[mirror_ubuntu-hirsute-kernel.git] / drivers / block / null_blk_zoned.c

// SPDX-License-Identifier: GPL-2.0
#include <linux/vmalloc.h>
#include <linux/bitmap.h>
#include "null_blk.h"

#define CREATE_TRACE_POINTS
#include "null_blk_trace.h"

/* zone_size in MBs to sectors. */
#define ZONE_SIZE_SHIFT		11

static inline unsigned int null_zone_no(struct nullb_device *dev, sector_t sect)
{
	return sect >> ilog2(dev->zone_size_sects);
}

int null_init_zoned_dev(struct nullb_device *dev, struct request_queue *q)
{
	sector_t dev_size = (sector_t)dev->size * 1024 * 1024;
	sector_t sector = 0;
	unsigned int i;

	if (!is_power_of_2(dev->zone_size)) {
		pr_err("zone_size must be power-of-two\n");
		return -EINVAL;
	}
	if (dev->zone_size > dev->size) {
		pr_err("Zone size larger than device capacity\n");
		return -EINVAL;
	}

	if (!dev->zone_capacity)
		dev->zone_capacity = dev->zone_size;

	if (dev->zone_capacity > dev->zone_size) {
		pr_err("null_blk: zone capacity (%lu MB) larger than zone size (%lu MB)\n",
					dev->zone_capacity, dev->zone_size);
		return -EINVAL;
	}

	dev->zone_size_sects = dev->zone_size << ZONE_SIZE_SHIFT;
	dev->nr_zones = dev_size >>
				(SECTOR_SHIFT + ilog2(dev->zone_size_sects));
	dev->zones = kvmalloc_array(dev->nr_zones, sizeof(struct blk_zone),
			GFP_KERNEL | __GFP_ZERO);
	if (!dev->zones)
		return -ENOMEM;

	/*
	 * With memory backing, the zone_lock spinlock needs to be temporarily
	 * released to avoid scheduling in atomic context. To guarantee zone
	 * information protection, use a bitmap to lock zones with
	 * wait_on_bit_lock_io(). Sleeping on the lock is OK as memory backing
	 * implies that the queue is marked with BLK_MQ_F_BLOCKING.
	 */
	spin_lock_init(&dev->zone_lock);
	if (dev->memory_backed) {
		dev->zone_locks = bitmap_zalloc(dev->nr_zones, GFP_KERNEL);
		if (!dev->zone_locks) {
			kvfree(dev->zones);
			return -ENOMEM;
		}
	}

	if (dev->zone_nr_conv >= dev->nr_zones) {
		dev->zone_nr_conv = dev->nr_zones - 1;
		pr_info("changed the number of conventional zones to %u",
			dev->zone_nr_conv);
	}

	/* Max active zones has to be < nbr of seq zones in order to be enforceable */
	if (dev->zone_max_active >= dev->nr_zones - dev->zone_nr_conv) {
		dev->zone_max_active = 0;
		pr_info("zone_max_active limit disabled, limit >= zone count\n");
	}

	/* Max open zones has to be <= max active zones */
	if (dev->zone_max_active && dev->zone_max_open > dev->zone_max_active) {
		dev->zone_max_open = dev->zone_max_active;
		pr_info("changed the maximum number of open zones to %u\n",
			dev->nr_zones);
	} else if (dev->zone_max_open >= dev->nr_zones - dev->zone_nr_conv) {
		dev->zone_max_open = 0;
		pr_info("zone_max_open limit disabled, limit >= zone count\n");
	}

	for (i = 0; i <  dev->zone_nr_conv; i++) {
		struct blk_zone *zone = &dev->zones[i];

		zone->start = sector;
		zone->len = dev->zone_size_sects;
		zone->capacity = zone->len;
		zone->wp = zone->start + zone->len;
		zone->type = BLK_ZONE_TYPE_CONVENTIONAL;
		zone->cond = BLK_ZONE_COND_NOT_WP;

		sector += dev->zone_size_sects;
	}

	for (i = dev->zone_nr_conv; i < dev->nr_zones; i++) {
		struct blk_zone *zone = &dev->zones[i];

		zone->start = zone->wp = sector;
		zone->len = dev->zone_size_sects;
		zone->capacity = dev->zone_capacity << ZONE_SIZE_SHIFT;
		zone->type = BLK_ZONE_TYPE_SEQWRITE_REQ;
		zone->cond = BLK_ZONE_COND_EMPTY;

		sector += dev->zone_size_sects;
	}

	q->limits.zoned = BLK_ZONED_HM;
	blk_queue_flag_set(QUEUE_FLAG_ZONE_RESETALL, q);
	blk_queue_required_elevator_features(q, ELEVATOR_F_ZBD_SEQ_WRITE);

	return 0;
}

int null_register_zoned_dev(struct nullb *nullb)
{
	struct nullb_device *dev = nullb->dev;
	struct request_queue *q = nullb->q;

	if (queue_is_mq(q)) {
		int ret = blk_revalidate_disk_zones(nullb->disk, NULL);

		if (ret)
			return ret;
	} else {
		blk_queue_chunk_sectors(q, dev->zone_size_sects);
		q->nr_zones = blkdev_nr_zones(nullb->disk);
	}

	blk_queue_max_zone_append_sectors(q, dev->zone_size_sects);
	blk_queue_max_open_zones(q, dev->zone_max_open);
	blk_queue_max_active_zones(q, dev->zone_max_active);

	return 0;
}

void null_free_zoned_dev(struct nullb_device *dev)
{
	bitmap_free(dev->zone_locks);
	kvfree(dev->zones);
}

static inline void null_lock_zone(struct nullb_device *dev, unsigned int zno)
{
	if (dev->memory_backed)
		wait_on_bit_lock_io(dev->zone_locks, zno, TASK_UNINTERRUPTIBLE);
	spin_lock_irq(&dev->zone_lock);
}

static inline void null_unlock_zone(struct nullb_device *dev, unsigned int zno)
{
	spin_unlock_irq(&dev->zone_lock);

	if (dev->memory_backed)
		clear_and_wake_up_bit(zno, dev->zone_locks);
}

int null_report_zones(struct gendisk *disk, sector_t sector,
		unsigned int nr_zones, report_zones_cb cb, void *data)
{
	struct nullb *nullb = disk->private_data;
	struct nullb_device *dev = nullb->dev;
	unsigned int first_zone, i, zno;
	struct blk_zone zone;
	int error;

	first_zone = null_zone_no(dev, sector);
	if (first_zone >= dev->nr_zones)
		return 0;

	nr_zones = min(nr_zones, dev->nr_zones - first_zone);
	trace_nullb_report_zones(nullb, nr_zones);

	zno = first_zone;
	for (i = 0; i < nr_zones; i++, zno++) {
		/*
		 * Stacked DM target drivers will remap the zone information by
		 * modifying the zone information passed to the report callback.
		 * So use a local copy to avoid corruption of the device zone
		 * array.
		 */
		null_lock_zone(dev, zno);
		memcpy(&zone, &dev->zones[zno], sizeof(struct blk_zone));
		null_unlock_zone(dev, zno);

		error = cb(&zone, i, data);
		if (error)
			return error;
	}

	return nr_zones;
}

/*
 * This is called in the case of memory backing from null_process_cmd()
 * with the target zone already locked.
 */
size_t null_zone_valid_read_len(struct nullb *nullb,
				sector_t sector, unsigned int len)
{
	struct nullb_device *dev = nullb->dev;
	struct blk_zone *zone = &dev->zones[null_zone_no(dev, sector)];
	unsigned int nr_sectors = len >> SECTOR_SHIFT;

	/* Read must be below the write pointer position */
	if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL ||
	    sector + nr_sectors <= zone->wp)
		return len;

	if (sector > zone->wp)
		return 0;

	return (zone->wp - sector) << SECTOR_SHIFT;
}

static blk_status_t null_close_zone(struct nullb_device *dev, struct blk_zone *zone)
{
	if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
		return BLK_STS_IOERR;

	switch (zone->cond) {
	case BLK_ZONE_COND_CLOSED:
		/* close operation on closed is not an error */
		return BLK_STS_OK;
	case BLK_ZONE_COND_IMP_OPEN:
		dev->nr_zones_imp_open--;
		break;
	case BLK_ZONE_COND_EXP_OPEN:
		dev->nr_zones_exp_open--;
		break;
	case BLK_ZONE_COND_EMPTY:
	case BLK_ZONE_COND_FULL:
	default:
		return BLK_STS_IOERR;
	}

	if (zone->wp == zone->start) {
		zone->cond = BLK_ZONE_COND_EMPTY;
	} else {
		zone->cond = BLK_ZONE_COND_CLOSED;
		dev->nr_zones_closed++;
	}

	return BLK_STS_OK;
}

static void null_close_first_imp_zone(struct nullb_device *dev)
{
	unsigned int i;

	for (i = dev->zone_nr_conv; i < dev->nr_zones; i++) {
		if (dev->zones[i].cond == BLK_ZONE_COND_IMP_OPEN) {
			null_close_zone(dev, &dev->zones[i]);
			return;
		}
	}
}

static blk_status_t null_check_active(struct nullb_device *dev)
{
	if (!dev->zone_max_active)
		return BLK_STS_OK;

	if (dev->nr_zones_exp_open + dev->nr_zones_imp_open +
			dev->nr_zones_closed < dev->zone_max_active)
		return BLK_STS_OK;

	return BLK_STS_ZONE_ACTIVE_RESOURCE;
}

static blk_status_t null_check_open(struct nullb_device *dev)
{
	if (!dev->zone_max_open)
		return BLK_STS_OK;

	if (dev->nr_zones_exp_open + dev->nr_zones_imp_open < dev->zone_max_open)
		return BLK_STS_OK;

	if (dev->nr_zones_imp_open) {
		if (null_check_active(dev) == BLK_STS_OK) {
			null_close_first_imp_zone(dev);
			return BLK_STS_OK;
		}
	}

	return BLK_STS_ZONE_OPEN_RESOURCE;
}

/*
 * This function matches the manage open zone resources function in the ZBC standard,
 * with the addition of max active zones support (added in the ZNS standard).
 *
 * The function determines if a zone can transition to implicit open or explicit open,
 * while maintaining the max open zone (and max active zone) limit(s). It may close an
 * implicit open zone in order to make additional zone resources available.
 *
 * ZBC states that an implicit open zone shall be closed only if there is not
 * room within the open limit. However, with the addition of an active limit,
 * it is not certain that closing an implicit open zone will allow a new zone
 * to be opened, since we might already be at the active limit capacity.
 */
static blk_status_t null_check_zone_resources(struct nullb_device *dev, struct blk_zone *zone)
{
	blk_status_t ret;

	switch (zone->cond) {
	case BLK_ZONE_COND_EMPTY:
		ret = null_check_active(dev);
		if (ret != BLK_STS_OK)
			return ret;
		fallthrough;
	case BLK_ZONE_COND_CLOSED:
		return null_check_open(dev);
	default:
		/* Should never be called for other states */
		WARN_ON(1);
		return BLK_STS_IOERR;
	}
}

static blk_status_t null_zone_write(struct nullb_cmd *cmd, sector_t sector,
				    unsigned int nr_sectors, bool append)
{
	struct nullb_device *dev = cmd->nq->dev;
	unsigned int zno = null_zone_no(dev, sector);
	struct blk_zone *zone = &dev->zones[zno];
	blk_status_t ret;

	trace_nullb_zone_op(cmd, zno, zone->cond);

	if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
		return null_process_cmd(cmd, REQ_OP_WRITE, sector, nr_sectors);

	null_lock_zone(dev, zno);

	switch (zone->cond) {
	case BLK_ZONE_COND_FULL:
		/* Cannot write to a full zone */
		ret = BLK_STS_IOERR;
		goto unlock;
	case BLK_ZONE_COND_EMPTY:
	case BLK_ZONE_COND_CLOSED:
		ret = null_check_zone_resources(dev, zone);
		if (ret != BLK_STS_OK)
			goto unlock;
		break;
	case BLK_ZONE_COND_IMP_OPEN:
	case BLK_ZONE_COND_EXP_OPEN:
		break;
	default:
		/* Invalid zone condition */
		ret = BLK_STS_IOERR;
		goto unlock;
	}

	/*
	 * Regular writes must be at the write pointer position.
	 * Zone append writes are automatically issued at the write
	 * pointer and the position returned using the request or BIO
	 * sector.
	 */
	if (append) {
		sector = zone->wp;
		if (cmd->bio)
			cmd->bio->bi_iter.bi_sector = sector;
		else
			cmd->rq->__sector = sector;
	} else if (sector != zone->wp) {
		ret = BLK_STS_IOERR;
		goto unlock;
	}

	if (zone->wp + nr_sectors > zone->start + zone->capacity) {
		ret = BLK_STS_IOERR;
		goto unlock;
	}

	if (zone->cond == BLK_ZONE_COND_CLOSED) {
		dev->nr_zones_closed--;
		dev->nr_zones_imp_open++;
	} else if (zone->cond == BLK_ZONE_COND_EMPTY) {
		dev->nr_zones_imp_open++;
	}
	if (zone->cond != BLK_ZONE_COND_EXP_OPEN)
		zone->cond = BLK_ZONE_COND_IMP_OPEN;

	/*
	 * Memory backing allocation may sleep: release the zone_lock spinlock
	 * to avoid scheduling in atomic context. Zone operation atomicity is
	 * still guaranteed through the zone_locks bitmap.
	 */
	if (dev->memory_backed)
		spin_unlock_irq(&dev->zone_lock);
	ret = null_process_cmd(cmd, REQ_OP_WRITE, sector, nr_sectors);
	if (dev->memory_backed)
		spin_lock_irq(&dev->zone_lock);

	if (ret != BLK_STS_OK)
		goto unlock;

	zone->wp += nr_sectors;
	if (zone->wp == zone->start + zone->capacity) {
		if (zone->cond == BLK_ZONE_COND_EXP_OPEN)
			dev->nr_zones_exp_open--;
		else if (zone->cond == BLK_ZONE_COND_IMP_OPEN)
			dev->nr_zones_imp_open--;
		zone->cond = BLK_ZONE_COND_FULL;
	}
	ret = BLK_STS_OK;

unlock:
	null_unlock_zone(dev, zno);

	return ret;
}

static blk_status_t null_open_zone(struct nullb_device *dev, struct blk_zone *zone)
{
	blk_status_t ret;

	if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
		return BLK_STS_IOERR;

	switch (zone->cond) {
	case BLK_ZONE_COND_EXP_OPEN:
		/* open operation on exp open is not an error */
		return BLK_STS_OK;
	case BLK_ZONE_COND_EMPTY:
		ret = null_check_zone_resources(dev, zone);
		if (ret != BLK_STS_OK)
			return ret;
		break;
	case BLK_ZONE_COND_IMP_OPEN:
		dev->nr_zones_imp_open--;
		break;
	case BLK_ZONE_COND_CLOSED:
		ret = null_check_zone_resources(dev, zone);
		if (ret != BLK_STS_OK)
			return ret;
		dev->nr_zones_closed--;
		break;
	case BLK_ZONE_COND_FULL:
	default:
		return BLK_STS_IOERR;
	}

	zone->cond = BLK_ZONE_COND_EXP_OPEN;
	dev->nr_zones_exp_open++;

	return BLK_STS_OK;
}

static blk_status_t null_finish_zone(struct nullb_device *dev, struct blk_zone *zone)
{
	blk_status_t ret;

	if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
		return BLK_STS_IOERR;

	switch (zone->cond) {
	case BLK_ZONE_COND_FULL:
		/* finish operation on full is not an error */
		return BLK_STS_OK;
	case BLK_ZONE_COND_EMPTY:
		ret = null_check_zone_resources(dev, zone);
		if (ret != BLK_STS_OK)
			return ret;
		break;
	case BLK_ZONE_COND_IMP_OPEN:
		dev->nr_zones_imp_open--;
		break;
	case BLK_ZONE_COND_EXP_OPEN:
		dev->nr_zones_exp_open--;
		break;
	case BLK_ZONE_COND_CLOSED:
		ret = null_check_zone_resources(dev, zone);
		if (ret != BLK_STS_OK)
			return ret;
		dev->nr_zones_closed--;
		break;
	default:
		return BLK_STS_IOERR;
	}

	zone->cond = BLK_ZONE_COND_FULL;
	zone->wp = zone->start + zone->len;

	return BLK_STS_OK;
}

static blk_status_t null_reset_zone(struct nullb_device *dev, struct blk_zone *zone)
{
	if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
		return BLK_STS_IOERR;

	switch (zone->cond) {
	case BLK_ZONE_COND_EMPTY:
		/* reset operation on empty is not an error */
		return BLK_STS_OK;
	case BLK_ZONE_COND_IMP_OPEN:
		dev->nr_zones_imp_open--;
		break;
	case BLK_ZONE_COND_EXP_OPEN:
		dev->nr_zones_exp_open--;
		break;
	case BLK_ZONE_COND_CLOSED:
		dev->nr_zones_closed--;
		break;
	case BLK_ZONE_COND_FULL:
		break;
	default:
		return BLK_STS_IOERR;
	}

	zone->cond = BLK_ZONE_COND_EMPTY;
	zone->wp = zone->start;

	return BLK_STS_OK;
}

static blk_status_t null_zone_mgmt(struct nullb_cmd *cmd, enum req_opf op,
				   sector_t sector)
{
	struct nullb_device *dev = cmd->nq->dev;
	unsigned int zone_no;
	struct blk_zone *zone;
	blk_status_t ret;
	size_t i;

	if (op == REQ_OP_ZONE_RESET_ALL) {
		for (i = dev->zone_nr_conv; i < dev->nr_zones; i++) {
			null_lock_zone(dev, i);
			zone = &dev->zones[i];
			if (zone->cond != BLK_ZONE_COND_EMPTY) {
				null_reset_zone(dev, zone);
				trace_nullb_zone_op(cmd, i, zone->cond);
			}
			null_unlock_zone(dev, i);
		}
		return BLK_STS_OK;
	}

	zone_no = null_zone_no(dev, sector);
	zone = &dev->zones[zone_no];

	null_lock_zone(dev, zone_no);

	switch (op) {
	case REQ_OP_ZONE_RESET:
		ret = null_reset_zone(dev, zone);
		break;
	case REQ_OP_ZONE_OPEN:
		ret = null_open_zone(dev, zone);
		break;
	case REQ_OP_ZONE_CLOSE:
		ret = null_close_zone(dev, zone);
		break;
	case REQ_OP_ZONE_FINISH:
		ret = null_finish_zone(dev, zone);
		break;
	default:
		ret = BLK_STS_NOTSUPP;
		break;
	}

	if (ret == BLK_STS_OK)
		trace_nullb_zone_op(cmd, zone_no, zone->cond);

	null_unlock_zone(dev, zone_no);

	return ret;
}

blk_status_t null_process_zoned_cmd(struct nullb_cmd *cmd, enum req_opf op,
				    sector_t sector, sector_t nr_sectors)
{
	struct nullb_device *dev = cmd->nq->dev;
	unsigned int zno = null_zone_no(dev, sector);
	blk_status_t sts;

	switch (op) {
	case REQ_OP_WRITE:
		sts = null_zone_write(cmd, sector, nr_sectors, false);
		break;
	case REQ_OP_ZONE_APPEND:
		sts = null_zone_write(cmd, sector, nr_sectors, true);
		break;
	case REQ_OP_ZONE_RESET:
	case REQ_OP_ZONE_RESET_ALL:
	case REQ_OP_ZONE_OPEN:
	case REQ_OP_ZONE_CLOSE:
	case REQ_OP_ZONE_FINISH:
		sts = null_zone_mgmt(cmd, op, sector);
		break;
	default:
		null_lock_zone(dev, zno);
		sts = null_process_cmd(cmd, op, sector, nr_sectors);
		null_unlock_zone(dev, zno);
	}

	return sts;
}
Commit	Line	Data
ca4b2a01 MB	1	// SPDX-License-Identifier: GPL-2.0
ca4b2a01 MB	2	#include <linux/vmalloc.h>
aa1c09cb	3	#include <linux/bitmap.h>
ca4b2a01 MB	4	#include "null_blk.h"
ca4b2a01 MB	5
766c3297 CK	6	#define CREATE_TRACE_POINTS
	7	#include "null_blk_trace.h"
	8
ca4b2a01 MB	9	/* zone_size in MBs to sectors. */
	10	#define ZONE_SIZE_SHIFT 11
	11
	12	static inline unsigned int null_zone_no(struct nullb_device *dev, sector_t sect)
	13	{
	14	return sect >> ilog2(dev->zone_size_sects);
	15	}
	16
d205bde7	17	int null_init_zoned_dev(struct nullb_device dev, struct request_queue q)
ca4b2a01 MB	18	{
	19	sector_t dev_size = (sector_t)dev->size * 1024 * 1024;
	20	sector_t sector = 0;
	21	unsigned int i;
	22
	23	if (!is_power_of_2(dev->zone_size)) {
9c7eddf1	24	pr_err("zone_size must be power-of-two\n");
ca4b2a01 MB	25	return -EINVAL;
ca4b2a01 MB	26	}
e2748325 CK	27	if (dev->zone_size > dev->size) {
	28	pr_err("Zone size larger than device capacity\n");
	29	return -EINVAL;
	30	}
ca4b2a01	31
089565fb AR	32	if (!dev->zone_capacity)
	33	dev->zone_capacity = dev->zone_size;
	34
	35	if (dev->zone_capacity > dev->zone_size) {
	36	pr_err("null_blk: zone capacity (%lu MB) larger than zone size (%lu MB)\n",
	37	dev->zone_capacity, dev->zone_size);
	38	return -EINVAL;
	39	}
	40
ca4b2a01 MB	41	dev->zone_size_sects = dev->zone_size << ZONE_SIZE_SHIFT;
	42	dev->nr_zones = dev_size >>
	43	(SECTOR_SHIFT + ilog2(dev->zone_size_sects));
	44	dev->zones = kvmalloc_array(dev->nr_zones, sizeof(struct blk_zone),
	45	GFP_KERNEL \| __GFP_ZERO);
	46	if (!dev->zones)
	47	return -ENOMEM;
	48
e1777d09 DLM	49	/*
	50	* With memory backing, the zone_lock spinlock needs to be temporarily
	51	* released to avoid scheduling in atomic context. To guarantee zone
	52	* information protection, use a bitmap to lock zones with
	53	* wait_on_bit_lock_io(). Sleeping on the lock is OK as memory backing
	54	* implies that the queue is marked with BLK_MQ_F_BLOCKING.
	55	*/
	56	spin_lock_init(&dev->zone_lock);
	57	if (dev->memory_backed) {
	58	dev->zone_locks = bitmap_zalloc(dev->nr_zones, GFP_KERNEL);
	59	if (!dev->zone_locks) {
	60	kvfree(dev->zones);
	61	return -ENOMEM;
	62	}
aa1c09cb DLM	63	}
aa1c09cb DLM	64
ea2c18e1 MS	65	if (dev->zone_nr_conv >= dev->nr_zones) {
ea2c18e1 MS	66	dev->zone_nr_conv = dev->nr_zones - 1;
9c7eddf1	67	pr_info("changed the number of conventional zones to %u",
ea2c18e1 MS	68	dev->zone_nr_conv);
	69	}
	70
dc4d137e NC	71	/* Max active zones has to be < nbr of seq zones in order to be enforceable */
	72	if (dev->zone_max_active >= dev->nr_zones - dev->zone_nr_conv) {
	73	dev->zone_max_active = 0;
	74	pr_info("zone_max_active limit disabled, limit >= zone count\n");
	75	}
	76
	77	/* Max open zones has to be <= max active zones */
	78	if (dev->zone_max_active && dev->zone_max_open > dev->zone_max_active) {
	79	dev->zone_max_open = dev->zone_max_active;
	80	pr_info("changed the maximum number of open zones to %u\n",
	81	dev->nr_zones);
	82	} else if (dev->zone_max_open >= dev->nr_zones - dev->zone_nr_conv) {
	83	dev->zone_max_open = 0;
	84	pr_info("zone_max_open limit disabled, limit >= zone count\n");
	85	}
	86
ea2c18e1 MS	87	for (i = 0; i < dev->zone_nr_conv; i++) {
	88	struct blk_zone *zone = &dev->zones[i];
	89
	90	zone->start = sector;
	91	zone->len = dev->zone_size_sects;
82394db7	92	zone->capacity = zone->len;
ea2c18e1 MS	93	zone->wp = zone->start + zone->len;
	94	zone->type = BLK_ZONE_TYPE_CONVENTIONAL;
	95	zone->cond = BLK_ZONE_COND_NOT_WP;
	96
	97	sector += dev->zone_size_sects;
	98	}
	99
	100	for (i = dev->zone_nr_conv; i < dev->nr_zones; i++) {
ca4b2a01 MB	101	struct blk_zone *zone = &dev->zones[i];
	102
	103	zone->start = zone->wp = sector;
	104	zone->len = dev->zone_size_sects;
089565fb	105	zone->capacity = dev->zone_capacity << ZONE_SIZE_SHIFT;
ca4b2a01 MB	106	zone->type = BLK_ZONE_TYPE_SEQWRITE_REQ;
	107	zone->cond = BLK_ZONE_COND_EMPTY;
	108
	109	sector += dev->zone_size_sects;
	110	}
	111
d205bde7 DLM	112	q->limits.zoned = BLK_ZONED_HM;
	113	blk_queue_flag_set(QUEUE_FLAG_ZONE_RESETALL, q);
	114	blk_queue_required_elevator_features(q, ELEVATOR_F_ZBD_SEQ_WRITE);
	115
	116	return 0;
	117	}
	118
	119	int null_register_zoned_dev(struct nullb *nullb)
	120	{
e0489ed5	121	struct nullb_device *dev = nullb->dev;
d205bde7 DLM	122	struct request_queue *q = nullb->q;
d205bde7 DLM	123
e0489ed5 DLM	124	if (queue_is_mq(q)) {
	125	int ret = blk_revalidate_disk_zones(nullb->disk, NULL);
	126
	127	if (ret)
	128	return ret;
	129	} else {
	130	blk_queue_chunk_sectors(q, dev->zone_size_sects);
	131	q->nr_zones = blkdev_nr_zones(nullb->disk);
	132	}
d205bde7	133
e0489ed5	134	blk_queue_max_zone_append_sectors(q, dev->zone_size_sects);
dc4d137e NC	135	blk_queue_max_open_zones(q, dev->zone_max_open);
dc4d137e NC	136	blk_queue_max_active_zones(q, dev->zone_max_active);
d205bde7	137
ca4b2a01 MB	138	return 0;
	139	}
	140
d205bde7	141	void null_free_zoned_dev(struct nullb_device *dev)
ca4b2a01	142	{
aa1c09cb	143	bitmap_free(dev->zone_locks);
ca4b2a01 MB	144	kvfree(dev->zones);
	145	}
	146
aa1c09cb DLM	147	static inline void null_lock_zone(struct nullb_device *dev, unsigned int zno)
aa1c09cb DLM	148	{
e1777d09 DLM	149	if (dev->memory_backed)
	150	wait_on_bit_lock_io(dev->zone_locks, zno, TASK_UNINTERRUPTIBLE);
	151	spin_lock_irq(&dev->zone_lock);
aa1c09cb DLM	152	}
	153
	154	static inline void null_unlock_zone(struct nullb_device *dev, unsigned int zno)
	155	{
e1777d09 DLM	156	spin_unlock_irq(&dev->zone_lock);
	157
	158	if (dev->memory_backed)
	159	clear_and_wake_up_bit(zno, dev->zone_locks);
aa1c09cb DLM	160	}
aa1c09cb DLM	161
7fc8fb51	162	int null_report_zones(struct gendisk *disk, sector_t sector,
d4100351	163	unsigned int nr_zones, report_zones_cb cb, void *data)
ca4b2a01	164	{
e76239a3 CH	165	struct nullb *nullb = disk->private_data;
e76239a3 CH	166	struct nullb_device *dev = nullb->dev;
aa1c09cb	167	unsigned int first_zone, i, zno;
d4100351 CH	168	struct blk_zone zone;
d4100351 CH	169	int error;
ca4b2a01	170
d4100351 CH	171	first_zone = null_zone_no(dev, sector);
	172	if (first_zone >= dev->nr_zones)
	173	return 0;
ca4b2a01	174
d4100351	175	nr_zones = min(nr_zones, dev->nr_zones - first_zone);
766c3297 CK	176	trace_nullb_report_zones(nullb, nr_zones);
766c3297 CK	177
aa1c09cb DLM	178	zno = first_zone;
aa1c09cb DLM	179	for (i = 0; i < nr_zones; i++, zno++) {
d4100351 CH	180	/*
	181	* Stacked DM target drivers will remap the zone information by
	182	* modifying the zone information passed to the report callback.
	183	* So use a local copy to avoid corruption of the device zone
	184	* array.
	185	*/
aa1c09cb DLM	186	null_lock_zone(dev, zno);
	187	memcpy(&zone, &dev->zones[zno], sizeof(struct blk_zone));
	188	null_unlock_zone(dev, zno);
35bc10b2	189
d4100351 CH	190	error = cb(&zone, i, data);
	191	if (error)
	192	return error;
	193	}
ca4b2a01	194
d4100351	195	return nr_zones;
ca4b2a01 MB	196	}
ca4b2a01 MB	197
aa1c09cb DLM	198	/*
	199	* This is called in the case of memory backing from null_process_cmd()
	200	* with the target zone already locked.
	201	*/
dd85b492 AJ	202	size_t null_zone_valid_read_len(struct nullb *nullb,
	203	sector_t sector, unsigned int len)
	204	{
	205	struct nullb_device *dev = nullb->dev;
	206	struct blk_zone *zone = &dev->zones[null_zone_no(dev, sector)];
	207	unsigned int nr_sectors = len >> SECTOR_SHIFT;
	208
	209	/* Read must be below the write pointer position */
	210	if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL \|\|
	211	sector + nr_sectors <= zone->wp)
	212	return len;
	213
	214	if (sector > zone->wp)
	215	return 0;
	216
	217	return (zone->wp - sector) << SECTOR_SHIFT;
	218	}
	219
dc4d137e NC	220	static blk_status_t null_close_zone(struct nullb_device dev, struct blk_zone zone)
	221	{
	222	if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
	223	return BLK_STS_IOERR;
	224
	225	switch (zone->cond) {
	226	case BLK_ZONE_COND_CLOSED:
	227	/* close operation on closed is not an error */
	228	return BLK_STS_OK;
	229	case BLK_ZONE_COND_IMP_OPEN:
	230	dev->nr_zones_imp_open--;
	231	break;
	232	case BLK_ZONE_COND_EXP_OPEN:
	233	dev->nr_zones_exp_open--;
	234	break;
	235	case BLK_ZONE_COND_EMPTY:
	236	case BLK_ZONE_COND_FULL:
	237	default:
	238	return BLK_STS_IOERR;
	239	}
	240
	241	if (zone->wp == zone->start) {
	242	zone->cond = BLK_ZONE_COND_EMPTY;
	243	} else {
	244	zone->cond = BLK_ZONE_COND_CLOSED;
	245	dev->nr_zones_closed++;
	246	}
	247
	248	return BLK_STS_OK;
	249	}
	250
	251	static void null_close_first_imp_zone(struct nullb_device *dev)
	252	{
	253	unsigned int i;
	254
	255	for (i = dev->zone_nr_conv; i < dev->nr_zones; i++) {
	256	if (dev->zones[i].cond == BLK_ZONE_COND_IMP_OPEN) {
	257	null_close_zone(dev, &dev->zones[i]);
	258	return;
	259	}
	260	}
	261	}
	262
fd78874b	263	static blk_status_t null_check_active(struct nullb_device *dev)
dc4d137e NC	264	{
dc4d137e NC	265	if (!dev->zone_max_active)
fd78874b KB	266	return BLK_STS_OK;
	267
	268	if (dev->nr_zones_exp_open + dev->nr_zones_imp_open +
	269	dev->nr_zones_closed < dev->zone_max_active)
	270	return BLK_STS_OK;
dc4d137e	271
fd78874b	272	return BLK_STS_ZONE_ACTIVE_RESOURCE;
dc4d137e NC	273	}
dc4d137e NC	274
fd78874b	275	static blk_status_t null_check_open(struct nullb_device *dev)
dc4d137e NC	276	{
dc4d137e NC	277	if (!dev->zone_max_open)
fd78874b	278	return BLK_STS_OK;
dc4d137e NC	279
dc4d137e NC	280	if (dev->nr_zones_exp_open + dev->nr_zones_imp_open < dev->zone_max_open)
fd78874b	281	return BLK_STS_OK;
dc4d137e	282
fd78874b KB	283	if (dev->nr_zones_imp_open) {
	284	if (null_check_active(dev) == BLK_STS_OK) {
	285	null_close_first_imp_zone(dev);
	286	return BLK_STS_OK;
	287	}
dc4d137e NC	288	}
dc4d137e NC	289
fd78874b	290	return BLK_STS_ZONE_OPEN_RESOURCE;
dc4d137e NC	291	}
	292
	293	/*
	294	* This function matches the manage open zone resources function in the ZBC standard,
	295	* with the addition of max active zones support (added in the ZNS standard).
	296	*
	297	* The function determines if a zone can transition to implicit open or explicit open,
	298	* while maintaining the max open zone (and max active zone) limit(s). It may close an
	299	* implicit open zone in order to make additional zone resources available.
	300	*
	301	* ZBC states that an implicit open zone shall be closed only if there is not
	302	* room within the open limit. However, with the addition of an active limit,
	303	* it is not certain that closing an implicit open zone will allow a new zone
	304	* to be opened, since we might already be at the active limit capacity.
	305	*/
fd78874b	306	static blk_status_t null_check_zone_resources(struct nullb_device dev, struct blk_zone zone)
dc4d137e	307	{
fd78874b KB	308	blk_status_t ret;
fd78874b KB	309
dc4d137e NC	310	switch (zone->cond) {
dc4d137e NC	311	case BLK_ZONE_COND_EMPTY:
fd78874b KB	312	ret = null_check_active(dev);
	313	if (ret != BLK_STS_OK)
	314	return ret;
dc4d137e NC	315	fallthrough;
dc4d137e NC	316	case BLK_ZONE_COND_CLOSED:
fd78874b	317	return null_check_open(dev);
dc4d137e NC	318	default:
	319	/* Should never be called for other states */
	320	WARN_ON(1);
fd78874b	321	return BLK_STS_IOERR;
dc4d137e NC	322	}
	323	}
	324
fceb5d1b	325	static blk_status_t null_zone_write(struct nullb_cmd *cmd, sector_t sector,
e0489ed5	326	unsigned int nr_sectors, bool append)
ca4b2a01 MB	327	{
ca4b2a01 MB	328	struct nullb_device *dev = cmd->nq->dev;
ca4b2a01 MB	329	unsigned int zno = null_zone_no(dev, sector);
ca4b2a01 MB	330	struct blk_zone *zone = &dev->zones[zno];
9dd44c7e DLM	331	blk_status_t ret;
	332
	333	trace_nullb_zone_op(cmd, zno, zone->cond);
	334
	335	if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
	336	return null_process_cmd(cmd, REQ_OP_WRITE, sector, nr_sectors);
ca4b2a01	337
aa1c09cb	338	null_lock_zone(dev, zno);
aa1c09cb	339
ca4b2a01 MB	340	switch (zone->cond) {
	341	case BLK_ZONE_COND_FULL:
	342	/* Cannot write to a full zone */
aa1c09cb DLM	343	ret = BLK_STS_IOERR;
aa1c09cb DLM	344	goto unlock;
ca4b2a01	345	case BLK_ZONE_COND_EMPTY:
dc4d137e	346	case BLK_ZONE_COND_CLOSED:
fd78874b KB	347	ret = null_check_zone_resources(dev, zone);
fd78874b KB	348	if (ret != BLK_STS_OK)
aa1c09cb	349	goto unlock;
dc4d137e	350	break;
ca4b2a01	351	case BLK_ZONE_COND_IMP_OPEN:
16c731fe	352	case BLK_ZONE_COND_EXP_OPEN:
dc4d137e NC	353	break;
	354	default:
	355	/* Invalid zone condition */
aa1c09cb DLM	356	ret = BLK_STS_IOERR;
aa1c09cb DLM	357	goto unlock;
dc4d137e NC	358	}
	359
	360	/*
	361	* Regular writes must be at the write pointer position.
	362	* Zone append writes are automatically issued at the write
	363	* pointer and the position returned using the request or BIO
	364	* sector.
	365	*/
	366	if (append) {
	367	sector = zone->wp;
	368	if (cmd->bio)
	369	cmd->bio->bi_iter.bi_sector = sector;
	370	else
	371	cmd->rq->__sector = sector;
	372	} else if (sector != zone->wp) {
aa1c09cb DLM	373	ret = BLK_STS_IOERR;
aa1c09cb DLM	374	goto unlock;
dc4d137e NC	375	}
dc4d137e NC	376
aa1c09cb DLM	377	if (zone->wp + nr_sectors > zone->start + zone->capacity) {
	378	ret = BLK_STS_IOERR;
	379	goto unlock;
	380	}
dc4d137e NC	381
	382	if (zone->cond == BLK_ZONE_COND_CLOSED) {
	383	dev->nr_zones_closed--;
	384	dev->nr_zones_imp_open++;
	385	} else if (zone->cond == BLK_ZONE_COND_EMPTY) {
	386	dev->nr_zones_imp_open++;
	387	}
	388	if (zone->cond != BLK_ZONE_COND_EXP_OPEN)
	389	zone->cond = BLK_ZONE_COND_IMP_OPEN;
	390
e1777d09 DLM	391	/*
	392	* Memory backing allocation may sleep: release the zone_lock spinlock
	393	* to avoid scheduling in atomic context. Zone operation atomicity is
	394	* still guaranteed through the zone_locks bitmap.
	395	*/
	396	if (dev->memory_backed)
	397	spin_unlock_irq(&dev->zone_lock);
dc4d137e	398	ret = null_process_cmd(cmd, REQ_OP_WRITE, sector, nr_sectors);
e1777d09 DLM	399	if (dev->memory_backed)
	400	spin_lock_irq(&dev->zone_lock);
	401
dc4d137e	402	if (ret != BLK_STS_OK)
aa1c09cb	403	goto unlock;
dc4d137e NC	404
	405	zone->wp += nr_sectors;
	406	if (zone->wp == zone->start + zone->capacity) {
	407	if (zone->cond == BLK_ZONE_COND_EXP_OPEN)
	408	dev->nr_zones_exp_open--;
	409	else if (zone->cond == BLK_ZONE_COND_IMP_OPEN)
	410	dev->nr_zones_imp_open--;
	411	zone->cond = BLK_ZONE_COND_FULL;
	412	}
aa1c09cb DLM	413	ret = BLK_STS_OK;
	414
	415	unlock:
aa1c09cb DLM	416	null_unlock_zone(dev, zno);
	417
	418	return ret;
dc4d137e NC	419	}
	420
	421	static blk_status_t null_open_zone(struct nullb_device dev, struct blk_zone zone)
	422	{
fd78874b KB	423	blk_status_t ret;
fd78874b KB	424
dc4d137e NC	425	if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
	426	return BLK_STS_IOERR;
	427
	428	switch (zone->cond) {
	429	case BLK_ZONE_COND_EXP_OPEN:
	430	/* open operation on exp open is not an error */
	431	return BLK_STS_OK;
	432	case BLK_ZONE_COND_EMPTY:
fd78874b KB	433	ret = null_check_zone_resources(dev, zone);
	434	if (ret != BLK_STS_OK)
	435	return ret;
dc4d137e NC	436	break;
	437	case BLK_ZONE_COND_IMP_OPEN:
	438	dev->nr_zones_imp_open--;
	439	break;
16c731fe	440	case BLK_ZONE_COND_CLOSED:
fd78874b KB	441	ret = null_check_zone_resources(dev, zone);
	442	if (ret != BLK_STS_OK)
	443	return ret;
dc4d137e NC	444	dev->nr_zones_closed--;
	445	break;
	446	case BLK_ZONE_COND_FULL:
	447	default:
	448	return BLK_STS_IOERR;
	449	}
	450
	451	zone->cond = BLK_ZONE_COND_EXP_OPEN;
	452	dev->nr_zones_exp_open++;
ca4b2a01	453
dc4d137e NC	454	return BLK_STS_OK;
	455	}
	456
	457	static blk_status_t null_finish_zone(struct nullb_device dev, struct blk_zone zone)
	458	{
fd78874b KB	459	blk_status_t ret;
fd78874b KB	460
dc4d137e NC	461	if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
	462	return BLK_STS_IOERR;
	463
	464	switch (zone->cond) {
	465	case BLK_ZONE_COND_FULL:
	466	/* finish operation on full is not an error */
	467	return BLK_STS_OK;
	468	case BLK_ZONE_COND_EMPTY:
fd78874b KB	469	ret = null_check_zone_resources(dev, zone);
	470	if (ret != BLK_STS_OK)
	471	return ret;
dc4d137e NC	472	break;
	473	case BLK_ZONE_COND_IMP_OPEN:
	474	dev->nr_zones_imp_open--;
	475	break;
	476	case BLK_ZONE_COND_EXP_OPEN:
	477	dev->nr_zones_exp_open--;
	478	break;
	479	case BLK_ZONE_COND_CLOSED:
fd78874b KB	480	ret = null_check_zone_resources(dev, zone);
	481	if (ret != BLK_STS_OK)
	482	return ret;
dc4d137e NC	483	dev->nr_zones_closed--;
	484	break;
	485	default:
	486	return BLK_STS_IOERR;
	487	}
089565fb	488
dc4d137e NC	489	zone->cond = BLK_ZONE_COND_FULL;
dc4d137e NC	490	zone->wp = zone->start + zone->len;
ca4b2a01	491
dc4d137e NC	492	return BLK_STS_OK;
	493	}
	494
	495	static blk_status_t null_reset_zone(struct nullb_device dev, struct blk_zone zone)
	496	{
	497	if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
	498	return BLK_STS_IOERR;
9dd44c7e	499
dc4d137e NC	500	switch (zone->cond) {
	501	case BLK_ZONE_COND_EMPTY:
	502	/* reset operation on empty is not an error */
9dd44c7e	503	return BLK_STS_OK;
dc4d137e NC	504	case BLK_ZONE_COND_IMP_OPEN:
	505	dev->nr_zones_imp_open--;
	506	break;
	507	case BLK_ZONE_COND_EXP_OPEN:
	508	dev->nr_zones_exp_open--;
	509	break;
	510	case BLK_ZONE_COND_CLOSED:
	511	dev->nr_zones_closed--;
	512	break;
	513	case BLK_ZONE_COND_FULL:
	514	break;
ca4b2a01	515	default:
fceb5d1b	516	return BLK_STS_IOERR;
ca4b2a01	517	}
dc4d137e NC	518
	519	zone->cond = BLK_ZONE_COND_EMPTY;
	520	zone->wp = zone->start;
	521
	522	return BLK_STS_OK;
ca4b2a01 MB	523	}
ca4b2a01 MB	524
da644b2c AJ	525	static blk_status_t null_zone_mgmt(struct nullb_cmd *cmd, enum req_opf op,
da644b2c AJ	526	sector_t sector)
ca4b2a01 MB	527	{
ca4b2a01 MB	528	struct nullb_device *dev = cmd->nq->dev;
aa1c09cb DLM	529	unsigned int zone_no;
	530	struct blk_zone *zone;
	531	blk_status_t ret;
a61dbfb1 CK	532	size_t i;
a61dbfb1 CK	533
aa1c09cb	534	if (op == REQ_OP_ZONE_RESET_ALL) {
f9c91042	535	for (i = dev->zone_nr_conv; i < dev->nr_zones; i++) {
aa1c09cb	536	null_lock_zone(dev, i);
f9c91042 DLM	537	zone = &dev->zones[i];
	538	if (zone->cond != BLK_ZONE_COND_EMPTY) {
	539	null_reset_zone(dev, zone);
	540	trace_nullb_zone_op(cmd, i, zone->cond);
	541	}
aa1c09cb	542	null_unlock_zone(dev, i);
f9c91042 DLM	543	}
f9c91042 DLM	544	return BLK_STS_OK;
aa1c09cb DLM	545	}
	546
	547	zone_no = null_zone_no(dev, sector);
	548	zone = &dev->zones[zone_no];
	549
	550	null_lock_zone(dev, zone_no);
aa1c09cb DLM	551
aa1c09cb DLM	552	switch (op) {
a61dbfb1	553	case REQ_OP_ZONE_RESET:
dc4d137e	554	ret = null_reset_zone(dev, zone);
a61dbfb1	555	break;
da644b2c	556	case REQ_OP_ZONE_OPEN:
dc4d137e	557	ret = null_open_zone(dev, zone);
da644b2c AJ	558	break;
da644b2c AJ	559	case REQ_OP_ZONE_CLOSE:
dc4d137e	560	ret = null_close_zone(dev, zone);
da644b2c AJ	561	break;
da644b2c AJ	562	case REQ_OP_ZONE_FINISH:
dc4d137e	563	ret = null_finish_zone(dev, zone);
da644b2c	564	break;
a61dbfb1	565	default:
aa1c09cb DLM	566	ret = BLK_STS_NOTSUPP;
aa1c09cb DLM	567	break;
ea2c18e1	568	}
766c3297	569
dc4d137e NC	570	if (ret == BLK_STS_OK)
	571	trace_nullb_zone_op(cmd, zone_no, zone->cond);
	572
aa1c09cb DLM	573	null_unlock_zone(dev, zone_no);
aa1c09cb DLM	574
dc4d137e	575	return ret;
fceb5d1b CK	576	}
fceb5d1b CK	577
9dd44c7e DLM	578	blk_status_t null_process_zoned_cmd(struct nullb_cmd *cmd, enum req_opf op,
9dd44c7e DLM	579	sector_t sector, sector_t nr_sectors)
fceb5d1b	580	{
35bc10b2	581	struct nullb_device *dev = cmd->nq->dev;
aa1c09cb DLM	582	unsigned int zno = null_zone_no(dev, sector);
aa1c09cb DLM	583	blk_status_t sts;
35bc10b2	584
fceb5d1b CK	585	switch (op) {
fceb5d1b CK	586	case REQ_OP_WRITE:
35bc10b2 KJ	587	sts = null_zone_write(cmd, sector, nr_sectors, false);
35bc10b2 KJ	588	break;
e0489ed5	589	case REQ_OP_ZONE_APPEND:
35bc10b2 KJ	590	sts = null_zone_write(cmd, sector, nr_sectors, true);
35bc10b2 KJ	591	break;
fceb5d1b CK	592	case REQ_OP_ZONE_RESET:
fceb5d1b CK	593	case REQ_OP_ZONE_RESET_ALL:
da644b2c AJ	594	case REQ_OP_ZONE_OPEN:
	595	case REQ_OP_ZONE_CLOSE:
	596	case REQ_OP_ZONE_FINISH:
35bc10b2 KJ	597	sts = null_zone_mgmt(cmd, op, sector);
35bc10b2 KJ	598	break;
fceb5d1b	599	default:
aa1c09cb	600	null_lock_zone(dev, zno);
35bc10b2	601	sts = null_process_cmd(cmd, op, sector, nr_sectors);
aa1c09cb	602	null_unlock_zone(dev, zno);
fceb5d1b	603	}
35bc10b2 KJ	604
35bc10b2 KJ	605	return sts;
ca4b2a01	606	}