[fio.git] / oslib / linux-blkzoned.c

/*
 * Copyright (C) 2020 Western Digital Corporation or its affiliates.
 *
 * This file is released under the GPL.
 */
#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 "file.h"
#include "fio.h"
#include "lib/pow2.h"
#include "log.h"
#include "oslib/asprintf.h"
#include "smalloc.h"
#include "verify.h"
#include "zbd_types.h"

#include <linux/blkzoned.h>

/*
 * If the uapi headers installed on the system lacks zone capacity support,
 * use our local versions. If the installed headers are recent enough to
 * support zone capacity, do not redefine any structs.
 */
#ifndef CONFIG_HAVE_REP_CAPACITY
#define BLK_ZONE_REP_CAPACITY	(1 << 0)

struct blk_zone_v2 {
	__u64	start;          /* Zone start sector */
	__u64	len;            /* Zone length in number of sectors */
	__u64	wp;             /* Zone write pointer position */
	__u8	type;           /* Zone type */
	__u8	cond;           /* Zone condition */
	__u8	non_seq;        /* Non-sequential write resources active */
	__u8	reset;          /* Reset write pointer recommended */
	__u8	resv[4];
	__u64	capacity;       /* Zone capacity in number of sectors */
	__u8	reserved[24];
};
#define blk_zone blk_zone_v2

struct blk_zone_report_v2 {
	__u64	sector;
	__u32	nr_zones;
	__u32	flags;
struct blk_zone zones[0];
};
#define blk_zone_report blk_zone_report_v2
#endif /* CONFIG_HAVE_REP_CAPACITY */

/*
 * 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);
}

/*
 * Get the value of a sysfs attribute for a block device.
 *
 * Returns NULL on failure.
 * Returns a pointer to a string on success.
 * The caller is responsible for freeing the memory.
 */
static char *blkzoned_get_sysfs_attr(const char *file_name, const char *attr)
{
	char *attr_path = 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;
	char *attr_str = 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(&attr_path,
		     "/sys/dev/block/%s/%s", sys_path, attr) < 0)
		goto out;

	attr_str = read_file(attr_path);
out:
	free(attr_path);
	free(part_str);
	free(part_attr_path);
	free(sys_devno_path);

	return attr_str;
}

int blkzoned_get_zoned_model(struct thread_data *td, struct fio_file *f,
			     enum zbd_zoned_model *model)
{
	char *model_str = NULL;

	if (f->filetype != FIO_TYPE_BLOCK)
		return -EINVAL;

	*model = ZBD_NONE;

	model_str = blkzoned_get_sysfs_attr(f->file_name, "queue/zoned");
	if (!model_str)
		return 0;

	dprint(FD_ZBD, "%s: zbd model string: %s\n", f->file_name, model_str);
	if (strcmp(model_str, "host-aware") == 0)
		*model = ZBD_HOST_AWARE;
	else if (strcmp(model_str, "host-managed") == 0)
		*model = ZBD_HOST_MANAGED;

	free(model_str);

	return 0;
}

int blkzoned_get_max_open_zones(struct thread_data *td, struct fio_file *f,
				unsigned int *max_open_zones)
{
	char *max_open_str;

	if (f->filetype != FIO_TYPE_BLOCK)
		return -EIO;

	max_open_str = blkzoned_get_sysfs_attr(f->file_name, "queue/max_open_zones");
	if (!max_open_str)
		return 0;

	dprint(FD_ZBD, "%s: max open zones supported by device: %s\n",
	       f->file_name, max_open_str);
	*max_open_zones = atoll(max_open_str);

	free(max_open_str);

	return 0;
}

static uint64_t zone_capacity(struct blk_zone_report *hdr,
			      struct blk_zone *blkz)
{
	if (hdr->flags & BLK_ZONE_REP_CAPACITY)
		return blkz->capacity << 9;
	return blkz->len << 9;
}

int blkzoned_report_zones(struct thread_data *td, struct fio_file *f,
			  uint64_t offset, struct zbd_zone *zones,
			  unsigned int nr_zones)
{
	struct blk_zone_report *hdr = NULL;
	struct blk_zone *blkz;
	struct zbd_zone *z;
	unsigned int i;
	int fd = -1, ret;

	fd = open(f->file_name, O_RDONLY | O_LARGEFILE);
	if (fd < 0)
		return -errno;

	hdr = calloc(1, sizeof(struct blk_zone_report) +
			nr_zones * sizeof(struct blk_zone));
	if (!hdr) {
		ret = -ENOMEM;
		goto out;
	}

	hdr->nr_zones = nr_zones;
	hdr->sector = offset >> 9;
	ret = ioctl(fd, BLKREPORTZONE, hdr);
	if (ret) {
		ret = -errno;
		goto out;
	}

	nr_zones = hdr->nr_zones;
	blkz = (void *) hdr + sizeof(*hdr);
	z = &zones[0];
	for (i = 0; i < nr_zones; i++, z++, blkz++) {
		z->start = blkz->start << 9;
		z->wp = blkz->wp << 9;
		z->len = blkz->len << 9;
		z->capacity = zone_capacity(hdr, blkz);

		switch (blkz->type) {
		case BLK_ZONE_TYPE_CONVENTIONAL:
			z->type = ZBD_ZONE_TYPE_CNV;
			break;
		case BLK_ZONE_TYPE_SEQWRITE_REQ:
			z->type = ZBD_ZONE_TYPE_SWR;
			break;
		case BLK_ZONE_TYPE_SEQWRITE_PREF:
			z->type = ZBD_ZONE_TYPE_SWP;
			break;
		default:
			td_verror(td, errno, "invalid zone type");
			log_err("%s: invalid type for zone at sector %llu.\n",
				f->file_name, (unsigned long long)offset >> 9);
			ret = -EIO;
			goto out;
		}

		switch (blkz->cond) {
		case BLK_ZONE_COND_NOT_WP:
			z->cond = ZBD_ZONE_COND_NOT_WP;
			break;
		case BLK_ZONE_COND_EMPTY:
			z->cond = ZBD_ZONE_COND_EMPTY;
			break;
		case BLK_ZONE_COND_IMP_OPEN:
			z->cond = ZBD_ZONE_COND_IMP_OPEN;
			break;
		case BLK_ZONE_COND_EXP_OPEN:
			z->cond = ZBD_ZONE_COND_EXP_OPEN;
			break;
		case BLK_ZONE_COND_CLOSED:
			z->cond = ZBD_ZONE_COND_CLOSED;
			break;
		case BLK_ZONE_COND_FULL:
			z->cond = ZBD_ZONE_COND_FULL;
			break;
		case BLK_ZONE_COND_READONLY:
		case BLK_ZONE_COND_OFFLINE:
		default:
			/* Treat all these conditions as offline (don't use!) */
			z->cond = ZBD_ZONE_COND_OFFLINE;
			z->wp = z->start;
		}
	}

	ret = nr_zones;
out:
	free(hdr);
	close(fd);

	return ret;
}

int blkzoned_reset_wp(struct thread_data *td, struct fio_file *f,
		      uint64_t offset, uint64_t length)
{
	struct blk_zone_range zr = {
		.sector         = offset >> 9,
		.nr_sectors     = length >> 9,
	};
	int fd, ret = 0;

	/* If the file is not yet opened, open it for this function. */
	fd = f->fd;
	if (fd < 0) {
		fd = open(f->file_name, O_RDWR | O_LARGEFILE);
		if (fd < 0)
			return -errno;
	}

	if (ioctl(fd, BLKRESETZONE, &zr) < 0)
		ret = -errno;

	if (f->fd < 0)
		close(fd);

	return ret;
}
Commit	Line	Data
b7694961 DLM	1	/*
	2	* Copyright (C) 2020 Western Digital Corporation or its affiliates.
	3	*
	4	* This file is released under the GPL.
	5	*/
	6	#include <errno.h>
	7	#include <string.h>
	8	#include <stdlib.h>
	9	#include <dirent.h>
	10	#include <fcntl.h>
	11	#include <sys/ioctl.h>
	12	#include <sys/stat.h>
	13	#include <unistd.h>
	14
	15	#include "file.h"
	16	#include "fio.h"
	17	#include "lib/pow2.h"
	18	#include "log.h"
	19	#include "oslib/asprintf.h"
	20	#include "smalloc.h"
	21	#include "verify.h"
	22	#include "zbd_types.h"
	23
	24	#include <linux/blkzoned.h>
	25
6ee607ba NC	26	/*
	27	* If the uapi headers installed on the system lacks zone capacity support,
	28	* use our local versions. If the installed headers are recent enough to
	29	* support zone capacity, do not redefine any structs.
	30	*/
	31	#ifndef CONFIG_HAVE_REP_CAPACITY
	32	#define BLK_ZONE_REP_CAPACITY (1 << 0)
	33
	34	struct blk_zone_v2 {
	35	__u64 start; /* Zone start sector */
	36	__u64 len; /* Zone length in number of sectors */
	37	__u64 wp; /* Zone write pointer position */
	38	__u8 type; /* Zone type */
	39	__u8 cond; /* Zone condition */
	40	__u8 non_seq; /* Non-sequential write resources active */
	41	__u8 reset; /* Reset write pointer recommended */
	42	__u8 resv[4];
	43	__u64 capacity; /* Zone capacity in number of sectors */
	44	__u8 reserved[24];
	45	};
	46	#define blk_zone blk_zone_v2
	47
	48	struct blk_zone_report_v2 {
	49	__u64 sector;
	50	__u32 nr_zones;
	51	__u32 flags;
	52	struct blk_zone zones[0];
	53	};
	54	#define blk_zone_report blk_zone_report_v2
	55	#endif /* CONFIG_HAVE_REP_CAPACITY */
	56
b7694961 DLM	57	/*
	58	* Read up to 255 characters from the first line of a file. Strip the trailing
	59	* newline.
	60	*/
	61	static char read_file(const char path)
	62	{
	63	char line[256], *p = line;
	64	FILE *f;
	65
	66	f = fopen(path, "rb");
	67	if (!f)
	68	return NULL;
	69	if (!fgets(line, sizeof(line), f))
	70	line[0] = '\0';
	71	strsep(&p, "\n");
	72	fclose(f);
	73
	74	return strdup(line);
	75	}
	76
eaa45783 NC	77	/*
	78	* Get the value of a sysfs attribute for a block device.
	79	*
	80	* Returns NULL on failure.
	81	* Returns a pointer to a string on success.
	82	* The caller is responsible for freeing the memory.
	83	*/
	84	static char blkzoned_get_sysfs_attr(const char file_name, const char *attr)
b7694961	85	{
eaa45783	86	char *attr_path = NULL;
b7694961 DLM	87	struct stat statbuf;
	88	char *sys_devno_path = NULL;
	89	char *part_attr_path = NULL;
	90	char *part_str = NULL;
	91	char sys_path[PATH_MAX];
	92	ssize_t sz;
	93	char *delim = NULL;
eaa45783	94	char *attr_str = NULL;
b7694961 DLM	95
	96	if (stat(file_name, &statbuf) < 0)
	97	goto out;
	98
	99	if (asprintf(&sys_devno_path, "/sys/dev/block/%d:%d",
	100	major(statbuf.st_rdev), minor(statbuf.st_rdev)) < 0)
	101	goto out;
	102
	103	sz = readlink(sys_devno_path, sys_path, sizeof(sys_path) - 1);
	104	if (sz < 0)
	105	goto out;
	106	sys_path[sz] = '\0';
	107
	108	/*
	109	* If the device is a partition device, cut the device name in the
	110	* canonical sysfs path to obtain the sysfs path of the holder device.
	111	* e.g.: /sys/devices/.../sda/sda1 -> /sys/devices/.../sda
	112	*/
	113	if (asprintf(&part_attr_path, "/sys/dev/block/%s/partition",
	114	sys_path) < 0)
	115	goto out;
	116	part_str = read_file(part_attr_path);
	117	if (part_str && *part_str == '1') {
	118	delim = strrchr(sys_path, '/');
	119	if (!delim)
	120	goto out;
	121	*delim = '\0';
	122	}
	123
eaa45783 NC	124	if (asprintf(&attr_path,
eaa45783 NC	125	"/sys/dev/block/%s/%s", sys_path, attr) < 0)
b7694961 DLM	126	goto out;
b7694961 DLM	127
eaa45783 NC	128	attr_str = read_file(attr_path);
	129	out:
	130	free(attr_path);
	131	free(part_str);
	132	free(part_attr_path);
	133	free(sys_devno_path);
	134
	135	return attr_str;
	136	}
	137
	138	int blkzoned_get_zoned_model(struct thread_data td, struct fio_file f,
	139	enum zbd_zoned_model *model)
	140	{
	141	char *model_str = NULL;
	142
2c7dd23e NC	143	if (f->filetype != FIO_TYPE_BLOCK)
2c7dd23e NC	144	return -EINVAL;
eaa45783 NC	145
	146	*model = ZBD_NONE;
	147
	148	model_str = blkzoned_get_sysfs_attr(f->file_name, "queue/zoned");
b7694961	149	if (!model_str)
eaa45783 NC	150	return 0;
	151
	152	dprint(FD_ZBD, "%s: zbd model string: %s\n", f->file_name, model_str);
b7694961 DLM	153	if (strcmp(model_str, "host-aware") == 0)
	154	*model = ZBD_HOST_AWARE;
	155	else if (strcmp(model_str, "host-managed") == 0)
	156	*model = ZBD_HOST_MANAGED;
eaa45783	157
b7694961	158	free(model_str);
eaa45783	159
b7694961 DLM	160	return 0;
	161	}
	162
d2f442bc NC	163	int blkzoned_get_max_open_zones(struct thread_data td, struct fio_file f,
	164	unsigned int *max_open_zones)
	165	{
	166	char *max_open_str;
	167
	168	if (f->filetype != FIO_TYPE_BLOCK)
	169	return -EIO;
	170
	171	max_open_str = blkzoned_get_sysfs_attr(f->file_name, "queue/max_open_zones");
	172	if (!max_open_str)
	173	return 0;
	174
	175	dprint(FD_ZBD, "%s: max open zones supported by device: %s\n",
	176	f->file_name, max_open_str);
	177	*max_open_zones = atoll(max_open_str);
	178
	179	free(max_open_str);
	180
	181	return 0;
	182	}
	183
236d23a8 SK	184	static uint64_t zone_capacity(struct blk_zone_report *hdr,
	185	struct blk_zone *blkz)
	186	{
236d23a8 SK	187	if (hdr->flags & BLK_ZONE_REP_CAPACITY)
236d23a8 SK	188	return blkz->capacity << 9;
236d23a8 SK	189	return blkz->len << 9;
	190	}
	191
b7694961 DLM	192	int blkzoned_report_zones(struct thread_data td, struct fio_file f,
	193	uint64_t offset, struct zbd_zone *zones,
	194	unsigned int nr_zones)
	195	{
	196	struct blk_zone_report *hdr = NULL;
	197	struct blk_zone *blkz;
	198	struct zbd_zone *z;
	199	unsigned int i;
	200	int fd = -1, ret;
	201
	202	fd = open(f->file_name, O_RDONLY \| O_LARGEFILE);
	203	if (fd < 0)
	204	return -errno;
	205
	206	hdr = calloc(1, sizeof(struct blk_zone_report) +
	207	nr_zones * sizeof(struct blk_zone));
	208	if (!hdr) {
	209	ret = -ENOMEM;
	210	goto out;
	211	}
	212
	213	hdr->nr_zones = nr_zones;
	214	hdr->sector = offset >> 9;
	215	ret = ioctl(fd, BLKREPORTZONE, hdr);
	216	if (ret) {
	217	ret = -errno;
	218	goto out;
	219	}
	220
	221	nr_zones = hdr->nr_zones;
f8779edf	222	blkz = (void ) hdr + sizeof(hdr);
b7694961 DLM	223	z = &zones[0];
	224	for (i = 0; i < nr_zones; i++, z++, blkz++) {
	225	z->start = blkz->start << 9;
	226	z->wp = blkz->wp << 9;
	227	z->len = blkz->len << 9;
236d23a8	228	z->capacity = zone_capacity(hdr, blkz);
b7694961 DLM	229
	230	switch (blkz->type) {
	231	case BLK_ZONE_TYPE_CONVENTIONAL:
	232	z->type = ZBD_ZONE_TYPE_CNV;
	233	break;
	234	case BLK_ZONE_TYPE_SEQWRITE_REQ:
	235	z->type = ZBD_ZONE_TYPE_SWR;
	236	break;
	237	case BLK_ZONE_TYPE_SEQWRITE_PREF:
	238	z->type = ZBD_ZONE_TYPE_SWP;
	239	break;
	240	default:
	241	td_verror(td, errno, "invalid zone type");
	242	log_err("%s: invalid type for zone at sector %llu.\n",
	243	f->file_name, (unsigned long long)offset >> 9);
	244	ret = -EIO;
	245	goto out;
	246	}
	247
	248	switch (blkz->cond) {
	249	case BLK_ZONE_COND_NOT_WP:
	250	z->cond = ZBD_ZONE_COND_NOT_WP;
	251	break;
	252	case BLK_ZONE_COND_EMPTY:
	253	z->cond = ZBD_ZONE_COND_EMPTY;
	254	break;
	255	case BLK_ZONE_COND_IMP_OPEN:
	256	z->cond = ZBD_ZONE_COND_IMP_OPEN;
	257	break;
	258	case BLK_ZONE_COND_EXP_OPEN:
	259	z->cond = ZBD_ZONE_COND_EXP_OPEN;
	260	break;
	261	case BLK_ZONE_COND_CLOSED:
	262	z->cond = ZBD_ZONE_COND_CLOSED;
	263	break;
	264	case BLK_ZONE_COND_FULL:
	265	z->cond = ZBD_ZONE_COND_FULL;
	266	break;
	267	case BLK_ZONE_COND_READONLY:
	268	case BLK_ZONE_COND_OFFLINE:
	269	default:
	270	/* Treat all these conditions as offline (don't use!) */
	271	z->cond = ZBD_ZONE_COND_OFFLINE;
59b67452	272	z->wp = z->start;
b7694961 DLM	273	}
	274	}
	275
	276	ret = nr_zones;
	277	out:
	278	free(hdr);
	279	close(fd);
	280
	281	return ret;
	282	}
	283
	284	int blkzoned_reset_wp(struct thread_data td, struct fio_file f,
	285	uint64_t offset, uint64_t length)
	286	{
	287	struct blk_zone_range zr = {
	288	.sector = offset >> 9,
	289	.nr_sectors = length >> 9,
	290	};
b4824992 SK	291	int fd, ret = 0;
	292
	293	/* If the file is not yet opened, open it for this function. */
	294	fd = f->fd;
	295	if (fd < 0) {
	296	fd = open(f->file_name, O_RDWR \| O_LARGEFILE);
	297	if (fd < 0)
	298	return -errno;
	299	}
b7694961	300
b4824992 SK	301	if (ioctl(fd, BLKRESETZONE, &zr) < 0)
b4824992 SK	302	ret = -errno;
b7694961	303
b4824992 SK	304	if (f->fd < 0)
	305	close(fd);
	306
	307	return ret;
b7694961	308	}