[thirdparty/linux.git] / drivers / block / null_blk_zoned.c

// SPDX-License-Identifier: GPL-2.0
#include <linux/vmalloc.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;
	}

	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;

	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);
	}

	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->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->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 request_queue *q = nullb->q;

	if (queue_is_mq(q))
		return blk_revalidate_disk_zones(nullb->disk);

	blk_queue_chunk_sectors(q, nullb->dev->zone_size_sects);
	q->nr_zones = blkdev_nr_zones(nullb->disk);

	return 0;
}

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

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;
	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);

	for (i = 0; i < nr_zones; i++) {
		/*
		 * 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.
		 */
		memcpy(&zone, &dev->zones[first_zone + i],
		       sizeof(struct blk_zone));
		error = cb(&zone, i, data);
		if (error)
			return error;
	}

	return nr_zones;
}

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_zone_write(struct nullb_cmd *cmd, sector_t sector,
		     unsigned int nr_sectors)
{
	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);

	switch (zone->cond) {
	case BLK_ZONE_COND_FULL:
		/* Cannot write to a full zone */
		return BLK_STS_IOERR;
	case BLK_ZONE_COND_EMPTY:
	case BLK_ZONE_COND_IMP_OPEN:
	case BLK_ZONE_COND_EXP_OPEN:
	case BLK_ZONE_COND_CLOSED:
		/* Writes must be at the write pointer position */
		if (sector != zone->wp)
			return BLK_STS_IOERR;

		if (zone->cond != BLK_ZONE_COND_EXP_OPEN)
			zone->cond = BLK_ZONE_COND_IMP_OPEN;

		ret = null_process_cmd(cmd, REQ_OP_WRITE, sector, nr_sectors);
		if (ret != BLK_STS_OK)
			return ret;

		zone->wp += nr_sectors;
		if (zone->wp == zone->start + zone->len)
			zone->cond = BLK_ZONE_COND_FULL;
		return BLK_STS_OK;
	default:
		/* Invalid zone condition */
		return BLK_STS_IOERR;
	}
}

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 = null_zone_no(dev, sector);
	struct blk_zone *zone = &dev->zones[zone_no];
	size_t i;

	switch (op) {
	case REQ_OP_ZONE_RESET_ALL:
		for (i = 0; i < dev->nr_zones; i++) {
			if (zone[i].type == BLK_ZONE_TYPE_CONVENTIONAL)
				continue;
			zone[i].cond = BLK_ZONE_COND_EMPTY;
			zone[i].wp = zone[i].start;
		}
		break;
	case REQ_OP_ZONE_RESET:
		if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
			return BLK_STS_IOERR;

		zone->cond = BLK_ZONE_COND_EMPTY;
		zone->wp = zone->start;
		break;
	case REQ_OP_ZONE_OPEN:
		if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
			return BLK_STS_IOERR;
		if (zone->cond == BLK_ZONE_COND_FULL)
			return BLK_STS_IOERR;

		zone->cond = BLK_ZONE_COND_EXP_OPEN;
		break;
	case REQ_OP_ZONE_CLOSE:
		if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
			return BLK_STS_IOERR;
		if (zone->cond == BLK_ZONE_COND_FULL)
			return BLK_STS_IOERR;

		if (zone->wp == zone->start)
			zone->cond = BLK_ZONE_COND_EMPTY;
		else
			zone->cond = BLK_ZONE_COND_CLOSED;
		break;
	case REQ_OP_ZONE_FINISH:
		if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
			return BLK_STS_IOERR;

		zone->cond = BLK_ZONE_COND_FULL;
		zone->wp = zone->start + zone->len;
		break;
	default:
		return BLK_STS_NOTSUPP;
	}

	trace_nullb_zone_op(cmd, zone_no, zone->cond);
	return BLK_STS_OK;
}

blk_status_t null_process_zoned_cmd(struct nullb_cmd *cmd, enum req_opf op,
				    sector_t sector, sector_t nr_sectors)
{
	switch (op) {
	case REQ_OP_WRITE:
		return null_zone_write(cmd, sector, nr_sectors);
	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:
		return null_zone_mgmt(cmd, op, sector);
	default:
		return null_process_cmd(cmd, op, sector, nr_sectors);
	}
}
Commit	Line	Data
ca4b2a01 MB	1	// SPDX-License-Identifier: GPL-2.0
	2	#include <linux/vmalloc.h>
	3	#include "null_blk.h"
	4
766c3297 CK	5	#define CREATE_TRACE_POINTS
	6	#include "null_blk_trace.h"
	7
ca4b2a01 MB	8	/* zone_size in MBs to sectors. */
	9	#define ZONE_SIZE_SHIFT 11
	10
	11	static inline unsigned int null_zone_no(struct nullb_device *dev, sector_t sect)
	12	{
	13	return sect >> ilog2(dev->zone_size_sects);
	14	}
	15
d205bde7	16	int null_init_zoned_dev(struct nullb_device dev, struct request_queue q)
ca4b2a01 MB	17	{
	18	sector_t dev_size = (sector_t)dev->size * 1024 * 1024;
	19	sector_t sector = 0;
	20	unsigned int i;
	21
	22	if (!is_power_of_2(dev->zone_size)) {
9c7eddf1	23	pr_err("zone_size must be power-of-two\n");
ca4b2a01 MB	24	return -EINVAL;
ca4b2a01 MB	25	}
e2748325 CK	26	if (dev->zone_size > dev->size) {
	27	pr_err("Zone size larger than device capacity\n");
	28	return -EINVAL;
	29	}
ca4b2a01 MB	30
	31	dev->zone_size_sects = dev->zone_size << ZONE_SIZE_SHIFT;
	32	dev->nr_zones = dev_size >>
	33	(SECTOR_SHIFT + ilog2(dev->zone_size_sects));
	34	dev->zones = kvmalloc_array(dev->nr_zones, sizeof(struct blk_zone),
	35	GFP_KERNEL \| __GFP_ZERO);
	36	if (!dev->zones)
	37	return -ENOMEM;
	38
ea2c18e1 MS	39	if (dev->zone_nr_conv >= dev->nr_zones) {
ea2c18e1 MS	40	dev->zone_nr_conv = dev->nr_zones - 1;
9c7eddf1	41	pr_info("changed the number of conventional zones to %u",
ea2c18e1 MS	42	dev->zone_nr_conv);
	43	}
	44
	45	for (i = 0; i < dev->zone_nr_conv; i++) {
	46	struct blk_zone *zone = &dev->zones[i];
	47
	48	zone->start = sector;
	49	zone->len = dev->zone_size_sects;
	50	zone->wp = zone->start + zone->len;
	51	zone->type = BLK_ZONE_TYPE_CONVENTIONAL;
	52	zone->cond = BLK_ZONE_COND_NOT_WP;
	53
	54	sector += dev->zone_size_sects;
	55	}
	56
	57	for (i = dev->zone_nr_conv; i < dev->nr_zones; i++) {
ca4b2a01 MB	58	struct blk_zone *zone = &dev->zones[i];
	59
	60	zone->start = zone->wp = sector;
	61	zone->len = dev->zone_size_sects;
	62	zone->type = BLK_ZONE_TYPE_SEQWRITE_REQ;
	63	zone->cond = BLK_ZONE_COND_EMPTY;
	64
	65	sector += dev->zone_size_sects;
	66	}
	67
d205bde7 DLM	68	q->limits.zoned = BLK_ZONED_HM;
	69	blk_queue_flag_set(QUEUE_FLAG_ZONE_RESETALL, q);
	70	blk_queue_required_elevator_features(q, ELEVATOR_F_ZBD_SEQ_WRITE);
	71
	72	return 0;
	73	}
	74
	75	int null_register_zoned_dev(struct nullb *nullb)
	76	{
	77	struct request_queue *q = nullb->q;
	78
	79	if (queue_is_mq(q))
	80	return blk_revalidate_disk_zones(nullb->disk);
	81
	82	blk_queue_chunk_sectors(q, nullb->dev->zone_size_sects);
	83	q->nr_zones = blkdev_nr_zones(nullb->disk);
	84
ca4b2a01 MB	85	return 0;
	86	}
	87
d205bde7	88	void null_free_zoned_dev(struct nullb_device *dev)
ca4b2a01 MB	89	{
	90	kvfree(dev->zones);
	91	}
	92
7fc8fb51	93	int null_report_zones(struct gendisk *disk, sector_t sector,
d4100351	94	unsigned int nr_zones, report_zones_cb cb, void *data)
ca4b2a01	95	{
e76239a3 CH	96	struct nullb *nullb = disk->private_data;
e76239a3 CH	97	struct nullb_device *dev = nullb->dev;
d4100351 CH	98	unsigned int first_zone, i;
	99	struct blk_zone zone;
	100	int error;
ca4b2a01	101
d4100351 CH	102	first_zone = null_zone_no(dev, sector);
	103	if (first_zone >= dev->nr_zones)
	104	return 0;
ca4b2a01	105
d4100351	106	nr_zones = min(nr_zones, dev->nr_zones - first_zone);
766c3297 CK	107	trace_nullb_report_zones(nullb, nr_zones);
766c3297 CK	108
d4100351 CH	109	for (i = 0; i < nr_zones; i++) {
	110	/*
	111	* Stacked DM target drivers will remap the zone information by
	112	* modifying the zone information passed to the report callback.
	113	* So use a local copy to avoid corruption of the device zone
	114	* array.
	115	*/
	116	memcpy(&zone, &dev->zones[first_zone + i],
	117	sizeof(struct blk_zone));
	118	error = cb(&zone, i, data);
	119	if (error)
	120	return error;
	121	}
ca4b2a01	122
d4100351	123	return nr_zones;
ca4b2a01 MB	124	}
ca4b2a01 MB	125
dd85b492 AJ	126	size_t null_zone_valid_read_len(struct nullb *nullb,
	127	sector_t sector, unsigned int len)
	128	{
	129	struct nullb_device *dev = nullb->dev;
	130	struct blk_zone *zone = &dev->zones[null_zone_no(dev, sector)];
	131	unsigned int nr_sectors = len >> SECTOR_SHIFT;
	132
	133	/* Read must be below the write pointer position */
	134	if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL \|\|
	135	sector + nr_sectors <= zone->wp)
	136	return len;
	137
	138	if (sector > zone->wp)
	139	return 0;
	140
	141	return (zone->wp - sector) << SECTOR_SHIFT;
	142	}
	143
fceb5d1b	144	static blk_status_t null_zone_write(struct nullb_cmd *cmd, sector_t sector,
b228ba1c	145	unsigned int nr_sectors)
ca4b2a01 MB	146	{
ca4b2a01 MB	147	struct nullb_device *dev = cmd->nq->dev;
ca4b2a01 MB	148	unsigned int zno = null_zone_no(dev, sector);
ca4b2a01 MB	149	struct blk_zone *zone = &dev->zones[zno];
9dd44c7e DLM	150	blk_status_t ret;
	151
	152	trace_nullb_zone_op(cmd, zno, zone->cond);
	153
	154	if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
	155	return null_process_cmd(cmd, REQ_OP_WRITE, sector, nr_sectors);
ca4b2a01 MB	156
	157	switch (zone->cond) {
	158	case BLK_ZONE_COND_FULL:
	159	/* Cannot write to a full zone */
fceb5d1b	160	return BLK_STS_IOERR;
ca4b2a01 MB	161	case BLK_ZONE_COND_EMPTY:
ca4b2a01 MB	162	case BLK_ZONE_COND_IMP_OPEN:
16c731fe DLM	163	case BLK_ZONE_COND_EXP_OPEN:
16c731fe DLM	164	case BLK_ZONE_COND_CLOSED:
ca4b2a01	165	/* Writes must be at the write pointer position */
fceb5d1b CK	166	if (sector != zone->wp)
fceb5d1b CK	167	return BLK_STS_IOERR;
ca4b2a01	168
16c731fe	169	if (zone->cond != BLK_ZONE_COND_EXP_OPEN)
ca4b2a01 MB	170	zone->cond = BLK_ZONE_COND_IMP_OPEN;
ca4b2a01 MB	171
9dd44c7e DLM	172	ret = null_process_cmd(cmd, REQ_OP_WRITE, sector, nr_sectors);
	173	if (ret != BLK_STS_OK)
	174	return ret;
	175
b228ba1c	176	zone->wp += nr_sectors;
ca4b2a01 MB	177	if (zone->wp == zone->start + zone->len)
ca4b2a01 MB	178	zone->cond = BLK_ZONE_COND_FULL;
9dd44c7e	179	return BLK_STS_OK;
ca4b2a01 MB	180	default:
ca4b2a01 MB	181	/* Invalid zone condition */
fceb5d1b	182	return BLK_STS_IOERR;
ca4b2a01 MB	183	}
	184	}
	185
da644b2c AJ	186	static blk_status_t null_zone_mgmt(struct nullb_cmd *cmd, enum req_opf op,
da644b2c AJ	187	sector_t sector)
ca4b2a01 MB	188	{
ca4b2a01 MB	189	struct nullb_device *dev = cmd->nq->dev;
766c3297 CK	190	unsigned int zone_no = null_zone_no(dev, sector);
766c3297 CK	191	struct blk_zone *zone = &dev->zones[zone_no];
a61dbfb1 CK	192	size_t i;
a61dbfb1 CK	193
da644b2c	194	switch (op) {
a61dbfb1 CK	195	case REQ_OP_ZONE_RESET_ALL:
	196	for (i = 0; i < dev->nr_zones; i++) {
	197	if (zone[i].type == BLK_ZONE_TYPE_CONVENTIONAL)
	198	continue;
	199	zone[i].cond = BLK_ZONE_COND_EMPTY;
	200	zone[i].wp = zone[i].start;
	201	}
	202	break;
	203	case REQ_OP_ZONE_RESET:
fceb5d1b CK	204	if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
fceb5d1b CK	205	return BLK_STS_IOERR;
ca4b2a01	206
a61dbfb1 CK	207	zone->cond = BLK_ZONE_COND_EMPTY;
	208	zone->wp = zone->start;
	209	break;
da644b2c AJ	210	case REQ_OP_ZONE_OPEN:
	211	if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
	212	return BLK_STS_IOERR;
	213	if (zone->cond == BLK_ZONE_COND_FULL)
	214	return BLK_STS_IOERR;
	215
	216	zone->cond = BLK_ZONE_COND_EXP_OPEN;
	217	break;
	218	case REQ_OP_ZONE_CLOSE:
	219	if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
	220	return BLK_STS_IOERR;
	221	if (zone->cond == BLK_ZONE_COND_FULL)
	222	return BLK_STS_IOERR;
	223
c7d776f8 DLM	224	if (zone->wp == zone->start)
	225	zone->cond = BLK_ZONE_COND_EMPTY;
	226	else
	227	zone->cond = BLK_ZONE_COND_CLOSED;
da644b2c AJ	228	break;
	229	case REQ_OP_ZONE_FINISH:
	230	if (zone->type == BLK_ZONE_TYPE_CONVENTIONAL)
	231	return BLK_STS_IOERR;
	232
	233	zone->cond = BLK_ZONE_COND_FULL;
	234	zone->wp = zone->start + zone->len;
	235	break;
a61dbfb1	236	default:
79a85e21	237	return BLK_STS_NOTSUPP;
ea2c18e1	238	}
766c3297 CK	239
766c3297 CK	240	trace_nullb_zone_op(cmd, zone_no, zone->cond);
fceb5d1b CK	241	return BLK_STS_OK;
	242	}
	243
9dd44c7e DLM	244	blk_status_t null_process_zoned_cmd(struct nullb_cmd *cmd, enum req_opf op,
9dd44c7e DLM	245	sector_t sector, sector_t nr_sectors)
fceb5d1b CK	246	{
	247	switch (op) {
	248	case REQ_OP_WRITE:
	249	return null_zone_write(cmd, sector, nr_sectors);
	250	case REQ_OP_ZONE_RESET:
	251	case REQ_OP_ZONE_RESET_ALL:
da644b2c AJ	252	case REQ_OP_ZONE_OPEN:
	253	case REQ_OP_ZONE_CLOSE:
	254	case REQ_OP_ZONE_FINISH:
	255	return null_zone_mgmt(cmd, op, sector);
fceb5d1b	256	default:
9dd44c7e	257	return null_process_cmd(cmd, op, sector, nr_sectors);
fceb5d1b	258	}
ca4b2a01	259	}