator;

	/*
	 * the time frame between a request being removed from the lists
	 * and to it is freed is accounted as io that is in progress at
	 * the driver side. note that we only account requests that the
	 * driver has seen (REQ_STARTED set), to avoid false accounting
	 * for request-request merges
	 */
	if (blk_account_rq(rq))
		q->in_flight++;

	/*
	 * the main clearing point for q->last_merge is on retrieval of
	 * request by driver (it calls elv_next_request()), but it _can_
	 * also happen here if a request is added to the queue but later
	 * deleted without ever being given to driver (merged with another
	 * request).
	 */
	if (rq == q->last_merge)
		q->last_merge = NULL;

	if (e->ops->elevator_remove_req_fn)
		e->ops->elevator_remove_req_fn(q, rq);
}

int elv_queue_empty(request_queue_t *q)
{
	elevator_t *e = q->elevator;

	if (e->ops->elevator_queue_empty_fn)
		return e->ops->elevator_queue_empty_fn(q);

	return list_empty(&q->queue_head);
}

struct request *elv_latter_request(request_queue_t *q, struct request *rq)
{
	struct list_head *next;

	elevator_t *e = q->elevator;

	if (e->ops->elevator_latter_req_fn)
		return e->ops->elevator_latter_req_fn(q, rq);

	next = rq->queuelist.next;
	if (next != &q->queue_head && next != &rq->queuelist)
		return list_entry_rq(next);

	return NULL;
}

struct request *elv_former_request(request_queue_t *q, struct request *rq)
{
	struct list_head *prev;

	elevator_t *e = q->elevator;

	if (e->ops->elevator_former_req_fn)
		return e->ops->elevator_former_req_fn(q, rq);

	prev = rq->queuelist.prev;
	if (prev != &q->queue_head && prev != &rq->queuelist)
		return list_entry_rq(prev);

	return NULL;
}

int elv_set_request(request_queue_t *q, struct request *rq, struct bio *bio,
		    int gfp_mask)
{
	elevator_t *e = q->elevator;

	if (e->ops->elevator_set_req_fn)
		return e->ops->elevator_set_req_fn(q, rq, bio, gfp_mask);

	rq->elevator_private = NULL;
	return 0;
}

void elv_put_request(request_queue_t *q, struct request *rq)
{
	elevator_t *e = q->elevator;

	if (e->ops->elevator_put_req_fn)
		e->ops->elevator_put_req_fn(q, rq);
}

int elv_may_queue(request_queue_t *q, int rw, struct bio *bio)
{
	elevator_t *e = q->elevator;

	if (e->ops->elevator_may_queue_fn)
		return e->ops->elevator_may_queue_fn(q, rw, bio);

	return ELV_MQUEUE_MAY;
}

void elv_completed_request(request_queue_t *q, struct request *rq)
{
	elevator_t *e = q->elevator;

	/*
	 * request is released from the driver, io must be done
	 */
	if (blk_account_rq(rq))
		q->in_flight--;

	if (e->ops->elevator_completed_req_fn)
		e->ops->elevator_completed_req_fn(q, rq);
}

int elv_register_queue(struct request_queue *q)
{
	elevator_t *e = q->elevator;

	e->kobj.parent = kobject_get(&q->kobj);
	if (!e->kobj.parent)
		return -EBUSY;

	snprintf(e->kobj.name, KOBJ_NAME_LEN, "%s", "iosched");
	e->kobj.ktype = e->elevator_type->elevator_ktype;

	return kobject_register(&e->kobj);
}

void elv_unregister_queue(struct request_queue *q)
{
	if (q) {
		elevator_t *e = q->elevator;
		kobject_unregister(&e->kobj);
		kobject_put(&q->kobj);
	}
}

int elv_register(struct elevator_type *e)
{
	if (elevator_find(e->elevator_name))
		BUG();

	spin_lock_irq(&elv_list_lock);
	list_add_tail(&e->list, &elv_list);
	spin_unlock_irq(&elv_list_lock);

	printk(KERN_INFO "io scheduler %s registered", e->elevator_name);
	if (!strcmp(e->elevator_name, chosen_elevator))
		printk(" (default)");
	printk("\n");
	return 0;
}
EXPORT_SYMBOL_GPL(elv_register);

void elv_unregister(struct elevator_type *e)
{
	spin_lock_irq(&elv_list_lock);
	list_del_init(&e->list);
	spin_unlock_irq(&elv_list_lock);
}
EXPORT_SYMBOL_GPL(elv_unregister);

/*
 * switch to new_e io scheduler. be careful not to introduce deadlocks -
 * we don't free the old io scheduler, before we have allocated what we
 * need for the new one. this way we have a chance of going back to the old
 * one, if the new one fails init for some reason. we also do an intermediate
 * switch to noop to ensure safety with stack-allocated requests, since they
 * don't originate from the block layer allocator. noop is safe here, because
 * it never needs to touch the elevator itself for completion events. DRAIN
 * flags will make sure we don't touch it for additions either.
 */
static void elevator_switch(request_queue_t *q, struct elevator_type *new_e)
{
	elevator_t *e = kmalloc(sizeof(elevator_t), GFP_KERNEL);
	struct elevator_type *noop_elevator = NULL;
	elevator_t *old_elevator;

	if (!e)
		goto error;

	/*
	 * first step, drain requests from the block freelist
	 */
	blk_wait_queue_drained(q, 0);

	/*
	 * unregister old elevator data
	 */
	elv_unregister_queue(q);
	old_elevator = q->elevator;

	/*
 	 * next step, switch to noop since it uses no private rq structures
	 * and doesn't allocate any memory for anything. then wait for any
	 * non-fs requests in-flight
 	 */
	noop_elevator = elevator_get("noop");
	spin_lock_irq(q->queue_lock);
	elevator_attach(q, noop_elevator, e);
	spin_unlock_irq(q->queue_lock);

	blk_wait_queue_drained(q, 1);

	/*
	 * attach and start new elevator
	 */
	if (elevator_attach(q, new_e, e))
		goto fail;

	if (elv_register_queue(q))
		goto fail_register;

	/*
	 * finally exit old elevator and start queue again
	 */
	elevator_exit(old_elevator);
	blk_finish_queue_drain(q);
	elevator_put(noop_elevator);
	return;

fail_register:
	/*
	 * switch failed, exit the new io scheduler and reattach the old
	 * one again (along with re-adding the sysfs dir)
	 */
	elevator_exit(e);
fail:
	q->elevator = old_elevator;
	elv_register_queue(q);
	blk_finish_queue_drain(q);
error:
	if (noop_elevator)
		elevator_put(noop_elevator);
	elevator_put(new_e);
	printk(KERN_ERR "elevator: switch to %s failed\n",new_e->elevator_name);
}

ssize_t elv_iosched_store(request_queue_t *q, const char *name, size_t count)
{
	char elevator_name[ELV_NAME_MAX];
	struct elevator_type *e;

	memset(elevator_name, 0, sizeof(elevator_name));
	strncpy(elevator_name, name, sizeof(elevator_name));

	if (elevator_name[strlen(elevator_name) - 1] == '\n')
		elevator_name[strlen(elevator_name) - 1] = '\0';

	e = elevator_get(elevator_name);
	if (!e) {
		printk(KERN_ERR "elevator: type %s not found\n", elevator_name);
		return -EINVAL;
	}

	if (!strcmp(elevator_name, q->elevator->elevator_type->elevator_name))
		return count;

	elevator_switch(q, e);
	return count;
}

ssize_t elv_iosched_show(request_queue_t *q, char *name)
{
	elevator_t *e = q->elevator;
	struct elevator_type *elv = e->elevator_type;
	struct list_head *entry;
	int len = 0;

	spin_lock_irq(q->queue_lock);
	list_for_each(entry, &elv_list) {
		struct elevator_type *__e;

		__e = list_entry(entry, struct elevator_type, list);
		if (!strcmp(elv->elevator_name, __e->elevator_name))
			len += sprintf(name+len, "[%s] ", elv->elevator_name);
		else
			len += sprintf(name+len, "%s ", __e->elevator_name);
	}
	spin_unlock_irq(q->queue_lock);

	len += sprintf(len+name, "\n");
	return len;
}

EXPORT_SYMBOL(elv_add_request);
EXPORT_SYMBOL(__elv_add_request);
EXPORT_SYMBOL(elv_requeue_request);
EXPORT_SYMBOL(elv_next_request);
EXPORT_SYMBOL(elv_remove_request);
EXPORT_SYMBOL(elv_queue_empty);
EXPORT_SYMBOL(elv_completed_request);
EXPORT_SYMBOL(elevator_exit);
EXPORT_SYMBOL(elevator_init);