mirror of
https://github.com/kdave/btrfs-progs
synced 2024-12-09 08:00:13 +00:00
e4df433b8a
As btrfs is specific to Linux, %m can be used instead of strerror(errno) in format strings. This has some size reduction benefits for embedded systems. glibc, musl, and uclibc-ng all support %m as a modifier to printf. A quick glance at the BIONIC libc source indicates that it has support for %m as well. BSDs and Windows do not but I do believe them to be beyond the scope of btrfs-progs. Compiled sizes on Ubuntu 16.04: Before: 3916512 btrfs 233688 libbtrfs.so.0.1 4899 bcp 2367672 btrfs-convert 2208488 btrfs-corrupt-block 13302 btrfs-debugfs 2152160 btrfs-debug-tree 2136024 btrfs-find-root 2287592 btrfs-image 2144600 btrfs-map-logical 2130760 btrfs-select-super 2152608 btrfstune 2131760 btrfs-zero-log 2277752 mkfs.btrfs 9166 show-blocks After: 3908744 btrfs 233256 libbtrfs.so.0.1 4899 bcp 2366560 btrfs-convert 2207432 btrfs-corrupt-block 13302 btrfs-debugfs 2151104 btrfs-debug-tree 2134968 btrfs-find-root 2281864 btrfs-image 2143536 btrfs-map-logical 2129704 btrfs-select-super 2151552 btrfstune 2130696 btrfs-zero-log 2276272 mkfs.btrfs 9166 show-blocks Total savings: 23928 (24 kilo)bytes Signed-off-by: Rosen Penev <rosenp@gmail.com> Signed-off-by: David Sterba <dsterba@suse.com>
1086 lines
26 KiB
C
1086 lines
26 KiB
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 <stdio.h>
|
|
#include <stdlib.h>
|
|
#include <string.h>
|
|
#include <unistd.h>
|
|
#include <sys/ioctl.h>
|
|
#include <errno.h>
|
|
#include <stdarg.h>
|
|
#include <getopt.h>
|
|
#include <fcntl.h>
|
|
|
|
#include "utils.h"
|
|
#include "kerncompat.h"
|
|
#include "ctree.h"
|
|
#include "string-table.h"
|
|
#include "cmds-fi-usage.h"
|
|
#include "commands.h"
|
|
#include "disk-io.h"
|
|
|
|
#include "version.h"
|
|
#include "help.h"
|
|
|
|
/*
|
|
* Add the chunk info to the chunk_info list
|
|
*/
|
|
static int add_info_to_list(struct chunk_info **info_ptr,
|
|
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 ((*info_ptr)[i].type == type &&
|
|
(*info_ptr)[i].devid == devid &&
|
|
(*info_ptr)[i].num_stripes == num_stripes ) {
|
|
p = (*info_ptr) + i;
|
|
break;
|
|
}
|
|
|
|
if (!p) {
|
|
int tmp = sizeof(struct btrfs_chunk) * (*info_count + 1);
|
|
struct chunk_info *res = realloc(*info_ptr, tmp);
|
|
|
|
if (!res) {
|
|
free(*info_ptr);
|
|
error("not enough memory");
|
|
return -ENOMEM;
|
|
}
|
|
|
|
*info_ptr = 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 **info_ptr, int *info_count)
|
|
{
|
|
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, e;
|
|
|
|
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);
|
|
e = errno;
|
|
if (e == EPERM)
|
|
return -e;
|
|
|
|
if (ret < 0) {
|
|
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(info_ptr, info_count, item);
|
|
if (ret) {
|
|
*info_ptr = 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(*info_ptr, *info_count, 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, 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("not enough memory");
|
|
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);
|
|
printf("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("not enough memory");
|
|
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;
|
|
}
|
|
|
|
/*
|
|
* This function computes the space occupied by a *single* RAID5/RAID6 chunk.
|
|
* The computation is performed on the basis of the number of stripes
|
|
* which compose the chunk, which could be different from the number of devices
|
|
* if a disk is added later.
|
|
*/
|
|
static void get_raid56_used(struct chunk_info *chunks, int chunkcount,
|
|
u64 *raid5_used, u64 *raid6_used)
|
|
{
|
|
struct chunk_info *info_ptr = chunks;
|
|
*raid5_used = 0;
|
|
*raid6_used = 0;
|
|
|
|
while (chunkcount-- > 0) {
|
|
if (info_ptr->type & BTRFS_BLOCK_GROUP_RAID5)
|
|
(*raid5_used) += info_ptr->size / (info_ptr->num_stripes - 1);
|
|
if (info_ptr->type & BTRFS_BLOCK_GROUP_RAID6)
|
|
(*raid6_used) += info_ptr->size / (info_ptr->num_stripes - 2);
|
|
info_ptr++;
|
|
}
|
|
}
|
|
|
|
#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,
|
|
char *path, unsigned unit_mode)
|
|
{
|
|
struct btrfs_ioctl_space_args *sargs = NULL;
|
|
int i;
|
|
int ret = 0;
|
|
int width = 10; /* default 10 for human units */
|
|
/*
|
|
* r_* prefix is for raw data
|
|
* l_* is for logical
|
|
*/
|
|
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_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 raid5_used = 0;
|
|
u64 raid6_used = 0;
|
|
u64 l_global_reserve = 0;
|
|
u64 l_global_reserve_used = 0;
|
|
u64 free_estimated = 0;
|
|
u64 free_min = 0;
|
|
int max_data_ratio = 1;
|
|
int mixed = 0;
|
|
|
|
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;
|
|
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_used(chunkinfo, chunkcount, &raid5_used, &raid6_used);
|
|
|
|
for (i = 0; i < sargs->total_spaces; i++) {
|
|
int ratio;
|
|
u64 flags = sargs->spaces[i].flags;
|
|
|
|
/*
|
|
* The raid5/raid6 ratio depends by the stripes number
|
|
* used by every chunk. It is computed separately
|
|
*/
|
|
if (flags & BTRFS_BLOCK_GROUP_RAID0)
|
|
ratio = 1;
|
|
else if (flags & BTRFS_BLOCK_GROUP_RAID1)
|
|
ratio = 2;
|
|
else if (flags & BTRFS_BLOCK_GROUP_RAID5)
|
|
ratio = 0;
|
|
else if (flags & BTRFS_BLOCK_GROUP_RAID6)
|
|
ratio = 0;
|
|
else if (flags & BTRFS_BLOCK_GROUP_DUP)
|
|
ratio = 2;
|
|
else if (flags & BTRFS_BLOCK_GROUP_RAID10)
|
|
ratio = 2;
|
|
else
|
|
ratio = 1;
|
|
|
|
if (!ratio)
|
|
warning("RAID56 detected, not implemented");
|
|
|
|
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 = 1;
|
|
}
|
|
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;
|
|
|
|
#if 0
|
|
/* add the raid5/6 allocated space */
|
|
total_chunks += raid5_used + raid6_used;
|
|
#endif
|
|
|
|
/*
|
|
* 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;
|
|
|
|
printf("Overall:\n");
|
|
|
|
printf(" Device size:\t\t%*s\n", width,
|
|
pretty_size_mode(r_total_size, unit_mode));
|
|
printf(" Device allocated:\t\t%*s\n", width,
|
|
pretty_size_mode(r_total_chunks, unit_mode));
|
|
printf(" Device unallocated:\t\t%*s\n", width,
|
|
pretty_size_mode(r_total_unused, unit_mode | UNITS_NEGATIVE));
|
|
printf(" Device missing:\t\t%*s\n", width,
|
|
pretty_size_mode(r_total_missing, unit_mode));
|
|
printf(" Used:\t\t\t%*s\n", width,
|
|
pretty_size_mode(r_total_used, unit_mode));
|
|
printf(" Free (estimated):\t\t%*s\t(",
|
|
width,
|
|
pretty_size_mode(free_estimated, unit_mode));
|
|
printf("min: %s)\n", pretty_size_mode(free_min, unit_mode));
|
|
printf(" Data ratio:\t\t\t%*.2f\n",
|
|
width, data_ratio);
|
|
printf(" Metadata ratio:\t\t%*.2f\n",
|
|
width, metadata_ratio);
|
|
printf(" 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));
|
|
|
|
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)
|
|
{
|
|
return strcmp(((struct device_info *)a)->path,
|
|
((struct device_info *)b)->path);
|
|
}
|
|
|
|
int dev_to_fsid(const char *dev, u8 *fsid)
|
|
{
|
|
struct btrfs_super_block *disk_super;
|
|
char buf[BTRFS_SUPER_INFO_SIZE];
|
|
int ret;
|
|
int fd;
|
|
|
|
fd = open(dev, O_RDONLY);
|
|
if (fd < 0) {
|
|
ret = -errno;
|
|
return ret;
|
|
}
|
|
|
|
disk_super = (struct btrfs_super_block *)buf;
|
|
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;
|
|
}
|
|
|
|
/*
|
|
* This function loads the device_info structure and put them in an array
|
|
*/
|
|
static int load_device_info(int fd, struct device_info **device_info_ptr,
|
|
int *device_info_count)
|
|
{
|
|
int ret, i, ndevs;
|
|
struct btrfs_ioctl_fs_info_args fi_args;
|
|
struct btrfs_ioctl_dev_info_args dev_info;
|
|
struct device_info *info;
|
|
u8 fsid[BTRFS_UUID_SIZE];
|
|
|
|
*device_info_count = 0;
|
|
*device_info_ptr = 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("not enough memory");
|
|
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,
|
|
(unsigned long long)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 = get_device_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).
|
|
* And we will skip only if dev_to_fsid 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 = dev_to_fsid((const char *)dev_info.path, fsid);
|
|
if (!ret && memcmp(fi_args.fsid, fsid, BTRFS_FSID_SIZE) != 0)
|
|
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 =
|
|
get_partition_size((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,
|
|
(unsigned long long)fi_args.num_devices);
|
|
goto out;
|
|
}
|
|
|
|
qsort(info, fi_args.num_devices,
|
|
sizeof(struct device_info), cmp_device_info);
|
|
|
|
*device_info_count = fi_args.num_devices;
|
|
*device_info_ptr = info;
|
|
|
|
return 0;
|
|
|
|
out:
|
|
free(info);
|
|
return ret;
|
|
}
|
|
|
|
int load_chunk_and_device_info(int fd, struct chunk_info **chunkinfo,
|
|
int *chunkcount, struct device_info **devinfo, int *devcount)
|
|
{
|
|
int ret;
|
|
|
|
ret = load_chunk_info(fd, chunkinfo, chunkcount);
|
|
if (ret == -EPERM) {
|
|
warning(
|
|
"cannot read detailed chunk info, RAID5/6 numbers will be incorrect, run as root");
|
|
} else if (ret) {
|
|
return ret;
|
|
}
|
|
|
|
ret = load_device_info(fd, devinfo, devcount);
|
|
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)
|
|
{
|
|
if (ci->type & BTRFS_BLOCK_GROUP_RAID0)
|
|
return ci->size / ci->num_stripes;
|
|
else if (ci->type & BTRFS_BLOCK_GROUP_RAID1)
|
|
return ci->size ;
|
|
else if (ci->type & BTRFS_BLOCK_GROUP_DUP)
|
|
return ci->size ;
|
|
else if (ci->type & BTRFS_BLOCK_GROUP_RAID5)
|
|
return ci->size / (ci->num_stripes -1);
|
|
else if (ci->type & BTRFS_BLOCK_GROUP_RAID6)
|
|
return ci->size / (ci->num_stripes -2);
|
|
else if (ci->type & BTRFS_BLOCK_GROUP_RAID10)
|
|
return ci->size / ci->num_stripes;
|
|
return ci->size;
|
|
}
|
|
|
|
/*
|
|
* 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 *device_info_ptr,
|
|
int device_info_count)
|
|
{
|
|
int i;
|
|
u64 total_unused = 0;
|
|
struct string_table *matrix = NULL;
|
|
int ncols, nrows;
|
|
int col;
|
|
int unallocated_col;
|
|
int spaceinfos_col;
|
|
const int vhdr_skip = 3; /* amount of vertical header space */
|
|
|
|
/* id, path, unallocated */
|
|
ncols = 3;
|
|
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 + device_info_count + 1 + 2;
|
|
|
|
matrix = table_create(ncols, nrows);
|
|
if (!matrix) {
|
|
error("not enough memory");
|
|
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;
|
|
|
|
table_printf(matrix, 0, 1, "<Id");
|
|
table_printf(matrix, 1, 1, "<Path");
|
|
table_printf(matrix, unallocated_col, 1, "<Unallocated");
|
|
|
|
/* body */
|
|
for (i = 0; i < device_info_count; i++) {
|
|
int k;
|
|
char *p;
|
|
|
|
u64 total_allocated = 0, unused;
|
|
|
|
p = strrchr(device_info_ptr[i].path, '/');
|
|
if (!p)
|
|
p = device_info_ptr[i].path;
|
|
else
|
|
p++;
|
|
|
|
table_printf(matrix, 0, vhdr_skip + i, ">%llu",
|
|
device_info_ptr[i].devid);
|
|
table_printf(matrix, 1, vhdr_skip + i, "<%s",
|
|
device_info_ptr[i].path);
|
|
|
|
for (col = spaceinfos_col, k = 0; k < sargs->total_spaces; k++) {
|
|
u64 flags = sargs->spaces[k].flags;
|
|
u64 devid = device_info_ptr[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 = get_partition_size(device_info_ptr[i].path)
|
|
- total_allocated;
|
|
|
|
table_printf(matrix, unallocated_col, vhdr_skip + i, ">%s",
|
|
pretty_size_mode(unused, unit_mode | UNITS_NEGATIVE));
|
|
total_unused += unused;
|
|
|
|
}
|
|
|
|
for (i = 0; i < spaceinfos_col; i++) {
|
|
table_printf(matrix, i, vhdr_skip - 1, "*-");
|
|
table_printf(matrix, i, vhdr_skip + device_info_count, "*-");
|
|
}
|
|
|
|
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 + device_info_count, "*-");
|
|
col++;
|
|
}
|
|
/* One for Unallocated */
|
|
table_printf(matrix, col, vhdr_skip - 1, "*-");
|
|
table_printf(matrix, col, vhdr_skip + device_info_count, "*-");
|
|
|
|
/* footer */
|
|
table_printf(matrix, 1, vhdr_skip + device_info_count + 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 + device_info_count + 1,
|
|
">%s",
|
|
pretty_size_mode(sargs->spaces[i].total_bytes, unit_mode));
|
|
}
|
|
|
|
table_printf(matrix, unallocated_col, vhdr_skip + device_info_count + 1,
|
|
">%s",
|
|
pretty_size_mode(total_unused, unit_mode | UNITS_NEGATIVE));
|
|
|
|
table_printf(matrix, 1, vhdr_skip + device_info_count + 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 + device_info_count + 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 *device_info_ptr,
|
|
int device_info_count,
|
|
unsigned unit_mode)
|
|
{
|
|
int i;
|
|
for (i = 0; i < device_info_count; i++) {
|
|
int j;
|
|
u64 total = 0;
|
|
|
|
for (j = 0; j < info_count; j++)
|
|
if (info_ptr[j].devid == device_info_ptr[i].devid)
|
|
total += calc_chunk_size(info_ptr+j);
|
|
|
|
printf(" %s\t%10s\n",
|
|
device_info_ptr[i].path,
|
|
pretty_size_mode(device_info_ptr[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 *device_info_ptr,
|
|
int device_info_count,
|
|
unsigned unit_mode)
|
|
{
|
|
int i;
|
|
|
|
for (i = 0; i < device_info_count; 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 != device_info_ptr[i].devid)
|
|
continue;
|
|
|
|
total += calc_chunk_size(&(chunks_info_ptr[j]));
|
|
//total += chunks_info_ptr[j].size;
|
|
}
|
|
|
|
if (total > 0)
|
|
printf(" %s\t%10s\n",
|
|
device_info_ptr[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 *device_info_ptr,
|
|
int device_info_count)
|
|
{
|
|
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);
|
|
|
|
printf("%s,%s: Size:%s, ",
|
|
description,
|
|
r_mode,
|
|
pretty_size_mode(sargs->spaces[i].total_bytes,
|
|
unit_mode));
|
|
printf("Used:%s\n",
|
|
pretty_size_mode(sargs->spaces[i].used_bytes, unit_mode));
|
|
print_chunk_device(flags, info_ptr, info_count,
|
|
device_info_ptr, device_info_count, unit_mode);
|
|
printf("\n");
|
|
}
|
|
|
|
printf("Unallocated:\n");
|
|
print_unused(info_ptr, info_count, device_info_ptr, device_info_count,
|
|
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,
|
|
char *path, unsigned unit_mode, int tabular)
|
|
{
|
|
struct btrfs_ioctl_space_args *sargs;
|
|
int ret = 0;
|
|
|
|
if (!chunkinfo)
|
|
return 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;
|
|
}
|
|
|
|
const char * const cmd_filesystem_usage_usage[] = {
|
|
"btrfs filesystem usage [options] <path> [<path>..]",
|
|
"Show detailed information about internal filesystem usage .",
|
|
HELPINFO_UNITS_SHORT_LONG,
|
|
"-T show data in tabular format",
|
|
NULL
|
|
};
|
|
|
|
int cmd_filesystem_usage(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);
|
|
|
|
while (1) {
|
|
int c;
|
|
|
|
c = getopt(argc, argv, "T");
|
|
if (c < 0)
|
|
break;
|
|
|
|
switch (c) {
|
|
case 'T':
|
|
tabular = 1;
|
|
break;
|
|
default:
|
|
usage(cmd_filesystem_usage_usage);
|
|
}
|
|
}
|
|
|
|
if (check_argc_min(argc - optind, 1))
|
|
usage(cmd_filesystem_usage_usage);
|
|
|
|
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)
|
|
printf("\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;
|
|
printf("\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;
|
|
}
|
|
|
|
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;
|
|
|
|
if (chunks_info_ptr[i].devid != devinfo->devid)
|
|
continue;
|
|
|
|
flags = chunks_info_ptr[i].type;
|
|
|
|
description = btrfs_group_type_str(flags);
|
|
r_mode = btrfs_group_profile_str(flags);
|
|
size = calc_chunk_size(chunks_info_ptr+i);
|
|
printf(" %s,%s:%*s%10s\n",
|
|
description,
|
|
r_mode,
|
|
(int)(20 - strlen(description) - strlen(r_mode)), "",
|
|
pretty_size_mode(size, unit_mode));
|
|
|
|
allocated += size;
|
|
|
|
}
|
|
printf(" 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)
|
|
{
|
|
printf(" Device size: %*s%10s\n",
|
|
(int)(20 - strlen("Device size")), "",
|
|
pretty_size_mode(devinfo->device_size, unit_mode));
|
|
printf(" Device slack: %*s%10s\n",
|
|
(int)(20 - strlen("Device slack")), "",
|
|
pretty_size_mode(devinfo->device_size > 0 ?
|
|
devinfo->device_size - devinfo->size : 0,
|
|
unit_mode));
|
|
}
|