btrfs-progs/cmds/filesystem-usage.c

/*
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public
 * License v2 as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public
 * License along with this program; if not, write to the
 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
 * Boston, MA 021110-1307, USA.
 */

#include "kerncompat.h"
#include <sys/ioctl.h>
#include <sys/statfs.h>
#include <linux/limits.h>
#include <stdbool.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <errno.h>
#include <fcntl.h>
#include <dirent.h>
#include <limits.h>
#include <uuid/uuid.h>
#include "kernel-lib/sizes.h"
#include "kernel-shared/ctree.h"
#include "kernel-shared/disk-io.h"
#include "kernel-shared/volumes.h"
#include "kernel-shared/accessors.h"
#include "kernel-shared/uapi/btrfs_tree.h"
#include "common/defs.h"
#include "common/utils.h"
#include "common/string-table.h"
#include "common/open-utils.h"
#include "common/units.h"
#include "common/help.h"
#include "common/device-utils.h"
#include "common/sysfs-utils.h"
#include "common/messages.h"
#include "common/path-utils.h"
#include "cmds/filesystem-usage.h"
#include "cmds/commands.h"

/*
 * Add the chunk info to the chunk_info list
 */
static int add_info_to_list(struct chunk_info **chunkinfo_ret,
			int *info_count,
			struct btrfs_chunk *chunk)
{

	u64 type = btrfs_stack_chunk_type(chunk);
	u64 size = btrfs_stack_chunk_length(chunk);
	int num_stripes = btrfs_stack_chunk_num_stripes(chunk);
	int j;

	for (j = 0 ; j < num_stripes ; j++) {
		int i;
		struct chunk_info *p = NULL;
		struct btrfs_stripe *stripe;
		u64    devid;

		stripe = btrfs_stripe_nr(chunk, j);
		devid = btrfs_stack_stripe_devid(stripe);

		for (i = 0 ; i < *info_count ; i++)
			if ((*chunkinfo_ret)[i].type == type &&
			    (*chunkinfo_ret)[i].devid == devid &&
			    (*chunkinfo_ret)[i].num_stripes == num_stripes ) {
				p = (*chunkinfo_ret) + i;
				break;
			}

		if (!p) {
			int tmp = sizeof(struct btrfs_chunk) * (*info_count + 1);
			struct chunk_info *res = realloc(*chunkinfo_ret, tmp);

			if (!res) {
				free(*chunkinfo_ret);
				error_msg(ERROR_MSG_MEMORY, NULL);
				return -ENOMEM;
			}

			*chunkinfo_ret = res;
			p = res + *info_count;
			(*info_count)++;

			p->devid = devid;
			p->type = type;
			p->size = 0;
			p->num_stripes = num_stripes;
		}

		p->size += size;

	}

	return 0;
}

/*
 *  Helper to sort the chunk type
 */
static int cmp_chunk_block_group(u64 f1, u64 f2)
{

	u64 mask;

	if ((f1 & BTRFS_BLOCK_GROUP_TYPE_MASK) ==
		(f2 & BTRFS_BLOCK_GROUP_TYPE_MASK))
			mask = BTRFS_BLOCK_GROUP_PROFILE_MASK;
	else if (f2 & BTRFS_BLOCK_GROUP_SYSTEM)
			return -1;
	else if (f1 & BTRFS_BLOCK_GROUP_SYSTEM)
			return +1;
	else
			mask = BTRFS_BLOCK_GROUP_TYPE_MASK;

	if ((f1 & mask) > (f2 & mask))
		return +1;
	else if ((f1 & mask) < (f2 & mask))
		return -1;
	else
		return 0;
}

/*
 * Helper to sort the chunk
 */
static int cmp_chunk_info(const void *a, const void *b)
{
	return cmp_chunk_block_group(
		((struct chunk_info *)a)->type,
		((struct chunk_info *)b)->type);
}

static int load_chunk_info(int fd, struct chunk_info **chunkinfo_ret,
		int *chunkcount_ret)
{
	int ret;
	struct btrfs_ioctl_search_args args;
	struct btrfs_ioctl_search_key *sk = &args.key;
	struct btrfs_ioctl_search_header *sh;
	unsigned long off = 0;
	int i;

	memset(&args, 0, sizeof(args));

	/*
	 * there may be more than one ROOT_ITEM key if there are
	 * snapshots pending deletion, we have to loop through
	 * them.
	 */
	sk->tree_id = BTRFS_CHUNK_TREE_OBJECTID;

	sk->min_objectid = 0;
	sk->max_objectid = (u64)-1;
	sk->max_type = 0;
	sk->min_type = (u8)-1;
	sk->min_offset = 0;
	sk->max_offset = (u64)-1;
	sk->min_transid = 0;
	sk->max_transid = (u64)-1;
	sk->nr_items = 4096;

	while (1) {
		ret = ioctl(fd, BTRFS_IOC_TREE_SEARCH, &args);
		if (ret < 0) {
			if (errno == EPERM)
				return -errno;
			error("cannot look up chunk tree info: %m");
			return 1;
		}
		/* the ioctl returns the number of item it found in nr_items */

		if (sk->nr_items == 0)
			break;

		off = 0;
		for (i = 0; i < sk->nr_items; i++) {
			struct btrfs_chunk *item;
			sh = (struct btrfs_ioctl_search_header *)(args.buf +
								  off);

			off += sizeof(*sh);
			item = (struct btrfs_chunk *)(args.buf + off);

			ret = add_info_to_list(chunkinfo_ret, chunkcount_ret, item);
			if (ret) {
				*chunkinfo_ret = NULL;
				return 1;
			}

			off += btrfs_search_header_len(sh);

			sk->min_objectid = btrfs_search_header_objectid(sh);
			sk->min_type = btrfs_search_header_type(sh);
			sk->min_offset = btrfs_search_header_offset(sh)+1;

		}
		if (!sk->min_offset)	/* overflow */
			sk->min_type++;
		else
			continue;

		if (!sk->min_type)
			sk->min_objectid++;
		 else
			continue;

		if (!sk->min_objectid)
			break;
	}

	qsort(*chunkinfo_ret, *chunkcount_ret, sizeof(struct chunk_info),
		cmp_chunk_info);

	return 0;
}

/*
 * Helper to sort the struct btrfs_ioctl_space_info
 */
static int cmp_btrfs_ioctl_space_info(const void *a, const void *b)
{
	return cmp_chunk_block_group(
		((struct btrfs_ioctl_space_info *)a)->flags,
		((struct btrfs_ioctl_space_info *)b)->flags);
}

/*
 * This function load all the information about the space usage
 */
static struct btrfs_ioctl_space_args *load_space_info(int fd, const char *path)
{
	struct btrfs_ioctl_space_args *sargs = NULL, *sargs_orig = NULL;
	int ret, count;

	sargs_orig = sargs = calloc(1, sizeof(struct btrfs_ioctl_space_args));
	if (!sargs) {
		error_msg(ERROR_MSG_MEMORY, NULL);
		return NULL;
	}

	sargs->space_slots = 0;
	sargs->total_spaces = 0;

	ret = ioctl(fd, BTRFS_IOC_SPACE_INFO, sargs);
	if (ret < 0) {
		error("cannot get space info on '%s': %m", path);
		free(sargs);
		return NULL;
	}
	if (!sargs->total_spaces) {
		free(sargs);
		pr_verbose(LOG_DEFAULT, "No chunks found\n");
		return NULL;
	}

	count = sargs->total_spaces;

	sargs = realloc(sargs, sizeof(struct btrfs_ioctl_space_args) +
			(count * sizeof(struct btrfs_ioctl_space_info)));
	if (!sargs) {
		free(sargs_orig);
		error_msg(ERROR_MSG_MEMORY, NULL);
		return NULL;
	}

	sargs->space_slots = count;
	sargs->total_spaces = 0;

	ret = ioctl(fd, BTRFS_IOC_SPACE_INFO, sargs);
	if (ret < 0) {
		error("cannot get space info with %u slots: %m",
			count);
		free(sargs);
		return NULL;
	}

	qsort(&(sargs->spaces), count, sizeof(struct btrfs_ioctl_space_info),
		cmp_btrfs_ioctl_space_info);

	return sargs;
}

/*
 * Compute the ratio between logical space used over logical space allocated
 * by profile basis
 */
static void get_raid56_logical_ratio(struct btrfs_ioctl_space_args *sargs,
				     u64 type, double *data_ratio,
				     double *metadata_ratio,
				     double *system_ratio)
{
	u64 l_data_chunk = 0, l_data_used = 0;
	u64 l_metadata_chunk = 0, l_metadata_used = 0;
	u64 l_system_chunk = 0, l_system_used = 0;
	int i;

	for (i = 0; i < sargs->total_spaces; i++) {
		u64 flags = sargs->spaces[i].flags;

		if (!(flags & type))
			continue;

		if (flags & BTRFS_BLOCK_GROUP_DATA) {
			l_data_used += sargs->spaces[i].used_bytes;
			l_data_chunk += sargs->spaces[i].total_bytes;
		} else if (flags & BTRFS_BLOCK_GROUP_METADATA) {
			l_metadata_used += sargs->spaces[i].used_bytes;
			l_metadata_chunk += sargs->spaces[i].total_bytes;
		} else if (flags & BTRFS_BLOCK_GROUP_SYSTEM) {
			l_system_used += sargs->spaces[i].used_bytes;
			l_system_chunk += sargs->spaces[i].total_bytes;
		}
	}

	*data_ratio = -1.0;
	*metadata_ratio = -1.0;
	*system_ratio = -1.0;

	if (l_data_chunk)
		*data_ratio = (double)l_data_used / l_data_chunk;
	if (l_metadata_chunk)
		*metadata_ratio = (double)l_metadata_used / l_metadata_chunk;
	if (l_system_chunk)
		*system_ratio = (double)l_system_used / l_system_chunk;
}

/*
 * Compute the "raw" space allocated for a chunk (r_*_chunks)
 * and the "raw" space used by a chunk (r_*_used)
 */
static void get_raid56_space_info(struct btrfs_ioctl_space_args *sargs,
				  struct chunk_info *chunkinfo, int chunkcount,
				  double *max_data_ratio,
				  u64 *r_data_chunks, u64 *r_data_used,
				  u64 *r_metadata_chunks, u64 *r_metadata_used,
				  u64 *r_system_chunks, u64 *r_system_used)
{
	struct chunk_info *info_ptr;
	double l_data_ratio_r5, l_metadata_ratio_r5, l_system_ratio_r5;
	double l_data_ratio_r6, l_metadata_ratio_r6, l_system_ratio_r6;

	get_raid56_logical_ratio(sargs, BTRFS_BLOCK_GROUP_RAID5,
		 &l_data_ratio_r5, &l_metadata_ratio_r5, &l_system_ratio_r5);
	get_raid56_logical_ratio(sargs, BTRFS_BLOCK_GROUP_RAID6,
		 &l_data_ratio_r6, &l_metadata_ratio_r6, &l_system_ratio_r6);

	for(info_ptr = chunkinfo; chunkcount > 0; chunkcount--, info_ptr++) {
		int parities_count;
		u64 size;
		double l_data_ratio, l_metadata_ratio, l_system_ratio, rt;

		parities_count = btrfs_bg_type_to_nparity(info_ptr->type);
		if (info_ptr->type & BTRFS_BLOCK_GROUP_RAID5) {
			l_data_ratio = l_data_ratio_r5;
			l_metadata_ratio = l_metadata_ratio_r5;
			l_system_ratio = l_system_ratio_r5;
		} else if (info_ptr->type & BTRFS_BLOCK_GROUP_RAID6) {
			l_data_ratio = l_data_ratio_r6;
			l_metadata_ratio = l_metadata_ratio_r6;
			l_system_ratio = l_system_ratio_r6;
		} else {
			continue;
		}

		rt = (double)info_ptr->num_stripes /
			(info_ptr->num_stripes - parities_count);
		if (rt > *max_data_ratio)
			*max_data_ratio = rt;

		/*
		 * size is the total disk(s) space occupied by a chunk
		 * the product of 'size' and  '*_ratio' is "in average"
		 * the disk(s) space used by the data
		 */
		size = info_ptr->size / (info_ptr->num_stripes - parities_count);

		if (info_ptr->type & BTRFS_BLOCK_GROUP_DATA) {
			UASSERT(l_data_ratio >= 0);
			*r_data_chunks += size;
			*r_data_used += size * l_data_ratio;
		} else if (info_ptr->type & BTRFS_BLOCK_GROUP_METADATA) {
			UASSERT(l_metadata_ratio >= 0);
			*r_metadata_chunks += size;
			*r_metadata_used += size * l_metadata_ratio;
		} else if (info_ptr->type & BTRFS_BLOCK_GROUP_SYSTEM) {
			UASSERT(l_system_ratio >= 0);
			*r_system_chunks += size;
			*r_system_used += size * l_system_ratio;
		}
	}
}

static u64 get_first_device_zone_size(int fd)
{
	int dirfd;
	DIR *dir;
	struct dirent *de;
	char name[NAME_MAX] = {0};
	u64 ret;

	dirfd = sysfs_open_fsid_dir(fd, "devices");
	if (dirfd < 0)
		return 0;
	dir = fdopendir(dirfd);
	if (!dir) {
		ret = 0;
		goto out;
	}
	while (1) {
		de = readdir(dir);
		if (strcmp(".", de->d_name) == 0 || strcmp("..", de->d_name) == 0)
			continue;
		strcpy(name, de->d_name);
		name[NAME_MAX - 1] = 0;
		break;
	}
	ret = device_get_zone_size(fd, name);
	ret *= 512;

out:
	closedir(dir);
	return ret;
}

static u64 calc_slack_size(const struct device_info *devinfo)
{
	if (devinfo->device_size > 0)
		return devinfo->device_size - devinfo->size;
	return 0;
}

#define	MIN_UNALOCATED_THRESH	SZ_16M
static int print_filesystem_usage_overall(int fd, struct chunk_info *chunkinfo,
		int chunkcount, struct device_info *devinfo, int devcount,
		const char *path, unsigned unit_mode)
{
	struct btrfs_ioctl_space_args *sargs = NULL;
	char *tmp;
	int i;
	int ret = 0;
	int width = 10;		/* default 10 for human units */
	/*
	 * r_* prefix is for raw data
	 * l_* prefix is for logical
	 * *_used suffix is for space used for data or metadata
	 * *_chunks suffix is for total space used by the chunk
	 */
	u64 r_total_size = 0;	/* filesystem size, sum of device sizes */
	u64 r_total_chunks = 0;	/* sum of chunks sizes on disk(s) */
	u64 r_total_used = 0;
	u64 r_total_unused = 0;
	u64 r_total_missing = 0;	/* sum of missing devices size */
	u64 r_total_slack = 0;
	u64 r_data_used = 0;
	u64 r_data_chunks = 0;
	u64 l_data_chunks = 0;
	u64 r_metadata_used = 0;
	u64 r_metadata_chunks = 0;
	u64 l_metadata_chunks = 0;
	u64 r_system_used = 0;
	u64 r_system_chunks = 0;
	double data_ratio;
	double metadata_ratio;
	/* logical */
	u64 l_global_reserve = 0;
	u64 l_global_reserve_used = 0;
	u64 free_estimated = 0;
	u64 free_min = 0;
	u64 zone_unusable = 0;
	double max_data_ratio = 1.0;
	bool mixed = false;
	struct statfs statfs_buf;
	struct btrfs_ioctl_feature_flags feature_flags;

	sargs = load_space_info(fd, path);
	if (!sargs) {
		ret = 1;
		goto exit;
	}

	r_total_size = 0;
	for (i = 0; i < devcount; i++) {
		r_total_size += devinfo[i].size;
		r_total_slack += calc_slack_size(&devinfo[i]);
		if (!devinfo[i].device_size)
			r_total_missing += devinfo[i].size;
	}

	if (r_total_size == 0) {
		error("cannot get space info on '%s': %m", path);

		ret = 1;
		goto exit;
	}

	get_raid56_space_info(sargs, chunkinfo, chunkcount, &max_data_ratio,
			      &r_data_chunks, &r_data_used,
			      &r_metadata_chunks, &r_metadata_used,
			      &r_system_chunks, &r_system_used);

	for (i = 0; i < sargs->total_spaces; i++) {
		int ratio;
		u64 flags = sargs->spaces[i].flags;

		ratio = btrfs_bg_type_to_ncopies(flags);

		/*
		 * The RAID5/6 ratio depends on the number of stripes and is
		 * computed separately. Setting ratio to 0 will not account
		 * the chunks in this loop.
		 */
		if (flags & BTRFS_BLOCK_GROUP_RAID56_MASK)
			ratio = 0;

		if (ratio > max_data_ratio)
			max_data_ratio = ratio;

		if (flags & BTRFS_SPACE_INFO_GLOBAL_RSV) {
			l_global_reserve = sargs->spaces[i].total_bytes;
			l_global_reserve_used = sargs->spaces[i].used_bytes;
		}
		if ((flags & (BTRFS_BLOCK_GROUP_DATA | BTRFS_BLOCK_GROUP_METADATA))
		    == (BTRFS_BLOCK_GROUP_DATA | BTRFS_BLOCK_GROUP_METADATA)) {
			mixed = true;
		} else {
			/*
			 * As mixed mode is not supported in zoned mode, this
			 * will account for all profile types
			 */
			u64 tmp;

			tmp = device_get_zone_unusable(fd, flags);
			if (tmp != DEVICE_ZONE_UNUSABLE_UNKNOWN)
				zone_unusable += tmp;
		}
		if (flags & BTRFS_BLOCK_GROUP_DATA) {
			r_data_used += sargs->spaces[i].used_bytes * ratio;
			r_data_chunks += sargs->spaces[i].total_bytes * ratio;
			l_data_chunks += sargs->spaces[i].total_bytes;
		}
		if (flags & BTRFS_BLOCK_GROUP_METADATA) {
			r_metadata_used += sargs->spaces[i].used_bytes * ratio;
			r_metadata_chunks += sargs->spaces[i].total_bytes * ratio;
			l_metadata_chunks += sargs->spaces[i].total_bytes;
		}
		if (flags & BTRFS_BLOCK_GROUP_SYSTEM) {
			r_system_used += sargs->spaces[i].used_bytes * ratio;
			r_system_chunks += sargs->spaces[i].total_bytes * ratio;
		}
	}

	r_total_chunks = r_data_chunks + r_system_chunks;
	r_total_used = r_data_used + r_system_used;
	if (!mixed) {
		r_total_chunks += r_metadata_chunks;
		r_total_used += r_metadata_used;
	}
	r_total_unused = r_total_size - r_total_chunks;

	/* Raw / Logical = raid factor, >= 1 */
	data_ratio = (double)r_data_chunks / l_data_chunks;
	if (mixed)
		metadata_ratio = data_ratio;
	else
		metadata_ratio = (double)r_metadata_chunks / l_metadata_chunks;

	/*
	 * We're able to fill at least DATA for the unused space
	 *
	 * With mixed raid levels, this gives a rough estimate but more
	 * accurate than just counting the logical free space
	 * (l_data_chunks - l_data_used)
	 *
	 * In non-mixed case there's no difference.
	 */
	free_estimated = (r_data_chunks - r_data_used) / data_ratio;
	/*
	 * For mixed-bg the metadata are left out in calculations thus global
	 * reserve would be lost. Part of it could be permanently allocated,
	 * we have to subtract the used bytes so we don't go under zero free.
	 */
	if (mixed)
		free_estimated -= l_global_reserve - l_global_reserve_used;
	free_min = free_estimated;

	/* Chop unallocatable space */
	/* FIXME: must be applied per device */
	if (r_total_unused >= MIN_UNALOCATED_THRESH) {
		free_estimated += r_total_unused / data_ratio;
		/* Match the calculation of 'df', use the highest raid ratio */
		free_min += r_total_unused / max_data_ratio;
	}

	if (unit_mode != UNITS_HUMAN)
		width = 18;

	ret = statfs(path, &statfs_buf);
	if (ret) {
		warning("cannot get space info with statfs() on '%s': %m", path);
		memset(&statfs_buf, 0, sizeof(statfs_buf));
		ret = 0;
	}

	pr_verbose(LOG_DEFAULT, "Overall:\n");

	pr_verbose(LOG_DEFAULT, "    Device size:\t\t%*s\n", width,
		pretty_size_mode(r_total_size, unit_mode));
	pr_verbose(LOG_DEFAULT, "    Device allocated:\t\t%*s\n", width,
		pretty_size_mode(r_total_chunks, unit_mode));
	pr_verbose(LOG_DEFAULT, "    Device unallocated:\t\t%*s\n", width,
		pretty_size_mode(r_total_unused, unit_mode | UNITS_NEGATIVE));
	pr_verbose(LOG_DEFAULT, "    Device missing:\t\t%*s\n", width,
		pretty_size_mode(r_total_missing, unit_mode));
	pr_verbose(LOG_DEFAULT, "    Device slack:\t\t%*s\n", width,
		pretty_size_mode(r_total_slack, unit_mode));
	ret = ioctl(fd, BTRFS_IOC_GET_FEATURES, &feature_flags);
	if (ret == 0 && (feature_flags.incompat_flags & BTRFS_FEATURE_INCOMPAT_ZONED)) {
		u64 zone_size;

		pr_verbose(LOG_DEFAULT, "    Device zone unusable:\t%*s\n", width,
			pretty_size_mode(zone_unusable, unit_mode));
		zone_size = get_first_device_zone_size(fd);
		pr_verbose(LOG_DEFAULT, "    Device zone size:\t\t%*s\n", width,
			pretty_size_mode(zone_size, unit_mode));
	}
	pr_verbose(LOG_DEFAULT, "    Used:\t\t\t%*s\n", width,
		pretty_size_mode(r_total_used, unit_mode));
	pr_verbose(LOG_DEFAULT, "    Free (estimated):\t\t%*s\t(",
		width,
		pretty_size_mode(free_estimated, unit_mode));
	pr_verbose(LOG_DEFAULT, "min: %s)\n", pretty_size_mode(free_min, unit_mode));
	pr_verbose(LOG_DEFAULT, "    Free (statfs, df):\t\t%*s\n", width,
		pretty_size_mode(statfs_buf.f_bavail * statfs_buf.f_bsize, unit_mode));
	pr_verbose(LOG_DEFAULT, "    Data ratio:\t\t\t%*.2f\n",
		width, data_ratio);
	pr_verbose(LOG_DEFAULT, "    Metadata ratio:\t\t%*.2f\n",
		width, metadata_ratio);
	pr_verbose(LOG_DEFAULT, "    Global reserve:\t\t%*s\t(used: %s)\n", width,
		pretty_size_mode(l_global_reserve, unit_mode),
		pretty_size_mode(l_global_reserve_used, unit_mode));
	tmp = btrfs_test_for_multiple_profiles(fd);
	if (tmp[0])
		pr_verbose(LOG_DEFAULT, "    Multiple profiles:\t\t%*s\t(%s)\n", width, "yes", tmp);
	else
		pr_verbose(LOG_DEFAULT, "    Multiple profiles:\t\t%*s\n", width, "no");
	free(tmp);

exit:

	if (sargs)
		free(sargs);

	return ret;
}

/*
 *  Helper to sort the device_info structure
 */
static int cmp_device_info(const void *a, const void *b)
{
	const struct device_info *deva = a;
	const struct device_info *devb = b;

	if (deva->devid < devb->devid)
		return -1;
	if (deva->devid > devb->devid)
		return 1;

	return 0;
}

int dev_to_fsid(const char *dev, u8 *fsid)
{
	struct btrfs_super_block disk_super;
	int ret;
	int fd;

	fd = open(dev, O_RDONLY);
	if (fd < 0) {
		ret = -errno;
		return ret;
	}

	ret = btrfs_read_dev_super(fd, &disk_super,
				   BTRFS_SUPER_INFO_OFFSET, SBREAD_DEFAULT);
	if (ret)
		goto out;

	memcpy(fsid, disk_super.fsid, BTRFS_FSID_SIZE);
	ret = 0;

out:
	close(fd);
	return ret;
}

static int device_is_seed(int fd, const char *dev_path, u64 devid, const u8 *mnt_fsid)
{
	char fsid_str[BTRFS_UUID_UNPARSED_SIZE];
	char fsid_path[PATH_MAX];
	char devid_str[20];
	u8 fsid[BTRFS_UUID_SIZE];
	int ret = -1;
	int sysfs_fd;

	snprintf(devid_str, 20, "%llu", devid);
	/* devinfo/<devid>/fsid */
	ret = path_cat3_out(fsid_path, "devinfo", devid_str, "fsid");
	if (ret < 0)
		return ret;

	/* /sys/fs/btrfs/<fsid>/devinfo/<devid>/fsid */
	sysfs_fd = sysfs_open_fsid_file(fd, fsid_path);
	if (sysfs_fd >= 0) {
		sysfs_read_file(sysfs_fd, fsid_str, BTRFS_UUID_UNPARSED_SIZE);
		fsid_str[BTRFS_UUID_UNPARSED_SIZE - 1] = 0;
		ret = uuid_parse(fsid_str, fsid);
		close(sysfs_fd);
	}

	if (ret || sysfs_fd < 0) {
		ret = dev_to_fsid(dev_path, fsid);
		if (ret)
			return ret;
	}

	if (memcmp(mnt_fsid, fsid, BTRFS_FSID_SIZE) != 0)
		return 0;

	return -1;
}

/*
 *  This function loads the device_info structure and put them in an array
 */
static int load_device_info(int fd, struct device_info **devinfo_ret,
			   int *devcount_ret)
{
	int ret, i, ndevs;
	struct btrfs_ioctl_fs_info_args fi_args;
	struct btrfs_ioctl_dev_info_args dev_info;
	struct device_info *info;

	*devcount_ret = 0;
	*devinfo_ret = NULL;

	ret = ioctl(fd, BTRFS_IOC_FS_INFO, &fi_args);
	if (ret < 0) {
		if (errno == EPERM)
			return -errno;
		error("cannot get filesystem info: %m");
		return 1;
	}

	info = calloc(fi_args.num_devices, sizeof(struct device_info));
	if (!info) {
		error_msg(ERROR_MSG_MEMORY, NULL);
		return 1;
	}

	for (i = 0, ndevs = 0 ; i <= fi_args.max_id ; i++) {
		if (ndevs >= fi_args.num_devices) {
			error("unexpected number of devices: %d >= %llu", ndevs,
				fi_args.num_devices);
			error(
		"if seed device is used, try running this command as root");
			goto out;
		}
		memset(&dev_info, 0, sizeof(dev_info));
		ret = device_get_info(fd, i, &dev_info);

		if (ret == -ENODEV)
			continue;
		if (ret) {
			error("cannot get info about device devid=%d", i);
			goto out;
		}

		/*
		 * Skip seed device by checking device's fsid (requires root if
		 * kernel is not patched to provide fsid from the sysfs).
		 * And we will skip only if device_is_seed is successful and dev
		 * is a seed device.
		 * Ignore any other error including -EACCES, which is seen when
		 * a non-root process calls dev_to_fsid(path)->open(path).
		 */
		ret = device_is_seed(fd, (const char *)dev_info.path, i, fi_args.fsid);
		if (!ret)
			continue;

		info[ndevs].devid = dev_info.devid;
		if (!dev_info.path[0]) {
			strcpy(info[ndevs].path, "missing");
		} else {
			strcpy(info[ndevs].path, (char *)dev_info.path);
			info[ndevs].device_size =
				device_get_partition_size((const char *)dev_info.path);
		}
		info[ndevs].size = dev_info.total_bytes;
		++ndevs;
	}

	if (ndevs != fi_args.num_devices) {
		error("unexpected number of devices: %d != %llu", ndevs,
				fi_args.num_devices);
		goto out;
	}

	qsort(info, fi_args.num_devices,
		sizeof(struct device_info), cmp_device_info);

	*devcount_ret = fi_args.num_devices;
	*devinfo_ret = info;

	return 0;

out:
	free(info);
	return ret;
}

int load_chunk_and_device_info(int fd, struct chunk_info **chunkinfo_ret,
		int *chunkcount_ret, struct device_info **devinfo_ret,
		int *devcount_ret)
{
	int ret;

	ret = load_chunk_info(fd, chunkinfo_ret, chunkcount_ret);
	if (ret == -EPERM) {
		warning(
"cannot read detailed chunk info, per-device usage will not be shown, run as root");
	} else if (ret) {
		return ret;
	}

	ret = load_device_info(fd, devinfo_ret, devcount_ret);
	if (ret == -EPERM) {
		warning(
		"cannot get filesystem info from ioctl(FS_INFO), run as root");
		ret = 0;
	}

	return ret;
}

/*
 *  This function computes the size of a chunk in a disk
 */
static u64 calc_chunk_size(struct chunk_info *ci)
{
	u32 div = 1;

	/*
	 * The formula doesn't work for RAID1/DUP types, we should just return the
	 * chunk size
	 */
	if (!(ci->type & (BTRFS_BLOCK_GROUP_RAID1_MASK|BTRFS_BLOCK_GROUP_DUP))) {
		/* No parity + sub_stripes, so order of "-" and "/" does not matter */
		div = (ci->num_stripes - btrfs_bg_type_to_nparity(ci->type)) /
			btrfs_bg_type_to_sub_stripes(ci->type);
	}

	return ci->size / div;
}

/*
 *  This function print the results of the command "btrfs fi usage"
 *  in tabular format
 */
static void _cmd_filesystem_usage_tabular(unsigned unit_mode,
					struct btrfs_ioctl_space_args *sargs,
					struct chunk_info *chunks_info_ptr,
					int chunks_info_count,
					struct device_info *devinfo,
					int devcount)
{
	int i;
	u64 total_unused = 0;
	u64 total_total = 0;
	u64 total_slack = 0;
	struct string_table *matrix = NULL;
	int  ncols, nrows;
	int col;
	int unallocated_col;
	int spaceinfos_col;
	int total_col;
	int slack_col;
	u64 slack;
	const int vhdr_skip = 3;	/* amount of vertical header space */

	/* id, path, unallocated, total, slack */
	ncols = 5;
	spaceinfos_col = 2;
	/* Properly count the real space infos */
	for (i = 0; i < sargs->total_spaces; i++) {
		if (sargs->spaces[i].flags & BTRFS_SPACE_INFO_GLOBAL_RSV)
			continue;
		ncols++;
	}

	/* 2 for header, empty line, devices, ===, total, used */
	nrows = vhdr_skip + devcount + 1 + 2;

	matrix = table_create(ncols, nrows);
	if (!matrix) {
		error_msg(ERROR_MSG_MEMORY, NULL);
		return;
	}

	/*
	 * We have to skip the global block reserve everywhere as it's an
	 * artificial blockgroup
	 */

	/* header */
	for (i = 0, col = spaceinfos_col; i < sargs->total_spaces; i++) {
		u64 flags = sargs->spaces[i].flags;

		if (flags & BTRFS_SPACE_INFO_GLOBAL_RSV)
			continue;

		table_printf(matrix, col, 0, "<%s",
				btrfs_group_type_str(flags));
		table_printf(matrix, col, 1, "<%s",
				btrfs_group_profile_str(flags));
		col++;
	}
	unallocated_col = col++;
	total_col = col++;
	slack_col = col++;

	table_printf(matrix, 0, 1, "<Id");
	table_printf(matrix, 1, 1, "<Path");
	table_printf(matrix, unallocated_col, 1, "<Unallocated");
	table_printf(matrix, total_col, 1, "<Total");
	table_printf(matrix, slack_col, 1, "<Slack");

	/* body */
	for (i = 0; i < devcount; i++) {
		int k;
		char *p;

		u64  total_allocated = 0, unused;

		p = strrchr(devinfo[i].path, '/');
		if (!p)
			p = devinfo[i].path;
		else
			p++;

		table_printf(matrix, 0, vhdr_skip + i, ">%llu",
				devinfo[i].devid);
		table_printf(matrix, 1, vhdr_skip + i, "<%s",
				devinfo[i].path);

		for (col = spaceinfos_col, k = 0; k < sargs->total_spaces; k++) {
			u64	flags = sargs->spaces[k].flags;
			u64 devid = devinfo[i].devid;
			int	j;
			u64 size = 0;

			if (flags & BTRFS_SPACE_INFO_GLOBAL_RSV)
				continue;

			for (j = 0 ; j < chunks_info_count ; j++) {
				if (chunks_info_ptr[j].type != flags )
						continue;
				if (chunks_info_ptr[j].devid != devid)
						continue;

				size += calc_chunk_size(chunks_info_ptr+j);
			}

			if (size)
				table_printf(matrix, col, vhdr_skip+ i,
					">%s", pretty_size_mode(size, unit_mode));
			else
				table_printf(matrix, col, vhdr_skip + i, ">-");

			total_allocated += size;
			col++;
		}

		unused = device_get_partition_size(devinfo[i].path)
				- total_allocated;
		unused = devinfo[i].size - total_allocated;

		table_printf(matrix, unallocated_col, vhdr_skip + i, ">%s",
			pretty_size_mode(unused, unit_mode | UNITS_NEGATIVE));
		table_printf(matrix, total_col, vhdr_skip + i, ">%s",
			pretty_size_mode(devinfo[i].size,
			unit_mode | UNITS_NEGATIVE));
		slack = calc_slack_size(&devinfo[i]);
		if (slack > 0) {
			table_printf(matrix, slack_col, vhdr_skip + i, ">%s",
				pretty_size_mode(slack,
				unit_mode | UNITS_NEGATIVE));
		} else {
			table_printf(matrix, slack_col, vhdr_skip + i, ">-");
		}
		total_unused += unused;
		total_slack += slack;
		total_total += devinfo[i].size;

	}

	for (i = 0; i < spaceinfos_col; i++) {
		table_printf(matrix, i, vhdr_skip - 1, "*-");
		table_printf(matrix, i, vhdr_skip + devcount, "*-");
	}

	for (i = 0, col = spaceinfos_col; i < sargs->total_spaces; i++) {
		if (sargs->spaces[i].flags & BTRFS_SPACE_INFO_GLOBAL_RSV)
			continue;

		table_printf(matrix, col, vhdr_skip - 1, "*-");
		table_printf(matrix, col, vhdr_skip + devcount, "*-");
		col++;
	}
	/* Line under Unallocated, Total, Slack */
	table_printf(matrix, col, vhdr_skip - 1, "*-");
	table_printf(matrix, col, vhdr_skip + devcount, "*-");
	table_printf(matrix, col + 1, vhdr_skip - 1, "*-");
	table_printf(matrix, col + 1, vhdr_skip + devcount, "*-");
	table_printf(matrix, col + 2, vhdr_skip - 1, "*-");
	table_printf(matrix, col + 2, vhdr_skip + devcount, "*-");

	/* footer */
	table_printf(matrix, 1, vhdr_skip + devcount + 1, "<Total");
	for (i = 0, col = spaceinfos_col; i < sargs->total_spaces; i++) {
		if (sargs->spaces[i].flags & BTRFS_SPACE_INFO_GLOBAL_RSV)
			continue;

		table_printf(matrix, col++, vhdr_skip + devcount + 1,
			">%s",
			pretty_size_mode(sargs->spaces[i].total_bytes, unit_mode));
	}

	table_printf(matrix, unallocated_col, vhdr_skip + devcount + 1,
		">%s",
		pretty_size_mode(total_unused, unit_mode | UNITS_NEGATIVE));
	table_printf(matrix, total_col, vhdr_skip + devcount + 1,
		">%s",
		pretty_size_mode(total_total, unit_mode | UNITS_NEGATIVE));
	table_printf(matrix, slack_col, vhdr_skip + devcount + 1,
		">%s",
		pretty_size_mode(total_slack, unit_mode | UNITS_NEGATIVE));

	table_printf(matrix, 1, vhdr_skip + devcount + 2, "<Used");
	for (i = 0, col = spaceinfos_col; i < sargs->total_spaces; i++) {
		if (sargs->spaces[i].flags & BTRFS_SPACE_INFO_GLOBAL_RSV)
			continue;

		table_printf(matrix, col++, vhdr_skip + devcount + 2,
			">%s",
			pretty_size_mode(sargs->spaces[i].used_bytes, unit_mode));
	}

	table_dump(matrix);
	table_free(matrix);
}

/*
 *  This function prints the unused space per every disk
 */
static void print_unused(struct chunk_info *info_ptr,
			  int info_count,
			  struct device_info *devinfo,
			  int devcount,
			  unsigned unit_mode)
{
	int i;
	for (i = 0; i < devcount; i++) {
		int	j;
		u64	total = 0;

		for (j = 0; j < info_count; j++)
			if (info_ptr[j].devid == devinfo[i].devid)
				total += calc_chunk_size(info_ptr+j);

		pr_verbose(LOG_DEFAULT, "   %s\t%10s\n",
			devinfo[i].path,
			pretty_size_mode(devinfo[i].size - total, unit_mode));
	}
}

/*
 *  This function prints the allocated chunk per every disk
 */
static void print_chunk_device(u64 chunk_type,
				struct chunk_info *chunks_info_ptr,
				int chunks_info_count,
				struct device_info *devinfo,
				int devcount,
				unsigned unit_mode)
{
	int i;

	for (i = 0; i < devcount; i++) {
		int	j;
		u64	total = 0;

		for (j = 0; j < chunks_info_count; j++) {

			if (chunks_info_ptr[j].type != chunk_type)
				continue;
			if (chunks_info_ptr[j].devid != devinfo[i].devid)
				continue;

			total += calc_chunk_size(&(chunks_info_ptr[j]));
			//total += chunks_info_ptr[j].size;
		}

		if (total > 0)
			pr_verbose(LOG_DEFAULT, "   %s\t%10s\n",
				devinfo[i].path,
				pretty_size_mode(total, unit_mode));
	}
}

/*
 *  This function print the results of the command "btrfs fi usage"
 *  in linear format
 */
static void _cmd_filesystem_usage_linear(unsigned unit_mode,
					struct btrfs_ioctl_space_args *sargs,
					struct chunk_info *info_ptr,
					int info_count,
					struct device_info *devinfo,
					int devcount)
{
	int i;

	for (i = 0; i < sargs->total_spaces; i++) {
		const char *description;
		const char *r_mode;
		u64 flags = sargs->spaces[i].flags;

		if (flags & BTRFS_SPACE_INFO_GLOBAL_RSV)
			continue;

		description = btrfs_group_type_str(flags);
		r_mode = btrfs_group_profile_str(flags);

		pr_verbose(LOG_DEFAULT, "%s,%s: Size:%s, ",
			description,
			r_mode,
			pretty_size_mode(sargs->spaces[i].total_bytes,
				unit_mode));
		pr_verbose(LOG_DEFAULT, "Used:%s (%.2f%%)\n",
			pretty_size_mode(sargs->spaces[i].used_bytes, unit_mode),
			100.0f * sargs->spaces[i].used_bytes /
			(sargs->spaces[i].total_bytes + 1));
		print_chunk_device(flags, info_ptr, info_count,
				devinfo, devcount, unit_mode);
		pr_verbose(LOG_DEFAULT, "\n");
	}

	if (info_count) {
		pr_verbose(LOG_DEFAULT, "Unallocated:\n");
		print_unused(info_ptr, info_count, devinfo,
				devcount, unit_mode | UNITS_NEGATIVE);
	}
}

static int print_filesystem_usage_by_chunk(int fd,
		struct chunk_info *chunkinfo, int chunkcount,
		struct device_info *devinfo, int devcount,
		const char *path, unsigned unit_mode, int tabular)
{
	struct btrfs_ioctl_space_args *sargs;
	int ret = 0;

	sargs = load_space_info(fd, path);
	if (!sargs) {
		ret = 1;
		goto out;
	}

	if (tabular)
		_cmd_filesystem_usage_tabular(unit_mode, sargs, chunkinfo,
				chunkcount, devinfo, devcount);
	else
		_cmd_filesystem_usage_linear(unit_mode, sargs, chunkinfo,
				chunkcount, devinfo, devcount);

	free(sargs);
out:
	return ret;
}

static const char * const cmd_filesystem_usage_usage[] = {
	"btrfs filesystem usage [options] <path> [<path>..]",
	"Show detailed information about internal filesystem usage .",
	"",
	HELPINFO_UNITS_SHORT_LONG,
	OPTLINE("-T", "show data in tabular format"),
	NULL
};

static int cmd_filesystem_usage(const struct cmd_struct *cmd,
				int argc, char **argv)
{
	int ret = 0;
	unsigned unit_mode;
	int i;
	int more_than_one = 0;
	int tabular = 0;

	unit_mode = get_unit_mode_from_arg(&argc, argv, 1);

	optind = 0;
	while (1) {
		int c;

		c = getopt(argc, argv, "T");
		if (c < 0)
			break;

		switch (c) {
		case 'T':
			tabular = 1;
			break;
		default:
			usage_unknown_option(cmd, argv);
		}
	}

	if (check_argc_min(argc - optind, 1))
		return 1;

	for (i = optind; i < argc; i++) {
		int fd;
		DIR *dirstream = NULL;
		struct chunk_info *chunkinfo = NULL;
		struct device_info *devinfo = NULL;
		int chunkcount = 0;
		int devcount = 0;

		fd = btrfs_open_dir(argv[i], &dirstream, 1);
		if (fd < 0) {
			ret = 1;
			goto out;
		}
		if (more_than_one)
			pr_verbose(LOG_DEFAULT, "\n");

		ret = load_chunk_and_device_info(fd, &chunkinfo, &chunkcount,
				&devinfo, &devcount);
		if (ret)
			goto cleanup;

		ret = print_filesystem_usage_overall(fd, chunkinfo, chunkcount,
				devinfo, devcount, argv[i], unit_mode);
		if (ret)
			goto cleanup;
		pr_verbose(LOG_DEFAULT, "\n");
		ret = print_filesystem_usage_by_chunk(fd, chunkinfo, chunkcount,
				devinfo, devcount, argv[i], unit_mode, tabular);
cleanup:
		close_file_or_dir(fd, dirstream);
		free(chunkinfo);
		free(devinfo);

		if (ret)
			goto out;
		more_than_one = 1;
	}

out:
	return !!ret;
}
DEFINE_SIMPLE_COMMAND(filesystem_usage, "usage");

void print_device_chunks(struct device_info *devinfo,
		struct chunk_info *chunks_info_ptr,
		int chunks_info_count, unsigned unit_mode)
{
	int i;
	u64 allocated = 0;

	for (i = 0 ; i < chunks_info_count ; i++) {
		const char *description;
		const char *r_mode;
		u64 flags;
		u64 size;
		u64 num_stripes;
		u64 profile;

		if (chunks_info_ptr[i].devid != devinfo->devid)
			continue;

		flags = chunks_info_ptr[i].type;
		profile = (flags & BTRFS_BLOCK_GROUP_PROFILE_MASK);

		description = btrfs_group_type_str(flags);
		r_mode = btrfs_group_profile_str(flags);
		size = calc_chunk_size(chunks_info_ptr+i);
		num_stripes = chunks_info_ptr[i].num_stripes;

		if (btrfs_bg_type_is_stripey(profile)) {
			pr_verbose(LOG_DEFAULT, "   %s,%s/%llu:%*s%10s\n",
				   description,
				   r_mode,
				   num_stripes,
				   (int)(20 - strlen(description) - strlen(r_mode)
						 - count_digits(num_stripes) - 1), "",
				   pretty_size_mode(size, unit_mode));
		} else {
			pr_verbose(LOG_DEFAULT, "   %s,%s:%*s%10s\n",
				   description,
				   r_mode,
				   (int)(20 - strlen(description) - strlen(r_mode)), "",
				   pretty_size_mode(size, unit_mode));
		}

		allocated += size;

	}
	pr_verbose(LOG_DEFAULT, "   Unallocated: %*s%10s\n",
		(int)(20 - strlen("Unallocated")), "",
		pretty_size_mode(devinfo->size - allocated,
			unit_mode | UNITS_NEGATIVE));
}

void print_device_sizes(struct device_info *devinfo, unsigned unit_mode)
{
	pr_verbose(LOG_DEFAULT, "   Device size: %*s%10s\n",
		(int)(20 - strlen("Device size")), "",
		pretty_size_mode(devinfo->device_size, unit_mode));
	pr_verbose(LOG_DEFAULT, "   Device slack: %*s%10s\n",
		(int)(20 - strlen("Device slack")), "",
		pretty_size_mode(calc_slack_size(devinfo), unit_mode));
}