[deliverable/linux.git] / drivers / block / null_blk.c

#include <linux/module.h>
#include <linux/moduleparam.h>
#include <linux/sched.h>
#include <linux/fs.h>
#include <linux/blkdev.h>
#include <linux/init.h>
#include <linux/slab.h>
#include <linux/blk-mq.h>
#include <linux/hrtimer.h>

struct nullb_cmd {
	struct list_head list;
	struct llist_node ll_list;
	struct call_single_data csd;
	struct request *rq;
	struct bio *bio;
	unsigned int tag;
	struct nullb_queue *nq;
};

struct nullb_queue {
	unsigned long *tag_map;
	wait_queue_head_t wait;
	unsigned int queue_depth;

	struct nullb_cmd *cmds;
};

struct nullb {
	struct list_head list;
	unsigned int index;
	struct request_queue *q;
	struct gendisk *disk;
	struct hrtimer timer;
	unsigned int queue_depth;
	spinlock_t lock;

	struct nullb_queue *queues;
	unsigned int nr_queues;
};

static LIST_HEAD(nullb_list);
static struct mutex lock;
static int null_major;
static int nullb_indexes;

struct completion_queue {
	struct llist_head list;
	struct hrtimer timer;
};

/*
 * These are per-cpu for now, they will need to be configured by the
 * complete_queues parameter and appropriately mapped.
 */
static DEFINE_PER_CPU(struct completion_queue, completion_queues);

enum {
	NULL_IRQ_NONE		= 0,
	NULL_IRQ_SOFTIRQ	= 1,
	NULL_IRQ_TIMER		= 2,

	NULL_Q_BIO		= 0,
	NULL_Q_RQ		= 1,
	NULL_Q_MQ		= 2,
};

static int submit_queues;
module_param(submit_queues, int, S_IRUGO);
MODULE_PARM_DESC(submit_queues, "Number of submission queues");

static int home_node = NUMA_NO_NODE;
module_param(home_node, int, S_IRUGO);
MODULE_PARM_DESC(home_node, "Home node for the device");

static int queue_mode = NULL_Q_MQ;
module_param(queue_mode, int, S_IRUGO);
MODULE_PARM_DESC(use_mq, "Use blk-mq interface (0=bio,1=rq,2=multiqueue)");

static int gb = 250;
module_param(gb, int, S_IRUGO);
MODULE_PARM_DESC(gb, "Size in GB");

static int bs = 512;
module_param(bs, int, S_IRUGO);
MODULE_PARM_DESC(bs, "Block size (in bytes)");

static int nr_devices = 2;
module_param(nr_devices, int, S_IRUGO);
MODULE_PARM_DESC(nr_devices, "Number of devices to register");

static int irqmode = NULL_IRQ_SOFTIRQ;
module_param(irqmode, int, S_IRUGO);
MODULE_PARM_DESC(irqmode, "IRQ completion handler. 0-none, 1-softirq, 2-timer");

static int completion_nsec = 10000;
module_param(completion_nsec, int, S_IRUGO);
MODULE_PARM_DESC(completion_nsec, "Time in ns to complete a request in hardware. Default: 10,000ns");

static int hw_queue_depth = 64;
module_param(hw_queue_depth, int, S_IRUGO);
MODULE_PARM_DESC(hw_queue_depth, "Queue depth for each hardware queue. Default: 64");

static bool use_per_node_hctx = true;
module_param(use_per_node_hctx, bool, S_IRUGO);
MODULE_PARM_DESC(use_per_node_hctx, "Use per-node allocation for hardware context queues. Default: true");

static void put_tag(struct nullb_queue *nq, unsigned int tag)
{
	clear_bit_unlock(tag, nq->tag_map);

	if (waitqueue_active(&nq->wait))
		wake_up(&nq->wait);
}

static unsigned int get_tag(struct nullb_queue *nq)
{
	unsigned int tag;

	do {
		tag = find_first_zero_bit(nq->tag_map, nq->queue_depth);
		if (tag >= nq->queue_depth)
			return -1U;
	} while (test_and_set_bit_lock(tag, nq->tag_map));

	return tag;
}

static void free_cmd(struct nullb_cmd *cmd)
{
	put_tag(cmd->nq, cmd->tag);
}

static struct nullb_cmd *__alloc_cmd(struct nullb_queue *nq)
{
	struct nullb_cmd *cmd;
	unsigned int tag;

	tag = get_tag(nq);
	if (tag != -1U) {
		cmd = &nq->cmds[tag];
		cmd->tag = tag;
		cmd->nq = nq;
		return cmd;
	}

	return NULL;
}

static struct nullb_cmd *alloc_cmd(struct nullb_queue *nq, int can_wait)
{
	struct nullb_cmd *cmd;
	DEFINE_WAIT(wait);

	cmd = __alloc_cmd(nq);
	if (cmd || !can_wait)
		return cmd;

	do {
		prepare_to_wait(&nq->wait, &wait, TASK_UNINTERRUPTIBLE);
		cmd = __alloc_cmd(nq);
		if (cmd)
			break;

		io_schedule();
	} while (1);

	finish_wait(&nq->wait, &wait);
	return cmd;
}

static void end_cmd(struct nullb_cmd *cmd)
{
	if (cmd->rq) {
		if (queue_mode == NULL_Q_MQ)
			blk_mq_end_io(cmd->rq, 0);
		else {
			INIT_LIST_HEAD(&cmd->rq->queuelist);
			blk_end_request_all(cmd->rq, 0);
		}
	} else if (cmd->bio)
		bio_endio(cmd->bio, 0);

	if (queue_mode != NULL_Q_MQ)
		free_cmd(cmd);
}

static enum hrtimer_restart null_cmd_timer_expired(struct hrtimer *timer)
{
	struct completion_queue *cq;
	struct llist_node *entry;
	struct nullb_cmd *cmd;

	cq = &per_cpu(completion_queues, smp_processor_id());

	while ((entry = llist_del_all(&cq->list)) != NULL) {
		do {
			cmd = container_of(entry, struct nullb_cmd, ll_list);
			end_cmd(cmd);
			entry = entry->next;
		} while (entry);
	}

	return HRTIMER_NORESTART;
}

static void null_cmd_end_timer(struct nullb_cmd *cmd)
{
	struct completion_queue *cq = &per_cpu(completion_queues, get_cpu());

	cmd->ll_list.next = NULL;
	if (llist_add(&cmd->ll_list, &cq->list)) {
		ktime_t kt = ktime_set(0, completion_nsec);

		hrtimer_start(&cq->timer, kt, HRTIMER_MODE_REL);
	}

	put_cpu();
}

static void null_softirq_done_fn(struct request *rq)
{
	blk_end_request_all(rq, 0);
}

#ifdef CONFIG_SMP

static void null_ipi_cmd_end_io(void *data)
{
	struct completion_queue *cq;
	struct llist_node *entry, *next;
	struct nullb_cmd *cmd;

	cq = &per_cpu(completion_queues, smp_processor_id());

	entry = llist_del_all(&cq->list);

	while (entry) {
		next = entry->next;
		cmd = llist_entry(entry, struct nullb_cmd, ll_list);
		end_cmd(cmd);
		entry = next;
	}
}

static void null_cmd_end_ipi(struct nullb_cmd *cmd)
{
	struct call_single_data *data = &cmd->csd;
	int cpu = get_cpu();
	struct completion_queue *cq = &per_cpu(completion_queues, cpu);

	cmd->ll_list.next = NULL;

	if (llist_add(&cmd->ll_list, &cq->list)) {
		data->func = null_ipi_cmd_end_io;
		data->flags = 0;
		__smp_call_function_single(cpu, data, 0);
	}

	put_cpu();
}

#endif /* CONFIG_SMP */

static inline void null_handle_cmd(struct nullb_cmd *cmd)
{
	/* Complete IO by inline, softirq or timer */
	switch (irqmode) {
	case NULL_IRQ_NONE:
		end_cmd(cmd);
		break;
	case NULL_IRQ_SOFTIRQ:
#ifdef CONFIG_SMP
		null_cmd_end_ipi(cmd);
#else
		end_cmd(cmd);
#endif
		break;
	case NULL_IRQ_TIMER:
		null_cmd_end_timer(cmd);
		break;
	}
}

static struct nullb_queue *nullb_to_queue(struct nullb *nullb)
{
	int index = 0;

	if (nullb->nr_queues != 1)
		index = raw_smp_processor_id() / ((nr_cpu_ids + nullb->nr_queues - 1) / nullb->nr_queues);

	return &nullb->queues[index];
}

static void null_queue_bio(struct request_queue *q, struct bio *bio)
{
	struct nullb *nullb = q->queuedata;
	struct nullb_queue *nq = nullb_to_queue(nullb);
	struct nullb_cmd *cmd;

	cmd = alloc_cmd(nq, 1);
	cmd->bio = bio;

	null_handle_cmd(cmd);
}

static int null_rq_prep_fn(struct request_queue *q, struct request *req)
{
	struct nullb *nullb = q->queuedata;
	struct nullb_queue *nq = nullb_to_queue(nullb);
	struct nullb_cmd *cmd;

	cmd = alloc_cmd(nq, 0);
	if (cmd) {
		cmd->rq = req;
		req->special = cmd;
		return BLKPREP_OK;
	}

	return BLKPREP_DEFER;
}

static void null_request_fn(struct request_queue *q)
{
	struct request *rq;

	while ((rq = blk_fetch_request(q)) != NULL) {
		struct nullb_cmd *cmd = rq->special;

		spin_unlock_irq(q->queue_lock);
		null_handle_cmd(cmd);
		spin_lock_irq(q->queue_lock);
	}
}

static int null_queue_rq(struct blk_mq_hw_ctx *hctx, struct request *rq)
{
	struct nullb_cmd *cmd = rq->special;

	cmd->rq = rq;
	cmd->nq = hctx->driver_data;

	null_handle_cmd(cmd);
	return BLK_MQ_RQ_QUEUE_OK;
}

static struct blk_mq_hw_ctx *null_alloc_hctx(struct blk_mq_reg *reg, unsigned int hctx_index)
{
	return kzalloc_node(sizeof(struct blk_mq_hw_ctx), GFP_KERNEL,
				hctx_index);
}

static void null_free_hctx(struct blk_mq_hw_ctx *hctx, unsigned int hctx_index)
{
	kfree(hctx);
}

static void null_init_queue(struct nullb *nullb, struct nullb_queue *nq)
{
	BUG_ON(!nullb);
	BUG_ON(!nq);

	init_waitqueue_head(&nq->wait);
	nq->queue_depth = nullb->queue_depth;
}

static int null_init_hctx(struct blk_mq_hw_ctx *hctx, void *data,
			  unsigned int index)
{
	struct nullb *nullb = data;
	struct nullb_queue *nq = &nullb->queues[index];

	hctx->driver_data = nq;
	null_init_queue(nullb, nq);
	nullb->nr_queues++;

	return 0;
}

static struct blk_mq_ops null_mq_ops = {
	.queue_rq       = null_queue_rq,
	.map_queue      = blk_mq_map_queue,
	.init_hctx	= null_init_hctx,
};

static struct blk_mq_reg null_mq_reg = {
	.ops		= &null_mq_ops,
	.queue_depth	= 64,
	.cmd_size	= sizeof(struct nullb_cmd),
	.flags		= BLK_MQ_F_SHOULD_MERGE,
};

static void null_del_dev(struct nullb *nullb)
{
	list_del_init(&nullb->list);

	del_gendisk(nullb->disk);
	if (queue_mode == NULL_Q_MQ)
		blk_mq_free_queue(nullb->q);
	else
		blk_cleanup_queue(nullb->q);
	put_disk(nullb->disk);
	kfree(nullb);
}

static int null_open(struct block_device *bdev, fmode_t mode)
{
	return 0;
}

static void null_release(struct gendisk *disk, fmode_t mode)
{
}

static const struct block_device_operations null_fops = {
	.owner =	THIS_MODULE,
	.open =		null_open,
	.release =	null_release,
};

static int setup_commands(struct nullb_queue *nq)
{
	struct nullb_cmd *cmd;
	int i, tag_size;

	nq->cmds = kzalloc(nq->queue_depth * sizeof(*cmd), GFP_KERNEL);
	if (!nq->cmds)
		return -ENOMEM;

	tag_size = ALIGN(nq->queue_depth, BITS_PER_LONG) / BITS_PER_LONG;
	nq->tag_map = kzalloc(tag_size * sizeof(unsigned long), GFP_KERNEL);
	if (!nq->tag_map) {
		kfree(nq->cmds);
		return -ENOMEM;
	}

	for (i = 0; i < nq->queue_depth; i++) {
		cmd = &nq->cmds[i];
		INIT_LIST_HEAD(&cmd->list);
		cmd->ll_list.next = NULL;
		cmd->tag = -1U;
	}

	return 0;
}

static void cleanup_queue(struct nullb_queue *nq)
{
	kfree(nq->tag_map);
	kfree(nq->cmds);
}

static void cleanup_queues(struct nullb *nullb)
{
	int i;

	for (i = 0; i < nullb->nr_queues; i++)
		cleanup_queue(&nullb->queues[i]);

	kfree(nullb->queues);
}

static int setup_queues(struct nullb *nullb)
{
	nullb->queues = kzalloc(submit_queues * sizeof(struct nullb_queue),
								GFP_KERNEL);
	if (!nullb->queues)
		return -ENOMEM;

	nullb->nr_queues = 0;
	nullb->queue_depth = hw_queue_depth;

	return 0;
}

static int init_driver_queues(struct nullb *nullb)
{
	struct nullb_queue *nq;
	int i, ret = 0;

	for (i = 0; i < submit_queues; i++) {
		nq = &nullb->queues[i];

		null_init_queue(nullb, nq);

		ret = setup_commands(nq);
		if (ret)
			goto err_queue;
		nullb->nr_queues++;
	}

	return 0;
err_queue:
	cleanup_queues(nullb);
	return ret;
}

static int null_add_dev(void)
{
	struct gendisk *disk;
	struct nullb *nullb;
	sector_t size;

	nullb = kzalloc_node(sizeof(*nullb), GFP_KERNEL, home_node);
	if (!nullb)
		return -ENOMEM;

	spin_lock_init(&nullb->lock);

	if (setup_queues(nullb))
		goto err;

	if (queue_mode == NULL_Q_MQ) {
		null_mq_reg.numa_node = home_node;
		null_mq_reg.queue_depth = hw_queue_depth;
		null_mq_reg.nr_hw_queues = submit_queues;

		if (use_per_node_hctx) {
			null_mq_reg.ops->alloc_hctx = null_alloc_hctx;
			null_mq_reg.ops->free_hctx = null_free_hctx;
		} else {
			null_mq_reg.ops->alloc_hctx = blk_mq_alloc_single_hw_queue;
			null_mq_reg.ops->free_hctx = blk_mq_free_single_hw_queue;
		}

		nullb->q = blk_mq_init_queue(&null_mq_reg, nullb);
	} else if (queue_mode == NULL_Q_BIO) {
		nullb->q = blk_alloc_queue_node(GFP_KERNEL, home_node);
		blk_queue_make_request(nullb->q, null_queue_bio);
		init_driver_queues(nullb);
	} else {
		nullb->q = blk_init_queue_node(null_request_fn, &nullb->lock, home_node);
		blk_queue_prep_rq(nullb->q, null_rq_prep_fn);
		if (nullb->q)
			blk_queue_softirq_done(nullb->q, null_softirq_done_fn);
		init_driver_queues(nullb);
	}

	if (!nullb->q)
		goto queue_fail;

	nullb->q->queuedata = nullb;
	queue_flag_set_unlocked(QUEUE_FLAG_NONROT, nullb->q);

	disk = nullb->disk = alloc_disk_node(1, home_node);
	if (!disk) {
queue_fail:
		if (queue_mode == NULL_Q_MQ)
			blk_mq_free_queue(nullb->q);
		else
			blk_cleanup_queue(nullb->q);
		cleanup_queues(nullb);
err:
		kfree(nullb);
		return -ENOMEM;
	}

	mutex_lock(&lock);
	list_add_tail(&nullb->list, &nullb_list);
	nullb->index = nullb_indexes++;
	mutex_unlock(&lock);

	blk_queue_logical_block_size(nullb->q, bs);
	blk_queue_physical_block_size(nullb->q, bs);

	size = gb * 1024 * 1024 * 1024ULL;
	sector_div(size, bs);
	set_capacity(disk, size);

	disk->flags |= GENHD_FL_EXT_DEVT;
	disk->major		= null_major;
	disk->first_minor	= nullb->index;
	disk->fops		= &null_fops;
	disk->private_data	= nullb;
	disk->queue		= nullb->q;
	sprintf(disk->disk_name, "nullb%d", nullb->index);
	add_disk(disk);
	return 0;
}

static int __init null_init(void)
{
	unsigned int i;

#if !defined(CONFIG_SMP)
	if (irqmode == NULL_IRQ_SOFTIRQ) {
		pr_warn("null_blk: softirq completions not available.\n");
		pr_warn("null_blk: using direct completions.\n");
		irqmode = NULL_IRQ_NONE;
	}
#endif

	if (queue_mode == NULL_Q_MQ && use_per_node_hctx)
		submit_queues = nr_online_nodes;
	else if (submit_queues > nr_cpu_ids)
		submit_queues = nr_cpu_ids;
	else if (!submit_queues)
		submit_queues = 1;

	mutex_init(&lock);

	/* Initialize a separate list for each CPU for issuing softirqs */
	for_each_possible_cpu(i) {
		struct completion_queue *cq = &per_cpu(completion_queues, i);

		init_llist_head(&cq->list);

		if (irqmode != NULL_IRQ_TIMER)
			continue;

		hrtimer_init(&cq->timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
		cq->timer.function = null_cmd_timer_expired;
	}

	null_major = register_blkdev(0, "nullb");
	if (null_major < 0)
		return null_major;

	for (i = 0; i < nr_devices; i++) {
		if (null_add_dev()) {
			unregister_blkdev(null_major, "nullb");
			return -EINVAL;
		}
	}

	pr_info("null: module loaded\n");
	return 0;
}

static void __exit null_exit(void)
{
	struct nullb *nullb;

	unregister_blkdev(null_major, "nullb");

	mutex_lock(&lock);
	while (!list_empty(&nullb_list)) {
		nullb = list_entry(nullb_list.next, struct nullb, list);
		null_del_dev(nullb);
	}
	mutex_unlock(&lock);
}

module_init(null_init);
module_exit(null_exit);

MODULE_AUTHOR("Jens Axboe <jaxboe@fusionio.com>");
MODULE_LICENSE("GPL");
Commit	Line	Data
f2298c04 JA	1	#include <linux/module.h>
	2	#include <linux/moduleparam.h>
	3	#include <linux/sched.h>
	4	#include <linux/fs.h>
	5	#include <linux/blkdev.h>
	6	#include <linux/init.h>
	7	#include <linux/slab.h>
	8	#include <linux/blk-mq.h>
	9	#include <linux/hrtimer.h>
	10
	11	struct nullb_cmd {
	12	struct list_head list;
	13	struct llist_node ll_list;
	14	struct call_single_data csd;
	15	struct request *rq;
	16	struct bio *bio;
	17	unsigned int tag;
	18	struct nullb_queue *nq;
	19	};
	20
	21	struct nullb_queue {
	22	unsigned long *tag_map;
	23	wait_queue_head_t wait;
	24	unsigned int queue_depth;
	25
	26	struct nullb_cmd *cmds;
	27	};
	28
	29	struct nullb {
	30	struct list_head list;
	31	unsigned int index;
	32	struct request_queue *q;
	33	struct gendisk *disk;
	34	struct hrtimer timer;
	35	unsigned int queue_depth;
	36	spinlock_t lock;
	37
	38	struct nullb_queue *queues;
	39	unsigned int nr_queues;
	40	};
	41
	42	static LIST_HEAD(nullb_list);
	43	static struct mutex lock;
	44	static int null_major;
	45	static int nullb_indexes;
	46
	47	struct completion_queue {
	48	struct llist_head list;
	49	struct hrtimer timer;
	50	};
	51
	52	/*
	53	* These are per-cpu for now, they will need to be configured by the
	54	* complete_queues parameter and appropriately mapped.
	55	*/
	56	static DEFINE_PER_CPU(struct completion_queue, completion_queues);
	57
	58	enum {
	59	NULL_IRQ_NONE = 0,
	60	NULL_IRQ_SOFTIRQ = 1,
	61	NULL_IRQ_TIMER = 2,
	62
	63	NULL_Q_BIO = 0,
	64	NULL_Q_RQ = 1,
65	NULL_Q_MQ = 2,
66	};
67
2d263a78	68	static int submit_queues;
f2298c04 JA	69	module_param(submit_queues, int, S_IRUGO);
	70	MODULE_PARM_DESC(submit_queues, "Number of submission queues");
	71
	72	static int home_node = NUMA_NO_NODE;
	73	module_param(home_node, int, S_IRUGO);
	74	MODULE_PARM_DESC(home_node, "Home node for the device");
	75
	76	static int queue_mode = NULL_Q_MQ;
	77	module_param(queue_mode, int, S_IRUGO);
	78	MODULE_PARM_DESC(use_mq, "Use blk-mq interface (0=bio,1=rq,2=multiqueue)");
	79
	80	static int gb = 250;
	81	module_param(gb, int, S_IRUGO);
	82	MODULE_PARM_DESC(gb, "Size in GB");
	83
	84	static int bs = 512;
	85	module_param(bs, int, S_IRUGO);
	86	MODULE_PARM_DESC(bs, "Block size (in bytes)");
	87
	88	static int nr_devices = 2;
	89	module_param(nr_devices, int, S_IRUGO);
	90	MODULE_PARM_DESC(nr_devices, "Number of devices to register");
	91
	92	static int irqmode = NULL_IRQ_SOFTIRQ;
	93	module_param(irqmode, int, S_IRUGO);
	94	MODULE_PARM_DESC(irqmode, "IRQ completion handler. 0-none, 1-softirq, 2-timer");
	95
	96	static int completion_nsec = 10000;
	97	module_param(completion_nsec, int, S_IRUGO);
	98	MODULE_PARM_DESC(completion_nsec, "Time in ns to complete a request in hardware. Default: 10,000ns");
	99
	100	static int hw_queue_depth = 64;
	101	module_param(hw_queue_depth, int, S_IRUGO);
	102	MODULE_PARM_DESC(hw_queue_depth, "Queue depth for each hardware queue. Default: 64");
	103
	104	static bool use_per_node_hctx = true;
	105	module_param(use_per_node_hctx, bool, S_IRUGO);
	106	MODULE_PARM_DESC(use_per_node_hctx, "Use per-node allocation for hardware context queues. Default: true");
	107
	108	static void put_tag(struct nullb_queue *nq, unsigned int tag)
	109	{
	110	clear_bit_unlock(tag, nq->tag_map);
	111
	112	if (waitqueue_active(&nq->wait))
	113	wake_up(&nq->wait);
	114	}
	115
	116	static unsigned int get_tag(struct nullb_queue *nq)
	117	{
	118	unsigned int tag;
	119
	120	do {
	121	tag = find_first_zero_bit(nq->tag_map, nq->queue_depth);
	122	if (tag >= nq->queue_depth)
	123	return -1U;
	124	} while (test_and_set_bit_lock(tag, nq->tag_map));
	125
	126	return tag;
	127	}
	128
	129	static void free_cmd(struct nullb_cmd *cmd)
	130	{
	131	put_tag(cmd->nq, cmd->tag);
	132	}
133
134	static struct nullb_cmd __alloc_cmd(struct nullb_queue nq)
135	{
136	struct nullb_cmd *cmd;
137	unsigned int tag;
138
139	tag = get_tag(nq);
140	if (tag != -1U) {
141	cmd = &nq->cmds[tag];
142	cmd->tag = tag;
143	cmd->nq = nq;
144	return cmd;
145	}
146
147	return NULL;
148	}
149
150	static struct nullb_cmd alloc_cmd(struct nullb_queue nq, int can_wait)
151	{
152	struct nullb_cmd *cmd;
153	DEFINE_WAIT(wait);
154
155	cmd = __alloc_cmd(nq);
156	if (cmd \|\| !can_wait)
157	return cmd;
158
159	do {
160	prepare_to_wait(&nq->wait, &wait, TASK_UNINTERRUPTIBLE);
161	cmd = __alloc_cmd(nq);
162	if (cmd)
163	break;
164
165	io_schedule();
166	} while (1);
167
168	finish_wait(&nq->wait, &wait);
169	return cmd;
170	}
171
172	static void end_cmd(struct nullb_cmd *cmd)
173	{
174	if (cmd->rq) {
175	if (queue_mode == NULL_Q_MQ)
176	blk_mq_end_io(cmd->rq, 0);
177	else {
178	INIT_LIST_HEAD(&cmd->rq->queuelist);
179	blk_end_request_all(cmd->rq, 0);
180	}
181	} else if (cmd->bio)
182	bio_endio(cmd->bio, 0);
183
184	if (queue_mode != NULL_Q_MQ)
185	free_cmd(cmd);
186	}
187
188	static enum hrtimer_restart null_cmd_timer_expired(struct hrtimer *timer)
189	{
190	struct completion_queue *cq;
191	struct llist_node *entry;
192	struct nullb_cmd *cmd;
193
194	cq = &per_cpu(completion_queues, smp_processor_id());
195
196	while ((entry = llist_del_all(&cq->list)) != NULL) {
197	do {
198	cmd = container_of(entry, struct nullb_cmd, ll_list);
199	end_cmd(cmd);
200	entry = entry->next;
201	} while (entry);
202	}
203
204	return HRTIMER_NORESTART;
205	}
206
207	static void null_cmd_end_timer(struct nullb_cmd *cmd)
208	{
209	struct completion_queue *cq = &per_cpu(completion_queues, get_cpu());
210
211	cmd->ll_list.next = NULL;
212	if (llist_add(&cmd->ll_list, &cq->list)) {
213	ktime_t kt = ktime_set(0, completion_nsec);
214
215	hrtimer_start(&cq->timer, kt, HRTIMER_MODE_REL);
216	}
217
218	put_cpu();
219	}
220
221	static void null_softirq_done_fn(struct request *rq)
222	{
223	blk_end_request_all(rq, 0);
224	}
225
044c8d4b	226	#ifdef CONFIG_SMP
f2298c04 JA	227
	228	static void null_ipi_cmd_end_io(void *data)
	229	{
	230	struct completion_queue *cq;
	231	struct llist_node entry, next;
	232	struct nullb_cmd *cmd;
	233
	234	cq = &per_cpu(completion_queues, smp_processor_id());
	235
	236	entry = llist_del_all(&cq->list);
	237
	238	while (entry) {
	239	next = entry->next;
	240	cmd = llist_entry(entry, struct nullb_cmd, ll_list);
	241	end_cmd(cmd);
	242	entry = next;
	243	}
	244	}
	245
	246	static void null_cmd_end_ipi(struct nullb_cmd *cmd)
	247	{
	248	struct call_single_data *data = &cmd->csd;
	249	int cpu = get_cpu();
	250	struct completion_queue *cq = &per_cpu(completion_queues, cpu);
	251
	252	cmd->ll_list.next = NULL;
	253
	254	if (llist_add(&cmd->ll_list, &cq->list)) {
	255	data->func = null_ipi_cmd_end_io;
	256	data->flags = 0;
	257	__smp_call_function_single(cpu, data, 0);
	258	}
	259
	260	put_cpu();
	261	}
	262
044c8d4b	263	#endif /* CONFIG_SMP */
f2298c04 JA	264
	265	static inline void null_handle_cmd(struct nullb_cmd *cmd)
	266	{
	267	/* Complete IO by inline, softirq or timer */
	268	switch (irqmode) {
	269	case NULL_IRQ_NONE:
	270	end_cmd(cmd);
	271	break;
	272	case NULL_IRQ_SOFTIRQ:
044c8d4b	273	#ifdef CONFIG_SMP
f2298c04 JA	274	null_cmd_end_ipi(cmd);
	275	#else
	276	end_cmd(cmd);
	277	#endif
	278	break;
	279	case NULL_IRQ_TIMER:
	280	null_cmd_end_timer(cmd);
	281	break;
	282	}
	283	}
	284
	285	static struct nullb_queue nullb_to_queue(struct nullb nullb)
	286	{
	287	int index = 0;
	288
	289	if (nullb->nr_queues != 1)
	290	index = raw_smp_processor_id() / ((nr_cpu_ids + nullb->nr_queues - 1) / nullb->nr_queues);
	291
	292	return &nullb->queues[index];
	293	}
	294
	295	static void null_queue_bio(struct request_queue q, struct bio bio)
	296	{
	297	struct nullb *nullb = q->queuedata;
	298	struct nullb_queue *nq = nullb_to_queue(nullb);
	299	struct nullb_cmd *cmd;
	300
	301	cmd = alloc_cmd(nq, 1);
	302	cmd->bio = bio;
	303
	304	null_handle_cmd(cmd);
	305	}
	306
	307	static int null_rq_prep_fn(struct request_queue q, struct request req)
	308	{
	309	struct nullb *nullb = q->queuedata;
	310	struct nullb_queue *nq = nullb_to_queue(nullb);
	311	struct nullb_cmd *cmd;
	312
	313	cmd = alloc_cmd(nq, 0);
	314	if (cmd) {
	315	cmd->rq = req;
	316	req->special = cmd;
	317	return BLKPREP_OK;
	318	}
	319
	320	return BLKPREP_DEFER;
	321	}
	322
	323	static void null_request_fn(struct request_queue *q)
	324	{
	325	struct request *rq;
	326
	327	while ((rq = blk_fetch_request(q)) != NULL) {
	328	struct nullb_cmd *cmd = rq->special;
	329
	330	spin_unlock_irq(q->queue_lock);
	331	null_handle_cmd(cmd);
	332	spin_lock_irq(q->queue_lock);
	333	}
	334	}
	335
	336	static int null_queue_rq(struct blk_mq_hw_ctx hctx, struct request rq)
	337	{
338	struct nullb_cmd *cmd = rq->special;
339
340	cmd->rq = rq;
341	cmd->nq = hctx->driver_data;
342
343	null_handle_cmd(cmd);
344	return BLK_MQ_RQ_QUEUE_OK;
345	}
346
347	static struct blk_mq_hw_ctx null_alloc_hctx(struct blk_mq_reg reg, unsigned int hctx_index)
348	{
349	return kzalloc_node(sizeof(struct blk_mq_hw_ctx), GFP_KERNEL,
350	hctx_index);
351	}
352
353	static void null_free_hctx(struct blk_mq_hw_ctx *hctx, unsigned int hctx_index)
354	{
355	kfree(hctx);
356	}
357
2d263a78 MB	358	static void null_init_queue(struct nullb nullb, struct nullb_queue nq)
	359	{
	360	BUG_ON(!nullb);
	361	BUG_ON(!nq);
	362
	363	init_waitqueue_head(&nq->wait);
	364	nq->queue_depth = nullb->queue_depth;
	365	}
	366
f2298c04 JA	367	static int null_init_hctx(struct blk_mq_hw_ctx hctx, void data,
	368	unsigned int index)
	369	{
	370	struct nullb *nullb = data;
	371	struct nullb_queue *nq = &nullb->queues[index];
	372
f2298c04	373	hctx->driver_data = nq;
2d263a78 MB	374	null_init_queue(nullb, nq);
2d263a78 MB	375	nullb->nr_queues++;
f2298c04 JA	376
	377	return 0;
	378	}
	379
	380	static struct blk_mq_ops null_mq_ops = {
	381	.queue_rq = null_queue_rq,
	382	.map_queue = blk_mq_map_queue,
	383	.init_hctx = null_init_hctx,
	384	};
	385
	386	static struct blk_mq_reg null_mq_reg = {
	387	.ops = &null_mq_ops,
	388	.queue_depth = 64,
	389	.cmd_size = sizeof(struct nullb_cmd),
	390	.flags = BLK_MQ_F_SHOULD_MERGE,
	391	};
	392
	393	static void null_del_dev(struct nullb *nullb)
	394	{
	395	list_del_init(&nullb->list);
	396
	397	del_gendisk(nullb->disk);
	398	if (queue_mode == NULL_Q_MQ)
	399	blk_mq_free_queue(nullb->q);
	400	else
	401	blk_cleanup_queue(nullb->q);
	402	put_disk(nullb->disk);
	403	kfree(nullb);
	404	}
	405
	406	static int null_open(struct block_device *bdev, fmode_t mode)
	407	{
	408	return 0;
	409	}
	410
	411	static void null_release(struct gendisk *disk, fmode_t mode)
	412	{
	413	}
	414
	415	static const struct block_device_operations null_fops = {
	416	.owner = THIS_MODULE,
	417	.open = null_open,
	418	.release = null_release,
	419	};
	420
	421	static int setup_commands(struct nullb_queue *nq)
	422	{
	423	struct nullb_cmd *cmd;
	424	int i, tag_size;
	425
	426	nq->cmds = kzalloc(nq->queue_depth * sizeof(*cmd), GFP_KERNEL);
	427	if (!nq->cmds)
2d263a78	428	return -ENOMEM;
f2298c04 JA	429
	430	tag_size = ALIGN(nq->queue_depth, BITS_PER_LONG) / BITS_PER_LONG;
	431	nq->tag_map = kzalloc(tag_size * sizeof(unsigned long), GFP_KERNEL);
	432	if (!nq->tag_map) {
	433	kfree(nq->cmds);
2d263a78	434	return -ENOMEM;
f2298c04 JA	435	}
	436
	437	for (i = 0; i < nq->queue_depth; i++) {
	438	cmd = &nq->cmds[i];
	439	INIT_LIST_HEAD(&cmd->list);
	440	cmd->ll_list.next = NULL;
	441	cmd->tag = -1U;
	442	}
	443
	444	return 0;
	445	}
	446
	447	static void cleanup_queue(struct nullb_queue *nq)
	448	{
	449	kfree(nq->tag_map);
	450	kfree(nq->cmds);
	451	}
	452
	453	static void cleanup_queues(struct nullb *nullb)
	454	{
	455	int i;
	456
	457	for (i = 0; i < nullb->nr_queues; i++)
	458	cleanup_queue(&nullb->queues[i]);
	459
	460	kfree(nullb->queues);
	461	}
	462
	463	static int setup_queues(struct nullb *nullb)
	464	{
2d263a78 MB	465	nullb->queues = kzalloc(submit_queues * sizeof(struct nullb_queue),
2d263a78 MB	466	GFP_KERNEL);
f2298c04	467	if (!nullb->queues)
2d263a78	468	return -ENOMEM;
f2298c04 JA	469
	470	nullb->nr_queues = 0;
	471	nullb->queue_depth = hw_queue_depth;
	472
2d263a78 MB	473	return 0;
	474	}
	475
	476	static int init_driver_queues(struct nullb *nullb)
	477	{
	478	struct nullb_queue *nq;
	479	int i, ret = 0;
f2298c04 JA	480
	481	for (i = 0; i < submit_queues; i++) {
	482	nq = &nullb->queues[i];
2d263a78 MB	483
	484	null_init_queue(nullb, nq);
	485
	486	ret = setup_commands(nq);
	487	if (ret)
	488	goto err_queue;
f2298c04 JA	489	nullb->nr_queues++;
	490	}
	491
2d263a78 MB	492	return 0;
2d263a78 MB	493	err_queue:
f2298c04	494	cleanup_queues(nullb);
2d263a78	495	return ret;
f2298c04 JA	496	}
	497
	498	static int null_add_dev(void)
	499	{
	500	struct gendisk *disk;
	501	struct nullb *nullb;
	502	sector_t size;
	503
	504	nullb = kzalloc_node(sizeof(*nullb), GFP_KERNEL, home_node);
	505	if (!nullb)
	506	return -ENOMEM;
	507
	508	spin_lock_init(&nullb->lock);
	509
	510	if (setup_queues(nullb))
	511	goto err;
	512
	513	if (queue_mode == NULL_Q_MQ) {
	514	null_mq_reg.numa_node = home_node;
	515	null_mq_reg.queue_depth = hw_queue_depth;
0c56010c	516	null_mq_reg.nr_hw_queues = submit_queues;
f2298c04 JA	517
	518	if (use_per_node_hctx) {
	519	null_mq_reg.ops->alloc_hctx = null_alloc_hctx;
	520	null_mq_reg.ops->free_hctx = null_free_hctx;
f2298c04 JA	521	} else {
	522	null_mq_reg.ops->alloc_hctx = blk_mq_alloc_single_hw_queue;
	523	null_mq_reg.ops->free_hctx = blk_mq_free_single_hw_queue;
f2298c04 JA	524	}
	525
	526	nullb->q = blk_mq_init_queue(&null_mq_reg, nullb);
	527	} else if (queue_mode == NULL_Q_BIO) {
	528	nullb->q = blk_alloc_queue_node(GFP_KERNEL, home_node);
	529	blk_queue_make_request(nullb->q, null_queue_bio);
2d263a78	530	init_driver_queues(nullb);
f2298c04 JA	531	} else {
	532	nullb->q = blk_init_queue_node(null_request_fn, &nullb->lock, home_node);
	533	blk_queue_prep_rq(nullb->q, null_rq_prep_fn);
	534	if (nullb->q)
	535	blk_queue_softirq_done(nullb->q, null_softirq_done_fn);
2d263a78	536	init_driver_queues(nullb);
f2298c04 JA	537	}
	538
	539	if (!nullb->q)
	540	goto queue_fail;
	541
	542	nullb->q->queuedata = nullb;
	543	queue_flag_set_unlocked(QUEUE_FLAG_NONROT, nullb->q);
	544
	545	disk = nullb->disk = alloc_disk_node(1, home_node);
	546	if (!disk) {
	547	queue_fail:
	548	if (queue_mode == NULL_Q_MQ)
	549	blk_mq_free_queue(nullb->q);
	550	else
	551	blk_cleanup_queue(nullb->q);
	552	cleanup_queues(nullb);
	553	err:
	554	kfree(nullb);
	555	return -ENOMEM;
	556	}
	557
	558	mutex_lock(&lock);
	559	list_add_tail(&nullb->list, &nullb_list);
	560	nullb->index = nullb_indexes++;
	561	mutex_unlock(&lock);
	562
	563	blk_queue_logical_block_size(nullb->q, bs);
	564	blk_queue_physical_block_size(nullb->q, bs);
	565
	566	size = gb * 1024 * 1024 * 1024ULL;
	567	sector_div(size, bs);
	568	set_capacity(disk, size);
	569
	570	disk->flags \|= GENHD_FL_EXT_DEVT;
	571	disk->major = null_major;
	572	disk->first_minor = nullb->index;
	573	disk->fops = &null_fops;
	574	disk->private_data = nullb;
	575	disk->queue = nullb->q;
	576	sprintf(disk->disk_name, "nullb%d", nullb->index);
	577	add_disk(disk);
	578	return 0;
	579	}
	580
	581	static int __init null_init(void)
	582	{
	583	unsigned int i;
	584
044c8d4b	585	#if !defined(CONFIG_SMP)
f2298c04 JA	586	if (irqmode == NULL_IRQ_SOFTIRQ) {
	587	pr_warn("null_blk: softirq completions not available.\n");
	588	pr_warn("null_blk: using direct completions.\n");
	589	irqmode = NULL_IRQ_NONE;
	590	}
	591	#endif
	592
2d263a78 MB	593	if (queue_mode == NULL_Q_MQ && use_per_node_hctx)
	594	submit_queues = nr_online_nodes;
	595	else if (submit_queues > nr_cpu_ids)
f2298c04 JA	596	submit_queues = nr_cpu_ids;
	597	else if (!submit_queues)
	598	submit_queues = 1;
	599
	600	mutex_init(&lock);
	601
	602	/* Initialize a separate list for each CPU for issuing softirqs */
	603	for_each_possible_cpu(i) {
	604	struct completion_queue *cq = &per_cpu(completion_queues, i);
	605
	606	init_llist_head(&cq->list);
	607
	608	if (irqmode != NULL_IRQ_TIMER)
	609	continue;
	610
	611	hrtimer_init(&cq->timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
	612	cq->timer.function = null_cmd_timer_expired;
	613	}
	614
	615	null_major = register_blkdev(0, "nullb");
	616	if (null_major < 0)
	617	return null_major;
	618
	619	for (i = 0; i < nr_devices; i++) {
	620	if (null_add_dev()) {
	621	unregister_blkdev(null_major, "nullb");
	622	return -EINVAL;
	623	}
	624	}
	625
	626	pr_info("null: module loaded\n");
	627	return 0;
	628	}
	629
	630	static void __exit null_exit(void)
	631	{
	632	struct nullb *nullb;
	633
	634	unregister_blkdev(null_major, "nullb");
	635
	636	mutex_lock(&lock);
	637	while (!list_empty(&nullb_list)) {
	638	nullb = list_entry(nullb_list.next, struct nullb, list);
	639	null_del_dev(nullb);
	640	}
	641	mutex_unlock(&lock);
	642	}
	643
	644	module_init(null_init);
	645	module_exit(null_exit);
	646
	647	MODULE_AUTHOR("Jens Axboe <jaxboe@fusionio.com>");
	648	MODULE_LICENSE("GPL");