X-Git-Url: https://git.kernel.dk/?p=fio.git;a=blobdiff_plain;f=zbd.c;h=a834f94c03fa0f8dd80ea5be6a1d685977c1192d;hp=0dd5a6191f8244f50e58470a0493ee1c5aae79a9;hb=59b67452b9dab3ad2bc9f1d1dd17c5e3993a66ee;hpb=6463db6c1d3a2a961008e87a86d464b596886f1a

diff --git a/zbd.c b/zbd.c
index 0dd5a619..a834f94c 100644
--- a/zbd.c
+++ b/zbd.c
@@ -7,13 +7,11 @@
 #include <errno.h>
 #include <string.h>
 #include <stdlib.h>
-#include <dirent.h>
 #include <fcntl.h>
-#include <sys/ioctl.h>
 #include <sys/stat.h>
 #include <unistd.h>
-#include <linux/blkzoned.h>
 
+#include "os/os.h"
 #include "file.h"
 #include "fio.h"
 #include "lib/pow2.h"
@@ -21,8 +19,100 @@
 #include "oslib/asprintf.h"
 #include "smalloc.h"
 #include "verify.h"
+#include "pshared.h"
 #include "zbd.h"
 
+/**
+ * zbd_get_zoned_model - Get a device zoned model
+ * @td: FIO thread data
+ * @f: FIO file for which to get model information
+ */
+int zbd_get_zoned_model(struct thread_data *td, struct fio_file *f,
+			enum zbd_zoned_model *model)
+{
+	int ret;
+
+	if (td->io_ops && td->io_ops->get_zoned_model)
+		ret = td->io_ops->get_zoned_model(td, f, model);
+	else
+		ret = blkzoned_get_zoned_model(td, f, model);
+	if (ret < 0) {
+		td_verror(td, errno, "get zoned model failed");
+		log_err("%s: get zoned model failed (%d).\n",
+			f->file_name, errno);
+	}
+
+	return ret;
+}
+
+/**
+ * zbd_report_zones - Get zone information
+ * @td: FIO thread data.
+ * @f: FIO file for which to get zone information
+ * @offset: offset from which to report zones
+ * @zones: Array of struct zbd_zone
+ * @nr_zones: Size of @zones array
+ *
+ * Get zone information into @zones starting from the zone at offset @offset
+ * for the device specified by @f.
+ *
+ * Returns the number of zones reported upon success and a negative error code
+ * upon failure. If the zone report is empty, always assume an error (device
+ * problem) and return -EIO.
+ */
+int zbd_report_zones(struct thread_data *td, struct fio_file *f,
+		     uint64_t offset, struct zbd_zone *zones,
+		     unsigned int nr_zones)
+{
+	int ret;
+
+	if (td->io_ops && td->io_ops->report_zones)
+		ret = td->io_ops->report_zones(td, f, offset, zones, nr_zones);
+	else
+		ret = blkzoned_report_zones(td, f, offset, zones, nr_zones);
+	if (ret < 0) {
+		td_verror(td, errno, "report zones failed");
+		log_err("%s: report zones from sector %llu failed (%d).\n",
+			f->file_name, (unsigned long long)offset >> 9, errno);
+	} else if (ret == 0) {
+		td_verror(td, errno, "Empty zone report");
+		log_err("%s: report zones from sector %llu is empty.\n",
+			f->file_name, (unsigned long long)offset >> 9);
+		ret = -EIO;
+	}
+
+	return ret;
+}
+
+/**
+ * zbd_reset_wp - reset the write pointer of a range of zones
+ * @td: FIO thread data.
+ * @f: FIO file for which to reset zones
+ * @offset: Starting offset of the first zone to reset
+ * @length: Length of the range of zones to reset
+ *
+ * Reset the write pointer of all zones in the range @offset...@offset+@length.
+ * Returns 0 upon success and a negative error code upon failure.
+ */
+int zbd_reset_wp(struct thread_data *td, struct fio_file *f,
+		 uint64_t offset, uint64_t length)
+{
+	int ret;
+
+	if (td->io_ops && td->io_ops->reset_wp)
+		ret = td->io_ops->reset_wp(td, f, offset, length);
+	else
+		ret = blkzoned_reset_wp(td, f, offset, length);
+	if (ret < 0) {
+		td_verror(td, errno, "resetting wp failed");
+		log_err("%s: resetting wp for %llu sectors at sector %llu failed (%d).\n",
+			f->file_name, (unsigned long long)length >> 9,
+			(unsigned long long)offset >> 9, errno);
+	}
+
+	return ret;
+}
+
 /**
  * zbd_zone_idx - convert an offset into a zone number
  * @f: file pointer.
@@ -41,6 +131,33 @@ static uint32_t zbd_zone_idx(const struct fio_file *f, uint64_t offset)
 	return min(zone_idx, f->zbd_info->nr_zones);
 }
 
+/**
+ * zbd_zone_swr - Test whether a zone requires sequential writes
+ * @z: zone info pointer.
+ */
+static inline bool zbd_zone_swr(struct fio_zone_info *z)
+{
+	return z->type == ZBD_ZONE_TYPE_SWR;
+}
+
+/**
+ * zbd_zone_end - Return zone end location
+ * @z: zone info pointer.
+ */
+static inline uint64_t zbd_zone_end(const struct fio_zone_info *z)
+{
+	return (z+1)->start;
+}
+
+/**
+ * zbd_zone_capacity_end - Return zone capacity limit end location
+ * @z: zone info pointer.
+ */
+static inline uint64_t zbd_zone_capacity_end(const struct fio_zone_info *z)
+{
+	return z->start + z->capacity;
+}
+
 /**
  * zbd_zone_full - verify whether a minimum number of bytes remain in a zone
  * @f: file pointer.
@@ -54,8 +171,40 @@ static bool zbd_zone_full(const struct fio_file *f, struct fio_zone_info *z,
 {
 	assert((required & 511) == 0);
 
-	return z->type == BLK_ZONE_TYPE_SEQWRITE_REQ &&
-		z->wp + required > z->start + f->zbd_info->zone_size;
+	return zbd_zone_swr(z) &&
+		z->wp + required > zbd_zone_capacity_end(z);
+}
+
+static void zone_lock(struct thread_data *td, struct fio_file *f, struct fio_zone_info *z)
+{
+	struct zoned_block_device_info *zbd = f->zbd_info;
+	uint32_t nz = z - zbd->zone_info;
+
+	/* A thread should never lock zones outside its working area. */
+	assert(f->min_zone <= nz && nz < f->max_zone);
+
+	/*
+	 * Lock the io_u target zone. The zone will be unlocked if io_u offset
+	 * is changed or when io_u completes and zbd_put_io() executed.
+	 * To avoid multiple jobs doing asynchronous I/Os from deadlocking each
+	 * other waiting for zone locks when building an io_u batch, first
+	 * only trylock the zone. If the zone is already locked by another job,
+	 * process the currently queued I/Os so that I/O progress is made and
+	 * zones unlocked.
+	 */
+	if (pthread_mutex_trylock(&z->mutex) != 0) {
+		if (!td_ioengine_flagged(td, FIO_SYNCIO))
+			io_u_quiesce(td);
+		pthread_mutex_lock(&z->mutex);
+	}
+}
+
+static inline void zone_unlock(struct fio_zone_info *z)
+{
+	int ret;
+
+	ret = pthread_mutex_unlock(&z->mutex);
+	assert(!ret);
 }
 
 static bool is_valid_offset(const struct fio_file *f, uint64_t offset)
@@ -63,6 +212,12 @@ static bool is_valid_offset(const struct fio_file *f, uint64_t offset)
 	return (uint64_t)(offset - f->file_offset) < f->io_size;
 }
 
+static inline struct fio_zone_info *get_zone(const struct fio_file *f,
+					     unsigned int zone_nr)
+{
+	return &f->zbd_info->zone_info[zone_nr];
+}
+
 /* Verify whether direct I/O is used for all host-managed zoned drives. */
 static bool zbd_using_direct_io(void)
 {
@@ -75,7 +230,7 @@ static bool zbd_using_direct_io(void)
 			continue;
 		for_each_file(td, f, j) {
 			if (f->zbd_info &&
-			    f->zbd_info->model == ZBD_DM_HOST_MANAGED)
+			    f->zbd_info->model == ZBD_HOST_MANAGED)
 				return false;
 		}
 	}
@@ -94,8 +249,7 @@ static bool zbd_is_seq_job(struct fio_file *f)
 	zone_idx_b = zbd_zone_idx(f, f->file_offset);
 	zone_idx_e = zbd_zone_idx(f, f->file_offset + f->io_size - 1);
 	for (zone_idx = zone_idx_b; zone_idx <= zone_idx_e; zone_idx++)
-		if (f->zbd_info->zone_info[zone_idx].type ==
-		    BLK_ZONE_TYPE_SEQWRITE_REQ)
+		if (zbd_zone_swr(get_zone(f, zone_idx)))
 			return true;
 
 	return false;
@@ -146,9 +300,10 @@ static bool zbd_verify_sizes(void)
 			}
 
 			zone_idx = zbd_zone_idx(f, f->file_offset);
-			z = &f->zbd_info->zone_info[zone_idx];
-			if (f->file_offset != z->start) {
-				new_offset = (z+1)->start;
+			z = get_zone(f, zone_idx);
+			if ((f->file_offset != z->start) &&
+			    (td->o.td_ddir != TD_DDIR_READ)) {
+				new_offset = zbd_zone_end(z);
 				if (new_offset >= f->file_offset + f->io_size) {
 					log_info("%s: io_size must be at least one zone\n",
 						 f->file_name);
@@ -161,9 +316,10 @@ static bool zbd_verify_sizes(void)
 				f->file_offset = new_offset;
 			}
 			zone_idx = zbd_zone_idx(f, f->file_offset + f->io_size);
-			z = &f->zbd_info->zone_info[zone_idx];
+			z = get_zone(f, zone_idx);
 			new_end = z->start;
-			if (f->file_offset + f->io_size != new_end) {
+			if ((td->o.td_ddir != TD_DDIR_READ) &&
+			    (f->file_offset + f->io_size != new_end)) {
 				if (new_end <= f->file_offset) {
 					log_info("%s: io_size must be at least one zone\n",
 						 f->file_name);
@@ -174,6 +330,10 @@ static bool zbd_verify_sizes(void)
 					 (unsigned long long) new_end - f->file_offset);
 				f->io_size = new_end - f->file_offset;
 			}
+
+			f->min_zone = zbd_zone_idx(f, f->file_offset);
+			f->max_zone = zbd_zone_idx(f, f->file_offset + f->io_size);
+			assert(f->min_zone < f->max_zone);
 		}
 	}
 
@@ -192,7 +352,7 @@ static bool zbd_verify_bs(void)
 			if (!f->zbd_info)
 				continue;
 			zone_size = f->zbd_info->zone_size;
-			for (k = 0; k < ARRAY_SIZE(td->o.bs); k++) {
+			for (k = 0; k < FIO_ARRAY_SIZE(td->o.bs); k++) {
 				if (td->o.verify != VERIFY_NONE &&
 				    zone_size % td->o.bs[k] != 0) {
 					log_info("%s: block size %llu is not a divisor of the zone size %d\n",
@@ -206,119 +366,6 @@ static bool zbd_verify_bs(void)
 	return true;
 }
 
-/*
- * Read zone information into @buf starting from sector @start_sector.
- * @fd is a file descriptor that refers to a block device and @bufsz is the
- * size of @buf.
- *
- * Returns 0 upon success and a negative error code upon failure.
- * If the zone report is empty, always assume an error (device problem) and
- * return -EIO.
- */
-static int read_zone_info(int fd, uint64_t start_sector,
-			  void *buf, unsigned int bufsz)
-{
-	struct blk_zone_report *hdr = buf;
-	int ret;
-
-	if (bufsz < sizeof(*hdr))
-		return -EINVAL;
-
-	memset(hdr, 0, sizeof(*hdr));
-
-	hdr->nr_zones = (bufsz - sizeof(*hdr)) / sizeof(struct blk_zone);
-	hdr->sector = start_sector;
-	ret = ioctl(fd, BLKREPORTZONE, hdr);
-	if (ret)
-		return -errno;
-	if (!hdr->nr_zones)
-		return -EIO;
-	return 0;
-}
-
-/*
- * Read up to 255 characters from the first line of a file. Strip the trailing
- * newline.
- */
-static char *read_file(const char *path)
-{
-	char line[256], *p = line;
-	FILE *f;
-
-	f = fopen(path, "rb");
-	if (!f)
-		return NULL;
-	if (!fgets(line, sizeof(line), f))
-		line[0] = '\0';
-	strsep(&p, "\n");
-	fclose(f);
-
-	return strdup(line);
-}
-
-static enum blk_zoned_model get_zbd_model(const char *file_name)
-{
-	enum blk_zoned_model model = ZBD_DM_NONE;
-	char *zoned_attr_path = NULL;
-	char *model_str = NULL;
-	struct stat statbuf;
-	char *sys_devno_path = NULL;
-	char *part_attr_path = NULL;
-	char *part_str = NULL;
-	char sys_path[PATH_MAX];
-	ssize_t sz;
-	char *delim = NULL;
-
-	if (stat(file_name, &statbuf) < 0)
-		goto out;
-
-	if (asprintf(&sys_devno_path, "/sys/dev/block/%d:%d",
-		     major(statbuf.st_rdev), minor(statbuf.st_rdev)) < 0)
-		goto out;
-
-	sz = readlink(sys_devno_path, sys_path, sizeof(sys_path) - 1);
-	if (sz < 0)
-		goto out;
-	sys_path[sz] = '\0';
-
-	/*
-	 * If the device is a partition device, cut the device name in the
-	 * canonical sysfs path to obtain the sysfs path of the holder device.
-	 *   e.g.:  /sys/devices/.../sda/sda1 -> /sys/devices/.../sda
-	 */
-	if (asprintf(&part_attr_path, "/sys/dev/block/%s/partition",
-		     sys_path) < 0)
-		goto out;
-	part_str = read_file(part_attr_path);
-	if (part_str && *part_str == '1') {
-		delim = strrchr(sys_path, '/');
-		if (!delim)
-			goto out;
-		*delim = '\0';
-	}
-
-	if (asprintf(&zoned_attr_path,
-		     "/sys/dev/block/%s/queue/zoned", sys_path) < 0)
-		goto out;
-
-	model_str = read_file(zoned_attr_path);
-	if (!model_str)
-		goto out;
-	dprint(FD_ZBD, "%s: zbd model string: %s\n", file_name, model_str);
-	if (strcmp(model_str, "host-aware") == 0)
-		model = ZBD_DM_HOST_AWARE;
-	else if (strcmp(model_str, "host-managed") == 0)
-		model = ZBD_DM_HOST_MANAGED;
-
-out:
-	free(model_str);
-	free(zoned_attr_path);
-	free(part_str);
-	free(part_attr_path);
-	free(sys_devno_path);
-	return model;
-}
-
 static int ilog2(uint64_t i)
 {
 	int log = -1;
@@ -339,8 +386,8 @@ static int init_zone_info(struct thread_data *td, struct fio_file *f)
 	uint32_t nr_zones;
 	struct fio_zone_info *p;
 	uint64_t zone_size = td->o.zone_size;
+	uint64_t zone_capacity = td->o.zone_capacity;
 	struct zoned_block_device_info *zbd_info = NULL;
-	pthread_mutexattr_t attr;
 	int i;
 
 	if (zone_size == 0) {
@@ -355,24 +402,33 @@ static int init_zone_info(struct thread_data *td, struct fio_file *f)
 		return 1;
 	}
 
+	if (zone_capacity == 0)
+		zone_capacity = zone_size;
+
+	if (zone_capacity > zone_size) {
+		log_err("%s: job parameter zonecapacity %llu is larger than zone size %llu\n",
+			f->file_name, (unsigned long long) td->o.zone_capacity,
+			(unsigned long long) td->o.zone_size);
+		return 1;
+	}
+
 	nr_zones = (f->real_file_size + zone_size - 1) / zone_size;
 	zbd_info = scalloc(1, sizeof(*zbd_info) +
 			   (nr_zones + 1) * sizeof(zbd_info->zone_info[0]));
 	if (!zbd_info)
 		return -ENOMEM;
 
-	pthread_mutexattr_init(&attr);
-	pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_RECURSIVE);
-	pthread_mutexattr_setpshared(&attr, true);
-	pthread_mutex_init(&zbd_info->mutex, &attr);
+	mutex_init_pshared(&zbd_info->mutex);
 	zbd_info->refcount = 1;
 	p = &zbd_info->zone_info[0];
 	for (i = 0; i < nr_zones; i++, p++) {
-		pthread_mutex_init(&p->mutex, &attr);
+		mutex_init_pshared_with_type(&p->mutex,
+					     PTHREAD_MUTEX_RECURSIVE);
 		p->start = i * zone_size;
-		p->wp = p->start + zone_size;
-		p->type = BLK_ZONE_TYPE_SEQWRITE_REQ;
-		p->cond = BLK_ZONE_COND_EMPTY;
+		p->wp = p->start;
+		p->type = ZBD_ZONE_TYPE_SWR;
+		p->cond = ZBD_ZONE_COND_EMPTY;
+		p->capacity = zone_capacity;
 	}
 	/* a sentinel */
 	p->start = nr_zones * zone_size;
@@ -380,58 +436,42 @@ static int init_zone_info(struct thread_data *td, struct fio_file *f)
 	f->zbd_info = zbd_info;
 	f->zbd_info->zone_size = zone_size;
 	f->zbd_info->zone_size_log2 = is_power_of_2(zone_size) ?
-		ilog2(zone_size) : -1;
+		ilog2(zone_size) : 0;
 	f->zbd_info->nr_zones = nr_zones;
-	pthread_mutexattr_destroy(&attr);
 	return 0;
 }
 
 /*
- * Parse the BLKREPORTZONE output and store it in f->zbd_info. Must be called
- * only for devices that support this ioctl, namely zoned block devices.
+ * Maximum number of zones to report in one operation.
+ */
+#define ZBD_REPORT_MAX_ZONES	8192U
+
+/*
+ * Parse the device zone report and store it in f->zbd_info. Must be called
+ * only for devices that are zoned, namely those with a model != ZBD_NONE.
  */
 static int parse_zone_info(struct thread_data *td, struct fio_file *f)
 {
-	const unsigned int bufsz = sizeof(struct blk_zone_report) +
-		4096 * sizeof(struct blk_zone);
-	uint32_t nr_zones;
-	struct blk_zone_report *hdr;
-	const struct blk_zone *z;
+	int nr_zones, nrz;
+	struct zbd_zone *zones, *z;
 	struct fio_zone_info *p;
-	uint64_t zone_size, start_sector;
+	uint64_t zone_size, offset;
 	struct zoned_block_device_info *zbd_info = NULL;
-	pthread_mutexattr_t attr;
-	void *buf;
-	int fd, i, j, ret = 0;
-
-	pthread_mutexattr_init(&attr);
-	pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_RECURSIVE);
-	pthread_mutexattr_setpshared(&attr, true);
+	int i, j, ret = -ENOMEM;
 
-	buf = malloc(bufsz);
-	if (!buf)
+	zones = calloc(ZBD_REPORT_MAX_ZONES, sizeof(struct zbd_zone));
+	if (!zones)
 		goto out;
 
-	fd = open(f->file_name, O_RDONLY | O_LARGEFILE);
-	if (fd < 0) {
-		ret = -errno;
-		goto free;
+	nrz = zbd_report_zones(td, f, 0, zones, ZBD_REPORT_MAX_ZONES);
+	if (nrz < 0) {
+		ret = nrz;
+		log_info("fio: report zones (offset 0) failed for %s (%d).\n",
+			 f->file_name, -ret);
+		goto out;
 	}
 
-	ret = read_zone_info(fd, 0, buf, bufsz);
-	if (ret < 0) {
-		log_info("fio: BLKREPORTZONE(%lu) failed for %s (%d).\n",
-			 0UL, f->file_name, -ret);
-		goto close;
-	}
-	hdr = buf;
-	if (hdr->nr_zones < 1) {
-		log_info("fio: %s has invalid zone information.\n",
-			 f->file_name);
-		goto close;
-	}
-	z = (void *)(hdr + 1);
-	zone_size = z->len << 9;
+	zone_size = zones[0].len;
 	nr_zones = (f->real_file_size + zone_size - 1) / zone_size;
 
 	if (td->o.zone_size == 0) {
@@ -441,7 +481,7 @@ static int parse_zone_info(struct thread_data *td, struct fio_file *f)
 			f->file_name, (unsigned long long) td->o.zone_size,
 			(unsigned long long) zone_size);
 		ret = -EINVAL;
-		goto close;
+		goto out;
 	}
 
 	dprint(FD_ZBD, "Device %s has %d zones of size %llu KB\n", f->file_name,
@@ -449,26 +489,27 @@ static int parse_zone_info(struct thread_data *td, struct fio_file *f)
 
 	zbd_info = scalloc(1, sizeof(*zbd_info) +
 			   (nr_zones + 1) * sizeof(zbd_info->zone_info[0]));
-	ret = -ENOMEM;
 	if (!zbd_info)
-		goto close;
-	pthread_mutex_init(&zbd_info->mutex, &attr);
+		goto out;
+	mutex_init_pshared(&zbd_info->mutex);
 	zbd_info->refcount = 1;
 	p = &zbd_info->zone_info[0];
-	for (start_sector = 0, j = 0; j < nr_zones;) {
-		z = (void *)(hdr + 1);
-		for (i = 0; i < hdr->nr_zones; i++, j++, z++, p++) {
-			pthread_mutex_init(&p->mutex, &attr);
-			p->start = z->start << 9;
+	for (offset = 0, j = 0; j < nr_zones;) {
+		z = &zones[0];
+		for (i = 0; i < nrz; i++, j++, z++, p++) {
+			mutex_init_pshared_with_type(&p->mutex,
+						     PTHREAD_MUTEX_RECURSIVE);
+			p->start = z->start;
+			p->capacity = z->capacity;
 			switch (z->cond) {
-			case BLK_ZONE_COND_NOT_WP:
-			case BLK_ZONE_COND_FULL:
-				p->wp = p->start + zone_size;
+			case ZBD_ZONE_COND_NOT_WP:
+			case ZBD_ZONE_COND_FULL:
+				p->wp = p->start + p->capacity;
 				break;
 			default:
 				assert(z->start <= z->wp);
-				assert(z->wp <= z->start + (zone_size >> 9));
-				p->wp = z->wp << 9;
+				assert(z->wp <= z->start + zone_size);
+				p->wp = z->wp;
 				break;
 			}
 			p->type = z->type;
@@ -477,38 +518,38 @@ static int parse_zone_info(struct thread_data *td, struct fio_file *f)
 				log_info("%s: invalid zone data\n",
 					 f->file_name);
 				ret = -EINVAL;
-				goto close;
+				goto out;
 			}
 		}
 		z--;
-		start_sector = z->start + z->len;
+		offset = z->start + z->len;
 		if (j >= nr_zones)
 			break;
-		ret = read_zone_info(fd, start_sector, buf, bufsz);
-		if (ret < 0) {
-			log_info("fio: BLKREPORTZONE(%llu) failed for %s (%d).\n",
-				 (unsigned long long) start_sector, f->file_name, -ret);
-			goto close;
+		nrz = zbd_report_zones(td, f, offset,
+					    zones, ZBD_REPORT_MAX_ZONES);
+		if (nrz < 0) {
+			ret = nrz;
+			log_info("fio: report zones (offset %llu) failed for %s (%d).\n",
+			 	 (unsigned long long)offset,
+				 f->file_name, -ret);
+			goto out;
 		}
 	}
+
 	/* a sentinel */
-	zbd_info->zone_info[nr_zones].start = start_sector << 9;
+	zbd_info->zone_info[nr_zones].start = offset;
 
 	f->zbd_info = zbd_info;
 	f->zbd_info->zone_size = zone_size;
 	f->zbd_info->zone_size_log2 = is_power_of_2(zone_size) ?
-		ilog2(zone_size) : -1;
+		ilog2(zone_size) : 0;
 	f->zbd_info->nr_zones = nr_zones;
 	zbd_info = NULL;
 	ret = 0;
 
-close:
-	sfree(zbd_info);
-	close(fd);
-free:
-	free(buf);
 out:
-	pthread_mutexattr_destroy(&attr);
+	sfree(zbd_info);
+	free(zones);
 	return ret;
 }
 
@@ -519,23 +560,35 @@ out:
  */
 static int zbd_create_zone_info(struct thread_data *td, struct fio_file *f)
 {
-	enum blk_zoned_model zbd_model;
-	int ret = 0;
+	enum zbd_zoned_model zbd_model;
+	int ret;
 
 	assert(td->o.zone_mode == ZONE_MODE_ZBD);
 
-	zbd_model = get_zbd_model(f->file_name);
+	ret = zbd_get_zoned_model(td, f, &zbd_model);
+	if (ret)
+		return ret;
+
 	switch (zbd_model) {
-	case ZBD_DM_HOST_AWARE:
-	case ZBD_DM_HOST_MANAGED:
+	case ZBD_IGNORE:
+		return 0;
+	case ZBD_HOST_AWARE:
+	case ZBD_HOST_MANAGED:
 		ret = parse_zone_info(td, f);
 		break;
-	case ZBD_DM_NONE:
+	case ZBD_NONE:
 		ret = init_zone_info(td, f);
 		break;
+	default:
+		td_verror(td, EINVAL, "Unsupported zoned model");
+		log_err("Unsupported zoned model\n");
+		return -EINVAL;
 	}
-	if (ret == 0)
+
+	if (ret == 0) {
 		f->zbd_info->model = zbd_model;
+		f->zbd_info->max_open_zones = td->o.max_open_zones;
+	}
 	return ret;
 }
 
@@ -543,8 +596,7 @@ void zbd_free_zone_info(struct fio_file *f)
 {
 	uint32_t refcount;
 
-	if (!f->zbd_info)
-		return;
+	assert(f->zbd_info);
 
 	pthread_mutex_lock(&f->zbd_info->mutex);
 	refcount = --f->zbd_info->refcount;
@@ -589,14 +641,17 @@ static int zbd_init_zone_info(struct thread_data *td, struct fio_file *file)
 	return ret;
 }
 
-int zbd_init(struct thread_data *td)
+static bool zbd_open_zone(struct thread_data *td, const struct fio_file *f,
+			  uint32_t zone_idx);
+static int zbd_reset_zone(struct thread_data *td, struct fio_file *f,
+			  struct fio_zone_info *z);
+
+int zbd_setup_files(struct thread_data *td)
 {
 	struct fio_file *f;
 	int i;
 
 	for_each_file(td, f, i) {
-		if (f->filetype != FIO_TYPE_BLOCK)
-			continue;
 		if (zbd_init_zone_info(td, f))
 			return 1;
 	}
@@ -612,86 +667,123 @@ int zbd_init(struct thread_data *td)
 	if (!zbd_verify_bs())
 		return 1;
 
+	for_each_file(td, f, i) {
+		struct zoned_block_device_info *zbd = f->zbd_info;
+		struct fio_zone_info *z;
+		int zi;
+
+		if (!zbd)
+			continue;
+
+		zbd->max_open_zones = zbd->max_open_zones ?: ZBD_MAX_OPEN_ZONES;
+
+		if (td->o.max_open_zones > 0 &&
+		    zbd->max_open_zones != td->o.max_open_zones) {
+			log_err("Different 'max_open_zones' values\n");
+			return 1;
+		}
+		if (zbd->max_open_zones > ZBD_MAX_OPEN_ZONES) {
+			log_err("'max_open_zones' value is limited by %u\n", ZBD_MAX_OPEN_ZONES);
+			return 1;
+		}
+
+		for (zi = f->min_zone; zi < f->max_zone; zi++) {
+			z = &zbd->zone_info[zi];
+			if (z->cond != ZBD_ZONE_COND_IMP_OPEN &&
+			    z->cond != ZBD_ZONE_COND_EXP_OPEN)
+				continue;
+			if (zbd_open_zone(td, f, zi))
+				continue;
+			/*
+			 * If the number of open zones exceeds specified limits,
+			 * reset all extra open zones.
+			 */
+			if (zbd_reset_zone(td, f, z) < 0) {
+				log_err("Failed to reest zone %d\n", zi);
+				return 1;
+			}
+		}
+	}
+
 	return 0;
 }
 
+static inline unsigned int zbd_zone_nr(const struct fio_file *f,
+				       struct fio_zone_info *zone)
+{
+	return zone - f->zbd_info->zone_info;
+}
+
 /**
- * zbd_reset_range - reset zones for a range of sectors
+ * zbd_reset_zone - reset the write pointer of a single zone
  * @td: FIO thread data.
- * @f: Fio file for which to reset zones
- * @sector: Starting sector in units of 512 bytes
- * @nr_sectors: Number of sectors in units of 512 bytes
+ * @f: FIO file associated with the disk for which to reset a write pointer.
+ * @z: Zone to reset.
  *
  * Returns 0 upon success and a negative error code upon failure.
+ *
+ * The caller must hold z->mutex.
  */
-static int zbd_reset_range(struct thread_data *td, const struct fio_file *f,
-			   uint64_t offset, uint64_t length)
+static int zbd_reset_zone(struct thread_data *td, struct fio_file *f,
+			  struct fio_zone_info *z)
 {
-	struct blk_zone_range zr = {
-		.sector         = offset >> 9,
-		.nr_sectors     = length >> 9,
-	};
-	uint32_t zone_idx_b, zone_idx_e;
-	struct fio_zone_info *zb, *ze, *z;
+	uint64_t offset = z->start;
+	uint64_t length = (z+1)->start - offset;
 	int ret = 0;
 
-	assert(f->fd != -1);
+	if (z->wp == z->start)
+		return 0;
+
 	assert(is_valid_offset(f, offset + length - 1));
+
+	dprint(FD_ZBD, "%s: resetting wp of zone %u.\n", f->file_name,
+		zbd_zone_nr(f, z));
 	switch (f->zbd_info->model) {
-	case ZBD_DM_HOST_AWARE:
-	case ZBD_DM_HOST_MANAGED:
-		ret = ioctl(f->fd, BLKRESETZONE, &zr);
-		if (ret < 0) {
-			td_verror(td, errno, "resetting wp failed");
-			log_err("%s: resetting wp for %llu sectors at sector %llu failed (%d).\n",
-				f->file_name, zr.nr_sectors, zr.sector, errno);
+	case ZBD_HOST_AWARE:
+	case ZBD_HOST_MANAGED:
+		ret = zbd_reset_wp(td, f, offset, length);
+		if (ret < 0)
 			return ret;
-		}
 		break;
-	case ZBD_DM_NONE:
+	default:
 		break;
 	}
 
-	zone_idx_b = zbd_zone_idx(f, offset);
-	zb = &f->zbd_info->zone_info[zone_idx_b];
-	zone_idx_e = zbd_zone_idx(f, offset + length);
-	ze = &f->zbd_info->zone_info[zone_idx_e];
-	for (z = zb; z < ze; z++) {
-		pthread_mutex_lock(&z->mutex);
-		pthread_mutex_lock(&f->zbd_info->mutex);
-		f->zbd_info->sectors_with_data -= z->wp - z->start;
-		pthread_mutex_unlock(&f->zbd_info->mutex);
-		z->wp = z->start;
-		z->verify_block = 0;
-		pthread_mutex_unlock(&z->mutex);
-	}
+	pthread_mutex_lock(&f->zbd_info->mutex);
+	f->zbd_info->sectors_with_data -= z->wp - z->start;
+	pthread_mutex_unlock(&f->zbd_info->mutex);
+	z->wp = z->start;
+	z->verify_block = 0;
 
-	td->ts.nr_zone_resets += ze - zb;
+	td->ts.nr_zone_resets++;
 
 	return ret;
 }
 
-static unsigned int zbd_zone_nr(struct zoned_block_device_info *zbd_info,
-				struct fio_zone_info *zone)
+/* The caller must hold f->zbd_info->mutex */
+static void zbd_close_zone(struct thread_data *td, const struct fio_file *f,
+			   unsigned int zone_idx)
 {
-	return zone - zbd_info->zone_info;
-}
+	uint32_t open_zone_idx = 0;
 
-/**
- * zbd_reset_zone - reset the write pointer of a single zone
- * @td: FIO thread data.
- * @f: FIO file associated with the disk for which to reset a write pointer.
- * @z: Zone to reset.
- *
- * Returns 0 upon success and a negative error code upon failure.
- */
-static int zbd_reset_zone(struct thread_data *td, const struct fio_file *f,
-			  struct fio_zone_info *z)
-{
-	dprint(FD_ZBD, "%s: resetting wp of zone %u.\n", f->file_name,
-		zbd_zone_nr(f->zbd_info, z));
+	for (; open_zone_idx < f->zbd_info->num_open_zones; open_zone_idx++) {
+		if (f->zbd_info->open_zones[open_zone_idx] == zone_idx)
+			break;
+	}
+	if (open_zone_idx == f->zbd_info->num_open_zones) {
+		dprint(FD_ZBD, "%s: zone %d is not open\n",
+		       f->file_name, zone_idx);
+		return;
+	}
 
-	return zbd_reset_range(td, f, z->start, (z+1)->start - z->start);
+	dprint(FD_ZBD, "%s: closing zone %d\n", f->file_name, zone_idx);
+	memmove(f->zbd_info->open_zones + open_zone_idx,
+		f->zbd_info->open_zones + open_zone_idx + 1,
+		(ZBD_MAX_OPEN_ZONES - (open_zone_idx + 1)) *
+		sizeof(f->zbd_info->open_zones[0]));
+	f->zbd_info->num_open_zones--;
+	td->num_open_zones--;
+	get_zone(f, zone_idx)->open = 0;
 }
 
 /*
@@ -712,24 +804,32 @@ static int zbd_reset_zones(struct thread_data *td, struct fio_file *f,
 	bool reset_wp;
 	int res = 0;
 
+	assert(min_bs);
+
 	dprint(FD_ZBD, "%s: examining zones %u .. %u\n", f->file_name,
-		zbd_zone_nr(f->zbd_info, zb), zbd_zone_nr(f->zbd_info, ze));
-	assert(f->fd != -1);
+		zbd_zone_nr(f, zb), zbd_zone_nr(f, ze));
 	for (z = zb; z < ze; z++) {
-		pthread_mutex_lock(&z->mutex);
-		if (z->type == BLK_ZONE_TYPE_SEQWRITE_REQ) {
-			reset_wp = all_zones ? z->wp != z->start :
-					(td->o.td_ddir & TD_DDIR_WRITE) &&
-					z->wp % min_bs != 0;
-			if (reset_wp) {
-				dprint(FD_ZBD, "%s: resetting zone %u\n",
-				       f->file_name,
-				       zbd_zone_nr(f->zbd_info, z));
-				if (zbd_reset_zone(td, f, z) < 0)
-					res = 1;
-			}
+		uint32_t nz = zbd_zone_nr(f, z);
+
+		if (!zbd_zone_swr(z))
+			continue;
+		zone_lock(td, f, z);
+		if (all_zones) {
+			pthread_mutex_lock(&f->zbd_info->mutex);
+			zbd_close_zone(td, f, nz);
+			pthread_mutex_unlock(&f->zbd_info->mutex);
+
+			reset_wp = z->wp != z->start;
+		} else {
+			reset_wp = z->wp % min_bs != 0;
+		}
+		if (reset_wp) {
+			dprint(FD_ZBD, "%s: resetting zone %u\n",
+			       f->file_name, zbd_zone_nr(f, z));
+			if (zbd_reset_zone(td, f, z) < 0)
+				res = 1;
 		}
-		pthread_mutex_unlock(&z->mutex);
+		zone_unlock(z);
 	}
 
 	return res;
@@ -739,14 +839,20 @@ static int zbd_reset_zones(struct thread_data *td, struct fio_file *f,
  * Reset zbd_info.write_cnt, the counter that counts down towards the next
  * zone reset.
  */
-static void zbd_reset_write_cnt(const struct thread_data *td,
-				const struct fio_file *f)
+static void _zbd_reset_write_cnt(const struct thread_data *td,
+				 const struct fio_file *f)
 {
 	assert(0 <= td->o.zrf.u.f && td->o.zrf.u.f <= 1);
 
-	pthread_mutex_lock(&f->zbd_info->mutex);
 	f->zbd_info->write_cnt = td->o.zrf.u.f ?
 		min(1.0 / td->o.zrf.u.f, 0.0 + UINT_MAX) : UINT_MAX;
+}
+
+static void zbd_reset_write_cnt(const struct thread_data *td,
+				const struct fio_file *f)
+{
+	pthread_mutex_lock(&f->zbd_info->mutex);
+	_zbd_reset_write_cnt(td, f);
 	pthread_mutex_unlock(&f->zbd_info->mutex);
 }
 
@@ -760,7 +866,7 @@ static bool zbd_dec_and_reset_write_cnt(const struct thread_data *td,
 	if (f->zbd_info->write_cnt)
 		write_cnt = --f->zbd_info->write_cnt;
 	if (write_cnt == 0)
-		zbd_reset_write_cnt(td, f);
+		_zbd_reset_write_cnt(td, f);
 	pthread_mutex_unlock(&f->zbd_info->mutex);
 
 	return write_cnt == 0;
@@ -777,9 +883,8 @@ static uint64_t zbd_process_swd(const struct fio_file *f, enum swd_action a)
 	struct fio_zone_info *zb, *ze, *z;
 	uint64_t swd = 0;
 
-	zb = &f->zbd_info->zone_info[zbd_zone_idx(f, f->file_offset)];
-	ze = &f->zbd_info->zone_info[zbd_zone_idx(f, f->file_offset +
-						  f->io_size)];
+	zb = get_zone(f, f->min_zone);
+	ze = get_zone(f, f->max_zone);
 	for (z = zb; z < ze; z++) {
 		pthread_mutex_lock(&z->mutex);
 		swd += z->wp - z->start;
@@ -795,7 +900,7 @@ static uint64_t zbd_process_swd(const struct fio_file *f, enum swd_action a)
 	}
 	pthread_mutex_unlock(&f->zbd_info->mutex);
 	for (z = zb; z < ze; z++)
-		pthread_mutex_unlock(&z->mutex);
+		zone_unlock(z);
 
 	return swd;
 }
@@ -830,14 +935,12 @@ static void zbd_init_swd(struct fio_file *f)
 void zbd_file_reset(struct thread_data *td, struct fio_file *f)
 {
 	struct fio_zone_info *zb, *ze;
-	uint32_t zone_idx_e;
 
-	if (!f->zbd_info)
+	if (!f->zbd_info || !td_write(td))
 		return;
 
-	zb = &f->zbd_info->zone_info[zbd_zone_idx(f, f->file_offset)];
-	zone_idx_e = zbd_zone_idx(f, f->file_offset + f->io_size);
-	ze = &f->zbd_info->zone_info[zone_idx_e];
+	zb = get_zone(f, f->min_zone);
+	ze = get_zone(f, f->max_zone);
 	zbd_init_swd(f);
 	/*
 	 * If data verification is enabled reset the affected zones before
@@ -845,7 +948,6 @@ void zbd_file_reset(struct thread_data *td, struct fio_file *f)
 	 * writing data, which causes data loss.
 	 */
 	zbd_reset_zones(td, f, zb, ze, td->o.verify != VERIFY_NONE &&
-			(td->o.td_ddir & TD_DDIR_WRITE) &&
 			td->runstate != TD_VERIFYING);
 	zbd_reset_write_cnt(td, f);
 }
@@ -857,8 +959,9 @@ static bool is_zone_open(const struct thread_data *td, const struct fio_file *f,
 	struct zoned_block_device_info *zbdi = f->zbd_info;
 	int i;
 
-	assert(td->o.max_open_zones <= ARRAY_SIZE(zbdi->open_zones));
-	assert(zbdi->num_open_zones <= td->o.max_open_zones);
+	assert(td->o.job_max_open_zones == 0 || td->num_open_zones <= td->o.job_max_open_zones);
+	assert(td->o.job_max_open_zones <= zbdi->max_open_zones);
+	assert(zbdi->num_open_zones <= zbdi->max_open_zones);
 
 	for (i = 0; i < zbdi->num_open_zones; i++)
 		if (zbdi->open_zones[i] == zone_idx)
@@ -873,15 +976,14 @@ static bool is_zone_open(const struct thread_data *td, const struct fio_file *f,
  * was not yet open and opening a new zone would cause the zone limit to be
  * exceeded.
  */
-static bool zbd_open_zone(struct thread_data *td, const struct io_u *io_u,
+static bool zbd_open_zone(struct thread_data *td, const struct fio_file *f,
 			  uint32_t zone_idx)
 {
 	const uint32_t min_bs = td->o.min_bs[DDIR_WRITE];
-	const struct fio_file *f = io_u->file;
-	struct fio_zone_info *z = &f->zbd_info->zone_info[zone_idx];
+	struct fio_zone_info *z = get_zone(f, zone_idx);
 	bool res = true;
 
-	if (z->cond == BLK_ZONE_COND_OFFLINE)
+	if (z->cond == ZBD_ZONE_COND_OFFLINE)
 		return false;
 
 	/*
@@ -891,18 +993,26 @@ static bool zbd_open_zone(struct thread_data *td, const struct io_u *io_u,
 	if (td->o.verify != VERIFY_NONE && zbd_zone_full(f, z, min_bs))
 		return false;
 
-	/* Zero means no limit */
-	if (!td->o.max_open_zones)
-		return true;
-
 	pthread_mutex_lock(&f->zbd_info->mutex);
-	if (is_zone_open(td, f, zone_idx))
+	if (is_zone_open(td, f, zone_idx)) {
+		/*
+		 * If the zone is already open and going to be full by writes
+		 * in-flight, handle it as a full zone instead of an open zone.
+		 */
+		if (z->wp >= zbd_zone_capacity_end(z))
+			res = false;
 		goto out;
+	}
 	res = false;
-	if (f->zbd_info->num_open_zones >= td->o.max_open_zones)
+	/* Zero means no limit */
+	if (td->o.job_max_open_zones > 0 &&
+	    td->num_open_zones >= td->o.job_max_open_zones)
+		goto out;
+	if (f->zbd_info->num_open_zones >= f->zbd_info->max_open_zones)
 		goto out;
 	dprint(FD_ZBD, "%s: opening zone %d\n", f->file_name, zone_idx);
 	f->zbd_info->open_zones[f->zbd_info->num_open_zones++] = zone_idx;
+	td->num_open_zones++;
 	z->open = 1;
 	res = true;
 
@@ -911,40 +1021,6 @@ out:
 	return res;
 }
 
-/* The caller must hold f->zbd_info->mutex */
-static void zbd_close_zone(struct thread_data *td, const struct fio_file *f,
-			   unsigned int open_zone_idx)
-{
-	uint32_t zone_idx;
-
-	assert(open_zone_idx < f->zbd_info->num_open_zones);
-	zone_idx = f->zbd_info->open_zones[open_zone_idx];
-	memmove(f->zbd_info->open_zones + open_zone_idx,
-		f->zbd_info->open_zones + open_zone_idx + 1,
-		(FIO_MAX_OPEN_ZBD_ZONES - (open_zone_idx + 1)) *
-		sizeof(f->zbd_info->open_zones[0]));
-	f->zbd_info->num_open_zones--;
-	f->zbd_info->zone_info[zone_idx].open = 0;
-}
-
-static void zone_lock(struct thread_data *td, struct fio_zone_info *z)
-{
-	/*
-	 * Lock the io_u target zone. The zone will be unlocked if io_u offset
-	 * is changed or when io_u completes and zbd_put_io() executed.
-	 * To avoid multiple jobs doing asynchronous I/Os from deadlocking each
-	 * other waiting for zone locks when building an io_u batch, first
-	 * only trylock the zone. If the zone is already locked by another job,
-	 * process the currently queued I/Os so that I/O progress is made and
-	 * zones unlocked.
-	 */
-	if (pthread_mutex_trylock(&z->mutex) != 0) {
-		if (!td_ioengine_flagged(td, FIO_SYNCIO))
-			io_u_quiesce(td);
-		pthread_mutex_lock(&z->mutex);
-	}
-}
-
 /* Anything goes as long as it is not a constant. */
 static uint32_t pick_random_zone_idx(const struct fio_file *f,
 				     const struct io_u *io_u)
@@ -963,15 +1039,16 @@ static struct fio_zone_info *zbd_convert_to_open_zone(struct thread_data *td,
 						      struct io_u *io_u)
 {
 	const uint32_t min_bs = td->o.min_bs[io_u->ddir];
-	const struct fio_file *f = io_u->file;
+	struct fio_file *f = io_u->file;
 	struct fio_zone_info *z;
 	unsigned int open_zone_idx = -1;
 	uint32_t zone_idx, new_zone_idx;
 	int i;
+	bool wait_zone_close;
 
 	assert(is_valid_offset(f, io_u->offset));
 
-	if (td->o.max_open_zones) {
+	if (td->o.max_open_zones || td->o.job_max_open_zones) {
 		/*
 		 * This statement accesses f->zbd_info->open_zones[] on purpose
 		 * without locking.
@@ -980,6 +1057,10 @@ static struct fio_zone_info *zbd_convert_to_open_zone(struct thread_data *td,
 	} else {
 		zone_idx = zbd_zone_idx(f, io_u->offset);
 	}
+	if (zone_idx < f->min_zone)
+		zone_idx = f->min_zone;
+	else if (zone_idx >= f->max_zone)
+		zone_idx = f->max_zone - 1;
 	dprint(FD_ZBD, "%s(%s): starting from zone %d (offset %lld, buflen %lld)\n",
 	       __func__, f->file_name, zone_idx, io_u->offset, io_u->buflen);
 
@@ -992,18 +1073,16 @@ static struct fio_zone_info *zbd_convert_to_open_zone(struct thread_data *td,
 	for (;;) {
 		uint32_t tmp_idx;
 
-		z = &f->zbd_info->zone_info[zone_idx];
+		z = get_zone(f, zone_idx);
 
-		zone_lock(td, z);
+		zone_lock(td, f, z);
 		pthread_mutex_lock(&f->zbd_info->mutex);
-		if (td->o.max_open_zones == 0)
+		if (td->o.max_open_zones == 0 && td->o.job_max_open_zones == 0)
 			goto examine_zone;
 		if (f->zbd_info->num_open_zones == 0) {
-			pthread_mutex_unlock(&f->zbd_info->mutex);
-			pthread_mutex_unlock(&z->mutex);
 			dprint(FD_ZBD, "%s(%s): no zones are open\n",
 			       __func__, f->file_name);
-			return NULL;
+			goto open_other_zone;
 		}
 
 		/*
@@ -1020,8 +1099,7 @@ static struct fio_zone_info *zbd_convert_to_open_zone(struct thread_data *td,
 			if (tmp_idx >= f->zbd_info->num_open_zones)
 				tmp_idx = 0;
 			tmpz = f->zbd_info->open_zones[tmp_idx];
-
-			if (is_valid_offset(f, f->zbd_info->zone_info[tmpz].start)) {
+			if (f->min_zone <= tmpz && tmpz < f->max_zone) {
 				open_zone_idx = tmp_idx;
 				goto found_candidate_zone;
 			}
@@ -1031,6 +1109,8 @@ static struct fio_zone_info *zbd_convert_to_open_zone(struct thread_data *td,
 
 		dprint(FD_ZBD, "%s(%s): no candidate zone\n",
 			__func__, f->file_name);
+		pthread_mutex_unlock(&f->zbd_info->mutex);
+		zone_unlock(z);
 		return NULL;
 
 found_candidate_zone:
@@ -1039,39 +1119,55 @@ found_candidate_zone:
 			break;
 		zone_idx = new_zone_idx;
 		pthread_mutex_unlock(&f->zbd_info->mutex);
-		pthread_mutex_unlock(&z->mutex);
+		zone_unlock(z);
 	}
 
 	/* Both z->mutex and f->zbd_info->mutex are held. */
 
 examine_zone:
-	if (z->wp + min_bs <= (z+1)->start) {
+	if (z->wp + min_bs <= zbd_zone_capacity_end(z)) {
 		pthread_mutex_unlock(&f->zbd_info->mutex);
 		goto out;
 	}
-	dprint(FD_ZBD, "%s(%s): closing zone %d\n", __func__, f->file_name,
-	       zone_idx);
-	if (td->o.max_open_zones)
-		zbd_close_zone(td, f, open_zone_idx);
+
+open_other_zone:
+	/* Check if number of open zones reaches one of limits. */
+	wait_zone_close =
+		f->zbd_info->num_open_zones == f->max_zone - f->min_zone ||
+		(td->o.max_open_zones &&
+		 f->zbd_info->num_open_zones == td->o.max_open_zones) ||
+		(td->o.job_max_open_zones &&
+		 td->num_open_zones == td->o.job_max_open_zones);
+
 	pthread_mutex_unlock(&f->zbd_info->mutex);
 
 	/* Only z->mutex is held. */
 
+	/*
+	 * When number of open zones reaches to one of limits, wait for
+	 * zone close before opening a new zone.
+	 */
+	if (wait_zone_close) {
+		dprint(FD_ZBD, "%s(%s): quiesce to allow open zones to close\n",
+		       __func__, f->file_name);
+		io_u_quiesce(td);
+	}
+
 	/* Zone 'z' is full, so try to open a new zone. */
 	for (i = f->io_size / f->zbd_info->zone_size; i > 0; i--) {
 		zone_idx++;
-		pthread_mutex_unlock(&z->mutex);
+		zone_unlock(z);
 		z++;
 		if (!is_valid_offset(f, z->start)) {
 			/* Wrap-around. */
-			zone_idx = zbd_zone_idx(f, f->file_offset);
-			z = &f->zbd_info->zone_info[zone_idx];
+			zone_idx = f->min_zone;
+			z = get_zone(f, zone_idx);
 		}
 		assert(is_valid_offset(f, z->start));
-		zone_lock(td, z);
+		zone_lock(td, f, z);
 		if (z->open)
 			continue;
-		if (zbd_open_zone(td, io_u, zone_idx))
+		if (zbd_open_zone(td, f, zone_idx))
 			goto out;
 	}
 
@@ -1081,18 +1177,20 @@ examine_zone:
 	pthread_mutex_lock(&f->zbd_info->mutex);
 	for (i = 0; i < f->zbd_info->num_open_zones; i++) {
 		zone_idx = f->zbd_info->open_zones[i];
+		if (zone_idx < f->min_zone || zone_idx >= f->max_zone)
+			continue;
 		pthread_mutex_unlock(&f->zbd_info->mutex);
-		pthread_mutex_unlock(&z->mutex);
+		zone_unlock(z);
 
-		z = &f->zbd_info->zone_info[zone_idx];
+		z = get_zone(f, zone_idx);
 
-		zone_lock(td, z);
-		if (z->wp + min_bs <= (z+1)->start)
+		zone_lock(td, f, z);
+		if (z->wp + min_bs <= zbd_zone_capacity_end(z))
 			goto out;
 		pthread_mutex_lock(&f->zbd_info->mutex);
 	}
 	pthread_mutex_unlock(&f->zbd_info->mutex);
-	pthread_mutex_unlock(&z->mutex);
+	zone_unlock(z);
 	dprint(FD_ZBD, "%s(%s): did not open another zone\n", __func__,
 	       f->file_name);
 	return NULL;
@@ -1112,15 +1210,15 @@ static struct fio_zone_info *zbd_replay_write_order(struct thread_data *td,
 	const struct fio_file *f = io_u->file;
 	const uint32_t min_bs = td->o.min_bs[DDIR_WRITE];
 
-	if (!zbd_open_zone(td, io_u, z - f->zbd_info->zone_info)) {
-		pthread_mutex_unlock(&z->mutex);
+	if (!zbd_open_zone(td, f, zbd_zone_nr(f, z))) {
+		zone_unlock(z);
 		z = zbd_convert_to_open_zone(td, io_u);
 		assert(z);
 	}
 
-	if (z->verify_block * min_bs >= f->zbd_info->zone_size)
+	if (z->verify_block * min_bs >= z->capacity)
 		log_err("%s: %d * %d >= %llu\n", f->file_name, z->verify_block,
-			min_bs, (unsigned long long) f->zbd_info->zone_size);
+			min_bs, (unsigned long long)z->capacity);
 	io_u->offset = z->start + z->verify_block++ * min_bs;
 	return z;
 }
@@ -1138,30 +1236,29 @@ zbd_find_zone(struct thread_data *td, struct io_u *io_u,
 	      struct fio_zone_info *zb, struct fio_zone_info *zl)
 {
 	const uint32_t min_bs = td->o.min_bs[io_u->ddir];
-	const struct fio_file *f = io_u->file;
+	struct fio_file *f = io_u->file;
 	struct fio_zone_info *z1, *z2;
-	const struct fio_zone_info *const zf =
-		&f->zbd_info->zone_info[zbd_zone_idx(f, f->file_offset)];
+	const struct fio_zone_info *const zf = get_zone(f, f->min_zone);
 
 	/*
 	 * Skip to the next non-empty zone in case of sequential I/O and to
 	 * the nearest non-empty zone in case of random I/O.
 	 */
 	for (z1 = zb + 1, z2 = zb - 1; z1 < zl || z2 >= zf; z1++, z2--) {
-		if (z1 < zl && z1->cond != BLK_ZONE_COND_OFFLINE) {
-			pthread_mutex_lock(&z1->mutex);
+		if (z1 < zl && z1->cond != ZBD_ZONE_COND_OFFLINE) {
+			zone_lock(td, f, z1);
 			if (z1->start + min_bs <= z1->wp)
 				return z1;
-			pthread_mutex_unlock(&z1->mutex);
+			zone_unlock(z1);
 		} else if (!td_random(td)) {
 			break;
 		}
 		if (td_random(td) && z2 >= zf &&
-		    z2->cond != BLK_ZONE_COND_OFFLINE) {
-			pthread_mutex_lock(&z2->mutex);
+		    z2->cond != ZBD_ZONE_COND_OFFLINE) {
+			zone_lock(td, f, z2);
 			if (z2->start + min_bs <= z2->wp)
 				return z2;
-			pthread_mutex_unlock(&z2->mutex);
+			zone_unlock(z2);
 		}
 	}
 	dprint(FD_ZBD, "%s: adjusting random read offset failed\n",
@@ -1169,6 +1266,28 @@ zbd_find_zone(struct thread_data *td, struct io_u *io_u,
 	return NULL;
 }
 
+/**
+ * zbd_end_zone_io - update zone status at command completion
+ * @io_u: I/O unit
+ * @z: zone info pointer
+ *
+ * If the write command made the zone full, close it.
+ *
+ * The caller must hold z->mutex.
+ */
+static void zbd_end_zone_io(struct thread_data *td, const struct io_u *io_u,
+			    struct fio_zone_info *z)
+{
+	const struct fio_file *f = io_u->file;
+
+	if (io_u->ddir == DDIR_WRITE &&
+	    io_u->offset + io_u->buflen >= zbd_zone_capacity_end(z)) {
+		pthread_mutex_lock(&f->zbd_info->mutex);
+		zbd_close_zone(td, f, zbd_zone_nr(f, z));
+		pthread_mutex_unlock(&f->zbd_info->mutex);
+	}
+}
+
 /**
  * zbd_queue_io - update the write pointer of a sequential zone
  * @io_u: I/O unit
@@ -1178,7 +1297,8 @@ zbd_find_zone(struct thread_data *td, struct io_u *io_u,
  * For write and trim operations, update the write pointer of the I/O unit
  * target zone.
  */
-static void zbd_queue_io(struct io_u *io_u, int q, bool success)
+static void zbd_queue_io(struct thread_data *td, struct io_u *io_u, int q,
+			 bool success)
 {
 	const struct fio_file *f = io_u->file;
 	struct zoned_block_device_info *zbd_info = f->zbd_info;
@@ -1191,9 +1311,9 @@ static void zbd_queue_io(struct io_u *io_u, int q, bool success)
 
 	zone_idx = zbd_zone_idx(f, io_u->offset);
 	assert(zone_idx < zbd_info->nr_zones);
-	z = &zbd_info->zone_info[zone_idx];
+	z = get_zone(f, zone_idx);
 
-	if (z->type != BLK_ZONE_TYPE_SEQWRITE_REQ)
+	if (!zbd_zone_swr(z))
 		return;
 
 	if (!success)
@@ -1206,7 +1326,7 @@ static void zbd_queue_io(struct io_u *io_u, int q, bool success)
 	switch (io_u->ddir) {
 	case DDIR_WRITE:
 		zone_end = min((uint64_t)(io_u->offset + io_u->buflen),
-			       (z + 1)->start);
+			       zbd_zone_capacity_end(z));
 		pthread_mutex_lock(&zbd_info->mutex);
 		/*
 		 * z->wp > zone_end means that one or more I/O errors
@@ -1224,10 +1344,13 @@ static void zbd_queue_io(struct io_u *io_u, int q, bool success)
 		break;
 	}
 
+	if (q == FIO_Q_COMPLETED && !io_u->error)
+		zbd_end_zone_io(td, io_u, z);
+
 unlock:
 	if (!success || q != FIO_Q_QUEUED) {
 		/* BUSY or COMPLETED: unlock the zone */
-		pthread_mutex_unlock(&z->mutex);
+		zone_unlock(z);
 		io_u->zbd_put_io = NULL;
 	}
 }
@@ -1236,7 +1359,7 @@ unlock:
  * zbd_put_io - Unlock an I/O unit target zone lock
  * @io_u: I/O unit
  */
-static void zbd_put_io(const struct io_u *io_u)
+static void zbd_put_io(struct thread_data *td, const struct io_u *io_u)
 {
 	const struct fio_file *f = io_u->file;
 	struct zoned_block_device_info *zbd_info = f->zbd_info;
@@ -1248,19 +1371,28 @@ static void zbd_put_io(const struct io_u *io_u)
 
 	zone_idx = zbd_zone_idx(f, io_u->offset);
 	assert(zone_idx < zbd_info->nr_zones);
-	z = &zbd_info->zone_info[zone_idx];
+	z = get_zone(f, zone_idx);
 
-	if (z->type != BLK_ZONE_TYPE_SEQWRITE_REQ)
+	if (!zbd_zone_swr(z))
 		return;
 
 	dprint(FD_ZBD,
 	       "%s: terminate I/O (%lld, %llu) for zone %u\n",
 	       f->file_name, io_u->offset, io_u->buflen, zone_idx);
 
-	assert(pthread_mutex_unlock(&z->mutex) == 0);
+	zbd_end_zone_io(td, io_u, z);
+
+	zone_unlock(z);
 	zbd_check_swd(f);
 }
 
+/*
+ * Windows and MacOS do not define this.
+ */
+#ifndef EREMOTEIO
+#define EREMOTEIO	121	/* POSIX value */
+#endif
+
 bool zbd_unaligned_write(int error_code)
 {
 	switch (error_code) {
@@ -1293,6 +1425,27 @@ void setup_zbd_zone_mode(struct thread_data *td, struct io_u *io_u)
 	assert(td->o.zone_mode == ZONE_MODE_ZBD);
 	assert(td->o.zone_size);
 
+	zone_idx = zbd_zone_idx(f, f->last_pos[ddir]);
+	z = get_zone(f, zone_idx);
+
+	/*
+	 * When the zone capacity is smaller than the zone size and the I/O is
+	 * sequential write, skip to zone end if the latest position is at the
+	 * zone capacity limit.
+	 */
+	if (z->capacity < f->zbd_info->zone_size && !td_random(td) &&
+	    ddir == DDIR_WRITE &&
+	    f->last_pos[ddir] >= zbd_zone_capacity_end(z)) {
+		dprint(FD_ZBD,
+		       "%s: Jump from zone capacity limit to zone end:"
+		       " (%llu -> %llu) for zone %u (%llu)\n",
+		       f->file_name, (unsigned long long) f->last_pos[ddir],
+		       (unsigned long long) zbd_zone_end(z), zone_idx,
+		       (unsigned long long) z->capacity);
+		td->io_skip_bytes += zbd_zone_end(z) - f->last_pos[ddir];
+		f->last_pos[ddir] = zbd_zone_end(z);
+	}
+
 	/*
 	 * zone_skip is valid only for sequential workloads.
 	 */
@@ -1306,11 +1459,8 @@ void setup_zbd_zone_mode(struct thread_data *td, struct io_u *io_u)
 	 * - For reads with td->o.read_beyond_wp == false, the last position
 	 *   reached the zone write pointer.
 	 */
-	zone_idx = zbd_zone_idx(f, f->last_pos[ddir]);
-	z = &f->zbd_info->zone_info[zone_idx];
-
 	if (td->zone_bytes >= td->o.zone_size ||
-	    f->last_pos[ddir] >= (z+1)->start ||
+	    f->last_pos[ddir] >= zbd_zone_end(z) ||
 	    (ddir == DDIR_READ &&
 	     (!td->o.read_beyond_wp) && f->last_pos[ddir] >= z->wp)) {
 		/*
@@ -1330,6 +1480,34 @@ void setup_zbd_zone_mode(struct thread_data *td, struct io_u *io_u)
 	}
 }
 
+/**
+ * zbd_adjust_ddir - Adjust an I/O direction for zonemode=zbd.
+ *
+ * @td: FIO thread data.
+ * @io_u: FIO I/O unit.
+ * @ddir: I/O direction before adjustment.
+ *
+ * Return adjusted I/O direction.
+ */
+enum fio_ddir zbd_adjust_ddir(struct thread_data *td, struct io_u *io_u,
+			      enum fio_ddir ddir)
+{
+	/*
+	 * In case read direction is chosen for the first random I/O, fio with
+	 * zonemode=zbd stops because no data can be read from zoned block
+	 * devices with all empty zones. Overwrite the first I/O direction as
+	 * write to make sure data to read exists.
+	 */
+	if (ddir != DDIR_READ || !td_rw(td))
+		return ddir;
+
+	if (io_u->file->zbd_info->sectors_with_data ||
+	    td->o.read_beyond_wp)
+		return DDIR_READ;
+
+	return DDIR_WRITE;
+}
+
 /**
  * zbd_adjust_block - adjust the offset and length as necessary for ZBD drives
  * @td: FIO thread data.
@@ -1341,7 +1519,7 @@ void setup_zbd_zone_mode(struct thread_data *td, struct io_u *io_u)
  */
 enum io_u_action zbd_adjust_block(struct thread_data *td, struct io_u *io_u)
 {
-	const struct fio_file *f = io_u->file;
+	struct fio_file *f = io_u->file;
 	uint32_t zone_idx_b;
 	struct fio_zone_info *zb, *zl, *orig_zb;
 	uint32_t orig_len = io_u->buflen;
@@ -1352,32 +1530,34 @@ enum io_u_action zbd_adjust_block(struct thread_data *td, struct io_u *io_u)
 	if (!f->zbd_info)
 		return io_u_accept;
 
+	assert(min_bs);
 	assert(is_valid_offset(f, io_u->offset));
 	assert(io_u->buflen);
 	zone_idx_b = zbd_zone_idx(f, io_u->offset);
-	zb = &f->zbd_info->zone_info[zone_idx_b];
+	zb = get_zone(f, zone_idx_b);
 	orig_zb = zb;
 
 	/* Accept the I/O offset for conventional zones. */
-	if (zb->type == BLK_ZONE_TYPE_CONVENTIONAL)
+	if (!zbd_zone_swr(zb))
 		return io_u_accept;
 
 	/*
 	 * Accept the I/O offset for reads if reading beyond the write pointer
 	 * is enabled.
 	 */
-	if (zb->cond != BLK_ZONE_COND_OFFLINE &&
+	if (zb->cond != ZBD_ZONE_COND_OFFLINE &&
 	    io_u->ddir == DDIR_READ && td->o.read_beyond_wp)
 		return io_u_accept;
 
 	zbd_check_swd(f);
 
-	zone_lock(td, zb);
+	zone_lock(td, f, zb);
 
 	switch (io_u->ddir) {
 	case DDIR_READ:
-		if (td->runstate == TD_VERIFYING) {
+		if (td->runstate == TD_VERIFYING && td_write(td)) {
 			zb = zbd_replay_write_order(td, io_u, zb);
+			zone_unlock(zb);
 			goto accept;
 		}
 		/*
@@ -1385,13 +1565,12 @@ enum io_u_action zbd_adjust_block(struct thread_data *td, struct io_u *io_u)
 		 * I/O of at least min_bs B. If there isn't, find a new zone for
 		 * the I/O.
 		 */
-		range = zb->cond != BLK_ZONE_COND_OFFLINE ?
+		range = zb->cond != ZBD_ZONE_COND_OFFLINE ?
 			zb->wp - zb->start : 0;
 		if (range < min_bs ||
 		    ((!td_random(td)) && (io_u->offset + min_bs > zb->wp))) {
-			pthread_mutex_unlock(&zb->mutex);
-			zl = &f->zbd_info->zone_info[zbd_zone_idx(f,
-						f->file_offset + f->io_size)];
+			zone_unlock(zb);
+			zl = get_zone(f, f->max_zone);
 			zb = zbd_find_zone(td, io_u, zb, zl);
 			if (!zb) {
 				dprint(FD_ZBD,
@@ -1437,12 +1616,12 @@ enum io_u_action zbd_adjust_block(struct thread_data *td, struct io_u *io_u)
 	case DDIR_WRITE:
 		if (io_u->buflen > f->zbd_info->zone_size)
 			goto eof;
-		if (!zbd_open_zone(td, io_u, zone_idx_b)) {
-			pthread_mutex_unlock(&zb->mutex);
+		if (!zbd_open_zone(td, f, zone_idx_b)) {
+			zone_unlock(zb);
 			zb = zbd_convert_to_open_zone(td, io_u);
 			if (!zb)
 				goto eof;
-			zone_idx_b = zb - f->zbd_info->zone_info;
+			zone_idx_b = zbd_zone_nr(f, zb);
 		}
 		/* Check whether the zone reset threshold has been exceeded */
 		if (td->o.zrf.u.f) {
@@ -1466,6 +1645,13 @@ enum io_u_action zbd_adjust_block(struct thread_data *td, struct io_u *io_u)
 			zb->reset_zone = 0;
 			if (zbd_reset_zone(td, f, zb) < 0)
 				goto eof;
+
+			if (zb->capacity < min_bs) {
+				log_err("zone capacity %llu smaller than minimum block size %d\n",
+					(unsigned long long)zb->capacity,
+					min_bs);
+				goto eof;
+			}
 		}
 		/* Make writes occur at the write pointer */
 		assert(!zbd_zone_full(f, zb, min_bs));
@@ -1481,7 +1667,7 @@ enum io_u_action zbd_adjust_block(struct thread_data *td, struct io_u *io_u)
 		 * small.
 		 */
 		new_len = min((unsigned long long)io_u->buflen,
-			      (zb + 1)->start - io_u->offset);
+			      zbd_zone_capacity_end(zb) - io_u->offset);
 		new_len = new_len / min_bs * min_bs;
 		if (new_len == io_u->buflen)
 			goto accept;
@@ -1492,7 +1678,7 @@ enum io_u_action zbd_adjust_block(struct thread_data *td, struct io_u *io_u)
 			goto accept;
 		}
 		log_err("Zone remainder %lld smaller than minimum block size %d\n",
-			((zb + 1)->start - io_u->offset),
+			(zbd_zone_capacity_end(zb) - io_u->offset),
 			min_bs);
 		goto eof;
 	case DDIR_TRIM:
@@ -1510,7 +1696,7 @@ enum io_u_action zbd_adjust_block(struct thread_data *td, struct io_u *io_u)
 
 accept:
 	assert(zb);
-	assert(zb->cond != BLK_ZONE_COND_OFFLINE);
+	assert(zb->cond != ZBD_ZONE_COND_OFFLINE);
 	assert(!io_u->zbd_queue_io);
 	assert(!io_u->zbd_put_io);
 	io_u->zbd_queue_io = zbd_queue_io;
@@ -1519,7 +1705,7 @@ accept:
 
 eof:
 	if (zb)
-		pthread_mutex_unlock(&zb->mutex);
+		zone_unlock(zb);
 	return io_u_eof;
 }