Viewing: lfs.c
// SPDX-License-Identifier: GPL-2.0-only
/*
* Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
* Use is subject to license terms.
*
* Copyright (c) 2011, 2017, Intel Corporation.
*/
/*
* This file is part of Lustre, http://www.lustre.org/
*
* lustre/utils/lfs.c
*
* Author: Peter J. Braam <braam@clusterfs.com>
* Author: Phil Schwan <phil@clusterfs.com>
* Author: Robert Read <rread@clusterfs.com>
*/
/* for O_DIRECTORY */
#ifndef _GNU_SOURCE
#define _GNU_SOURCE
#endif
#include <asm/byteorder.h>
#include <ctype.h>
#include <dirent.h>
#include <err.h>
#include <errno.h>
#include <fcntl.h>
#include <getopt.h>
#include <grp.h>
#include <inttypes.h>
#include <libgen.h>
#include <math.h>
#include <mntent.h>
#include <pwd.h>
#include <regex.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/ioctl.h>
#include <sys/mman.h>
#include <sys/param.h>
#include <sys/quota.h>
#include <sys/stat.h>
#include <sys/time.h>
#include <sys/types.h>
#include <sys/xattr.h>
#include <time.h>
#include <unistd.h>
#include <zlib.h>
#include <libcfs/util/ioctl.h>
#include <libcfs/util/param.h>
#include <libcfs/util/parser.h>
#include <libcfs/util/string.h>
#include <linux/lnet/nidstr.h>
#include <linux/lustre/lustre_idl.h>
#include <linux/lustre/lustre_param.h>
#include <linux/lustre/lustre_ver.h>
#include <lnetconfig/cyaml.h>
#include <lustre/lustreapi.h>
#include "callvpe.h"
#include "lfs_project.h"
#include "lstddef.h"
#include "lustreapi_internal.h"
/* all functions */
static int lfs_find(int argc, char **argv);
static int lfs_getstripe(int argc, char **argv);
static int lfs_getdirstripe(int argc, char **argv);
static int lfs_setdirstripe(int argc, char **argv);
static int lfs_rmentry(int argc, char **argv);
static int lfs_unlink_foreign(int argc, char **argv);
static int lfs_osts(int argc, char **argv);
static int lfs_mdts(int argc, char **argv);
static int lfs_df(int argc, char **argv);
static int lfs_getname(int argc, char **argv);
static int lfs_check(int argc, char **argv);
#ifdef HAVE_SYS_QUOTA_H
struct quota_param {
char *qp_delim;
__u32 qp_valid;
unsigned int qp_verbose:1;
unsigned int qp_quiet:1;
unsigned int qp_human_readable:1;
unsigned int qp_show_default:1;
unsigned int qp_show_pools:1;
unsigned int qp_show_qid:1;
unsigned int qp_show_title:1;
unsigned int qp_show_qid_num:1;
unsigned int qp_show_root:1;
__u32 qp_detail;
};
static int lfs_setquota(int argc, char **argv);
static int lfs_quota(int argc, char **argv);
static int lfs_project(int argc, char **argv);
#endif
static int lfs_flushctx(int argc, char **argv);
static int lfs_poollist(int argc, char **argv);
static int lfs_changelog(int argc, char **argv);
static int lfs_changelog_clear(int argc, char **argv);
static int lfs_fid2path(int argc, char **argv);
static int lfs_path2fid(int argc, char **argv);
static int lfs_rmfid(int argc, char **argv);
static int lfs_data_version(int argc, char **argv);
static int lfs_hsm(int argc, char **argv);
static int lfs_hsm_state(int argc, char **argv);
static int lfs_hsm_set(int argc, char **argv);
static int lfs_hsm_clear(int argc, char **argv);
static int lfs_hsm_action(int argc, char **argv);
static int lfs_hsm_archive(int argc, char **argv);
static int lfs_hsm_restore(int argc, char **argv);
static int lfs_hsm_release(int argc, char **argv);
static int lfs_hsm_remove(int argc, char **argv);
static int lfs_hsm_cancel(int argc, char **argv);
static int lfs_swap_layouts(int argc, char **argv);
static int lfs_mv(int argc, char **argv);
static int lfs_ladvise(int argc, char **argv);
static int lfs_getsom(int argc, char **argv);
static int lfs_somsync(int argc, char **argv);
static int lfs_heat_get(int argc, char **argv);
static int lfs_heat_set(int argc, char **argv);
static int lfs_mirror(int argc, char **argv);
static inline int lfs_mirror_resync(int argc, char **argv);
static inline int lfs_mirror_verify(int argc, char **argv);
static inline int lfs_mirror_read(int argc, char **argv);
static inline int lfs_mirror_write(int argc, char **argv);
static inline int lfs_mirror_copy(int argc, char **argv);
static inline int lfs_mirror_verify_file(const char *fname, __u16 *mirror_ids,
int ids_nr, int verbose, int stale);
static int lfs_pcc_attach(int argc, char **argv);
static int lfs_pcc_attach_fid(int argc, char **argv);
static int lfs_pcc_detach(int argc, char **argv);
static int lfs_pcc_detach_fid(int argc, char **argv);
static int lfs_pcc_pin(int argc, char **argv);
static int lfs_pcc_unpin(int argc, char **argv);
static int lfs_pcc_state(int argc, char **argv);
static int lfs_pcc_delete(int argc, char **argv);
static int lfs_pcc(int argc, char **argv);
static int lfs_migrate_to_dom(int fd_src, int fd_dst, char *name,
enum llapi_migration_flags migration_flags,
unsigned long long bandwidth_bytes_sec,
long stats_interval_sec);
struct pool_to_id_cbdata {
const char *pool;
__u32 id;
};
static int find_comp_id_by_pool(struct llapi_layout *layout, void *cbdata);
static int find_mirror_id_by_pool(struct llapi_layout *layout, void *cbdata);
enum setstripe_origin {
SO_SETSTRIPE,
SO_MIGRATE,
SO_MIGRATE_MDT,
SO_MIRROR_CREATE,
SO_MIRROR_EXTEND,
SO_MIRROR_SPLIT,
SO_MIRROR_DELETE,
};
static int lfs_setstripe_internal(int argc, char **argv,
enum setstripe_origin opc);
static inline int lfs_setstripe(int argc, char **argv)
{
return lfs_setstripe_internal(argc, argv, SO_SETSTRIPE);
}
static inline int lfs_setstripe_migrate(int argc, char **argv)
{
return lfs_setstripe_internal(argc, argv, SO_MIGRATE);
}
static inline int lfs_mirror_create(int argc, char **argv)
{
return lfs_setstripe_internal(argc, argv, SO_MIRROR_CREATE);
}
static inline int lfs_mirror_extend(int argc, char **argv)
{
return lfs_setstripe_internal(argc, argv, SO_MIRROR_EXTEND);
}
static inline int lfs_mirror_split(int argc, char **argv)
{
return lfs_setstripe_internal(argc, argv, SO_MIRROR_SPLIT);
}
static inline int lfs_mirror_delete(int argc, char **argv)
{
return lfs_setstripe_internal(argc, argv, SO_MIRROR_DELETE);
}
#define SSM_SETSTRIPE_OPT \
"[--component-add|--component-del|--delete|-d]\n" \
"\t\t[--comp-set --comp-id|-I COMP_ID|--comp-flags=COMP_FLAGS]\n" \
"\t\t[--component-end|-E END_OFFSET]\n" \
"\t\t[--copy=SOURCE_LAYOUT_FILE]|--yaml|-y YAML_TEMPLATE_FILE]\n" \
"\t\t[--ec|--ec-stripe-count DATA+PARITY]\n" \
"\t\t[--ec-expert DATA+PARITY]\n" \
"\t\t[--extension-size|--ext-size|-z EXT_SIZE]\n" \
"\t\t[--help|-h]\n" \
"\t\t[--foreign=FOREIGN_TYPE --xattr|-x LAYOUT]\n" \
"\t\t[--layout|-L PATTERN] [--mode FILE_MODE]\n" \
"\t\t[--mirror-count|-N[MIRROR_COUNT]]\n" \
"\t\t[--ost|-o OST_INDEX[,OST_INDEX,...]]\n" \
"\t\t[--overstripe-count|-C STRIPE_COUNT]\n" \
"\t\t[--pool|-p POOL_NAME]\n" \
"\t\t[--stripe-count|-c STRIPE_COUNT]\n" \
"\t\t[--stripe-index|-i START_OST_IDX]\n" \
"\t\t[--stripe-size|-S STRIPE_SIZE]"
/* Setstripe and migrate share mostly the same parameters */
#define SSM_CMD_COMMON(cmd) \
"Usage: " cmd \
" " SSM_SETSTRIPE_OPT "\n"
#define SETSTRIPE_USAGE \
SSM_CMD_COMMON("setstripe ") \
"\t\tFILENAME|DIRECTORY\n"
#define MIGRATE_USAGE \
SSM_CMD_COMMON("migrate ") \
"\t\t[--bandwidth|-W BANDWIDTH_MB[MG]]\n" \
"\t\t[--auto|--auto-stripe|-A]\n" \
"\t\t[--block|-b] [--non-block|-n]\n" \
"\t\t[--lustre-dir=LUSTRE_MOUNT_POINT --fid]\n" \
"\t\t[--non-direct|-D] [--verbose|-v] FILENAME\n" \
"\t\t[--restripe] [--stats-interval SECONDS]\n" \
"\t\t-0|--null|--files-from=LIST_FILE|FILENAME ...\n"
#define SETDIRSTRIPE_USAGE \
"\t\t[--mdt-count|-c stripe_count>\n" \
"\t\t[--help|-h] [--mdt-hash|-H mdt_hash]\n" \
"\t\t[--mdt-index|-i mdt_index[,mdt_index,...]\n" \
"\t\t[--mdt-overcount|-C stripe_count>\n" \
"\t\t[--default|-D] [--mode|-o mode]\n" \
"\t\t[--max-inherit|-X max_inherit]\n" \
"\t\t[--max-inherit-rr max_inherit_rr] <dir>\n" \
"To create dir with a foreign (free format) layout :\n" \
"setdirstripe|mkdir --foreign[=FOREIGN_TYPE] -x|-xattr STRING " \
" [--mode|-o MODE] [--flags HEX] DIRECTORY\n"
/*
* LFS_SUBCMD() - Parse and execute lfs subcommands.
* @argc: The count of lfs subcommand line arguments.
* @argv: Array of strings for lfs subcommand line arguments.
*
* This function parses lfs subcommands and performs the
* corresponding functions specified in name##_cmdlist[].
*
* Return: 0 on success or an error code on failure.
*/
#define LFS_SUBCMD(name) \
static int lfs_##name(int argc, char **argv) \
{ \
char cmd[PATH_MAX]; \
int rc = 0; \
\
setlinebuf(stdout); \
\
snprintf(cmd, sizeof(cmd), "%s %s", \
program_invocation_short_name, argv[0]); \
program_invocation_short_name = cmd; \
rc = cfs_parser(argc, argv, name##_cmdlist); \
\
return rc < 0 ? -rc : rc; \
}
/*
* command_t mirror_cmdlist - lfs mirror commands.
*/
command_t mirror_cmdlist[] = {
{ .pc_name = "create", .pc_func = lfs_mirror_create,
.pc_help = "Create a mirrored file.\n"
"usage: lfs mirror create --mirror-count|-N[MIRROR_COUNT]\n"
" [SETSTRIPE_OPTIONS] ... FILENAME|DIRECTORY ...\n" },
{ .pc_name = "delete", .pc_func = lfs_mirror_delete,
.pc_help = "Delete a mirror from a file.\n"
"usage: lfs mirror delete {--mirror-id <mirror_id> |\n"
"\t --component-id|--comp-id|-I COMP_ID |\n"
"\t -p <pool>} MIRRORED_FILE ...\n"
},
{ .pc_name = "extend", .pc_func = lfs_mirror_extend,
.pc_help = "Extend a mirrored file.\n"
"Usage: lfs mirror extend [--mirror-count|-N[MIRROR_COUNT]]\n"
"\t\t[--no-verify] [--stats|--stats-interval=STATS_INTERVAL]\n"
"\t\t[--bandwidth|-W BANDWIDTH_MB[MG]]\n"
"\t\t[--file|-f VICTIM_FILE]\n"
"\t\t" SSM_SETSTRIPE_OPT "]\n"
"\t\t[--fid [--lustre-dir=LUSTRE_MOUNT_POINT]]\n"
"\t\t-0|--null|--files-from=LIST_FILE|FILENAME ...\n" },
{ .pc_name = "split", .pc_func = lfs_mirror_split,
.pc_help = "Split a mirrored file.\n"
"usage: lfs mirror split {--mirror-id MIRROR_ID |\n"
"\t --component-id|-I COMP_ID|-p POOL} [--destroy|-d]\n"
"\t [-f NEW_FILE] MIRRORED_FILE ...\n" },
{ .pc_name = "read", .pc_func = lfs_mirror_read,
.pc_help = "Read the content of a specified mirror of a file.\n"
"usage: lfs mirror read --mirror-id|-N MIRROR_ID\n"
"\t\t[--outfile|-o OUTPUT_FILE] MIRRORED_FILE\n" },
{ .pc_name = "write", .pc_func = lfs_mirror_write,
.pc_help = "Write to a specified mirror of a file.\n"
"usage: lfs mirror write {--mirror-id|-N MIRROR_ID}\n"
"\t\t[--inputfile|-i INPUT_FILE] MIRRORED_FILE\n" },
{ .pc_name = "copy", .pc_func = lfs_mirror_copy,
.pc_help = "Copy a specified mirror to other mirror(s) of a file.\n"
"usage: lfs mirror copy --read-mirror|-i SRC_MIRROR_ID0\n"
"\t\t--write-mirror|-o TGT_MIRROR_ID1[,...] MIRROR_FILE\n" },
{ .pc_name = "resync", .pc_func = lfs_mirror_resync,
.pc_help = "Resynchronizes out-of-sync mirrored file(s).\n"
"usage: lfs mirror resync [--only MIRROR_ID[,...]]|\n"
"\t\t[--stats|--stats-interval=SECONDS]\n"
"\t\t[--bandwidth|-W BANDWIDTH_MB[MG]]\n"
"\t\t[--force-resync|-f]\n"
"\t\tMIRRORED_FILE [MIRRORED_FILE2...]\n" },
{ .pc_name = "verify", .pc_func = lfs_mirror_verify,
.pc_help = "Verify mirrored file(s).\n"
"usage: lfs mirror verify [--only MIRROR_ID[,...]]\n"
"\t\t[--stale|-s] [--verbose|-v]\n"
"\t\tMIRRORED_FILE [MIRRORED_FILE2 ...]\n" },
{ .pc_help = NULL }
};
LFS_SUBCMD(mirror);
/*
* command_t pcc_cmdlist - lfs pcc commands.
*/
command_t pcc_cmdlist[] = {
{ .pc_name = "attach", .pc_func = lfs_pcc_attach,
.pc_help = "Attach given files to the Persistent Client Cache.\n"
"usage: lfs pcc attach [--id|-i ID] [--readonly|-r] [--write|-w] FILE [...]\n"
"\t-i: archive ID for PCC\n"
"\t-r: readonly attach\n"
"\t-w: writeable attach\n" },
{ .pc_name = "attach_fid", .pc_func = lfs_pcc_attach_fid,
.pc_help = "Attach given files into PCC by FID(s).\n"
"usage: lfs pcc attach_id [--id|-i ID] {--mnt|-m MOUNTPOINT} [--readonly|-r] [--write|-w] FID [...]\n"
"\t-i: archive ID for PCC\n"
"\t-m: Lustre mount point\n"
"\t-r: readonly attach\n"
"\t-w: writeable attach\n" },
{ .pc_name = "state", .pc_func = lfs_pcc_state,
.pc_help = "Display the PCC state for given files.\n"
"usage: lfs pcc state FILE [...]\n" },
{ .pc_name = "status", .pc_func = lfs_pcc_state,
.pc_help = "Display the PCC state for given files.\n"
"usage: lfs pcc status FILE [...]\n" },
{ .pc_name = "detach", .pc_func = lfs_pcc_detach,
.pc_help = "Detach given files from the Persistent Client Cache.\n"
"usage: lfs pcc detach [-k] FILE [...]\n" },
{ .pc_name = "detach_fid", .pc_func = lfs_pcc_detach_fid,
.pc_help = "Detach given files from PCC by FID(s).\n"
"usage: lfs pcc detach_fid {--mnt|-m MOUNTPATH} FID...\n" },
{ .pc_name = "delete", .pc_func = lfs_pcc_delete,
.pc_help = "Delete the PCC layout component for given files.\n"
"usage: lfs pcc delete <FILE> ...\n" },
{ .pc_name = "pin", .pc_func = lfs_pcc_pin,
.pc_help = "Pin files to prevent them from being removed from PCC.\n"
"usage: lfs pcc pin [--id|-i ID] FILE ...\n"
"\t-i: archive ID for PCC\n"},
{ .pc_name = "unpin", .pc_func = lfs_pcc_unpin,
.pc_help = "Un-pin files so that they can be removed from PCC.\n"
"usage: lfs pcc unpin [--id|-i ID] FILE ...\n"},
{ .pc_help = NULL }
};
LFS_SUBCMD(pcc);
/*
* command_t hsm_cmdlist - lfs hsm commands.
*/
command_t hsm_cmdlist[] = {
{.pc_name = "state", .pc_func = lfs_hsm_state,
.pc_help = "Display the HSM information for given files.\n"
"usage: hsm state FILE"},
{.pc_name = "set", .pc_func = lfs_hsm_set,
.pc_help = "Set HSM user flag on specified files.\n"
"usage: hsm set [--norelease] [--noarchive] [--dirty] [--exists] "
"[--archived] [--lost] [--archive-id NUM] FILE"},
{.pc_name = "clear", .pc_func = lfs_hsm_clear,
.pc_help = "Clear HSM user flag on specified files.\n"
"usage: hsm clear [--norelease] [--noarchive] [--dirty] [--exists] "
"[--archived] [--lost] FILE"},
{.pc_name = "action", .pc_func = lfs_hsm_action,
.pc_help = "Display current HSM request for given files.\n"
"usage: hsm action FILE"},
{.pc_name = "archive", .pc_func = lfs_hsm_archive,
.pc_help = "Archive file to external storage.\n"
"usage: hsm archive [--filelist FILELIST] [--data DATA]\n"
" [--archive NUM] FILE"},
{.pc_name = "restore", .pc_func = lfs_hsm_restore,
.pc_help = "Restore file from external storage.\n"
"usage: hsm restore [--filelist FILELIST] [--data DATA] FILE"},
{.pc_name = "release", .pc_func = lfs_hsm_release,
.pc_help = "Release files from Lustre.\n"
"usage: hsm release [--filelist FILELIST] [--data DATA] FILE"},
{.pc_name = "remove", .pc_func = lfs_hsm_remove,
.pc_help = "Remove file copy from external storage.\n"
"usage: hsm remove [--filelist FILELIST] [--data DATA]\n"
" [--archive NUM]\n"
" {FILE | --mntpath MOUNTPATH FID}\n\n"
"Note: To remove an archived copy of a file already deleted from a\n"
"Lustre FS, --mntpath option and a list of FIDs must be specified."},
{.pc_name = "cancel", .pc_func = lfs_hsm_cancel,
.pc_help = "Cancel requests related to specified files.\n"
"usage: hsm cancel [--filelist FILELIST] [--data DATA] FILE"},
{.pc_help = NULL}
};
LFS_SUBCMD(hsm);
/* all available commands */
command_t cmdlist[] = {
{"setstripe", lfs_setstripe, 0,
"Create a file with specified striping/composite layout, or\n"
"set the default layout on an existing directory.\n"
SETSTRIPE_USAGE},
{"getstripe", lfs_getstripe, 0,
"List the layout pattern for a given file or files in a\n"
"directory or recursively for all files in a directory tree.\n"
"Usage: getstripe [--ost|-O OST_NAME] [--quiet|-q] [--verbose|-v]\n"
" [--stripe-count|-c] [--stripe-index|-i] [--fid|-F]\n"
" [--pool|-p] [--stripe-size|-S] [--directory|-d]\n"
" [--mdt-index|-m] [--recursive|-r] [--raw|-R]\n"
" [--layout|-L] [--generation|-g] [--yaml|-y]\n"
" [--help|-h] [--hex-idx]\n"
" [--component-id|-I[=COMP_ID]]\n"
" [--component-flags[=COMP_FLAGS]]\n"
" [--component-count]\n"
" [--extension-size|--ext-size|-z]\n"
" [--component-start[=[+-]START_OFFSET]]\n"
" [--component-end|-E[[+-]END_OFFSET]]\n"
" [[!] --mirror-index=[+-]MIRROR_INDEX |\n"
" [!] --mirror-id=[+-]MIRROR_ID] [--mirror-count|-N]\n"
" [--no-follow]\n"
" FILENAME|DIRECTORY"},
{"setdirstripe", lfs_setdirstripe, 0,
"Create striped directory on specified MDT, same as mkdir.\n"
"May be restricted to root or group users, depending on settings.\n"
"usage: setdirstripe [OPTION] <directory>\n"
SETDIRSTRIPE_USAGE},
{"getdirstripe", lfs_getdirstripe, 0,
"To list the layout pattern info for a given directory\n"
"or recursively for all directories in a directory tree.\n"
"usage: getdirstripe [--mdt-count|-c] [--default|-D] [--fid|-F]\n"
" [--help|-h] [--hex-idx] [--mdt-hash|-H]\n"
" [--mdt-index|-m|-i] [--obd|-O UUID]\n"
" [--recursive|-r] [--raw|-R]\n"
" [--verbose|-v] [--max-inherit|-X]\n"
" [--max-inherit-rr] [--yaml|-y] <dir> ...\n"},
{"mkdir", lfs_setdirstripe, 0,
"Create striped directory on specified MDT, same as setdirstripe.\n"
"usage: mkdir [OPTION] <directory>\n"
SETDIRSTRIPE_USAGE},
{"rm_entry", lfs_rmentry, 0,
"To remove the name entry of the remote directory. Note: This\n"
"command will only delete the name entry, i.e. the remote directory\n"
"will become inaccessable after this command. This can only be done\n"
"by the administrator\n"
"usage: rm_entry <dir>\n"},
{"rmentry", lfs_rmentry, 0, "remove a dir entry, same as 'rm_entry'\n"},
{"unlink_foreign", lfs_unlink_foreign, 0,
"To remove the foreign file/dir.\n"
"Note: This is for files/dirs prevented to be removed using\n"
"unlink/rmdir, but works also for regular ones\n"
"usage: unlink_foreign <foreign_dir/file> [<foreign_dir/file> ...]\n"},
{"pool_list", lfs_poollist, 0,
"List pools or pool OSTs\n"
"usage: pool_list <fsname>[.<pool>] | <pathname>\n"},
{"find", lfs_find, 0,
"find files matching given attributes recursively in directory tree.\n"
"usage: find <directory|filename> ...\n"
" [[!] --atime|-A [+-]N[smhdwy]] [[!] --btime|-B [+-]N[smhdwy]]\n"
" [[!] --ctime|-C [+-]N[smhdwy]] [[!] --mtime|-M [+-]N[smhdwy]]\n"
" [[!] --attrs=[^]ATTR[,...]]\n"
" [[!] --blocks|-b N] [[!] --component-count [+-]<comp_cnt>]\n"
" [[!] --component-start [+-]N[kMGTPE]]\n"
" [[!] --component-end|-E [+-]N[kMGTPE]]\n"
" [[!] --component-flags {init,stale,prefer,prefrd,prefwr,offline,nosync,extension}]\n"
" [[!] --extension-size|--ext-size|-z [+-]N[kMGT]]\n"
" [[!] --foreign[=<foreign_type>]]\n"
" [[!] --gid|-g|--group|-G <gid>|<gname>] [--help|-h]\n"
" [[!] --layout|-L released,raid0,mdt] [--lazy|-l] [[!] --links [+-]n]\n"
" [--maxdepth|-D N] [--mindepth|-d N]\n"
" [[!] --mdt-count|-T [+-]<stripes>]\n"
" [[!] --mdt-hash|-H <[^][blm],[^]fnv_1a_64,all_char,crush,...>\n"
" [[!] --mdt-index|--mdt|-m <uuid|index,...>]\n"
" [[!] --mirror-count|-N [+-]<n>]\n"
" [[!] --mirror-state <[^]state>]\n"
" [[!] --name|-n <pattern>] [[!] --newer[XY] <reference>]\n"
" [[!] --ost|-O <uuid|index,...>] [[!] --perm [/-]mode]\n"
" [[!] --pool <pool>] [--print|-P] [--print0|-0] [--printf <format>]\n"
" [[!] --projid <projid>] [[!] --size|-s [+-]N[bkMGTPE]]\n"
" [--skip|-k PERCENT] [[!] --stripe-count|-c [+-]<stripes>]\n"
" [[!] --stripe-index|-i <index,...>]\n"
" [[!] --stripe-size|-S [+-]N[kMGT]] [--threads N]\n"
" [[!] --type|-t <filetype>]\n"
" [[!] --uid|-u|--user|-U <uid>|<uname>]\n"
"\t !: used before an option indicates 'NOT' requested attribute\n"
"\t -: used before a value indicates less than requested value\n"
"\t +: used before a value indicates more than requested value\n"
"\t ^: used before a flag indicates to exclude it\n"},
{"check", lfs_check, 0,
"Display the status of MGTs, MDTs or OSTs (as specified in the command)\n"
"or all the servers (MGTs, MDTs and OSTs) [for specified path only].\n"
"usage: check {mgts|osts|mdts|all} [PATH]"},
{"osts", lfs_osts, 0, "list OSTs connected to client "
"[for specified path only]\n" "usage: osts [PATH]"},
{"mdts", lfs_mdts, 0, "list MDTs connected to client "
"[for specified path only]\n" "usage: mdts [PATH]"},
{"df", lfs_df, 0,
"report filesystem disk space usage or inodes usage "
"of each MDS and all OSDs or a batch belonging to a specific pool.\n"
"Usage: df [--inodes|-i] [--human-readable|-h] [--lazy|-l]\n"
"[--mdt|-m[INDEX]] [--ost|-o[INDEX]] [--output|-u FIELD_LIST] [--no-header|-N]\n"
"[--only-summary|-s] [--pool|-p FSNAME[.POOL]] [PATH]"},
{"getname", lfs_getname, 0,
"list instances and specified mount points [for specified path only]\n"
"Usage: getname [--help|-h] [--instance|-i] [--fsname|-n] [--uuid|-u]\n"
" [PATH ...]"},
#ifdef HAVE_SYS_QUOTA_H
{"setquota", lfs_setquota, 0, "Set filesystem quotas.\n"
"usage: setquota [-t] {-u|-U|-g|-G|-p|-P ID} {-b|-B|-i|-I LIMIT}\n"
" [--lqa LQA_NAME] [--pool POOL] MOUNT_POINT\n"
" setquota {-u|-g|-p ID} {--default|--delete} MOUNT_POINT\n"},
{"quota", lfs_quota, 0, "Display disk usage and limits.\n"
"usage: quota [-q] [-v] [-h] [-o OBD_UUID|-o OST_IDX|-m MDT_IDX]\n"
" [{-u|-g|-p} UNAME|UID|GNAME|GID|PROJID]\n"
" [--blocks|--busage|--space]\n"
" [--block-softlimit|--bsoftlimit]\n"
" [--block-hardlimit|--bhardlimit]\n"
" [--block-grace|--bgrace|--btime]\n"
" [--delimiter DELIMITER]\n"
" [--filesystem|--mount-point]\n"
" [--inodes|--iusage]\n"
" [--inode-softlimit|--isoftlimit]\n"
" [--inode-hardlimit|--ihardlimit]\n"
" [--inode-grace|--igrace|--itime]\n"
" [--lqa LQA_NAME]\n"
" [--pool OST_POOL_NAME]\n"
" [MOUNT_POINT ...]\n"
" quota -t {-u|-g|-p} [--pool OST_POOL_NAME] [MOUNT_POINT ...]\n"
" quota [-hqv] {-U|-G|-P} [--pool OST_POOL_NAME] [MOUNT_POINT ...]\n"
" quota -a {-u|-g|-p} [-n] [-s START_QID] [-e END_QID]\n"
" [--show-root] [MOUNT_POINT ...]\n"},
{"project", lfs_project, 0,
"Change or list project attribute for specified file or directory.\n"
"usage: project [-d|--directory] [-r|--recursive] FILE|DIRECTORY...\n"
" list project ID and flags on file(s) or directories\n"
" project [-p|--proj-inherit|--inherit ID] [-s|--set-inherit]\n"
" [-d|--directory] [-r|--recursive] FILE|DIRECTORY...\n"
" set project ID and/or inherit flag for specified file(s) or directories\n"
" project -c|--check [-p|--proj-inherit|--inherit ID]\n"
" [-0|--print0] [-d|--directory] [-r|--recursive]\n"
" FILE|DIRECTORY...\n"
" check project ID and flags on file(s) or directories, print outliers\n"
" project -C|--clear [-k|--keep] [-d|--directory]\n"
" [-r|--recursive] FILE|DIRECTORY...\n"
" clear the project inherit flag and ID on the file or directory\n"
},
#endif
{"flushctx", lfs_flushctx, 0,
"Flush security context for current user.\n"
"usage: flushctx [-k] [-r] [mountpoint...]"},
{"changelog", lfs_changelog, 0,
"Show the metadata changes on an MDT.\n"
"usage: changelog [--follow] [--user USERNAME] [--mask MASK] "
"<mdtname> [startrec [endrec]]"},
{"changelog_clear", lfs_changelog_clear, 0,
"Indicate that old changelog records up to <endrec> are no longer of "
"interest to consumer <id>, allowing the system to free up space.\n"
"An <endrec> of 0 means all records.\n"
"usage: changelog_clear <mdtname> <id> <endrec>"},
{"fid2path", lfs_fid2path, 0,
"Resolve the full path(s) for given FID(s). For a specific hardlink "
"specify link number <linkno>.\n"
"usage: fid2path [--print0|-0] [--print-fid|-f] [--print-link|-c] "
"[--link|-l <linkno>] [--name|-n] <fsname|root> <fid>..."},
{"path2fid", lfs_path2fid, 0, "Display the fid(s) for a given path(s).\n"
"usage: path2fid [--parents] <path> ..."},
{"rmfid", lfs_rmfid, 0, "Remove file(s) by FID(s)\n"
"usage: rmfid <fsname|rootpath> <fid> ..."},
{"data_version", lfs_data_version, 0, "Display file data version, "
"and optionally set the data version in the HSM xattr for FILENAME.\n"
"usage: data_version {--nosync|--read|--write} [--set-hsm] FILENAME"},
{"hsm_state", lfs_hsm_state, 0,
"Display the HSM information for given files.\n"
"usage: hsm_state FILE"},
{"hsm_set", lfs_hsm_set, 0,
"Set HSM user flag on specified files.\n"
"usage: hsm_set [--norelease] [--noarchive] [--dirty] [--exists]\n"
" [--archived] [--lost] [--archive-id NUM] FILE"},
{"hsm_clear", lfs_hsm_clear, 0,
"Clear HSM user flag on specified files.\n"
"usage: hsm_clear [--norelease] [--noarchive] [--dirty] [--exists]\n"
" [--archived] [--lost] FILE"},
{"hsm_action", lfs_hsm_action, 0,
"Display current HSM request for given files.\n"
"usage: hsm_action FILE"},
{"hsm_archive", lfs_hsm_archive, 0,
"Archive file to external storage.\n"
"usage: hsm_archive [--filelist FILELIST] [--data DATA]\n"
" [--archive NUM] FILE"},
{"hsm_restore", lfs_hsm_restore, 0,
"Restore file from external storage.\n"
"usage: hsm_restore [--filelist FILELIST] [--data DATA] FILE"},
{"hsm_release", lfs_hsm_release, 0,
"Release files from Lustre.\n"
"usage: hsm_release [--filelist FILELIST] [--data DATA] FILE"},
{"hsm_remove", lfs_hsm_remove, 0,
"Remove file copy from external storage.\n"
"usage: hsm_remove [--filelist FILELIST] [--data DATA]\n"
" [--archive NUM]\n"
" {FILE | --mntpath MOUNTPATH FID}\n\n"
"Note: To remove an archived copy of a file already deleted from a "
"Lustre FS, --mntpath option and a list of FIDs must be specified"},
{"hsm_cancel", lfs_hsm_cancel, 0,
"Cancel requests related to specified files.\n"
"usage: hsm_cancel [--filelist FILELIST] [--data DATA] FILE"},
{"hsm", lfs_hsm, hsm_cmdlist, ""},
{"swap_layouts", lfs_swap_layouts, 0, "Swap layouts between 2 files.\n"
"usage: swap_layouts <path1> <path2>"},
{"migrate", lfs_setstripe_migrate, 0,
"migrate directories and their inodes between MDTs.\n"
"usage: migrate [--mdt-count|-c STRIPE_COUNT]\n"
" [--mdt-overcount|-C OVERSTRIPE_COUNT\n"
" [--directory|-d] [--mdt-hash|-H HASH_TYPE]\n"
" [--mdt-index|-m START_MDT_INDEX] [--verbose|-v]\n"
" [--clear-fixed]\n"
" DIRECTORY\n"
"\n"
"migrate file objects from one OST layout to another\n"
"(may be not safe with concurent writes).\n"
MIGRATE_USAGE },
{"mv", lfs_mv, 0,
"To move directories between MDTs. This command is deprecated, "
"use \"migrate\" instead.\n"
"usage: mv <directory|filename> [--mdt-index|-m MDT_INDEX] "
"[--verbose|-v]\n"},
{"ladvise", lfs_ladvise, 0,
"Provide servers with advice about access patterns for a file.\n"
"usage: ladvise [--advice|-a ADVICE] [--start|-s START[kMGT]]\n"
" [--background|-b] [--unset|-u]\n\n"
" {--end|-e END[kMGT]|--length|-l LENGTH[kMGT]}\n"
" {[--mode|-m [READ,WRITE]}\n"
" <file> ...\n"},
{"mirror", lfs_mirror, mirror_cmdlist,
"lfs commands used to manage files with mirrored components:\n"
"lfs mirror create - create a mirrored file or directory\n"
"lfs mirror extend - add mirror(s) to an existing file\n"
"lfs mirror split - split a mirror from an existing mirrored file\n"
"lfs mirror resync - resynchronize out-of-sync mirrored file(s)\n"
"lfs mirror read - read a mirror content of a mirrored file\n"
"lfs mirror write - write to a mirror of a mirrored file\n"
"lfs mirror copy - copy a mirror to other mirror(s) of a file\n"
"lfs mirror verify - verify mirrored file(s)\n"},
{"getsom", lfs_getsom, 0, "To list the SOM info for a given file.\n"
"usage: getsom [-s] [-b] [-f] <path>\n"
"\t-s: Only show the size value of the SOM data for a given file\n"
"\t-b: Only show the blocks value of the SOM data for a given file\n"
"\t-f: Only show the flags value of the SOM data for a given file\n"},
{"somsync", lfs_somsync, 0,
"Synchronize SOM xattr(s) for given file(s) or FID(s).\n"
"usage: somsync FILE ...\n"
" somsync --by-fid MOUNT FID ...\n"},
{"heat_get", lfs_heat_get, 0,
"To get heat of files.\n"
"usage: heat_get <file> ...\n"},
{"heat_set", lfs_heat_set, 0,
"To set heat flags of files.\n"
"usage: heat_set [--clear|-c] [--off|-o] [--on|-O] <file> ...\n"
"\t--clear|-c: Clear file heat for given files\n"
"\t--off|-o: Turn off file heat for given files\n"
"\t--on|-O: Turn on file heat for given files\n"},
{"pcc", lfs_pcc, pcc_cmdlist,
"lfs commands used to interact with PCC features:\n"
"lfs pcc attach - attach given files to Persistent Client Cache\n"
"lfs pcc attach_fid - attach given files into PCC by FID(s)\n"
"lfs pcc state - display the PCC state for given files\n"
"lfs pcc detach - detach given files from Persistent Client Cache\n"
"lfs pcc detach_fid - detach given files from PCC by FID(s)\n"
"lfs pcc delete - delete the PCC layout componenet for given files\n"
"lfs pcc pin - pin give files for PCC\n"
"lfs pcc unpin - unpin given files for PCC\n"},
{ 0, 0, 0, NULL }
};
/*
* Convert input string (dir hash) to its number value
*
* Return 0 on unrecognized hash (failure) or positive value on success
*/
static enum lmv_hash_type check_hashtype(const char *hashtype)
{
int type_num;
int i;
if (!hashtype)
goto out;
type_num = atoi(hashtype);
/* numeric hash type */
if (lmv_is_known_hash_type(type_num))
return type_num;
/* string hash type */
for (i = LMV_HASH_TYPE_ALL_CHARS; i < ARRAY_SIZE(mdt_hash_name); i++)
if (strcmp(hashtype, mdt_hash_name[i]) == 0)
return i;
out:
return LMV_HASH_TYPE_UNKNOWN;
}
static uint32_t check_foreign_type_name(const char *foreign_type_name)
{
uint32_t i;
for (i = 0; i < LU_FOREIGN_TYPE_UNKNOWN; i++) {
if (!lu_foreign_types[i].lft_name)
break;
if (strcmp(foreign_type_name,
lu_foreign_types[i].lft_name) == 0)
return lu_foreign_types[i].lft_type;
}
return LU_FOREIGN_TYPE_UNKNOWN;
}
static int
migrate_open_files(const char *name, enum llapi_migration_flags migration_flags,
const struct llapi_stripe_param *param,
struct llapi_layout *layout, int *fd_src_ptr,
int *fd_dst_ptr, char **err_str)
{
int fd_src = -1;
int fd_dst = -1;
int rflags;
int mdt_index;
int random_value;
char parent[PATH_MAX];
char volatile_file[PATH_MAX];
char *ptr;
int rc;
struct stat st;
struct stat stv;
/*
* Ensure param and layout are mutually exclusive.
* If both are set, it indicates a programming error in the caller.
*/
if (param && layout) {
*err_str = "both param and layout specified";
return -EINVAL;
}
/* search for file directory pathname */
if (strlen(name) > sizeof(parent) - 1) {
*err_str = "source file name";
return -ERANGE;
}
strncpy(parent, name, sizeof(parent));
ptr = strrchr(parent, '/');
if (!ptr) {
if (!getcwd(parent, sizeof(parent))) {
*err_str = "getcwd";
return -errno;
}
} else {
if (ptr == parent) /* leading '/' */
ptr = parent + 1;
*ptr = '\0';
}
/* even if the file is only read, WR mode is nedeed to allow
* layout swap on fd
*/
/* Allow migrating even without the key on encrypted files */
rflags = O_RDWR | O_NOATIME | O_CIPHERTEXT;
if (!(migration_flags & LLAPI_MIGRATION_NONDIRECT))
rflags |= O_DIRECT;
source_open:
fd_src = open(name, rflags);
if (fd_src < 0) {
/* If encrypted file without the key,
* retry mirror extend in O_DIRECT.
*/
if (errno == ENOKEY && !(rflags & O_DIRECT) &&
migration_flags & LLAPI_MIGRATION_MIRROR) {
rflags |= O_DIRECT;
goto source_open;
}
rc = -errno;
*err_str = "cannot open source file";
return rc;
}
rc = llapi_file_fget_mdtidx(fd_src, &mdt_index);
if (rc < 0) {
*err_str = "cannot get MDT index";
goto out;
}
do {
int open_flags = O_WRONLY | O_CREAT | O_EXCL | O_NOFOLLOW |
/* Allow migrating without the key on encrypted files */
O_CIPHERTEXT;
mode_t open_mode = S_IRUSR | S_IWUSR;
if (rflags & O_DIRECT)
open_flags |= O_DIRECT;
random_value = random();
rc = snprintf(volatile_file, sizeof(volatile_file),
"%s/%s:%.4X:%.4X:fd=%.2d", parent,
LUSTRE_VOLATILE_HDR, mdt_index,
random_value, fd_src);
if (rc >= sizeof(volatile_file)) {
rc = -ENAMETOOLONG;
break;
}
/* create, open a volatile file, use caching (ie no directio) */
if (layout) {
/* Returns -1 and sets errno on error: */
fd_dst = llapi_layout_file_open(volatile_file,
open_flags, open_mode,
layout);
if (fd_dst < 0)
fd_dst = -errno;
} else if (param) {
/* Does the right thing on error: */
fd_dst = llapi_file_open_param(volatile_file,
open_flags,
open_mode, param);
} else {
/* Neither layout nor param specified - use default
* striping from parent directory
*/
fd_dst = open(volatile_file, open_flags, open_mode);
if (fd_dst < 0)
fd_dst = -errno;
}
} while (fd_dst < 0 && (rc = fd_dst) == -EEXIST);
if (rc < 0) {
*err_str = "cannot create volatile file";
goto out;
}
/*
* Not-owner (root?) special case.
* Need to set owner/group of volatile file like original.
* This will allow to pass related check during layout_swap.
*/
rc = fstat(fd_src, &st);
if (rc != 0) {
rc = -errno;
*err_str = "cannot stat source file";
goto out;
}
rc = fstat(fd_dst, &stv);
if (rc != 0) {
rc = -errno;
*err_str = "cannot stat volatile";
goto out;
}
if (st.st_uid != stv.st_uid || st.st_gid != stv.st_gid) {
rc = fchown(fd_dst, st.st_uid, st.st_gid);
if (rc != 0) {
rc = -errno;
*err_str = "cannot change ownwership of volatile";
goto out;
}
}
out:
if (rc < 0) {
if (fd_src > 0)
close(fd_src);
if (fd_dst > 0)
close(fd_dst);
} else {
*fd_src_ptr = fd_src;
*fd_dst_ptr = fd_dst;
}
return rc;
}
static int migrate_copy_data(int fd_src, int fd_dst, int (*check_file)(int),
unsigned long long bandwidth_bytes_sec,
long stats_interval_sec, off_t file_size_bytes)
{
struct llapi_layout *layout;
size_t buf_size = DEFAULT_IO_BUFLEN;
uint64_t stripe_size = ONE_MB;
void *buf = NULL;
off_t pos = 0;
off_t data_end = 0;
ssize_t page_size;
bool sparse;
int rc;
size_t write_bytes = 0;
ssize_t read_bytes = 0;
struct timespec start_time;
struct timespec now;
struct timespec last_bw_print;
layout = llapi_layout_get_by_fd(fd_src, 0);
if (layout) {
rc = llapi_layout_stripe_size_get(layout, &stripe_size);
if (rc == 0) {
/* We like big bufs */
if (stripe_size > buf_size)
buf_size = stripe_size;
else
/* Trim to stripe_size multiple */
buf_size -= buf_size % stripe_size;
}
llapi_layout_free(layout);
}
/* limit transfer size to what can be sent in one second */
if (bandwidth_bytes_sec && bandwidth_bytes_sec < buf_size)
buf_size = (bandwidth_bytes_sec + stripe_size - 1) &
~(stripe_size - 1);
page_size = sysconf(_SC_PAGESIZE);
if (page_size < 0) {
rc = -errno;
return rc;
}
/* Use a page-aligned buffer for direct I/O */
rc = posix_memalign(&buf, page_size, buf_size);
if (rc != 0)
return -rc;
(void)mlock(buf, buf_size);
sparse = llapi_file_is_sparse(fd_src);
if (sparse) {
rc = ftruncate(fd_dst, pos);
if (rc < 0) {
rc = -errno;
llapi_error(LLAPI_MSG_ERROR, rc,
"fail to ftruncate dst file to %ld", pos);
goto out_free;
}
}
clock_gettime(CLOCK_MONOTONIC, &start_time);
now = last_bw_print = start_time;
while (1) {
off_t data_off;
size_t to_read, to_write;
ssize_t rsize;
if (sparse && pos >= data_end) {
size_t data_size;
data_off = llapi_data_seek(fd_src, pos, &data_size);
if (data_off < 0) {
/* Non-fatal, switch to full copy */
sparse = false;
continue;
}
/* hole at the end of file, truncate up to it */
if (!data_size) {
rc = ftruncate(fd_dst, data_off);
if (rc < 0) {
llapi_error(LLAPI_MSG_ERROR, rc,
"fail to ftruncate dst file to %ld",
data_off);
goto out;
}
}
pos = data_off & ~(page_size - 1);
data_end = data_off + data_size;
to_read = ((data_end - pos - 1) | (page_size - 1)) + 1;
to_read = MIN(to_read, buf_size);
} else {
to_read = buf_size;
}
if (check_file) {
rc = check_file(fd_src);
if (rc < 0) {
llapi_error(LLAPI_MSG_ERROR, rc,
"error checking src file");
goto out;
}
}
rsize = pread(fd_src, buf, to_read, pos);
read_bytes += rsize;
if (rsize < 0) {
rc = -errno;
llapi_error(LLAPI_MSG_ERROR, rc,
"error reading src bytes %ld-%ld",
pos, to_read);
goto out;
}
/* EOF */
if (rsize == 0)
break;
to_write = rsize;
while (to_write > 0) {
ssize_t written;
written = pwrite(fd_dst, buf, to_write, pos);
if (written < 0) {
rc = -errno;
llapi_error(LLAPI_MSG_ERROR, rc,
"error writing dst bytes %ld-%ld",
pos, to_write);
goto out;
}
pos += written;
to_write -= written;
write_bytes += written;
if (bandwidth_bytes_sec == 0)
continue;
clock_gettime(CLOCK_MONOTONIC, &now);
llapi_bandwidth_throttle(&now, &start_time,
bandwidth_bytes_sec,
write_bytes);
}
if (stats_interval_sec && write_bytes != file_size_bytes) {
clock_gettime(CLOCK_MONOTONIC, &now);
llapi_stats_log(&now, &start_time, &last_bw_print,
stats_interval_sec, read_bytes,
write_bytes, pos, file_size_bytes);
}
if (rc || rsize < to_read)
break;
}
/* Output at least one log, regardless of stats_interval */
if (stats_interval_sec) {
clock_gettime(CLOCK_MONOTONIC, &now);
llapi_stats_log(&now, &start_time, &last_bw_print,
stats_interval_sec, read_bytes, write_bytes,
file_size_bytes, file_size_bytes);
}
rc = fsync(fd_dst);
if (rc < 0) {
rc = -errno;
llapi_error(LLAPI_MSG_ERROR, rc,
"failed to fsync dst file");
}
out:
/* Try to avoid page cache pollution after migration. */
(void)posix_fadvise(fd_src, 0, 0, POSIX_FADV_DONTNEED);
(void)posix_fadvise(fd_dst, 0, 0, POSIX_FADV_DONTNEED);
out_free:
(void)munlock(buf, buf_size);
free(buf);
return rc;
}
static int migrate_set_timestamps(int fd, const struct stat *st)
{
struct timeval tv[2] = {
{.tv_sec = st->st_atime},
{.tv_sec = st->st_mtime}
};
return futimes(fd, tv);
}
static int migrate_block(int fd_src, int fd_dst,
unsigned long long bandwidth_bytes_sec,
long stats_interval_sec, char **err_str)
{
struct stat st;
__u64 dv1;
__u64 dv2;
int gid;
int rc;
int rc2;
do
gid = random();
while (gid == 0);
/* The grouplock blocks all concurrent accesses to the file. */
rc = llapi_group_lock(fd_src, gid);
if (rc < 0) {
*err_str = "cannot get group lock";
return rc;
}
rc = fstat(fd_src, &st);
if (rc < 0) {
*err_str = "cannot stat source file";
rc = -errno;
goto out_unlock;
}
/*
* LL_DV_RD_FLUSH should not be set, otherwise the servers will try to
* get extent locks on the OST objects. This will conflict with our
* extent group locks.
*/
rc = llapi_get_data_version(fd_src, &dv1, 0);
if (rc < 0) {
*err_str = "cannot get initial dataversion";
goto out_unlock;
}
rc = migrate_copy_data(fd_src, fd_dst, NULL, bandwidth_bytes_sec,
stats_interval_sec, st.st_size);
if (rc < 0) {
*err_str = "data copy failed";
goto out_unlock;
}
/* Make sure we keep original atime/mtime values */
rc = migrate_set_timestamps(fd_dst, &st);
if (rc < 0) {
*err_str = "set target file timestamp failed";
goto out_unlock;
}
rc = llapi_get_data_version(fd_dst, &dv2, LL_DV_RD_FLUSH);
if (rc < 0) {
*err_str = "cannot get dataversion after copy";
goto out_unlock;
}
/*
* swap layouts
* for a migration we need to check data version on file did
* not change.
*
* Pass in gid=0 since we already own grouplock.
*/
rc = llapi_fswap_layouts_grouplock(fd_src, fd_dst, dv1, dv2, 0,
SWAP_LAYOUTS_CHECK_DV1);
if (rc == -EAGAIN) {
*err_str = "file changed";
goto out_unlock;
} else if (rc < 0) {
*err_str = "cannot swap layout";
goto out_unlock;
}
out_unlock:
rc2 = llapi_group_unlock(fd_src, gid);
if (rc2 < 0 && rc == 0) {
*err_str = "unlock group lock";
rc = rc2;
}
return rc;
}
/**
* check_lease() - Helper for migrate_copy_data
* @fd: File descriptor on which to check the lease.
*
* Internal helper for migrate_copy_data(). Check lease and report error if
* need be.
*
* Return:
* * %0 Migration can keep on going.
* * %-errno Error occurred, abort migration.
*/
static int check_lease(int fd)
{
int rc;
rc = llapi_lease_check(fd);
if (rc > 0)
return 0; /* llapi_check_lease returns > 0 on success. */
return -EBUSY;
}
static int migrate_nonblock(int fd_src, int fd_dst,
unsigned long long bandwidth_bytes_sec,
long stats_interval_sec, __u64 *dv_src,
char **err_str)
{
struct stat st;
__u64 dv1;
__u64 dv2;
int rc;
rc = fstat(fd_src, &st);
if (rc < 0) {
*err_str = "cannot stat source file";
return -errno;
}
rc = llapi_get_data_version(fd_src, &dv1, LL_DV_RD_FLUSH);
if (rc < 0) {
*err_str = "cannot get data version";
return rc;
}
rc = migrate_copy_data(fd_src, fd_dst, check_lease,
bandwidth_bytes_sec,
stats_interval_sec, st.st_size);
if (rc < 0) {
*err_str = "data copy failed";
return rc;
}
rc = llapi_get_data_version(fd_src, &dv2, LL_DV_RD_FLUSH);
if (rc != 0) {
*err_str = "cannot get data version";
return rc;
}
if (dv_src)
*dv_src = dv2;
if (dv1 != dv2) {
rc = -EAGAIN;
*err_str = "source file changed";
return rc;
}
/* Make sure we keep original atime/mtime values */
rc = migrate_set_timestamps(fd_dst, &st);
if (rc < 0) {
*err_str = "set target file timestamp failed";
return -errno;
}
return 0;
}
static
int lfs_layout_compid_by_pool(char *fname, const char *pool, int *comp_id)
{
struct pool_to_id_cbdata data = { .pool = pool };
struct llapi_layout *layout = NULL;
int rc;
layout = llapi_layout_get_by_path(fname, 0);
if (!layout) {
fprintf(stderr,
"error %s: file '%s' couldn't get layout: rc=%d\n",
progname, fname, errno);
rc = -errno;
goto free_layout;
}
rc = llapi_layout_sanity(layout, false, true);
if (rc < 0) {
llapi_layout_sanity_perror(errno);
goto free_layout;
}
rc = llapi_layout_comp_iterate(layout, find_comp_id_by_pool, &data);
if (rc < 0)
goto free_layout;
*comp_id = data.id;
rc = 0;
free_layout:
if (layout)
llapi_layout_free(layout);
return rc;
}
static int lfs_component_set(char *fname, int comp_id, const char *pool,
__u32 flags, __u32 neg_flags)
{
__u32 ids[2];
__u32 flags_array[2];
size_t count = 0;
int rc;
if (!comp_id) {
if (pool == NULL) {
fprintf(stderr,
"error %s: neither component id nor pool is specified\n",
progname);
return -EINVAL;
}
rc = lfs_layout_compid_by_pool(fname, pool, &comp_id);
if (rc)
return rc;
}
if (flags) {
/* Check if trying to set prefer flags on parity components */
if (flags & LCME_FL_PREF_RW) {
struct llapi_layout *layout;
__u32 comp_flags = 0;
__u32 id;
bool is_parity_comp = false;
layout = llapi_layout_get_by_path(fname, 0);
if (!layout) {
fprintf(stderr,
"error %s: file '%s' couldn't get layout: rc=%d\n",
progname, fname, errno);
return -errno;
}
/* Find the component and check its flags */
rc = llapi_layout_comp_use(layout,
LLAPI_LAYOUT_COMP_USE_FIRST);
while (rc == 0) {
rc = llapi_layout_comp_id_get(layout, &id);
if (rc < 0)
break;
if (id == comp_id) {
rc = llapi_layout_comp_flags_get(layout,
&comp_flags);
if (rc < 0)
break;
is_parity_comp = !!(comp_flags &
LCME_FL_PARITY);
break;
}
rc = llapi_layout_comp_use(layout,
LLAPI_LAYOUT_COMP_USE_NEXT);
}
llapi_layout_free(layout);
if (is_parity_comp) {
fprintf(stderr,
"%s: cannot set prefer flags on parity component '%#x' of file '%s'\n",
progname, comp_id, fname);
return -EINVAL;
}
}
ids[count] = comp_id;
flags_array[count] = flags;
++count;
}
if (neg_flags) {
if (neg_flags & LCME_FL_STALE) {
fprintf(stderr,
"%s: cannot clear 'stale' flags from component. Please use lfs-mirror-resync(1) instead\n",
progname);
return -EINVAL;
}
ids[count] = comp_id;
flags_array[count] = neg_flags | LCME_FL_NEG;
++count;
}
rc = llapi_layout_file_comp_set(fname, ids, flags_array, count);
if (rc) {
if (errno == EUCLEAN) {
rc = -errno;
fprintf(stderr,
"%s: cannot set 'stale' flag on component '%#x' of the last non-stale mirror of '%s'\n",
progname, comp_id, fname);
} else {
fprintf(stderr,
"%s: cannot change the flags of component '%#x' of file '%s': %x / ^(%x)\n",
progname, comp_id, fname, flags, neg_flags);
}
}
return rc;
}
static int lfs_component_set_by_mirror(char *fname, __u16 mirror_id,
__u32 flags, __u32 neg_flags)
{
struct llapi_layout *layout = NULL;
int rc = 0;
int comp_count = 0;
layout = llapi_layout_get_by_path(fname, 0);
if (!layout) {
fprintf(stderr, "%s: cannot get layout of '%s': %s\n",
progname, fname, strerror(errno));
return -errno;
}
rc = llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_FIRST);
if (rc < 0) {
fprintf(stderr, "%s: cannot move to first component of '%s': %s\n",
progname, fname, strerror(-rc));
goto free_layout;
}
/* Count components and find matching mirror components */
while (rc == 0) {
uint32_t comp_mirror_id;
uint32_t comp_id;
rc = llapi_layout_mirror_id_get(layout, &comp_mirror_id);
if (rc < 0)
break;
if (comp_mirror_id == mirror_id) {
rc = llapi_layout_comp_id_get(layout, &comp_id);
if (rc < 0)
break;
/* Set flags on this component */
rc = lfs_component_set(fname, comp_id, NULL, flags,
neg_flags);
if (rc < 0) {
fprintf(stderr, "%s: cannot set flags on component %u of mirror %u in '%s': %s\n",
progname, comp_id, mirror_id, fname,
strerror(-rc));
goto free_layout;
}
comp_count++;
}
rc = llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_NEXT);
if (rc < 0) {
rc = 0; /* End of components */
break;
}
}
if (comp_count == 0) {
fprintf(stderr, "%s: no components found for mirror %u in '%s'\n",
progname, mirror_id, fname);
rc = -EINVAL;
} else {
printf("Set flags on %d component(s) of mirror %u in '%s'\n",
comp_count, mirror_id, fname);
rc = 0;
}
free_layout:
llapi_layout_free(layout);
return rc;
}
static int lfs_component_del(char *fname, __u32 comp_id,
__u32 flags, __u32 neg_flags)
{
int rc = 0;
if (flags && neg_flags) {
fprintf(stderr,
"%s: cannot specify both positive and negative flags\n",
progname);
return -EINVAL;
}
if (!flags && neg_flags)
flags = neg_flags | LCME_FL_NEG;
if (flags && comp_id) {
fprintf(stderr,
"%s: cannot specify component ID and flags at the same time\n",
progname);
return -EINVAL;
}
if (!flags && !comp_id) {
fprintf(stderr,
"%s: neither flags nor component ID is specified\n",
progname);
return -EINVAL;
}
if (flags) {
if (flags & ~LCME_KNOWN_FLAGS) {
fprintf(stderr,
"%s setstripe: unknown flags %#x\n",
progname, flags);
return -EINVAL;
}
} else if (comp_id > LCME_ID_MAX) {
fprintf(stderr, "%s setstripe: invalid component id %u\n",
progname, comp_id);
return -EINVAL;
}
rc = llapi_layout_file_comp_del(fname, comp_id, flags);
if (rc)
fprintf(stderr,
"%s setstripe: cannot delete component %#x from '%s': %s\n",
progname, comp_id, fname, strerror(errno));
return rc;
}
static int lfs_component_add(char *fname, struct llapi_layout *layout)
{
int rc;
if (!layout)
return -EINVAL;
rc = llapi_layout_file_comp_add(fname, layout);
if (rc)
fprintf(stderr, "Add layout component(s) to %s failed. %s\n",
fname, strerror(errno));
return rc;
}
static int lfs_component_create(char *fname, int open_flags, mode_t open_mode,
struct llapi_layout *layout)
{
struct stat st;
int fd;
if (!layout)
return -EINVAL;
fd = lstat(fname, &st);
if (fd == 0 && S_ISDIR(st.st_mode))
open_flags = O_DIRECTORY | O_RDONLY;
fd = llapi_layout_file_open(fname, open_flags, open_mode, layout);
if (fd < 0)
llapi_error(LLAPI_MSG_ERROR, errno, "%s: %s '%s'\n",
progname, S_ISDIR(st.st_mode) ?
"cannot set default composite layout for" :
"cannot create composite file", fname);
return fd;
}
static int lfs_migrate(char *name, enum llapi_migration_flags migration_flags,
struct llapi_stripe_param *param,
struct llapi_layout *layout,
unsigned long long bandwidth_bytes_sec,
long stats_interval_sec)
{
struct llapi_layout *existing;
uint64_t dom_new, dom_cur;
__u64 dv_src = 0;
__u64 dv_dst = 0;
int fd_src = -1;
int fd_dst = -1;
char *err_str = "syserror";
int rc;
rc = migrate_open_files(name, migration_flags, param, layout,
&fd_src, &fd_dst, &err_str);
if (rc < 0)
goto out;
rc = llapi_layout_dom_size(layout, &dom_new);
if (rc) {
err_str = "cannot get new layout DoM size";
goto out;
}
/* special case for migration to DOM layout*/
existing = llapi_layout_get_by_fd(fd_src, 0);
if (!existing) {
err_str = "cannot get existing layout";
goto out;
}
rc = llapi_layout_dom_size(existing, &dom_cur);
if (rc) {
err_str = "cannot get current layout DoM size";
goto out;
}
/*
* if file has DoM layout already then migration is possible to
* the new layout with the same DoM component via swap layout,
* if new layout used bigger DOM size, then mirroring is used
*/
if (dom_new > dom_cur) {
rc = lfs_migrate_to_dom(fd_src, fd_dst, name,
migration_flags,
bandwidth_bytes_sec,
stats_interval_sec);
if (rc)
err_str = "cannot migrate to DOM layout";
goto out_closed;
}
if (stats_interval_sec)
printf("%s:\n", name);
if (!(migration_flags & LLAPI_MIGRATION_NONBLOCK)) {
/*
* Blocking mode (forced if servers do not support file lease).
* It is also the default mode, since we cannot distinguish
* between a broken lease and a server that does not support
* atomic swap/close (LU-6785)
*/
rc = migrate_block(fd_src, fd_dst, bandwidth_bytes_sec,
stats_interval_sec, &err_str);
goto out;
}
rc = llapi_lease_acquire(fd_src, LL_LEASE_RDLCK);
if (rc < 0) {
err_str = "cannot get lease";
goto out;
}
rc = migrate_nonblock(fd_src, fd_dst, bandwidth_bytes_sec,
stats_interval_sec, &dv_src, &err_str);
if (rc < 0) {
llapi_lease_release(fd_src);
goto out;
}
rc = llapi_get_data_version(fd_dst, &dv_dst, LL_DV_RD_FLUSH);
if (rc != 0) {
err_str = "cannot get data version";
return rc;
}
/*
* Atomically put lease, swap layouts and close.
* for a migration we need to check data version on file did
* not change.
*/
rc = llapi_fswap_layouts(fd_src, fd_dst, dv_src, dv_dst,
SWAP_LAYOUTS_CLOSE);
if (rc < 0) {
err_str = "cannot swap layout";
goto out;
}
out:
if (fd_src >= 0)
close(fd_src);
if (fd_dst >= 0)
close(fd_dst);
out_closed:
if (rc < 0)
fprintf(stderr, "error: %s: %s: %s: %s\n",
progname, name, err_str, strerror(-rc));
else if (migration_flags & LLAPI_MIGRATION_VERBOSE)
printf("%s\n", name);
return rc;
}
static int comp_str2flags(char *string, __u32 *flags, __u32 *neg_flags)
{
char *name;
char *dup_string = NULL;
int rc = 0;
*flags = 0;
*neg_flags = 0;
if (!string || !string[0])
return -EINVAL;
dup_string = strdup(string);
if (!dup_string) {
llapi_printf(LLAPI_MSG_ERROR,
"%s: insufficient memory\n",
progname);
return -ENOMEM;
}
for (name = strtok(dup_string, ","); name; name = strtok(NULL, ",")) {
bool found = false;
int i;
for (i = 0; i < ARRAY_SIZE(comp_flags_table); i++) {
__u32 comp_flag = comp_flags_table[i].cfn_flag;
const char *comp_name = comp_flags_table[i].cfn_name;
if (strcmp(name, comp_name) == 0) {
*flags |= comp_flag;
found = true;
} else if (strncmp(name, "^", 1) == 0 &&
strcmp(name + 1, comp_name) == 0) {
*neg_flags |= comp_flag;
found = true;
}
}
if (!found) {
llapi_printf(LLAPI_MSG_ERROR,
"%s: component flag '%s' not supported\n",
progname, name);
rc = -EINVAL;
goto out_free;
}
}
if (!*flags && !*neg_flags)
rc = -EINVAL;
/* don't allow to set and exclude the same flag */
if (*flags & *neg_flags)
rc = -EINVAL;
out_free:
free(dup_string);
return rc;
}
static int mdthash_input(char *string, __u32 *inflags,
__u32 *exflags, __u32 *type)
{
char *name;
struct mhf_list {
char *name;
__u32 flag;
} mhflist[] = {
{"migrating", LMV_HASH_FLAG_MIGRATION},
{"bad_type", LMV_HASH_FLAG_BAD_TYPE},
{"badtype", LMV_HASH_FLAG_BAD_TYPE},
{"lost_lmv", LMV_HASH_FLAG_LOST_LMV},
{"lostlmv", LMV_HASH_FLAG_LOST_LMV},
};
if (string == NULL)
return -EINVAL;
*inflags = 0;
*exflags = 0;
*type = 0;
for (name = strtok(string, ","); name; name = strtok(NULL, ",")) {
bool found = false;
int i;
for (i = 0; i < ARRAY_SIZE(mhflist); i++) {
if (strcmp(name, mhflist[i].name) == 0 ||
name[0] == mhflist[i].name[0]) {
*inflags |= mhflist[i].flag;
found = true;
} else if (name[0] == '^' &&
(strcmp(name + 1, mhflist[i].name) == 0 ||
name[1] == mhflist[i].name[0])) {
*exflags |= mhflist[i].flag;
found = true;
}
}
if (!found) {
i = check_hashtype(name);
if (i > 0) {
*type |= 1 << i;
continue;
}
llapi_printf(LLAPI_MSG_ERROR,
"%s: invalid mdt_hash value '%s'\n",
progname, name);
return -EINVAL;
}
}
/* don't allow to include and exclude the same flag */
if (*inflags & *exflags) {
llapi_printf(LLAPI_MSG_ERROR,
"%s: include and exclude same flag '%s'\n",
progname, string);
return -EINVAL;
}
return 0;
}
static int mirror_str2state(char *string, __u16 *state, __u16 *neg_state)
{
if (!string)
return -EINVAL;
*state = 0;
*neg_state = 0;
if (strncmp(string, "^", 1) == 0) {
*neg_state = llapi_layout_string_flags(string + 1);
if (*neg_state != 0)
return 0;
} else {
*state = llapi_layout_string_flags(string);
if (*state != 0)
return 0;
}
llapi_printf(LLAPI_MSG_ERROR,
"%s: mirrored file state '%s' not supported\n",
progname, string);
return -EINVAL;
}
/*
* struct mirror_args - Command-line arguments for mirror(s).
* @m_count: Number of mirrors to be created with this layout.
* @m_flags: Mirror level flags, only 'prefer' is supported.
* @m_layout: Mirror layout.
* @m_file: A victim file. Its layout will be split and used as a mirror.
* @m_next: Point to the next node of the list.
* @m_inherit: Whether to inherit unspecified attributes from previous mirror.
* @m_has_ec: True if this layout already contains EC parity components.
*
* Command-line arguments for mirror(s) will be parsed and stored in
* a linked list that consists of this structure.
*/
struct mirror_args {
__u32 m_count;
enum lov_comp_md_entry_flags m_flags;
struct llapi_layout *m_layout;
const char *m_file;
struct mirror_args *m_next;
bool m_inherit;
bool m_has_ec;
};
/*
* enum mirror_flags - Flags for extending a mirrored file.
* @MF_NO_VERIFY: indicates to not verify the mirror(s) from victim file(s)
* and the user asserts the victim file(s) contains the same
* data as the original mirrored file.
* @MF_DESTROY: indicates to delete the mirror from the mirrored file.
* @MF_COMP_ID: specified component id instead of mirror id
* @MF_COMP_POOL: specified component pool instead of mirror id
* @MF_FOREIGN: specified foreign component instead of mirror id
*
* Flags for extending a mirrored file.
*/
enum mirror_flags {
MF_NO_VERIFY = 0x01,
MF_DESTROY = 0x02,
MF_COMP_ID = 0x04,
MF_COMP_POOL = 0x08,
MF_FOREIGN = 0x10,
};
/**
* mirror_create_sanity_check() - Check mirror list.
* @fname: file to be created
* @list: A linked list that stores the mirror arguments.
*
* This function does a sanity check on @list for creating
* a mirrored file.
*
* Return: 0 on success or a negative error code on failure.
*/
static int mirror_create_sanity_check(const char *fname,
struct mirror_args *list)
{
int rc = 0;
bool has_m_file = false;
char fsname[MAX_OBD_NAME + 1] = { 0 };
bool has_m_layout = false;
if (!list)
return -EINVAL;
if (fname) {
rc = llapi_search_fsname(fname, fsname);
if (rc) {
fprintf(stderr,
"error: %s: file '%s' has no fsname\n",
progname, fname);
return rc;
}
}
while (list) {
if (list->m_file) {
has_m_file = true;
llapi_layout_free(list->m_layout);
list->m_layout =
llapi_layout_get_by_path(list->m_file, 0);
if (!list->m_layout) {
fprintf(stderr,
"error: %s: file '%s' has no layout\n",
progname, list->m_file);
return -ENODATA;
}
} else {
has_m_layout = true;
if (!list->m_layout) {
fprintf(stderr, "error: %s: no mirror layout\n",
progname);
return -EINVAL;
}
}
/* Skip validation for layouts with EC parity - they will
* be validated after merging in mirror_create().
*/
if (!list->m_has_ec) {
rc = llapi_layout_v2_sanity(list->m_layout, false, true,
fsname);
if (rc) {
llapi_layout_sanity_perror(rc);
return rc;
}
}
list = list->m_next;
}
if (has_m_file && has_m_layout) {
fprintf(stderr,
"error: %s: -f <victim_file> option should not be specified with setstripe options\n",
progname);
return -EINVAL;
}
return 0;
}
static int mirror_set_flags(struct llapi_layout *layout, void *cbdata)
{
__u32 mirror_flags = *(__u32 *)cbdata;
uint32_t flags;
int rc;
rc = llapi_layout_comp_flags_get(layout, &flags);
if (rc < 0)
return rc;
if (!flags) {
rc = llapi_layout_comp_flags_set(layout, mirror_flags);
if (rc)
return rc;
}
return LLAPI_LAYOUT_ITER_CONT;
}
/**
* mirror_create() - Create a mirrored file.
* @fname: The file to be created.
* @mirror_list: A linked list that stores the mirror arguments.
*
* This function creates a mirrored file @fname with the mirror(s)
* from @mirror_list.
*
* Return: 0 on success or a negative error code on failure.
*/
static int mirror_create(char *fname, struct mirror_args *mirror_list)
{
struct llapi_layout *layout = NULL;
struct mirror_args *cur_mirror = NULL;
uint16_t mirror_count = 0;
int i = 0;
int rc = 0;
rc = mirror_create_sanity_check(fname, mirror_list);
if (rc)
return rc;
cur_mirror = mirror_list;
while (cur_mirror) {
rc = llapi_layout_comp_iterate(cur_mirror->m_layout,
mirror_set_flags,
&cur_mirror->m_flags);
if (rc) {
rc = -errno;
fprintf(stderr, "%s: failed to set mirror flags\n",
progname);
goto error;
}
/* If this layout already contains EC parity components,
* the layout contains multiple mirrors (data + parity).
* We need to merge it m_count times, and each merge adds
* 2 mirrors (data + parity), so the total mirror count
* is m_count * 2.
*/
for (i = 0; i < cur_mirror->m_count; i++) {
rc = llapi_layout_merge(&layout, cur_mirror->m_layout);
if (rc) {
rc = -errno;
fprintf(stderr,
"error: %s: merge layout failed: %s\n",
progname, strerror(errno));
goto error;
}
}
if (cur_mirror->m_has_ec) {
/* Layout contains 2 mirrors (data + parity), so
* each merge adds 2 mirrors.
*/
mirror_count += cur_mirror->m_count * 2;
} else {
mirror_count += cur_mirror->m_count;
}
cur_mirror = cur_mirror->m_next;
}
if (!layout) {
fprintf(stderr, "error: %s: layout is NULL\n", progname);
return -EINVAL;
}
rc = llapi_layout_mirror_count_set(layout, mirror_count);
if (rc) {
rc = -errno;
fprintf(stderr, "error: %s: set mirror count failed: %s\n",
progname, strerror(errno));
goto error;
}
/* Validate the merged layout */
rc = llapi_layout_sanity(layout, false, true);
if (rc) {
llapi_layout_sanity_perror(rc);
rc = -EINVAL;
goto error;
}
rc = lfs_component_create(fname, O_CREAT | O_WRONLY, 0666,
layout);
if (rc >= 0) {
close(rc);
rc = 0;
}
error:
llapi_layout_free(layout);
return rc;
}
/**
* mirror_file_compare() - Compare files and check lease on @fd_src
* @fd_src: Source file
* @fd_dst: Destination file
*
* Return bytes number of bytes are the same
*/
static ssize_t mirror_file_compare(int fd_src, int fd_dst)
{
const size_t buflen = DEFAULT_IO_BUFLEN;
void *buf;
ssize_t bytes_done = 0;
int rc;
/* first half of buffer used for read, the other half for compare */
rc = posix_memalign(&buf, sysconf(_SC_PAGESIZE), buflen * 2);
if (rc) {
fprintf(stderr, "%s: posix_memalign() failed: %s\n",
progname, strerror(rc));
return -rc;
}
(void)mlock(buf, buflen * 2);
while (1) {
ssize_t bytes_read;
if (!llapi_lease_check(fd_src)) {
bytes_done = -EBUSY;
break;
}
bytes_read = read(fd_src, buf, buflen);
if (bytes_read <= 0)
break;
if (bytes_read != read(fd_dst, buf + buflen, buflen))
break;
/*
* XXX: should compute the checksum on each buffer and then
* compare checksum to avoid cache collision
*/
if (memcmp(buf, buf + buflen, bytes_read))
break;
bytes_done += bytes_read;
}
(void)munlock(buf, buflen * 2);
free(buf);
return bytes_done;
}
static int
open_by_fid_str(const char *fid_str, const char *path, int *fdp, int flags)
{
char mntdir[PATH_MAX] = {'\0'};
struct lu_fid fid;
int fd;
int rc;
rc = llapi_fid_parse(fid_str, &fid, NULL);
if (rc != 0)
return rc;
rc = llapi_search_mounts(path, 0, mntdir, NULL);
if (rc < 0 || mntdir[0] == '\0') {
fprintf(stderr, "Cannot find mounted Lustre filesystem: %s\n",
(rc < 0) ? strerror(-rc) : strerror(ENODEV));
return rc != 0 ? rc : -ENODEV;
}
fd = llapi_open_by_fid(mntdir, &fid, flags);
if (fd < 0)
return fd;
*fdp = fd;
return rc;
}
/*
* if @param name_or_fid matches the pattern of fid, @param path is any path in
* the lustre fs.
* if @param name_or_fid does not match the pattern of fid, it should be the
* path to file, @param path and @param flags are ignored.
*/
static struct llapi_layout*
layout_get_by_name_or_fid(const char *name_or_fid, const char *path,
enum llapi_layout_get_flags layout_flags, int flags)
{
int rc;
/* Check if name or fid */
if (isdigit(*name_or_fid) || *name_or_fid == '[') {
int fd;
rc = open_by_fid_str(name_or_fid, path, &fd, flags);
if (rc == 0) {
struct llapi_layout *layout;
layout = llapi_layout_get_by_fd(fd, layout_flags);
close(fd);
if (layout != NULL)
return layout;
}
}
/* Then try getting by name */
return llapi_layout_get_by_path(name_or_fid, layout_flags);
}
static int mirror_extend_file(const char *fname, const char *victim_file,
enum mirror_flags mirror_flags,
enum llapi_migration_flags migration_flags)
{
int fd = -1;
int fdv = -1;
struct stat stbuf;
struct stat stbuf_v;
struct ll_ioc_lease *data = NULL;
char *err_str = "syserror";
int open_flags = O_RDWR;
int rc;
if (!(migration_flags & LLAPI_MIGRATION_NONDIRECT))
open_flags |= O_DIRECT;
fd = open(fname, open_flags);
if (fd < 0) {
err_str = "open source file";
rc = -errno;
goto out;
}
fdv = open(victim_file, open_flags);
if (fdv < 0) {
err_str = "open target file";
rc = -errno;
goto out;
}
if (fstat(fd, &stbuf) || fstat(fdv, &stbuf_v)) {
err_str = "stat source or target file";
rc = -errno;
goto out;
}
if (stbuf.st_dev != stbuf_v.st_dev) {
err_str = "stat source and target file";
rc = -EXDEV;
goto out;
}
/* mirrors should be of the same size */
if (stbuf.st_size != stbuf_v.st_size) {
err_str = "file sizes don't match";
rc = -EINVAL;
goto out;
}
rc = llapi_lease_acquire(fd, LL_LEASE_RDLCK);
if (rc < 0) {
err_str = "cannot get lease";
goto out;
}
if (!(mirror_flags & MF_NO_VERIFY)) {
ssize_t ret;
/* mirrors should have the same contents */
ret = mirror_file_compare(fd, fdv);
if (ret != stbuf.st_size) {
err_str = "file busy or contents don't match";
rc = ret < 0 ? ret : -EINVAL;
goto out;
}
}
/* Get rid of caching pages from clients */
rc = llapi_file_flush(fd);
if (rc < 0) {
err_str = "cannot get data version";
goto out;
}
rc = llapi_file_flush(fdv);
if (rc < 0) {
err_str = "cannot get data version";
goto out;
}
rc = migrate_set_timestamps(fd, &stbuf);
if (rc < 0) {
err_str = "cannot set source file timestamp";
goto out;
}
/* Atomically put lease, merge layouts and close. */
data = calloc(1, offsetof(typeof(*data), lil_ids[1]));
if (!data) {
err_str = "memory allocation";
goto out;
}
data->lil_mode = LL_LEASE_UNLCK;
data->lil_flags = LL_LEASE_LAYOUT_MERGE;
data->lil_count = 1;
data->lil_ids[0] = fdv;
rc = llapi_lease_set(fd, data);
if (rc < 0) {
err_str = "cannot merge layout";
goto out;
} else if (rc == 0) {
rc = -EBUSY;
err_str = "lost lease lock";
goto out;
}
rc = 0;
out:
if (data)
free(data);
if (fd >= 0)
close(fd);
if (fdv >= 0)
close(fdv);
if (!rc)
(void) unlink(victim_file);
if (rc < 0)
fprintf(stderr, "error: %s: %s: %s: %s\n",
progname, fname, err_str, strerror(-rc));
return rc;
}
static int mirror_extend_layout(char *name, struct llapi_layout *m_layout,
bool inherit, uint32_t comp_flags,
enum llapi_migration_flags migration_flags,
unsigned long long bandwidth_bytes_sec,
long stats_interval_sec)
{
struct llapi_layout *f_layout = NULL;
struct ll_ioc_lease *data = NULL;
int fd_src = -1;
int fd_dst = -1;
struct stat st;
char *err_str = "syserror";
int rc = 0;
if (inherit) {
f_layout = llapi_layout_get_by_path(name, 0);
if (!f_layout) {
rc = -EINVAL;
err_str = "cannot get layout";
goto out;
}
rc = llapi_layout_get_last_init_comp(f_layout);
if (rc) {
err_str = "cannot get the last init comp";
goto out;
}
rc = llapi_layout_mirror_inherit(f_layout, m_layout);
if (rc) {
err_str = "cannot inherit from the last init comp";
goto out;
}
}
llapi_layout_comp_flags_set(m_layout, comp_flags);
rc = migrate_open_files(name, migration_flags | LLAPI_MIGRATION_MIRROR,
NULL, m_layout, &fd_src, &fd_dst, &err_str);
if (rc < 0)
goto out;
rc = llapi_lease_acquire(fd_src, LL_LEASE_RDLCK);
if (rc < 0) {
err_str = "cannot get lease";
goto out;
}
rc = fstat(fd_src, &st);
if (rc < 0) {
err_str = "cannot stat source file";
goto out;
}
if (stats_interval_sec)
printf("%s:\n", name);
rc = migrate_nonblock(fd_src, fd_dst, bandwidth_bytes_sec,
stats_interval_sec, NULL, &err_str);
if (rc < 0) {
llapi_lease_release(fd_src);
goto out;
}
rc = migrate_set_timestamps(fd_src, &st);
if (rc < 0) {
err_str = "cannot set source file timestamp";
goto out;
}
/* Atomically put lease, merge layouts and close. */
data = calloc(1, offsetof(typeof(*data), lil_ids[1]));
if (!data) {
err_str = "memory allocation";
goto out;
}
data->lil_mode = LL_LEASE_UNLCK;
data->lil_flags = LL_LEASE_LAYOUT_MERGE;
data->lil_count = 1;
data->lil_ids[0] = fd_dst;
rc = llapi_lease_set(fd_src, data);
if (rc < 0) {
err_str = "cannot merge layout";
goto out;
} else if (rc == 0) {
rc = -EBUSY;
err_str = "lost lease lock";
goto out;
}
rc = 0;
out:
if (data)
free(data);
if (fd_src >= 0)
close(fd_src);
if (fd_dst >= 0)
close(fd_dst);
if (rc < 0)
fprintf(stderr, "error: %s: %s: %s: %s\n",
progname, name, err_str, strerror(-rc));
return rc;
}
static int mirror_extend(char *fname, struct mirror_args *mirror_list,
enum mirror_flags mirror_flags,
enum llapi_migration_flags migration_flags,
unsigned long long bandwidth_bytes_sec,
long stats_interval_sec)
{
int rc = 0;
while (mirror_list) {
if (mirror_list->m_file) {
rc = mirror_extend_file(fname, mirror_list->m_file,
mirror_flags, migration_flags);
} else {
__u32 mirror_count = mirror_list->m_count;
while (mirror_count > 0) {
rc = mirror_extend_layout(fname,
mirror_list->m_layout,
mirror_list->m_inherit,
mirror_list->m_flags,
migration_flags,
bandwidth_bytes_sec,
stats_interval_sec);
if (rc)
break;
--mirror_count;
}
}
if (rc)
break;
mirror_list = mirror_list->m_next;
}
return rc;
}
static int find_foreign_id(struct llapi_layout *layout, void *cbdata)
{
uint64_t pattern;
uint32_t id;
int rc;
rc = llapi_layout_pattern_get(layout, &pattern);
if (rc < 0)
return rc;
if (pattern == LLAPI_LAYOUT_FOREIGN) {
rc = llapi_layout_mirror_id_get(layout, &id);
if (rc < 0)
return rc;
*(uint32_t *)cbdata = id;
return LLAPI_LAYOUT_ITER_STOP;
}
return LLAPI_LAYOUT_ITER_CONT;
}
static int find_mirror_id(struct llapi_layout *layout, void *cbdata)
{
uint32_t id;
int rc;
rc = llapi_layout_mirror_id_get(layout, &id);
if (rc < 0)
return rc;
if ((__u16)id == *(__u16 *)cbdata)
return LLAPI_LAYOUT_ITER_STOP;
return LLAPI_LAYOUT_ITER_CONT;
}
static int find_comp_id(struct llapi_layout *layout, void *cbdata)
{
uint32_t id;
int rc;
rc = llapi_layout_comp_id_get(layout, &id);
if (rc < 0)
return rc;
if (id == *(__u32 *)cbdata)
return LLAPI_LAYOUT_ITER_STOP;
return LLAPI_LAYOUT_ITER_CONT;
}
static int find_mirror_id_by_pool(struct llapi_layout *layout, void *cbdata)
{
char buf[LOV_MAXPOOLNAME + 1];
struct pool_to_id_cbdata *d = (void *)cbdata;
uint32_t id;
int rc;
rc = llapi_layout_pool_name_get(layout, buf, sizeof(buf));
if (rc < 0)
return rc;
if (strcmp(d->pool, buf))
return LLAPI_LAYOUT_ITER_CONT;
rc = llapi_layout_mirror_id_get(layout, &id);
if (rc < 0)
return rc;
d->id = id;
return LLAPI_LAYOUT_ITER_STOP;
}
struct parity_check_cbdata {
uint32_t mirror_id;
bool is_parity;
};
static int check_parity_mirror(struct llapi_layout *layout, void *cbdata)
{
struct parity_check_cbdata *d = cbdata;
uint32_t mirror_id;
uint32_t flags;
int rc;
rc = llapi_layout_mirror_id_get(layout, &mirror_id);
if (rc < 0)
return rc;
if (mirror_id != d->mirror_id)
return LLAPI_LAYOUT_ITER_CONT;
rc = llapi_layout_comp_flags_get(layout, &flags);
if (rc < 0)
return rc;
d->is_parity = !!(flags & LCME_FL_PARITY);
return LLAPI_LAYOUT_ITER_STOP;
}
/**
* is_parity_mirror() - Check if a mirror is a parity mirror.
* @layout: Layout to check.
* @mirror_id: Mirror ID to check.
*
* A parity mirror has the LCME_FL_PARITY flag set on all its components.
* We only need to check the first component since either all components
* in a mirror are parity or none are (mixed parity mirrors are invalid).
*
* Return: true if mirror is a parity mirror, false otherwise.
*/
static bool is_parity_mirror(struct llapi_layout *layout, uint32_t mirror_id)
{
struct parity_check_cbdata d = {
.mirror_id = mirror_id,
.is_parity = false,
};
int rc;
rc = llapi_layout_comp_iterate(layout, check_parity_mirror, &d);
if (rc < 0)
return false;
return d.is_parity;
}
static int find_comp_id_by_pool(struct llapi_layout *layout, void *cbdata)
{
char buf[LOV_MAXPOOLNAME + 1];
struct pool_to_id_cbdata *d = (void *)cbdata;
uint32_t id;
int rc;
rc = llapi_layout_pool_name_get(layout, buf, sizeof(buf));
if (rc < 0)
return rc;
if (strcmp(d->pool, buf))
return LLAPI_LAYOUT_ITER_CONT;
rc = llapi_layout_comp_id_get(layout, &id);
if (rc < 0)
return rc;
d->id = id;
return LLAPI_LAYOUT_ITER_STOP;
}
struct collect_ids_data {
__u16 *cid_ids;
int cid_count;
__u16 cid_exclude;
};
static int collect_mirror_id(struct llapi_layout *layout, void *cbdata)
{
struct collect_ids_data *cid = cbdata;
uint32_t id;
int rc;
rc = llapi_layout_mirror_id_get(layout, &id);
if (rc < 0)
return rc;
if ((__u16)id != cid->cid_exclude) {
int i;
for (i = 0; i < cid->cid_count; i++) {
/* already collected the mirror id */
if (id == cid->cid_ids[i])
return LLAPI_LAYOUT_ITER_CONT;
}
cid->cid_ids[cid->cid_count] = id;
cid->cid_count++;
}
return LLAPI_LAYOUT_ITER_CONT;
}
/**
* last_non_stale_mirror() - Check if a mirror is the last non-stale mirror.
* @mirror_id: Mirror id to be checked.
* @layout: Mirror component list.
*
* This function checks if a mirror with specified @mirror_id is the last
* non-stale mirror of a layout @layout.
*
* Return: true or false.
*/
static inline
bool last_non_stale_mirror(__u16 mirror_id, struct llapi_layout *layout)
{
__u16 mirror_ids[128] = { 0 };
struct collect_ids_data cid = { .cid_ids = mirror_ids,
.cid_count = 0,
.cid_exclude = mirror_id, };
int i;
llapi_layout_comp_iterate(layout, collect_mirror_id, &cid);
for (i = 0; i < cid.cid_count; i++) {
struct llapi_resync_comp comp_array[1024] = { { 0 } };
int comp_size = 0;
comp_size = llapi_mirror_find_stale(layout, comp_array,
ARRAY_SIZE(comp_array),
&mirror_ids[i], 1);
if (comp_size == 0)
return false;
}
return true;
}
static int mirror_split(const char *fname, __u32 id, const char *pool,
enum mirror_flags mflags, const char *victim_file)
{
struct llapi_layout *layout;
char parent[PATH_MAX];
char victim[PATH_MAX];
int flags = O_CREAT | O_EXCL | O_LOV_DELAY_CREATE | O_NOFOLLOW;
char *ptr;
struct ll_ioc_lease *data;
uint16_t mirror_count;
__u32 mirror_id;
int mdt_index;
int fd, fdv;
bool purge = true; /* delete mirror by setting fdv=fd */
bool is_encrypted;
int rc;
if (victim_file && (strcmp(fname, victim_file) == 0)) {
fprintf(stderr,
"error %s: the source file '%s' and -f file are the same\n",
progname, fname);
return -EINVAL;
}
/* check fname contains mirror with mirror_id/comp_id */
layout = llapi_layout_get_by_path(fname, 0);
if (!layout) {
fprintf(stderr,
"error %s: file '%s' couldn't get layout\n",
progname, fname);
return -EINVAL;
}
rc = llapi_layout_sanity(layout, false, true);
if (rc) {
llapi_layout_sanity_perror(rc);
goto free_layout;
}
rc = llapi_layout_mirror_count_get(layout, &mirror_count);
if (rc) {
fprintf(stderr,
"error %s: file '%s' couldn't get mirror count\n",
progname, fname);
goto free_layout;
}
if (mirror_count < 2) {
fprintf(stderr,
"error %s: file '%s' has %d component, cannot split\n",
progname, fname, mirror_count);
goto free_layout;
}
if (mflags & MF_COMP_POOL) {
struct pool_to_id_cbdata data = { .pool = pool };
rc = llapi_layout_comp_iterate(layout, find_mirror_id_by_pool,
&data);
mirror_id = data.id;
} else if (mflags & MF_COMP_ID) {
rc = llapi_layout_comp_iterate(layout, find_comp_id, &id);
mirror_id = mirror_id_of(id);
} else if (mflags & MF_FOREIGN) {
rc = llapi_layout_comp_iterate(layout, find_foreign_id, &id);
mirror_id = id;
} else {
rc = llapi_layout_comp_iterate(layout, find_mirror_id, &id);
mirror_id = id;
}
if (rc < 0) {
fprintf(stderr, "error %s: failed to iterate layout of '%s'\n",
progname, fname);
goto free_layout;
} else if (rc == LLAPI_LAYOUT_ITER_CONT) {
if (mflags & MF_COMP_POOL) {
fprintf(stderr,
"error %s: file '%s' does not contain mirror with pool '%s'\n",
progname, fname, pool);
goto free_layout;
} else if (mflags & MF_COMP_ID) {
fprintf(stderr,
"error %s: file '%s' does not contain mirror with comp-id %u\n",
progname, fname, id);
goto free_layout;
} else if (mflags & MF_FOREIGN) {
fprintf(stderr,
"error %s: file '%s' does not contain foreign component\n",
progname, fname);
goto free_layout;
} else {
fprintf(stderr,
"error %s: file '%s' does not contain mirror with id %u\n",
progname, fname, id);
goto free_layout;
}
}
/* Check if this is a parity mirror */
if (is_parity_mirror(layout, mirror_id)) {
/* Parity mirrors can only be split with -d (destroy) */
if (!(mflags & MF_DESTROY)) {
fprintf(stderr,
"error %s: mirror %u is a parity mirror and can only be destroyed with -d flag\n",
progname, mirror_id);
rc = -EINVAL;
goto free_layout;
}
}
if (!victim_file && mflags & MF_DESTROY)
/* Allow mirror split even without the key on encrypted files,
* and in this case of a 'split -d', open file with O_DIRECT
* (no IOs will be done).
*/
fd = open(fname, O_RDWR | O_DIRECT | O_CIPHERTEXT);
else
fd = open(fname, O_RDWR);
if (fd < 0) {
fprintf(stderr,
"error %s: open file '%s' failed: %s\n",
progname, fname, strerror(errno));
goto free_layout;
}
/* get victim file directory pathname */
if (strlen(fname) > sizeof(parent) - 1) {
fprintf(stderr, "error %s: file name of '%s' too long\n",
progname, fname);
rc = -ERANGE;
goto close_fd;
}
strncpy(parent, fname, sizeof(parent));
ptr = strrchr(parent, '/');
if (!ptr) {
if (!getcwd(parent, sizeof(parent))) {
fprintf(stderr, "error %s: getcwd failed: %s\n",
progname, strerror(errno));
rc = -errno;
goto close_fd;
}
} else {
if (ptr == parent)
ptr = parent + 1;
*ptr = '\0';
}
rc = llapi_file_fget_mdtidx(fd, &mdt_index);
if (rc < 0) {
fprintf(stderr, "%s: cannot get MDT index of '%s'\n",
progname, fname);
goto close_fd;
}
rc = llapi_file_is_encrypted(fd);
if (rc < 0) {
fprintf(stderr, "%s: cannot get flags of '%s': %d\n",
progname, fname, rc);
goto close_fd;
}
is_encrypted = rc;
again:
if (!victim_file) {
/* use a temp file to store the splitted layout */
if (mflags & MF_DESTROY) {
char file_path[PATH_MAX];
unsigned int rnumber;
int open_flags;
if (purge) {
/* don't use volatile file for mirror destroy */
fdv = fd;
} else {
/**
* try the old way to delete mirror using
* volatile file.
*/
do {
rnumber = random();
rc = snprintf(file_path,
sizeof(file_path),
"%s/" LUSTRE_VOLATILE_HDR ":%.4X:%.4X:fd=%.2d",
parent, mdt_index,
rnumber, fd);
if (rc < 0 ||
rc >= sizeof(file_path)) {
fdv = -ENAMETOOLONG;
break;
}
open_flags = O_RDWR |
(O_LOV_DELAY_CREATE & ~O_ACCMODE) |
O_CREAT | O_EXCL | O_NOFOLLOW |
/* O_DIRECT for mirror split -d */
O_DIRECT |
/* Allow split without the key */
O_CIPHERTEXT;
fdv = open(file_path, open_flags,
S_IRUSR | S_IWUSR);
if (fdv < 0)
rc = -errno;
} while (fdv < 0 && rc == -EEXIST);
}
} else {
if (is_encrypted) {
rc = -1;
fprintf(stderr,
"error %s: not permitted on encrypted file '%s': %d\n",
progname, fname, rc);
goto close_fd;
}
snprintf(victim, sizeof(victim), "%s.mirror~%u",
fname, mirror_id);
fdv = open(victim, flags, S_IRUSR | S_IWUSR);
}
} else {
/* user specified victim file */
if (is_encrypted) {
rc = -1;
fprintf(stderr,
"error %s: not permitted on encrypted file '%s': %d\n",
progname, fname, rc);
goto close_fd;
}
fdv = open(victim_file, flags, S_IRUSR | S_IWUSR);
}
if (fdv < 0) {
fprintf(stderr,
"error %s: create victim file failed: %s\n",
progname, strerror(errno));
goto close_fd;
}
/* get lease lock of fname */
rc = llapi_lease_acquire(fd, LL_LEASE_WRLCK);
if (rc < 0) {
fprintf(stderr,
"error %s: cannot get lease of file '%s': %d\n",
progname, fname, rc);
goto close_victim;
}
/* Atomatically put lease, split layouts and close. */
data = malloc(offsetof(typeof(*data), lil_ids[2]));
if (!data) {
rc = -ENOMEM;
goto close_victim;
}
data->lil_mode = LL_LEASE_UNLCK;
data->lil_flags = LL_LEASE_LAYOUT_SPLIT;
data->lil_count = 2;
data->lil_ids[0] = fdv;
data->lil_ids[1] = mirror_id;
rc = llapi_lease_set(fd, data);
if (rc <= 0) {
if ((rc == -EINVAL || rc == -EBUSY) && purge) {
/* could be old MDS which prohibit fd==fdv */
purge = false;
goto again;
}
if (rc == 0) /* lost lease lock */
rc = -EBUSY;
fprintf(stderr,
"error %s: cannot split '%s': %s\n",
progname, fname, strerror(-rc));
} else {
rc = 0;
}
free(data);
close_victim:
if (!purge)
close(fdv);
close_fd:
close(fd);
free_layout:
llapi_layout_free(layout);
return rc;
}
static inline
int calc_stripe(const char *filename, off_t *size, off_t *obj_max_kb,
int cap, int min_free, int max_free, const char *pool)
{
off_t filekb = *size / 1024;
off_t filegb = filekb / 1048576;
off_t stripe_count = 1;
off_t ost_max_count = 0;
/* Files up to 1GB will have 1 stripe */
if (filegb < 1)
return 1;
/* Calculate stripe count using sqrt(size_in_GB) + 1 */
stripe_count = (off_t)(sqrt((double)filegb) + 1);
if (*obj_max_kb == 0) {
char mntdir[PATH_MAX] = {'\0'}, fsname[PATH_MAX] = "";
const char *poolname = NULL;
struct obd_statfs stat_buf;
struct obd_uuid uuid_buf;
off_t ost_min_kb = LLONG_MAX;
__u32 index;
int fd, rc;
off_t avail;
/* Calculate cap on object size at 1% of smallest OST
* but only include OSTs that have 256MB+ available space
*/
rc = llapi_search_mounts(filename, 0, mntdir, fsname);
if (rc < 0) {
fprintf(stderr,
"warning: unable to find mount point for '%s'\n",
filename);
return stripe_count;
}
fd = open(mntdir, O_RDONLY);
if (fd < 0) {
fprintf(stderr,
"warning: unable to open mount point '%s': %s\n",
mntdir, strerror(errno));
return stripe_count;
}
if (pool) {
poolname = strchr(pool, '.');
if (poolname) {
if (strncmp(fsname, pool,
strlen(fsname)) != 0) {
fprintf(stderr,
"warning: filesystem name mismatch in pool '%s'\n",
pool);
close(fd);
return stripe_count;
}
poolname++;
} else {
poolname = pool;
}
}
for (index = 0; index < LOV_ALL_STRIPES; index++) {
memset(&stat_buf, 0, sizeof(struct obd_statfs));
memset(&uuid_buf, 0, sizeof(struct obd_uuid));
rc = llapi_obd_fstatfs(fd, LL_STATFS_LOV, index,
&stat_buf, &uuid_buf);
if (rc == -ENODEV)
break;
if (rc == -EAGAIN || rc == -ENODATA)
continue;
if (rc < 0)
continue;
if (poolname && llapi_search_ost(fsname, poolname,
obd_uuid2str(&uuid_buf)) != 1)
continue;
avail = (stat_buf.os_bavail * stat_buf.os_bsize) >> 10;
if (max_free && avail > max_free)
avail = max_free;
if (avail >= min_free) {
ost_max_count++;
if (avail < ost_min_kb)
ost_min_kb = avail;
}
}
close(fd);
if (ost_max_count == 0)
return -1;
if (ost_min_kb == LLONG_MAX) {
fprintf(stderr,
"warning: unable to determine minimum OST size, object size not capped\n");
*obj_max_kb = 0;
} else {
*obj_max_kb = ost_min_kb / cap;
}
}
/* Check if obj_max_kb is 0 after calculation attempt */
if (*obj_max_kb == 0) {
fprintf(stderr,
"warning: unable to determine minimum OST size, object size not capped\n");
return stripe_count;
}
/* If disk usage would exceed the cap, increase the number of stripes.
* Round up to the nearest MB to ensure file will fit.
*/
if (*obj_max_kb > 0 && filekb > stripe_count * (*obj_max_kb))
stripe_count = (filekb + (*obj_max_kb) - 1024) / (*obj_max_kb);
/* Limit the count to the number of eligible OSTs */
if (ost_max_count > 0 && stripe_count > ost_max_count)
return ost_max_count;
return stripe_count >= 1 ? stripe_count : 1;
}
static inline
int lfs_mirror_resync_file(const char *fname, struct ll_ioc_lease *ioc,
__u16 *mirror_ids, int ids_nr,
long stats_interval_sec, long bandwidth_bytes_sec,
bool force_resync);
static int lfs_migrate_to_dom(int fd_src, int fd_dst, char *name,
enum llapi_migration_flags migration_flags,
unsigned long long bandwidth_bytes_sec,
long stats_interval_sec)
{
struct ll_ioc_lease *data = NULL;
char *err_str = "syserror";
int rc;
rc = llapi_lease_acquire(fd_src, LL_LEASE_RDLCK);
if (rc < 0) {
err_str = "cannot get lease";
goto out_close;
}
if (stats_interval_sec)
printf("%s:\n", name);
rc = migrate_nonblock(fd_src, fd_dst, bandwidth_bytes_sec,
stats_interval_sec, NULL, &err_str);
if (rc < 0)
goto out_release;
/* Atomically put lease, merge layouts, resync and close. */
data = calloc(1, offsetof(typeof(*data), lil_ids[1]));
if (!data) {
err_str = "memory allocation";
goto out_release;
}
data->lil_mode = LL_LEASE_UNLCK;
data->lil_flags = LL_LEASE_LAYOUT_MERGE;
data->lil_count = 1;
data->lil_ids[0] = fd_dst;
rc = llapi_lease_set(fd_src, data);
if (rc < 0) {
err_str = "cannot merge layout";
goto out_close;
} else if (rc == 0) {
rc = -EBUSY;
err_str = "lost lease lock";
goto out_close;
}
close(fd_src);
close(fd_dst);
rc = lfs_mirror_resync_file(name, data, NULL, 0,
stats_interval_sec,
bandwidth_bytes_sec, false);
if (rc) {
err_str = "cannot resync file";
goto out;
}
/* delete first mirror now */
rc = mirror_split(name, 1, NULL, MF_DESTROY, NULL);
if (rc < 0)
err_str = "cannot delete old layout";
goto out;
out_release:
llapi_lease_release(fd_src);
out_close:
close(fd_src);
close(fd_dst);
out:
if (rc < 0)
fprintf(stderr, "error: %s: %s: %s: %s\n",
progname, name, err_str, strerror(-rc));
else if (migration_flags & LLAPI_MIGRATION_VERBOSE)
printf("%s\n", name);
if (data)
free(data);
return rc;
}
/**
* parse_targets() - Parse a string containing an target index list into an
* array of integers.
* @tgts: array to store indices in [out]
* @size: size of @tgts array
* @offset: starting index in @tgts
* @arg: string containing OST index list
* @duplicates: tell caller list contains duplicates [out]
*
* The input string contains a comma delimited list of individual
* indices and ranges, for example "1,2-4,7". Add the indices into the
* @tgts array and remove duplicates.
*
* Return:
* * %>0 number of indices in @tgts
* * %-EINVAL unable to parse @arg
*/
static int parse_targets(__u32 *tgts, int size, int offset, char *arg,
bool *duplicates)
{
int rc;
int nr = offset;
int slots = size - offset;
char *ptr = NULL;
bool end_of_loop;
if (!arg)
return -EINVAL;
end_of_loop = false;
while (!end_of_loop) {
int start_index = 0;
int end_index = 0;
int i;
char *endptr = NULL;
rc = -EINVAL;
ptr = strchrnul(arg, ',');
end_of_loop = *ptr == '\0';
*ptr = '\0';
errno = 0;
start_index = strtol(arg, &endptr, 0);
if (endptr == arg) /* no data at all */
break;
if (errno != 0 || start_index < -1 ||
(*endptr != '-' && *endptr != '\0'))
break;
end_index = start_index;
if (*endptr == '-') {
errno = 0;
end_index = strtol(endptr + 1, &endptr, 0);
if (errno != 0 || *endptr != '\0' || end_index < -1)
break;
if (end_index < start_index)
break;
}
for (i = start_index; i <= end_index && slots > 0; i++) {
int j;
/* note presence of duplicates */
for (j = 0; j < offset; j++) {
if (tgts[j] == i)
*duplicates = true;
}
j = offset;
if (j == offset) { /* check complete */
tgts[nr++] = i;
--slots;
}
}
if (slots == 0 && i < end_index)
break;
*ptr = ',';
arg = ++ptr;
offset = nr;
rc = 0;
}
if (!end_of_loop && ptr)
*ptr = ',';
return rc < 0 ? rc : nr;
}
/* EC specification for a component */
struct ec_spec {
int comp_index; /* Which component this applies to (-1 = not set) */
uint8_t data_count; /* k in k+m */
uint8_t parity_count; /* m in k+m */
};
struct lfs_setstripe_args {
unsigned long long lsa_comp_end;
unsigned long long lsa_stripe_size;
unsigned long long lsa_extension_size;
long long lsa_stripe_count;
long long lsa_stripe_off;
__u32 lsa_comp_flags;
__u32 lsa_comp_neg_flags;
unsigned long long lsa_pattern;
enum lmv_hash_type lsa_hash;
unsigned int lsa_mirror_count;
int lsa_nr_tgts;
bool lsa_first_comp;
bool lsa_extension_comp;
__u32 *lsa_tgts;
char *lsa_pool_name;
/* EC support */
struct ec_spec *lsa_ec_specs; /* Array of EC specifications */
int lsa_ec_spec_count; /* Number of EC specs */
int lsa_comp_count; /* Total components added */
bool lsa_has_ec; /* Whether --ec was specified */
};
static inline void setstripe_args_init(struct lfs_setstripe_args *lsa)
{
struct ec_spec *ec_specs = lsa->lsa_ec_specs;
int ec_spec_count = lsa->lsa_ec_spec_count;
int comp_count = lsa->lsa_comp_count;
bool has_ec = lsa->lsa_has_ec;
memset(lsa, 0, sizeof(*lsa));
lsa->lsa_stripe_size = LLAPI_LAYOUT_DEFAULT;
lsa->lsa_stripe_count = LLAPI_LAYOUT_DEFAULT;
lsa->lsa_stripe_off = LLAPI_LAYOUT_DEFAULT;
lsa->lsa_pattern = LLAPI_LAYOUT_RAID0;
lsa->lsa_hash = LMV_HASH_TYPE_UNKNOWN;
lsa->lsa_pool_name = NULL;
lsa->lsa_ec_specs = ec_specs;
lsa->lsa_ec_spec_count = ec_spec_count;
lsa->lsa_comp_count = comp_count;
lsa->lsa_has_ec = has_ec;
}
/**
* setstripe_args_init_inherit() - Initialize and inherit stripe options.
* @lsa: Stripe options to be initialized and inherited.
*
* This function initializes stripe options in @lsa and inherit
* stripe_size, stripe_count and OST pool_name options.
*
* Return: void.
*/
static inline void setstripe_args_init_inherit(struct lfs_setstripe_args *lsa)
{
unsigned long long stripe_size;
long long stripe_count;
char *pool_name = NULL;
unsigned int mirror_count = lsa->lsa_mirror_count;
bool first_comp = lsa->lsa_first_comp;
if (lsa->lsa_pattern == LLAPI_LAYOUT_MDT)
stripe_size = LLAPI_LAYOUT_DEFAULT;
else
stripe_size = lsa->lsa_stripe_size;
stripe_count = lsa->lsa_stripe_count;
pool_name = lsa->lsa_pool_name;
setstripe_args_init(lsa);
lsa->lsa_stripe_size = stripe_size;
lsa->lsa_stripe_count = stripe_count;
lsa->lsa_pool_name = pool_name;
lsa->lsa_mirror_count = mirror_count;
lsa->lsa_first_comp = first_comp;
}
static inline bool setstripe_args_specified(struct lfs_setstripe_args *lsa)
{
return (lsa->lsa_stripe_size != LLAPI_LAYOUT_DEFAULT ||
lsa->lsa_stripe_count != LLAPI_LAYOUT_DEFAULT ||
lsa->lsa_stripe_off != LLAPI_LAYOUT_DEFAULT ||
lsa->lsa_pattern != LLAPI_LAYOUT_RAID0 ||
lsa->lsa_hash != LMV_HASH_TYPE_UNKNOWN ||
lsa->lsa_comp_end != 0);
}
static int lsa_args_stripe_count_check(struct lfs_setstripe_args *lsa)
{
if (lsa->lsa_nr_tgts) {
if (lsa->lsa_nr_tgts < 0 ||
lsa->lsa_nr_tgts >= LOV_MAX_STRIPE_COUNT) {
fprintf(stderr, "Invalid nr_tgts(%d)\n",
lsa->lsa_nr_tgts);
errno = EINVAL;
return -1;
}
if (lsa->lsa_stripe_count > 0 &&
lsa->lsa_stripe_count != LLAPI_LAYOUT_DEFAULT &&
!(lsa->lsa_stripe_count >= LLAPI_LAYOUT_WIDE_MIN &&
lsa->lsa_stripe_count <= LLAPI_LAYOUT_WIDE_MAX) &&
lsa->lsa_nr_tgts != lsa->lsa_stripe_count) {
fprintf(stderr, "stripe_count(%lld) != nr_tgts(%d)\n",
lsa->lsa_stripe_count,
lsa->lsa_nr_tgts);
errno = EINVAL;
return -1;
}
}
return 0;
}
/**
* comp_args_to_layout() - Create or extend a composite layout.
* @composite: Pointer to the composite layout.
* @lsa: Stripe options for the new component.
* @set_extent: If %true then extend else if %false create
*
* This function creates or extends a composite layout by adding a new
* component with stripe options from @lsa.
*
* When modified, adjust llapi_stripe_param_verify() if needed as well.
*
* Return: 0 on success or an error code on failure.
*/
static int comp_args_to_layout(struct llapi_layout **composite,
struct lfs_setstripe_args *lsa,
bool set_extent)
{
struct llapi_layout *layout = *composite;
uint64_t prev_end = 0;
uint64_t size;
int i = 0, rc;
new_comp:
if (!layout) {
layout = llapi_layout_alloc();
if (!layout) {
fprintf(stderr, "Alloc llapi_layout failed. %s\n",
strerror(errno));
errno = ENOMEM;
return -1;
}
*composite = layout;
lsa->lsa_first_comp = true;
} else {
uint64_t start;
/*
* Get current component extent, current component
* must be the tail component.
*/
rc = llapi_layout_comp_extent_get(layout, &start, &prev_end);
if (rc) {
fprintf(stderr, "Get comp extent failed. %s\n",
strerror(errno));
return rc;
}
if (lsa->lsa_first_comp) {
prev_end = 0;
rc = llapi_layout_add_first_comp(layout);
} else {
rc = llapi_layout_comp_add(layout);
}
if (rc) {
fprintf(stderr, "Add component failed. %s\n",
strerror(errno));
return rc;
}
}
rc = llapi_layout_comp_flags_set(layout, lsa->lsa_comp_flags);
if (rc) {
fprintf(stderr, "Set flags 0x%x failed: %s\n",
lsa->lsa_comp_flags, strerror(errno));
return rc;
}
if (set_extent) {
uint64_t comp_end = lsa->lsa_comp_end;
/*
* The extendable component is 0-length, so it can be removed
* if there is insufficient space to extend it.
*/
if (lsa->lsa_extension_comp)
comp_end = prev_end;
rc = llapi_layout_comp_extent_set(layout, prev_end,
comp_end);
if (rc) {
fprintf(stderr, "Set extent [%lu, %lu) failed. %s\n",
prev_end, comp_end, strerror(errno));
return rc;
}
}
/* reset lsa_first_comp */
lsa->lsa_first_comp = false;
/* Data-on-MDT component setting */
if (lsa->lsa_pattern == LLAPI_LAYOUT_MDT) {
/* Yaml support */
if (lsa->lsa_stripe_count == 0)
lsa->lsa_stripe_count = LLAPI_LAYOUT_DEFAULT;
if (lsa->lsa_stripe_size == lsa->lsa_comp_end)
lsa->lsa_stripe_size = LLAPI_LAYOUT_DEFAULT;
if (lsa->lsa_stripe_off == -1 ||
lsa->lsa_stripe_off == 0)
lsa->lsa_stripe_off = LLAPI_LAYOUT_DEFAULT;
/*
* In case of Data-on-MDT patterns the only extra option
* applicable is stripe size option.
*/
if (lsa->lsa_stripe_count != LLAPI_LAYOUT_DEFAULT) {
fprintf(stderr,
"Option 'stripe-count' can't be specified with Data-on-MDT component: %lld\n",
lsa->lsa_stripe_count);
errno = EINVAL;
return -1;
}
if (lsa->lsa_stripe_size != LLAPI_LAYOUT_DEFAULT &&
lsa->lsa_stripe_size != lsa->lsa_comp_end - prev_end) {
fprintf(stderr,
"Option 'stripe-size' can't be specified with Data-on-MDT component: %llu\n",
lsa->lsa_stripe_size);
errno = EINVAL;
return -1;
}
if (lsa->lsa_nr_tgts != 0) {
fprintf(stderr,
"Option 'ost-list' can't be specified with Data-on-MDT component: '%i'\n",
lsa->lsa_nr_tgts);
errno = EINVAL;
return -1;
}
if (lsa->lsa_stripe_off != LLAPI_LAYOUT_DEFAULT) {
fprintf(stderr,
"Option 'stripe-offset' can't be specified with Data-on-MDT component: %lld\n",
lsa->lsa_stripe_off);
errno = EINVAL;
return -1;
}
if (lsa->lsa_pool_name != 0) {
fprintf(stderr,
"Option 'pool' can't be specified with Data-on-MDT component: '%s'\n",
lsa->lsa_pool_name);
errno = EINVAL;
return -1;
}
rc = llapi_layout_pattern_set(layout, lsa->lsa_pattern);
if (rc) {
fprintf(stderr, "Set stripe pattern %#llx failed. %s\n",
lsa->lsa_pattern,
strerror(errno));
return rc;
}
/* Data-on-MDT component has always single stripe up to end */
lsa->lsa_stripe_size = lsa->lsa_comp_end;
} else if (lsa->lsa_pattern == LLAPI_LAYOUT_OVERSTRIPING) {
rc = llapi_layout_pattern_set(layout, lsa->lsa_pattern);
if (rc) {
fprintf(stderr, "Set stripe pattern %#llx failed. %s\n",
lsa->lsa_pattern,
strerror(errno));
return rc;
}
}
size = lsa->lsa_comp_flags & LCME_FL_EXTENSION ?
lsa->lsa_extension_size : lsa->lsa_stripe_size;
if (lsa->lsa_comp_flags & LCME_FL_EXTENSION)
rc = llapi_layout_extension_size_set(layout, size);
else
rc = llapi_layout_stripe_size_set(layout, size);
if (rc) {
fprintf(stderr, "Set stripe size %lu failed: %s\n",
size, strerror(errno));
return rc;
}
rc = llapi_layout_stripe_count_set(layout, lsa->lsa_stripe_count);
if (rc) {
fprintf(stderr, "Set stripe count %lld failed: %s\n",
lsa->lsa_stripe_count, strerror(errno));
return rc;
}
if (lsa->lsa_pool_name) {
rc = llapi_layout_pool_name_set(layout, lsa->lsa_pool_name);
if (rc) {
fprintf(stderr, "Set pool name: %s failed. %s\n",
lsa->lsa_pool_name, strerror(errno));
return rc;
}
} else {
rc = llapi_layout_pool_name_set(layout, "");
if (rc) {
fprintf(stderr, "Clear pool name failed: %s\n",
strerror(errno));
return rc;
}
}
rc = lsa_args_stripe_count_check(lsa);
if (rc)
return rc;
if (lsa->lsa_nr_tgts > 0) {
bool found = false;
for (i = 0; i < lsa->lsa_nr_tgts; i++) {
rc = llapi_layout_ost_index_set(layout, i,
lsa->lsa_tgts[i]);
if (rc)
break;
/* Make sure stripe offset is in OST list. */
if (lsa->lsa_tgts[i] == lsa->lsa_stripe_off)
found = true;
}
if (!found) {
fprintf(stderr, "Invalid stripe offset '%lld', not in the target list",
lsa->lsa_stripe_off);
errno = EINVAL;
return -1;
}
} else if (lsa->lsa_stripe_off != LLAPI_LAYOUT_DEFAULT &&
lsa->lsa_stripe_off != -1) {
rc = llapi_layout_ost_index_set(layout, 0, lsa->lsa_stripe_off);
}
if (rc) {
fprintf(stderr, "Set ost index %d failed. %s\n",
i, strerror(errno));
return rc;
}
/* Create the second, virtual component of extension space */
if (lsa->lsa_extension_comp) {
lsa->lsa_comp_flags |= LCME_FL_EXTENSION;
lsa->lsa_extension_comp = false;
goto new_comp;
}
return rc;
}
static int build_component(struct llapi_layout **layout,
struct lfs_setstripe_args *lsa, bool set_extent)
{
int rc;
rc = comp_args_to_layout(layout, lsa, set_extent);
if (rc)
return rc;
if (lsa->lsa_mirror_count > 0) {
rc = llapi_layout_mirror_count_set(*layout,
lsa->lsa_mirror_count);
if (rc)
return rc;
rc = llapi_layout_flags_set(*layout, LCM_FL_RDONLY);
if (rc)
return rc;
lsa->lsa_mirror_count = 0;
}
return rc;
}
static int build_prev_component(struct llapi_layout **layout,
struct lfs_setstripe_args *prev,
struct lfs_setstripe_args *lsa)
{
int extension = lsa->lsa_comp_flags & LCME_FL_EXTENSION;
int rc;
if (prev->lsa_stripe_size) {
if (extension) {
prev->lsa_comp_end = lsa->lsa_comp_end;
prev->lsa_extension_size = lsa->lsa_extension_size;
prev->lsa_extension_comp = true;
}
rc = build_component(layout, prev, true);
if (rc)
return rc;
}
/*
* Copy lsa to previous lsa;
* if this is an extension component, make the previous invalid;
*/
if (extension)
prev->lsa_stripe_size = 0;
else
*prev = *lsa;
return 0;
}
static int build_layout_from_yaml_node(struct cYAML *node,
struct llapi_layout **layout,
struct lfs_setstripe_args *lsa,
struct lfs_setstripe_args *prevp)
{
struct lfs_setstripe_args prev = { 0 };
__u32 *osts = lsa->lsa_tgts;
char *string;
int rc = 0;
if (!prevp)
prevp = &prev;
while (node) {
string = node->cy_string;
if (node->cy_type == CYAML_TYPE_OBJECT ||
node->cy_type == CYAML_TYPE_ARRAY) {
/* go deep to sub blocks */
rc = build_layout_from_yaml_node(node->cy_child, layout,
lsa, prevp);
if (rc)
return rc;
} else {
if (!node->cy_string)
return -EINVAL;
if (strcmp(string, "lcme_id") == 0 &&
lsa->lsa_stripe_count != LLAPI_LAYOUT_DEFAULT) {
rc = build_prev_component(layout, prevp, lsa);
if (rc)
return rc;
/* initialize lsa. */
setstripe_args_init(lsa);
lsa->lsa_first_comp = false;
lsa->lsa_tgts = osts;
}
/* skip leading lmm_ if present, to simplify parsing */
if (strncmp(string, "lmm_", 4) == 0)
string += 4;
if (node->cy_type == CYAML_TYPE_STRING) {
if (!strcmp(string, "lcme_extent.e_end")) {
if (!strcmp(node->cy_valuestring, "EOF") ||
!strcmp(node->cy_valuestring, "eof"))
lsa->lsa_comp_end = LUSTRE_EOF;
} else if (!strcmp(string, "pool")) {
lsa->lsa_pool_name = node->cy_valuestring;
} else if (!strcmp(string, "pattern")) {
if (!strcmp(node->cy_valuestring, "mdt"))
lsa->lsa_pattern = LLAPI_LAYOUT_MDT;
if (!strcmp(node->cy_valuestring,
"raid0,overstriped"))
lsa->lsa_pattern =
LLAPI_LAYOUT_OVERSTRIPING;
} else if (!strcmp(string, "lcme_flags")) {
rc = comp_str2flags(node->cy_valuestring,
&lsa->lsa_comp_flags,
&lsa->lsa_comp_neg_flags);
if (rc)
return rc;
/*
* Only template flags have meaning in
* the layout for a new file
*/
lsa->lsa_comp_flags &= LCME_TEMPLATE_FLAGS;
}
} else if (node->cy_type == CYAML_TYPE_NUMBER) {
if (!strcmp(string, "lcm_mirror_count")) {
lsa->lsa_mirror_count = node->cy_valueint;
} else if (!strcmp(string, "lcme_extent.e_start")) {
if (node->cy_valueint == 0)
lsa->lsa_first_comp = true;
} else if (!strcmp(string, "lcme_extent.e_end")) {
if (node->cy_valueint == -1)
lsa->lsa_comp_end = LUSTRE_EOF;
else
lsa->lsa_comp_end = node->cy_valueint;
} else if (!strcmp(string, "stripe_count")) {
lsa->lsa_stripe_count = node->cy_valueint;
} else if (!strcmp(string, "stripe_size")) {
lsa->lsa_stripe_size = node->cy_valueint;
} else if (!strcmp(string, "extension_size")) {
lsa->lsa_extension_size = node->cy_valueint;
lsa->lsa_extension_comp = true;
} else if (!strcmp(string, "stripe_offset")) {
lsa->lsa_stripe_off = node->cy_valueint;
} else if (!strcmp(string, "l_ost_idx")) {
osts[lsa->lsa_nr_tgts] = node->cy_valueint;
lsa->lsa_nr_tgts++;
}
}
}
node = node->cy_next;
}
if (prevp == &prev) {
rc = build_prev_component(layout, prevp, lsa);
if (rc)
return rc;
if (!(lsa->lsa_comp_flags & LCME_FL_EXTENSION))
rc = build_component(layout, lsa, *layout != NULL);
}
return rc;
}
static int lfs_comp_create_from_yaml(char *template,
struct llapi_layout **layout,
struct lfs_setstripe_args *lsa,
__u32 *osts)
{
struct cYAML *tree = NULL, *err_rc = NULL;
int rc = 0;
tree = cYAML_build_tree(template, NULL, 0, &err_rc, false);
if (!tree) {
fprintf(stderr, "%s: cannot parse YAML file %s\n",
progname, template);
cYAML_build_error(-EINVAL, -1, "yaml", "from comp yaml",
"can't parse", &err_rc);
cYAML_print_tree2file(stderr, err_rc);
cYAML_free_tree(err_rc);
rc = -EINVAL;
goto err;
}
/* initialize lsa for plain file */
setstripe_args_init(lsa);
lsa->lsa_tgts = osts;
rc = build_layout_from_yaml_node(tree, layout, lsa, NULL);
if (rc) {
fprintf(stderr, "%s: cannot build layout from YAML file %s.\n",
progname, template);
goto err;
}
/* clean clean lsa */
setstripe_args_init(lsa);
err:
if (tree)
cYAML_free_tree(tree);
return rc;
}
/**
* layout_extend_comp() - Get the extension size from the next (SEL) component
* and extend the current component on it. The start of
* the next component is to be adjusted as well.
* @layout: the current layout
* @start: the start of the current component
* @end: the end of the current component [in, out]
* @offset: the offset to adjust the end position to instead of extension size
*
* Return:
* * %0 extended successfully
* * %negative on error
*/
static int layout_extend_comp(struct llapi_layout *layout,
uint64_t start, uint64_t *end,
uint64_t offset)
{
uint64_t size, next_start, next_end;
int rc;
rc = llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_NEXT);
if (rc < 0) {
fprintf(stderr,
"%s setstripe: cannot move component cursor: %s\n",
progname, strerror(errno));
return rc;
}
/*
* Even if the @size will not be used below, this will fail if
* this is not a SEL component - a good confirmation we are
* working on right components.
*/
rc = llapi_layout_extension_size_get(layout, &size);
if (rc < 0) {
fprintf(stderr,
"%s setstripe: cannot get component ext size: %s\n",
progname, strerror(errno));
return rc;
}
rc = llapi_layout_comp_extent_get(layout, &next_start, &next_end);
if (rc) {
fprintf(stderr, "%s setstripe: cannot get extent: %s\n",
progname, strerror(errno));
return rc;
}
next_start += offset ?: size;
rc = llapi_layout_comp_extent_set(layout, next_start, next_end);
if (rc) {
fprintf(stderr, "%s setstripe: cannot set extent: %s\n",
progname, strerror(errno));
return rc;
}
rc = llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_PREV);
if (rc < 0) {
fprintf(stderr,
"%s setstripe: cannot move component cursor: %s\n",
progname, strerror(errno));
return rc;
}
*end += offset ?: size;
rc = llapi_layout_comp_extent_set(layout, start, *end);
if (rc) {
fprintf(stderr, "%s setstripe: cannot set extent: %s\n",
progname, strerror(errno));
return rc;
}
return 0;
}
/*
* In 'lfs setstripe --component-add' mode, we need to fetch the extent
* end of the last component in the existing file, and adjust the
* first extent start of the components to be added accordingly.
*
* In the create mode, we need to check if the first component is an extendable
* SEL component and extend its length to the extension size (first component
* of the PFL file is initialised at the create time, cannot be 0-lenght.
*/
static int layout_adjust_first_extent(char *fname, struct llapi_layout *layout,
bool comp_add)
{
struct llapi_layout *head;
uint64_t start = 0, prev_end = 0;
uint64_t end;
int rc, ret = 0;
if (!layout || !(comp_add || llapi_layout_is_composite(layout)))
return 0;
errno = 0;
while (comp_add) {
head = llapi_layout_get_by_path(fname, 0);
if (!head) {
fprintf(stderr,
"%s setstripe: cannot read layout from '%s': %s\n",
progname, fname, strerror(errno));
return -EINVAL;
} else if (errno == ENODATA) {
/*
* file without LOVEA, this component-add will be turned
* into a component-create.
*/
llapi_layout_free(head);
ret = -ENODATA;
/*
* the new layout will be added to an empty one, it
* still needs to be adjusted below
*/
comp_add = 0;
break;
} else if (!llapi_layout_is_composite(head)) {
fprintf(stderr,
"%s setstripe: '%s' not a composite file\n",
progname, fname);
llapi_layout_free(head);
return -EINVAL;
}
rc = llapi_layout_comp_extent_get(head, &start, &prev_end);
if (rc) {
fprintf(stderr,
"%s setstripe: cannot get prev extent: %s\n",
progname, strerror(errno));
llapi_layout_free(head);
return rc;
}
llapi_layout_free(head);
break;
}
/* Make sure we use the first component of the layout to be added. */
rc = llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_FIRST);
if (rc < 0) {
fprintf(stderr,
"%s setstripe: cannot move component cursor: %s\n",
progname, strerror(errno));
return rc;
}
rc = llapi_layout_comp_extent_get(layout, &start, &end);
if (rc) {
fprintf(stderr, "%s setstripe: cannot get extent: %s\n",
progname, strerror(errno));
return rc;
}
if (start == 0 && end == 0) {
rc = layout_extend_comp(layout, start, &end,
comp_add ? prev_end : 0);
if (rc)
return rc;
}
if (start > prev_end || end < prev_end) {
fprintf(stderr,
"%s setstripe: first extent [%lu, %lu) not adjacent with extent end %lu\n",
progname, start, end, prev_end);
return -EINVAL;
}
rc = llapi_layout_comp_extent_set(layout, prev_end, end);
if (rc) {
fprintf(stderr,
"%s setstripe: cannot set component extent [%lu, %lu): %s\n",
progname, prev_end, end, strerror(errno));
return rc;
}
return ret;
}
static int mirror_adjust_first_extents(struct mirror_args *list)
{
int rc = 0;
if (!list)
return 0;
while (list) {
rc = layout_adjust_first_extent(NULL, list->m_layout, false);
if (rc)
break;
list = list->m_next;
}
return rc;
}
static inline bool arg_is_eof(char *arg)
{
return !strncmp(arg, "-1", strlen("-1")) ||
!strncmp(arg, "EOF", strlen("EOF")) ||
!strncmp(arg, "eof", strlen("eof"));
}
/* Forward declaration */
static int apply_ec_specs(struct ec_spec *ec_specs, int ec_spec_count,
int comp_count, int **ec_map);
/**
* lfs_mirror_alloc() - Allocate a mirror argument structure.
*
* Return: Valid mirror_args pointer on success and
* NULL if memory allocation fails.
*/
static struct mirror_args *lfs_mirror_alloc(void)
{
struct mirror_args *mirror = NULL;
while (1) {
mirror = calloc(1, sizeof(*mirror));
if (mirror) {
mirror->m_inherit = false;
break;
}
sleep(1);
}
return mirror;
}
/**
* lfs_mirror_free() - Free memory allocated for a mirror argument
* structure.
* @mirror: Previously allocated mirror argument structure by
* lfs_mirror_alloc().
*
* Free memory allocated for @mirror.
*
* Return: void.
*/
static void lfs_mirror_free(struct mirror_args *mirror)
{
if (mirror->m_layout)
llapi_layout_free(mirror->m_layout);
free(mirror);
}
/**
* lfs_mirror_list_free() - Free memory allocated for a mirror list.
* @mirror_list: Previously allocated mirror list.
*
* Free memory allocated for @mirror_list.
*
* Return: void.
*/
static void lfs_mirror_list_free(struct mirror_args *mirror_list)
{
struct mirror_args *next_mirror = NULL;
while (mirror_list) {
next_mirror = mirror_list->m_next;
lfs_mirror_free(mirror_list);
mirror_list = next_mirror;
}
}
enum {
LFS_SETQUOTA_DELETE = (CHAR_MAX + 1),
LFS_POOL_OPT,
LFS_COMP_COUNT_OPT,
LFS_COMP_START_OPT,
LFS_COMP_FLAGS_OPT,
LFS_COMP_DEL_OPT,
LFS_COMP_SET_OPT,
LFS_COMP_ADD_OPT,
LFS_COMP_NO_VERIFY_OPT,
LFS_EC_STRIPE_COUNT_OPT,
LFS_EC_EXPERT_OPT,
LFS_PROJID_OPT,
LFS_LAYOUT_FLAGS_OPT, /* used for mirror and foreign flags */
LFS_MIRROR_ID_OPT,
LFS_MIRROR_STATE_OPT,
LFS_LAYOUT_COPY,
LFS_MIRROR_INDEX_OPT,
LFS_LAYOUT_FOREIGN_OPT,
LFS_MODE_OPT,
LFS_NEWERXY_OPT,
LFS_INHERIT_RR_OPT,
LFS_FIND_PERM,
LFS_PRINTF_OPT,
LFS_PRINTF_LS,
LFS_NO_FOLLOW_OPT,
LFS_HEX_IDX_OPT,
LFS_RESTRIPE_OPT,
LFS_STATS_OPT,
LFS_STATS_INTERVAL_OPT,
LFS_LINKS_OPT,
LFS_ATTRS_OPT,
LFS_XATTRS_MATCH_OPT,
LFS_MIGRATE_NOFIX,
LFS_QUOTA_FILESYSTEM_OPT,
LFS_QUOTA_SPACE_OPT,
LFS_QUOTA_BGRACE_OPT,
LFS_QUOTA_INODES_OPT,
LFS_QUOTA_ISOFTLIMIT_OPT,
LFS_QUOTA_IHARDLIMIT_OPT,
LFS_QUOTA_IGRACE_OPT,
LFS_QUOTA_SHOW_ROOT_OPT,
LFS_FILES_FROM,
LFS_THREAD_OPT,
LFS_LUSTRE_DIR,
LFS_MIN_FREE_OPT,
LFS_MAX_FREE_OPT,
LFS_LQA_OPT,
LFS_QUOTA_DEFAULT_OPT,
};
#ifndef LCME_USER_MIRROR_FLAGS
/* The mirror flags can be set by users at creation time. */
#define LCME_USER_MIRROR_FLAGS (LCME_FL_PREF_RW)
#endif
/* functions */
/**
* apply_ec_specs() - Apply EC specifications to all components.
* @ec_specs: Array of EC specifications.
* @ec_spec_count: Number of EC specifications.
* @comp_count: Total number of components.
* @ec_map: Output array mapping component index to EC spec index.
*
* Apply EC specifications to components: each --ec value applies to its
* component and inherits to subsequent components until replaced by another
* --ec. The first --ec also applies to any earlier components.
*
* Return: 0 on success, -ENOMEM on failure.
*/
static int apply_ec_specs(struct ec_spec *ec_specs, int ec_spec_count,
int comp_count, int **ec_map)
{
int *map;
int i, j;
if (ec_spec_count == 0 || comp_count == 0)
return -EINVAL;
/* Allocate map: component index -> EC spec index (-1 = none) */
map = malloc(comp_count * sizeof(int));
if (!map)
return -ENOMEM;
/* Initialize all to -1 (no EC) */
for (i = 0; i < comp_count; i++)
map[i] = -1;
/* Single EC spec: applies to all components */
if (ec_spec_count == 1) {
for (i = 0; i < comp_count; i++)
map[i] = 0;
*ec_map = map;
return 0;
}
/*
* Multiple EC specs: each inherits to subsequent components.
* Each --ec value applies to its component and all following
* components until replaced by the next --ec (or end of components).
*/
for (i = 0; i < ec_spec_count; i++) {
int comp_idx = ec_specs[i].comp_index;
int next_boundary = comp_count;
/* Find next EC boundary */
for (j = i + 1; j < ec_spec_count; j++) {
if (ec_specs[j].comp_index > comp_idx) {
next_boundary = ec_specs[j].comp_index;
break;
}
}
/* Apply from comp_idx to next_boundary */
for (j = comp_idx; j < next_boundary; j++)
map[j] = i;
}
/*
* Apply first --ec to any earlier components.
* Fill in any remaining gaps (map[i] == -1) with the first
* EC spec found to the right.
*/
for (i = 0; i < comp_count; i++) {
if (map[i] == -1) {
/* Find first EC spec to the right */
for (j = i + 1; j < comp_count; j++) {
if (map[j] != -1) {
map[i] = map[j];
break;
}
}
}
}
*ec_map = map;
return 0;
}
/**
* create_ec_parity_mirror() - Create parity mirror from data mirror.
* @layout: Layout containing the data mirror.
* @ec_specs: Array of EC specifications.
* @ec_spec_count: Number of EC specifications.
*
* Create a parity mirror by iterating through the data mirror components,
* extracting their properties (extent, stripe_size, pool), and creating
* matching EC parity components with the appropriate EC parameters.
*
* The caller is responsible for setting the mirror count on the layout.
* In mirror mode, mirror_create() sets the mirror count. In non-mirror mode,
* the caller should set the mirror count to 2 (data + parity) after calling
* this function.
*
* Return: 0 on success, negative error code on failure.
*/
static int create_ec_parity_mirror(struct llapi_layout *layout,
struct ec_spec *ec_specs,
int ec_spec_count)
{
struct {
uint64_t start;
uint64_t end;
uint64_t stripe_count;
uint32_t mirror_id;
int ec_idx;
} *comp_info = NULL;
int *ec_map = NULL;
int rc, i, comp_count;
if (!layout || !ec_specs || ec_spec_count == 0)
return -EINVAL;
/* Move to first component of data mirror */
rc = llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_FIRST);
if (rc) {
fprintf(stderr, "error: cannot move to first component: %s\n",
strerror(errno));
return rc;
}
/* Count components by iterating through them */
comp_count = 1;
while (llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_NEXT) == 0)
comp_count++;
/*
* Normalize EC spec comp_index values to be relative to this mirror.
* The comp_index values are global (across all mirrors), but we need
* them to be local (starting from 0) for propagation to work correctly.
*/
int min_comp_index = INT_MAX;
for (i = 0; i < ec_spec_count; i++) {
if (ec_specs[i].comp_index < min_comp_index)
min_comp_index = ec_specs[i].comp_index;
}
/* Create a temporary copy of EC specs with normalized comp_index */
struct ec_spec *normalized_specs = calloc(ec_spec_count,
sizeof(struct ec_spec));
if (!normalized_specs) {
fprintf(stderr, "error: cannot allocate normalized EC specs\n");
return -ENOMEM;
}
for (i = 0; i < ec_spec_count; i++) {
normalized_specs[i] = ec_specs[i];
normalized_specs[i].comp_index -= min_comp_index;
}
/* Apply EC specs to get mapping for each component */
rc = apply_ec_specs(normalized_specs, ec_spec_count, comp_count,
&ec_map);
free(normalized_specs);
if (rc) {
fprintf(stderr, "error: failed to apply EC specs: %s\n",
strerror(-rc));
return rc;
}
rc = llapi_layout_mirror_count_sync(layout);
if (rc) {
fprintf(stderr, "error: cannot sync mirror count: %s\n",
strerror(errno));
goto out_free;
}
/* Allocate array to store component info */
comp_info = calloc(comp_count, sizeof(*comp_info));
if (!comp_info) {
rc = -ENOMEM;
goto out_free;
}
/* Move back to first component */
rc = llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_FIRST);
if (rc) {
fprintf(stderr, "error: cannot move to first component: %s\n",
strerror(errno));
goto out_free;
}
/* Collect extent information from all data mirror components */
for (i = 0; i < comp_count; i++) {
int ec_idx = ec_map[i];
if (ec_idx < 0 || ec_idx >= ec_spec_count) {
fprintf(stderr,
"error: invalid EC mapping for component %d\n",
i);
rc = -EINVAL;
goto out_free;
}
/* Get component extent */
rc = llapi_layout_comp_extent_get(layout,
&comp_info[i].start,
&comp_info[i].end);
if (rc) {
fprintf(stderr,
"error: cannot get component %d extent: %s\n",
i, strerror(errno));
goto out_free;
}
/* Get component stripe count */
rc = llapi_layout_stripe_count_get(layout,
&comp_info[i].stripe_count);
if (rc) {
fprintf(stderr,
"error: cannot get component %d stripe count: %s\n",
i, strerror(errno));
goto out_free;
}
/* Get component mirror ID */
rc = llapi_layout_mirror_id_get(layout,
&comp_info[i].mirror_id);
if (rc) {
fprintf(stderr,
"error: cannot get component %d mirror ID: %s\n",
i, strerror(errno));
goto out_free;
}
comp_info[i].ec_idx = ec_idx;
/* Move to next component for next iteration */
if (i < comp_count - 1) {
rc = llapi_layout_comp_use(layout,
LLAPI_LAYOUT_COMP_USE_NEXT);
if (rc) {
fprintf(stderr,
"error: cannot move to next component: %s\n",
strerror(errno));
goto out_free;
}
}
}
/* Now create EC components for each data component */
for (i = 0; i < comp_count; i++) {
uint8_t data_count, parity_count;
/* Get EC parameters for this component */
data_count = ec_specs[comp_info[i].ec_idx].data_count;
parity_count = ec_specs[comp_info[i].ec_idx].parity_count;
/*
* Reduce EC data count if the actual stripe count is less
* than k. E.g.: -E 128M -c 4 -E -1 -c 16 --ec 8+2
* creates EC 4+2 for first component and EC 8+2 for second.
*/
if (comp_info[i].stripe_count < data_count)
data_count = comp_info[i].stripe_count;
/* Create EC parity component */
rc = llapi_layout_comp_add_ec(layout, comp_info[i].mirror_id,
comp_info[i].start,
comp_info[i].end,
data_count, parity_count);
if (rc) {
fprintf(stderr,
"error: cannot add EC component [%llu, %llu] to protect mirror id %u with EC(%u+%u): %s\n",
(unsigned long long)comp_info[i].start,
(unsigned long long)comp_info[i].end,
comp_info[i].mirror_id, data_count,
parity_count, strerror(errno));
goto out_free;
}
}
/* Validate the layout */
rc = llapi_layout_sanity(layout, false, false);
if (rc)
goto out_free;
rc = 0;
out_free:
free(comp_info);
free(ec_map);
return rc;
}
/* EC stripe count limits for non-expert mode */
#define LOV_EC_DATA_STRIPES_RECOMMENDED 32
#define LOV_EC_CODING_STRIPES_RECOMMENDED 4
/**
* parse_ec_stripe_count() - Parse EC stripe count in k+m or k:m format.
* @optarg: String containing EC stripe count (e.g., "4+2" or "4:2").
* @data_count: Pointer to store data stripe count (k).
* @parity_count: Pointer to store parity stripe count (m).
* @ec_expert: Whether --ec-expert flag was specified.
*
* Parse the EC stripe count string in "k+m" or "k:m" format where k is the
* number of data stripes and m is the number of parity stripes.
*
* Return: 0 on success, -EINVAL on failure.
*/
static int parse_ec_stripe_count(const char *optarg, uint8_t *data_count,
uint8_t *parity_count, bool ec_expert)
{
char *end;
unsigned long data, parity;
const char *sep;
/* Find the '+' or ':' separator */
sep = strchr(optarg, '+');
if (!sep)
sep = strchr(optarg, ':');
if (!sep) {
fprintf(stderr,
"error: invalid EC stripe count '%s', expected format: DATA+PARITY or DATA:PARITY\n",
optarg);
return -EINVAL;
}
/* Parse data count */
errno = 0;
data = strtoul(optarg, &end, 0);
if (errno != 0 || end != sep || data == 0 || data < 2 ||
data > UINT8_MAX) {
fprintf(stderr,
"error: invalid data stripe count in '%s'\n", optarg);
return -EINVAL;
}
/* Parse parity count */
errno = 0;
parity = strtoul(sep + 1, &end, 0);
if (errno != 0 || *end != '\0' || parity == 0 || parity > UINT8_MAX) {
fprintf(stderr,
"error: invalid parity stripe count in '%s'\n",
optarg);
return -EINVAL;
}
/* Validate EC parameters */
if (parity > data) {
fprintf(stderr,
"error: parity count (%lu) must be less than or equal to data count (%lu)\n",
parity, data);
return -EINVAL;
}
if (data + parity > LOV_EC_MAX_TOTAL_STRIPES) {
fprintf(stderr,
"error: data+parity (%lu+%lu=%lu) exceeds %u (Cauchy-matrix-over-GF(2^8) limit)\n",
data, parity, data + parity,
LOV_EC_MAX_TOTAL_STRIPES);
return -EINVAL;
}
/* Check supported limits unless --ec-expert is specified */
if (!ec_expert) {
if (data > LOV_EC_DATA_STRIPES_RECOMMENDED) {
fprintf(stderr,
"error: data stripe count (%lu) exceeds supported limit (%u). Use --ec-expert to override.\n",
data, LOV_EC_DATA_STRIPES_RECOMMENDED);
return -EINVAL;
}
if (parity > LOV_EC_CODING_STRIPES_RECOMMENDED) {
fprintf(stderr,
"error: parity stripe count (%lu) exceeds supported limit (%u). Use --ec-expert to override.\n",
parity, LOV_EC_CODING_STRIPES_RECOMMENDED);
return -EINVAL;
}
}
*data_count = (uint8_t)data;
*parity_count = (uint8_t)parity;
return 0;
}
static int guess_only_lustre_mount_root(char *mntdir)
{
int rc;
char buf[PATH_MAX] = {0};
mntdir[0] = '\0';
rc = llapi_search_mounts(NULL, 0, mntdir, NULL);
if (rc != 0) {
llapi_error(LLAPI_MSG_DEBUG, rc,
"no lustre mount point");
return rc;
}
rc = llapi_search_mounts(NULL, 1, buf, NULL);
if (rc == 0) {
llapi_error(LLAPI_MSG_DEBUG, EEXIST,
"multiple lustre mount points: %s %s", mntdir, buf);
return -EEXIST;
}
return 0;
}
static int lfs_setstripe_internal(int argc, char **argv,
enum setstripe_origin opc)
{
struct lfs_setstripe_args lsa = { 0 };
struct llapi_stripe_param *param = NULL;
struct find_param migrate_mdt_param = {
.fp_max_depth = -1,
.fp_mdt_index = -1,
};
char *fname;
int result = 0;
int result2 = 0;
char *end;
int c;
int delete = 0;
unsigned long long size_units = 1;
bool migrate_mode = false;
bool migrate_mdt_mode = false;
bool setstripe_mode = false;
bool migration_block = false;
enum llapi_migration_flags migration_flags = 0;
__u32 tgts[LOV_MAX_STRIPE_COUNT] = { 0 };
int comp_del = 0, comp_set = 0;
int comp_add = 0;
__u32 comp_id = 0;
struct llapi_layout *layout = NULL;
struct llapi_layout **lpp = &layout;
bool mirror_mode = false;
bool mirror_total_mode = false;
bool has_m_file = false;
__u32 mirror_count = 0;
enum mirror_flags mirror_flags = 0;
struct mirror_args *mirror_list = NULL;
struct mirror_args *new_mirror = NULL;
struct mirror_args *last_mirror = NULL;
__u16 mirror_id = 0;
char cmd[PATH_MAX];
bool from_yaml = false;
bool from_copy = false;
char *template = NULL;
bool foreign_mode = false;
char *xattr = NULL;
bool overstriped = false;
bool clear_hash_fixed = false;
uint32_t type = LU_FOREIGN_TYPE_NONE, flags = 0;
char *mode_opt = NULL;
mode_t previous_umask = 0;
mode_t mode = S_IRUSR | S_IWUSR | S_IRGRP | S_IWGRP | S_IROTH | S_IWOTH;
unsigned long long bandwidth_bytes_sec = 0;
unsigned long long bandwidth_unit = ONE_MB;
long stats_interval_sec = 0;
bool null_mode = false;
bool stripe_count_set = false;
bool min_free_set = false;
bool cap_set = false;
bool comp_end_set = false;
bool auto_stripe = false;
off_t obj_max_kb = 0;
int cap = 100;
int long_idx;
bool restripe = false;
int min_free = 256 * 1024;
int max_free = 0;
const char *files_from = NULL;
FILE *files_from_fp = NULL;
int delim = '\n';
char *buf = NULL;
size_t bufsize = 0;
char *lustre_dir = NULL;
char mntdir[PATH_MAX];
int lustre_dir_fd = -1;
bool fid_mode = false;
struct lu_fid fid;
struct option long_opts[] = {
{ .val = LFS_COMP_ADD_OPT,
.name = "comp-add", .has_arg = no_argument },
{ .val = LFS_COMP_ADD_OPT,
.name = "component-add", .has_arg = no_argument },
{ .val = LFS_COMP_DEL_OPT,
.name = "comp-del", .has_arg = no_argument },
{ .val = LFS_COMP_DEL_OPT,
.name = "component-del", .has_arg = no_argument },
{ .val = LFS_COMP_FLAGS_OPT,
.name = "comp-flags", .has_arg = required_argument },
{ .val = LFS_COMP_FLAGS_OPT,
.name = "component-flags",
.has_arg = required_argument },
{ .val = LFS_COMP_SET_OPT,
.name = "comp-set", .has_arg = no_argument },
{ .val = LFS_COMP_SET_OPT,
.name = "component-set",
.has_arg = no_argument},
{ .val = LFS_COMP_NO_VERIFY_OPT,
.name = "no-verify", .has_arg = no_argument},
{ .val = LFS_LAYOUT_FLAGS_OPT,
.name = "flags", .has_arg = required_argument},
{ .val = LFS_LAYOUT_FOREIGN_OPT,
.name = "foreign", .has_arg = optional_argument},
{ .val = LFS_MIRROR_ID_OPT,
.name = "mirror-id", .has_arg = required_argument},
{ .val = LFS_MODE_OPT,
.name = "mode", .has_arg = required_argument},
{ .val = LFS_LAYOUT_COPY,
.name = "copy", .has_arg = required_argument},
{ .val = LFS_RESTRIPE_OPT,
.name = "restripe", .has_arg = no_argument},
{ .val = LFS_STATS_OPT,
.name = "stats", .has_arg = no_argument},
{ .val = LFS_STATS_INTERVAL_OPT,
.name = "stats-interval",
.has_arg = required_argument},
{ .val = LFS_FILES_FROM,
.name = "files-from", .has_arg = required_argument},
{ .val = LFS_LUSTRE_DIR,
.name = "lustre-dir", .has_arg = required_argument},
{ .val = '0', .name = "null", .has_arg = no_argument },
{ .val = 'A', .name = "auto-stripe", .has_arg = no_argument },
/* find { .val = 'A', .name = "atime", .has_arg = required_argument }*/
/* --block is only valid in migrate mode */
{ .val = 'b', .name = "block", .has_arg = no_argument },
/* find { .val = 'B', .name = "btime", .has_arg = required_argument }*/
{ .val = 'c', .name = "stripe-count", .has_arg = required_argument},
{ .val = 'c', .name = "stripe_count", .has_arg = required_argument},
{ .val = 'c', .name = "mdt-count", .has_arg = required_argument},
{ .val = 'C', .name = "overstripe-count",
.has_arg = required_argument},
{ .val = 'C', .name = "mdt-overcount",
.has_arg = required_argument},
{ .val = 'd', .name = "delete", .has_arg = no_argument},
{ .val = 'd', .name = "destroy", .has_arg = no_argument},
/* used with "lfs migrate -m" */
{ .val = 'd', .name = "directory", .has_arg = no_argument},
/* --non-direct is only valid in migrate and mirror mode */
{ .val = 'D', .name = "non-direct", .has_arg = no_argument },
{ .val = 'E', .name = "comp-end", .has_arg = required_argument},
{ .val = 'E', .name = "component-end",
.has_arg = required_argument},
{ .val = LFS_EC_STRIPE_COUNT_OPT,
.name = "ec", .has_arg = required_argument},
{ .val = LFS_EC_STRIPE_COUNT_OPT,
.name = "ec-stripe-count", .has_arg = required_argument},
{ .val = LFS_EC_EXPERT_OPT,
.name = "ec-expert", .has_arg = required_argument},
{ .val = 'f', .name = "file", .has_arg = required_argument },
{ .val = 'F', .name = "fid", .has_arg = no_argument },
/* find { .val = 'g', .name = "gid", .has_arg = no_argument }, */
/* find { .val = 'G', .name = "group", .has_arg = required_argument }*/
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'H', .name = "mdt-hash", .has_arg = required_argument},
{ .val = 'i', .name = "stripe-index", .has_arg = required_argument},
{ .val = 'i', .name = "stripe_index", .has_arg = required_argument},
{ .val = 'I', .name = "comp-id", .has_arg = required_argument},
{ .val = 'I', .name = "component-id", .has_arg = required_argument},
{ .val = 'K', .name = "auto-cap", .has_arg = required_argument},
/* find { .val = 'l', .name = "lazy", .has_arg = no_argument }, */
{ .val = 'L', .name = "layout", .has_arg = required_argument},
{ .val = 'm', .name = "mdt", .has_arg = required_argument},
{ .val = 'm', .name = "mdt-index", .has_arg = required_argument},
{ .val = 'm', .name = "mdt_index", .has_arg = required_argument},
{ .val = 'M', .name = "min-free", .has_arg = required_argument},
/* --non-block is only valid in migrate mode */
{ .val = 'n', .name = "non-block", .has_arg = no_argument },
{ .val = 'N', .name = "mirror-count", .has_arg = optional_argument},
{ .val = 'o', .name = "ost", .has_arg = required_argument },
#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(3, 0, 53, 0)
{ .val = 'o', .name = "ost-list", .has_arg = required_argument },
{ .val = 'o', .name = "ost_list", .has_arg = required_argument },
#endif
{ .val = 'p', .name = "pool", .has_arg = required_argument },
/* find { .val = 'P', .name = "print", .has_arg = no_argument }, */
/* getstripe { .val = 'q', .name = "quiet", .has_arg = no_argument }, */
/* getstripe { .val = 'R', .name = "raw", .has_arg = no_argument }, */
{ .val = 'S', .name = "stripe-size", .has_arg = required_argument },
{ .val = 'S', .name = "stripe_size", .has_arg = required_argument },
/* find { .val = 't', .name = "type", .has_arg = required_argument }*/
/* dirstripe { .val = 'T', .name = "mdt-count", .has_arg = required_argument }*/
/* find { .val = 'u', .name = "uid", .has_arg = required_argument }*/
/* find { .val = 'U', .name = "user", .has_arg = required_argument }*/
/* --verbose is only valid in migrate mode */
{ .val = 'v', .name = "verbose", .has_arg = no_argument},
{ .val = 'W', .name = "bandwidth-limit", .has_arg = required_argument},
{ .val = 'x', .name = "xattr", .has_arg = required_argument },
{ .val = 'X', .name = "max-free", .has_arg = required_argument},
/* dirstripe { .val = 'X',.name = "max-inherit",.has_arg = required_argument }*/
{ .val = 'y', .name = "yaml", .has_arg = required_argument},
{ .val = 'z', .name = "ext-size", .has_arg = required_argument},
{ .val = 'z', .name = "extension-size", .has_arg = required_argument},
{ .val = LFS_MIGRATE_NOFIX, .name = "clear-fixed", .has_arg = no_argument},
{ .name = NULL } };
setstripe_args_init(&lsa);
migrate_mode = (opc == SO_MIGRATE);
mirror_mode = (opc == SO_MIRROR_CREATE || opc == SO_MIRROR_EXTEND);
setstripe_mode = (opc == SO_SETSTRIPE);
if (opc == SO_MIRROR_DELETE) {
delete = 1;
mirror_flags = MF_DESTROY;
}
snprintf(cmd, sizeof(cmd), "%s %s", progname, argv[0]);
progname = cmd;
/* pre-allocate buf */
bufsize = PATH_MAX;
buf = malloc(bufsize);
if (buf == NULL) {
result = -errno;
fprintf(stderr, "%s %s: failed to allocate memory\n",
progname, argv[0]);
goto error;
}
while ((c = getopt_long(argc, argv,
"0Abc:C:dDE:f:FhH:i:I:K:m:M:N::no:p:L:s:S:vx:X:W:y:z:",
long_opts, &long_idx)) >= 0) {
size_units = 1;
switch (c) {
case 0:
/* Long options. */
break;
case LFS_COMP_ADD_OPT:
comp_add = 1;
break;
case LFS_COMP_DEL_OPT:
comp_del = 1;
break;
case LFS_COMP_FLAGS_OPT:
result = comp_str2flags(optarg, &lsa.lsa_comp_flags,
&lsa.lsa_comp_neg_flags);
if (result != 0)
goto usage_error;
if (mirror_mode && lsa.lsa_comp_neg_flags) {
fprintf(stderr,
"%s: inverted flags are not supported\n",
progname);
goto usage_error;
}
break;
case LFS_COMP_SET_OPT:
comp_set = 1;
break;
case LFS_COMP_NO_VERIFY_OPT:
mirror_flags |= MF_NO_VERIFY;
break;
case LFS_MIRROR_ID_OPT: {
unsigned long int id;
errno = 0;
id = strtoul(optarg, &end, 0);
if (errno != 0 || *end != '\0' || id == 0 ||
id > UINT16_MAX) {
fprintf(stderr,
"%s %s: invalid mirror ID '%s'\n",
progname, argv[0], optarg);
goto usage_error;
}
mirror_id = (__u16)id;
break;
}
case LFS_LAYOUT_FLAGS_OPT: {
uint32_t neg_flags;
/* check for numeric flags (foreign and mirror cases) */
if (setstripe_mode && !mirror_mode && !last_mirror) {
errno = 0;
flags = strtoul(optarg, &end, 16);
if (errno != 0 || *end != '\0' ||
flags >= UINT32_MAX) {
fprintf(stderr,
"%s %s: invalid hex flags '%s'\n",
progname, argv[0], optarg);
return CMD_HELP;
}
if (!foreign_mode) {
fprintf(stderr,
"%s %s: hex flags must be specified with --foreign option\n",
progname, argv[0]);
return CMD_HELP;
}
break;
}
if (!mirror_mode || !last_mirror) {
fprintf(stderr,
"error: %s: --flags must be specified with --mirror-count|-N option\n",
progname);
goto usage_error;
}
result = comp_str2flags(optarg, &last_mirror->m_flags,
&neg_flags);
if (result != 0)
goto usage_error;
if (neg_flags) {
fprintf(stderr,
"%s: inverted flags are not supported\n",
progname);
result = -EINVAL;
goto usage_error;
}
if (last_mirror->m_flags & ~LCME_USER_MIRROR_FLAGS) {
fprintf(stderr,
"%s: unsupported mirror flags: %s\n",
progname, optarg);
result = -EINVAL;
goto error;
}
break;
}
case LFS_RESTRIPE_OPT:
if (auto_stripe)
goto restripe_error;
restripe = true;
break;
case LFS_LAYOUT_FOREIGN_OPT:
if (restripe)
goto restripe_error;
if (optarg) {
/* check pure numeric */
type = strtoul(optarg, &end, 0);
if (*end) {
/* check name */
type = check_foreign_type_name(optarg);
if (type == LU_FOREIGN_TYPE_UNKNOWN) {
fprintf(stderr,
"%s %s: unrecognized foreign type '%s'\n",
progname, argv[0],
optarg);
return CMD_HELP;
}
} else if (type >= UINT32_MAX) {
fprintf(stderr,
"%s %s: invalid foreign type '%s'\n",
progname, argv[0], optarg);
return CMD_HELP;
}
}
foreign_mode = true;
break;
case LFS_MODE_OPT:
mode_opt = optarg;
if (mode_opt) {
mode = strtoul(mode_opt, &end, 8);
if (*end != '\0') {
fprintf(stderr,
"%s %s: bad mode '%s'\n",
progname, argv[0], mode_opt);
return CMD_HELP;
}
previous_umask = umask(0);
}
break;
case LFS_LAYOUT_COPY:
if (restripe)
goto restripe_error;
from_copy = true;
template = optarg;
break;
case LFS_STATS_OPT:
stats_interval_sec = 5;
break;
case LFS_STATS_INTERVAL_OPT:
stats_interval_sec = strtol(optarg, &end, 0);
if (stats_interval_sec == 0 && errno) {
fprintf(stderr,
"%s %s: invalid stats interval %s\n",
progname, argv[0], optarg);
goto usage_error;
}
break;
case LFS_MIGRATE_NOFIX:
if (!migrate_mode) {
fprintf(stderr,
"%s %s: --clear-fixed valid only for migrate command\n",
progname, argv[0]);
goto usage_error;
}
clear_hash_fixed = true;
break;
case LFS_FILES_FROM:
files_from = optarg;
break;
case LFS_LUSTRE_DIR:
lustre_dir = optarg;
break;
case '0':
null_mode = true;
break;
case 'A':
if (restripe)
goto restripe_error;
auto_stripe = true;
break;
case 'b':
if (!migrate_mode) {
fprintf(stderr,
"%s %s: -b|--block valid only for migrate command\n",
progname, argv[0]);
goto usage_error;
}
migration_block = true;
break;
case 'C':
if (lsa.lsa_pattern == LLAPI_LAYOUT_MDT) {
fprintf(stderr,
"%s %s: -C|--overstripe-count incompatible with DoM layout\n",
progname, argv[0]);
goto usage_error;
}
overstriped = true;
lsa.lsa_pattern = LLAPI_LAYOUT_OVERSTRIPING;
fallthrough;
case 'c':
if (auto_stripe || restripe) {
fprintf(stderr,
"%s %s: -c|--stripe-count incompatible with -A|auto-stripe or --restripe\n",
progname, argv[0]);
goto usage_error;
}
errno = 0;
lsa.lsa_stripe_count = strtoul(optarg, &end, 0);
/* only allow count -2..-32 for overstriped files */
if (errno != 0 || *end != '\0' || optarg == end ||
lsa.lsa_stripe_count <
(overstriped ? LLAPI_OVERSTRIPE_COUNT_MAX :
LLAPI_OVERSTRIPE_COUNT_MIN) ||
lsa.lsa_stripe_count > LOV_MAX_STRIPE_COUNT) {
fprintf(stderr,
"%s %s: invalid stripe count '%s'\n",
progname, argv[0], optarg);
goto usage_error;
}
if (lsa.lsa_stripe_count <=
LLAPI_OVERSTRIPE_COUNT_MIN &&
lsa.lsa_stripe_count >=
LLAPI_OVERSTRIPE_COUNT_MAX) {
lsa.lsa_stripe_count = LLAPI_LAYOUT_WIDE_MIN -
(lsa.lsa_stripe_count + 1);
}
stripe_count_set = true;
break;
case 'd':
if (migrate_mode) {
migrate_mdt_param.fp_max_depth = 1;
} else {
/* delete the default striping pattern */
delete = 1;
if (opc == SO_MIRROR_SPLIT) {
if (has_m_file) {
fprintf(stderr,
"%s %s: -d cannot used with -f\n",
progname, argv[0]);
goto usage_error;
}
mirror_flags |= MF_DESTROY;
}
}
break;
case 'D':
if (!migrate_mode && !mirror_mode) {
fprintf(stderr,
"%s %s: -D|--non-direct is valid only for migrate or mirror command\n",
progname, argv[0]);
goto usage_error;
}
migration_flags |= LLAPI_MIGRATION_NONDIRECT;
break;
case 'E':
if (auto_stripe || restripe) {
fprintf(stderr,
"%s %s: -E|--component-end incompatible with -A|--auto-stripe or --restripe\n",
progname, argv[0]);
goto usage_error;
}
if (lsa.lsa_comp_end != 0) {
result = comp_args_to_layout(lpp, &lsa, true);
if (result) {
fprintf(stderr, "%s: invalid layout\n",
progname);
goto usage_error;
}
lsa.lsa_comp_count++;
setstripe_args_init_inherit(&lsa);
}
if (arg_is_eof(optarg)) {
lsa.lsa_comp_end = LUSTRE_EOF;
} else {
result = llapi_parse_size(optarg,
&lsa.lsa_comp_end,
&size_units, 0);
/* assume units of KB if too small */
if (lsa.lsa_comp_end < 4096)
lsa.lsa_comp_end *= 1024;
if (result ||
lsa.lsa_comp_end & (LOV_MIN_STRIPE_SIZE - 1)) {
fprintf(stderr,
"%s %s: invalid component end '%s'\n",
progname, argv[0], optarg);
goto usage_error;
}
}
comp_end_set = true;
break;
case LFS_EC_STRIPE_COUNT_OPT:
case LFS_EC_EXPERT_OPT: {
uint8_t data_count, parity_count;
struct ec_spec *new_specs;
bool ec_expert = (c == LFS_EC_EXPERT_OPT);
/* Parse EC stripe count */
result = parse_ec_stripe_count(optarg, &data_count,
&parity_count,
ec_expert);
if (result)
goto usage_error;
/* Allocate or expand EC specs array */
new_specs = realloc(lsa.lsa_ec_specs,
(lsa.lsa_ec_spec_count + 1) *
sizeof(struct ec_spec));
if (!new_specs) {
fprintf(stderr,
"%s %s: cannot allocate EC spec array: %s\n",
progname, argv[0],
strerror(ENOMEM));
result = -ENOMEM;
goto error;
}
lsa.lsa_ec_specs = new_specs;
lsa.lsa_ec_specs[lsa.lsa_ec_spec_count].comp_index =
lsa.lsa_comp_count;
lsa.lsa_ec_specs[lsa.lsa_ec_spec_count].data_count =
data_count;
lsa.lsa_ec_specs[lsa.lsa_ec_spec_count].parity_count =
parity_count;
lsa.lsa_ec_spec_count++;
lsa.lsa_has_ec = true;
break;
}
case 'F':
fid_mode = true;
break;
case 'H':
if (!migrate_mode) {
fprintf(stderr,
"--mdt-hash is valid only for migrate command\n");
return CMD_HELP;
}
lsa.lsa_hash = check_hashtype(optarg);
if (lsa.lsa_hash == LMV_HASH_TYPE_UNKNOWN) {
fprintf(stderr,
"%s %s: bad stripe hash type '%s'\n",
progname, argv[0], optarg);
return CMD_HELP;
}
break;
case 'i':
errno = 0;
lsa.lsa_stripe_off = strtol(optarg, &end, 0);
if (errno != 0 || *end != '\0' || optarg == end ||
lsa.lsa_stripe_off < (__s16)LOV_ALL_STRIPES ||
lsa.lsa_stripe_off > LOV_V1_INSANE_STRIPE_INDEX) {
fprintf(stderr,
"%s %s: invalid stripe offset '%s'\n",
progname, argv[0], optarg);
goto usage_error;
}
if (lsa.lsa_stripe_off == -1)
lsa.lsa_stripe_off = LLAPI_LAYOUT_DEFAULT;
break;
case 'I':
comp_id = strtoul(optarg, &end, 0);
if (*end != '\0' || comp_id == 0 ||
comp_id > LCME_ID_MAX) {
fprintf(stderr,
"%s %s: invalid component ID '%s'\n",
progname, argv[0], optarg);
goto usage_error;
}
break;
case 'f':
if (opc != SO_MIRROR_EXTEND && opc != SO_MIRROR_SPLIT) {
fprintf(stderr,
"error: %s: invalid option: %s\n",
progname, argv[optopt + 1]);
goto usage_error;
}
if (opc == SO_MIRROR_EXTEND) {
if (!last_mirror) {
fprintf(stderr,
"error: %s: '-N' must exist in front of '%s'\n",
progname, argv[optopt + 1]);
goto usage_error;
}
last_mirror->m_file = optarg;
last_mirror->m_count = 1;
} else {
/* mirror split */
if (!mirror_list)
mirror_list = lfs_mirror_alloc();
mirror_list->m_file = optarg;
}
has_m_file = true;
break;
case 'K':
cap = atoi(optarg);
cap_set = true;
if (cap <= 0) {
fprintf(stderr,
"%s %s: invalid -K|--auto-cap value '%s'\n",
progname, argv[0], optarg);
goto usage_error;
}
break;
case 'L':
if (restripe)
goto restripe_error;
if (strcmp(argv[optind - 1], "mdt") == 0) {
/* Can be only the first component */
if (layout) {
result = -EINVAL;
fprintf(stderr,
"error: 'mdt' layout can be only the first one\n");
goto error;
}
if (lsa.lsa_comp_end > (1ULL << 30)) { /* 1Gb */
result = -EFBIG;
fprintf(stderr,
"error: 'mdt' layout size is too big\n");
goto error;
}
lsa.lsa_pattern = LLAPI_LAYOUT_MDT;
lsa.lsa_stripe_size = LLAPI_LAYOUT_DEFAULT;
} else if (strcmp(argv[optind - 1], "raid0") != 0) {
result = -EINVAL;
fprintf(stderr,
"error: layout '%s' is unknown, supported layouts are: 'mdt', 'raid0'\n",
argv[optind]);
goto error;
}
break;
case 'm':
if (!migrate_mode) {
fprintf(stderr,
"%s %s: -m|--mdt-index is valid only for migrate command\n",
progname, argv[0]);
goto usage_error;
}
migrate_mdt_mode = true;
lsa.lsa_nr_tgts = parse_targets(tgts,
sizeof(tgts) / sizeof(__u32),
lsa.lsa_nr_tgts, optarg,
&overstriped);
if (lsa.lsa_nr_tgts < 0) {
fprintf(stderr,
"%s: invalid MDT target(s) '%s'\n",
progname, optarg);
goto usage_error;
}
lsa.lsa_tgts = tgts;
if (lsa.lsa_stripe_off == LLAPI_LAYOUT_DEFAULT)
lsa.lsa_stripe_off = tgts[0];
break;
case 'M':
min_free = atoi(optarg);
min_free_set = true;
if (min_free < 0) {
fprintf(stderr,
"%s %s: invalid min-free value '%s'\n",
progname, argv[0], optarg);
goto usage_error;
}
break;
case 'n':
if (!migrate_mode) {
fprintf(stderr,
"%s %s: -n|--non-block valid only for migrate command\n",
progname, argv[0]);
goto usage_error;
}
migration_flags |= LLAPI_MIGRATION_NONBLOCK;
break;
case 'N':
if (restripe)
goto restripe_error;
create_mirror:
if (opc == SO_SETSTRIPE) {
opc = SO_MIRROR_CREATE;
mirror_mode = true;
}
mirror_total_mode = false;
if (optarg) {
if (optarg[0] == '=') {
mirror_total_mode = true;
optarg++; /* skip '=' */
}
errno = 0;
mirror_count = strtoul(optarg, &end, 0);
if (errno != 0 || *end != '\0' ||
mirror_count == 0 ||
mirror_count > LUSTRE_MIRROR_COUNT_MAX) {
fprintf(stderr,
"error: %s: bad mirror count: %s\n",
progname, optarg);
result = -EINVAL;
goto error;
}
} else if (optind < argc && (*argv[optind] >= '0') &&
(*argv[optind] <= '9')) {
errno = 0;
mirror_count = strtoul(argv[optind], &end, 0);
if (errno != 0 || *end != '\0' ||
mirror_count == 0 ||
mirror_count > LUSTRE_MIRROR_COUNT_MAX) {
fprintf(stderr,
"error: %s: bad mirror count: %s\n",
progname, argv[optind]);
result = -EINVAL;
goto error;
}
optind++;
} else {
mirror_count = 1;
}
if (mirror_total_mode) {
char *path = argv[argc - 1];
struct llapi_layout *f_layout;
uint16_t have_mirrors;
f_layout = llapi_layout_get_by_path(path, 0);
if (!f_layout) {
result = -errno;
fprintf(stderr,
"error: %s: cannot get layout from %s: %s\n",
progname, path,
strerror(errno));
goto error;
}
llapi_layout_mirror_count_get(f_layout,
&have_mirrors);
llapi_layout_free(f_layout);
if (have_mirrors >= mirror_count)
mirror_count = 0;
else
mirror_count -= have_mirrors;
if (mirror_count == 0) {
fprintf(stderr,
"warning: '%s' already has %d mirrors, no new mirrors will be created\n",
path, have_mirrors);
break;
}
}
new_mirror = lfs_mirror_alloc();
new_mirror->m_count = mirror_count;
if (!mirror_list)
mirror_list = new_mirror;
if (last_mirror) {
/* wrap up last mirror */
if (!setstripe_args_specified(&lsa))
last_mirror->m_inherit = true;
if (lsa.lsa_comp_end == 0)
lsa.lsa_comp_end = LUSTRE_EOF;
result = comp_args_to_layout(lpp, &lsa, true);
if (result) {
lfs_mirror_free(new_mirror);
goto error;
}
lsa.lsa_comp_count++;
/* Create EC parity mirror if this mirror has EC */
if (lsa.lsa_has_ec &&
lsa.lsa_ec_spec_count > 0) {
result = create_ec_parity_mirror(
last_mirror->m_layout,
lsa.lsa_ec_specs,
lsa.lsa_ec_spec_count);
if (result) {
fprintf(stderr,
"error: %s: failed to create EC parity mirror\n",
progname);
lfs_mirror_free(new_mirror);
goto error;
}
/* Mark that this layout already contains
* EC parity components. The layout now
* contains 2 mirrors (data + parity).
* We keep m_count at its original value
* (e.g., 1 for -N, 2 for -N2, etc.), and
* set m_has_ec=true so that mirror_create()
* knows to multiply the mirror count by 2.
*/
last_mirror->m_has_ec = true;
/* Clear EC specs for next mirror */
free(lsa.lsa_ec_specs);
lsa.lsa_ec_specs = NULL;
lsa.lsa_ec_spec_count = 0;
lsa.lsa_has_ec = false;
}
last_mirror->m_next = new_mirror;
setstripe_args_init_inherit(&lsa);
}
last_mirror = new_mirror;
lpp = &last_mirror->m_layout;
break;
case 'o':
#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(3, 0, 53, 0)
if (strcmp(argv[optind - 1], "--ost-list") == 0)
fprintf(stderr,
"warning: '--ost-list' is deprecated, use '--ost' instead\n");
#endif
if (restripe)
goto restripe_error;
if (lsa.lsa_pattern == LLAPI_LAYOUT_MDT) {
fprintf(stderr,
"%s %s: -o|--ost incompatible with DoM layout\n",
progname, argv[0]);
goto usage_error;
}
/*
* -o allows overstriping, and must note it because
* parse_targets is shared with MDT striping, which
* does not allow duplicates
*/
lsa.lsa_nr_tgts = parse_targets(tgts,
sizeof(tgts) / sizeof(__u32),
lsa.lsa_nr_tgts, optarg,
&overstriped);
if (lsa.lsa_nr_tgts < 0) {
fprintf(stderr,
"%s %s: invalid OST target(s) '%s'\n",
progname, argv[0], optarg);
goto usage_error;
}
if (overstriped)
lsa.lsa_pattern = LLAPI_LAYOUT_OVERSTRIPING;
lsa.lsa_tgts = tgts;
if (lsa.lsa_stripe_off == LLAPI_LAYOUT_DEFAULT)
lsa.lsa_stripe_off = tgts[0];
break;
case 'p':
if (restripe)
goto restripe_error;
if (!optarg)
goto usage_error;
if (optarg[0] == '\0' || lov_pool_is_inherited(optarg))
lsa.lsa_pool_name = NULL;
else
lsa.lsa_pool_name = optarg;
break;
case 'S':
if (restripe)
goto restripe_error;
result = llapi_parse_size(optarg, &lsa.lsa_stripe_size,
&size_units, 0);
/* assume units of KB if too small to be valid */
if (lsa.lsa_stripe_size < 4096)
lsa.lsa_stripe_size *= 1024;
if (result ||
lsa.lsa_stripe_size & (LOV_MIN_STRIPE_SIZE - 1)) {
fprintf(stderr,
"%s %s: invalid stripe size '%s'\n",
progname, argv[0], optarg);
goto usage_error;
}
break;
case 'v':
if (!migrate_mode) {
fprintf(stderr,
"%s %s: -v|--verbose valid only for migrate command\n",
progname, argv[0]);
goto usage_error;
}
migrate_mdt_param.fp_verbose = VERBOSE_DETAIL;
migration_flags |= LLAPI_MIGRATION_VERBOSE;
break;
case 'x':
xattr = optarg;
break;
case 'W':
if (!migrate_mode && !mirror_mode) {
fprintf(stderr,
"--bandwidth is valid only for migrate and mirror mode\n");
goto error;
}
if (llapi_parse_size(optarg, &bandwidth_bytes_sec,
&bandwidth_unit, 0) < 0) {
fprintf(stderr,
"error: %s: bad value for bandwidth '%s'\n",
argv[0], optarg);
goto error;
}
break;
case 'X':
max_free = atoi(optarg);
if (max_free < 0) {
fprintf(stderr,
"%s %s: invalid max-free value '%s'\n",
progname, argv[0], optarg);
goto usage_error;
}
break;
case 'y':
if (restripe)
goto restripe_error;
from_yaml = true;
template = optarg;
break;
case 'z':
result = llapi_parse_size(optarg,
&lsa.lsa_extension_size,
&size_units, 0);
if (result) {
fprintf(stderr,
"%s %s: invalid extension size '%s'\n",
progname, argv[0], optarg);
goto usage_error;
}
lsa.lsa_extension_comp = true;
break;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
case 'h':
goto usage_error;
}
}
fname = argv[optind];
if (cap_set && (!migrate_mode || !auto_stripe)) {
fprintf(stderr,
"%s %s: -K|--auto-cap valid only for migrate command with --auto-stripe\n",
progname, argv[0]);
goto usage_error;
}
if (max_free > 0 && (!migrate_mode || !auto_stripe)) {
fprintf(stderr,
"%s %s: --max-free valid only for migrate command with --auto-stripe\n",
progname, argv[0]);
goto usage_error;
}
if (min_free_set && (!migrate_mode || !auto_stripe)) {
fprintf(stderr,
"%s %s: --min-free valid only for migrate command with --auto-stripe\n",
progname, argv[0]);
goto usage_error;
}
if (stripe_count_set && auto_stripe) {
fprintf(stderr,
"%s %s: -c|--stripe-count incompatible with -A|--auto-stripe\n",
progname, argv[0]);
goto usage_error;
}
if (comp_end_set && auto_stripe) {
fprintf(stderr,
"%s %s: -E|--component-end incompatible with -A|--auto-stripe\n",
progname, argv[0]);
goto usage_error;
}
/* for 'lfs migrate' and 'lfs mirror extend' command,
*
* at least one of FILE/--null/--files-from=LIST_FILE must be specified.
* If both --null and --files-from=LIST_FILE are specified, read
* filenames from LIST_FILE and use '\0' as delimiter.
*
* --lustre-dir= and --fid may be specified at the same time, so that
* FID is provided on command line or file/stdin instead of file path.
*/
if (opc == SO_MIGRATE || opc == SO_MIRROR_EXTEND) {
int num = 0;
if (optind < argc)
num++;
if (null_mode) {
if (files_from_fp == NULL)
files_from_fp = stdin;
delim = 0;
num++;
}
if (files_from != NULL) {
if (strcmp("-", files_from) == 0)
files_from_fp = stdin;
else
files_from_fp = fopen(files_from, "r");
if (files_from_fp == NULL) {
result = -errno;
fprintf(stderr, "%s %s: failed to open filelist file '%s'\n",
progname, argv[0], files_from);
goto error;
}
num++;
}
if (num < 1) {
fprintf(stderr, "%s %s: at least one of FILE/--null/--files-from=LIST_FILE must be specified\n",
progname, argv[0]);
goto usage_error;
}
/* check fid mode is set correctly */
if (fid_mode) {
int rc;
struct lu_fid fid;
if (lustre_dir == NULL) {
rc = guess_only_lustre_mount_root(mntdir);
if (rc != 0) {
fprintf(stderr,
"%s %s: not able to guess lustre mount point, please specify --lustre-dir option\n",
progname, argv[0]);
goto usage_error;
}
lustre_dir = mntdir;
}
lustre_dir_fd = open(lustre_dir, O_RDONLY);
if (lustre_dir_fd < 0) {
result = -errno;
fprintf(stderr,
"%s %s: failed to open lustre dir: %s\n",
progname, argv[0], lustre_dir);
goto error;
}
rc = llapi_fd2fid(lustre_dir_fd, &fid);
if (rc < 0) {
result = rc;
fprintf(stderr,
"%s %s: path is not in a lustre file system: %s\n",
progname, argv[0], lustre_dir);
goto error;
}
}
} else if (optind == argc) {
fprintf(stderr, "%s %s: FILE must be specified\n",
progname, argv[0]);
goto usage_error;
}
/* lfs migrate $filename should keep the file's layout by default */
if (migrate_mode && !layout && !from_yaml && !restripe &&
!auto_stripe && !setstripe_args_specified(&lsa) &&
!lsa.lsa_pool_name)
from_copy = true;
if (xattr && !foreign_mode) {
/*
* only print a warning as this is harmless and will be ignored
*/
fprintf(stderr,
"%s %s: xattr has been specified for non-foreign layout\n",
progname, argv[0]);
} else if (foreign_mode && !xattr) {
fprintf(stderr,
"%s %s: xattr must be provided in foreign mode\n",
progname, argv[0]);
goto usage_error;
}
if (foreign_mode && (!setstripe_mode || comp_add | comp_del ||
comp_set || comp_id || delete || from_copy ||
setstripe_args_specified(&lsa) || lsa.lsa_nr_tgts ||
lsa.lsa_tgts)) {
fprintf(stderr,
"%s %s: only --xattr/--flags/--mode options are valid with --foreign\n",
progname, argv[0]);
return CMD_HELP;
}
/* Validate EC options */
#if LUSTRE_VERSION_CODE > OBD_OCD_VERSION(2, 17, 90, 0)
#error "Remove LFS_EC_OK check for 2.18 release"
#endif
if (lsa.lsa_has_ec) {
static bool ec_warning_printed;
if (!ec_warning_printed &&
getenv("LFS_EC_OK") == NULL) {
fprintf(stderr,
"WARNING: Erasure coding implementation is incomplete and not yet intended for production use. This message will be removed once erasure coding is complete. To turn on erasure coding support, set LFS_EC_OK=yes to acknowledge this limitation.\n");
ec_warning_printed = true;
result = -EOPNOTSUPP;
goto error;
}
if (foreign_mode) {
fprintf(stderr,
"error: %s: --ec cannot be used with --foreign\n",
progname);
goto usage_error;
}
if (comp_add || comp_del || comp_set) {
fprintf(stderr,
"error: %s: --ec cannot be used with --component-add/del/set\n",
progname);
goto usage_error;
}
/*
* Automatically create a [0,EOF] component if --ec was
* specified without -E, similar to how compression and
* mirror addition work.
*/
if (lsa.lsa_comp_end == 0 && !mirror_mode)
lsa.lsa_comp_end = LUSTRE_EOF;
}
if (mirror_mode && (!mirror_total_mode || mirror_count)) {
if (mirror_count == 0)
goto create_mirror;
if (!setstripe_args_specified(&lsa))
last_mirror->m_inherit = true;
if (lsa.lsa_comp_end == 0)
lsa.lsa_comp_end = LUSTRE_EOF;
}
if (lsa.lsa_comp_end != 0) {
result = comp_args_to_layout(lpp, &lsa, true);
if (result) {
fprintf(stderr, "error: %s: invalid layout\n",
progname);
result = -EINVAL;
goto error;
}
lsa.lsa_comp_count++;
}
/* Create EC parity mirror if --ec without -N was specified */
if (lsa.lsa_has_ec && lsa.lsa_ec_spec_count > 0) {
if (mirror_mode && last_mirror) {
result = create_ec_parity_mirror(
last_mirror->m_layout,
lsa.lsa_ec_specs,
lsa.lsa_ec_spec_count);
if (result) {
fprintf(stderr,
"error: %s: failed to create EC parity mirror\n",
progname);
goto error;
}
/* Mark that this layout already contains EC parity
* components
*/
last_mirror->m_has_ec = true;
} else if (layout) {
/* In non-mirror mode, add parity mirror to layout */
uint16_t mirror_count = 0;
/* Verify layout doesn't already have multiple mirrors.
* A freshly created layout has mirror_count = 1.
*/
result = llapi_layout_mirror_count_get(layout,
&mirror_count);
if (result) {
fprintf(stderr,
"error: %s: failed to get mirror count: %s\n",
progname, strerror(errno));
goto error;
}
if (mirror_count > 1) {
fprintf(stderr,
"error: %s: layout already has %u mirrors, cannot add EC\n",
progname, mirror_count);
result = -EINVAL;
goto error;
}
result = create_ec_parity_mirror(layout,
lsa.lsa_ec_specs,
lsa.lsa_ec_spec_count);
if (result) {
fprintf(stderr,
"error: %s: failed to create EC parity mirror\n",
progname);
goto error;
}
/* Set mirror count to 2 (data + parity) */
result = llapi_layout_mirror_count_set(layout, 2);
if (result) {
fprintf(stderr,
"error: %s: failed to set mirror count: %s\n",
progname, strerror(errno));
goto error;
}
}
}
if (mirror_flags & MF_NO_VERIFY) {
if (opc != SO_MIRROR_EXTEND) {
fprintf(stderr,
"error: %s: --no-verify is valid only for lfs mirror extend command\n",
progname);
result = -EINVAL;
goto error;
} else if (!has_m_file) {
fprintf(stderr,
"error: %s: --no-verify must be specified with -f <victim_file> option\n",
progname);
result = -EINVAL;
goto error;
}
}
if (comp_set && !comp_id && !lsa.lsa_pool_name && !mirror_id) {
fprintf(stderr,
"%s %s: --component-set doesn't have component-id or mirror-id set\n",
progname, argv[0]);
goto usage_error;
}
if ((delete + comp_set + comp_del + comp_add) > 1) {
fprintf(stderr,
"%s %s: options --component-set, --component-del, --component-add and -d are mutually exclusive\n",
progname, argv[0]);
goto usage_error;
}
if (delete && (setstripe_args_specified(&lsa) || comp_id != 0 ||
lsa.lsa_comp_flags != 0 || layout != NULL)) {
fprintf(stderr,
"%s %s: option -d is mutually exclusive with -s, -c, -o, -p, -I, -F and -E options\n",
progname, argv[0]);
goto usage_error;
}
if ((comp_set || comp_del) &&
(setstripe_args_specified(&lsa) || layout != NULL)) {
fprintf(stderr,
"%s %s: options --component-del and --component-set are mutually exclusive when used with -c, -E, -o, -p, or -s\n",
progname, argv[0]);
goto usage_error;
}
if (comp_del && comp_id != 0 && lsa.lsa_comp_flags != 0) {
fprintf(stderr,
"%s %s: options -I and -F are mutually exclusive when used with --component-del\n",
progname, argv[0]);
goto usage_error;
}
if (comp_add || comp_del) {
struct stat st;
result = lstat(fname, &st);
if (result == 0 && S_ISDIR(st.st_mode)) {
fprintf(stderr,
"%s setstripe: cannot use --component-add or --component-del for directory\n",
progname);
goto usage_error;
}
if (mirror_mode) {
fprintf(stderr,
"error: %s: can't use --component-add or --component-del for mirror operation\n",
progname);
goto usage_error;
}
}
if (comp_add) {
if (!layout) {
fprintf(stderr,
"%s %s: option -E must be specified with --component-add\n",
progname, argv[0]);
goto usage_error;
}
}
if (from_yaml && from_copy) {
fprintf(stderr,
"%s: can't specify --yaml and --copy together\n",
progname);
goto error;
}
if ((from_yaml || from_copy) &&
(setstripe_args_specified(&lsa) || layout != NULL)) {
fprintf(stderr,
"error: %s: can't specify --yaml or --copy with -c, -S, -i, -o, -p or -E options.\n",
argv[0]);
goto error;
}
if ((migration_flags & LLAPI_MIGRATION_NONBLOCK) && migration_block) {
fprintf(stderr,
"%s %s: options --non-block and --block are mutually exclusive\n",
progname, argv[0]);
goto usage_error;
}
if (!comp_del && !comp_set && opc != SO_MIRROR_SPLIT &&
opc != SO_MIRROR_DELETE && comp_id != 0) {
fprintf(stderr,
"%s: option -I can only be used with --component-del or --component-set or lfs mirror split\n",
progname);
goto usage_error;
}
if (migrate_mdt_mode) {
struct lmv_user_md *lmu;
/* initialize migrate mdt parameters */
lmu = calloc(1, lmv_user_md_size(lsa.lsa_nr_tgts,
LMV_USER_MAGIC_SPECIFIC));
if (!lmu) {
fprintf(stderr,
"%s %s: cannot allocate memory for lmv_user_md: %s\n",
progname, argv[0], strerror(ENOMEM));
result = -ENOMEM;
goto error;
}
if (lsa.lsa_stripe_count != LLAPI_LAYOUT_DEFAULT)
lmu->lum_stripe_count = lsa.lsa_stripe_count;
if (lsa.lsa_stripe_off == LLAPI_LAYOUT_DEFAULT) {
fprintf(stderr,
"%s %s: migrate should specify MDT index\n",
progname, argv[0]);
free(lmu);
goto usage_error;
}
lmu->lum_stripe_offset = lsa.lsa_stripe_off;
lmu->lum_hash_type = lsa.lsa_hash;
if (overstriped)
lmu->lum_hash_type |= LMV_HASH_FLAG_OVERSTRIPED;
if (!clear_hash_fixed)
lmu->lum_hash_type |= LMV_HASH_FLAG_FIXED;
if (lsa.lsa_pool_name)
snprintf(lmu->lum_pool_name, sizeof(lmu->lum_pool_name),
"%s", lsa.lsa_pool_name);
if (lsa.lsa_nr_tgts > 1) {
int i;
if (lsa.lsa_stripe_count > 0 &&
lsa.lsa_stripe_count != LLAPI_LAYOUT_DEFAULT &&
lsa.lsa_stripe_count != lsa.lsa_nr_tgts) {
fprintf(stderr,
"error: %s: stripe count %lld doesn't match the number of MDTs: %d\n",
progname, lsa.lsa_stripe_count,
lsa.lsa_nr_tgts);
free(lmu);
goto usage_error;
}
lmu->lum_magic = LMV_USER_MAGIC_SPECIFIC;
lmu->lum_stripe_count = lsa.lsa_nr_tgts;
for (i = 0; i < lsa.lsa_nr_tgts; i++)
lmu->lum_objects[i].lum_mds = lsa.lsa_tgts[i];
} else {
lmu->lum_magic = LMV_USER_MAGIC;
}
migrate_mdt_param.fp_lmv_md = lmu;
migrate_mdt_param.fp_migrate = 1;
} else if (!layout && !from_yaml && !from_copy) {
if (lsa_args_stripe_count_check(&lsa))
goto usage_error;
/* initialize stripe parameters */
param = calloc(1, offsetof(typeof(*param),
lsp_osts[lsa.lsa_nr_tgts]));
if (!param) {
fprintf(stderr,
"%s %s: cannot allocate memory for parameters: %s\n",
progname, argv[0], strerror(ENOMEM));
result = -ENOMEM;
goto error;
}
if (lsa.lsa_stripe_size != LLAPI_LAYOUT_DEFAULT)
param->lsp_stripe_size = lsa.lsa_stripe_size;
if (lsa.lsa_stripe_count != LLAPI_LAYOUT_DEFAULT) {
if (lsa.lsa_stripe_count >= LLAPI_LAYOUT_WIDE_MIN &&
lsa.lsa_stripe_count <= LLAPI_LAYOUT_WIDE_MAX)
param->lsp_stripe_count =
LLAPI_LAYOUT_WIDE_MIN -
(lsa.lsa_stripe_count + 1);
else
param->lsp_stripe_count = lsa.lsa_stripe_count;
}
if (lsa.lsa_stripe_off == LLAPI_LAYOUT_DEFAULT)
param->lsp_stripe_offset = -1;
else
param->lsp_stripe_offset = lsa.lsa_stripe_off;
param->lsp_stripe_pattern =
llapi_pattern_to_lov(lsa.lsa_pattern);
if (param->lsp_stripe_pattern == EINVAL) {
fprintf(stderr, "error: %s: invalid stripe pattern\n",
argv[0]);
free(param);
goto usage_error;
}
param->lsp_pool = lsa.lsa_pool_name;
param->lsp_is_specific = false;
if (lsa.lsa_nr_tgts > 0) {
param->lsp_is_specific = true;
param->lsp_stripe_count = lsa.lsa_nr_tgts;
memcpy(param->lsp_osts, tgts,
sizeof(*tgts) * lsa.lsa_nr_tgts);
}
}
if (from_yaml) {
/* generate a layout from a YAML template */
result = lfs_comp_create_from_yaml(template, &layout,
&lsa, tgts);
if (result) {
fprintf(stderr,
"error: %s: can't create composite layout from template file %s\n",
argv[0], template);
goto error;
}
}
if (layout != NULL || mirror_list != NULL) {
if (mirror_list)
result = mirror_adjust_first_extents(mirror_list);
else
result = layout_adjust_first_extent(fname, layout,
comp_add);
if (result == -ENODATA)
comp_add = 0;
else if (result != 0) {
fprintf(stderr, "error: %s: invalid layout\n",
progname);
goto error;
}
}
while (true) {
if (files_from_fp == NULL) {
/* file names from arguments */
fname = argv[optind++];
if (optind > argc || fname == NULL)
break;
} else {
/* file names from file/stdin */
ssize_t len;
errno = 0;
len = getdelim(&buf, &bufsize, delim, files_from_fp);
if (len == -1) {
if (errno != 0) { /* error */
result = -errno;
fprintf(stderr, "%s %s: failed to read from list file\n",
progname, argv[0]);
goto error;
} else { /* EOF */
break;
}
}
/* remove possible trailing '\n' */
if (buf[len - 1] == '\n')
buf[len - 1] = '\0';
/* skip empty line or comment line starting with # */
if (strlen(buf) == 0 || buf[0] == '#')
continue;
fname = buf;
}
/* if fid_mode is true, fname is actually fid string,
* parse the fid and fill real fname
*/
if (fid_mode) {
int len;
char *endptr = NULL;
result = llapi_fid_parse(fname, &fid, &endptr);
if (result) {
fprintf(stderr,
"%s %s: invalid FID: %s\n",
progname, argv[0], fname);
if (result2 == 0)
result2 = result;
continue;
}
strncpy(buf, lustre_dir, PATH_MAX);
buf[PATH_MAX - 2] = '\0';
len = strlen(buf);
if (lustre_dir[len - 1] != '/') {
buf[len] = '/';
buf[len + 1] = '\0';
len++;
}
result = llapi_fid2path_at(lustre_dir_fd, &fid,
buf + len, bufsize - len,
NULL, NULL);
if (result < 0) {
fprintf(stderr,
"%s %s: failed to lookup path for FID:" DFID_NOBRACE "\n",
progname, argv[0], PFID(&fid));
/* record the errno and continue for next FID */
if (result2 == 0)
result2 = result;
continue;
}
llapi_printf(migration_flags & LLAPI_MIGRATION_VERBOSE ?
LLAPI_MSG_NORMAL : LLAPI_MSG_DEBUG,
DFID_NOBRACE " => %s\n",
PFID(&fid), buf);
fname = buf;
}
/* Handle restripe */
if (restripe) {
llapi_layout_free(layout);
layout = NULL;
free(param);
param = NULL;
setstripe_args_init(&lsa);
} else if (from_copy) {
layout = layout_get_by_name_or_fid(template ?: fname,
fname, 0, O_RDONLY);
if (!layout) {
fprintf(stderr,
"%s: can't create composite layout from file %s: %s\n",
progname, template ?: fname,
strerror(errno));
result = -errno;
goto error;
}
}
if (migrate_mdt_mode) {
result = llapi_migrate_mdt(fname, &migrate_mdt_param);
} else if (migrate_mode) {
if (from_copy) {
/*
* Strip the source layout of specific
* OST object/index values.
*/
result = llapi_layout_ost_index_reset(layout);
if (result) {
fprintf(stderr,
"%s: set default ost index failed: %s\n",
progname, strerror(errno));
result = -errno;
goto error;
}
} else if (auto_stripe) {
struct stat st;
int calc_stripe_count;
result = stat(fname, &st);
if (result) {
fprintf(stderr,
"%s: cannot stat file '%s': %s\n",
progname, fname,
strerror(errno));
result = -errno;
goto error;
}
/*
* obj_max_kb is passed by reference to allow
* calc_stripe() to calculate and return the
* maximum object size per OST for subsequent
* calls to this function.
*/
calc_stripe_count =
calc_stripe(fname, &st.st_size,
&obj_max_kb, cap, min_free,
max_free,
lsa.lsa_pool_name);
if (calc_stripe_count < 0) {
fprintf(stderr,
"%s: auto-striping failed for '%s'\n",
progname, fname);
result = -EINVAL;
goto error;
}
/* Update stripe count in param */
param->lsp_stripe_count = calc_stripe_count;
if (layout) {
llapi_layout_free(layout);
layout = NULL;
}
if (migration_flags & LLAPI_MIGRATION_VERBOSE)
printf("%s: auto-striping set stripe count to %d\n",
fname, calc_stripe_count);
}
result = lfs_migrate(fname, migration_flags, param,
layout, bandwidth_bytes_sec,
stats_interval_sec);
} else if (comp_set != 0) {
if (mirror_id != 0)
result = lfs_component_set_by_mirror(fname,
mirror_id,
lsa.lsa_comp_flags,
lsa.lsa_comp_neg_flags);
else
result = lfs_component_set(fname, comp_id,
lsa.lsa_pool_name,
lsa.lsa_comp_flags,
lsa.lsa_comp_neg_flags);
} else if (comp_del != 0) {
result = lfs_component_del(fname, comp_id,
lsa.lsa_comp_flags,
lsa.lsa_comp_neg_flags);
} else if (comp_add != 0) {
result = lfs_component_add(fname, layout);
} else if (opc == SO_MIRROR_CREATE) {
result = mirror_create(fname, mirror_list);
} else if (opc == SO_MIRROR_EXTEND) {
result = mirror_extend(fname, mirror_list,
mirror_flags,
migration_flags,
bandwidth_bytes_sec,
stats_interval_sec);
} else if (opc == SO_MIRROR_SPLIT || opc == SO_MIRROR_DELETE) {
if (!mirror_id && !comp_id && !lsa.lsa_pool_name) {
fprintf(stderr,
"%s: no mirror id, component id, or pool name specified to delete from '%s'\n",
progname, fname);
goto usage_error;
}
if (lsa.lsa_pool_name)
mirror_flags |= MF_COMP_POOL;
else if (mirror_id != 0)
comp_id = mirror_id;
else
mirror_flags |= MF_COMP_ID;
if (has_m_file && !strcmp(fname, mirror_list->m_file)) {
fprintf(stderr,
"%s: the file specified by -f cannot be same as the source file '%s'\n",
progname, fname);
goto usage_error;
}
/* If the mirror is the only non-stale mirror,
* do resync before mirror_split().
*/
result = 0;
if (!layout)
layout = layout_get_by_name_or_fid(template ?:
fname, fname, 0, O_RDONLY);
if (last_non_stale_mirror(mirror_id, layout)) {
struct ll_ioc_lease *ioc = NULL;
ioc = calloc(1, sizeof(*ioc) +
sizeof(__u32) * IOC_IDS_MAX);
if (ioc) {
result = lfs_mirror_resync_file(fname,
ioc, NULL, 0,
stats_interval_sec,
bandwidth_bytes_sec,
false);
if (result)
fprintf(stderr,
"Cannot resync file\n");
free(ioc);
}
}
if (!result)
result = mirror_split(fname, comp_id,
lsa.lsa_pool_name, mirror_flags,
has_m_file ?
mirror_list->m_file : NULL);
} else if (layout) {
result = lfs_component_create(fname, O_CREAT | O_WRONLY,
mode, layout);
if (result >= 0) {
close(result);
result = 0;
}
} else if (foreign_mode) {
result = llapi_file_create_foreign(fname, mode, type,
flags, xattr);
if (result >= 0) {
close(result);
result = 0;
}
} else {
result = llapi_file_open_param(fname,
O_CREAT | O_WRONLY,
mode, param);
if (result >= 0) {
close(result);
result = 0;
}
}
if (result) {
/* Save the first error encountered. */
if (result2 == 0)
result2 = result;
continue;
}
}
if (mode_opt)
umask(previous_umask);
free(param);
free(migrate_mdt_param.fp_lmv_md);
llapi_layout_free(layout);
lfs_mirror_list_free(mirror_list);
return result2;
restripe_error:
{
int idx = -1;
char optbuf[128];
if (c < 256) {
for (int i = 0; long_opts[i].name != NULL; i++) {
if (long_opts[i].val == c) {
idx = i;
break;
}
}
} else {
idx = long_idx;
}
if (idx >= 0) {
if (c < 256)
snprintf(optbuf, sizeof(optbuf),
"-%c|--%s", c, long_opts[idx].name);
else
snprintf(optbuf, sizeof(optbuf),
"--%s", long_opts[idx].name);
fprintf(stderr,
"%s %s: %s incompatible with --restripe\n",
progname, argv[0], optbuf);
}
goto usage_error;
}
usage_error:
result = CMD_HELP;
error:
free(param);
llapi_layout_free(layout);
lfs_mirror_list_free(mirror_list);
if (files_from_fp != NULL && files_from_fp != stdin)
fclose(files_from_fp);
if (lustre_dir_fd >= 0)
close(lustre_dir_fd);
free(buf);
free(lsa.lsa_ec_specs);
return result;
}
static int lfs_poollist(int argc, char **argv)
{
if (argc != 2)
return CMD_HELP;
return llapi_poollist(argv[1]);
}
#define FP_DEFAULT_TIME_MARGIN (24 * 60 * 60)
static int set_time(struct find_param *param, time_t *time, time_t *set_t,
int *sign_t, char *str)
{
long long t = 0;
int sign = 0;
char *endptr = "AD";
char *timebuf;
time_t prev_set = 0;
int prev_sign = 0;
if (str[0] == '+')
sign = 1;
else if (str[0] == '-')
sign = -1;
if (sign)
str++;
if (*set_t) {
prev_set = *set_t;
prev_sign = *sign_t;
}
for (timebuf = str; *endptr && *(endptr + 1); timebuf = endptr + 1) {
long long val = strtoll(timebuf, &endptr, 0);
int unit = 1;
switch (*endptr) {
case 'y':
unit *= 52; /* 52 weeks + 1 day below */
fallthrough;
case 'w':
unit *= 7;
if (param->fp_time_margin == FP_DEFAULT_TIME_MARGIN)
param->fp_time_margin *= (1 + unit / 52);
unit += (*endptr == 'y'); /* +1 day for 365 days/year */
fallthrough;
case '\0': /* days are default unit if none used */
fallthrough;
case 'd':
unit *= 24;
fallthrough;
case 'h':
unit *= 60;
fallthrough;
case 'm':
unit *= 60;
fallthrough;
case 's':
break;
/* don't need to multiply by 1 for seconds */
default:
fprintf(stderr,
"%s find: bad time string '%s': %s\n",
progname, timebuf, strerror(EINVAL));
return INT_MAX;
}
if (param->fp_time_margin == 0 ||
(*endptr && unit < param->fp_time_margin))
param->fp_time_margin = unit;
t += val * unit;
}
if (*time < t) {
if (sign != 0)
str--;
fprintf(stderr, "%s find: bad time '%s': too large\n",
progname, str);
return INT_MAX;
}
*set_t = *time - t;
/* if user requested a time range via "-xtime +M -xtime -N" then
* use the largest time and increase margin to cover the difference
* This will occur as long as the signs differ.
*/
if (sign * prev_sign == -1) {
param->fp_time_margin = abs(*set_t - prev_set);
if (prev_set > *set_t)
*set_t = prev_set;
return 0;
}
return sign;
}
static int str2quotaid(__u32 *id, const char *arg)
{
unsigned long id_tmp = 0;
char *endptr = NULL;
id_tmp = strtoul(arg, &endptr, 10);
if (*endptr != '\0')
return -EINVAL;
/* UINT32_MAX is not allowed - see projid_valid()/INVALID_PROJID */
if (id_tmp >= MDT_INVALID_PROJID)
return -ERANGE;
*id = id_tmp;
return 0;
}
static int name2uid(__u32 *id, const char *name)
{
struct passwd *passwd;
passwd = getpwnam(name);
if (!passwd)
return -ENOENT;
*id = passwd->pw_uid;
return 0;
}
static int name2gid(__u32 *id, const char *name)
{
struct group *group;
group = getgrnam(name);
if (!group)
return -ENOENT;
*id = group->gr_gid;
return 0;
}
static inline int name2projid(__u32 *id, const char *name)
{
struct ll_project prj = { .lprj_valid = LPRJ_VALID_SIZE,
.lprj_size = sizeof(prj) };
int rc;
const char *name2;
name2 = strchr(name, ':');
if (name2) {
name2++;
rc = str2quotaid(id, name2);
if (!rc)
return 0;
if (strncmp(name, "u:", 2) == 0 ||
strncmp(name, "user:", 5) == 0)
return name2uid(id, name2);
if (strncmp(name, "g:", 2) == 0 ||
strncmp(name, "group:", 5) == 0)
return name2gid(id, name2);
return -EINVAL;
}
rc = llapi_project_getnam(&prj, name);
if (rc)
return rc;
if (prj.lprj_valid & LPRJ_VALID_ID)
*id = prj.lprj_projid;
else
rc = -ENOENT;
return rc;
}
#define DEF_PW_SIZE_MAX 4096 /* minimum size to avoid malloc() handling */
static size_t getpw_size_max(void)
{
static size_t pw_size_max;
if (!pw_size_max) {
pw_size_max = sysconf(_SC_GETPW_R_SIZE_MAX);
if (pw_size_max <= 0)
pw_size_max = DEF_PW_SIZE_MAX;
}
return pw_size_max;
}
static int uid2name(char *name, size_t name_max, __u32 id)
{
struct passwd pwdbuf;
struct passwd *passwd;
int rc;
rc = getpwuid_r(id, &pwdbuf, name, name_max, &passwd);
if (passwd == NULL || rc)
return -ENOENT;
return 0;
}
static inline int gid2name(char *name, size_t name_max, __u32 id)
{
struct group grpbuf;
struct group *group;
int rc;
rc = getgrgid_r(id, &grpbuf, name, name_max, &group);
if (group == NULL || rc)
return -ENOENT;
return 0;
}
static inline int prjid2name(char *name, size_t name_max, __u32 id)
{
struct ll_project prj = { .lprj_valid = LPRJ_VALID_SIZE,
.lprj_size = sizeof(prj) };
int rc;
rc = llapi_project_getprjid(&prj, id);
if (rc)
return rc;
if (prj.lprj_valid & LPRJ_VALID_ID)
strncpy(name, prj.lprj_projname, name_max);
else
rc = -ENOENT;
return rc;
}
static int name2attrs(char *name, __u64 *attrs, __u64 *neg_attrs)
{
char *ptr, *attr_name = name;
struct attrs_name *ap;
int islongopt = 0; /* 1 true; 0 not known yet; -1 false. */
*attrs = 0;
*neg_attrs = 0;
if (strchr(name, ','))
islongopt = 1;
for (ptr = name; ; ptr = NULL) {
if (islongopt != -1)
attr_name = strtok(ptr, ",");
else
attr_name = attr_name + 1;
if (!attr_name || *attr_name == '\0')
break;
for (ap = (struct attrs_name *)attrs_array;
ap->an_attr != 0;
ap++) {
if (islongopt != -1 &&
strcmp(attr_name, ap->an_name) == 0) {
*attrs |= ap->an_attr;
islongopt = 1;
break;
} else if (islongopt != -1 && attr_name[0] == '^' &&
strcmp(attr_name + 1, ap->an_name) == 0) {
*neg_attrs |= ap->an_attr;
islongopt = 1;
break;
} else if (islongopt != 1 &&
*attr_name == ap->an_shortname) {
*attrs |= ap->an_attr;
islongopt = -1;
break;
} else if (islongopt != 1 && *attr_name == '^' &&
attr_name[1] == ap->an_shortname) {
*neg_attrs |= ap->an_attr;
islongopt = -1;
attr_name++;
break;
}
}
if (ap->an_attr == 0) {
/* provided attr is unknown */
fprintf(stderr, "error: bad attribute name '%s'\n",
attr_name);
return -1;
}
}
return 0;
}
/**
* xattr_match_info_append() - add the supplied name and value regex patterns
* to the supplied xattr_match_info struct.
* @xmi: key-value store (pointer to struct xattr_match_info)
* @exclude: If %true exclude extended attr, else if %false include attr
* @name_pattern: extended attributes name
* @value_pattern: extended attributes value
*
* Return: %0 for success, nonzero if any errors encountered.
*/
int xattr_match_info_append(struct xattr_match_info *xmi, bool exclude,
char *name_pattern, char *value_pattern)
{
int flags = REG_EXTENDED;
char *err_buf;
int err_len;
void *nptr;
int ret;
int n;
if (xmi->xattr_name_buf == NULL) {
xmi->xattr_name_buf = malloc(XATTR_LIST_MAX);
if (xmi->xattr_name_buf == NULL)
goto err_out;
}
if (xmi->xattr_value_buf == NULL) {
/*
* an xattr value need not be null-terminated, so allocate an
* extra byte to append a '\0', since regexec() expects a null-
* terminated string.
*/
xmi->xattr_value_buf = malloc(XATTR_SIZE_MAX + 1);
if (xmi->xattr_value_buf == NULL)
goto err_out;
}
n = ++xmi->xattr_regex_count;
nptr = realloc(xmi->xattr_regex_matched, n * sizeof(bool));
if (nptr == NULL)
goto err_out;
xmi->xattr_regex_matched = nptr;
nptr = realloc(xmi->xattr_regex_exclude, n * sizeof(bool));
if (nptr == NULL)
goto err_out;
xmi->xattr_regex_exclude = nptr;
nptr = realloc(xmi->xattr_regex_name, n * sizeof(regex_t *));
if (nptr == NULL)
goto err_out;
xmi->xattr_regex_name = nptr;
nptr = realloc(xmi->xattr_regex_value, n * sizeof(regex_t *));
if (nptr == NULL)
goto err_out;
xmi->xattr_regex_value = nptr;
n--;
xmi->xattr_regex_exclude[n] = exclude;
xmi->xattr_regex_name[n] = malloc(sizeof(regex_t));
if (xmi->xattr_regex_name[n] == NULL)
goto err_out;
ret = regcomp(xmi->xattr_regex_name[n], name_pattern, flags);
if (ret) {
err_len = regerror(ret, xmi->xattr_regex_name[n], NULL, 0);
err_buf = malloc(err_len);
if (err_buf == NULL)
goto err_out;
regerror(ret, xmi->xattr_regex_name[n], err_buf, err_len);
fprintf(stderr, "%s: %s: %s\n",
progname, name_pattern, err_buf);
free(err_buf);
return ret;
}
if (value_pattern && value_pattern[0] != '\0') {
xmi->xattr_regex_value[n] = malloc(sizeof(regex_t));
ret = regcomp(xmi->xattr_regex_value[n], value_pattern, flags);
if (ret) {
err_len = regerror(ret, xmi->xattr_regex_value[n],
NULL, 0);
err_buf = malloc(err_len);
if (err_buf == NULL)
goto err_out;
regerror(ret, xmi->xattr_regex_value[n], err_buf,
err_len);
fprintf(stderr, "%s: %s: %s\n",
progname, value_pattern, err_buf);
free(err_buf);
return ret;
}
} else {
xmi->xattr_regex_value[n] = NULL;
}
return 0;
err_out:
fprintf(stderr, "%s: %s\n", progname, strerror(ENOMEM));
return -ENOMEM;
}
void xattr_match_info_free(struct xattr_match_info *xmi)
{
int i;
free(xmi->xattr_regex_exclude);
xmi->xattr_regex_exclude = NULL;
free(xmi->xattr_regex_matched);
xmi->xattr_regex_matched = NULL;
for (i = 0; i < xmi->xattr_regex_count; i++) {
if (xmi->xattr_regex_name[i]) {
regfree(xmi->xattr_regex_name[i]);
free(xmi->xattr_regex_name[i]);
}
if (xmi->xattr_regex_value[i]) {
regfree(xmi->xattr_regex_value[i]);
free(xmi->xattr_regex_value[i]);
}
}
xmi->xattr_regex_count = 0;
free(xmi->xattr_regex_name);
xmi->xattr_regex_name = NULL;
free(xmi->xattr_regex_value);
xmi->xattr_regex_value = NULL;
free(xmi->xattr_name_buf);
xmi->xattr_name_buf = NULL;
free(xmi->xattr_value_buf);
xmi->xattr_value_buf = NULL;
}
/**
* compile_xattr_match_regex() - Compile regexes for matching xattr names and
* values, returning an error if either fails to compile.
* @optarg: Input (example NAME=VALUE)
* @exclude: If %true exclude extended attr, else if %false include attr
* @param: Pointer to struct find_param
*
* The argument should be in the form "NAME=VALUE". The first '=' found
* is assumed to be the separator between the name regex and the value regex.
*
* VALUE may be empty. If it is empty, it is not compiled and left NULL.
* NAME must not be empty.
*
* Return: 0 if argument string is succesfully processed, nonzero if any
* errors encountered.
*/
static int compile_xattr_match_regex(char *optarg, bool exclude,
struct find_param *param)
{
char *sep;
sep = strchr(optarg, '=');
if (sep)
*sep = '\0';
/* error if no NAME pattern specified */
if (*optarg == '\0') {
fprintf(stderr, "%s: must specify xattr pattern\n", progname);
return CMD_HELP;
}
/* if first -xattr option seen */
if (param->fp_xattr_match_info == NULL) {
param->fp_xattr_match_info = calloc(1,
sizeof(struct xattr_match_info));
if (param->fp_xattr_match_info == NULL) {
fprintf(stderr, "%s: %s\n", progname, strerror(ENOMEM));
return -ENOMEM;
}
}
/*
* if '=' was not provided, or if there is no value after the '=',
* then pass NULL to xattr_match_info_append() so that no VALUE regex
* is compiled.
*/
if (sep) {
sep++;
if (*sep == '\0')
sep = NULL;
}
return xattr_match_info_append(param->fp_xattr_match_info, exclude,
optarg, sep);
}
static int parse_symbolic(const char *input, mode_t *outmode, const char **end)
{
int loop;
int user, group, other;
int who, all;
char c, op;
mode_t perm;
mode_t usermask;
mode_t previous_flags;
user = group = other = 0;
all = 0;
loop = 1;
perm = 0;
previous_flags = 0;
*end = input;
usermask = 0;
while (loop) {
switch (*input) {
case 'u':
user = 1;
break;
case 'g':
group = 1;
break;
case 'o':
other = 1;
break;
case 'a':
user = group = other = 1;
all = 1;
break;
default:
loop = 0;
}
if (loop)
input++;
}
who = user || group || other;
if (!who) {
/* get the umask */
usermask = umask(0022);
umask(usermask);
usermask &= 07777;
}
if (*input == '-' || *input == '+' || *input == '=')
op = *input++;
else
/* operation is required */
return -1;
/* get the flags in *outmode */
switch (*input) {
case 'u':
previous_flags = (*outmode & 0700);
perm |= user ? previous_flags : 0;
perm |= group ? (previous_flags >> 3) : 0;
perm |= other ? (previous_flags >> 6) : 0;
input++;
goto write_perm;
case 'g':
previous_flags = (*outmode & 0070);
perm |= user ? (previous_flags << 3) : 0;
perm |= group ? previous_flags : 0;
perm |= other ? (previous_flags >> 3) : 0;
input++;
goto write_perm;
case 'o':
previous_flags = (*outmode & 0007);
perm |= user ? (previous_flags << 6) : 0;
perm |= group ? (previous_flags << 3) : 0;
perm |= other ? previous_flags : 0;
input++;
goto write_perm;
default:
break;
}
/* this part is optional,
* if empty perm = 0 and *outmode is not modified
*/
loop = 1;
while (loop) {
c = *input;
switch (c) {
case 'r':
perm |= user ? 0400 : 0;
perm |= group ? 0040 : 0;
perm |= other ? 0004 : 0;
/* set read permission for uog except for umask's
* permissions
*/
perm |= who ? 0 : (0444 & ~usermask);
break;
case 'w':
perm |= user ? 0200 : 0;
perm |= group ? 0020 : 0;
perm |= other ? 0002 : 0;
/* set write permission for uog except for umask'
* permissions
*/
perm |= who ? 0 : (0222 & ~usermask);
break;
case 'x':
perm |= user ? 0100 : 0;
perm |= group ? 0010 : 0;
perm |= other ? 0001 : 0;
/* set execute permission for uog except for umask'
* permissions
*/
perm |= who ? 0 : (0111 & ~usermask);
break;
case 'X':
/*
* Adds execute permission to 'u', 'g' and/or 'g' if
* specified and either 'u', 'g' or 'o' already has
* execute permissions.
*/
if ((*outmode & 0111) != 0) {
perm |= user ? 0100 : 0;
perm |= group ? 0010 : 0;
perm |= other ? 0001 : 0;
perm |= !who ? 0111 : 0;
}
break;
case 's':
/* s is ignored if o is given, but it's not an error */
if (other && !group && !user)
break;
perm |= user ? S_ISUID : 0;
perm |= group ? S_ISGID : 0;
break;
case 't':
/* 't' should be used when 'a' is given
* or who is empty
*/
perm |= (!who || all) ? S_ISVTX : 0;
/* using ugo with t is not an error */
break;
default:
loop = 0;
break;
}
if (loop)
input++;
}
write_perm:
/* uog flags should be only one character long */
if (previous_flags && (*input != '\0' && *input != ','))
return -1;
switch (op) {
case '-':
/* remove the flags from outmode */
*outmode &= ~perm;
break;
case '+':
/* add the flags to outmode */
*outmode |= perm;
break;
case '=':
/* set the flags of outmode to perm */
if (perm != 0)
*outmode = perm;
break;
}
*end = input;
return 0;
}
static int str2mode_t(const char *input, mode_t *outmode)
{
int ret;
const char *iter;
ret = 0;
if (*input >= '0' && *input <= '7') {
/* parse octal representation */
char *end;
iter = input;
/* look for invalid digits in octal representation */
while (isdigit(*iter))
if (*iter++ > '7')
return -1;
errno = 0;
*outmode = strtoul(input, &end, 8);
if (errno != 0 || *outmode > 07777) {
*outmode = 0;
ret = -1;
}
} else if (*input == '8' || *input == '9') {
/* error: invalid octal number */
ret = -1;
} else {
/* parse coma seperated list of symbolic representation */
int rc;
const char *end;
*outmode = 0;
rc = 0;
end = NULL;
do {
rc = parse_symbolic(input, outmode, &end);
if (rc)
return -1;
input = end+1;
} while (*end == ',');
if (*end != '\0')
ret = -1;
}
return ret;
}
/*
* Get the number of CPUs configured for Lustre CPTs
* by reading cpu_partition_table parameter.
*/
static int get_lustre_cpu_count(void)
{
glob_t paths;
FILE *fp;
char line[256];
int cpu_count = 0;
int rc;
/* Try to get cpu_partition_table parameter */
rc = cfs_get_param_paths(&paths, "cpu_partition_table");
if (rc != 0)
return 0;
if (paths.gl_pathc == 0) {
cfs_free_param_data(&paths);
return 0;
}
fp = fopen(paths.gl_pathv[0], "r");
if (fp == NULL) {
cfs_free_param_data(&paths);
return 0;
}
/* Count CPUs in cpu_partition_table
* Format: "cpu_partition_table=0 : 0 1 2 3..."
* "1 : 16 17 18..."
* Parse each line and count the CPU numbers listed
*/
while (fgets(line, sizeof(line), fp) != NULL) {
char *ptr = line;
char *colon_ptr;
/* Skip whitespace */
while (*ptr == ' ' || *ptr == '\t')
ptr++;
/* Look for lines with CPU partition info */
if (strncmp(ptr, "cpu_partition_table=", 20) == 0 ||
isdigit(*ptr)) {
/* Find the colon separator */
colon_ptr = strchr(ptr, ':');
if (colon_ptr) {
colon_ptr++;
while ((colon_ptr = strchr(colon_ptr,
' ')) != NULL) {
cpu_count++;
colon_ptr++;
}
}
}
}
fclose(fp);
cfs_free_param_data(&paths);
return cpu_count;
}
/*
* Calculate default thread count for lfs find based on MDT count
* and CPU count. Uses the formula:
* threads = min(MDT_COUNT * 4, CPU_COUNT / 2)
* Falls back to 4 if unable to determine parameters.
*/
static int calculate_default_thread_count(const char *path)
{
char mntdir[PATH_MAX] = "";
char fsname[PATH_MAX] = "";
int default_threads = 4;
int mdt_count = 0;
int cpu_count = 0;
int rc;
/* Get mount point and filesystem name */
rc = llapi_search_mounts(path, 0, mntdir, fsname);
if (rc < 0) {
/* Fallback to conservative default */
return default_threads;
}
/* Get MDT count */
rc = llapi_get_obd_count(mntdir, &mdt_count, 1);
if (rc < 0 || mdt_count <= 0)
mdt_count = 1; /* assume single MDT */
/* Get CPU count */
cpu_count = get_lustre_cpu_count();
if (cpu_count <= 0)
cpu_count = 1; /* fallback */
/* Calculate: min(MDT_COUNT * 4, CPU_COUNT / 2) */
default_threads = mdt_count * 4;
if (default_threads > cpu_count / 2)
default_threads = cpu_count / 2;
/* Ensure minimum of 4 */
if (default_threads < 4)
default_threads = 4;
return default_threads;
}
static int lfs_find(int argc, char **argv)
{
int c, rc;
int ret = 0;
time_t t;
struct find_param param = {
.fp_max_depth = -1,
.fp_min_depth = 0,
.fp_quiet = 1,
.fp_time_margin = FP_DEFAULT_TIME_MARGIN,
};
struct option long_opts[] = {
{ .val = 'A', .name = "atime", .has_arg = required_argument },
{ .val = LFS_ATTRS_OPT,
.name = "attrs", .has_arg = required_argument },
{ .val = 'b', .name = "blocks", .has_arg = required_argument },
{ .val = 'B', .name = "btime", .has_arg = required_argument },
{ .val = 'B', .name = "Btime", .has_arg = required_argument },
{ .val = 'B', .name = "crtime", .has_arg = required_argument },
{ .val = LFS_COMP_COUNT_OPT,
.name = "comp-count", .has_arg = required_argument },
{ .val = LFS_COMP_COUNT_OPT,
.name = "component-count",
.has_arg = required_argument },
{ .val = LFS_COMP_FLAGS_OPT,
.name = "comp-flags", .has_arg = required_argument },
{ .val = LFS_COMP_FLAGS_OPT,
.name = "component-flags",
.has_arg = required_argument },
{ .val = LFS_COMP_START_OPT,
.name = "comp-start", .has_arg = required_argument },
{ .val = LFS_COMP_START_OPT,
.name = "component-start",
.has_arg = required_argument },
{ .val = LFS_MIRROR_STATE_OPT,
.name = "mirror-state", .has_arg = required_argument },
{ .val = LFS_NEWERXY_OPT,
.name = "newer", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "neweraa", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "neweram", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newerac", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newerab", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newerma", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newermm", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newermc", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newermb", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newerca", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newercm", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newercc", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newercb", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newerba", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newerbm", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newerbc", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newerbb", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newerBa", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newerBm", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newerBc", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newerBB", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newerat", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newermt", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newerct", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newerbt", .has_arg = required_argument},
{ .val = LFS_NEWERXY_OPT,
.name = "newerBt", .has_arg = required_argument},
{ .val = 'c', .name = "stripe-count", .has_arg = required_argument },
{ .val = 'c', .name = "stripe_count", .has_arg = required_argument },
{ .val = 'C', .name = "ctime", .has_arg = required_argument },
/* getstripe { .val = 'd', .name = "directory", .has_arg = no_argument }, */
{ .val = 'd', .name = "mindepth", .has_arg = required_argument },
{ .val = 'D', .name = "maxdepth", .has_arg = required_argument },
{ .val = 'E', .name = "comp-end", .has_arg = required_argument },
{ .val = 'E', .name = "component-end",
.has_arg = required_argument },
/* find { .val = 'F', .name = "fid", .has_arg = no_argument }, */
{ .val = LFS_LAYOUT_FOREIGN_OPT,
.name = "foreign", .has_arg = optional_argument},
{ .val = 'g', .name = "gid", .has_arg = required_argument },
{ .val = 'G', .name = "group", .has_arg = required_argument },
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'H', .name = "mdt-hash", .has_arg = required_argument },
{ .val = 'i', .name = "stripe-index", .has_arg = required_argument },
{ .val = 'i', .name = "stripe_index", .has_arg = required_argument },
/* getstripe { .val = 'I', .name = "comp-id", .has_arg = required_argument }*/
{ .val = 'k', .name = "skip", .has_arg = required_argument },
{ .val = 'l', .name = "lazy", .has_arg = no_argument },
{ .val = 'L', .name = "layout", .has_arg = required_argument },
{ .val = LFS_LINKS_OPT,
.name = "links", .has_arg = required_argument },
{ .val = LFS_PRINTF_LS,
.name = "ls", .has_arg = no_argument },
{ .val = 'm', .name = "mdt", .has_arg = required_argument },
{ .val = 'm', .name = "mdt-index", .has_arg = required_argument },
{ .val = 'm', .name = "mdt_index", .has_arg = required_argument },
{ .val = 'M', .name = "mtime", .has_arg = required_argument },
{ .val = 'n', .name = "name", .has_arg = required_argument },
{ .val = 'N', .name = "mirror-count", .has_arg = required_argument },
/* find { .val = 'o' .name = "or", .has_arg = no_argument }, like find(1) */
{ .val = 'O', .name = "obd", .has_arg = required_argument },
{ .val = 'O', .name = "ost", .has_arg = required_argument },
{ .val = LFS_FIND_PERM,
.name = "perm", .has_arg = required_argument },
/* no short option for pool yet, can be 'p' after 2.18 */
{ .val = LFS_POOL_OPT,
.name = "pool", .has_arg = required_argument },
{ .val = '0', .name = "print0", .has_arg = no_argument },
{ .val = LFS_THREAD_OPT,
.name = "threads", .has_arg = required_argument },
{ .val = 'P', .name = "print", .has_arg = no_argument },
{ .val = LFS_PRINTF_OPT,
.name = "printf", .has_arg = required_argument },
{ .val = LFS_PROJID_OPT,
.name = "projid", .has_arg = required_argument },
/* getstripe { .val = 'q', .name = "quiet", .has_arg = no_argument }, */
/* getstripe { .val = 'r', .name = "recursive", .has_arg = no_argument }, */
/* getstripe { .val = 'R', .name = "raw", .has_arg = no_argument }, */
{ .val = 's', .name = "size", .has_arg = required_argument },
{ .val = 'S', .name = "stripe-size", .has_arg = required_argument },
{ .val = 'S', .name = "stripe_size", .has_arg = required_argument },
{ .val = 't', .name = "type", .has_arg = required_argument },
{ .val = 'T', .name = "mdt-count", .has_arg = required_argument },
{ .val = 'u', .name = "uid", .has_arg = required_argument },
{ .val = 'U', .name = "user", .has_arg = required_argument },
/* getstripe { .val = 'v', .name = "verbose", .has_arg = no_argument }, */
/*migrate{.val = 'W', .name = "bandwidth-limit",.has_arg = required_argument},*/
{ .val = LFS_XATTRS_MATCH_OPT,
.name = "xattr", .has_arg = required_argument },
{ .val = 'z', .name = "extension-size",
.has_arg = required_argument },
{ .val = 'z', .name = "ext-size", .has_arg = required_argument },
{ .name = NULL } };
int prev_optind = optind;
int optidx = 0;
int pathstart = -1;
int pathend = -1;
int neg_opt = 0;
time_t *xtime;
int *xsign;
int isoption;
char *endptr;
time(&t);
/* when getopt_long_only() hits '!' it returns 1, puts "!" in optarg */
while ((c = getopt_long_only(argc, argv,
"-0A:b:B:c:C:d:D:E:g:G:hH:i:k:lL:m:M:n:N:O:Ppqrs:S:t:T:u:U:z:",
long_opts, &optidx)) >= 0) {
xtime = NULL;
xsign = NULL;
if (neg_opt)
--neg_opt;
/* '!' is part of option */
/*
* when getopt_long_only() finds a string which is not
* an option nor a known option argument it returns 1
* in that case if we already have found pathstart and pathend
* (i.e. we have the list of pathnames),
* the only supported value is "!"
*/
isoption = (c != 1) || (strcmp(optarg, "!") == 0);
if (!isoption && pathend != -1) {
fprintf(stderr,
"err: %s: filename|dirname must either precede options or follow options\n",
argv[0]);
ret = CMD_HELP;
goto err;
}
if (!isoption && pathstart == -1)
pathstart = prev_optind;
if (isoption && pathstart != -1 && pathend == -1)
pathend = prev_optind;
prev_optind = optind;
switch (c) {
case 0:
/* Long options. */
break;
case 1:
/*
* unknown; opt is "!" or path component,
* checking done above.
*/
if (strcmp(optarg, "!") == 0)
neg_opt = 2;
break;
case 'A':
xtime = ¶m.fp_atime;
xsign = ¶m.fp_asign;
param.fp_exclude_atime = !!neg_opt;
/* no break, this falls through to 'B' for btime */
fallthrough;
case 'B':
if (c == 'B') {
xtime = ¶m.fp_btime;
xsign = ¶m.fp_bsign;
param.fp_exclude_btime = !!neg_opt;
}
/* no break, this falls through to 'C' for ctime */
fallthrough;
case 'C':
if (c == 'C') {
xtime = ¶m.fp_ctime;
xsign = ¶m.fp_csign;
param.fp_exclude_ctime = !!neg_opt;
}
/* no break, this falls through to 'M' for mtime */
fallthrough;
case 'M':
if (c == 'M') {
xtime = ¶m.fp_mtime;
xsign = ¶m.fp_msign;
param.fp_exclude_mtime = !!neg_opt;
}
rc = set_time(¶m, &t, xtime, xsign, optarg);
if (rc == INT_MAX) {
ret = -1;
goto err;
}
*xsign = rc;
break;
case LFS_ATTRS_OPT:
ret = name2attrs(optarg, ¶m.fp_attrs,
¶m.fp_neg_attrs);
if (ret)
goto err;
param.fp_exclude_attrs = !!neg_opt;
break;
case 'b':
if (optarg[0] == '+') {
param.fp_blocks_sign = -1;
optarg++;
} else if (optarg[0] == '-') {
param.fp_blocks_sign = 1;
optarg++;
}
param.fp_blocks_units = 512;
ret = llapi_parse_size(optarg, ¶m.fp_blocks,
¶m.fp_blocks_units, 0);
if (ret) {
fprintf(stderr, "error: bad blocks '%s'\n",
optarg);
goto err;
}
param.fp_check_blocks = 1;
param.fp_exclude_blocks = !!neg_opt;
break;
case LFS_COMP_COUNT_OPT:
if (optarg[0] == '+') {
param.fp_comp_count_sign = -1;
optarg++;
} else if (optarg[0] == '-') {
param.fp_comp_count_sign = 1;
optarg++;
}
errno = 0;
param.fp_comp_count = strtoul(optarg, &endptr, 0);
if (errno != 0 || *endptr != '\0' ||
param.fp_comp_count > UINT32_MAX) {
fprintf(stderr,
"error: bad component count '%s'\n",
optarg);
goto err;
}
param.fp_check_comp_count = 1;
param.fp_exclude_comp_count = !!neg_opt;
break;
case LFS_COMP_FLAGS_OPT:
rc = comp_str2flags(optarg, ¶m.fp_comp_flags,
¶m.fp_comp_neg_flags);
if (rc) {
fprintf(stderr,
"error: bad component flags '%s'\n",
optarg);
goto err;
}
param.fp_check_comp_flags = 1;
if (neg_opt) {
__u32 flags = param.fp_comp_neg_flags;
param.fp_comp_neg_flags = param.fp_comp_flags;
param.fp_comp_flags = flags;
}
break;
case LFS_COMP_START_OPT:
if (optarg[0] == '+') {
param.fp_comp_start_sign = -1;
optarg++;
} else if (optarg[0] == '-') {
param.fp_comp_start_sign = 1;
optarg++;
}
rc = llapi_parse_size(optarg, ¶m.fp_comp_start,
¶m.fp_comp_start_units, 0);
if (rc) {
fprintf(stderr,
"error: bad component start '%s'\n",
optarg);
goto err;
}
param.fp_check_comp_start = 1;
param.fp_exclude_comp_start = !!neg_opt;
break;
case LFS_MIRROR_STATE_OPT:
rc = mirror_str2state(optarg, ¶m.fp_mirror_state,
¶m.fp_mirror_neg_state);
if (rc) {
fprintf(stderr,
"error: bad mirrored file state '%s'\n",
optarg);
goto err;
}
param.fp_check_mirror_state = 1;
if (neg_opt) {
__u16 state = param.fp_mirror_neg_state;
param.fp_mirror_neg_state =
param.fp_mirror_state;
param.fp_mirror_state = state;
}
break;
case 'c':
if (optarg[0] == '+') {
param.fp_stripe_count_sign = -1;
optarg++;
} else if (optarg[0] == '-') {
param.fp_stripe_count_sign = 1;
optarg++;
}
errno = 0;
param.fp_stripe_count = strtoul(optarg, &endptr, 0);
if (errno != 0 || *endptr != '\0' ||
param.fp_stripe_count > LOV_MAX_STRIPE_COUNT) {
fprintf(stderr,
"error: bad stripe_count '%s'\n",
optarg);
ret = -1;
goto err;
}
param.fp_check_stripe_count = 1;
param.fp_exclude_stripe_count = !!neg_opt;
break;
case 'd':
errno = 0;
param.fp_min_depth = strtoul(optarg, 0, 0);
if (errno != 0 || param.fp_min_depth > PATH_MAX / 2) {
fprintf(stderr,
"error: bad mindepth '%s'\n",
optarg);
ret = -1;
goto err;
}
break;
case 'D':
errno = 0;
param.fp_max_depth = strtol(optarg, 0, 0);
if (errno != 0 || param.fp_max_depth < 0) {
fprintf(stderr,
"error: bad maxdepth '%s'\n",
optarg);
ret = -1;
goto err;
}
break;
case 'E':
if (optarg[0] == '+') {
param.fp_comp_end_sign = -1;
optarg++;
} else if (optarg[0] == '-') {
param.fp_comp_end_sign = 1;
optarg++;
}
if (arg_is_eof(optarg)) {
param.fp_comp_end = LUSTRE_EOF;
param.fp_comp_end_units = 1;
rc = 0;
} else {
rc = llapi_parse_size(optarg,
¶m.fp_comp_end,
¶m.fp_comp_end_units, 0);
/* assume units of KB if too small */
if (param.fp_comp_end < 4096)
param.fp_comp_end *= 1024;
}
if (rc) {
fprintf(stderr,
"error: bad component end '%s'\n",
optarg);
goto err;
}
param.fp_check_comp_end = 1;
param.fp_exclude_comp_end = !!neg_opt;
break;
case LFS_LAYOUT_FOREIGN_OPT: {
/* all types by default */
uint32_t type = LU_FOREIGN_TYPE_UNKNOWN;
if (optarg) {
/* check pure numeric */
type = strtoul(optarg, &endptr, 0);
if (*endptr) {
/* check name */
type = check_foreign_type_name(optarg);
if (type == LU_FOREIGN_TYPE_UNKNOWN) {
fprintf(stderr,
"%s %s: unknown foreign type '%s'\n",
progname, argv[0],
optarg);
return CMD_HELP;
}
} else if (type >= UINT32_MAX) {
fprintf(stderr,
"%s %s: invalid foreign type '%s'\n",
progname, argv[0], optarg);
return CMD_HELP;
}
}
param.fp_foreign_type = type;
param.fp_check_foreign = 1;
param.fp_exclude_foreign = !!neg_opt;
break;
}
case LFS_NEWERXY_OPT: {
char x = 'm';
char y = 'm';
int xidx;
int negidx;
time_t *newery;
time_t ref = time(NULL);
/* no need to check bad options, they won't get here */
if (strlen(long_opts[optidx].name) == 7) {
x = long_opts[optidx].name[5];
y = long_opts[optidx].name[6];
}
if (y == 't') {
static const char *const fmts[] = {
"%Y-%m-%d %H:%M:%S",
"%Y-%m-%d %H:%M",
"%Y-%m-%d",
"%H:%M:%S", /* sometime today */
"%H:%M",
"@%s",
"%s",
NULL };
struct tm tm;
bool found = false;
int i;
for (i = 0; fmts[i] != NULL; i++) {
char *ptr;
/* Init for times relative to today */
if (strncmp(fmts[i], "%H", 2) == 0) {
localtime_r(&ref, &tm);
} else {
memset(&tm, 0, sizeof(tm));
tm.tm_isdst = -1;
}
ptr = strptime(optarg, fmts[i], &tm);
/* Skip spaces */
while (ptr && isspace(*ptr))
ptr++;
if (ptr == optarg + strlen(optarg)) {
found = true;
break;
}
}
if (!found) {
fprintf(stderr,
"%s: invalid time '%s'\n",
progname, optarg);
fprintf(stderr,
"supported formats are:\n ");
for (i = 0; fmts[i] != NULL; i++)
fprintf(stderr, "'%s', ",
fmts[i]);
fprintf(stderr, "\n");
ret = -EINVAL;
goto err;
}
ref = mktime(&tm);
} else if (y == 'b' || y == 'B') {
lstatx_t stx;
rc = llapi_get_lum_file(optarg, NULL, &stx,
NULL, 0);
if (rc || !(stx.stx_mask & STATX_BTIME)) {
if (!(stx.stx_mask & STATX_BTIME))
ret = -EOPNOTSUPP;
else
ret = -errno;
fprintf(stderr,
"%s: get btime failed '%s': %s\n",
progname, optarg,
strerror(-ret));
goto err;
}
ref = stx.stx_btime.tv_sec;
} else {
struct stat statbuf;
if (stat(optarg, &statbuf) < 0) {
fprintf(stderr,
"%s: cannot stat file '%s': %s\n",
progname, optarg,
strerror(errno));
ret = -errno;
goto err;
}
switch (y) {
case 'a':
ref = statbuf.st_atime;
break;
case 'm':
ref = statbuf.st_mtime;
break;
case 'c':
ref = statbuf.st_ctime;
break;
default:
fprintf(stderr,
"%s: invalid Y argument: '%c'\n",
progname, x);
ret = -EINVAL;
goto err;
}
}
switch (x) {
case 'a':
xidx = NEWERXY_ATIME;
break;
case 'm':
xidx = NEWERXY_MTIME;
break;
case 'c':
xidx = NEWERXY_CTIME;
break;
case 'b':
case 'B':
xidx = NEWERXY_BTIME;
break;
default:
fprintf(stderr,
"%s: invalid X argument: '%c'\n",
progname, x);
ret = -EINVAL;
goto err;
}
negidx = !!neg_opt;
newery = ¶m.fp_newery[xidx][negidx];
if (*newery == 0) {
*newery = ref;
} else {
if (negidx)
*newery = *newery > ref ? ref : *newery;
else
*newery = *newery > ref ? *newery : ref;
}
param.fp_newerxy = 1;
break;
}
case 'g':
case 'G':
rc = name2gid(¶m.fp_gid, optarg);
if (rc && str2quotaid(¶m.fp_gid, optarg)) {
fprintf(stderr, "invalid group/GID '%s'\n",
optarg);
ret = -1;
goto err;
}
param.fp_exclude_gid = !!neg_opt;
param.fp_check_gid = 1;
break;
case 'H':
rc = mdthash_input(optarg, ¶m.fp_hash_inflags,
¶m.fp_hash_exflags,
¶m.fp_hash_type);
if (rc) {
ret = -1;
goto err;
}
if (param.fp_hash_inflags || param.fp_hash_exflags)
param.fp_check_hash_flag = 1;
param.fp_exclude_hash_type = !!neg_opt;
break;
case 'k': {
unsigned long pct = strtoul(optarg, &endptr, 10);
if (pct < 0 || pct >= 100 ||
(*endptr != '\0' && *endptr != '%')) {
ret = -1;
fprintf(stderr,
"error: invalid skip percentage '%s'\n",
optarg);
goto err;
}
param.fp_skip_percent = pct;
break;
}
case 'l':
param.fp_lazy = 1;
break;
case 'L':
ret = llapi_lov_string_pattern(optarg, ¶m.fp_layout);
if (ret)
goto err;
param.fp_exclude_layout = !!neg_opt;
param.fp_check_layout = 1;
break;
case LFS_LINKS_OPT:
if (optarg[0] == '+') {
param.fp_nlink_sign = -1;
optarg++;
} else if (optarg[0] == '-') {
param.fp_nlink_sign = 1;
optarg++;
}
errno = 0;
param.fp_nlink = strtoul(optarg, &endptr, 0);
if (errno != 0 || *endptr != '\0' || !param.fp_nlink) {
fprintf(stderr, "error: bad link count '%s'\n",
optarg);
ret = -1;
goto err;
}
param.fp_exclude_nlink = !!neg_opt;
break;
case 'u':
case 'U':
rc = name2uid(¶m.fp_uid, optarg);
if (rc && str2quotaid(¶m.fp_uid, optarg)) {
fprintf(stderr, "invalid username/UID '%s'\n",
optarg);
ret = -1;
goto err;
}
param.fp_exclude_uid = !!neg_opt;
param.fp_check_uid = 1;
break;
case 'n':
param.fp_pattern = (char *)optarg;
param.fp_exclude_pattern = !!neg_opt;
break;
case 'N':
if (optarg[0] == '+') {
param.fp_mirror_count_sign = -1;
optarg++;
} else if (optarg[0] == '-') {
param.fp_mirror_count_sign = 1;
optarg++;
}
errno = 0;
param.fp_mirror_count = strtoul(optarg, &endptr, 0);
if (errno != 0 || *endptr != '\0' ||
param.fp_mirror_count > LUSTRE_MIRROR_COUNT_MAX) {
fprintf(stderr,
"error: bad mirror count '%s'\n",
optarg);
goto err;
}
param.fp_check_mirror_count = 1;
param.fp_exclude_mirror_count = !!neg_opt;
break;
case 'm':
case 'i':
case 'O': {
int len, rc2;
int *p_num, *p_alloc;
struct obd_uuid **pp_uuid, *tmp;
const char *p1 = optarg, *p2;
char buf[UUID_MAX];
const char *pattern = "^([0-9]+)-([0-9]+)$";
regex_t reg;
regmatch_t pmatch[3];
param.fp_exclude_obd = !!neg_opt;
if (c == 'm') {
p_num = ¶m.fp_num_mdts;
p_alloc = ¶m.fp_num_alloc_mdts;
pp_uuid = ¶m.fp_mdt_uuid;
} else {
p_num = ¶m.fp_num_obds;
p_alloc = ¶m.fp_num_alloc_obds;
pp_uuid = ¶m.fp_obd_uuid;
}
regcomp(®, pattern, REG_EXTENDED);
while (p1 && *p1 != '\0') {
/* grab one uuid/idx/idx_range */
p2 = strchr(p1, ',');
if (p2 == NULL)
p2 = p1 + strlen(p1);
len = p2 - p1;
if (len >= sizeof(buf)) {
regfree(®);
ret = -E2BIG;
goto err;
}
strncpy(buf, p1, len);
buf[len] = '\0';
if (*p2 == '\0')
p1 = p2;
else
p1 = p2 + 1;
/* extend array if necessary */
if (*p_num >= *p_alloc) {
tmp = realloc(*pp_uuid,
(*p_alloc + 16) *
sizeof((*pp_uuid)[0]));
if (tmp == NULL) {
regfree(®);
ret = -ENOMEM;
goto err;
}
*pp_uuid = tmp;
*p_alloc += 16;
}
/* check pattern */
rc2 = regexec(®, buf, 3, pmatch, 0);
if (rc2 == 0) {
/* idx range such as 0-3 */
int start, end;
start = atoi(&buf[pmatch[1].rm_so]);
end = atoi(&buf[pmatch[2].rm_so]);
for ( ; start <= end; start++) {
if (*p_num >= *p_alloc) {
tmp = realloc(*pp_uuid,
(*p_alloc + 16) *
sizeof((*pp_uuid)[0]));
if (tmp == NULL) {
regfree(®);
ret = -ENOMEM;
goto err;
}
*pp_uuid = tmp;
*p_alloc += 16;
}
sprintf(buf, "%d", start);
strcpy((*pp_uuid)[(*p_num)++].uuid, buf);
}
} else if (rc2 == REG_NOMATCH) {
/* single idx or uuid */
strcpy((*pp_uuid)[(*p_num)++].uuid, buf);
} else {
regfree(®);
ret = -errno;
goto err;
}
}
regfree(®);
break;
}
#if LUSTRE_VERSION_CODE >= OBD_OCD_VERSION(2, 18, 53, 0)
case 'p':
#endif
case LFS_POOL_OPT:
if (strlen(optarg) > LOV_MAXPOOLNAME) {
fprintf(stderr,
"Pool name %s is too long (max %d)\n",
optarg, LOV_MAXPOOLNAME);
ret = -1;
goto err;
}
/*
* We do check for empty pool because empty pool
* is used to find V1 LOV attributes
*/
snprintf(param.fp_poolname, sizeof(param.fp_poolname),
"%s", optarg);
param.fp_exclude_pool = !!neg_opt;
param.fp_check_pool = 1;
break;
case '0':
param.fp_zero_end = 1;
break;
case LFS_THREAD_OPT:
param.fp_thread_count = strtol(optarg, &endptr, 0);
if (*endptr != '\0' || param.fp_thread_count < 1) {
fprintf(stderr,
"error: bad thread count '%s'\n",
optarg);
ret = -1;
goto err;
}
break;
case 'P': /* we always print, this option is a no-op */
break;
case LFS_PRINTF_OPT:
param.fp_format_printf_str = strdup(optarg);
break;
case LFS_PRINTF_LS:
param.fp_format_printf_str =
strdup("%i\t%k\t%M\t%n\t%u\t%g\t%s\t%t\t%p\n");
break;
case LFS_PROJID_OPT:
rc = name2projid(¶m.fp_projid, optarg);
if (rc && str2quotaid(¶m.fp_projid, optarg)) {
fprintf(stderr, "invalid project id: '%s'\n",
optarg);
ret = -1;
goto err;
}
param.fp_exclude_projid = !!neg_opt;
param.fp_check_projid = 1;
break;
case 's':
if (optarg[0] == '+') {
param.fp_size_sign = -1;
optarg++;
} else if (optarg[0] == '-') {
param.fp_size_sign = 1;
optarg++;
}
param.fp_size_units = 512;
ret = llapi_parse_size(optarg, ¶m.fp_size,
¶m.fp_size_units, 0);
if (ret) {
fprintf(stderr, "error: bad file size '%s'\n",
optarg);
goto err;
}
param.fp_check_size = 1;
param.fp_exclude_size = !!neg_opt;
break;
case 'S':
if (optarg[0] == '+') {
param.fp_stripe_size_sign = -1;
optarg++;
} else if (optarg[0] == '-') {
param.fp_stripe_size_sign = 1;
optarg++;
}
ret = llapi_parse_size(optarg, ¶m.fp_stripe_size,
¶m.fp_stripe_size_units, 0);
/* assume units of KB if too small to be valid */
if (param.fp_stripe_size < 4096)
param.fp_stripe_size *= 1024;
if (ret) {
fprintf(stderr, "error: bad stripe_size '%s'\n",
optarg);
goto err;
}
param.fp_check_stripe_size = 1;
param.fp_exclude_stripe_size = !!neg_opt;
break;
case 't':
param.fp_exclude_type = !!neg_opt;
switch (optarg[0]) {
case 'b':
param.fp_type = S_IFBLK;
break;
case 'c':
param.fp_type = S_IFCHR;
break;
case 'd':
param.fp_type = S_IFDIR;
break;
case 'f':
param.fp_type = S_IFREG;
break;
case 'l':
param.fp_type = S_IFLNK;
break;
case 'p':
param.fp_type = S_IFIFO;
break;
case 's':
param.fp_type = S_IFSOCK;
break;
default:
fprintf(stderr, "%s: bad type '%s'\n",
progname, optarg);
ret = CMD_HELP;
goto err;
}
break;
case LFS_FIND_PERM:
param.fp_exclude_perm = !!neg_opt;
param.fp_perm_sign = LFS_FIND_PERM_EXACT;
if (*optarg == '/') {
param.fp_perm_sign = LFS_FIND_PERM_ANY;
optarg++;
} else if (*optarg == '-') {
param.fp_perm_sign = LFS_FIND_PERM_ALL;
optarg++;
}
if (str2mode_t(optarg, ¶m.fp_perm)) {
fprintf(stderr, "error: invalid mode '%s'\n",
optarg);
ret = -1;
goto err;
}
break;
case 'T':
if (optarg[0] == '+') {
param.fp_mdt_count_sign = -1;
optarg++;
} else if (optarg[0] == '-') {
param.fp_mdt_count_sign = 1;
optarg++;
}
errno = 0;
param.fp_mdt_count = strtoul(optarg, &endptr, 0);
if (errno != 0 || *endptr != '\0' ||
param.fp_mdt_count >= UINT32_MAX) {
fprintf(stderr, "error: bad mdt_count '%s'\n",
optarg);
ret = -1;
goto err;
}
param.fp_check_mdt_count = 1;
param.fp_exclude_mdt_count = !!neg_opt;
break;
case LFS_XATTRS_MATCH_OPT:
ret = compile_xattr_match_regex(optarg, neg_opt,
¶m);
if (ret)
goto err;
break;
case 'z':
if (optarg[0] == '+') {
param.fp_ext_size_sign = -1;
optarg++;
} else if (optarg[0] == '-') {
param.fp_ext_size_sign = 1;
optarg++;
}
ret = llapi_parse_size(optarg, ¶m.fp_ext_size,
¶m.fp_ext_size_units, 0);
if (ret) {
fprintf(stderr, "error: bad ext-size '%s'\n",
optarg);
goto err;
}
param.fp_ext_size /= SEL_UNIT_SIZE;
param.fp_ext_size_units /= SEL_UNIT_SIZE;
param.fp_check_ext_size = 1;
param.fp_exclude_ext_size = !!neg_opt;
break;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
case 'h':
ret = CMD_HELP;
goto err;
}
}
if (!param.fp_verbose)
param.fp_verbose = VERBOSE_DEFAULT;
if (pathstart == -1) {
fprintf(stderr, "error: %s: no filename|pathname\n",
argv[0]);
ret = CMD_HELP;
goto err;
} else if (pathend == -1) {
/* no options */
pathend = argc;
}
if (param.fp_min_depth > param.fp_max_depth) {
fprintf(stderr, "error: %s: mindepth %u > maxdepth %u\n",
argv[0], param.fp_min_depth, param.fp_max_depth);
ret = CMD_HELP;
goto err;
}
/* Set default thread count if not specified */
if (param.fp_thread_count == 0) {
param.fp_thread_count =
calculate_default_thread_count(argv[pathstart]);
}
do {
rc = llapi_find(argv[pathstart], ¶m);
if (rc) {
if (!ret)
ret = rc;
fprintf(stderr, "%s: failed for '%s': %s\n",
progname, argv[pathstart], strerror(-rc));
}
} while (++pathstart < pathend);
err:
if (param.fp_obd_uuid && param.fp_num_alloc_obds)
free(param.fp_obd_uuid);
if (param.fp_mdt_uuid && param.fp_num_alloc_mdts)
free(param.fp_mdt_uuid);
if (param.fp_format_printf_str)
free(param.fp_format_printf_str);
if (param.fp_xattr_match_info) {
xattr_match_info_free(param.fp_xattr_match_info);
free(param.fp_xattr_match_info);
param.fp_xattr_match_info = NULL;
}
return ret;
}
static int lfs_getstripe_internal(int argc, char **argv,
struct find_param *param)
{
struct option long_opts[] = {
/* find { .val = 'A', .name = "atime", .has_arg = required_argument }*/
/* find { .val = 'b', .name = "blocks", .has_arg = required_argument }*/
/* find { .val = 'B', .name = "btime", .has_arg = required_argument }*/
/* find { .val = 'B', .name = "Btime", .has_arg = required_argument }*/
{ .val = LFS_COMP_COUNT_OPT,
.name = "comp-count", .has_arg = no_argument },
{ .val = LFS_COMP_COUNT_OPT,
.name = "component-count", .has_arg = no_argument },
{ .val = LFS_COMP_FLAGS_OPT,
.name = "comp-flags", .has_arg = optional_argument },
{ .val = LFS_COMP_FLAGS_OPT,
.name = "component-flags", .has_arg = optional_argument },
{ .val = LFS_COMP_START_OPT,
.name = "comp-start", .has_arg = optional_argument },
{ .val = LFS_COMP_START_OPT,
.name = "component-start", .has_arg = optional_argument },
{ .val = 'c', .name = "stripe-count", .has_arg = no_argument },
{ .val = 'c', .name = "stripe_count", .has_arg = no_argument },
/* find { .val = 'C', .name = "ctime", .has_arg = required_argument }*/
{ .val = 'd', .name = "directory", .has_arg = no_argument },
{ .val = 'D', .name = "default", .has_arg = no_argument },
{ .val = 'E', .name = "comp-end", .has_arg = optional_argument },
{ .val = 'E', .name = "component-end", .has_arg = optional_argument },
{ .val = 'F', .name = "fid", .has_arg = no_argument },
{ .val = 'g', .name = "generation", .has_arg = no_argument },
/* find { .val = 'G', .name = "group", .has_arg = required_argument }*/
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = LFS_HEX_IDX_OPT,
.name = "hex-idx", .has_arg = no_argument },
/* dirstripe { .val = 'H', .name = "mdt-hash", .has_arg = required_argument }*/
{ .val = 'i', .name = "stripe-index", .has_arg = no_argument },
{ .val = 'i', .name = "stripe_index", .has_arg = no_argument },
{ .val = 'I', .name = "comp-id", .has_arg = optional_argument },
{ .val = 'I', .name = "component-id", .has_arg = optional_argument },
/* find { .val = 'l', .name = "lazy", .has_arg = no_argument }, */
{ .val = 'L', .name = "layout", .has_arg = no_argument },
{ .val = 'm', .name = "mdt", .has_arg = no_argument },
{ .val = 'm', .name = "mdt-index", .has_arg = no_argument },
{ .val = 'm', .name = "mdt_index", .has_arg = no_argument },
/* find { .val = 'M', .name = "mtime", .has_arg = required_argument }*/
/* find { .val = 'n', .name = "name", .has_arg = required_argument }*/
{ .val = 'N', .name = "mirror-count", .has_arg = no_argument },
{ .val = LFS_MIRROR_INDEX_OPT,
.name = "mirror-index", .has_arg = required_argument },
{ .val = LFS_MIRROR_ID_OPT,
.name = "mirror-id", .has_arg = required_argument },
{ .val = LFS_NO_FOLLOW_OPT,
.name = "no-follow", .has_arg = no_argument },
{ .val = 'O', .name = "obd", .has_arg = required_argument },
{ .val = 'O', .name = "ost", .has_arg = required_argument },
{ .val = 'p', .name = "pool", .has_arg = no_argument },
/* find { .val = 'P', .name = "print", .has_arg = no_argument }, */
{ .val = 'q', .name = "quiet", .has_arg = no_argument },
{ .val = 'r', .name = "recursive", .has_arg = no_argument },
{ .val = 'R', .name = "raw", .has_arg = no_argument },
{ .val = 'S', .name = "stripe-size", .has_arg = no_argument },
{ .val = 'S', .name = "stripe_size", .has_arg = no_argument },
/* find { .val = 't', .name = "type", .has_arg = required_argument }*/
/* dirstripe { .val = 'T', .name = "mdt-count", .has_arg = required_argument }*/
/* find { .val = 'u', .name = "uid", .has_arg = required_argument }*/
/* find { .val = 'U', .name = "user", .has_arg = required_argument }*/
{ .val = 'v', .name = "verbose", .has_arg = no_argument },
/* dirstripe { .val = 'X',.name = "max-inherit",.has_arg = required_argument }*/
/* migrate{.val = 'W', .name = "bandwidth-limit",.has_arg = required_argument}*/
{ .val = 'y', .name = "yaml", .has_arg = no_argument },
{ .val = 'z', .name = "extension-size", .has_arg = no_argument },
{ .val = 'z', .name = "ext-size", .has_arg = no_argument },
{ .name = NULL } };
int c, rc = 0;
int neg_opt = 0;
int pathstart = -1, pathend = -1;
int isoption;
char *end, *tmp;
while ((c = getopt_long(argc, argv,
"-cdDE::FghiI::LmMNO:pqrRSvyz",
long_opts, NULL)) != -1) {
if (neg_opt)
--neg_opt;
/* '!' is part of option */
isoption = (c != 1) || (strcmp(optarg, "!") == 0);
if (!isoption && pathend != -1) {
fprintf(stderr,
"error: %s: filename|dirname must either precede options or follow options\n",
argv[0]);
return CMD_HELP;
}
if (!isoption && pathstart == -1)
pathstart = optind - 1;
if (isoption && pathstart != -1 && pathend == -1)
pathend = optind - 2;
switch (c) {
case 1:
/* unknown: opt is "!" */
if (strcmp(optarg, "!") == 0)
neg_opt = 2;
break;
case 'c':
if (!(param->fp_verbose & VERBOSE_DETAIL)) {
param->fp_verbose |= VERBOSE_STRIPE_COUNT;
param->fp_max_depth = 0;
}
break;
case LFS_COMP_COUNT_OPT:
param->fp_verbose |= VERBOSE_COMP_COUNT;
param->fp_max_depth = 0;
break;
case LFS_COMP_FLAGS_OPT:
if (optarg) {
rc = comp_str2flags(optarg,
¶m->fp_comp_flags,
¶m->fp_comp_neg_flags);
if (rc != 0) {
fprintf(stderr,
"error: %s bad component flags '%s'.\n",
argv[0], optarg);
return CMD_HELP;
}
param->fp_check_comp_flags = 1;
} else {
param->fp_verbose |= VERBOSE_COMP_FLAGS;
param->fp_max_depth = 0;
}
break;
case LFS_COMP_START_OPT:
if (optarg) {
tmp = optarg;
if (tmp[0] == '+') {
param->fp_comp_start_sign = -1;
tmp++;
} else if (tmp[0] == '-') {
param->fp_comp_start_sign = 1;
tmp++;
}
rc = llapi_parse_size(tmp,
¶m->fp_comp_start,
¶m->fp_comp_start_units, 0);
if (rc != 0) {
fprintf(stderr,
"error: %s bad component start '%s'.\n",
argv[0], tmp);
return CMD_HELP;
}
param->fp_check_comp_start = 1;
} else {
param->fp_verbose |= VERBOSE_COMP_START;
param->fp_max_depth = 0;
}
break;
case LFS_MIRROR_INDEX_OPT: {
unsigned long int mirror_index;
if (optarg[0] == '+') {
param->fp_mirror_index_sign = -1;
optarg++;
} else if (optarg[0] == '-') {
param->fp_mirror_index_sign = 1;
optarg++;
}
errno = 0;
mirror_index = strtoul(optarg, &end, 0);
if (errno != 0 || *end != '\0' ||
mirror_index > UINT16_MAX || (mirror_index == 0 &&
param->fp_mirror_index_sign == 0 && neg_opt == 0)) {
fprintf(stderr,
"%s %s: invalid mirror index '%s'\n",
progname, argv[0], optarg);
return CMD_HELP;
}
param->fp_mirror_index = (__u16)mirror_index;
if (param->fp_mirror_id != 0) {
fprintf(stderr,
"%s %s: can't specify both mirror index and mirror ID\n",
progname, argv[0]);
return CMD_HELP;
}
param->fp_check_mirror_index = 1;
param->fp_exclude_mirror_index = !!neg_opt;
break;
}
case LFS_MIRROR_ID_OPT: {
unsigned long int mirror_id;
if (optarg[0] == '+') {
param->fp_mirror_id_sign = -1;
optarg++;
} else if (optarg[0] == '-') {
param->fp_mirror_id_sign = 1;
optarg++;
}
errno = 0;
mirror_id = strtoul(optarg, &end, 0);
if (errno != 0 || *end != '\0' ||
mirror_id > UINT16_MAX || (mirror_id == 0 &&
param->fp_mirror_id_sign == 0 && neg_opt == 0)) {
fprintf(stderr,
"%s %s: invalid mirror ID '%s'\n",
progname, argv[0], optarg);
return CMD_HELP;
}
param->fp_mirror_id = (__u16)mirror_id;
if (param->fp_mirror_index != 0) {
fprintf(stderr,
"%s %s: can't specify both mirror index and mirror ID\n",
progname, argv[0]);
return CMD_HELP;
}
param->fp_check_mirror_id = 1;
param->fp_exclude_mirror_id = !!neg_opt;
break;
}
case LFS_NO_FOLLOW_OPT:
param->fp_no_follow = true;
break;
case LFS_HEX_IDX_OPT:
param->fp_hex_idx = true;
break;
case 'd':
param->fp_max_depth = 0;
break;
case 'D':
param->fp_get_default_lmv = 1;
break;
case 'E':
if (optarg) {
tmp = optarg;
if (tmp[0] == '+') {
param->fp_comp_end_sign = -1;
tmp++;
} else if (tmp[0] == '-') {
param->fp_comp_end_sign = 1;
tmp++;
}
if (arg_is_eof(tmp)) {
param->fp_comp_end = LUSTRE_EOF;
param->fp_comp_end_units = 1;
rc = 0;
} else {
rc = llapi_parse_size(tmp,
¶m->fp_comp_end,
¶m->fp_comp_end_units, 0);
/* assume units of KB if too small */
if (param->fp_comp_end < 4096)
param->fp_comp_end *= 1024;
}
if (rc != 0) {
fprintf(stderr,
"error: %s bad component end '%s'.\n",
argv[0], tmp);
return CMD_HELP;
}
param->fp_check_comp_end = 1;
} else {
param->fp_verbose |= VERBOSE_COMP_END;
param->fp_max_depth = 0;
}
break;
case 'F':
if (!(param->fp_verbose & VERBOSE_DETAIL)) {
param->fp_verbose |= VERBOSE_DFID;
param->fp_max_depth = 0;
}
break;
case 'g':
if (!(param->fp_verbose & VERBOSE_DETAIL)) {
param->fp_verbose |= VERBOSE_GENERATION;
param->fp_max_depth = 0;
}
break;
case 'i':
if (!(param->fp_verbose & VERBOSE_DETAIL)) {
param->fp_verbose |= VERBOSE_STRIPE_OFFSET;
param->fp_max_depth = 0;
}
break;
case 'I':
if (optarg) {
param->fp_comp_id = strtoul(optarg, &end, 0);
if (*end != '\0' || param->fp_comp_id == 0 ||
param->fp_comp_id > LCME_ID_MAX) {
fprintf(stderr,
"error: %s bad component id '%s'\n",
argv[0], optarg);
return CMD_HELP;
}
param->fp_check_comp_id = 1;
} else {
param->fp_max_depth = 0;
param->fp_verbose |= VERBOSE_COMP_ID;
}
break;
case 'L':
if (!(param->fp_verbose & VERBOSE_DETAIL)) {
param->fp_verbose |= VERBOSE_PATTERN;
param->fp_max_depth = 0;
}
break;
#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(3, 0, 53, 0)
case 'M':
fprintf(stderr,
"warning: '-M' deprecated, use '--mdt-index' or '-m' instead\n");
#endif
case 'm':
if (!(param->fp_verbose & VERBOSE_DETAIL))
param->fp_max_depth = 0;
param->fp_verbose |= VERBOSE_MDTINDEX;
break;
case 'N':
if (!(param->fp_verbose & VERBOSE_DETAIL)) {
param->fp_verbose |= VERBOSE_MIRROR_COUNT;
param->fp_max_depth = 0;
}
break;
case 'O':
if (param->fp_obd_uuid) {
fprintf(stderr,
"error: %s: only one obduuid allowed",
argv[0]);
return CMD_HELP;
}
param->fp_obd_uuid = (struct obd_uuid *)optarg;
break;
case 'p':
if (!(param->fp_verbose & VERBOSE_DETAIL)) {
param->fp_verbose |= VERBOSE_POOL;
param->fp_max_depth = 0;
}
break;
case 'q':
param->fp_quiet++;
break;
case 'r':
param->fp_recursive = 1;
break;
case 'R':
param->fp_raw = 1;
break;
case 'S':
if (!(param->fp_verbose & VERBOSE_DETAIL)) {
param->fp_verbose |= VERBOSE_STRIPE_SIZE;
param->fp_max_depth = 0;
}
break;
case 'v':
param->fp_verbose = VERBOSE_DEFAULT | VERBOSE_DETAIL;
break;
case 'y':
param->fp_yaml = 1;
break;
case 'z':
if (!(param->fp_verbose & VERBOSE_DETAIL)) {
param->fp_verbose |= VERBOSE_EXT_SIZE;
param->fp_max_depth = 0;
}
break;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
case 'h':
return CMD_HELP;
}
}
if (pathstart == -1) {
fprintf(stderr, "error: %s: no filename|pathname\n",
argv[0]);
return CMD_HELP;
} else if (pathend == -1) {
/* no options */
pathend = argc;
}
if (pathend > argc)
return CMD_HELP;
if (param->fp_recursive)
param->fp_max_depth = -1;
else if (param->fp_verbose & VERBOSE_DETAIL)
param->fp_max_depth = 1;
if (!param->fp_verbose)
param->fp_verbose = VERBOSE_DEFAULT;
if (param->fp_quiet)
param->fp_verbose = VERBOSE_OBJID;
do {
int rc2;
rc2 = llapi_getstripe(argv[pathstart], param);
if (rc2) {
fprintf(stderr, "%s: %s for '%s' failed: %s\n",
progname, argv[0], argv[pathstart],
strerror(-rc2));
if (!rc)
rc = rc2;
}
} while (++pathstart < pathend);
return rc;
}
static int lfs_tgts(int argc, char **argv)
{
char mntdir[PATH_MAX] = {'\0'}, path[PATH_MAX] = {'\0'};
struct find_param param;
int index = 0, rc = 0;
if (argc > 2)
return CMD_HELP;
if (argc == 2 && !realpath(argv[1], path)) {
rc = -errno;
fprintf(stderr, "error: invalid path '%s': %s\n",
argv[1], strerror(-rc));
return rc;
}
while (!llapi_search_mounts(path, index++, mntdir, NULL)) {
/* Check if we have a mount point */
if (mntdir[0] == '\0')
continue;
memset(¶m, 0, sizeof(param));
if (!strcmp(argv[0], "mdts"))
param.fp_get_lmv = 1;
rc = llapi_ostlist(mntdir, ¶m);
if (rc) {
fprintf(stderr, "error: %s: failed on %s\n",
argv[0], mntdir);
}
if (path[0] != '\0')
break;
memset(mntdir, 0, PATH_MAX);
}
return rc;
}
static int lfs_getstripe(int argc, char **argv)
{
struct find_param param = { 0 };
param.fp_max_depth = 1;
return lfs_getstripe_internal(argc, argv, ¶m);
}
/* functions */
static int lfs_getdirstripe(int argc, char **argv)
{
struct find_param param = { 0 };
struct option long_opts[] = {
{ .val = 'c', .name = "mdt-count", .has_arg = no_argument },
{ .val = 'D', .name = "default", .has_arg = no_argument },
{ .val = 'F', .name = "fid", .has_arg = no_argument },
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'H', .name = "mdt-hash", .has_arg = no_argument },
{ .val = LFS_HEX_IDX_OPT,
.name = "hex-idx", .has_arg = no_argument },
{ .val = 'i', .name = "mdt-index", .has_arg = no_argument },
{ .val = 'm', .name = "mdt-index", .has_arg = no_argument },
{ .val = 'O', .name = "obd", .has_arg = required_argument },
{ .val = 'r', .name = "recursive", .has_arg = no_argument },
{ .val = 'R', .name = "raw", .has_arg = no_argument },
{ .val = 'T', .name = "mdt-count", .has_arg = no_argument },
{ .val = 'v', .name = "verbose", .has_arg = no_argument },
{ .val = 'X', .name = "max-inherit", .has_arg = no_argument },
{ .val = LFS_INHERIT_RR_OPT,
.name = "max-inherit-rr", .has_arg = no_argument },
{ .val = 'y', .name = "yaml", .has_arg = no_argument },
{ .name = NULL } };
int c, rc = 0;
param.fp_get_lmv = 1;
while ((c = getopt_long(argc, argv,
"cDFhHimO:rRtTvXy", long_opts, NULL)) != -1) {
switch (c) {
case 'c':
case 'T':
param.fp_verbose |= VERBOSE_STRIPE_COUNT;
break;
case 'D':
param.fp_get_default_lmv = 1;
break;
case 'F':
param.fp_verbose |= VERBOSE_DFID;
break;
#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(3, 0, 53, 0)
case 't':
fprintf(stderr,
"warning: '-t' deprecated, use '--mdt-hash' or '-H' instead\n");
fallthrough;
#endif
case 'H':
param.fp_verbose |= VERBOSE_HASH_TYPE;
break;
case LFS_HEX_IDX_OPT:
param.fp_hex_idx = 1;
break;
case 'i':
fallthrough;
case 'm':
param.fp_verbose |= VERBOSE_STRIPE_OFFSET;
break;
case 'O':
if (param.fp_obd_uuid) {
fprintf(stderr,
"%s: only one obduuid allowed",
progname);
return CMD_HELP;
}
param.fp_obd_uuid = (struct obd_uuid *)optarg;
break;
case 'r':
param.fp_recursive = 1;
break;
case 'R':
param.fp_raw = 1;
break;
case 'v':
param.fp_verbose |= VERBOSE_DEFAULT | VERBOSE_DETAIL;
break;
case 'X':
param.fp_verbose |= VERBOSE_INHERIT;
break;
case LFS_INHERIT_RR_OPT:
param.fp_verbose |= VERBOSE_INHERIT_RR;
break;
case 'y':
param.fp_yaml = 1;
break;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
return CMD_HELP;
}
}
if (optind >= argc)
return CMD_HELP;
if (param.fp_recursive)
param.fp_max_depth = -1;
if (!param.fp_verbose)
param.fp_verbose = VERBOSE_DEFAULT;
do {
int rc2;
rc2 = llapi_getstripe(argv[optind], ¶m);
if (rc2) {
fprintf(stderr, "%s: %s for '%s' failed: %s\n",
progname, argv[0], argv[optind],
strerror(-rc2));
if (!rc)
rc = rc2;
}
} while (++optind < argc);
return rc;
}
enum mntdf_flags {
MNTDF_INODES = 0x0001,
MNTDF_COOKED = 0x0002,
MNTDF_LAZY = 0x0004,
MNTDF_VERBOSE = 0x0008,
MNTDF_SHOW = 0x0010,
MNTDF_DECIMAL = 0x0020,
MNTDF_NO_HEADER = 0x0040,
MNTDF_ONLY_SUMMARY = 0x0080,
};
enum showdf_fields {
SHOWDF_UUID = 0x0001,
SHOWDF_BTOTAL = 0x0002,
SHOWDF_BUSED = 0x0004,
SHOWDF_BFREE = 0x0008,
SHOWDF_BAVAIL = 0x0010,
SHOWDF_BPCT = 0x0020,
SHOWDF_BLOCKS = (SHOWDF_BTOTAL|SHOWDF_BUSED|SHOWDF_BFREE|SHOWDF_BAVAIL|
SHOWDF_BPCT),
SHOWDF_ITOTAL = 0x0040,
SHOWDF_IUSED = 0x0080,
SHOWDF_IFREE = 0x0100,
SHOWDF_IPCT = 0x0200,
SHOWDF_INODES = (SHOWDF_ITOTAL|SHOWDF_IUSED|SHOWDF_IFREE|SHOWDF_IPCT),
SHOWDF_MNTDIR = 0x0400,
SHOWDF_DEVICE = 0x0800,
SHOWDF_STATE = 0x1000,
};
struct showdf_field_name {
const char *sdfn_name;
enum showdf_fields sdfn_field;
};
static const struct showdf_field_name showdf_field_names[] = {
{ .sdfn_name = "uuid", .sdfn_field = SHOWDF_UUID },
{ .sdfn_name = "source", .sdfn_field = SHOWDF_UUID },
{ .sdfn_name = "device", .sdfn_field = SHOWDF_UUID },
{ .sdfn_name = "btotal", .sdfn_field = SHOWDF_BTOTAL },
{ .sdfn_name = "size", .sdfn_field = SHOWDF_BTOTAL },
{ .sdfn_name = "total", .sdfn_field = SHOWDF_BTOTAL },
{ .sdfn_name = "bused", .sdfn_field = SHOWDF_BUSED },
{ .sdfn_name = "used", .sdfn_field = SHOWDF_BUSED },
{ .sdfn_name = "bfree", .sdfn_field = SHOWDF_BFREE },
{ .sdfn_name = "free", .sdfn_field = SHOWDF_BFREE },
{ .sdfn_name = "bavail", .sdfn_field = SHOWDF_BAVAIL },
{ .sdfn_name = "avail", .sdfn_field = SHOWDF_BAVAIL },
{ .sdfn_name = "bpct", .sdfn_field = SHOWDF_BPCT },
{ .sdfn_name = "pcent", .sdfn_field = SHOWDF_BPCT },
{ .sdfn_name = "usepct", .sdfn_field = SHOWDF_BPCT },
{ .sdfn_name = "pct", .sdfn_field = SHOWDF_BPCT },
{ .sdfn_name = "itotal", .sdfn_field = SHOWDF_ITOTAL },
{ .sdfn_name = "inodes", .sdfn_field = SHOWDF_ITOTAL },
{ .sdfn_name = "iused", .sdfn_field = SHOWDF_IUSED },
{ .sdfn_name = "ifree", .sdfn_field = SHOWDF_IFREE },
{ .sdfn_name = "iavail", .sdfn_field = SHOWDF_IFREE },
{ .sdfn_name = "ipct", .sdfn_field = SHOWDF_IPCT },
{ .sdfn_name = "ipcent", .sdfn_field = SHOWDF_IPCT },
{ .sdfn_name = "iusepct", .sdfn_field = SHOWDF_IPCT },
{ .sdfn_name = "target", .sdfn_field = SHOWDF_MNTDIR },
{ .sdfn_name = "state", .sdfn_field = SHOWDF_STATE },
};
static enum showdf_fields showdf_str2field(const char *name)
{
int i;
for (i = 0; i < ARRAY_SIZE(showdf_field_names); i++) {
if (strcmp(name, showdf_field_names[i].sdfn_name) == 0)
return showdf_field_names[i].sdfn_field;
}
return 0;
}
static void showdf_print_available_fields(void)
{
int i;
fprintf(stderr, "Available fields:");
for (i = 0; i < ARRAY_SIZE(showdf_field_names); i++)
fprintf(stderr, " %s", showdf_field_names[i].sdfn_name);
fprintf(stderr, "\n");
}
#define COOK(value, base) \
({ \
int radix = 0; \
while (value > base) { \
value /= base; \
radix++; \
} \
radix; \
})
#define UUF "%-20s"
#define CSF "%11s"
#define CDF "%11llu"
#define HDF "%8.1f%c"
#define RSF "%4s"
#define RDF "%3d%%"
static inline int obd_statfs_ratio(const struct obd_statfs *st, bool inodes)
{
double avail, used, ratio = 0;
if (inodes) {
avail = st->os_ffree;
used = st->os_files - st->os_ffree;
} else {
avail = st->os_bavail;
used = st->os_blocks - st->os_bfree;
}
if (avail + used > 0)
ratio = used / (used + avail) * 100;
/* Round up to match df(1) usage percentage */
return (ratio - (int)ratio) > 0 ? (int)(ratio + 1) : (int)ratio;
}
/* Helper function to format and print a value with optional cooking */
static void print_field_value(long long value, enum mntdf_flags flags, int base,
char *suffix)
{
if (flags & MNTDF_COOKED) {
double cook_val = (double)value;
int i = COOK(cook_val, base);
if (i > 0)
printf(HDF, cook_val, suffix[i - 1]);
else
printf(CDF, value);
} else {
printf(CDF, value);
}
}
static void print_os_state(struct obd_statfs *stat, enum mntdf_flags flags)
{
if (stat->os_state) {
__u32 state = stat->os_state;
while (state != 0) {
const struct obd_statfs_state_name *osn;
osn = obd_statfs_state_name_find(state);
if (!osn) {
/* Unknown flag(s) for remainder.
* Print in octal to avoid confusion
* with existing 'a' and 'f' flags
* if printed in hex.
*/
printf("(%#o)", state);
break;
}
if (osn->osn_err || flags & MNTDF_VERBOSE)
printf("%c", osn->osn_name);
state ^= osn->osn_state;
}
}
}
static int showdf(char *mntdir, struct obd_statfs *stat,
const char *uuid, enum mntdf_flags flags,
char *type, int index, int rc, enum showdf_fields fields,
enum showdf_fields *field_order, int field_count)
{
int base = flags & MNTDF_DECIMAL ? 1000 : 1024;
char *suffix = flags & MNTDF_DECIMAL ? "kMGTPEZY" : "KMGTPEZY";
int shift = flags & MNTDF_COOKED ? 0 : 10;
long long btotal, bused, bfree, bavail;
long long itotal, iused, ifree;
int ratio, iratio;
int i;
if (!uuid || !stat)
return -EINVAL;
switch (rc) {
case 0:
/* Print fields in specified order */
for (i = 0; i < field_count; i++) {
enum showdf_fields field = field_order[i];
fields |= field;
if (i > 0)
printf(" ");
switch (field) {
case SHOWDF_UUID:
printf(UUF, uuid);
break;
case SHOWDF_BTOTAL:
btotal = (stat->os_blocks *
stat->os_bsize) >> shift;
print_field_value(btotal, flags, base, suffix);
break;
case SHOWDF_BUSED:
bused = ((stat->os_blocks - stat->os_bfree) *
stat->os_bsize) >> shift;
print_field_value(bused, flags, base, suffix);
break;
case SHOWDF_BFREE:
bfree = (stat->os_bfree *
stat->os_bsize) >> shift;
print_field_value(bfree, flags, base, suffix);
break;
case SHOWDF_BAVAIL:
bavail = (stat->os_bavail *
stat->os_bsize) >> shift;
print_field_value(bavail, flags, base, suffix);
break;
case SHOWDF_BPCT:
ratio = obd_statfs_ratio(stat, false);
printf(RDF, ratio);
break;
case SHOWDF_ITOTAL:
itotal = stat->os_files;
print_field_value(itotal, flags, base, suffix);
break;
case SHOWDF_IUSED:
iused = stat->os_files - stat->os_ffree;
print_field_value(iused, flags, base, suffix);
break;
case SHOWDF_IFREE:
ifree = stat->os_ffree;
print_field_value(ifree, flags, base, suffix);
break;
case SHOWDF_IPCT:
iratio = obd_statfs_ratio(stat, true);
printf(RDF, iratio);
break;
case SHOWDF_MNTDIR:
printf(" %-s", mntdir);
if (type && fields)
printf("[%s:%d]", type, index);
break;
case SHOWDF_STATE:
print_os_state(stat, flags);
break;
default:
break;
}
}
printf("\n");
break;
case -ENODATA:
printf(UUF": inactive device\n", uuid);
break;
default:
printf(UUF": %s\n", uuid, strerror(-rc));
break;
}
return 0;
}
struct ll_stat_type {
int st_op;
char *st_name;
};
#define LL_STATFS_MAX LOV_MAX_STRIPE_COUNT
struct ll_statfs_data {
int sd_index;
struct obd_statfs sd_st;
};
struct ll_statfs_buf {
int sb_count;
struct ll_statfs_data sb_buf[LL_STATFS_MAX];
};
static int mntdf(char *mntdir, char *fsname, char *pool, enum mntdf_flags flags,
int ops, struct ll_statfs_buf *lsb, int mdt_idx, int ost_idx,
enum showdf_fields fields, enum showdf_fields *field_order,
int field_count)
{
struct obd_statfs stat_buf, sum = { .os_bsize = 1 };
struct obd_uuid uuid_buf;
char *poolname = NULL;
struct ll_stat_type types[] = {
{ .st_op = LL_STATFS_LMV, .st_name = "MDT" },
{ .st_op = LL_STATFS_LOV, .st_name = "OST" },
{ .st_name = NULL } };
struct ll_stat_type *tp;
__u64 ost_files = 0;
__u64 ost_ffree = 0;
__u32 index;
__u32 type;
int fd;
int rc = 0;
int rc2;
bool show_headers;
bool only_summary;
if (pool) {
poolname = strchr(pool, '.');
if (poolname) {
if (strncmp(fsname, pool, strlen(fsname))) {
fprintf(stderr, "filesystem name incorrect\n");
return -ENODEV;
}
poolname++;
} else {
poolname = pool;
}
}
fd = open(mntdir, O_RDONLY);
if (fd < 0) {
rc = -errno;
fprintf(stderr, "%s: cannot open '%s': %s\n", progname, mntdir,
strerror(errno));
return rc;
}
/* Set default fields and field order if none specified */
if (fields == 0) {
if (flags & MNTDF_INODES) {
field_count = 0;
field_order[field_count++] = SHOWDF_UUID;
field_order[field_count++] = SHOWDF_ITOTAL;
field_order[field_count++] = SHOWDF_IUSED;
field_order[field_count++] = SHOWDF_IFREE;
field_order[field_count++] = SHOWDF_IPCT;
field_order[field_count++] = SHOWDF_MNTDIR;
field_order[field_count++] = SHOWDF_STATE;
} else {
field_count = 0;
field_order[field_count++] = SHOWDF_UUID;
field_order[field_count++] = SHOWDF_BTOTAL;
field_order[field_count++] = SHOWDF_BUSED;
field_order[field_count++] = SHOWDF_BAVAIL;
field_order[field_count++] = SHOWDF_BPCT;
field_order[field_count++] = SHOWDF_MNTDIR;
field_order[field_count++] = SHOWDF_STATE;
}
}
/* Print headers based on --output usage:
* - For --output: show headers when no specific OST or MDT index is
* specified
* - For --only-summary: never show headers
* - For default: always show headers
* - Never show headers if --no-header is specified
*/
show_headers = (ost_idx == -1 && mdt_idx == -1 &&
!(flags & MNTDF_NO_HEADER) &&
!(flags & MNTDF_ONLY_SUMMARY));
only_summary = (ost_idx == -1 && mdt_idx == -1 &&
(flags & MNTDF_ONLY_SUMMARY));
if (show_headers) {
/* Print headers in specified order */
int i;
for (i = 0; i < field_count; i++) {
enum showdf_fields field = field_order[i];
switch (field) {
case SHOWDF_UUID:
printf(UUF" ", "UUID");
break;
case SHOWDF_BTOTAL:
printf(CSF" ",
flags & MNTDF_COOKED ?
"bytes" : "1K-blocks");
break;
case SHOWDF_BUSED:
printf(CSF" ", "Used");
break;
case SHOWDF_BFREE:
printf(CSF" ", "Free");
break;
case SHOWDF_BAVAIL:
printf(CSF" ", "Available");
break;
case SHOWDF_BPCT:
printf(RSF" ", "Use%");
break;
case SHOWDF_ITOTAL:
printf(CSF" ", "Inodes");
break;
case SHOWDF_IUSED:
printf(CSF" ", "IUsed");
break;
case SHOWDF_IFREE:
printf(CSF" ", "IFree");
break;
case SHOWDF_IPCT:
printf(RSF" ", "IUse%");
break;
case SHOWDF_MNTDIR:
printf(" %-s", "Mounted on");
break;
default:
break;
}
}
printf("\n");
}
for (tp = types; tp->st_name != NULL; tp++) {
bool have_ost = false;
int index_max = LOV_ALL_STRIPES;
if (!(tp->st_op & ops))
continue;
for (index = 0; index < index_max &&
(!lsb || lsb->sb_count < LL_STATFS_MAX); index++) {
/* Skip indices that don't match the requested one */
if (tp->st_op == LL_STATFS_LMV && mdt_idx >= 0) {
if (index != mdt_idx)
continue;
index_max = 0; /* force last loop */
}
if (tp->st_op == LL_STATFS_LOV && ost_idx >= 0) {
if (index != ost_idx)
continue;
index_max = 0; /* force last loop */
}
memset(&stat_buf, 0, sizeof(struct obd_statfs));
memset(&uuid_buf, 0, sizeof(struct obd_uuid));
type = flags & MNTDF_LAZY ?
tp->st_op | LL_STATFS_NODELAY : tp->st_op;
rc2 = llapi_obd_fstatfs(fd, type, index,
&stat_buf, &uuid_buf);
if (rc2 == -ENODEV)
break;
if (rc2 == -EAGAIN)
continue;
if (rc2 == -ENODATA) { /* Inactive device, OK. */
if (!(flags & MNTDF_VERBOSE))
continue;
} else if (rc2 < 0 && rc == 0) {
rc = rc2;
}
/*
* If we have OSTs then don't report MDT block counts.
* For MDT-only filesystems the expectation is that all
* layouts have a DoM component. For filesystems with
* OSTs, files are not necessarily going to store data
* on MDTs, and MDT space is limited to a fraction of
* OST space, so don't include it in the summary.
*/
if (tp->st_op == LL_STATFS_LOV && !have_ost) {
have_ost = true;
sum.os_blocks = 0;
sum.os_bfree = 0;
sum.os_bavail = 0;
}
if (poolname && tp->st_op == LL_STATFS_LOV &&
llapi_search_ost(fsname, poolname,
obd_uuid2str(&uuid_buf)) != 1)
continue;
/*
* the llapi_obd_fstatfs() call may have returned with
* an error, but if it filled in uuid_buf we will at
* lease use that to print out a message for that OBD.
* If we didn't get anything in the uuid_buf, then fill
* it in so that we can print an error message.
*/
if (uuid_buf.uuid[0] == '\0')
snprintf(uuid_buf.uuid, sizeof(uuid_buf.uuid),
"%s%04x", tp->st_name, index);
if (!rc && lsb) {
lsb->sb_buf[lsb->sb_count].sd_index = index;
lsb->sb_buf[lsb->sb_count].sd_st = stat_buf;
lsb->sb_count++;
}
if (flags & MNTDF_SHOW && !only_summary)
showdf(mntdir, &stat_buf,
obd_uuid2str(&uuid_buf), flags,
tp->st_name, index, rc2, fields,
field_order, field_count);
if (rc2)
continue;
if (tp->st_op == LL_STATFS_LMV) {
sum.os_ffree += stat_buf.os_ffree;
sum.os_files += stat_buf.os_files;
} else /* if (tp->st_op == LL_STATFS_LOV) */ {
ost_files += stat_buf.os_files;
ost_ffree += stat_buf.os_ffree;
}
sum.os_blocks += stat_buf.os_blocks *
stat_buf.os_bsize;
sum.os_bfree += stat_buf.os_bfree *
stat_buf.os_bsize;
sum.os_bavail += stat_buf.os_bavail *
stat_buf.os_bsize;
}
}
close(fd);
/*
* If we have _some_ OSTs, but don't have as many free objects on the
* OST as inodes on the MDTs, reduce the reported number of inodes
* to compensate, so that the "inodes in use" number is correct.
* This should be kept in sync with ll_statfs_internal().
*/
if (ost_files && ost_ffree < sum.os_ffree) {
sum.os_files = (sum.os_files - sum.os_ffree) + ost_ffree;
sum.os_ffree = ost_ffree;
}
if (((flags & MNTDF_SHOW) && show_headers) || only_summary) {
if (!only_summary)
printf("\n");
showdf(mntdir, &sum, "filesystem_summary:", flags, NULL, 0, 0,
fields, field_order, field_count);
if (!only_summary)
printf("\n");
}
return rc;
}
enum {
LAYOUT_INHERIT_UNSET = -2,
};
/* functions */
static int lfs_setdirstripe(int argc, char **argv)
{
char *dname;
struct lfs_setstripe_args lsa = { 0 };
struct llapi_stripe_param *param = NULL;
__u32 mdts[LMV_MAX_STRIPE_COUNT] = { 0 };
char *end;
int c;
char *mode_opt = NULL;
bool default_stripe = false;
bool delete = false;
bool foreign_mode = false;
bool mdt_count_set = false;
bool overstriped = false;
mode_t mode = S_IRWXU | S_IRWXG | S_IRWXO;
mode_t previous_mode = 0;
char *xattr = NULL;
__u32 type = LU_FOREIGN_TYPE_SYMLINK, flags = 0;
int max_inherit = LAYOUT_INHERIT_UNSET;
int max_inherit_rr = LAYOUT_INHERIT_UNSET;
struct option long_opts[] = {
{ .val = 'c', .name = "count", .has_arg = required_argument },
{ .val = 'c', .name = "mdt-count", .has_arg = required_argument },
{ .val = 'C', .name = "mdt-overcount", .has_arg = required_argument },
{ .val = 'd', .name = "delete", .has_arg = no_argument },
{ .val = 'D', .name = "default", .has_arg = no_argument },
{ .val = 'D', .name = "default_stripe", .has_arg = no_argument },
{ .val = LFS_LAYOUT_FLAGS_OPT,
.name = "flags", .has_arg = required_argument },
{ .val = LFS_LAYOUT_FOREIGN_OPT,
.name = "foreign", .has_arg = optional_argument},
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'H', .name = "mdt-hash", .has_arg = required_argument },
#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(2, 17, 53, 0)
{ .val = 'i', .name = "mdt-index", .has_arg = required_argument },
{ .val = 'i', .name = "mdt", .has_arg = required_argument },
#else
/* find { .val = 'l', .name = "lazy", .has_arg = no_argument }, */
{ .val = 'm', .name = "mdt-index", .has_arg = required_argument },
{ .val = 'm', .name = "mdt", .has_arg = required_argument },
#endif
#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(3, 0, 53, 0)
{ .val = 'i', .name = "index", .has_arg = required_argument },
#endif
{ .val = 'o', .name = "mode", .has_arg = required_argument },
#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(3, 0, 53, 0)
{ .val = 't', .name = "hash-type", .has_arg = required_argument },
#endif
{ .val = 'T', .name = "mdt-count", .has_arg = required_argument },
{ .val = 'x', .name = "xattr", .has_arg = required_argument },
{ .val = 'X', .name = "max-inherit", .has_arg = required_argument },
{ .val = LFS_INHERIT_RR_OPT,
.name = "max-inherit-rr", .has_arg = required_argument},
/* setstripe { .val = 'y', .name = "yaml", .has_arg = no_argument }, */
/* setstripe { .val = 'W', .name = "bandwidth", .has_arg = required_argument }, */
{ .name = NULL } };
int result = 0;
setstripe_args_init(&lsa);
while ((c = getopt_long(argc, argv, "c:C:dDi:hH:m:o:t:T:x:X:",
long_opts, NULL)) >= 0) {
switch (c) {
case 0:
/* Long options. */
break;
case 'C':
overstriped = true;
fallthrough;
case 'c':
fallthrough;
case 'T':
errno = 0;
lsa.lsa_stripe_count = strtoul(optarg, &end, 0);
/* only allow count -1..-5 for overstriped dirs */
if (errno != 0 || *end != '\0' ||
lsa.lsa_stripe_count <
(overstriped ? LMV_OVERSTRIPE_COUNT_MAX :
LLAPI_OVERSTRIPE_COUNT_MIN) ||
lsa.lsa_stripe_count > LMV_MAX_STRIPE_COUNT) {
fprintf(stderr,
"%s: invalid stripe count '%s'\n",
progname, optarg);
return CMD_HELP;
}
mdt_count_set = true;
break;
case 'd':
delete = true;
default_stripe = true;
break;
case 'D':
default_stripe = true;
break;
case LFS_LAYOUT_FOREIGN_OPT:
if (optarg) {
/* check pure numeric */
type = strtoul(optarg, &end, 0);
if (*end) {
/* check name */
type = check_foreign_type_name(optarg);
if (type == LU_FOREIGN_TYPE_UNKNOWN) {
fprintf(stderr,
"%s %s: unknown foreign type '%s'\n",
progname, argv[0],
optarg);
return CMD_HELP;
}
} else if (type >= UINT32_MAX) {
fprintf(stderr,
"%s %s: invalid foreign type '%s'\n",
progname, argv[0], optarg);
return CMD_HELP;
}
}
foreign_mode = true;
break;
case LFS_LAYOUT_FLAGS_OPT:
errno = 0;
flags = strtoul(optarg, &end, 16);
if (errno != 0 || *end != '\0' ||
flags >= UINT32_MAX) {
fprintf(stderr,
"%s %s: invalid hex flags '%s'\n",
progname, argv[0], optarg);
return CMD_HELP;
}
if (!foreign_mode) {
fprintf(stderr,
"%s %s: hex flags must be specified with --foreign option\n",
progname, argv[0]);
return CMD_HELP;
}
break;
#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(3, 0, 53, 0)
case 't':
fprintf(stderr,
"warning: '--hash-type' and '-t' deprecated, use '--mdt-hash' or '-H' instead\n");
fallthrough;
#endif
case 'H':
lsa.lsa_hash = check_hashtype(optarg);
if (lsa.lsa_hash == LMV_HASH_TYPE_UNKNOWN) {
fprintf(stderr,
"%s %s: bad directory hash type '%s'\n",
progname, argv[0], optarg);
return CMD_HELP;
}
break;
case 'i':
#if LUSTRE_VERSION_CODE >= OBD_OCD_VERSION(2, 17, 53, 0)
case 'm':
#endif
#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(3, 0, 53, 0)
if (strcmp(argv[optind - 1], "--index") == 0)
fprintf(stderr,
"%s %s: warning: '--index' deprecated, use '--mdt-index' instead\n",
progname, argv[0]);
#endif
lsa.lsa_nr_tgts = parse_targets(mdts,
sizeof(mdts) / sizeof(__u32),
lsa.lsa_nr_tgts, optarg,
&overstriped);
if (lsa.lsa_nr_tgts < 0) {
fprintf(stderr,
"%s %s: invalid MDT target(s) '%s'\n",
progname, argv[0], optarg);
return CMD_HELP;
}
lsa.lsa_tgts = mdts;
if (lsa.lsa_stripe_off == LLAPI_LAYOUT_DEFAULT)
lsa.lsa_stripe_off = mdts[0];
break;
case 'o':
mode_opt = optarg;
break;
case 'x':
xattr = optarg;
break;
case 'X':
errno = 0;
max_inherit = strtol(optarg, &end, 10);
if (errno != 0 || *end != '\0' || max_inherit < -2) {
fprintf(stderr,
"%s %s: invalid max-inherit '%s'\n",
progname, argv[0], optarg);
return CMD_HELP;
}
if (max_inherit == 0) {
max_inherit = LMV_INHERIT_NONE;
} else if (max_inherit == -1) {
max_inherit = LMV_INHERIT_UNLIMITED;
} else if (max_inherit > LMV_INHERIT_MAX) {
fprintf(stderr,
"%s %s: max-inherit %d exceeds maximum %u\n",
progname, argv[0], max_inherit,
LMV_INHERIT_MAX);
return CMD_HELP;
}
break;
case LFS_INHERIT_RR_OPT:
if (!default_stripe) {
fprintf(stderr,
"%s %s: '--max-inherit-rr' must be specified with '-D'\n",
progname, argv[0]);
return CMD_HELP;
}
errno = 0;
max_inherit_rr = strtol(optarg, &end, 10);
if (errno != 0 || *end != '\0' || max_inherit_rr < -2) {
fprintf(stderr,
"%s %s: invalid max-inherit-rr '%s'\n",
progname, argv[0], optarg);
return CMD_HELP;
}
if (max_inherit_rr == 0) {
max_inherit_rr = LMV_INHERIT_RR_NONE;
} else if (max_inherit_rr == -1) {
max_inherit_rr = LMV_INHERIT_RR_UNLIMITED;
} else if (max_inherit_rr > LMV_INHERIT_RR_MAX) {
fprintf(stderr,
"%s %s: max-inherit-rr %d exceeds maximum %u\n",
progname, argv[0], max_inherit_rr,
LMV_INHERIT_RR_MAX);
return CMD_HELP;
}
break;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
return CMD_HELP;
}
}
if (optind == argc) {
fprintf(stderr, "%s %s: DIR must be specified\n",
progname, argv[0]);
return CMD_HELP;
}
if (xattr && !foreign_mode) {
/*
* only print a warning as this is armless and will be
* ignored
*/
fprintf(stderr,
"%s %s: xattr has been specified for non-foreign layout\n",
progname, argv[0]);
} else if (foreign_mode && !xattr) {
fprintf(stderr,
"%s %s: xattr must be provided in foreign mode\n",
progname, argv[0]);
return CMD_HELP;
}
if (foreign_mode && (delete || default_stripe || lsa.lsa_nr_tgts ||
lsa.lsa_tgts || setstripe_args_specified(&lsa))) {
fprintf(stderr,
"%s %s: only --xattr/--flags/--mode options are valid with --foreign\n",
progname, argv[0]);
return CMD_HELP;
}
if (!delete && lsa.lsa_stripe_off == LLAPI_LAYOUT_DEFAULT &&
lsa.lsa_stripe_count == LLAPI_LAYOUT_DEFAULT && !foreign_mode) {
/* if no parameters set, create directory on least-used MDTs */
lsa.lsa_stripe_off = LMV_OFFSET_DEFAULT;
lsa.lsa_stripe_count = 1;
}
if (delete &&
(lsa.lsa_stripe_off != LLAPI_LAYOUT_DEFAULT ||
lsa.lsa_stripe_count != LLAPI_LAYOUT_DEFAULT)) {
fprintf(stderr,
"%s %s: cannot specify -d with -c or -i options\n",
progname, argv[0]);
return CMD_HELP;
}
if (mode_opt) {
mode = strtoul(mode_opt, &end, 8);
if (*end != '\0') {
fprintf(stderr,
"%s %s: bad MODE '%s'\n",
progname, argv[0], mode_opt);
return CMD_HELP;
}
previous_mode = umask(0);
}
/* check max-inherit and warn user in some cases */
if (default_stripe &&
(lsa.lsa_stripe_count < 0 || lsa.lsa_stripe_count > 1)) {
if (max_inherit == LMV_INHERIT_UNLIMITED)
fprintf(stderr,
"%s %s: unrecommended max-inherit=-1 when default stripe-count=%lld\n",
progname, argv[0], lsa.lsa_stripe_count);
else if (max_inherit > LMV_INHERIT_DEFAULT_STRIPED + 2 &&
max_inherit != LMV_INHERIT_NONE)
fprintf(stderr,
"%s %s: unrecommended max-inherit=%d when default stripe-count=%lld\n",
progname, argv[0], max_inherit,
lsa.lsa_stripe_count);
}
if (default_stripe && lsa.lsa_nr_tgts > 1 && !mdt_count_set) {
fprintf(stderr,
"%s %s: trying to create unrecommended default striped directory layout,\n"
" '-D -i x,y,z' will stripe every new directory across all MDTs,\n"
" add -c with the number of MDTs to do this anyway\n",
progname, argv[0]);
return CMD_HELP;
}
if (max_inherit_rr != LAYOUT_INHERIT_UNSET &&
lsa.lsa_stripe_off != LLAPI_LAYOUT_DEFAULT &&
lsa.lsa_stripe_off != LMV_OFFSET_DEFAULT) {
fprintf(stderr,
"%s %s: max-inherit-rr needs mdt-index=-1, not %lld\n",
progname, argv[0], lsa.lsa_stripe_off);
return CMD_HELP;
}
/* foreign LMV/dir case */
if (foreign_mode) {
if (argc > optind + 1) {
fprintf(stderr,
"%s %s: cannot specify multiple foreign dirs\n",
progname, argv[0]);
return CMD_HELP;
}
dname = argv[optind];
result = llapi_dir_create_foreign(dname, mode, type, flags,
xattr);
if (result != 0)
fprintf(stderr,
"%s mkdir: can't create foreign dir '%s': %s\n",
progname, dname, strerror(-result));
return result;
}
/*
* initialize stripe parameters, in case param is converted to specific,
* i.e, 'lfs mkdir -i -1 -c N', always allocate space for lsp_tgts.
*/
if (lsa.lsa_stripe_count == LLAPI_LAYOUT_DEFAULT ||
lsa.lsa_stripe_count <= LLAPI_OVERSTRIPE_COUNT_MIN) {
param = calloc(1, offsetof(typeof(*param),
lsp_tgts[lsa.lsa_nr_tgts]));
} else {
param = calloc(1, offsetof(typeof(*param),
lsp_tgts[lsa.lsa_stripe_count]));
}
if (!param) {
fprintf(stderr,
"%s %s: cannot allocate memory for parameters: %s\n",
progname, argv[0], strerror(ENOMEM));
return CMD_HELP;
}
/* if "lfs setdirstripe -D -i -1" is used, assume 1-stripe directory */
if (default_stripe && lsa.lsa_stripe_off == LMV_OFFSET_DEFAULT &&
(lsa.lsa_stripe_count == LLAPI_LAYOUT_DEFAULT ||
lsa.lsa_stripe_count == 0))
lsa.lsa_stripe_count = 1;
if (lsa.lsa_stripe_count != LLAPI_LAYOUT_DEFAULT)
param->lsp_stripe_count = lsa.lsa_stripe_count;
if (lsa.lsa_stripe_off == LLAPI_LAYOUT_DEFAULT)
param->lsp_stripe_offset = LMV_OFFSET_DEFAULT;
else
param->lsp_stripe_offset = lsa.lsa_stripe_off;
param->lsp_stripe_pattern = lsa.lsa_hash;
if (overstriped) {
param->lsp_stripe_pattern |= LMV_HASH_FLAG_OVERSTRIPED;
max_inherit = LMV_INHERIT_DEFAULT_OVERSTRIPED;
}
param->lsp_pool = lsa.lsa_pool_name;
param->lsp_is_specific = false;
if (max_inherit == LAYOUT_INHERIT_UNSET) {
if (lsa.lsa_stripe_count == 0 || lsa.lsa_stripe_count == 1 ||
lsa.lsa_stripe_count == LLAPI_LAYOUT_DEFAULT)
max_inherit = LMV_INHERIT_DEFAULT_PLAIN;
else
max_inherit = LMV_INHERIT_DEFAULT_STRIPED;
}
param->lsp_max_inherit = max_inherit;
if (default_stripe) {
if (max_inherit_rr == LAYOUT_INHERIT_UNSET)
max_inherit_rr = LMV_INHERIT_RR_DEFAULT;
param->lsp_max_inherit_rr = max_inherit_rr;
}
if (strcmp(argv[0], "mkdir") == 0)
param->lsp_is_create = true;
if (lsa.lsa_nr_tgts > 1) {
if (lsa.lsa_stripe_count > 0 &&
lsa.lsa_stripe_count != LLAPI_LAYOUT_DEFAULT &&
lsa.lsa_stripe_count != lsa.lsa_nr_tgts) {
fprintf(stderr,
"error: %s: stripe count %lld doesn't match the number of MDTs: %d\n",
argv[0], lsa.lsa_stripe_count,
lsa.lsa_nr_tgts);
free(param);
return CMD_HELP;
}
param->lsp_is_specific = true;
param->lsp_stripe_count = lsa.lsa_nr_tgts;
memcpy(param->lsp_tgts, mdts, sizeof(*mdts) * lsa.lsa_nr_tgts);
}
dname = argv[optind];
do {
int rc;
if (default_stripe) {
rc = llapi_dir_set_default_lmv(dname, param);
if (rc)
fprintf(stderr,
"%s setdirstripe: cannot set default stripe on dir '%s': %s\n",
progname, dname, strerror(-rc));
if (!result)
result = rc; /* save first error for return */
continue;
}
rc = llapi_dir_create(dname, mode, param);
if (rc)
fprintf(stderr,
"%s setdirstripe: cannot create dir '%s': %s\n",
progname, dname, strerror(-rc));
if (!result)
result = rc; /* save first error for return */
} while ((dname = argv[++optind]));
if (mode_opt)
umask(previous_mode);
free(param);
return result;
}
static int lfs_rmentry(int argc, char **argv)
{
char *dname;
int index;
int result = 0;
if (argc <= 1) {
fprintf(stderr, "error: %s: missing dirname\n",
argv[0]);
return CMD_HELP;
}
index = 1;
dname = argv[index];
while (dname) {
int rc2;
rc2 = llapi_direntry_remove(dname);
if (rc2) {
fprintf(stderr,
"%s %s: remove dir entry '%s' failed: %s\n",
progname, argv[0], dname, strerror(-rc2));
if (!result)
result = rc2;
}
dname = argv[++index];
}
return result;
}
static int lfs_unlink_foreign(int argc, char **argv)
{
char *name;
int index;
int result = 0;
if (argc <= 1) {
fprintf(stderr, "error: %s: missing pathname\n",
argv[0]);
return CMD_HELP;
}
index = 1;
name = argv[index];
while (name != NULL) {
result = llapi_unlink_foreign(name);
if (result) {
fprintf(stderr,
"error: %s: unlink foreign entry '%s' failed\n",
argv[0], name);
break;
}
name = argv[++index];
}
return result;
}
static int lfs_mv(int argc, char **argv)
{
struct lmv_user_md lmu = { LMV_USER_MAGIC };
struct find_param param = {
.fp_max_depth = -1,
.fp_mdt_index = -1,
};
char *end;
int c;
int rc = 0;
struct option long_opts[] = {
{ .val = 'm', .name = "mdt", .has_arg = required_argument },
{ .val = 'm', .name = "mdt-index", .has_arg = required_argument },
{ .val = 'v', .name = "verbose", .has_arg = no_argument },
{ .name = NULL } };
while ((c = getopt_long(argc, argv, "m:M:v", long_opts, NULL)) != -1) {
switch (c) {
#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(3, 0, 53, 0)
case 'M':
fprintf(stderr,
"warning: '-M' deprecated, use '--mdt-index' or '-m' instead\n");
#endif
case 'm':
errno = 0;
lmu.lum_stripe_offset = strtoul(optarg, &end, 0);
if (errno != 0 || *end != '\0' ||
lmu.lum_stripe_offset >= UINT32_MAX) {
fprintf(stderr, "%s mv: bad MDT index '%s'\n",
progname, optarg);
return CMD_HELP;
}
break;
case 'v':
param.fp_verbose = VERBOSE_DETAIL;
break;
default:
fprintf(stderr, "%s mv: unrecognized option '%s'\n",
progname, argv[optind - 1]);
return CMD_HELP;
}
}
if (lmu.lum_stripe_offset == LMV_OFFSET_DEFAULT) {
fprintf(stderr, "%s mv: MDT index must be specified\n",
progname);
return CMD_HELP;
}
if (optind >= argc) {
fprintf(stderr, "%s mv: DIR must be specified\n", progname);
return CMD_HELP;
}
lmu.lum_hash_type = LMV_HASH_TYPE_UNKNOWN;
/* initialize migrate mdt parameters */
param.fp_lmv_md = &lmu;
param.fp_migrate = 1;
rc = llapi_migrate_mdt(argv[optind], ¶m);
if (rc != 0)
fprintf(stderr, "%s mv: cannot migrate '%s' to MDT%04x: %s\n",
progname, argv[optind], lmu.lum_stripe_offset,
strerror(-rc));
return rc;
}
static int lfs_osts(int argc, char **argv)
{
return lfs_tgts(argc, argv);
}
static int lfs_mdts(int argc, char **argv)
{
return lfs_tgts(argc, argv);
}
static int lfs_df(int argc, char **argv)
{
char mntdir[PATH_MAX] = {'\0'}, path[PATH_MAX] = {'\0'};
enum mntdf_flags flags = MNTDF_SHOW;
int ops = 0;
int c, rc = 0, rc1 = 0, index = 0, arg_idx = 0;
enum showdf_fields fields = 0;
enum showdf_fields field_order[16]; /* Store field order */
int field_count = 0;
char fsname[PATH_MAX] = "", *pool_name = NULL;
int mdt_idx = -1;
int ost_idx = -1;
struct option long_opts[] = {
{ .val = 'h', .name = "human-readable", .has_arg = no_argument },
{ .val = 'H', .name = "si", .has_arg = no_argument },
{ .val = 'i', .name = "inodes", .has_arg = no_argument },
{ .val = 'l', .name = "lazy", .has_arg = no_argument },
{ .val = 'm', .name = "mdt", .has_arg = optional_argument },
{ .val = 'N', .name = "no-header", .has_arg = no_argument },
{ .val = 's', .name = "only-summary", .has_arg = no_argument },
{ .val = 'o', .name = "ost", .has_arg = optional_argument },
{ .val = 'u', .name = "output", .has_arg = required_argument},
{ .val = 'p', .name = "pool", .has_arg = required_argument },
{ .val = 'v', .name = "verbose", .has_arg = no_argument },
{ .name = NULL} };
while ((c = getopt_long(argc, argv, "hHilm::Nso::p:u:v",
long_opts, NULL)) != -1) {
switch (c) {
case 'h':
flags = (flags & ~MNTDF_DECIMAL) | MNTDF_COOKED;
break;
case 'H':
flags |= MNTDF_COOKED | MNTDF_DECIMAL;
break;
case 'i':
flags |= MNTDF_INODES;
break;
case 'l':
flags |= MNTDF_LAZY;
break;
case 'm':
ops |= LL_STATFS_LMV;
if (optarg) {
char *end;
errno = 0;
mdt_idx = strtol(optarg, &end, 0);
if (errno != 0 || *end != '\0' || mdt_idx < 0 ||
mdt_idx > LOV_V1_INSANE_STRIPE_INDEX) {
fprintf(stderr,
"%s: invalid MDT index '%s'\n",
progname, optarg);
return CMD_HELP;
}
}
break;
case 'N':
flags |= MNTDF_NO_HEADER;
break;
case 's':
flags |= MNTDF_ONLY_SUMMARY;
break;
case 'o':
ops |= LL_STATFS_LOV;
if (optarg) {
char *end;
errno = 0;
ost_idx = strtol(optarg, &end, 0);
if (errno != 0 || *end != '\0' || ost_idx < 0 ||
ost_idx > LOV_V1_INSANE_STRIPE_INDEX) {
fprintf(stderr,
"%s: invalid OST index '%s'\n",
progname, optarg);
return CMD_HELP;
}
}
break;
case 'p':
pool_name = optarg;
break;
case 'u':
if (optarg) {
char *opt;
char *saveptr;
opt = strtok_r(optarg, ",", &saveptr);
while (opt != NULL) {
enum showdf_fields field;
field = showdf_str2field(opt);
if (field == 0) {
fprintf(stderr,
"%s: unknown field '%s' for --output option\n",
progname, opt);
showdf_print_available_fields();
return CMD_HELP;
}
fields |= field;
field_order[field_count++] = field;
opt = strtok_r(NULL, ",", &saveptr);
}
}
break;
case 'v':
flags |= MNTDF_VERBOSE;
break;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
return CMD_HELP;
}
}
/* Handle case where neither MDT nor OST flag is specified */
if (!ops)
ops |= LL_STATFS_LMV | LL_STATFS_LOV;
/* Handle case where path is not specified */
if (optind == argc) {
while (!llapi_search_mounts(path, index++, mntdir, fsname)) {
/* Check if we have a mount point */
if (mntdir[0] == '\0')
continue;
rc = mntdf(mntdir, fsname, pool_name, flags, ops, NULL,
mdt_idx, ost_idx, fields, field_order,
field_count);
if (rc || path[0] != '\0')
break;
fsname[0] = '\0'; /* avoid matching in next loop */
mntdir[0] = '\0'; /* avoid matching in next loop */
path[0] = '\0'; /* clean for next loop */
}
return rc;
}
/* Loop through all the remaining arguments. These are Lustre FS
* paths.
*/
for (arg_idx = optind; arg_idx <= argc - 1; arg_idx++) {
bool valid = false;
fsname[0] = '\0'; /* start clean */
mntdir[0] = '\0'; /* start clean */
path[0] = '\0'; /* start clean */
/* path does not exists at all */
if (!realpath(argv[arg_idx], path)) {
rc = -errno;
fprintf(stderr, "error: invalid path '%s': %s\n",
argv[arg_idx], strerror(-rc));
/* save first seen error */
if (!rc1)
rc1 = rc;
continue;
}
/* path exists but may not be a Lustre filesystem */
while (!llapi_search_mounts(path, index++, mntdir, fsname)) {
/* Check if we have a mount point */
if (mntdir[0] == '\0')
continue;
rc = mntdf(mntdir, fsname, pool_name, flags, ops, NULL,
mdt_idx, ost_idx, fields, field_order,
field_count);
if (rc || path[0] != '\0') {
valid = true;
/* save first seen error */
if (!rc1)
rc1 = rc;
break;
}
}
if (!valid) {
llapi_printf(LLAPI_MSG_ERROR,
"%s:%s Not a Lustre filesystem\n",
argv[0], argv[arg_idx]);
/* save first seen error */
if (!rc1)
rc1 = -EOPNOTSUPP;
}
}
return rc1;
}
static int print_instance(const char *mntdir, char *buf, size_t buflen,
bool opt_instance, bool opt_fsname, bool opt_uuid,
bool opt_mntdir)
{
struct obd_uuid uuid;
char *tmp = buf;
int rc = 0;
if (opt_uuid) {
rc = llapi_file_get_type_uuid(mntdir, CLI_TYPE, &uuid);
tmp = uuid.uuid;
} else if (opt_fsname == opt_instance) { /* both true or both false */
rc = llapi_getname(mntdir, buf, buflen);
} else if (opt_fsname) {
/*
* llapi_search_mounts() fills @buf with fsname, but that is not
* called if explicit paths are specified on the command-line
*/
if (buf[0] == '\0')
rc = llapi_get_fsname(mntdir, buf, buflen);
} else /* if (opt_instance) */ {
rc = llapi_get_instance(mntdir, buf, buflen);
}
if (rc < 0) {
fprintf(stderr, "cannot get instance for '%s': %s\n",
mntdir, strerror(-rc));
return rc;
}
if (opt_mntdir)
printf("%s %s\n", tmp, mntdir);
else
printf("%s\n", tmp);
return 0;
}
static int lfs_getname(int argc, char **argv)
{
struct option long_opts[] = {
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'i', .name = "instance", .has_arg = no_argument },
{ .val = 'n', .name = "fsname", .has_arg = no_argument },
{ .val = 'u', .name = "uuid", .has_arg = no_argument },
{ .name = NULL} };
bool opt_instance = false, opt_fsname = false, opt_uuid = false;
char fsname[PATH_MAX] = { 0 };
int rc = 0, rc2, c;
while ((c = getopt_long(argc, argv, "hinu", long_opts, NULL)) != -1) {
switch (c) {
case 'i':
opt_instance = true;
break;
case 'n':
opt_fsname = true;
break;
case 'u':
opt_uuid = true;
break;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
return CMD_HELP;
}
}
if (optind == argc) { /* no paths specified, get all paths. */
char mntdir[PATH_MAX] = { 0 };
char path[PATH_MAX] = { 0 };
int index = 0;
while (!llapi_search_mounts(path, index++, mntdir, fsname)) {
rc2 = print_instance(mntdir, fsname, sizeof(fsname),
opt_instance, opt_fsname, opt_uuid,
true);
if (!rc)
rc = rc2;
path[0] = fsname[0] = mntdir[0] = '\0';
}
} else { /* paths specified, only attempt to search these. */
bool opt_mntdir = ((argc - optind) != 1);
/* if only one path is given, print only requested info */
for (; optind < argc; optind++) {
rc2 = print_instance(argv[optind], fsname,
sizeof(fsname), opt_instance,
opt_fsname, opt_uuid, opt_mntdir);
if (!rc)
rc = rc2;
fsname[0] = '\0';
}
}
return rc;
}
static int lfs_check(int argc, char **argv)
{
char mntdir[PATH_MAX] = {'\0'}, path[PATH_MAX] = {'\0'};
int num_types = 1;
char *obd_types[3];
char obd_type1[4];
char obd_type2[4];
char obd_type3[4];
int rc;
if (argc < 2 || argc > 3) {
fprintf(stderr, "%s check: server type must be specified\n",
progname);
return CMD_HELP;
}
obd_types[0] = obd_type1;
obd_types[1] = obd_type2;
obd_types[2] = obd_type3;
if (strcmp(argv[1], "osts") == 0) {
strcpy(obd_types[0], "osc");
} else if (strcmp(argv[1], "mdts") == 0 ||
strcmp(argv[1], "mds") == 0) {
strcpy(obd_types[0], "mdc");
} else if (strcmp(argv[1], "mgts") == 0) {
strcpy(obd_types[0], "mgc");
} else if (strcmp(argv[1], "all") == 0 ||
strcmp(argv[1], "servers") == 0) {
num_types = 3;
strcpy(obd_types[0], "osc");
strcpy(obd_types[1], "mdc");
strcpy(obd_types[2], "mgc");
} else {
fprintf(stderr, "%s check: unrecognized option '%s'\n",
progname, argv[1]);
return CMD_HELP;
}
if (argc >= 3 && !realpath(argv[2], path)) {
rc = -errno;
fprintf(stderr, "error: invalid path '%s': %s\n",
argv[2], strerror(-rc));
return rc;
}
rc = llapi_search_mounts(path, 0, mntdir, NULL);
if (rc < 0 || mntdir[0] == '\0') {
fprintf(stderr,
"%s %s: cannot find mounted Lustre filesystem: %s\n",
progname, argv[0],
(rc < 0) ? strerror(-rc) : strerror(ENODEV));
return rc;
}
rc = llapi_target_check(num_types, obd_types, path);
if (rc)
fprintf(stderr, "%s %s: cannot check target '%s': %s\n",
progname, argv[0], argv[1], strerror(-rc));
return rc;
}
#ifdef HAVE_SYS_QUOTA_H
#define ADD_OVERFLOW(a, b) \
((((a) + (b)) < (a)) ? \
((a) = ULONG_MAX) : ((a) = (a) + (b)))
/* Convert format time string "XXwXXdXXhXXmXXs" into seconds value
* returns the value or ULONG_MAX on integer overflow or incorrect format
* Notes:
* 1. the order of specifiers is arbitrary (may be: 5w3s or 3s5w)
* 2. specifiers may be encountered multiple times (2s3s is 5 seconds)
* 3. empty integer value is interpreted as 0
*/
static unsigned long str2sec(const char *timestr)
{
const char spec[] = "smhdw";
const unsigned long mult[] = {1, 60, 60*60, 24*60*60, 7*24*60*60};
unsigned long val = 0;
char *tail;
if (strpbrk(timestr, spec) == NULL) {
/*
* no specifiers inside the time string,
* should treat it as an integer value
*/
val = strtoul(timestr, &tail, 10);
return *tail ? ULONG_MAX : val;
}
/* format string is XXwXXdXXhXXmXXs */
while (*timestr) {
unsigned long v;
int ind;
const char *ptr;
v = strtoul(timestr, &tail, 10);
if (v == ULONG_MAX || *tail == '\0')
/*
* value too large (ULONG_MAX or more)
* or missing specifier
*/
goto error;
ptr = strchr(spec, *tail);
if (!ptr)
/* unknown specifier */
goto error;
ind = ptr - spec;
/* check if product will overflow the type */
if (!(v < ULONG_MAX / mult[ind]))
goto error;
ADD_OVERFLOW(val, mult[ind] * v);
if (val == ULONG_MAX)
goto error;
timestr = tail + 1;
}
return val;
error:
return ULONG_MAX;
}
#define ARG2ULL(nr, str, def_units) \
do { \
unsigned long long limit, units = def_units; \
int rc; \
\
rc = llapi_parse_size(str, &limit, &units, 1); \
if (rc < 0) { \
fprintf(stderr, "%s: invalid limit '%s'\n", \
progname, str); \
return CMD_HELP; \
} \
nr = limit; \
} while (0)
static inline int has_times_option(int argc, char **argv)
{
int i;
for (i = 1; i < argc; i++)
if (!strcmp(argv[i], "-t") || !strcmp(argv[i], "--times"))
return 1;
return 0;
}
static inline int has_lqa_option(int argc, char **argv)
{
int i;
for (i = 1; i < argc; i++)
if (!strcmp(argv[i], "--lqa"))
return 1;
return 0;
}
/* return true if arg is insane */
static inline bool lfs_arg_insane(const char *arg, int len, const char *name)
{
const char *c;
if (strnlen(arg, len + 1) > len) {
fprintf(stderr, "%s name '%.*s' is longer than %d\n", name, len,
arg, len);
return true;
}
for (c = arg; *c != '\0'; c++) {
if (isalnum(*c) || *c == '_')
continue;
fprintf(stderr, "%s name '%.*s' has illegal character '%c'(0x%02x)\n",
name, len, arg, isprint(*c) ? *c : ' ', *c);
return true;
}
return false;
}
#define lfs_poolarg_insane(pool) lfs_arg_insane(pool, LOV_MAXPOOLNAME, "Pool")
#define lfs_lqaarg_insane(lqa) lfs_arg_insane(lqa, LQA_NAME_MAX, "LQA")
/* special grace time, only notify the user when its quota is over soft limit
* but doesn't block new writes until the hard limit is reached.
*/
#define NOTIFY_GRACE "notify"
#define NOTIFY_GRACE_TIME LQUOTA_GRACE_MASK
static int lfs_setquota_times(int argc, char **argv, struct if_quotactl *qctl)
{
int c, rc;
char *mnt, *obd_type = (char *)qctl->obd_type;
struct obd_dqblk *dqb = &qctl->qc_dqblk;
struct obd_dqinfo *dqi = &qctl->qc_dqinfo;
struct option long_opts[] = {
{ .val = 'b', .name = "block-grace", .has_arg = required_argument },
{ .val = 'g', .name = "group", .has_arg = no_argument },
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'i', .name = "inode-grace", .has_arg = required_argument },
{ .val = LFS_LQA_OPT, .name = "lqa", .has_arg = required_argument },
{ .val = LFS_POOL_OPT, .name = "pool", .has_arg = required_argument },
{ .val = 'p', .name = "projid", .has_arg = no_argument },
{ .val = 't', .name = "times", .has_arg = no_argument },
{ .val = 'u', .name = "user", .has_arg = no_argument },
{ .name = NULL } };
int qtype;
qctl->qc_cmd = LUSTRE_Q_SETINFO;
qctl->qc_type = ALLQUOTA;
while ((c = getopt_long(argc, argv, "b:ghi:ptu",
long_opts, NULL)) != -1) {
switch (c) {
case 'b':
if (strncmp(optarg, NOTIFY_GRACE,
strlen(NOTIFY_GRACE)) == 0) {
dqi->dqi_bgrace = NOTIFY_GRACE_TIME;
} else {
dqi->dqi_bgrace = str2sec(optarg);
if (dqi->dqi_bgrace >= NOTIFY_GRACE_TIME) {
fprintf(stderr,
"%s: bad block-grace: %s\n",
progname, optarg);
return CMD_HELP;
}
}
dqb->dqb_valid |= QIF_BTIME;
break;
case 'g':
qtype = GRPQUOTA;
goto quota_type;
case 'i':
if (strncmp(optarg, NOTIFY_GRACE,
strlen(NOTIFY_GRACE)) == 0) {
dqi->dqi_igrace = NOTIFY_GRACE_TIME;
} else {
dqi->dqi_igrace = str2sec(optarg);
if (dqi->dqi_igrace >= NOTIFY_GRACE_TIME) {
fprintf(stderr,
"%s: bad inode-grace: %s\n",
progname, optarg);
return CMD_HELP;
}
}
dqb->dqb_valid |= QIF_ITIME;
break;
case 'p':
qtype = PRJQUOTA;
goto quota_type;
case LFS_LQA_OPT:
if (lfs_lqaarg_insane(optarg))
return -1;
snprintf(qctl->qc_lqaname, LQA_NAME_MAX + 1, "%s",
optarg);
qctl->qc_cmd = LUSTRE_Q_SETINFOLQA;
break;
case LFS_POOL_OPT:
if (lfs_poolarg_insane(optarg))
return -1;
snprintf(qctl->qc_poolname, LOV_MAXPOOLNAME + 1, "%s",
optarg);
qctl->qc_cmd = LUSTRE_Q_SETINFOPOOL;
break;
case 't': /* Yes, of course! */
break;
case 'u':
qtype = USRQUOTA;
quota_type:
if (qctl->qc_type != ALLQUOTA) {
fprintf(stderr,
"%s: -u/g/p cannot be used more than once\n",
progname);
return CMD_HELP;
}
qctl->qc_type = qtype;
break;
/* getopt prints error message for us when opterr != 0 */
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
return CMD_HELP;
}
}
if (qctl->qc_type == ALLQUOTA) {
fprintf(stderr, "%s: neither -u, -g nor -p specified\n",
progname);
return CMD_HELP;
}
if (optind != argc - 1) {
fprintf(stderr, "%s: unexpected parameter '%s'\n",
progname, argv[optind + 1]);
return CMD_HELP;
}
mnt = argv[optind];
rc = llapi_quotactl(mnt, qctl);
if (rc) {
if (*obd_type)
fprintf(stderr, "%s %s ", obd_type,
obd_uuid2str(&qctl->obd_uuid));
fprintf(stderr, "setquota failed: %s\n", strerror(-rc));
return rc;
}
return 0;
}
static int lfs_reset_quota(char *mnt, struct if_quotactl *qctl)
{
struct if_quotactl tmp_qctl;
int index, md_count, dt_count;
int wait_phase = 0, wait_index = 0, wait_count = 0;
int rc, rc2;
/* reset the quota ID, the existing quota setting will be returned */
rc = llapi_quotactl(mnt, qctl);
if (rc)
return rc;
/* sanity check */
if ((qctl->qc_dqblk.dqb_valid & QIF_LIMITS) != QIF_LIMITS) {
fprintf(stderr,
"the existing quota settings are not returned!\n");
return -EINVAL;
}
rc = llapi_get_obd_count(mnt, &md_count, 1);
if (rc) {
fprintf(stderr, "can not get mdt count: %s\n", strerror(-rc));
return rc;
}
rc = llapi_get_obd_count(mnt, &dt_count, 0);
if (rc) {
fprintf(stderr, "can not get ost count: %s\n", strerror(-rc));
return rc;
}
memset(&tmp_qctl, 0, sizeof(tmp_qctl));
tmp_qctl.qc_type = qctl->qc_type;
tmp_qctl.qc_id = qctl->qc_id;
tmp_qctl.qc_cmd = LUSTRE_Q_GETQUOTA;
retry:
if (wait_phase == 0) {
for (index = wait_index; index < md_count; index++) {
tmp_qctl.qc_idx = index;
tmp_qctl.qc_valid = QC_MDTIDX;
rc = llapi_quotactl(mnt, &tmp_qctl);
if (rc == -ENODEV || rc == -ENODATA)
continue;
if (rc) {
fprintf(stderr, "quotactl mdt%d failed: %s\n",
index, strerror(-rc));
break;
}
/* check whether the md quota grant is reset */
if (tmp_qctl.qc_dqblk.dqb_valid & QIF_LIMITS &&
tmp_qctl.qc_dqblk.dqb_ihardlimit != 0)
break;
}
if (index < md_count) {
wait_phase = 0;
wait_index = index;
goto wait;
}
} else {
for (index = wait_index; index < dt_count; index++) {
tmp_qctl.qc_idx = index;
tmp_qctl.qc_valid = QC_OSTIDX;
rc = llapi_quotactl(mnt, &tmp_qctl);
if (rc == -ENODEV || rc == -ENODATA)
continue;
if (rc) {
fprintf(stderr, "quotactl mdt%d failed: %s\n",
index, strerror(-rc));
break;
}
/* check whether the dt quota grant is reset */
if (tmp_qctl.qc_dqblk.dqb_valid & QIF_LIMITS &&
tmp_qctl.qc_dqblk.dqb_bhardlimit != 0)
break;
}
if (index < dt_count) {
wait_phase = 1;
wait_index = index;
goto wait;
}
}
if (wait_phase == 0) {
wait_phase = 1;
goto retry;
}
goto out;
wait:
if (rc || wait_count > 30) {
fprintf(stderr, "fail to reset the quota ID %d on OBDs\n",
qctl->qc_id);
goto out;
}
wait_count++;
sleep(1);
fprintf(stdout, "wait %d seconds for OBDs to reset the quota ID %u\n",
wait_count, qctl->qc_id);
goto retry;
out:
/* restore the quota setting */
if (qctl->qc_dqblk.dqb_isoftlimit == 0 &&
qctl->qc_dqblk.dqb_ihardlimit == 0 &&
qctl->qc_dqblk.dqb_bsoftlimit == 0 &&
qctl->qc_dqblk.dqb_bhardlimit == 0)
return rc;
memcpy(&tmp_qctl, qctl, sizeof(tmp_qctl));
tmp_qctl.qc_cmd = LUSTRE_Q_SETQUOTA;
rc2 = llapi_quotactl(mnt, &tmp_qctl);
if (!rc2)
return rc;
fprintf(stderr,
"fail to restore the quota setting: %s, please restore it manually by\n lfs setquota %s %d",
strerror(-rc2),
qctl->qc_type == USRQUOTA ? "-u" :
(qctl->qc_type == GRPQUOTA ? "-g" : "-p"),
qctl->qc_id);
if (qctl->qc_dqblk.dqb_isoftlimit != 0)
fprintf(stderr, " -i %llu",
(unsigned long long)qctl->qc_dqblk.dqb_isoftlimit);
if (qctl->qc_dqblk.dqb_ihardlimit != 0)
fprintf(stderr, " -I %llu",
(unsigned long long)qctl->qc_dqblk.dqb_ihardlimit);
if (qctl->qc_dqblk.dqb_bsoftlimit != 0)
fprintf(stderr, " -b %llu",
(unsigned long long)qctl->qc_dqblk.dqb_bsoftlimit);
if (qctl->qc_dqblk.dqb_bhardlimit != 0)
fprintf(stderr, " -B %llu",
(unsigned long long)qctl->qc_dqblk.dqb_bhardlimit);
fprintf(stderr, " %s\n", mnt);
if (!rc)
rc = rc2;
return rc;
}
#define BSLIMIT (1 << 0)
#define BHLIMIT (1 << 1)
#define ISLIMIT (1 << 2)
#define IHLIMIT (1 << 3)
int lfs_setquota(int argc, char **argv)
{
int c, rc = 0;
struct if_quotactl *qctl;
char *mnt, *obd_type;
struct obd_dqblk *dqb;
struct option long_opts[] = {
{ .val = 'b', .name = "block-softlimit",
.has_arg = required_argument },
{ .val = 'B', .name = "block-hardlimit",
.has_arg = required_argument },
{ .val = 'D', .name = "default", .has_arg = no_argument },
{ .val = LFS_SETQUOTA_DELETE,
.name = "delete", .has_arg = no_argument },
{ .val = 'g', .name = "group", .has_arg = required_argument },
{ .val = 'G', .name = "default-grp", .has_arg = no_argument },
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'i', .name = "inode-softlimit",
.has_arg = required_argument },
{ .val = 'I', .name = "inode-hardlimit",
.has_arg = required_argument },
{ .val = 'p', .name = "projid", .has_arg = required_argument },
{ .val = 'P', .name = "default-prj", .has_arg = no_argument },
{ .val = 'r', .name = "reset", .has_arg = no_argument },
{ .val = 'u', .name = "user", .has_arg = required_argument },
{ .val = 'U', .name = "default-usr", .has_arg = no_argument },
{ .val = LFS_LQA_OPT,
.name = "lqa", .has_arg = required_argument },
{ .val = LFS_POOL_OPT,
.name = "pool", .has_arg = required_argument },
{ .name = NULL } };
unsigned int limit_mask = 0;
bool use_default = false;
int qtype, qctl_len;
/* By default LOV_MAXPOOLNAME and LQA_NAME_MAX should be equal. Handle a
* case if it would be changed in future.
*/
qctl_len = sizeof(*qctl) + max(LOV_MAXPOOLNAME + 1, LQA_NAME_MAX + 1);
qctl = malloc(qctl_len);
if (!qctl)
return -ENOMEM;
memset(qctl, 0, qctl_len);
obd_type = (char *)qctl->obd_type;
dqb = &qctl->qc_dqblk;
if (has_times_option(argc, argv)) {
rc = lfs_setquota_times(argc, argv, qctl);
goto out;
}
qctl->qc_cmd = LUSTRE_Q_SETQUOTA;
qctl->qc_type = ALLQUOTA; /* ALLQUOTA makes no sense for setquota,
* so it can be used as a marker that qc_type
* isn't reinitialized from command line
*/
while ((c = getopt_long(argc, argv, "b:B:dDg:Ghi:I:p:Pru:U",
long_opts, NULL)) != -1) {
switch (c) {
case 'U':
qctl->qc_cmd = LUSTRE_Q_SETDEFAULT;
qtype = USRQUOTA;
qctl->qc_id = 0;
goto quota_type_def;
case 'u':
qtype = USRQUOTA;
rc = name2uid(&qctl->qc_id, optarg);
goto quota_type;
case 'G':
qctl->qc_cmd = LUSTRE_Q_SETDEFAULT;
qtype = GRPQUOTA;
qctl->qc_id = 0;
goto quota_type_def;
case 'g':
qtype = GRPQUOTA;
rc = name2gid(&qctl->qc_id, optarg);
goto quota_type;
case 'P':
qctl->qc_cmd = LUSTRE_Q_SETDEFAULT;
qtype = PRJQUOTA;
qctl->qc_id = 0;
goto quota_type_def;
case 'p':
qtype = PRJQUOTA;
rc = name2projid(&qctl->qc_id, optarg);
quota_type:
if (rc && str2quotaid(&qctl->qc_id, optarg)) {
fprintf(stderr,
"%s setquota: invalid %s '%s'\n",
progname,
qtype == USRQUOTA ? "user id" :
qtype == GRPQUOTA ? "group id" :
"project id",
optarg);
rc = -1;
goto out;
}
if (qctl->qc_id == 0) {
fprintf(stderr,
"%s setquota: can't set quota for root usr/group/project.\n",
progname);
rc = -1;
goto out;
}
quota_type_def:
if (qctl->qc_type != ALLQUOTA) {
fprintf(stderr,
"%s setquota: only one of -u, -U, -g, -G, -p or -P may be specified\n",
progname);
rc = CMD_HELP;
goto out;
}
qctl->qc_type = qtype;
break;
#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(2, 22, 53, 0)
case 'd':
fprintf(stderr,
"%s setquota: '-d' deprecated, use '-D' or '--default'\n",
progname);
fallthrough;
#endif
case 'D':
use_default = true;
qctl->qc_cmd = LUSTRE_Q_SETDEFAULT;
break;
case LFS_SETQUOTA_DELETE:
qctl->qc_cmd = LUSTRE_Q_DELETEQID;
break;
case 'b':
ARG2ULL(dqb->dqb_bsoftlimit, optarg, 1024);
dqb->dqb_bsoftlimit >>= 10;
limit_mask |= BSLIMIT;
if (dqb->dqb_bsoftlimit &&
dqb->dqb_bsoftlimit <= 1024) /* <= 1M? */
fprintf(stderr,
"%s setquota: warning: block softlimit '%llu' smaller than minimum qunit size\nSee '%s help setquota' or Lustre manual for details\n",
progname,
(unsigned long long)dqb->dqb_bsoftlimit,
progname);
break;
case 'B':
ARG2ULL(dqb->dqb_bhardlimit, optarg, 1024);
dqb->dqb_bhardlimit >>= 10;
limit_mask |= BHLIMIT;
if (dqb->dqb_bhardlimit &&
dqb->dqb_bhardlimit <= 1024) /* <= 1M? */
fprintf(stderr,
"%s setquota: warning: block hardlimit '%llu' smaller than minimum qunit size\n"
"See '%s help setquota' or Lustre manual for details\n",
progname,
(unsigned long long)dqb->dqb_bhardlimit,
progname);
break;
case 'i':
ARG2ULL(dqb->dqb_isoftlimit, optarg, 1);
limit_mask |= ISLIMIT;
if (dqb->dqb_isoftlimit &&
dqb->dqb_isoftlimit <= 1024) /* <= 1K inodes? */
fprintf(stderr,
"%s setquota: warning: inode softlimit '%llu' smaller than minimum qunit size\nSee '%s help setquota' or Lustre manual for details\n",
progname,
(unsigned long long)dqb->dqb_isoftlimit,
progname);
break;
case 'I':
ARG2ULL(dqb->dqb_ihardlimit, optarg, 1);
limit_mask |= IHLIMIT;
if (dqb->dqb_ihardlimit &&
dqb->dqb_ihardlimit <= 1024) /* <= 1K inodes? */
fprintf(stderr,
"%s setquota: warning: inode hardlimit '%llu' smaller than minimum qunit size\nSee '%s help setquota' or Lustre manual for details\n",
progname,
(unsigned long long)dqb->dqb_ihardlimit,
progname);
break;
case LFS_LQA_OPT:
if (qctl->qc_type == ALLQUOTA || qctl->qc_id != 0) {
fprintf(stderr, "Specify LQA quota type with -U,-G, or -P\n");
rc = CMD_HELP;
goto out;
}
if (lfs_lqaarg_insane(optarg)) {
rc = -1;
goto out;
}
snprintf(qctl->qc_lqaname, LQA_NAME_MAX + 1, "%s",
optarg);
qctl->qc_cmd = LUSTRE_Q_SETQUOTALQA;
qctl->qc_id = 1;
break;
case LFS_POOL_OPT:
if (lfs_poolarg_insane(optarg)) {
rc = -1;
goto out;
}
snprintf(qctl->qc_poolname, LOV_MAXPOOLNAME + 1, "%s",
optarg);
qctl->qc_cmd = qctl->qc_cmd == LUSTRE_Q_SETDEFAULT ?
LUSTRE_Q_SETDEFAULT_POOL :
LUSTRE_Q_SETQUOTAPOOL;
break;
case 'r':
qctl->qc_cmd = LUSTRE_Q_RESETQID;
break;
default:
fprintf(stderr,
"%s setquota: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
rc = CMD_HELP;
goto out;
}
}
if (LUSTRE_Q_CMD_IS_POOL(qctl->qc_cmd) &&
limit_mask & (IHLIMIT | ISLIMIT)) {
fprintf(stderr,
"%s setquota: inode limits are not supported with Pool Quotas\n",
progname);
rc = -EINVAL;
goto out;
}
if (qctl->qc_type == ALLQUOTA) {
fprintf(stderr,
"%s setquota: either -u or -g must be specified\n",
progname);
rc = CMD_HELP;
goto out;
}
if (!use_default && qctl->qc_cmd != LUSTRE_Q_DELETEQID &&
qctl->qc_cmd != LUSTRE_Q_RESETQID && limit_mask == 0) {
fprintf(stderr,
"%s setquota: at least one limit must be specified\n",
progname);
rc = CMD_HELP;
goto out;
}
if ((use_default || qctl->qc_cmd == LUSTRE_Q_DELETEQID ||
qctl->qc_cmd == LUSTRE_Q_RESETQID) && limit_mask != 0) {
fprintf(stderr,
"%s setquota: limits should not be specified when using default quota, deleting or resetting quota ID\n",
progname);
rc = CMD_HELP;
goto out;
}
if (use_default && qctl->qc_id == 0) {
fprintf(stderr,
"%s setquota: can not set default quota for root user/group/project\n",
progname);
rc = CMD_HELP;
goto out;
}
if ((qctl->qc_cmd == LUSTRE_Q_DELETEQID ||
qctl->qc_cmd == LUSTRE_Q_RESETQID) && qctl->qc_id == 0) {
fprintf(stderr,
"%s setquota: can not delete or reset root user/group/project\n",
progname);
rc = CMD_HELP;
goto out;
}
if (optind != argc - 1) {
fprintf(stderr,
"%s setquota: filesystem not specified or unexpected argument '%s'\n",
progname, argv[optind]);
rc = CMD_HELP;
goto out;
}
mnt = argv[optind];
if (use_default) {
dqb->dqb_bhardlimit = 0;
dqb->dqb_bsoftlimit = 0;
dqb->dqb_ihardlimit = 0;
dqb->dqb_isoftlimit = 0;
dqb->dqb_itime = 0;
dqb->dqb_btime = 0;
dqb->dqb_valid |= QIF_LIMITS | QIF_TIMES;
/* do not set inode limits for Pool Quotas */
if (qctl->qc_cmd == LUSTRE_Q_SETDEFAULT_POOL)
dqb->dqb_valid ^= QIF_ILIMITS | QIF_ITIME;
} else if ((!(limit_mask & BHLIMIT) ^ !(limit_mask & BSLIMIT)) ||
(!(limit_mask & IHLIMIT) ^ !(limit_mask & ISLIMIT))) {
/* sigh, we can't just set blimits/ilimits */
struct if_quotactl *tmp_qctl;
tmp_qctl = calloc(1, qctl_len);
if (!tmp_qctl)
goto out;
if (qctl->qc_cmd == LUSTRE_Q_SETQUOTAPOOL) {
tmp_qctl->qc_cmd = LUSTRE_Q_GETQUOTAPOOL;
snprintf(tmp_qctl->qc_poolname, LOV_MAXPOOLNAME + 1,
"%s", qctl->qc_poolname);
} else if (qctl->qc_cmd == LUSTRE_Q_SETQUOTALQA) {
tmp_qctl->qc_cmd = LUSTRE_Q_GETQUOTALQA;
snprintf(tmp_qctl->qc_lqaname, LQA_NAME_MAX + 1,
"%s", qctl->qc_lqaname);
} else {
tmp_qctl->qc_cmd = LUSTRE_Q_GETQUOTA;
}
tmp_qctl->qc_type = qctl->qc_type;
tmp_qctl->qc_id = qctl->qc_id;
rc = llapi_quotactl(mnt, tmp_qctl);
if (rc < 0) {
free(tmp_qctl);
goto out;
}
if (!(limit_mask & BHLIMIT))
dqb->dqb_bhardlimit = tmp_qctl->qc_dqblk.dqb_bhardlimit;
if (!(limit_mask & BSLIMIT))
dqb->dqb_bsoftlimit = tmp_qctl->qc_dqblk.dqb_bsoftlimit;
if (!(limit_mask & IHLIMIT))
dqb->dqb_ihardlimit = tmp_qctl->qc_dqblk.dqb_ihardlimit;
if (!(limit_mask & ISLIMIT))
dqb->dqb_isoftlimit = tmp_qctl->qc_dqblk.dqb_isoftlimit;
/* Keep grace times if we have got no softlimit arguments */
if ((limit_mask & BHLIMIT) && !(limit_mask & BSLIMIT)) {
dqb->dqb_valid |= QIF_BTIME;
dqb->dqb_btime = tmp_qctl->qc_dqblk.dqb_btime;
}
if ((limit_mask & IHLIMIT) && !(limit_mask & ISLIMIT)) {
dqb->dqb_valid |= QIF_ITIME;
dqb->dqb_itime = tmp_qctl->qc_dqblk.dqb_itime;
}
free(tmp_qctl);
}
dqb->dqb_valid |= (limit_mask & (BHLIMIT | BSLIMIT)) ? QIF_BLIMITS : 0;
dqb->dqb_valid |= (limit_mask & (IHLIMIT | ISLIMIT)) ? QIF_ILIMITS : 0;
if (qctl->qc_cmd == LUSTRE_Q_RESETQID)
rc = lfs_reset_quota(mnt, qctl);
else
rc = llapi_quotactl(mnt, qctl);
if (rc) {
if (*obd_type)
fprintf(stderr,
"%s setquota: cannot quotactl '%s' '%s': %s\n",
progname, obd_type,
obd_uuid2str(&qctl->obd_uuid), strerror(-rc));
else
fprintf(stderr,
"%s setquota: quotactl failed: %s\n",
progname, strerror(-rc));
}
out:
if (rc)
fprintf(stderr, "setquota failed: %s\n", strerror(-rc));
free(qctl);
return rc;
}
/* Converts seconds value into format string
* result is returned in buf
* Notes:
* 1. result is in descenting order: 1w2d3h4m5s
* 2. zero fields are not filled (except for p. 3): 5d1s
* 3. zero seconds value is presented as "0s"
*/
static char *__sec2str(time_t seconds, char *buf)
{
const char spec[] = "smhdw";
const unsigned long mult[] = {1, 60, 60*60, 24*60*60, 7*24*60*60};
unsigned long c;
char *tail = buf;
int i;
for (i = ARRAY_SIZE(mult) - 1 ; i >= 0; i--) {
c = seconds / mult[i];
if (c > 0 || (i == 0 && buf == tail))
tail += scnprintf(tail, 40-(tail-buf), "%lu%c", c,
spec[i]);
seconds %= mult[i];
}
return tail;
}
static void sec2str(time_t seconds, char *buf, int rc)
{
char *tail = buf;
if (rc)
*tail++ = '[';
tail = __sec2str(seconds, tail);
if (rc && tail - buf < 39) {
*tail++ = ']';
*tail++ = 0;
}
}
static void diff2str(time_t seconds, char *buf, time_t now)
{
buf[0] = 0;
if (!seconds)
return;
if (seconds <= now) {
strcpy(buf, "expired");
return;
}
__sec2str(seconds - now, buf);
}
static void print_quota_val(const char *val, int cols, bool print_over,
struct quota_param *param)
{
if ((param->qp_detail & (param->qp_detail - 1)) == 0 &&
!param->qp_show_qid) /* single value */
printf("%s", val);
else if (param->qp_delim[0])
printf("%*s%s", cols, val, param->qp_delim);
else if (print_over) /* add a space to avoid big values connecting */
printf("%*s* ", cols - 2, val);
else
printf("%*s ", cols - 1, val);
}
static void print_quota_title(char *name, struct if_quotactl *qctl,
struct quota_param *param)
{
if (param->qp_quiet ||
qctl->qc_cmd == LUSTRE_Q_GETINFO ||
qctl->qc_cmd == LUSTRE_Q_GETINFOPOOL ||
qctl->qc_cmd == Q_GETOINFO)
return;
if (param->qp_show_qid) {
printf("Disk %s quotas\n", qtype_name(qctl->qc_type));
} else if (param->qp_show_default) {
printf("Disk default %s quota:\n", qtype_name(qctl->qc_type));
} else {
printf("Disk quotas for %s %s (%cid %u):\n",
qtype_name(qctl->qc_type), name,
*qtype_name(qctl->qc_type), qctl->qc_id);
}
if (param->qp_detail & QIF_FILESYSTEM)
print_quota_val("Filesystem", 16, false, param);
if (param->qp_show_qid)
print_quota_val("quota_id", 10, false, param);
if ((param->qp_detail & QIF_SPACE) && !param->qp_show_default)
print_quota_val(param->qp_human_readable ? "used" : "kbytes",
8, false, param);
if (param->qp_detail & QIF_BSOFTLIMIT)
print_quota_val("bquota", 8, false, param);
if (param->qp_detail & QIF_BHARDLIMIT)
print_quota_val("blimit", 8, false, param);
if (param->qp_detail & QIF_BTIME)
print_quota_val("bgrace", 8, false, param);
if ((param->qp_detail & QIF_INODES) && !param->qp_show_default)
print_quota_val("files", 8, false, param);
if (param->qp_detail & QIF_ISOFTLIMIT)
print_quota_val("iquota", 8, false, param);
if (param->qp_detail & QIF_IHARDLIMIT)
print_quota_val("ilimit", 8, false, param);
if (param->qp_detail & QIF_ITIME)
print_quota_val("igrace", 8, false, param);
printf("\n");
}
static void kbytes2str(__u64 num, char *buf, int buflen, bool h)
{
if (!h) {
snprintf(buf, buflen, "%ju", (uintmax_t)num);
} else {
if (num >> 40)
snprintf(buf, buflen, "%5.4gP",
(double)num / ((__u64)1 << 40));
else if (num >> 30)
snprintf(buf, buflen, "%5.4gT",
(double)num / (1 << 30));
else if (num >> 20)
snprintf(buf, buflen, "%5.4gG",
(double)num / (1 << 20));
else if (num >> 10)
snprintf(buf, buflen, "%5.4gM",
(double)num / (1 << 10));
else
snprintf(buf, buflen, "%ju%s", (uintmax_t)num, "k");
}
}
#ifdef HAVE_NATIVE_CLIENT
/* In the current Lustre implementation, the grace time is either the time
* or the timestamp to be used after some quota ID exceeds the soft limt,
* 48 bits should be enough, its high 16 bits can be used as quota flags.
*/
#define LQUOTA_GRACE_BITS 48
#define LQUOTA_GRACE_MASK ((1ULL << LQUOTA_GRACE_BITS) - 1)
#define LQUOTA_GRACE_MAX LQUOTA_GRACE_MASK
#define LQUOTA_GRACE(t) (t & LQUOTA_GRACE_MASK)
#define LQUOTA_FLAG(t) (t >> LQUOTA_GRACE_BITS)
#define LQUOTA_GRACE_FLAG(t, f) ((__u64)t | (__u64)f << LQUOTA_GRACE_BITS)
#endif
#define STRBUF_LEN 24
static void print_quota(char *mnt, struct if_quotactl *qctl, int type,
int rc, struct quota_param *param)
{
time_t now;
time(&now);
if (qctl->qc_cmd == LUSTRE_Q_GETQUOTA || qctl->qc_cmd == Q_GETOQUOTA ||
qctl->qc_cmd == LUSTRE_Q_GETQUOTAPOOL ||
qctl->qc_cmd == LUSTRE_Q_GETDEFAULT ||
qctl->qc_cmd == LUSTRE_Q_GETDEFAULT_POOL ||
qctl->qc_cmd == LUSTRE_Q_GETQUOTALQA) {
int bover = 0, iover = 0;
struct obd_dqblk *dqb = &qctl->qc_dqblk;
char numbuf[3][STRBUF_LEN + 2]; /* 2 for brackets or wildcard */
char timebuf[40];
char strbuf[STRBUF_LEN];
char *tmp;
dqb->dqb_btime &= LQUOTA_GRACE_MASK;
dqb->dqb_itime &= LQUOTA_GRACE_MASK;
if (dqb->dqb_bhardlimit &&
stoqb(dqb->dqb_curspace) >= dqb->dqb_bhardlimit) {
bover = 1;
} else if (dqb->dqb_bsoftlimit && dqb->dqb_btime) {
if (dqb->dqb_btime > now)
bover = 2;
else
bover = 3;
}
if (dqb->dqb_ihardlimit &&
dqb->dqb_curinodes >= dqb->dqb_ihardlimit) {
iover = 1;
} else if (dqb->dqb_isoftlimit && dqb->dqb_itime) {
if (dqb->dqb_itime > now)
iover = 2;
else
iover = 3;
}
tmp = strstr(mnt, "_UUID");
if (tmp)
*tmp = '\0';
if (param->qp_detail & QIF_FILESYSTEM) /* Filesystem */
print_quota_val(mnt, 16, false, param);
if (param->qp_show_qid) {
char namebuf[DEF_PW_SIZE_MAX];
char *name = NULL;
size_t name_max;
int rc2 = 0;
name_max = getpw_size_max();
if (name_max > sizeof(namebuf))
name = malloc(name_max);
if (!name) {
name = namebuf;
name_max = sizeof(namebuf);
}
if (param->qp_show_qid_num) {
name[0] = '\0';
goto use_qid_value;
}
if (qctl->qc_type == USRQUOTA) {
rc2 = uid2name(name, name_max, qctl->qc_id);
} else if (qctl->qc_type == GRPQUOTA) {
rc2 = gid2name(name, name_max, qctl->qc_id);
} else if (qctl->qc_type == PRJQUOTA) {
rc2 = prjid2name(name, name_max, qctl->qc_id);
}
use_qid_value:
if (rc2 || name[0] == '\0')
snprintf(name, sizeof(namebuf), "%u",
qctl->qc_id);
print_quota_val(name, 10, false, param);
if (name != namebuf)
free(name);
}
if (param->qp_show_default)
snprintf(timebuf, sizeof(timebuf), "%llu",
(unsigned long long)dqb->dqb_btime);
else if (bover)
diff2str(dqb->dqb_btime, timebuf, now);
kbytes2str(stoqb(dqb->dqb_curspace),
strbuf, sizeof(strbuf), param->qp_human_readable);
if (rc == -EREMOTEIO)
sprintf(numbuf[0], "%s*", strbuf);
else
sprintf(numbuf[0], (dqb->dqb_valid & QIF_SPACE) ?
"%s" : "[%s]", strbuf);
kbytes2str(dqb->dqb_bsoftlimit, strbuf, sizeof(strbuf),
param->qp_human_readable);
if (type == QC_GENERAL)
sprintf(numbuf[1], (dqb->dqb_valid & QIF_BLIMITS) ?
"%s" : "[%s]", strbuf);
else
sprintf(numbuf[1], "%s", "-");
kbytes2str(dqb->dqb_bhardlimit, strbuf, sizeof(strbuf),
param->qp_human_readable);
sprintf(numbuf[2], (dqb->dqb_valid & QIF_BLIMITS) ?
"%s" : "[%s]", strbuf);
if (bover <= 1)
sprintf(timebuf, "-");
if ((param->qp_detail & QIF_SPACE) && !param->qp_show_default)
print_quota_val(numbuf[0], 8, bover, param);
if (param->qp_detail & QIF_BSOFTLIMIT)
print_quota_val(numbuf[1], 8, false, param);
if (param->qp_detail & QIF_BHARDLIMIT)
print_quota_val(numbuf[2], 8, false, param);
if (param->qp_detail & QIF_BTIME)
print_quota_val(timebuf, 8, false, param);
if (param->qp_show_default)
snprintf(timebuf, sizeof(timebuf), "%llu",
(unsigned long long)dqb->dqb_itime);
else if (iover)
diff2str(dqb->dqb_itime, timebuf, now);
snprintf(numbuf[0], sizeof(numbuf),
(dqb->dqb_valid & QIF_INODES) ? "%ju" : "[%ju]",
(uintmax_t)dqb->dqb_curinodes);
if (type == QC_GENERAL)
sprintf(numbuf[1], (dqb->dqb_valid & QIF_ILIMITS) ?
"%ju" : "[%ju]",
(uintmax_t)dqb->dqb_isoftlimit);
else
sprintf(numbuf[1], "%s", "-");
sprintf(numbuf[2], (dqb->dqb_valid & QIF_ILIMITS) ?
"%ju" : "[%ju]", (uintmax_t)dqb->dqb_ihardlimit);
if (type == QC_OSTIDX || iover <= 1)
sprintf(timebuf, "-");
if ((param->qp_detail & QIF_INODES) && !param->qp_show_default)
print_quota_val(numbuf[0], 8, bover, param);
if (param->qp_detail & QIF_ISOFTLIMIT)
print_quota_val(numbuf[1], 8, false, param);
if (param->qp_detail & QIF_IHARDLIMIT)
print_quota_val(numbuf[2], 8, false, param);
if (param->qp_detail & QIF_ITIME)
print_quota_val(timebuf, 8, false, param);
printf("\n");
} else if (qctl->qc_cmd == LUSTRE_Q_GETINFO ||
qctl->qc_cmd == LUSTRE_Q_GETINFOPOOL ||
qctl->qc_cmd == LUSTRE_Q_GETINFOLQA ||
qctl->qc_cmd == Q_GETOINFO) {
char bgtimebuf[40];
char igtimebuf[40];
if (qctl->qc_dqinfo.dqi_bgrace == NOTIFY_GRACE_TIME)
strncpy(bgtimebuf, NOTIFY_GRACE, 40);
else
sec2str(qctl->qc_dqinfo.dqi_bgrace, bgtimebuf, rc);
if (qctl->qc_dqinfo.dqi_igrace == NOTIFY_GRACE_TIME)
strncpy(igtimebuf, NOTIFY_GRACE, 40);
else
sec2str(qctl->qc_dqinfo.dqi_igrace, igtimebuf, rc);
printf("Block grace time: %s; Inode grace time: %s\n",
bgtimebuf, igtimebuf);
}
}
static int tgt_name2index(const char *tgtname, unsigned int *idx)
{
const char *dash;
char *endp;
/* format is "lustre-OST0001" */
dash = memchr(tgtname, '-', LUSTRE_MAXFSNAME + 1);
if (!dash) {
fprintf(stderr, "wrong tgtname format '%s'\n", tgtname);
return -EINVAL;
}
dash += 4;
*idx = strtoul(dash, &endp, 16);
if (*idx > LOV_V1_INSANE_STRIPE_INDEX) {
fprintf(stderr, "wrong index %s\n", tgtname);
return -ERANGE;
}
return 0;
}
static int print_obd_quota(char *mnt, struct if_quotactl *qctl, int is_mdt,
struct quota_param *param, __u64 *total)
{
int rc = 0, rc1 = 0, count = 0, i = 0;
char **list = NULL, *buffer = NULL;
__u32 valid = qctl->qc_valid;
if (qctl->qc_cmd == LUSTRE_Q_GETQUOTAPOOL && is_mdt)
return 0;
/* Is it correct for the case OST0000, OST0002, OST0003 -
* we will ask OST0001 that is absent and won't ask OST0003? */
rc = llapi_get_obd_count(mnt, &count, is_mdt);
if (rc) {
fprintf(stderr, "can not get %s count: %s\n",
is_mdt ? "mdt" : "ost", strerror(-rc));
return rc;
}
if (qctl->qc_cmd == LUSTRE_Q_GETQUOTAPOOL) {
char fname[PATH_MAX];
char fsname[LUSTRE_MAXFSNAME + 1];
int bufsize = sizeof(struct obd_uuid) * count;
rc = llapi_search_fsname(mnt, fsname);
if (rc) {
fprintf(stderr, "cannot get fsname for mountpoint %s\n",
mnt);
goto out;
}
buffer = malloc(bufsize + sizeof(*list) * count);
if (!buffer)
return -ENOMEM;
list = (char **)(buffer + bufsize);
snprintf(fname, PATH_MAX, "%s.%s", fsname, qctl->qc_poolname);
count = llapi_get_poolmembers(fname, list, count,
buffer, bufsize);
if (count <= 0)
goto out;
}
for (i = 0; i < count; i++) {
char name[UUID_MAX+8];
if (qctl->qc_cmd == LUSTRE_Q_GETQUOTAPOOL) {
unsigned int index;
if (tgt_name2index(list[i], &index))
continue;
qctl->qc_idx = index;
} else {
qctl->qc_idx = i;
}
qctl->qc_valid = is_mdt ? QC_MDTIDX : QC_OSTIDX;
rc = llapi_quotactl(mnt, qctl);
if (rc) {
/* It is remote client case. */
if (rc == -EOPNOTSUPP) {
rc = 0;
goto out;
}
/* no target for this index yet */
if (rc == -ENODEV) {
rc = 0;
continue;
}
/* inactive target */
if (rc == -ENODATA) {
snprintf(name, sizeof(name), "%s[inact]",
obd_uuid2str(&qctl->obd_uuid));
memset(&qctl->qc_dqinfo, 0,
sizeof(qctl->qc_dqinfo));
memset(&qctl->qc_dqblk, 0,
sizeof(qctl->qc_dqblk));
print_quota(name, qctl, qctl->qc_valid, 0,
param);
rc = 0;
continue;
}
if (!rc1)
rc1 = rc;
fprintf(stderr, "quotactl %s%d failed.\n",
is_mdt ? "mdt" : "ost", qctl->qc_idx);
continue;
}
snprintf(name, sizeof(name), "%s",
obd_uuid2str(&qctl->obd_uuid));
print_quota(name, qctl, qctl->qc_valid, 0, param);
*total += is_mdt ? qctl->qc_dqblk.dqb_ihardlimit :
qctl->qc_dqblk.dqb_bhardlimit;
}
out:
if (buffer)
free(buffer);
qctl->qc_valid = valid;
return rc ? : rc1;
}
static int print_one_quota(char *mnt, char *name, struct if_quotactl *qctl,
struct quota_param *param, int rc)
{
int rc1 = 0, rc2 = 0;
char *obd_type = (char *)qctl->obd_type;
char *obd_uuid = (char *)qctl->obd_uuid.uuid;
__u64 total_ialloc = 0, total_balloc = 0;
int inacc;
if (qctl->qc_dqblk.dqb_valid & QIF_BTIME &&
LQUOTA_FLAG(qctl->qc_dqblk.dqb_btime) & LQUOTA_FLAG_DEFAULT)
qctl->qc_dqblk.dqb_btime &= LQUOTA_GRACE_MASK;
if (qctl->qc_dqblk.dqb_valid & QIF_ITIME &&
LQUOTA_FLAG(qctl->qc_dqblk.dqb_itime) & LQUOTA_FLAG_DEFAULT)
qctl->qc_dqblk.dqb_itime &= LQUOTA_GRACE_MASK;
if (rc && *obd_type)
fprintf(stderr, "%s %s ", obd_type, obd_uuid);
if (qctl->qc_valid != QC_GENERAL)
mnt = "";
inacc = (qctl->qc_cmd == LUSTRE_Q_GETQUOTA ||
qctl->qc_cmd == LUSTRE_Q_GETQUOTAPOOL) &&
((qctl->qc_dqblk.dqb_valid & (QIF_LIMITS|QIF_USAGE)) !=
(QIF_LIMITS|QIF_USAGE));
inacc |= (qctl->qc_cmd == LUSTRE_Q_GETQUOTALQA &&
!(qctl->qc_dqblk.dqb_valid & QIF_LIMITS));
print_quota(mnt, qctl, QC_GENERAL, rc, param);
if (!param->qp_show_qid && !param->qp_show_default &&
param->qp_verbose && qctl->qc_valid == QC_GENERAL &&
qctl->qc_cmd != LUSTRE_Q_GETINFO &&
qctl->qc_cmd != LUSTRE_Q_GETINFOPOOL &&
qctl->qc_cmd != LUSTRE_Q_GETINFOLQA) {
char strbuf[STRBUF_LEN];
rc1 = print_obd_quota(mnt, qctl, 1, param,
&total_ialloc);
rc2 = print_obd_quota(mnt, qctl, 0, param,
&total_balloc);
kbytes2str(total_balloc, strbuf, sizeof(strbuf),
param->qp_human_readable);
printf("Total allocated inode limit: %ju, total allocated block limit: %s\n",
(uintmax_t)total_ialloc, strbuf);
}
if (!param->qp_show_qid && (rc || rc1 || rc2 || inacc))
printf("%d Some errors happened when getting quota info. Some devices may be not working or deactivated. The data in \"[]\" is inaccurate.\n", inacc);
if (rc)
return rc;
if (rc1)
return rc1;
if (rc2)
return rc2;
if (inacc)
return -EIO;
return 0;
}
static int iter_all_quota(char *mnt, struct if_quotactl *qctl,
struct quota_param *param)
{
struct if_quotactl qctl_tmp, *qctl_iter;
void *buffer = NULL;
__u64 mark;
__u64 cur, buflen = 0;
int rc = 0;
memcpy(&qctl_tmp, qctl, sizeof(struct if_quotactl));
qctl_tmp.qc_cmd = LUSTRE_Q_ITERQUOTA;
rc = llapi_quotactl(mnt, &qctl_tmp);
if (rc)
goto out;
buflen = qctl_tmp.qc_allquota_count * sizeof(struct if_quotactl);
buffer = malloc(buflen);
if (buffer == NULL) {
rc = -ENOMEM;
goto out;
}
mark = qctl_tmp.qc_allquota_mark;
memcpy(&qctl_tmp, qctl, sizeof(struct if_quotactl));
qctl_tmp.qc_cmd = LUSTRE_Q_GETALLQUOTA;
qctl_tmp.qc_allquota_buffer = (__u64)buffer;
qctl_tmp.qc_allquota_buflen = buflen;
qctl_tmp.qc_allquota_mark = mark;
rc = llapi_quotactl(mnt, &qctl_tmp);
if (rc)
goto out;
cur = 0;
while (cur < buflen) {
if ((buflen - cur) < sizeof(struct if_quotactl)) {
rc = -EFAULT;
break;
}
qctl_iter = buffer + cur;
qctl_iter->qc_cmd = LUSTRE_Q_GETQUOTA;
cur += sizeof(struct if_quotactl);
/* Skip root quota entry unless --show-root is specified */
if (qctl_iter->qc_id == 0 && !param->qp_show_root)
continue;
/* Is no file created for this quota ID yet? */
if ((qctl_iter->qc_dqblk.dqb_valid & QIF_USAGE) != QIF_USAGE)
qctl_iter->qc_dqblk.dqb_valid |= QIF_USAGE;
print_one_quota(mnt, NULL, qctl_iter, param, 0);
}
out:
if (buffer != NULL)
free(buffer);
if (rc)
fprintf(stderr, "get all quota failed %d\n", rc);
return rc;
}
static int get_print_quota(char *mnt, struct if_quotactl *qctl,
struct quota_param *param)
{
char *name = NULL;
int rc;
rc = llapi_quotactl(mnt, qctl);
if (rc < 0) {
switch (rc) {
case -ESRCH:
fprintf(stderr, "%s quotas are not enabled.\n",
qtype_name(qctl->qc_type));
break;
case -EPERM:
fprintf(stderr, "Permission denied.\n");
case -ENODEV:
case -ENOENT:
/* We already got error message. */
break;
default:
fprintf(stderr, "Unexpected quotactl error: %s\n",
strerror(-rc));
}
return rc;
}
return print_one_quota(mnt, name, qctl, param, rc);
}
static int lfs_project(int argc, char **argv)
{
int ret = 0, err = 0, c, i;
struct project_handle_control phc = { 0 };
enum lfs_project_ops_t op;
struct option long_opts[] = {
{ .val = '0', .name = "print0", .has_arg = no_argument },
{ .val = 'c', .name = "check", .has_arg = no_argument },
{ .val = 'C', .name = "clear", .has_arg = no_argument },
{ .val = 'd', .name = "directory", .has_arg = no_argument },
{ .val = 'k', .name = "keep", .has_arg = no_argument },
{ .val = 'p', .name = "proj-inherit", .has_arg = required_argument },
{ .val = 'r', .name = "recursive", .has_arg = no_argument },
{ .val = 's', .name = "set-inherit", .has_arg = no_argument },
{ .val = 's', .name = "inherit", .has_arg = no_argument },
};
phc.newline = true;
phc.assign_projid = false;
/* default action */
op = LFS_PROJECT_LIST;
while ((c = getopt_long(argc, argv, "0cCdkp:rs",
long_opts, NULL)) != -1) {
switch (c) {
case '0':
phc.newline = false;
break;
case 'c':
if (op != LFS_PROJECT_LIST) {
fprintf(stderr,
"%s: cannot specify '-c' '-C' '-s' together\n",
progname);
return CMD_HELP;
}
op = LFS_PROJECT_CHECK;
break;
case 'C':
if (op != LFS_PROJECT_LIST) {
fprintf(stderr,
"%s: cannot specify '-c' '-C' '-s' together\n",
progname);
return CMD_HELP;
}
op = LFS_PROJECT_CLEAR;
break;
case 'd':
phc.dironly = true;
break;
case 'k':
phc.keep_projid = true;
break;
case 'p':
if (name2projid(&phc.projid, optarg) &&
str2quotaid(&phc.projid, optarg)) {
fprintf(stderr, "invalid project id '%s'\n",
optarg);
return CMD_HELP;
}
phc.assign_projid = true;
break;
case 'r':
phc.recursive = true;
break;
case 's':
if (op != LFS_PROJECT_LIST) {
fprintf(stderr,
"%s: cannot specify '-c' '-C' '-s' together\n",
progname);
return CMD_HELP;
}
phc.set_inherit = true;
op = LFS_PROJECT_SET;
break;
default:
fprintf(stderr, "%s: invalid option '%c'\n",
progname, optopt);
return CMD_HELP;
}
}
if (phc.assign_projid && op == LFS_PROJECT_LIST) {
op = LFS_PROJECT_SET;
phc.set_projid = true;
} else if (phc.assign_projid && op == LFS_PROJECT_SET) {
phc.set_projid = true;
}
switch (op) {
case LFS_PROJECT_CHECK:
if (phc.keep_projid) {
fprintf(stderr,
"%s: '-k' is useless together with '-c'\n",
progname);
return CMD_HELP;
}
break;
case LFS_PROJECT_CLEAR:
if (!phc.newline) {
fprintf(stderr,
"%s: '-0' is useless together with '-C'\n",
progname);
return CMD_HELP;
}
if (phc.assign_projid) {
fprintf(stderr,
"%s: '-p' is useless together with '-C'\n",
progname);
return CMD_HELP;
}
break;
case LFS_PROJECT_SET:
if (!phc.newline) {
fprintf(stderr,
"%s: '-0' is useless together with '-s'\n",
progname);
return CMD_HELP;
}
if (phc.keep_projid) {
fprintf(stderr,
"%s: '-k' is useless together with '-s'\n",
progname);
return CMD_HELP;
}
break;
default:
if (!phc.newline) {
fprintf(stderr,
"%s: '-0' is useless for list operations\n",
progname);
return CMD_HELP;
}
break;
}
argv += optind;
argc -= optind;
if (argc == 0) {
fprintf(stderr, "%s: missing file or directory target(s)\n",
progname);
return CMD_HELP;
}
for (i = 0; i < argc; i++) {
switch (op) {
case LFS_PROJECT_CHECK:
err = lfs_project_check(argv[i], &phc);
break;
case LFS_PROJECT_LIST:
err = lfs_project_list(argv[i], &phc);
break;
case LFS_PROJECT_CLEAR:
err = lfs_project_clear(argv[i], &phc);
break;
case LFS_PROJECT_SET:
err = lfs_project_set(argv[i], &phc);
break;
default:
break;
}
if (err && !ret)
ret = err;
}
return ret;
}
static int do_quota_op(char *mnt, struct if_quotactl *qctl,
struct quota_param *param)
{
struct if_quotactl *qctl_tmp;
char **poollist = NULL;
char *buf = NULL;
int poolcount, extra, i, rc = 0;
extra = max(LOV_MAXPOOLNAME + 1, LQA_NAME_MAX + 1);
/* avoid modifying the original qctl */
qctl_tmp = malloc(sizeof(*qctl_tmp) + extra);
memcpy(qctl_tmp, qctl, sizeof(*qctl_tmp) + extra);
if (qctl_tmp->qc_cmd == LUSTRE_Q_ITERQUOTA) {
rc = iter_all_quota(mnt, qctl_tmp, param);
goto out;
}
if (param->qp_show_pools) {
char *p;
rc = llapi_get_poolbuf(mnt, &buf, &poollist, &poolcount);
if (rc)
goto out;
for (i = 0; i < poolcount; i++) {
p = memchr(poollist[i], '.', MAXNAMLEN);
if (!p) {
fprintf(stderr, "bad string format %.*s\n",
MAXNAMLEN, poollist[i]);
rc = -EINVAL;
goto out;
}
p++;
printf("Quotas for pool: %s\n", p);
snprintf(qctl_tmp->qc_poolname, LOV_MAXPOOLNAME + 1,
"%s", p);
rc = get_print_quota(mnt, qctl_tmp, param);
if (rc)
goto out;
}
}
rc = get_print_quota(mnt, qctl_tmp, param);
goto out;
out:
free(qctl_tmp);
free(buf);
return rc;
}
static int lfs_quota(int argc, char **argv)
{
struct quota_param param = {
.qp_valid = QC_GENERAL,
.qp_delim = "",
};
struct if_quotactl *qctl;
char *obd_uuid, *endp, *name = NULL;
char namebuf[DEF_PW_SIZE_MAX];
size_t name_max;
__u32 start_qid = 0, end_qid = 0;
int c, qtype, qctl_len, rc = 0;
long idx = 0;
bool all = false;
struct option long_opts[] = {
{ .val = 'a', .name = "all", .has_arg = required_argument },
{ .val = 'e', .name = "end-qid", .has_arg = required_argument },
{ .val = 'd', .name = "delimiter", .has_arg = required_argument },
{ .val = LFS_QUOTA_DEFAULT_OPT,
.name = "default", .has_arg = no_argument },
{ .val = 'g', .name = "group", .has_arg = required_argument },
{ .val = 'G', .name = "default-grp", .has_arg = no_argument },
{ .val = 'h', .name = "human-readable", .has_arg = no_argument },
/* It is unfortunate that '-i' was used for mdt-index, and '-I' for
* ost-index, because '-i' is used for ost-index everywhere else. '-i'
* and '-I' are also used for soft/hard inode quotas in lfs_setquota().
* These options have been this way since ancient days, but I suspect
* that they are not often used. Prefer --ost and --mdt instead.
*/
{ .val = 'm', .name = "mdt-index", .has_arg = required_argument },
{ .val = 'm', .name = "mdt", .has_arg = required_argument },
{ .val = 'n', .name = "num", .has_arg = no_argument },
{ .val = 'o', .name = "ost-index", .has_arg = required_argument },
{ .val = 'o', .name = "ost", .has_arg = required_argument },
{ .val = LFS_LQA_OPT, .name = "lqa", .has_arg = required_argument },
{ .val = LFS_POOL_OPT, .name = "pool", .has_arg = optional_argument },
{ .val = 'p', .name = "projid", .has_arg = required_argument },
{ .val = 'P', .name = "default-prj", .has_arg = no_argument },
{ .val = 'q', .name = "quiet", .has_arg = no_argument },
{ .val = 's', .name = "start-qid", .has_arg = required_argument },
{ .val = 't', .name = "times", .has_arg = no_argument },
{ .val = 'u', .name = "user", .has_arg = required_argument },
{ .val = 'U', .name = "default-usr", .has_arg = required_argument },
{ .val = 'v', .name = "verbose", .has_arg = no_argument },
{ .val = LFS_QUOTA_FILESYSTEM_OPT,
.name = "filesystem", .has_arg = no_argument },
{ .val = LFS_QUOTA_FILESYSTEM_OPT,
.name = "mount-point", .has_arg = no_argument },
{ .val = LFS_QUOTA_SPACE_OPT,
.name = "blocks", .has_arg = no_argument },
{ .val = LFS_QUOTA_SPACE_OPT,
.name = "busage", .has_arg = no_argument },
{ .val = LFS_QUOTA_SPACE_OPT,
.name = "space", .has_arg = no_argument },
{ .val = LFS_QUOTA_SPACE_OPT,
.name = "kbytes", .has_arg = no_argument },
{ .val = 'b', .name = "block-softlimit", .has_arg = no_argument },
{ .val = 'b', .name = "bsoftlimit", .has_arg = no_argument },
{ .val = 'b', .name = "bquota", .has_arg = no_argument },
{ .val = 'B', .name = "block-hardlimit", .has_arg = no_argument },
{ .val = 'B', .name = "bhardlimit", .has_arg = no_argument },
{ .val = 'B', .name = "blimit", .has_arg = no_argument },
{ .val = LFS_QUOTA_BGRACE_OPT,
.name = "block-grace", .has_arg = no_argument },
{ .val = LFS_QUOTA_BGRACE_OPT,
.name = "bgrace", .has_arg = no_argument },
{ .val = LFS_QUOTA_BGRACE_OPT,
.name = "btime", .has_arg = no_argument },
{ .val = LFS_QUOTA_INODES_OPT,
.name = "inodes", .has_arg = no_argument },
{ .val = LFS_QUOTA_INODES_OPT,
.name = "iusage", .has_arg = no_argument },
{ .val = LFS_QUOTA_INODES_OPT,
.name = "files", .has_arg = no_argument },
{ .val = LFS_QUOTA_ISOFTLIMIT_OPT,
.name = "inode-softlimit", .has_arg = no_argument },
{ .val = LFS_QUOTA_ISOFTLIMIT_OPT,
.name = "isoftlimit", .has_arg = no_argument },
{ .val = LFS_QUOTA_ISOFTLIMIT_OPT,
.name = "iquota", .has_arg = no_argument },
{ .val = LFS_QUOTA_IHARDLIMIT_OPT,
.name = "inode-hardlimit", .has_arg = no_argument },
{ .val = LFS_QUOTA_IHARDLIMIT_OPT,
.name = "ihardlimit", .has_arg = no_argument },
{ .val = LFS_QUOTA_IHARDLIMIT_OPT,
.name = "ilimit", .has_arg = no_argument },
{ .val = LFS_QUOTA_IGRACE_OPT,
.name = "inode-grace", .has_arg = no_argument },
{ .val = LFS_QUOTA_IGRACE_OPT,
.name = "igrace", .has_arg = no_argument },
{ .val = LFS_QUOTA_IGRACE_OPT,
.name = "itime", .has_arg = no_argument },
{ .val = LFS_QUOTA_SHOW_ROOT_OPT,
.name = "show-root", .has_arg = no_argument },
{ .name = NULL } };
qctl_len = sizeof(*qctl) + max(LOV_MAXPOOLNAME + 1, LQA_NAME_MAX + 1);
qctl = calloc(1, qctl_len);
if (!qctl)
return -ENOMEM;
name_max = getpw_size_max();
if (name_max > sizeof(namebuf))
name = malloc(name_max);
if (!name) {
name = namebuf;
name_max = sizeof(namebuf);
}
qctl->qc_cmd = LUSTRE_Q_GETQUOTA;
qctl->qc_type = ALLQUOTA;
obd_uuid = (char *)qctl->obd_uuid.uuid;
while ((c = getopt_long(argc, argv, "abBe:gGhi:I:m:no:pPqs:tuUv",
long_opts, NULL)) != -1) {
switch (c) {
case 'a':
param.qp_show_qid = 1;
qctl->qc_cmd = LUSTRE_Q_ITERQUOTA;
break;
case 'd':
if (optarg == NULL || *optarg == '\0') {
fprintf(stderr,
"%s quota: invalid delimiter\n",
progname);
rc = CMD_HELP;
goto out;
}
param.qp_delim = optarg;
break;
case 'b':
param.qp_detail |= QIF_BSOFTLIMIT;
break;
case 'B':
param.qp_detail |= QIF_BHARDLIMIT;
break;
case 'e':
if (optarg == NULL || *optarg == '\0') {
fprintf(stderr,
"%s quota: invalid end quota ID\n",
progname);
rc = CMD_HELP;
goto out;
}
end_qid = strtoul(optarg, NULL, 0);
break;
case LFS_QUOTA_DEFAULT_OPT:
param.qp_show_default = 1;
break;
case 'G':
param.qp_show_default = 1;
/* fallthrough */
case 'g':
qtype = GRPQUOTA;
goto quota_type;
case 'h':
param.qp_human_readable = 1;
break;
#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(2, 22, 53, 0)
case 'i':
fprintf(stderr,
"'-i' deprecated, use '--ost' or '--mdt'\n");
/* fallthrough */
#endif
case 'm':
errno = 0;
idx = strtol(optarg, &endp, 0);
if (errno != 0 || idx > LOV_V1_INSANE_STRIPE_INDEX ||
idx < 0 || *endp != '\0') {
fprintf(stderr,
"%s quota: invalid MDT index '%s'\n",
progname, optarg);
rc = CMD_HELP;
goto out;
}
param.qp_valid = qctl->qc_valid = QC_MDTIDX;
qctl->qc_idx = idx;
break;
case 'n':
param.qp_show_qid_num = 1;
break;
#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(2, 22, 53, 0)
case 'I':
fprintf(stderr, "'-I' deprecated, use '--ost'\n");
/* fallthrough */
#endif
case 'o':
errno = 0;
idx = strtol(optarg, &endp, 0);
/* simple digit, treat it as a numerical OST index */
if (*endp == '\0') {
if (idx > LOV_V1_INSANE_STRIPE_INDEX ||
idx < 0) {
fprintf(stderr,
"%s quota: invalid OST index '%s'\n",
progname, optarg);
rc = CMD_HELP;
goto out;
}
param.qp_valid = qctl->qc_valid = QC_OSTIDX;
qctl->qc_idx = idx;
break;
}
/* need to also handle a UUID for compatibility */
param.qp_valid = qctl->qc_valid = QC_UUID;
snprintf(obd_uuid, UUID_MAX, "%s", optarg);
break;
case 'P':
param.qp_show_default = 1;
/* fallthrough */
case 'p':
qtype = PRJQUOTA;
goto quota_type;
case LFS_LQA_OPT:
if (!param.qp_show_default ||
qctl->qc_type == ALLQUOTA) {
fprintf(stderr, "Specify LQA quota type with -U,-G, or -P\n");
rc = CMD_HELP;
goto out;
}
if (lfs_lqaarg_insane(optarg)) {
rc = -1;
goto out;
}
snprintf(qctl->qc_lqaname, LQA_NAME_MAX + 1, "%s",
optarg);
qctl->qc_cmd = LUSTRE_Q_GETQUOTALQA;
qctl->qc_id = 1;
break;
case LFS_POOL_OPT:
if ((!optarg) && (argv[optind] != NULL) &&
(argv[optind][0] != '-') &&
(argv[optind][0] != '/')) {
optarg = argv[optind++];
if (lfs_poolarg_insane(optarg)) {
rc = -EINVAL;
goto out;
}
snprintf(qctl->qc_poolname,
LOV_MAXPOOLNAME + 1, "%s", optarg);
if (qctl->qc_cmd == LUSTRE_Q_GETINFO)
qctl->qc_cmd = LUSTRE_Q_GETINFOPOOL;
else
qctl->qc_cmd = LUSTRE_Q_GETQUOTAPOOL;
break;
}
/* optarg is NULL */
param.qp_show_pools = 1;
qctl->qc_cmd = LUSTRE_Q_GETQUOTAPOOL;
break;
case 'q':
param.qp_quiet = 1;
break;
case 's':
if (optarg == NULL || *optarg == '\0') {
fprintf(stderr,
"%s quota: invalid start quota ID\n",
progname);
rc = CMD_HELP;
goto out;
}
start_qid = strtoul(optarg, NULL, 0);
break;
case 't':
qctl->qc_cmd = LUSTRE_Q_GETINFO;
break;
case 'U':
param.qp_show_default = 1;
/* fallthrough */
case 'u':
qtype = USRQUOTA;
quota_type:
/*
* since ID is not required for when -a or -t is used
* it is only set after all options have been processed
*/
if (qctl->qc_type != ALLQUOTA) {
fprintf(stderr,
"%s quota: only one of -u, -g, or -p may be specified\n",
progname);
rc = CMD_HELP;
goto out;
}
qctl->qc_type = qtype;
break;
case 'v':
param.qp_verbose = 1;
break;
case LFS_QUOTA_FILESYSTEM_OPT:
param.qp_detail |= QIF_FILESYSTEM;
break;
case LFS_QUOTA_SPACE_OPT:
param.qp_detail |= QIF_SPACE;
break;
case LFS_QUOTA_BGRACE_OPT:
param.qp_detail |= QIF_BTIME;
break;
case LFS_QUOTA_INODES_OPT:
param.qp_detail |= QIF_INODES;
break;
case LFS_QUOTA_ISOFTLIMIT_OPT:
param.qp_detail |= QIF_ISOFTLIMIT;
break;
case LFS_QUOTA_IHARDLIMIT_OPT:
param.qp_detail |= QIF_IHARDLIMIT;
break;
case LFS_QUOTA_IGRACE_OPT:
param.qp_detail |= QIF_ITIME;
break;
case LFS_QUOTA_SHOW_ROOT_OPT:
param.qp_show_root = 1;
break;
default:
fprintf(stderr, "%s quota: unrecognized option '%s'\n",
progname, argv[optind - 1]);
rc = CMD_HELP;
goto out;
}
}
if (!param.qp_detail)
param.qp_detail = QIF_ALL_DETAIL;
if (param.qp_show_root && qctl->qc_cmd != LUSTRE_Q_ITERQUOTA) {
fprintf(stderr, "%s quota: --show-root must be used with -a\n",
progname);
rc = CMD_HELP;
goto out;
}
if (qctl->qc_cmd == LUSTRE_Q_ITERQUOTA) {
if (qctl->qc_type == ALLQUOTA) {
fprintf(stderr, "%s quota: no quota type to iterate\n",
progname);
rc = CMD_HELP;
goto out;
}
if (end_qid != 0 && start_qid > end_qid) {
fprintf(stderr,
"%s quota: end qid is smaller than start qid\n",
progname);
rc = CMD_HELP;
goto out;
}
qctl->qc_allquota_qid_start = start_qid;
qctl->qc_allquota_qid_end = end_qid;
} else if (qctl->qc_type != ALLQUOTA &&
(qctl->qc_cmd == LUSTRE_Q_GETQUOTA ||
qctl->qc_cmd == LUSTRE_Q_GETQUOTAPOOL)) {
char *argname = "<unknown>";
if (!param.qp_show_default) {
if (optind >= argc) {
fprintf(stderr,
"%s quota: u/g/p-name is required\n",
progname);
rc = CMD_HELP;
goto out;
}
argname = argv[optind++];
switch (qctl->qc_type) {
case USRQUOTA:
rc = name2uid(&qctl->qc_id, argname);
break;
case GRPQUOTA:
rc = name2gid(&qctl->qc_id, argname);
break;
case PRJQUOTA:
rc = name2projid(&qctl->qc_id, argname);
break;
default:
rc = -ENOTSUP;
break;
}
} else {
int cmd = qctl->qc_cmd;
qctl->qc_valid = QC_GENERAL;
if (cmd != LUSTRE_Q_GETQUOTALQA)
qctl->qc_cmd = cmd == LUSTRE_Q_GETQUOTAPOOL ?
LUSTRE_Q_GETDEFAULT_POOL :
LUSTRE_Q_GETDEFAULT;
}
if (rc) {
if (str2quotaid(&qctl->qc_id, argname)) {
fprintf(stderr, "%s quota: invalid id '%s'\n",
progname, argname);
rc = CMD_HELP;
goto out;
}
}
} else if (qctl->qc_type == ALLQUOTA) {
all = true;
qctl->qc_type = USRQUOTA;
}
do {
qctl->qc_valid = param.qp_valid;
switch (qctl->qc_type) {
case USRQUOTA:
if (all)
qctl->qc_id = geteuid();
rc = uid2name(name, name_max, qctl->qc_id);
break;
case GRPQUOTA:
if (all)
qctl->qc_id = getegid();
rc = gid2name(name, name_max, qctl->qc_id);
break;
case PRJQUOTA:
if (all)
qctl->qc_id = geteuid();
rc = prjid2name(name, name_max, qctl->qc_id);
break;
default:
sprintf(name, "<unknown>");
/* will print root quota */
memset(&qctl->qc_dqblk, 0,
sizeof(qctl->qc_dqblk));
break;
}
if (rc)
sprintf(name, "<unknown>");
print_quota_title(name, qctl, ¶m);
if (optind == argc) {
char mnt[PATH_MAX];
int i = 0;
while (!llapi_search_mounts(NULL, i++, mnt, NULL)) {
if (mnt[0] == '\0')
continue;
rc = do_quota_op(mnt, qctl, ¶m);
if (rc)
break;
mnt[0] = '\0'; /* avoid matching in next loop */
}
} else {
int i = optind;
while (i < argc) {
rc = do_quota_op(argv[i++], qctl, ¶m);
if (rc)
break;
}
}
} while (all && ++qctl->qc_type <= GRPQUOTA);
out:
if (name != namebuf)
free(name);
free(qctl);
return rc;
}
#endif /* HAVE_SYS_QUOTA_H! */
static int flushctx_ioctl(char *mp)
{
int fd, rc;
fd = open(mp, O_RDONLY);
if (fd == -1) {
fprintf(stderr, "flushctx: error open %s: %s\n",
mp, strerror(errno));
return -1;
}
rc = ioctl(fd, LL_IOC_FLUSHCTX);
if (rc == -1)
fprintf(stderr, "flushctx: error ioctl %s: %s\n",
mp, strerror(errno));
close(fd);
return rc;
}
static int lfs_flushctx(int argc, char **argv)
{
int kdestroy = 0, reap = 0, c;
char **mnts = NULL, **mnt_ptr;
int mnt_num = 1, index = 0, rc = 0, rc2;
extern char **environ;
while ((c = getopt(argc, argv, "kr")) != -1) {
switch (c) {
case 'k':
kdestroy = 1;
break;
case 'r':
reap = 1;
break;
default:
fprintf(stderr,
"error: %s: option '-%c' unrecognized\n",
argv[0], c);
return CMD_HELP;
}
}
if (optind >= argc) {
/* flush for all lustre mount points */
again:
mnt_ptr = realloc(mnts, mnt_num * sizeof(char *));
if (!mnt_ptr) {
mnt_num--;
rc = -ENOMEM;
goto reap;
}
mnts = mnt_ptr;
mnts[mnt_num - 1] = (char *)calloc(PATH_MAX + 1, sizeof(char));
if (!mnts[mnt_num - 1]) {
rc = -ENOMEM;
goto reap;
}
next:
if (!llapi_search_mounts(NULL, index++,
mnts[mnt_num - 1], NULL)) {
if (*mnts[mnt_num - 1] == '\0')
goto next;
mnt_num++;
goto again;
} else {
*mnts[mnt_num - 1] = '\0';
}
mnt_ptr = mnts;
index = 0;
} else {
/* flush for mounts as specified on command line */
mnt_ptr = argv + optind;
mnt_num = argc - optind;
}
for (index = 0; index < mnt_num; index++) {
/* Check if we have a mount point */
if (*mnt_ptr[index] == '\0')
continue;
rc2 = flushctx_ioctl(mnt_ptr[index]);
if (rc2) {
rc2 = -errno;
fprintf(stderr,
"error flushing contexts on mount point %s: %s\n",
mnt_ptr[index], strerror(errno));
rc = rc ? rc : rc2;
}
}
reap:
if (reap) {
static char *args[] = { "keyctl", "reap", NULL };
/* use callvpe to bypass the shell */
rc2 = callvpe("keyctl", args, environ);
if (rc2) {
rc2 = WEXITSTATUS(rc2);
fprintf(stderr, "error reaping keyring: %d\n", rc2);
}
}
if (kdestroy) {
static char *args[] = { "kdestroy", NULL };
/* use callvpe to bypass the shell */
rc2 = callvpe("kdestroy", args, environ);
if (rc2) {
rc2 = WEXITSTATUS(rc2);
fprintf(stderr, "error destroying tickets: %d\n", rc2);
}
}
if (mnts) {
for (index = 0; index < mnt_num; index++)
free(mnts[index]);
free(mnts);
}
return rc;
}
static int lfs_changelog(int argc, char **argv)
{
void *changelog_priv;
struct changelog_rec *rec;
long long startrec = 0, endrec = 0;
char *clid = NULL;
char *mask_str = NULL;
__u64 mask = 0;
char *mdd;
struct option long_opts[] = {
{ .val = 'f', .name = "follow", .has_arg = no_argument },
{ .val = 'm', .name = "mask", .has_arg = required_argument },
{ .val = 'u', .name = "user", .has_arg = required_argument },
{ .name = NULL } };
char short_opts[] = "fm:u:";
int rc, follow = 0;
while ((rc = getopt_long(argc, argv, short_opts,
long_opts, NULL)) != -1) {
switch (rc) {
case 'f':
follow++;
break;
case 'm':
mask_str = optarg;
/* convert mask string to numeric value */
mask = CHANGELOG_DEFMASK;
rc = llapi_convert_str2mask(mask_str,
changelog_type2str,
&mask, CHANGELOG_MINMASK,
CHANGELOG_ALLMASK,
CHANGELOG_DEFMASK);
if (rc != 0) {
llapi_error(LLAPI_MSG_ERROR, rc,
"invalid changelog mask '%s'\n",
mask_str);
return CMD_HELP;
}
break;
case 'u':
clid = optarg;
break;
default:
fprintf(stderr,
"%s changelog: unrecognized option '%s'\n",
progname, argv[optind - 1]);
return CMD_HELP;
}
}
if (optind >= argc) {
fprintf(stderr, "%s changelog: mdtname must be specified\n",
progname);
return CMD_HELP;
}
mdd = argv[optind++];
if (argc > optind) {
errno = 0;
startrec = strtoll(argv[optind++], NULL, 10);
if (errno != 0 || startrec < 0) {
fprintf(stderr,
"%s changelog: bad startrec\n",
progname);
return CMD_HELP;
}
}
if (argc > optind) {
errno = 0;
endrec = strtoll(argv[optind++], NULL, 10);
if (errno != 0 || endrec < 0) {
fprintf(stderr,
"%s changelog: bad endrec\n",
progname);
return CMD_HELP;
}
}
if (clid)
rc = llapi_changelog_start_user(&changelog_priv,
CHANGELOG_FLAG_BLOCK |
CHANGELOG_FLAG_JOBID |
CHANGELOG_FLAG_NID_BE |
CHANGELOG_FLAG_EXTRA_FLAGS |
(follow ?
CHANGELOG_FLAG_FOLLOW : 0),
mdd, clid, mask, startrec);
else
rc = llapi_changelog_start(&changelog_priv,
CHANGELOG_FLAG_BLOCK |
CHANGELOG_FLAG_JOBID |
CHANGELOG_FLAG_NID_BE |
CHANGELOG_FLAG_EXTRA_FLAGS |
(follow ? CHANGELOG_FLAG_FOLLOW : 0),
mdd, startrec);
if (rc < 0) {
fprintf(stderr, "%s changelog: cannot start changelog: %s\n",
progname, strerror(errno = -rc));
return rc;
}
rc = llapi_changelog_set_xflags(changelog_priv,
CHANGELOG_EXTRA_FLAG_UIDGID |
CHANGELOG_EXTRA_FLAG_NID |
CHANGELOG_EXTRA_FLAG_OMODE |
CHANGELOG_EXTRA_FLAG_XATTR);
if (rc < 0) {
fprintf(stderr,
"%s changelog: cannot set xflags for changelog: %s\n",
progname, strerror(errno = -rc));
return rc;
}
while ((rc = llapi_changelog_recv(changelog_priv, &rec)) == 0) {
time_t secs;
struct tm ts;
if (endrec && rec->cr_index > endrec) {
llapi_changelog_free(&rec);
break;
}
if (rec->cr_index < startrec) {
llapi_changelog_free(&rec);
continue;
}
secs = rec->cr_time >> 30;
gmtime_r(&secs, &ts);
printf("%ju %02d%-5s %02d:%02d:%02d.%09d %04d.%02d.%02d "
"0x%x t="DFID, (uintmax_t)rec->cr_index, rec->cr_type,
changelog_type2str(rec->cr_type),
ts.tm_hour, ts.tm_min, ts.tm_sec,
(int)(rec->cr_time & ((1 << 30) - 1)),
ts.tm_year + 1900, ts.tm_mon + 1, ts.tm_mday,
rec->cr_flags & CLF_FLAGMASK, PFID(&rec->cr_tfid));
if (rec->cr_flags & CLF_JOBID) {
struct changelog_ext_jobid *jid =
changelog_rec_jobid(rec);
if (jid->cr_jobid[0] != '\0')
printf(" j=%s", jid->cr_jobid);
}
if (rec->cr_flags & CLF_EXTRA_FLAGS) {
struct changelog_ext_extra_flags *ef =
changelog_rec_extra_flags(rec);
printf(" ef=0x%llx",
(unsigned long long)ef->cr_extra_flags);
if (ef->cr_extra_flags & CLFE_UIDGID) {
struct changelog_ext_uidgid *uidgid =
changelog_rec_uidgid(rec);
printf(" u=%llu:%llu",
(unsigned long long)uidgid->cr_uid,
(unsigned long long)uidgid->cr_gid);
}
if (ef->cr_extra_flags & CLFE_NID) {
if (ef->cr_extra_flags & CLFE_NID_BE) {
struct lnet_nid *nid =
(void *)changelog_rec_nid(rec);
printf(" nid=%s", libcfs_nidstr(nid));
} else {
struct changelog_ext_nid *nid =
changelog_rec_nid(rec);
printf(" nid=%s",
libcfs_nid2str(nid->cr_nid));
}
}
if (ef->cr_extra_flags & CLFE_OPEN) {
struct changelog_ext_openmode *omd =
changelog_rec_openmode(rec);
char mode[] = "---";
/* exec mode must be exclusive */
if (omd->cr_openflags & MDS_FMODE_EXEC) {
mode[2] = 'x';
} else {
if (omd->cr_openflags & MDS_FMODE_READ)
mode[0] = 'r';
if (omd->cr_openflags &
(MDS_FMODE_WRITE |
MDS_OPEN_TRUNC |
MDS_OPEN_APPEND))
mode[1] = 'w';
}
if (strcmp(mode, "---") != 0)
printf(" m=%s", mode);
}
if (ef->cr_extra_flags & CLFE_XATTR) {
struct changelog_ext_xattr *xattr =
changelog_rec_xattr(rec);
if (xattr->cr_xattr[0] != '\0')
printf(" x=%s", xattr->cr_xattr);
}
}
if (!fid_is_zero(&rec->cr_pfid))
printf(" p="DFID, PFID(&rec->cr_pfid));
if (rec->cr_namelen)
printf(" %.*s", rec->cr_namelen,
changelog_rec_name(rec));
if (rec->cr_flags & CLF_RENAME) {
struct changelog_ext_rename *rnm =
changelog_rec_rename(rec);
if (!fid_is_zero(&rnm->cr_sfid))
printf(" s="DFID" sp="DFID" %.*s",
PFID(&rnm->cr_sfid),
PFID(&rnm->cr_spfid),
(int)changelog_rec_snamelen(rec),
changelog_rec_sname(rec));
}
printf("\n");
llapi_changelog_free(&rec);
}
llapi_changelog_fini(&changelog_priv);
if (rc < 0)
fprintf(stderr, "%s changelog: cannot access changelog: %s\n",
progname, strerror(errno = -rc));
return (rc == 1 ? 0 : rc);
}
static int lfs_changelog_clear(int argc, char **argv)
{
long long endrec;
int rc;
if (argc != 4)
return CMD_HELP;
errno = 0;
endrec = strtoll(argv[3], NULL, 10);
if (errno != 0 || endrec < 0) {
fprintf(stderr,
"%s: bad endrec '%s'\n",
argv[0], argv[3]);
return CMD_HELP;
}
rc = llapi_changelog_clear(argv[1], argv[2], endrec);
if (rc == -EINVAL)
fprintf(stderr, "%s: record out of range: %llu\n",
argv[0], endrec);
else if (rc == -ENOENT)
fprintf(stderr, "%s: no changelog user: %s\n",
argv[0], argv[2]);
else if (rc)
fprintf(stderr, "%s error: %s\n", argv[0],
strerror(-rc));
if (rc)
errno = -rc;
return rc;
}
static void rstripc(char *str, int c)
{
char *end = str + strlen(str);
for (; str < end && end[-1] == c; --end)
end[-1] = '\0';
}
/* Helper function to lfs_fid2path. To print out only the file names and
* not the full path. Do not call OBD_IOC_FID2PATH for every file. Instead
* read the trusted.link xattr and loop over all the records to get all the
* file names.
*/
static int lfs_fid2path_prn_name(char *mnt_dir, char *path_buf,
bool print_linkno, bool print_fid, char *ptr,
const char *fid_str, int linktmp)
{
char buf[65536]; /* BUFFER_SIZE 65536 */
char full_path[PATH_MAX * 2 + 2];
struct link_ea_header *leh;
struct link_ea_entry *lee;
ssize_t size;
int reclen, i, rc = 0;
/* Generate full_path */
snprintf(full_path, sizeof(full_path) - 1, "%s/%s", mnt_dir, path_buf);
size = getxattr(full_path, "trusted.link", buf, sizeof(buf));
if (size < 0) {
fprintf(stderr, "%s: failed to read %s xattr: %s\n", path_buf,
"trusted.link", strerror(errno));
rc = -errno;
goto fail;
}
leh = (struct link_ea_header *)buf;
if (leh->leh_magic == __swab32(LINK_EA_MAGIC))
leh->leh_reccount = __swab32(leh->leh_reccount);
lee = (struct link_ea_entry *)(leh + 1);
for (i = 0; i < leh->leh_reccount; i++) {
reclen = (lee->lee_reclen[0] << 8) | lee->lee_reclen[1];
/* handle -n -l case */
if (print_linkno) {
ptr = strrchr(path_buf, '/');
if (!ptr)
ptr = path_buf;
else
ptr = ptr + 1;
if (strcmp(ptr, lee->lee_name) == 0) {
if (print_fid)
printf("%s ", fid_str);
printf("%d ", linktmp);
printf("%s\n", lee->lee_name);
break;
}
} else {
if (print_fid)
printf("%s ", fid_str);
printf("%s\n", lee->lee_name);
}
/* Get next record */
lee = (struct link_ea_entry *)((char *)lee + reclen);
}
fail:
return rc;
}
static int lfs_fid2path(int argc, char **argv)
{
struct option long_opts[] = {
{ .val = '0', .name = "print0", .has_arg = no_argument },
{ .val = 'c', .name = "cur", .has_arg = no_argument },
{ .val = 'c', .name = "current", .has_arg = no_argument },
{ .val = 'c', .name = "print-link", .has_arg = no_argument },
{ .val = 'f', .name = "print-fid", .has_arg = no_argument },
{ .val = 'l', .name = "link", .has_arg = required_argument },
{ .val = 'n', .name = "name", .has_arg = no_argument },
{ .name = NULL } };
char short_opts[] = "0cfl:pr:n";
bool print_only_fname = false;
bool print_linkno = false;
bool print_link = false;
bool print_fid = false;
bool print_mnt_dir;
char *mnt_dir = NULL;
int mnt_dir_len = PATH_MAX + 1;
int mnt_fd = -1;
char *path_or_fsname;
long long recno = -1;
int linkno = -1;
char *endptr = NULL;
char link_separator = '\n';
int rc = 0;
int c;
int i;
while ((c = getopt_long(argc, argv, short_opts,long_opts, NULL)) !=
-1) {
switch (c) {
case '0':
link_separator = '\0';
break;
case 'c':
print_link = true;
break;
case 'f':
print_fid = true;
break;
case 'l':
errno = 0;
linkno = strtol(optarg, &endptr, 10);
if (errno != 0 || *endptr != '\0' || linkno < 0) {
fprintf(stderr,
"%s fid2path: invalid linkno '%s'\n",
progname, optarg);
return CMD_HELP;
}
print_linkno = true;
break;
case 'n':
/* Bypass the full parent path if true
* only print the final filename */
print_only_fname = true;
break;
case 'r':
/* recno is something to do with changelogs
* that was never implemented. We just pass it
* through for the MDT to ignore.
*/
errno = 0;
recno = strtoll(optarg, &endptr, 10);
if (errno != 0 || *endptr != '\0' || recno < 0) {
fprintf(stderr,
"%s fid2path: invalid recno '%s'\n",
progname, optarg);
return CMD_HELP;
}
break;
default:
fprintf(stderr,
"%s fid2path: unrecognized option '%s'\n",
progname, argv[optind - 1]);
return CMD_HELP;
}
}
if (argc - optind < 2) {
fprintf(stderr,
"Usage: %s fid2path FSNAME|ROOT FID...\n",
progname);
return CMD_HELP;
}
path_or_fsname = argv[optind];
if (path_or_fsname && strlen(path_or_fsname))
mnt_dir_len = strlen(path_or_fsname) + 1;
mnt_dir = malloc(mnt_dir_len + 1);
if (!mnt_dir) {
rc = -ENOMEM;
goto out;
}
if (path_or_fsname && *path_or_fsname == '/') {
print_mnt_dir = true;
rc = llapi_search_mounts(path_or_fsname, 0, mnt_dir, NULL);
} else {
print_mnt_dir = false;
rc = llapi_search_rootpath(mnt_dir, path_or_fsname);
}
if (rc < 0) {
fprintf(stderr,
"%s fid2path: cannot resolve mount point for '%s': %s\n",
progname, path_or_fsname, strerror(-rc));
goto out;
}
mnt_fd = open(mnt_dir, O_RDONLY | O_DIRECTORY);
if (mnt_fd < 0) {
fprintf(stderr,
"%s fid2path: cannot open mount point for '%s': %s\n",
progname, path_or_fsname, strerror(-rc));
goto out;
}
/* Strip trailing slashes from mnt_dir. */
rstripc(mnt_dir + 1, '/');
for (i = optind + 1; i < argc; i++) {
const char *fid_str = argv[i];
int path_len = PATH_MAX;
char *path_buf;
struct lu_fid fid;
char *ptr = NULL;
int rc2;
rc2 = llapi_fid_parse(fid_str, &fid, NULL);
if (rc2 < 0) {
fprintf(stderr,
"%s fid2path: invalid FID '%s'\n",
progname, fid_str);
if (rc == 0)
rc = rc2;
continue;
}
int linktmp = (linkno >= 0) ? linkno : 0;
path_buf = malloc(path_len);
if (!path_buf) {
if (rc == 0)
rc = -ENOMEM;
continue;
}
while (1) {
int oldtmp = linktmp;
long long rectmp = recno;
fid2path:
rc2 = llapi_fid2path_at(mnt_fd, &fid, path_buf,
path_len, &rectmp,
&linktmp);
if (rc2 < 0) {
if (rc2 == -ERANGE) {
char *tmpbuf;
path_len += PATH_MAX;
tmpbuf = realloc(path_buf, path_len);
if (!tmpbuf) {
if (rc == 0)
rc = -ENOMEM;
break;
}
path_buf = tmpbuf;
goto fid2path;
}
fprintf(stderr,
"%s fid2path: cannot find %s %s: %s\n",
progname, path_or_fsname, fid_str,
strerror(-rc2));
if (rc == 0)
rc = rc2;
break;
}
if (print_only_fname && !print_link) {
/* '-n' is passed as option here.
* For all other cases of -c fall back
* to default(else) path as to get the link
* count associated with the file name call
* to OBD_IOC_FID2PATH is required
*/
rc = lfs_fid2path_prn_name(mnt_dir,
path_buf,
print_linkno,
print_fid, ptr,
fid_str, linktmp);
/* llapi_fid2path_at() is already called once
* in this case. No need to call it again.
* Break out as we have all the filenames.
*/
break;
}
if (print_fid)
printf("%s ", fid_str);
if (print_link)
printf("%d ", linktmp);
/* You may think this looks wrong or weird (and it is!)
* but we are actually trying to preserve the old quirky
* behaviors (enforced by our old quirky tests!) that
* make lfs so much fun to work on:
*
* lustre 0x200000007:0x1:0x0 => "/"
* /mnt/lustre 0x200000007:0x1:0x0 => "/mnt/lustre//"
*
* Note that llapi_fid2path() returns "" for the root
* FID. */
if (!print_only_fname) {
printf("%s%s%s%c",
print_mnt_dir ? mnt_dir : "",
(print_mnt_dir || *path_buf == '\0') ?
"/" : "", path_buf, link_separator);
} else {
ptr = strrchr(path_buf, '/');
if (!ptr)
printf("%s\n", path_buf);
else
printf("%s\n", ptr + 1);
}
if (linkno >= 0)
/* specified linkno */
break;
if (oldtmp == linktmp)
/* no more links */
break;
}
free(path_buf);
}
out:
if (!(mnt_fd < 0))
close(mnt_fd);
free(mnt_dir);
return rc;
}
static int lfs_path2fid(int argc, char **argv)
{
struct option long_opts[] = {
{ .val = 'p', .name = "parents", .has_arg = no_argument },
{ .name = NULL } };
char **path;
const char short_opts[] = "p";
const char *sep = "";
struct lu_fid fid;
int rc = 0;
bool show_parents = false;
while ((rc = getopt_long(argc, argv, short_opts,
long_opts, NULL)) != -1) {
switch (rc) {
case 'p':
show_parents = true;
break;
default:
fprintf(stderr,
"%s path2fid: unrecognized option '%s'\n",
progname, argv[optind - 1]);
return CMD_HELP;
}
}
if (optind > argc - 1) {
fprintf(stderr, "%s path2fid: FILE... must be specified\n",
progname);
return CMD_HELP;
} else if (optind < argc - 1) {
sep = ": ";
}
rc = 0;
for (path = argv + optind; optind < argc; path++, optind++) {
int err = 0;
if (!show_parents) {
err = llapi_path2fid(*path, &fid);
if (!err)
printf("%s%s"DFID"\n",
*sep != '\0' ? *path : "", sep,
PFID(&fid));
} else {
char name[NAME_MAX + 1];
unsigned int linkno = 0;
while ((err = llapi_path2parent(*path, linkno, &fid,
name, sizeof(name))) == 0) {
if (*sep != '\0' && linkno == 0)
printf("%s%s", *path, sep);
printf("%s"DFID"/%s", linkno != 0 ? "\t" : "",
PFID(&fid), name);
linkno++;
}
/* err == -ENODATA is end-of-loop */
if (linkno > 0 && err == -ENODATA) {
printf("\n");
err = 0;
}
}
if (err) {
fprintf(stderr,
"%s path2fid: cannot get %sfid for '%s': %s\n",
progname, show_parents ? "parent " : "", *path,
strerror(-err));
if (rc == 0) {
rc = err;
errno = -err;
}
}
}
return rc;
}
#define MAX_ERRNO 4095
#define IS_ERR_VALUE(x) ((unsigned long)(x) >= (unsigned long)-MAX_ERRNO)
static int lfs_rmfid_and_show_errors(int rootfd, struct fid_array *fa)
{
int rc, rc2, k;
rc = llapi_rmfid_at(rootfd, fa);
if (rc < 0) {
fprintf(stderr, "%s rmfid: cannot remove FIDs: %s\n",
progname, strerror(-rc));
return rc;
}
for (k = 0; k < fa->fa_nr; k++) {
rc2 = (__s32)fa->fa_fids[k].f_ver;
if (!IS_ERR_VALUE(rc2))
continue;
if (rc == 0)
rc = rc2;
fa->fa_fids[k].f_ver = 0;
fprintf(stderr, "%s rmfid: cannot remove "DFID": %s\n",
progname, PFID(&fa->fa_fids[k]), strerror(-rc2));
}
return rc;
}
static int lfs_rmfid(int argc, char **argv)
{
int rc = 0, rc2, rc3 = 0, nr;
struct fid_array *fa;
const char *device;
char *fidstr;
int rootfd;
/* Interactive mode: Adjust optind */
if (!optind)
optind++;
device = argv[optind++];
if (optind > argc - 1) {
fprintf(stderr, "%s rmfid: missing dirname\n", progname);
return CMD_HELP;
}
nr = argc - optind;
rc = llapi_root_path_open(device, &rootfd);
if (rc < 0) {
fprintf(stderr,
"%s rmfid: error opening device/fsname '%s': %s\n",
progname, device, strerror(-rc));
return -rc;
}
fa = malloc(offsetof(struct fid_array, fa_fids[nr + 1]));
if (!fa) {
rc = -errno ?: -ENOMEM;
fprintf(stderr, "%s rmfid: error allocating %zd bytes: %s\n",
progname, offsetof(struct fid_array, fa_fids[nr + 1]),
strerror(-rc));
goto out_close;
}
fa->fa_nr = 0;
rc = 0;
while (optind < argc) {
char *origfidstr;
int found;
origfidstr = fidstr = argv[optind++];
while (*fidstr == '[')
fidstr++;
found = sscanf(fidstr, SFID, RFID(&fa->fa_fids[fa->fa_nr]));
if (found != 3) {
fprintf(stderr, "lfs rmfid: '%s': Wrong FID format\n",
origfidstr);
if (!rc3)
rc3 = -EINVAL; /* Invalid argument */
continue;
}
fa->fa_nr++;
if (fa->fa_nr == OBD_MAX_FIDS_IN_ARRAY) {
/* start another batch */
rc2 = lfs_rmfid_and_show_errors(rootfd, fa);
if (rc2 && !rc)
rc = rc2;
if (rc3)
rc = rc3;
fa->fa_nr = 0;
}
}
if (fa->fa_nr) {
rc2 = lfs_rmfid_and_show_errors(rootfd, fa);
if (rc2 && !rc)
rc = rc2;
if (rc3)
rc = rc3;
}
if (fa) {
free(fa);
fa = NULL;
}
out_close:
close(rootfd);
return rc;
}
static int lfs_data_version(int argc, char **argv)
{
struct option long_opts[] = {
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'n', .name = "nosync", .has_arg = no_argument },
{ .val = 'r', .name = "read-lock", .has_arg = no_argument },
{ .val = 's', .name = "set-hsm", .has_arg = no_argument },
{ .val = 'w', .name = "write-lock", .has_arg = no_argument },
{ .name = NULL } };
int data_version_flags = LL_DV_RD_FLUSH; /* Read by default */
__u64 data_version;
char *path;
bool hsm_sync = false;
int fd;
int rc;
int c;
while ((c = getopt_long(argc, argv, "hnrsw", long_opts, NULL)) != -1) {
switch (c) {
case 'n':
data_version_flags = 0;
break;
case 'r':
data_version_flags |= LL_DV_RD_FLUSH;
break;
case 's':
hsm_sync = true;
break;
case 'w':
data_version_flags |= LL_DV_WR_FLUSH;
break;
default:
fprintf(stderr,
"%s data_version: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
return CMD_HELP;
}
}
if (optind == argc) {
fprintf(stderr, "%s data_version: FILENAME must be specified\n",
progname);
return CMD_HELP;
}
path = argv[optind];
fd = open(path, O_RDONLY);
if (fd < 0) {
rc = -errno;
fprintf(stderr, "%s data_version: cannot open file '%s': %s\n",
progname, path, strerror(-rc));
return rc;
}
rc = llapi_get_data_version(fd, &data_version, data_version_flags);
if (rc < 0) {
fprintf(stderr,
"%s data_version: cannot get version for '%s': %s\n",
progname, path, strerror(-rc));
} else {
printf("%ju" "\n", (uintmax_t)data_version);
if (hsm_sync) {
rc = llapi_hsm_data_version_set(fd, data_version);
if (rc < 0)
fprintf(stderr,
"%s data_version: cannot set version %llu for"
" '%s': %s\n", progname,
(unsigned long long)data_version, path,
strerror(-rc));
}
}
close(fd);
return rc;
}
static int lfs_hsm_state(int argc, char **argv)
{
int rc = 0;
int i = 1;
char *path;
struct hsm_user_state hus;
if (argc < 2)
return CMD_HELP;
do {
int rc2;
path = argv[i];
rc2 = llapi_hsm_state_get(path, &hus);
if (rc2) {
fprintf(stderr,
"%s %s: get HSM state for '%s' failed: %s\n",
progname, argv[0], path, strerror(-rc2));
if (!rc)
rc = rc2;
continue;
}
/* Display path name and status flags */
printf("%s: (0x%08x)", path, hus.hus_states);
if (hus.hus_states & HS_RELEASED)
printf(" released");
if (hus.hus_states & HS_EXISTS)
printf(" exists");
if (hus.hus_states & HS_DIRTY)
printf(" dirty");
if (hus.hus_states & HS_ARCHIVED)
printf(" archived");
/* Display user-settable flags */
if (hus.hus_states & HS_NORELEASE)
printf(" never_release");
if (hus.hus_states & HS_NOARCHIVE)
printf(" never_archive");
if (hus.hus_states & HS_LOST)
printf(" lost_from_hsm");
if (hus.hus_archive_id != 0)
printf(", archive_id:%d", hus.hus_archive_id);
printf("\n");
} while (++i < argc);
return rc;
}
#define LFS_HSM_SET 0
#define LFS_HSM_CLEAR 1
/*
* Generic function to set or clear HSM flags.
* Used by hsm_set and hsm_clear.
*
* @mode if LFS_HSM_SET, set the flags, if LFS_HSM_CLEAR, clear the flags.
*/
static int lfs_hsm_change_flags(int argc, char **argv, int mode)
{
struct option long_opts[] = {
{ .val = 'A', .name = "archived", .has_arg = no_argument },
{ .val = 'a', .name = "noarchive", .has_arg = no_argument },
{ .val = 'd', .name = "dirty", .has_arg = no_argument },
{ .val = 'e', .name = "exists", .has_arg = no_argument },
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'i', .name = "archive-id", .has_arg = required_argument },
{ .val = 'l', .name = "lost", .has_arg = no_argument },
{ .val = 'r', .name = "norelease", .has_arg = no_argument },
{ .name = NULL } };
__u64 mask = 0;
int c, rc = 0;
char *path;
__u32 archive_id = 0;
char *end = NULL;
if (argc < 3)
return CMD_HELP;
while ((c = getopt_long(argc, argv, "aAdehi:lr",
long_opts, NULL)) != -1) {
switch (c) {
case 'l':
mask |= HS_LOST;
break;
case 'a':
mask |= HS_NOARCHIVE;
break;
case 'A':
mask |= HS_ARCHIVED;
break;
case 'r':
mask |= HS_NORELEASE;
break;
case 'd':
mask |= HS_DIRTY;
break;
case 'e':
mask |= HS_EXISTS;
break;
case 'i':
errno = 0;
archive_id = strtol(optarg, &end, 10);
if (errno != 0 || *end != '\0' || archive_id < 0) {
fprintf(stderr,
"%s: invalid archive_id: '%s'\n",
progname, end);
return CMD_HELP;
}
break;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
return CMD_HELP;
}
}
/* User should have specified a flag */
if (mask == 0)
return CMD_HELP;
while (optind < argc) {
int rc2;
path = argv[optind];
/* If mode == 0, this means we apply the mask. */
if (mode == LFS_HSM_SET)
rc2 = llapi_hsm_state_set(path, mask, 0, archive_id);
else
rc2 = llapi_hsm_state_set(path, 0, mask, 0);
if (rc2) {
fprintf(stderr,
"%s %s: change hsm flags for '%s' failed: %s\n",
progname, argv[0], path, strerror(-rc2));
if (!rc)
rc = rc2;
}
optind++;
}
return rc;
}
static int lfs_hsm_action(int argc, char **argv)
{
struct hsm_current_action hca;
struct hsm_extent he;
enum hsm_user_action hua;
enum hsm_progress_states hps;
int rc = 0;
int i = 1;
char *path;
if (argc < 2)
return CMD_HELP;
do {
int rc2;
path = argv[i];
rc2 = llapi_hsm_current_action(path, &hca);
if (rc2) {
fprintf(stderr,
"%s %s: get hsm action for '%s' failed: %s\n",
progname, argv[0], path, strerror(-rc2));
if (!rc)
rc = rc2;
continue;
}
he = hca.hca_location;
hua = hca.hca_action;
hps = hca.hca_state;
printf("%s: %s", path, hsm_user_action2name(hua));
/* Skip file without action */
if (hca.hca_action == HUA_NONE) {
printf("\n");
continue;
}
printf(" %s ", hsm_progress_state2name(hps));
if ((hps == HPS_RUNNING) &&
(hua == HUA_ARCHIVE || hua == HUA_RESTORE))
printf("(%llu bytes moved)\n",
(unsigned long long)he.length);
else if ((he.offset + he.length) == LUSTRE_EOF)
printf("(from %llu to EOF)\n",
(unsigned long long)he.offset);
else
printf("(from %llu to %llu)\n",
(unsigned long long)he.offset,
(unsigned long long)(he.offset + he.length));
} while (++i < argc);
return rc;
}
static int lfs_hsm_set(int argc, char **argv)
{
return lfs_hsm_change_flags(argc, argv, LFS_HSM_SET);
}
static int lfs_hsm_clear(int argc, char **argv)
{
return lfs_hsm_change_flags(argc, argv, LFS_HSM_CLEAR);
}
/**
* lfs_hsm_prepare_file() - Check file state and return its fid,
* to be used by lfs_hsm_request().
* @file: Path to file to check
* @fid: Pointer to allocated lu_fid struct. [in, out]
* @last_dev: Pointer to last device id used. [in, out]
*
* Return:
* * %0 on success
* * %negative on failure
*/
static int lfs_hsm_prepare_file(const char *file, struct lu_fid *fid,
dev_t *last_dev)
{
struct stat st;
int rc;
rc = lstat(file, &st);
if (rc) {
fprintf(stderr, "Cannot stat %s: %s\n", file, strerror(errno));
return -errno;
}
/*
* Checking for regular file as archiving as posix copytool
* rejects archiving files other than regular files
*/
if (!S_ISREG(st.st_mode)) {
fprintf(stderr, "error: \"%s\" is not a regular file\n", file);
return CMD_HELP;
}
/* A request should be ... */
if (*last_dev != st.st_dev && *last_dev != 0) {
fprintf(stderr,
"All files should be on the same filesystem: %s\n",
file);
return -EINVAL;
}
*last_dev = st.st_dev;
rc = llapi_path2fid(file, fid);
if (rc) {
fprintf(stderr, "Cannot read FID of %s: %s\n",
file, strerror(-rc));
return rc;
}
return 0;
}
/* Fill an HSM HUR item with a given file name.
*
* If mntpath is set, then the filename is actually a FID, and no
* lookup on the filesystem will be performed.
*
* \param[in] hur the user request to fill
* \param[in] idx index of the item inside the HUR to fill
* \param[in] mntpath mountpoint of Lustre
* \param[in] fname filename (if mtnpath is NULL)
* or FID (if mntpath is set)
* \param[in] last_dev pointer to last device id used
*
* \retval 0 on success
* \retval CMD_HELP or a negative errno on error
*/
static int fill_hur_item(struct hsm_user_request *hur, unsigned int idx,
const char *mntpath, const char *fname,
dev_t *last_dev)
{
struct hsm_user_item *hui = &hur->hur_user_item[idx];
int rc;
hui->hui_extent.length = -1;
if (mntpath) {
rc = llapi_fid_parse(fname, &hui->hui_fid, NULL);
if (rc)
fprintf(stderr, "hsm: '%s' is not a valid FID\n",
fname);
} else {
rc = lfs_hsm_prepare_file(fname, &hui->hui_fid, last_dev);
}
if (rc == 0)
hur->hur_request.hr_itemcount++;
return rc;
}
static int lfs_hsm_request(int argc, char **argv, int action)
{
struct option long_opts[] = {
{ .val = 'a', .name = "archive", .has_arg = required_argument },
{ .val = 'D', .name = "data", .has_arg = required_argument },
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'l', .name = "filelist", .has_arg = required_argument },
{ .val = 'm', .name = "mntpath", .has_arg = required_argument },
{ .name = NULL } };
dev_t last_dev = 0;
struct hsm_user_request *hur, *oldhur;
int c, i;
size_t len;
int nbfile;
char *line = NULL;
char *filelist = NULL;
char fullpath[PATH_MAX];
char *opaque = NULL;
int opaque_len = 0;
int archive_id = 0;
FILE *fp;
int nbfile_alloc = 0;
char *some_file = NULL;
char *mntpath = NULL;
int rc;
if (argc < 2) {
rc = CMD_HELP;
goto out_cmd_help;
}
while ((c = getopt_long(argc, argv, "a:D:hl:m:",
long_opts, NULL)) != -1) {
switch (c) {
case 'l':
filelist = optarg;
break;
case 'D':
opaque = optarg;
break;
case 'a':
if (action != HUA_ARCHIVE &&
action != HUA_REMOVE) {
fprintf(stderr,
"error: -a is supported only when archiving or removing\n");
rc = CMD_HELP;
goto out_cmd_help;
}
archive_id = atoi(optarg);
break;
case 'm':
if (!some_file) {
mntpath = optarg;
some_file = strdup(optarg);
}
break;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
rc = CMD_HELP;
goto out_cmd_help;
}
}
/* All remaining args are files, so we have at least nbfile */
nbfile = argc - optind;
if ((nbfile == 0) && (!filelist)) {
rc = errno;
goto out_errno;
}
if (opaque)
opaque_len = strlen(opaque);
/*
* Alloc the request structure with enough place to store all files
* from command line.
*/
hur = llapi_hsm_user_request_alloc(nbfile, opaque_len);
if (!hur) {
fprintf(stderr, "Cannot create the request: %s\n",
strerror(errno));
rc = errno;
goto out_errno;
}
nbfile_alloc = nbfile;
hur->hur_request.hr_action = action;
hur->hur_request.hr_archive_id = archive_id;
hur->hur_request.hr_flags = 0;
/* All remaining args are files, add them */
if (nbfile != 0 && some_file == NULL)
some_file = strdup(argv[optind]);
for (i = 0; i < nbfile; i++) {
rc = fill_hur_item(hur, i, mntpath, argv[optind + i],
&last_dev);
if (rc)
goto out_hur;
}
/* from here stop using nb_file, use hur->hur_request.hr_itemcount */
/* If a filelist was specified, read the filelist from it. */
if (filelist) {
fp = fopen(filelist, "r");
if (!fp) {
fprintf(stderr, "Cannot read the file list %s: %s\n",
filelist, strerror(errno));
rc = -errno;
goto out_hur;
}
while ((rc = getline(&line, &len, fp)) != -1) {
/*
* If allocated buffer was too small, get something
* larger
*/
if (nbfile_alloc <= hur->hur_request.hr_itemcount) {
ssize_t size;
nbfile_alloc = nbfile_alloc * 2 + 1;
oldhur = hur;
hur = llapi_hsm_user_request_alloc(nbfile_alloc,
opaque_len);
if (!hur) {
fprintf(stderr,
"hsm: cannot allocate the request: %s\n",
strerror(errno));
hur = oldhur;
rc = -errno;
fclose(fp);
goto out_hur;
}
size = hur_len(oldhur);
if (size < 0) {
fprintf(stderr,
"hsm: cannot allocate %u files + %u bytes data\n",
oldhur->hur_request.hr_itemcount,
oldhur->hur_request.hr_data_len);
free(hur);
hur = oldhur;
rc = -E2BIG;
fclose(fp);
goto out_hur;
}
memcpy(hur, oldhur, size);
free(oldhur);
}
/* Chop CR */
if (line[strlen(line) - 1] == '\n')
line[strlen(line) - 1] = '\0';
rc = fill_hur_item(hur, hur->hur_request.hr_itemcount,
mntpath, line, &last_dev);
if (rc) {
fclose(fp);
goto out_hur;
}
if (!some_file) {
some_file = line;
line = NULL;
}
}
rc = fclose(fp);
free(line);
}
/* If a --data was used, add it to the request */
hur->hur_request.hr_data_len = opaque_len;
if (opaque)
memcpy(hur_data(hur), opaque, opaque_len);
/* Send the HSM request */
if (realpath(some_file, fullpath) == NULL) {
fprintf(stderr, "Could not find path '%s': %s\n",
some_file, strerror(errno));
}
rc = llapi_hsm_request(fullpath, hur);
if (rc)
fprintf(stderr, "Cannot send HSM request (use of %s): %s\n",
some_file, strerror(-rc));
out_hur:
free(hur);
out_errno:
free(some_file);
out_cmd_help:
return rc;
}
static int lfs_hsm_archive(int argc, char **argv)
{
return lfs_hsm_request(argc, argv, HUA_ARCHIVE);
}
static int lfs_hsm_restore(int argc, char **argv)
{
return lfs_hsm_request(argc, argv, HUA_RESTORE);
}
static int lfs_hsm_release(int argc, char **argv)
{
return lfs_hsm_request(argc, argv, HUA_RELEASE);
}
static int lfs_hsm_remove(int argc, char **argv)
{
return lfs_hsm_request(argc, argv, HUA_REMOVE);
}
static int lfs_hsm_cancel(int argc, char **argv)
{
return lfs_hsm_request(argc, argv, HUA_CANCEL);
}
static int lfs_swap_layouts(int argc, char **argv)
{
int noxtime = 0;
if (argc == 4 && !strcmp(argv[1], "-n"))
noxtime = 1;
else if (argc != 3)
return CMD_HELP;
return llapi_swap_layouts(argv[1+noxtime], argv[2+noxtime],
0, 0, noxtime ? 0 :
(SWAP_LAYOUTS_KEEP_MTIME |
SWAP_LAYOUTS_KEEP_ATIME));
}
static const char *const ladvise_names[] = LU_LADVISE_NAMES;
static const char *const lock_mode_names[] = LOCK_MODE_NAMES;
static int lfs_get_mode(const char *string)
{
enum lock_mode_user mode;
for (mode = 0; mode < ARRAY_SIZE(lock_mode_names); mode++) {
if (lock_mode_names[mode] == NULL)
continue;
if (strcasecmp(string, lock_mode_names[mode]) == 0)
return mode;
}
return -EINVAL;
}
static enum lu_ladvise_type lfs_get_ladvice(const char *string)
{
enum lu_ladvise_type advice;
for (advice = 0;
advice < ARRAY_SIZE(ladvise_names); advice++) {
if (ladvise_names[advice] == NULL)
continue;
if (strcmp(string, ladvise_names[advice]) == 0)
return advice;
}
return LU_LADVISE_INVALID;
}
static int lfs_ladvise(int argc, char **argv)
{
struct option long_opts[] = {
{ .val = 'a', .name = "advice", .has_arg = required_argument },
{ .val = 'b', .name = "background", .has_arg = no_argument },
{ .val = 'e', .name = "end", .has_arg = required_argument },
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'l', .name = "length", .has_arg = required_argument },
{ .val = 'm', .name = "mode", .has_arg = required_argument },
{ .val = 's', .name = "start", .has_arg = required_argument },
{ .val = 'u', .name = "unset", .has_arg = no_argument },
{ .name = NULL } };
struct llapi_lu_ladvise advice;
enum lu_ladvise_type advice_type = LU_LADVISE_INVALID;
unsigned long long start = 0;
unsigned long long end = LUSTRE_EOF;
unsigned long long length = 0;
unsigned long long size_units;
unsigned long long flags = 0;
int c, fd, rc = 0;
const char *path;
int mode = 0;
optind = 0;
while ((c = getopt_long(argc, argv, "a:be:hl:m:s:u",
long_opts, NULL)) != -1) {
switch (c) {
case 'a':
advice_type = lfs_get_ladvice(optarg);
if (advice_type == LU_LADVISE_INVALID) {
fprintf(stderr,
"%s: invalid advice type '%s'\n",
progname, optarg);
fprintf(stderr, "Valid types:");
for (advice_type = 0;
advice_type < ARRAY_SIZE(ladvise_names);
advice_type++) {
if (ladvise_names[advice_type] == NULL)
continue;
fprintf(stderr, " %s",
ladvise_names[advice_type]);
}
fprintf(stderr, "\n");
return CMD_HELP;
}
break;
case 'b':
flags |= LF_ASYNC;
break;
case 'u':
flags |= LF_UNSET;
break;
case 'e':
size_units = 1;
rc = llapi_parse_size(optarg, &end,
&size_units, 0);
if (rc) {
fprintf(stderr, "%s: bad end offset '%s'\n",
argv[0], optarg);
return CMD_HELP;
}
break;
case 's':
size_units = 1;
rc = llapi_parse_size(optarg, &start,
&size_units, 0);
if (rc) {
fprintf(stderr,
"%s: bad start offset '%s'\n",
argv[0], optarg);
return CMD_HELP;
}
break;
case 'l':
size_units = 1;
rc = llapi_parse_size(optarg, &length,
&size_units, 0);
if (rc) {
fprintf(stderr, "%s: bad length '%s'\n",
argv[0], optarg);
return CMD_HELP;
}
break;
case 'm':
mode = lfs_get_mode(optarg);
if (mode < 0) {
fprintf(stderr,
"%s: bad mode '%s', valid modes are READ or WRITE\n",
argv[0], optarg);
return CMD_HELP;
}
break;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
return CMD_HELP;
}
}
if (advice_type == LU_LADVISE_INVALID) {
fprintf(stderr, "%s: please give an advice type\n", argv[0]);
fprintf(stderr, "Valid types:");
for (advice_type = 0; advice_type < ARRAY_SIZE(ladvise_names);
advice_type++) {
if (ladvise_names[advice_type] == NULL)
continue;
fprintf(stderr, " %s", ladvise_names[advice_type]);
}
fprintf(stderr, "\n");
return CMD_HELP;
}
if (advice_type == LU_LADVISE_LOCKNOEXPAND) {
fprintf(stderr,
"%s: Lock no expand advice is a per file descriptor advice, so when called from lfs, it does nothing.\n",
argv[0]);
return CMD_HELP;
}
if (argc <= optind) {
fprintf(stderr, "%s: please give one or more file names\n",
argv[0]);
return CMD_HELP;
}
if (end != LUSTRE_EOF && length != 0 && end != start + length) {
fprintf(stderr, "%s: conflicting arguments of -l and -e\n",
argv[0]);
return CMD_HELP;
}
if (end == LUSTRE_EOF && length != 0)
end = start + length;
if (end <= start) {
fprintf(stderr, "%s: range [%llu, %llu] is invalid\n",
argv[0], start, end);
return CMD_HELP;
}
if (advice_type != LU_LADVISE_LOCKAHEAD && mode != 0) {
fprintf(stderr, "%s: mode is only valid with lockahead\n",
argv[0]);
return CMD_HELP;
}
if (advice_type == LU_LADVISE_LOCKAHEAD && mode == 0) {
fprintf(stderr, "%s: mode is required with lockahead\n",
argv[0]);
return CMD_HELP;
}
while (optind < argc) {
int rc2;
path = argv[optind++];
fd = open(path, O_RDONLY);
if (fd < 0) {
rc2 = -errno;
fprintf(stderr, "%s: cannot open file '%s': %s\n",
argv[0], path, strerror(-rc2));
if (!rc)
rc = rc2;
continue;
}
advice.lla_start = start;
advice.lla_end = end;
advice.lla_advice = advice_type;
advice.lla_value1 = 0;
advice.lla_value2 = 0;
advice.lla_value3 = 0;
advice.lla_value4 = 0;
if (advice_type == LU_LADVISE_LOCKAHEAD) {
advice.lla_lockahead_mode = mode;
advice.lla_peradvice_flags = flags;
}
rc2 = llapi_ladvise(fd, flags, 1, &advice);
close(fd);
if (rc2 < 0) {
fprintf(stderr,
"%s: cannot give advice '%s' to file '%s': %s\n",
argv[0], ladvise_names[advice_type],
path, strerror(errno));
if (!rc)
rc = rc2;
continue;
}
}
return rc;
}
static const char *const heat_names[] = LU_HEAT_NAMES;
static int lfs_heat_get(int argc, char **argv)
{
struct lu_heat *heat;
int rc = 0, rc2;
char *path;
int fd;
int i;
if (argc <= 1)
return CMD_HELP;
heat = calloc(sizeof(*heat) + sizeof(__u64) * OBD_HEAT_COUNT, 1);
if (!heat) {
fprintf(stderr, "%s: memory allocation failed\n", argv[0]);
return -ENOMEM;
}
optind = 1;
while (optind < argc) {
path = argv[optind++];
fd = open(path, O_RDONLY);
if (fd < 0) {
fprintf(stderr, "%s: cannot open file '%s': %s\n",
argv[0], path, strerror(errno));
rc2 = -errno;
goto next;
}
heat->lh_count = OBD_HEAT_COUNT;
rc2 = llapi_heat_get(fd, heat);
close(fd);
if (rc2 < 0) {
fprintf(stderr,
"%s: cannot get heat of file '%s': %s\n",
argv[0], path, strerror(errno));
goto next;
}
printf("flags: %x\n", heat->lh_flags);
for (i = 0; i < heat->lh_count; i++)
printf("%s: %llu\n", heat_names[i],
(unsigned long long)heat->lh_heat[i]);
next:
if (rc == 0 && rc2 < 0)
rc = rc2;
}
free(heat);
return rc;
}
static int lfs_heat_set(int argc, char **argv)
{
struct option long_opts[] = {
{ .val = 'c', .name = "clear", .has_arg = no_argument },
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'o', .name = "off", .has_arg = no_argument },
{ .val = 'O', .name = "on", .has_arg = no_argument },
{ .name = NULL } };
enum lu_heat_flag flags = 0;
int rc = 0, rc2;
char *path;
int fd;
int c;
if (argc <= 1)
return CMD_HELP;
optind = 0;
while ((c = getopt_long(argc, argv, "choO", long_opts, NULL)) != -1) {
switch (c) {
case 'c':
flags |= LU_HEAT_FLAG_CLEAR;
break;
case 'o':
flags |= LU_HEAT_FLAG_CLEAR;
flags |= LU_HEAT_FLAG_OFF;
break;
case 'O':
flags &= ~LU_HEAT_FLAG_OFF;
break;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
return CMD_HELP;
}
}
if (argc <= optind) {
fprintf(stderr, "%s: please give one or more file names\n",
argv[0]);
return CMD_HELP;
}
while (optind < argc) {
path = argv[optind++];
fd = open(path, O_RDONLY);
if (fd < 0) {
fprintf(stderr, "%s: cannot open file '%s': %s\n",
argv[0], path, strerror(errno));
rc2 = -errno;
goto next;
}
rc2 = llapi_heat_set(fd, flags);
close(fd);
if (rc2 < 0) {
fprintf(stderr,
"%s: cannot setflags heat of file '%s': %s\n",
argv[0], path, strerror(errno));
goto next;
}
next:
if (rc == 0 && rc2 < 0)
rc = rc2;
}
return rc;
}
/**
* parse_mirror_ids() - Parse mirror ids
* @ids: parsed mirror ids [out]
* @size: maximum size of ids(mirror)
* @arg: input string (example "1,2-4,7")
*
* The input string contains a comma delimited list of component ids and
* ranges, for example "1,2-4,7".
*
* * Return:
* * %>=0 number of mirror ids parsed into @ids
* * %-EINVAL malformed @arg
*/
static int parse_mirror_ids(__u16 *ids, int size, char *arg)
{
bool end_of_loop = false;
char *ptr = NULL;
int nr = 0;
int rc;
if (!arg)
return -EINVAL;
while (!end_of_loop) {
int start_index;
int end_index;
int i;
char *endptr = NULL;
rc = -EINVAL;
ptr = strchrnul(arg, ',');
end_of_loop = *ptr == '\0';
*ptr = '\0';
start_index = strtol(arg, &endptr, 0);
if (endptr == arg) /* no data at all */
break;
if (*endptr != '-' && *endptr != '\0') /* has invalid data */
break;
if (start_index < 0)
break;
end_index = start_index;
if (*endptr == '-') {
end_index = strtol(endptr + 1, &endptr, 0);
if (*endptr != '\0')
break;
if (end_index < start_index)
break;
}
for (i = start_index; i <= end_index && size > 0; i++) {
int j;
/* remove duplicate */
for (j = 0; j < nr; j++) {
if (ids[j] == i)
break;
}
if (j == nr) { /* no duplicate */
ids[nr++] = i;
--size;
}
}
if (size == 0 && i < end_index)
break;
*ptr = ',';
arg = ++ptr;
rc = 0;
}
if (!end_of_loop && ptr)
*ptr = ',';
return rc < 0 ? rc : nr;
}
/**
* struct verify_mirror_id - Mirror id to be verified.
* @mirror_id: A specified mirror id.
* @is_valid_id: @mirror_id is valid or not in the mirrored file.
*/
struct verify_mirror_id {
__u16 mirror_id;
bool is_valid_id;
};
/**
* compare_mirror_ids() - Compare mirror ids.
* @layout: Mirror component list.
* @cbdata: Callback data in verify_mirror_id structure.
*
* This is a callback function called by llapi_layout_comp_iterate()
* to compare the specified mirror id with the one in the current
* component of @layout. If they are the same, then the specified
* mirror id is valid.
*
* Return: a negative error code on failure or
* LLAPI_LAYOUT_ITER_CONT: Proceed iteration
* LLAPI_LAYOUT_ITER_STOP: Stop iteration
*/
static inline
int compare_mirror_ids(struct llapi_layout *layout, void *cbdata)
{
struct verify_mirror_id *mirror_id_cbdata =
(struct verify_mirror_id *)cbdata;
uint32_t mirror_id;
int rc = 0;
rc = llapi_layout_mirror_id_get(layout, &mirror_id);
if (rc < 0) {
rc = -errno;
fprintf(stderr,
"%s: llapi_layout_mirror_id_get failed: %s.\n",
progname, strerror(errno));
return rc;
}
if (mirror_id_cbdata->mirror_id == mirror_id) {
mirror_id_cbdata->is_valid_id = true;
return LLAPI_LAYOUT_ITER_STOP;
}
return LLAPI_LAYOUT_ITER_CONT;
}
/**
* verify_mirror_ids() - Verify specified mirror ids.
* @fname: Mirrored file name.
* @mirror_ids: Specified mirror ids to be verified.
* @ids_nr: Number of specified mirror ids.
*
* This function verifies that specified @mirror_ids are valid
* in the mirrored file @fname.
*
* Return: 0 on success or a negative error code on failure.
*/
static inline
int verify_mirror_ids(const char *fname, __u16 *mirror_ids, int ids_nr)
{
struct llapi_layout *layout = NULL;
struct verify_mirror_id mirror_id_cbdata = { 0 };
struct stat stbuf;
uint32_t flr_state;
int i;
int fd;
int rc = 0;
int rc2 = 0;
if (ids_nr <= 0)
return -EINVAL;
if (stat(fname, &stbuf) < 0) {
fprintf(stderr, "%s: cannot stat file '%s': %s.\n",
progname, fname, strerror(errno));
rc = -errno;
goto error;
}
if (!S_ISREG(stbuf.st_mode)) {
fprintf(stderr, "%s: '%s' is not a regular file.\n",
progname, fname);
rc = -EINVAL;
goto error;
}
fd = open(fname, O_DIRECT | O_RDONLY);
if (fd < 0) {
fprintf(stderr, "%s: cannot open '%s': %s.\n",
progname, fname, strerror(errno));
rc = -errno;
goto error;
}
rc = llapi_lease_acquire(fd, LL_LEASE_RDLCK);
if (rc < 0) {
fprintf(stderr, "%s: '%s' llapi_lease_acquire failed: %s.\n",
progname, fname, strerror(errno));
goto close_fd;
}
layout = llapi_layout_get_by_fd(fd, 0);
if (!layout) {
fprintf(stderr, "%s: '%s' llapi_layout_get_by_fd failed: %s.\n",
progname, fname, strerror(errno));
rc = -errno;
llapi_lease_release(fd);
goto close_fd;
}
rc = llapi_layout_flags_get(layout, &flr_state);
if (rc < 0) {
fprintf(stderr, "%s: '%s' llapi_layout_flags_get failed: %s.\n",
progname, fname, strerror(errno));
rc = -errno;
goto free_layout;
}
flr_state &= LCM_FL_FLR_MASK;
switch (flr_state) {
case LCM_FL_NONE:
rc = -EINVAL;
fprintf(stderr, "%s: '%s' file state error: %s.\n",
progname, fname, llapi_layout_flags_string(flr_state));
goto free_layout;
default:
break;
}
rc2 = 0;
for (i = 0; i < ids_nr; i++) {
mirror_id_cbdata.mirror_id = mirror_ids[i];
mirror_id_cbdata.is_valid_id = false;
rc = llapi_layout_comp_iterate(layout, compare_mirror_ids,
&mirror_id_cbdata);
if (rc < 0) {
rc = -errno;
fprintf(stderr,
"%s: '%s' failed to verify mirror id: %u.\n",
progname, fname, mirror_ids[i]);
goto free_layout;
}
if (!mirror_id_cbdata.is_valid_id) {
rc2 = -EINVAL;
fprintf(stderr,
"%s: '%s' invalid specified mirror id: %u.\n",
progname, fname, mirror_ids[i]);
}
}
rc = rc2;
free_layout:
llapi_layout_free(layout);
llapi_lease_release(fd);
close_fd:
close(fd);
error:
return rc;
}
/*
* Returns number of components to resync, or %negative error code
*/
static int lfs_mirror_force_resync(const char *fname,
struct llapi_layout *layout,
struct llapi_resync_comp *comp_array,
__u16 *mirror_ids, int ids_nr)
{
int idx = 0;
uint16_t source_mirror = 0;
int rc;
/* First pass: find a source mirror (not in target list) */
rc = llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_FIRST);
if (rc < 0) {
fprintf(stderr, "%s: failed to position to first component: %s\n",
progname, strerror(-rc));
return rc;
}
while (rc == 0) {
uint32_t mirror_id, comp_flags;
int i, found = 0;
rc = llapi_layout_mirror_id_get(layout, &mirror_id);
if (rc < 0)
break;
/* Check if this mirror is in our target list */
for (i = 0; i < ids_nr; i++) {
if (mirror_ids[i] == mirror_id) {
found = 1;
break;
}
}
/* If not in target list, use as source */
if (!found) {
rc = llapi_layout_comp_flags_get(layout, &comp_flags);
if (rc == 0 && !(comp_flags & LCME_FL_STALE)) {
source_mirror = mirror_id;
break;
}
}
rc = llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_NEXT);
if (rc < 0) {
rc = 0;
break;
}
}
if (source_mirror == 0) {
/* No non-target source found, use target mirror as source
* and mark other mirrors as stale
*/
source_mirror = mirror_ids[0];
fprintf(stdout,
"%s: using target mirror %u as source (no other non-stale mirrors found).\n",
progname, source_mirror);
}
fprintf(stdout,
"%s: using mirror %u as source for force resync.\n",
progname, source_mirror);
/* Second pass: find components to mark as stale */
rc = llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_FIRST);
if (rc < 0) {
fprintf(stderr, "%s: failed to position to first component: %s\n",
progname, strerror(-rc));
return rc;
}
while (idx < 1024) { /* comp_array size limit */
uint32_t mirror_id, comp_id;
uint64_t start, end;
int i, is_target = 0;
rc = llapi_layout_mirror_id_get(layout, &mirror_id);
if (rc < 0)
break;
/* Check if this mirror is in our target list */
for (i = 0; i < ids_nr; i++) {
if (mirror_ids[i] == mirror_id) {
is_target = 1;
break;
}
}
bool collect_this = false;
if (source_mirror == mirror_ids[0]) {
/* Target is source, collect non-target mirrors */
collect_this = !is_target;
} else {
/* Non-target is source, collect target mirrors */
collect_this = is_target;
}
if (collect_this) {
rc = llapi_layout_comp_id_get(layout, &comp_id);
if (rc < 0)
break;
rc = llapi_layout_comp_extent_get(layout, &start, &end);
if (rc < 0)
break;
comp_array[idx].lrc_mirror_id = mirror_id;
comp_array[idx].lrc_id = comp_id;
comp_array[idx].lrc_start = start;
comp_array[idx].lrc_end = end;
idx++;
}
rc = llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_NEXT);
if (rc != 0) {
rc = 0;
break;
}
}
if (idx <= 0) {
fprintf(stderr,
"%s: no components found for specified mirrors.\n",
progname);
return -EINVAL;
}
/* Mark the collected components as stale */
int stale_count = 0;
for (int i = 0; i < idx; i++) {
rc = lfs_component_set((char *)fname, comp_array[i].lrc_id,
NULL, LCME_FL_STALE, 0);
if (rc < 0) {
fprintf(stderr,
"%s: failed to set stale flag on component %u: %s\n",
progname, comp_array[i].lrc_id, strerror(-rc));
/* Continue with other components */
} else {
stale_count++;
}
}
fprintf(stdout,
"%s: marked %d components as stale for force resync.\n",
progname, stale_count);
fprintf(stdout,
"%s: proceeding with resync of %d components.\n",
progname, idx);
return idx; /* Return number of components to resync */
}
int lfs_mirror_resync_file(const char *fname, struct ll_ioc_lease *ioc,
__u16 *mirror_ids, int ids_nr,
long stats_interval_sec, long bandwidth_bytes_sec,
bool force_resync)
{
struct llapi_resync_comp stale_comp_array[1024] = { { 0 } };
struct llapi_layout *layout;
struct stat stbuf;
int stale_data_comp_count = 0;
int stale_ec_comp_count = 0;
uint32_t flr_state;
uint64_t start;
uint64_t end;
int idx;
int fd;
int rc;
int rc2;
if (stat(fname, &stbuf) < 0) {
fprintf(stderr, "%s: cannot stat file '%s': %s.\n", progname,
fname, strerror(errno));
rc = -errno;
goto error;
}
if (!S_ISREG(stbuf.st_mode)) {
fprintf(stderr, "%s: '%s' is not a regular file.\n", progname,
fname);
rc = -EINVAL;
goto error;
}
/* Allow mirror resync even without the key on encrypted files */
fd = open(fname, O_DIRECT | O_RDWR | O_CIPHERTEXT);
if (fd < 0) {
fprintf(stderr, "%s: cannot open '%s': %s.\n", progname, fname,
strerror(errno));
rc = -errno;
goto error;
}
layout = llapi_layout_get_by_fd(fd, 0);
if (!layout) {
fprintf(stderr, "%s: '%s' llapi_layout_get_by_fd failed: %s.\n",
progname, fname, strerror(errno));
rc = -errno;
goto close_fd;
}
rc = llapi_layout_flags_get(layout, &flr_state);
if (rc) {
fprintf(stderr, "%s: '%s' llapi_layout_flags_get failed: %s.\n",
progname, fname, strerror(errno));
rc = -errno;
goto free_layout;
}
flr_state &= LCM_FL_FLR_MASK;
if (flr_state == LCM_FL_NONE) {
rc = -EINVAL;
fprintf(stderr, "%s: '%s' is not a FLR file.\n", progname,
fname);
goto free_layout;
}
/* get stale component info for regular mirrors */
stale_data_comp_count = llapi_mirror_find_stale(
layout, stale_comp_array, ARRAY_SIZE(stale_comp_array),
mirror_ids, ids_nr);
if (stale_data_comp_count < 0) {
rc = stale_data_comp_count;
goto free_layout;
}
/* get stale component info for ec mirrors */
stale_ec_comp_count = llapi_ec_find_stale(
layout, &stale_comp_array[stale_data_comp_count],
ARRAY_SIZE(stale_comp_array) - stale_data_comp_count,
mirror_ids, ids_nr);
if (stale_ec_comp_count < 0) {
rc = stale_ec_comp_count;
goto free_layout;
}
if (stale_data_comp_count + stale_ec_comp_count == 0) {
if (force_resync && ids_nr > 0) {
stale_data_comp_count = lfs_mirror_force_resync(
fname, layout, stale_comp_array, mirror_ids,
ids_nr);
if (stale_data_comp_count < 0) {
rc = stale_data_comp_count;
goto free_layout;
}
} else if (force_resync) {
/* --force-resync requires --only option */
rc = -EINVAL;
fprintf(stderr,
"%s: --force-resync requires --only option to specify target mirrors.\n",
progname);
goto free_layout;
} else if (ids_nr > 0) {
/* --only specified but no stale components and
* no --force-resync
* Check if mirrors are consistent - if not, error
*/
rc = lfs_mirror_verify_file(fname, NULL, 0, 0, 0);
if (rc < 0) {
/* mirrors inconsistent but no stale flags */
fprintf(stderr,
"%s: mirrors are inconsistent but no stale components marked in '%s'. Use 'lfs mirror resync --only %d --force-resync' to force resync.\n",
progname, fname, mirror_ids[0]);
goto free_layout;
}
/* Mirrors are consistent - succeed as no-op */
rc = 0;
goto free_layout;
} else {
/* Normal resync with no stale components and no --only
* Check if mirrors are consistent - if not, error
*/
rc = lfs_mirror_verify_file(fname, NULL, 0, 0, 0);
if (rc < 0) {
/* file has inconsistency but no stale flags */
fprintf(stderr,
"%s: mirrors are inconsistent but no stale components marked in '%s'. Use 'lfs mirror resync --only N --force-resync' to force resync specific mirrors.\n",
progname, fname);
goto free_layout;
}
/* Mirrors are consistent - succeed as no-op */
rc = 0;
goto free_layout;
}
}
ioc->lil_mode = LL_LEASE_WRLCK;
ioc->lil_flags = LL_LEASE_RESYNC;
rc = llapi_lease_set(fd, ioc);
if (rc < 0) {
if (rc == -EALREADY) {
rc = 0;
} else {
fprintf(stderr,
"%s: '%s' llapi_lease_get_ext resync failed: %s.\n",
progname, fname, strerror(-rc));
goto free_layout;
}
}
/*
* There are no data replicas to resync but there might be ec mirrors
* that need to be re-synced.
*/
if (!stale_data_comp_count)
goto ec_resync;
/* get the read range [start, end) */
start = stale_comp_array[0].lrc_start;
end = stale_comp_array[0].lrc_end;
for (idx = 1; idx < stale_data_comp_count; idx++) {
if (stale_comp_array[idx].lrc_start < start)
start = stale_comp_array[idx].lrc_start;
if (end < stale_comp_array[idx].lrc_end)
end = stale_comp_array[idx].lrc_end;
}
rc = llapi_lease_check(fd);
if (rc != LL_LEASE_WRLCK) {
fprintf(stderr, "%s: '%s' lost lease lock.\n", progname, fname);
goto free_layout;
}
rc = llapi_mirror_resync_many_params(fd, layout, stale_comp_array,
stale_data_comp_count, start, end,
stats_interval_sec,
bandwidth_bytes_sec);
/* If resync succeeded, manually set lrc_synced for all components */
if (rc == 0)
for (int i = 0; i < stale_data_comp_count; i++)
stale_comp_array[i].lrc_synced = 1;
if (rc < 0) {
llapi_error(LLAPI_MSG_ERROR, rc, "fail to mirror resync '%s'\n",
fname);
goto free_layout;
}
ec_resync:
/*
* If we have stale ec comps then resync them, otherwise
* skip to updating the timestamps and finishing the resync.
*/
if (!stale_ec_comp_count)
goto finish_resync;
rc = llapi_ec_resync_many_params(
fd, layout, &stale_comp_array[stale_data_comp_count],
stale_ec_comp_count, stats_interval_sec, bandwidth_bytes_sec);
if (rc < 0) {
llapi_error(LLAPI_MSG_ERROR, rc, "fail to ec resync '%s'\n",
fname);
goto free_layout;
}
finish_resync:
rc2 = migrate_set_timestamps(fd, &stbuf);
if (rc2 < 0) {
fprintf(stderr, "%s: '%s' cannot set timestamps: %s\n",
progname, fname, strerror(-rc2));
if (!rc)
rc = rc2;
goto free_layout;
}
/* need to do the lease unlock even resync fails */
ioc->lil_mode = LL_LEASE_UNLCK;
ioc->lil_flags = LL_LEASE_RESYNC_DONE;
ioc->lil_count = 0;
for (idx = 0; idx < stale_data_comp_count + stale_ec_comp_count;
idx++) {
if (stale_comp_array[idx].lrc_synced) {
ioc->lil_ids[ioc->lil_count] =
stale_comp_array[idx].lrc_id;
ioc->lil_count++;
}
}
rc2 = llapi_lease_set(fd, ioc);
/**
* llapi_lease_set returns lease mode when it request to unlock
* the lease lock.
*/
if (rc2 <= 0) {
/* rc2 == 0 means lost lease lock */
if (rc2 == 0 && rc == 0)
rc = -EBUSY;
else
rc = rc2;
fprintf(stderr, "%s: resync file '%s' failed: %s.\n", progname,
fname, rc2 == 0 ? "lost lease lock" : strerror(-rc2));
llapi_lease_release(fd);
goto free_layout;
}
free_layout:
llapi_layout_free(layout);
close_fd:
close(fd);
error:
return rc;
}
static inline int lfs_mirror_resync(int argc, char **argv)
{
struct option long_opts[] = {
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'o', .name = "only", .has_arg = required_argument },
{ .val = 'W', .name = "bandwidth-limit", .has_arg = required_argument},
{ .val = LFS_STATS_OPT, .name = "stats", .has_arg = no_argument},
{ .val = LFS_STATS_INTERVAL_OPT,
.name = "stats-interval", .has_arg = required_argument},
{ .val = 'f', .name = "force-resync", .has_arg = no_argument },
{ .name = NULL } };
struct ll_ioc_lease *ioc = NULL;
__u16 mirror_ids[128] = { 0 };
unsigned int stats_interval_sec = 0;
unsigned long long bandwidth_bytes_sec = 0;
unsigned long long bandwidth_unit = ONE_MB;
int ids_nr = 0;
int c;
int rc = 0;
bool force_resync = false;
while ((c = getopt_long(argc, argv, "ho:W:f", long_opts, NULL)) >= 0) {
char *end;
switch (c) {
case 'o':
rc = parse_mirror_ids(mirror_ids,
sizeof(mirror_ids) / sizeof(__u16),
optarg);
if (rc < 0) {
fprintf(stderr,
"%s: bad mirror ids '%s'.\n",
argv[0], optarg);
goto error;
}
ids_nr = rc;
break;
case 'W':
if (llapi_parse_size(optarg, &bandwidth_bytes_sec,
&bandwidth_unit, 0) < 0) {
fprintf(stderr,
"error: %s: bad value for bandwidth '%s'\n",
argv[0], optarg);
goto error;
}
break;
case LFS_STATS_OPT:
stats_interval_sec = 5;
break;
case LFS_STATS_INTERVAL_OPT:
stats_interval_sec = strtol(optarg, &end, 0);
break;
case 'f':
force_resync = true;
break;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
rc = CMD_HELP;
goto error;
}
}
if (argc == optind) {
fprintf(stderr, "%s: no file name given.\n", argv[0]);
rc = CMD_HELP;
goto error;
}
if (ids_nr > 0 && argc > optind + 1) {
fprintf(stderr,
"%s: option '--only' cannot be used upon multiple files.\n",
argv[0]);
rc = CMD_HELP;
goto error;
}
if (ids_nr > 0) {
rc = verify_mirror_ids(argv[optind], mirror_ids, ids_nr);
if (rc < 0)
goto error;
}
/* set the lease on the file */
ioc = calloc(1, sizeof(*ioc) + sizeof(__u32) * IOC_IDS_MAX);
if (!ioc) {
fprintf(stderr, "%s: cannot alloc id array for ioc: %s.\n",
argv[0], strerror(errno));
rc = -errno;
goto error;
}
for (; optind < argc; optind++) {
rc = lfs_mirror_resync_file(argv[optind], ioc,
mirror_ids, ids_nr,
stats_interval_sec,
bandwidth_bytes_sec, force_resync);
/* ignore previous file's error, continue with next file */
/* reset ioc */
memset(ioc, 0, sizeof(*ioc) + sizeof(__u32) * IOC_IDS_MAX);
}
free(ioc);
error:
return rc;
}
static inline int verify_mirror_id_by_fd(int fd, __u16 mirror_id)
{
struct llapi_layout *layout;
int rc;
layout = llapi_layout_get_by_fd(fd, 0);
if (!layout) {
fprintf(stderr, "could not get layout.\n");
return -EINVAL;
}
rc = llapi_layout_comp_iterate(layout, find_mirror_id, &mirror_id);
if (rc < 0) {
fprintf(stderr, "failed to iterate layout\n");
llapi_layout_free(layout);
return rc;
} else if (rc == LLAPI_LAYOUT_ITER_CONT) {
fprintf(stderr, "does not find mirror with ID %u\n", mirror_id);
llapi_layout_free(layout);
return -EINVAL;
}
llapi_layout_free(layout);
return 0;
}
static inline int lfs_somsync_by_fd(int fd)
{
struct stat st;
int rc = 0;
/* flush dirty pages from clients */
rc = llapi_fsync(fd);
if (rc < 0)
goto out;
rc = fstat(fd, &st);
if (rc < 0)
rc = -errno;
/*
* After call fstat(), it already gets OST attrs to the client,
* when close the file, MDS will update the LSOM data itself
* according the size and blocks information from the client.
*/
out:
close(fd);
return rc;
}
static inline int lfs_somsync_by_path(const char *fname)
{
int fd;
int rc = 0;
fd = open(fname, O_RDONLY | O_NOATIME);
if (fd < 0) {
rc = -errno;
fprintf(stderr,
"%s somsync: cannot open '%s': %s\n",
progname, fname, strerror(errno));
return rc;
}
rc = lfs_somsync_by_fd(fd);
if (rc < 0) {
fprintf(stderr,
"%s somsync: cannot synchronize SOM data of '%s': %s\n",
progname, fname, strerror(-rc));
return rc;
}
return 0;
}
static inline int lfs_somsync_by_fid(const char *lustre_dir,
const struct lu_fid *fid)
{
int fd = -1;
char fidstr[FID_LEN];
int rc = 0;
snprintf(fidstr, sizeof(fidstr), DFID, PFID(fid));
fd = llapi_open_by_fid(lustre_dir, fid, O_RDONLY | O_NOATIME);
if (fd < 0) {
rc = -errno;
fprintf(stderr,
"%s somsync: cannot open '%s': %s\n",
progname, fidstr, strerror(-rc));
return rc;
}
rc = lfs_somsync_by_fd(fd);
if (rc < 0) {
fprintf(stderr,
"%s somsync: cannot synchronize SOM data of '%s': %s\n",
progname, fidstr, strerror(-rc));
return rc;
}
return 0;
}
enum {
LFS_SOMSYNC_CLIENT_MOUNT = 1,
};
static int lfs_somsync(int argc, char **argv)
{
struct option long_opts[] = {
{ "by-fid", required_argument, NULL, LFS_SOMSYNC_CLIENT_MOUNT },
{ NULL },
};
const char *client_mount = NULL;
int c;
int rc = 0, rc1;
while ((c = getopt_long(argc, argv, "", long_opts, NULL)) != -1) {
switch (c) {
case LFS_SOMSYNC_CLIENT_MOUNT:
client_mount = optarg;
break;
default:
fprintf(stderr,
"%s somsync: unrecognized option '%s'\n",
progname, argv[optind - 1]);
return CMD_HELP;
}
}
if (client_mount != NULL) {
/* lfs somsync --by-fid MOUNT FID ... */
char mntdir[PATH_MAX];
struct lu_fid fid;
char *fidstr;
int found;
if (argc == optind) {
fprintf(stderr, "%s somsync: missing FID\n", progname);
return CMD_HELP;
}
rc = llapi_search_mounts(client_mount, 0, mntdir, NULL);
if (rc < 0) {
fprintf(stderr,
"%s somsync: invalid MOUNT '%s': %s\n",
progname, client_mount, strerror(-rc));
return rc;
}
rc = 0;
while (optind < argc) {
found = 0;
fidstr = argv[optind++];
while (*fidstr == '[')
fidstr++;
found = sscanf(fidstr, SFID, RFID(&fid));
if (found != 3) {
fprintf(stderr,
"%s somsync: unrecognized FID: %s\n",
progname, argv[optind - 1]);
return -EINVAL;
}
rc1 = lfs_somsync_by_fid(mntdir, &fid);
if (rc1 && !rc)
rc = rc1;
}
return rc;
}
/* lfs somsync FILE ... */
if (argc == optind) {
fprintf(stderr, "%s somsync: missing FILE\n", progname);
return CMD_HELP;
}
rc = 0;
while (optind < argc) {
rc1 = lfs_somsync_by_path(argv[optind++]);
if (rc1 && !rc)
rc = rc1;
}
return rc;
}
/**
* check_same_file() - Check whether two files are the same file
* @fd: File descriptor of file1
* @f2: Path name of file2
*
* Return:
* * %0 same file
* * %1 not the same file
* * %negative error code
*/
static inline int check_same_file(int fd, const char *f2)
{
struct stat stbuf1;
struct stat stbuf2;
if (fstat(fd, &stbuf1) < 0)
return -errno;
if (stat(f2, &stbuf2) < 0)
return 1;
if (stbuf1.st_rdev == stbuf2.st_rdev &&
stbuf1.st_ino == stbuf2.st_ino)
return 0;
return 1;
}
static inline int lfs_mirror_read(int argc, char **argv)
{
int rc = CMD_HELP;
__u16 mirror_id = 0;
const char *outfile = NULL;
char *fname;
int fd = 0;
int outfd;
int c;
void *buf;
const size_t buflen = DEFAULT_IO_BUFLEN;
ssize_t page_size;
off_t pos;
struct option long_opts[] = {
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'N', .name = "mirror-id", .has_arg = required_argument },
{ .val = 'o', .name = "outfile", .has_arg = required_argument },
{ .name = NULL } };
while ((c = getopt_long(argc, argv, "hN:o:", long_opts, NULL)) >= 0) {
char *end;
switch (c) {
case 'N': {
unsigned long int id;
errno = 0;
id = strtoul(optarg, &end, 0);
if (errno != 0 || *end != '\0' || id == 0 ||
id > UINT16_MAX) {
fprintf(stderr,
"%s %s: invalid mirror ID '%s'\n",
progname, argv[0], optarg);
return rc;
}
mirror_id = (__u16)id;
break;
}
case 'o':
outfile = optarg;
break;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
return CMD_HELP;
}
}
if (argc == optind) {
fprintf(stderr, "%s %s: no mirrored file provided\n",
progname, argv[0]);
return rc;
} else if (argc > optind + 1) {
fprintf(stderr, "%s %s: too many files\n", progname, argv[0]);
return rc;
}
if (mirror_id == 0) {
fprintf(stderr, "%s %s: no valid mirror ID is provided\n",
progname, argv[0]);
return rc;
}
/* open mirror file */
fname = argv[optind];
fd = open(fname, O_DIRECT | O_RDONLY);
if (fd < 0) {
fprintf(stderr, "%s %s: cannot open '%s': %s\n",
progname, argv[0], fname, strerror(errno));
return rc;
}
/* verify mirror id */
rc = verify_mirror_id_by_fd(fd, mirror_id);
if (rc) {
fprintf(stderr,
"%s %s: cannot find mirror with ID %u in '%s'\n",
progname, argv[0], mirror_id, fname);
goto close_fd;
}
/* open output file - O_EXCL ensures output is not the same as input */
if (outfile) {
outfd = open(outfile, O_EXCL | O_WRONLY | O_CREAT, 0644);
if (outfd < 0) {
fprintf(stderr, "%s %s: cannot create file '%s': %s\n",
progname, argv[0], outfile, strerror(errno));
rc = -errno;
goto close_fd;
}
} else {
outfd = STDOUT_FILENO;
}
page_size = sysconf(_SC_PAGESIZE);
if (page_size < 0) {
rc = -errno;
goto close_fd;
}
/* allocate buffer */
rc = posix_memalign(&buf, page_size, buflen);
if (rc) {
fprintf(stderr, "%s %s: posix_memalign() failed: %s\n",
progname, argv[0], strerror(rc));
goto close_outfd;
}
(void)mlock(buf, buflen);
pos = 0;
while (1) {
ssize_t bytes_read;
ssize_t written = 0;
bytes_read = llapi_mirror_read(fd, mirror_id, buf, buflen, pos);
if (bytes_read < 0) {
rc = bytes_read;
fprintf(stderr,
"%s %s: fail to read data from mirror %u: %s\n",
progname, argv[0], mirror_id, strerror(-rc));
goto free_buf;
}
/* EOF reached */
if (bytes_read == 0)
break;
while (written < bytes_read) {
ssize_t written2;
written2 = write(outfd, buf + written,
bytes_read - written);
if (written2 < 0) {
fprintf(stderr,
"%s %s: fail to write %s: %s\n",
progname, argv[0], outfile ? : "STDOUT",
strerror(errno));
rc = -errno;
goto free_buf;
}
written += written2;
}
if (written != bytes_read) {
fprintf(stderr,
"%s %s: written %ld bytes does not match with %ld read.\n",
progname, argv[0], written, bytes_read);
rc = -EIO;
goto free_buf;
}
pos += bytes_read;
}
fsync(outfd);
rc = 0;
free_buf:
(void)munlock(buf, buflen);
free(buf);
close_outfd:
if (outfile)
close(outfd);
close_fd:
close(fd);
return rc;
}
static inline int lfs_mirror_write(int argc, char **argv)
{
int rc = CMD_HELP;
__u16 mirror_id = 0;
const char *inputfile = NULL;
char *fname;
int fd = 0;
int inputfd;
int c;
void *buf;
const size_t buflen = DEFAULT_IO_BUFLEN;
off_t pos;
ssize_t page_size = sysconf(_SC_PAGESIZE);
struct ll_ioc_lease_id ioc;
struct option long_opts[] = {
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'i', .name = "inputfile", .has_arg = required_argument },
{ .val = 'N', .name = "mirror-id", .has_arg = required_argument },
{ .name = NULL } };
while ((c = getopt_long(argc, argv, "hi:N:", long_opts, NULL)) >= 0) {
char *end;
switch (c) {
case 'N': {
unsigned long int id;
errno = 0;
id = strtoul(optarg, &end, 0);
if (errno != 0 || *end != '\0' || id == 0 ||
id > UINT16_MAX) {
fprintf(stderr,
"%s %s: invalid mirror ID '%s'\n",
progname, argv[0], optarg);
return rc;
}
mirror_id = (__u16)id;
break;
}
case 'i':
inputfile = optarg;
break;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
return CMD_HELP;
}
}
if (argc == optind) {
fprintf(stderr, "%s %s: no mirrored file provided\n",
progname, argv[0]);
return rc;
} else if (argc > optind + 1) {
fprintf(stderr, "%s %s: too many files\n", progname, argv[0]);
return rc;
}
if (mirror_id == 0) {
fprintf(stderr, "%s %s: no valid mirror ID is provided\n",
progname, argv[0]);
return rc;
}
/* open mirror file */
fname = argv[optind];
fd = open(fname, O_DIRECT | O_WRONLY);
if (fd < 0) {
fprintf(stderr, "%s %s: cannot open '%s': %s\n",
progname, argv[0], fname, strerror(errno));
return rc;
}
/* verify mirror id */
rc = verify_mirror_id_by_fd(fd, mirror_id);
if (rc) {
fprintf(stderr,
"%s %s: cannot find mirror with ID %u in '%s'\n",
progname, argv[0], mirror_id, fname);
goto close_fd;
}
/* open input file */
if (inputfile) {
rc = check_same_file(fd, inputfile);
if (rc == 0) {
fprintf(stderr,
"%s %s: input file cannot be the mirrored file '%s'\n",
progname, argv[0], fname);
goto close_fd;
}
if (rc < 0)
goto close_fd;
inputfd = open(inputfile, O_RDONLY, 0644);
if (inputfd < 0) {
fprintf(stderr, "%s %s: cannot open file '%s': %s\n",
progname, argv[0], inputfile, strerror(errno));
rc = -errno;
goto close_fd;
}
} else {
inputfd = STDIN_FILENO;
}
/* allocate buffer */
rc = posix_memalign(&buf, page_size, buflen);
if (rc) {
fprintf(stderr, "%s %s: posix_memalign returns %d\n",
progname, argv[0], rc);
goto close_inputfd;
}
(void)mlock(buf, buflen);
/* prepare target mirror components instantiation */
ioc.lil_mode = LL_LEASE_WRLCK;
ioc.lil_flags = LL_LEASE_RESYNC;
ioc.lil_mirror_id = mirror_id;
rc = llapi_lease_set(fd, (struct ll_ioc_lease *)&ioc);
if (rc < 0) {
fprintf(stderr,
"%s %s: '%s' llapi_lease_get_ext failed: %s\n",
progname, argv[0], fname, strerror(errno));
goto free_buf;
}
pos = 0;
while (1) {
ssize_t bytes_read;
ssize_t written;
size_t to_write;
rc = llapi_lease_check(fd);
if (rc != LL_LEASE_WRLCK) {
fprintf(stderr, "%s %s: '%s' lost lease lock\n",
progname, argv[0], fname);
goto free_buf;
}
bytes_read = read(inputfd, buf, buflen);
if (bytes_read < 0) {
rc = bytes_read;
fprintf(stderr,
"%s %s: fail to read data from '%s': %s\n",
progname, argv[0], inputfile ? : "STDIN",
strerror(errno));
rc = -errno;
goto free_buf;
}
/* EOF reached */
if (bytes_read == 0)
break;
/* round up to page align to make direct IO happy. */
to_write = (bytes_read + page_size - 1) & ~(page_size - 1);
written = llapi_mirror_write(fd, mirror_id, buf, to_write,
pos);
if (written < 0) {
rc = written;
fprintf(stderr,
"%s %s: fail to write to mirror %u: %s\n",
progname, argv[0], mirror_id,
strerror(-rc));
goto free_buf;
}
pos += bytes_read;
}
if (pos & (page_size - 1)) {
rc = llapi_mirror_truncate(fd, mirror_id, pos);
if (rc < 0) {
llapi_error(LLAPI_MSG_ERROR, rc,
"fail to trucate mirror %u of file '%s' to %ld",
mirror_id, fname, pos);
goto free_buf;
}
}
ioc.lil_mode = LL_LEASE_UNLCK;
ioc.lil_flags = LL_LEASE_RESYNC_DONE;
ioc.lil_count = 0;
rc = llapi_lease_set(fd, (struct ll_ioc_lease *)&ioc);
if (rc <= 0) {
if (rc == 0)
rc = -EBUSY;
fprintf(stderr,
"%s %s: release lease lock of '%s' failed: %s\n",
progname, argv[0], fname, strerror(-rc));
goto free_buf;
}
rc = 0;
free_buf:
(void)munlock(buf, buflen);
free(buf);
close_inputfd:
if (inputfile)
close(inputfd);
close_fd:
close(fd);
return rc;
}
static inline int get_other_mirror_ids(int fd, __u16 *ids, __u16 exclude_id)
{
struct llapi_layout *layout;
struct collect_ids_data cid = { .cid_ids = ids,
.cid_count = 0,
.cid_exclude = exclude_id, };
int rc;
layout = llapi_layout_get_by_fd(fd, 0);
if (!layout) {
fprintf(stderr, "could not get layout\n");
return -EINVAL;
}
rc = llapi_layout_comp_iterate(layout, collect_mirror_id, &cid);
if (rc < 0) {
fprintf(stderr, "failed to iterate layout\n");
llapi_layout_free(layout);
return rc;
}
llapi_layout_free(layout);
return cid.cid_count;
}
#ifndef MIRROR_ID_NEG
#define MIRROR_ID_NEG 0x8000
#endif
static inline int lfs_mirror_copy(int argc, char **argv)
{
int rc = CMD_HELP;
__u16 read_mirror_id = 0;
__u16 ids[128] = { 0 };
int count = 0;
struct llapi_layout *layout = NULL;
struct llapi_resync_comp comp_array[1024] = { { 0 } };
int comp_size = 0;
char *fname;
int fd = 0;
int c;
int i;
ssize_t copied;
struct ll_ioc_lease *ioc = NULL;
struct ll_ioc_lease_id *resync_ioc;
struct option long_opts[] = {
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'i', .name = "read-mirror", .has_arg = required_argument },
{ .val = 'o', .name = "write-mirror", .has_arg = required_argument },
{ .name = NULL } };
char cmd[PATH_MAX];
snprintf(cmd, sizeof(cmd), "%s %s", progname, argv[0]);
progname = cmd;
while ((c = getopt_long(argc, argv, "hi:o:", long_opts, NULL)) >= 0) {
char *end;
switch (c) {
case 'i': {
unsigned long int id;
errno = 0;
id = strtoul(optarg, &end, 0);
if (errno != 0 || *end != '\0' || id == 0 ||
id > UINT16_MAX) {
fprintf(stderr,
"%s: invalid read mirror ID '%s'\n",
progname, optarg);
return rc;
}
read_mirror_id = (__u16)id;
break;
}
case 'o':
if (!strcmp(optarg, "-1")) {
/* specify all other mirrors */
ids[0] = (__u16)-1;
count = 1;
} else {
count = parse_mirror_ids((__u16 *)ids,
ARRAY_SIZE(ids),
optarg);
if (count < 0)
return rc;
}
break;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
return CMD_HELP;
}
}
if (argc == optind) {
fprintf(stderr, "%s %s: no mirrored file provided\n",
progname, argv[0]);
return rc;
} else if (argc > optind + 1) {
fprintf(stderr, "%s %s: too many files\n", progname, argv[0]);
return rc;
}
if (read_mirror_id == 0) {
fprintf(stderr,
"%s %s: no valid read mirror ID %d is provided\n",
progname, argv[0], read_mirror_id);
return rc;
}
if (count == 0) {
fprintf(stderr,
"%s %s: no write mirror ID is provided\n",
progname, argv[0]);
return rc;
}
for (i = 0; i < count; i++) {
if (read_mirror_id == ids[i]) {
fprintf(stderr,
"%s %s: read and write mirror ID cannot be the same\n",
progname, argv[0]);
return rc;
}
}
/* open mirror file */
fname = argv[optind];
fd = open(fname, O_DIRECT | O_RDWR);
if (fd < 0) {
fprintf(stderr, "%s %s: cannot open '%s': %s\n",
progname, argv[0], fname, strerror(errno));
return rc;
}
/* write to all other mirrors */
if (ids[0] == (__u16)-1) {
count = get_other_mirror_ids(fd, ids, read_mirror_id);
if (count <= 0) {
rc = count;
fprintf(stderr,
"%s %s: failed to get other mirror ids in '%s': %d\n",
progname, argv[0], fname, rc);
goto close_fd;
}
}
/* verify mirror id */
rc = verify_mirror_id_by_fd(fd, read_mirror_id);
if (rc) {
fprintf(stderr,
"%s %s: cannot find mirror with ID %u in '%s'\n",
progname, argv[0], read_mirror_id, fname);
goto close_fd;
}
for (i = 0; i < count; i++) {
rc = verify_mirror_id_by_fd(fd, ids[i]);
if (rc) {
fprintf(stderr,
"%s %s: cannot find mirror with ID %u in '%s'\n",
progname, argv[0], ids[i], fname);
goto close_fd;
}
}
ioc = calloc(1, sizeof(*ioc) + sizeof(__u32) * IOC_IDS_MAX);
if (!ioc) {
fprintf(stderr,
"%s %s: cannot alloc comp id array for ioc: %s\n",
progname, argv[0], strerror(errno));
rc = -errno;
goto close_fd;
}
/* get stale component info */
layout = llapi_layout_get_by_fd(fd, 0);
if (!layout) {
fprintf(stderr, "%s %s: failed to get layout of '%s': %s\n",
progname, argv[0], fname, strerror(errno));
rc = -errno;
goto free_ioc;
}
comp_size = llapi_mirror_find_stale(layout, comp_array,
ARRAY_SIZE(comp_array),
ids, count);
llapi_layout_free(layout);
if (comp_size < 0) {
rc = comp_size;
goto free_ioc;
}
/* prepare target mirror components instantiation */
resync_ioc = (struct ll_ioc_lease_id *)ioc;
resync_ioc->lil_mode = LL_LEASE_WRLCK;
resync_ioc->lil_flags = LL_LEASE_RESYNC;
if (count == 1)
resync_ioc->lil_mirror_id = ids[0];
else
resync_ioc->lil_mirror_id = read_mirror_id | MIRROR_ID_NEG;
rc = llapi_lease_set(fd, ioc);
if (rc < 0) {
fprintf(stderr,
"%s %s: '%s' llapi_lease_get_ext failed: %s\n",
progname, argv[0], fname, strerror(errno));
goto free_ioc;
}
copied = llapi_mirror_copy_many(fd, read_mirror_id, ids, count);
if (copied < 0) {
rc = copied;
fprintf(stderr, "%s %s: copy error: %d\n",
progname, argv[0], rc);
goto free_ioc;
}
fprintf(stdout, "mirror copied successfully: ");
for (i = 0; i < copied; i++)
fprintf(stdout, "%d ", ids[i]);
fprintf(stdout, "\n");
ioc->lil_mode = LL_LEASE_UNLCK;
ioc->lil_flags = LL_LEASE_RESYNC_DONE;
ioc->lil_count = 0;
for (i = 0; i < comp_size; i++) {
int j;
for (j = 0; j < copied; j++) {
if (comp_array[i].lrc_mirror_id != ids[j])
continue;
ioc->lil_ids[ioc->lil_count] = comp_array[i].lrc_id;
ioc->lil_count++;
}
}
rc = llapi_lease_set(fd, ioc);
if (rc <= 0) {
if (rc == 0)
rc = -EBUSY;
fprintf(stderr,
"%s %s: release lease lock of '%s' failed: %s\n",
progname, argv[0], fname, strerror(errno));
goto free_ioc;
}
rc = 0;
free_ioc:
free(ioc);
close_fd:
close(fd);
return rc;
}
/**
* struct verify_chunk - Mirror chunk to be verified.
* @chunk: [start, end) of the chunk.
* @mirror_count: Number of mirror ids in @mirror_id array.
* @mirror_id: Array of valid mirror ids that cover the chunk.
*/
struct verify_chunk {
struct lu_extent chunk;
unsigned int mirror_count;
__u16 mirror_id[LUSTRE_MIRROR_COUNT_MAX];
};
/**
* print_chunks() - Print chunk information.
* @fname: Mirrored file name.
* @chunks: Array of chunks.
* @chunk_count: Number of chunks in @chunks array.
*
* This function prints [start, end) of each chunk in @chunks
* for mirrored file @fname, and also prints the valid mirror ids
* that cover the chunk.
*
* Return: void.
*/
static inline
void print_chunks(const char *fname, struct verify_chunk *chunks,
int chunk_count)
{
int i;
int j;
fprintf(stdout, "Chunks to be verified in %s:\n", fname);
for (i = 0; i < chunk_count; i++) {
fprintf(stdout, DEXT, PEXT(&chunks[i].chunk));
if (chunks[i].mirror_count == 0)
fprintf(stdout, "\t[");
else {
fprintf(stdout, "\t[%u", chunks[i].mirror_id[0]);
for (j = 1; j < chunks[i].mirror_count; j++)
fprintf(stdout, ", %u", chunks[i].mirror_id[j]);
}
fprintf(stdout, "]\t%u\n", chunks[i].mirror_count);
}
fprintf(stdout, "\n");
}
/**
* print_checksums() - Print CRC-32 checksum values.
* @chunk: A chunk and its corresponding valid mirror ids.
* @crc: CRC-32 checksum values on the chunk for each valid mirror.
* @pos: Start offset of the chunk.
* @len: Length of the chunk.
* @layout: Mirror component layout.
*
* This function prints CRC-32 checksum values on @chunk for
* each valid mirror that covers it.
*
* Return: void.
*/
static inline
void print_checksums(struct verify_chunk *chunk, unsigned long *crc,
unsigned long long pos, unsigned long long len,
struct llapi_layout *layout)
{
int i;
fprintf(stdout,
"CRC-32 checksum value for chunk "DEXT":\n", pos, pos + len);
for (i = 0; i < chunk->mirror_count; i++) {
bool is_stale = false;
/* Check if this mirror is stale */
if (layout) {
int rc = llapi_layout_comp_use(layout,
LLAPI_LAYOUT_COMP_USE_FIRST);
while (rc == 0) {
uint32_t mirror_id, comp_flags;
rc = llapi_layout_mirror_id_get(layout,
&mirror_id);
if (rc == 0 &&
mirror_id == chunk->mirror_id[i]) {
rc = llapi_layout_comp_flags_get(layout,
&comp_flags);
if (rc == 0 &&
(comp_flags & LCME_FL_STALE))
is_stale = true;
break;
}
rc = llapi_layout_comp_use(layout,
LLAPI_LAYOUT_COMP_USE_NEXT);
if (rc < 0)
break;
}
}
if (is_stale)
fprintf(stdout, "Mirror %u:\t%#lx (stale)\n",
chunk->mirror_id[i], crc[i]);
else
fprintf(stdout, "Mirror %u:\t%#lx\n",
chunk->mirror_id[i], crc[i]);
}
fprintf(stdout, "\n");
}
/**
* print_stale_mirrors() - Print stale mirrors with checksums.
* @fd: File descriptor of the mirrored file.
* @layout: Mirror component layout.
* @file_size: Size of the file.
* @verbose: Verbose mode.
*
* This function finds stale mirrors and displays them with checksums
* in the same format as normal verification output.
*
* Return: number of stale mirrors found, or negative error code.
*/
static inline
int print_stale_mirrors(int fd, struct llapi_layout *layout,
size_t file_size, int verbose)
{
int stale_count = 0;
int rc = 0;
bool header_printed = false;
if (!layout)
goto check_only;
rc = llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_FIRST);
while (rc == 0) {
uint32_t mirror_id, comp_flags;
rc = llapi_layout_mirror_id_get(layout, &mirror_id);
if (rc < 0)
break;
rc = llapi_layout_comp_flags_get(layout, &comp_flags);
if (rc < 0)
break;
if (comp_flags & LCME_FL_STALE) {
bool is_nosync = comp_flags & LCME_FL_NOSYNC;
const char *flag_str = is_nosync ?
"nosync,stale" : "stale";
/* Print header only once */
if (!header_printed) {
fprintf(stdout,
"CRC-32 checksum value for chunk [0, 0x%lx):\n",
file_size);
header_printed = true;
}
/* Calculate and display checksum for stale mirror */
const size_t buflen = 4 * 1024 * 1024; /* 4M */
void *buf;
ssize_t bytes_read;
unsigned long crc;
if (posix_memalign(&buf, sysconf(_SC_PAGESIZE),
buflen) == 0) {
bytes_read = llapi_mirror_read(fd,
mirror_id, buf,
MIN(buflen,
file_size),
0);
if (bytes_read > 0) {
crc = crc32(crc32(0L, Z_NULL, 0), buf,
bytes_read);
fprintf(stdout,
"Mirror %u:\t%#lx (%s)\n",
mirror_id, crc, flag_str);
} else {
fprintf(stdout,
"Mirror %u:\t(%s - read error)\n",
mirror_id, flag_str);
}
free(buf);
} else {
fprintf(stdout,
"Mirror %u:\t(%s - alloc error)\n",
mirror_id, flag_str);
}
/* Only count non-nosync stale as errors */
if (!is_nosync)
stale_count++;
}
rc = llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_NEXT);
if (rc < 0) {
rc = 0; /* End of components */
break;
}
}
if (header_printed)
fprintf(stdout, "\n");
return stale_count;
check_only:
/* Just count stale mirrors without printing */
rc = llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_FIRST);
while (rc == 0) {
uint32_t comp_flags;
rc = llapi_layout_comp_flags_get(layout, &comp_flags);
if (rc < 0)
break;
if (comp_flags & LCME_FL_STALE)
stale_count++;
rc = llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_NEXT);
if (rc < 0) {
rc = 0; /* End of components */
break;
}
}
return stale_count;
}
/**
* filter_mirror_id() - Filter specified mirror ids.
* @chunks: Array of chunks.
* @chunk_count: Number of chunks in @chunks array.
* @mirror_ids: Specified mirror ids to be verified.
* @ids_nr: Number of specified mirror ids.
*
* This function scans valid mirror ids that cover each chunk in @chunks
* and filters specified mirror ids.
*
* Return: void.
*/
static inline
void filter_mirror_id(struct verify_chunk *chunks, int chunk_count,
__u16 *mirror_ids, int ids_nr)
{
int i;
int j;
int k;
__u16 valid_id[LUSTRE_MIRROR_COUNT_MAX] = { 0 };
unsigned int valid_count = 0;
for (i = 0; i < chunk_count; i++) {
if (chunks[i].mirror_count == 0)
continue;
valid_count = 0;
for (j = 0; j < ids_nr; j++) {
for (k = 0; k < chunks[i].mirror_count; k++) {
if (chunks[i].mirror_id[k] == mirror_ids[j]) {
valid_id[valid_count] = mirror_ids[j];
valid_count++;
break;
}
}
}
memcpy(chunks[i].mirror_id, valid_id,
sizeof(__u16) * valid_count);
chunks[i].mirror_count = valid_count;
}
}
/**
* lfs_mirror_prepare_chunk() - Find mirror chunks to be verified.
* @layout: Mirror component list.
* @chunks: Array of chunks.
* @chunks_size: Array size of @chunks.
* @include_stale: Whether to include stale mirrors in chunks.
*
* This function scans the components in @layout from offset 0 to LUSTRE_EOF
* to find out chunk segments and store them in @chunks array.
*
* The @mirror_id array in each element of @chunks will store the valid
* mirror ids that cover the chunk. If @include_stale is false and a mirror
* component covering the chunk has LCME_FL_STALE or LCME_FL_OFFLINE flag,
* then the mirror id will not be stored into the @mirror_id array, and the
* chunk for that mirror will not be verified. If @include_stale is true,
* stale mirrors will be included for verification.
*
* The @mirror_count in each element of @chunks will store the number of
* mirror ids in @mirror_id array. If @mirror_count is 0, it indicates the
* chunk is invalid in all of the mirrors. And if @mirror_count is 1, it
* indicates the chunk is valid in only one mirror. In both cases, the
* chunk will not be verified.
*
* Here is an example:
*
* 0 1M 2M 3M 4M EOF
* +------+-------------+--------------------+
* | | | S | mirror1
* +------+------+------+------+-------------+
* | | S | S | | mirror2
* +-------------+------+------+-------------+
*
* prepared @chunks array will contain 5 elements:
* (([0, 1M), [1, 2], 2),
* ([1M, 2M), [1, 2], 2),
* ([2M, 3M), [1], 1),
* ([3M, 4M], [], 0),
* ([4M, EOF), [2], 1))
*
* Return: the actual array size of @chunks on success
* or a negative error code on failure.
*/
static inline
int lfs_mirror_prepare_chunk(struct llapi_layout *layout,
struct verify_chunk *chunks,
size_t chunks_size, bool include_stale)
{
uint64_t start;
uint64_t end;
uint32_t mirror_id;
uint32_t flags;
int idx = 0;
int i = 0;
int rc = 0;
memset(chunks, 0, sizeof(*chunks) * chunks_size);
while (1) {
rc = llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_FIRST);
if (rc < 0) {
fprintf(stderr,
"%s: move to the first layout component: %s.\n",
progname, strerror(errno));
goto error;
}
i = 0;
rc = 0;
chunks[idx].chunk.e_end = LUSTRE_EOF;
while (rc == 0) {
rc = llapi_layout_comp_extent_get(layout, &start, &end);
if (rc < 0) {
fprintf(stderr,
"%s: llapi_layout_comp_extent_get failed: %s.\n",
progname, strerror(errno));
goto error;
}
if (start > chunks[idx].chunk.e_start ||
end <= chunks[idx].chunk.e_start)
goto next;
if (end < chunks[idx].chunk.e_end)
chunks[idx].chunk.e_end = end;
rc = llapi_layout_comp_flags_get(layout, &flags);
if (rc < 0) {
fprintf(stderr,
"%s: llapi_layout_comp_flags_get failed: %s.\n",
progname, strerror(errno));
goto error;
}
/* Skip nosync components - they are intentionally
* excluded from resync so should not be verified
*/
if (flags & LCME_FL_NOSYNC)
goto next;
if (flags & LCME_FL_PARITY)
goto next;
if (!include_stale && (flags & LCME_FL_STALE ||
flags & LCME_FL_OFFLINE))
goto next;
rc = llapi_layout_mirror_id_get(layout, &mirror_id);
if (rc < 0) {
fprintf(stderr,
"%s: llapi_layout_mirror_id_get failed: %s.\n",
progname, strerror(errno));
goto error;
}
if (i >= ARRAY_SIZE(chunks[idx].mirror_id)) {
fprintf(stderr,
"%s: mirror_id array is too small.\n",
progname);
rc = -EINVAL;
goto error;
}
chunks[idx].mirror_id[i] = mirror_id;
i++;
next:
rc = llapi_layout_comp_use(layout,
LLAPI_LAYOUT_COMP_USE_NEXT);
if (rc < 0) {
fprintf(stderr,
"%s: move to the next layout component: %s.\n",
progname, strerror(errno));
goto error;
}
} /* loop through all components */
chunks[idx].mirror_count = i;
if (chunks[idx].chunk.e_end == LUSTRE_EOF)
break;
idx++;
if (idx >= chunks_size) {
fprintf(stderr, "%s: chunks array is too small.\n",
progname);
rc = -EINVAL;
goto error;
}
chunks[idx].chunk.e_start = chunks[idx - 1].chunk.e_end;
}
error:
return rc < 0 ? rc : idx + 1;
}
/**
* lfs_mirror_verify_chunk() - Verify a chunk.
* @fd: File descriptor of the mirrored file.
* @file_size: Size of the mirrored file.
* @chunk: A chunk and its corresponding valid mirror ids.
* @verbose: Verbose mode.
* @layout: Mirror component layout.
*
* This function verifies a @chunk contains exactly the same data
* ammong the mirrors that cover it.
*
* If @verbose is specified, then the function will print where the
* differences are if the data do not match. Otherwise, it will
* just return an error in that case.
*
* Return: 0 on success or a negative error code on failure.
*/
static inline
int lfs_mirror_verify_chunk(int fd, size_t file_size,
struct verify_chunk *chunk, int verbose,
struct llapi_layout *layout)
{
const size_t buflen = DEFAULT_IO_BUFLEN;
void *buf;
size_t page_size;
ssize_t bytes_read;
ssize_t bytes_done;
size_t count;
off_t pos;
unsigned long crc;
unsigned long crc_array[LUSTRE_MIRROR_COUNT_MAX] = { 0 };
int i;
int rc = 0;
if (file_size == 0)
return 0;
page_size = sysconf(_SC_PAGESIZE);
if (page_size < 0) {
rc = -errno;
return rc;
}
rc = posix_memalign(&buf, page_size, buflen);
if (rc) /* error code is returned directly */
return -rc;
(void)mlock(buf, buflen);
if (verbose > 1) {
fprintf(stdout, "Verifying chunk "DEXT" on mirror:",
PEXT(&chunk->chunk));
for (i = 0; i < chunk->mirror_count; i++)
fprintf(stdout, " %u", chunk->mirror_id[i]);
fprintf(stdout, "\n");
}
bytes_done = 0;
count = MIN(chunk->chunk.e_end, file_size) - chunk->chunk.e_start;
pos = chunk->chunk.e_start;
while (bytes_done < count) {
bool print = false;
/* compute initial CRC-32 checksum */
crc = crc32(0L, Z_NULL, 0);
memset(crc_array, 0, sizeof(crc_array));
bytes_read = 0;
for (i = 0; i < chunk->mirror_count; i++) {
bytes_read = llapi_mirror_read(fd, chunk->mirror_id[i],
buf, MIN(buflen, count),
pos);
if (bytes_read < 0) {
rc = bytes_read;
fprintf(stderr,
"%s: error reading from mirror %u: %s\n",
progname, chunk->mirror_id[i],
strerror(-rc));
goto error;
}
/* compute new CRC-32 checksum */
crc_array[i] = crc32(crc, buf, bytes_read);
}
/* compare CRC-32 checksum values */
for (i = 1; i < chunk->mirror_count; i++) {
if (crc_array[i] != crc_array[0]) {
rc = -EUCLEAN;
fprintf(stderr,
"%s: chunk "DEXT" has different checksum value on mirror %u:%lx and mirror %u:%lx: %s\n",
progname, PEXT(&chunk->chunk),
chunk->mirror_id[0], crc_array[0],
chunk->mirror_id[i], crc_array[i],
strerror(-rc));
print = true;
}
}
if (verbose || print)
print_checksums(chunk, crc_array, pos, bytes_read,
layout);
pos += bytes_read;
bytes_done += bytes_read;
}
if (verbose > 1 && rc == 0) {
fprintf(stdout, "Verifying chunk "DEXT" on mirror:",
PEXT(&chunk->chunk));
for (i = 0; i < chunk->mirror_count; i++)
fprintf(stdout, " %u", chunk->mirror_id[i]);
fprintf(stdout, " PASS\n\n");
}
error:
(void)munlock(buf, buflen);
free(buf);
return rc;
}
struct mirror_info {
uint32_t mirror_id;
uint32_t flags;
unsigned long checksum;
bool has_prefwr;
bool has_prefrd;
};
/**
* compare_mirror_checksums() - Compare checksums of all mirrors for a chunk.
* @fd: File descriptor of the mirrored file.
* @chunk: A chunk and its corresponding valid mirror ids.
* @mirrors: Array to store mirror information including checksums.
* @mirror_count: Number of mirrors.
*
* This function reads data from each mirror and computes checksums to
* determine which mirrors have consistent vs inconsistent data.
*
* Return: 0 on success or a negative error code on failure.
*/
static int compare_mirror_checksums(int fd, struct verify_chunk *chunk,
struct mirror_info *mirrors,
int mirror_count)
{
const size_t buflen = DEFAULT_IO_BUFLEN;
void *buf;
size_t page_size;
ssize_t bytes_read;
off_t pos;
unsigned long crc;
int i;
int rc = 0;
page_size = sysconf(_SC_PAGESIZE);
if (page_size < 0)
return -errno;
rc = posix_memalign(&buf, page_size, buflen);
if (rc)
return -rc;
(void)mlock(buf, buflen);
/* Read from each mirror and compute checksum */
pos = chunk->chunk.e_start;
crc = crc32(0L, Z_NULL, 0);
for (i = 0; i < mirror_count; i++) {
bytes_read = llapi_mirror_read(fd, mirrors[i].mirror_id,
buf, MIN(buflen,
chunk->chunk.e_end - pos), pos);
if (bytes_read < 0) {
rc = bytes_read;
fprintf(stderr,
"%s: error reading from mirror %u: %s\n",
progname, mirrors[i].mirror_id,
strerror(-rc));
goto out;
}
mirrors[i].checksum = crc32(crc, buf, bytes_read);
}
out:
(void)munlock(buf, buflen);
free(buf);
return rc;
}
/**
* get_mirror_properties() - Get properties of all mirrors.
* @layout: Mirror component layout.
* @mirrors: Array to store mirror information.
* @max_mirrors: Maximum number of mirrors to process.
*
* This function iterates through all mirror components and extracts
* their properties including mirror IDs and preference flags.
*
* Return: Number of mirrors found or a negative error code on failure.
*/
static int get_mirror_properties(struct llapi_layout *layout,
struct mirror_info *mirrors, int max_mirrors)
{
int mirror_count = 0;
uint32_t mirror_id, flags;
int rc;
rc = llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_FIRST);
if (rc < 0)
return rc;
do {
rc = llapi_layout_mirror_id_get(layout, &mirror_id);
if (rc < 0)
return rc;
rc = llapi_layout_comp_flags_get(layout, &flags);
if (rc < 0)
return rc;
/* Check if this mirror ID is already in our array */
int found = -1;
for (int i = 0; i < mirror_count; i++) {
if (mirrors[i].mirror_id == mirror_id) {
found = i;
break;
}
}
if (found == -1) {
if (mirror_count >= max_mirrors)
return -E2BIG;
found = mirror_count++;
mirrors[found].mirror_id = mirror_id;
mirrors[found].flags = 0;
mirrors[found].has_prefwr = false;
mirrors[found].has_prefrd = false;
}
mirrors[found].flags |= flags;
if (flags & LCME_FL_PREF_WR)
mirrors[found].has_prefwr = true;
if (flags & LCME_FL_PREF_RD)
mirrors[found].has_prefrd = true;
rc = llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_NEXT);
} while (rc == 0);
return mirror_count;
}
/**
* auto_mark_stale_mirrors() - Intelligently mark inconsistent mirrors as stale.
* @fd: File descriptor of the mirrored file.
* @layout: Mirror component layout.
* @fname: File name for error reporting.
* @verbose: Verbosity level for output.
*
* This function analyzes mirror inconsistencies and marks only the mirrors
* that should be stale based on intelligent heuristics:
* - For 2 mirrors with different checksums: mark non-prefwr mirror as stale
* - For multiple mirrors: mark mirrors that don't match the majority checksum
* - Use prefwr/prefrd flags as tie-breakers when determining majority
* - Only processes chunks where mirrors have different checksums
*
* Return: 0 on success or a negative error code on failure.
*/
static inline
int auto_mark_stale_mirrors(int fd, struct llapi_layout *layout,
const char *fname, int verbose)
{
struct verify_chunk chunks_array[1024] = { };
struct mirror_info mirrors[LUSTRE_MIRROR_COUNT_MAX];
int chunk_count = 0;
int mirror_count;
int rc;
/* Get all mirror properties */
mirror_count = get_mirror_properties(layout, mirrors,
LUSTRE_MIRROR_COUNT_MAX);
if (mirror_count < 0) {
fprintf(stderr,
"%s: failed to get mirror properties for '%s': %s\n",
progname, fname, strerror(-mirror_count));
return mirror_count;
}
if (mirror_count < 2) {
fprintf(stderr, "%s: file '%s' has less than 2 mirrors\n",
progname, fname);
return -EINVAL;
}
chunk_count = lfs_mirror_prepare_chunk(layout, chunks_array,
ARRAY_SIZE(chunks_array),
false);
if (chunk_count < 0) {
fprintf(stderr, "%s: failed to prepare chunks for '%s': %s\n",
progname, fname, strerror(-chunk_count));
return chunk_count;
}
/* For each chunk, compare mirror checksums */
for (int i = 0; i < chunk_count; i++) {
struct verify_chunk *chunk = &chunks_array[i];
unsigned long checksums[LUSTRE_MIRROR_COUNT_MAX];
int checksum_groups[LUSTRE_MIRROR_COUNT_MAX];
int group_counts[LUSTRE_MIRROR_COUNT_MAX] = {0};
int num_groups = 0;
int majority_group = -1;
int max_count = 0;
/* Compare checksums for this chunk */
rc = compare_mirror_checksums(fd, chunk, mirrors, mirror_count);
if (rc < 0) {
fprintf(stderr,
"%s: failed to compare checksums for '%s': %s\n",
progname, fname, strerror(-rc));
return rc;
}
/* Group mirrors by checksum */
for (int j = 0; j < mirror_count; j++) {
checksums[j] = mirrors[j].checksum;
checksum_groups[j] = -1;
/* Find existing group with same checksum */
for (int g = 0; g < num_groups; g++) {
for (int k = 0; k < j; k++) {
if (checksum_groups[k] == g &&
checksums[k] == checksums[j]) {
checksum_groups[j] = g;
group_counts[g]++;
goto next_mirror;
}
}
}
/* Create new group if not found */
if (checksum_groups[j] == -1) {
checksum_groups[j] = num_groups;
group_counts[num_groups] = 1;
num_groups++;
}
next_mirror:
continue;
}
/* Find majority group, with tie-breaking */
for (int g = 0; g < num_groups; g++) {
if (group_counts[g] > max_count) {
max_count = group_counts[g];
majority_group = g;
} else if (group_counts[g] == max_count &&
max_count > 1) {
/* prefer group with prefwr mirrors */
bool current_has_prefwr = false;
bool candidate_has_prefwr = false;
/* Check if current majority group has
* prefwr mirrors
*/
for (int k = 0; k < mirror_count; k++) {
if (checksum_groups[k] == majority_group &&
mirrors[k].has_prefwr) {
current_has_prefwr = true;
break;
}
}
/* Check if candidate group has
* prefwr mirrors
*/
for (int k = 0; k < mirror_count; k++) {
if (checksum_groups[k] == g &&
mirrors[k].has_prefwr) {
candidate_has_prefwr = true;
break;
}
}
/* Prefer group with prefwr mirrors */
if (candidate_has_prefwr &&
!current_has_prefwr) {
majority_group = g;
}
}
}
/* If all mirrors have same checksum, nothing to mark */
if (num_groups == 1)
continue;
/* Validate that we found a majority group */
if (majority_group == -1) {
fprintf(stderr,
"%s: failed to determine majority group for chunk %d in '%s'\n",
progname, i, fname);
return -EINVAL;
}
if (verbose > 1) {
printf("Found %d checksum groups for chunk %d, majority group %d has %d mirrors\n",
num_groups, i, majority_group, max_count);
}
/* Special case for 2 mirrors with different checksums */
if (mirror_count == 2 && num_groups == 2) {
int prefwr_mirror = -1;
int non_prefwr_mirror = -1;
/* Find prefwr vs non-prefwr mirrors */
for (int j = 0; j < mirror_count; j++) {
if (mirrors[j].has_prefwr)
prefwr_mirror = j;
else
non_prefwr_mirror = j;
}
/* mark non-prefwr as stale */
if (prefwr_mirror >= 0 && non_prefwr_mirror >= 0) {
printf("Marking non-prefwr mirror %u as stale\n",
mirrors[non_prefwr_mirror].mirror_id);
rc = lfs_component_set_by_mirror((char *)fname,
mirrors[non_prefwr_mirror].mirror_id,
LCME_FL_STALE, 0);
if (rc != 0) {
fprintf(stderr,
"%s: failed to mark mirror %u stale in '%s': %s\n",
progname,
mirrors[non_prefwr_mirror].mirror_id,
fname, strerror(-rc));
return rc;
}
continue;
}
}
/* Mark mirrors that don't match majority as stale */
for (int j = 0; j < mirror_count; j++) {
if (checksum_groups[j] != majority_group) {
/* This mirror is inconsistent, mark it stale */
printf("Marking mirror %u as stale (has_prefwr=%s)\n",
mirrors[j].mirror_id,
mirrors[j].has_prefwr ? "yes" : "no");
rc = lfs_component_set_by_mirror((char *)fname,
mirrors[j].mirror_id,
LCME_FL_STALE, 0);
if (rc != 0) {
fprintf(stderr,
"%s: failed to mark mirror %u stale in '%s': %s\n",
progname, mirrors[j].mirror_id,
fname, strerror(-rc));
return rc;
}
}
}
}
if (verbose > 0) {
printf("Successfully marked inconsistent mirrors as stale in '%s'\n",
fname);
}
return 0;
}
/**
* lfs_mirror_prepare_ec() - Find EC parity components to be verified.
* @layout: Mirror component list.
* @ecs: Output array of EC parity component ids.
* @ecs_size: Capacity of the @ecs array.
*
* Walks @layout and collects the ids of every EC parity component
* (LCME_FL_PARITY) that is eligible for verification. NOSYNC components are
* skipped silently; STALE parity components are treated as an error since
* their on-disk parities cannot be meaningfully compared against the data.
*
* Return: number of EC parity components written to @ecs on success, or a
* negative errno on failure.
*/
static inline int lfs_mirror_prepare_ec(struct llapi_layout *layout, __u32 *ecs,
size_t ecs_size)
{
int i, rc;
memset(ecs, 0, sizeof(*ecs) * ecs_size);
rc = llapi_layout_comp_use(layout, LLAPI_LAYOUT_COMP_USE_FIRST);
if (rc < 0) {
fprintf(stderr,
"%s: move to the first layout component: %s.\n",
progname, strerror(errno));
goto error;
}
i = 0;
rc = 0;
while (rc == 0) {
uint32_t id, flags;
rc = llapi_layout_comp_flags_get(layout, &flags);
if (rc < 0) {
fprintf(stderr,
"%s: llapi_layout_comp_flags_get failed: %s.\n",
progname, strerror(errno));
goto error;
}
if (!(flags & LCME_FL_PARITY))
goto next;
/* Skip nosync components - they are intentionally kept stale */
if (flags & LCME_FL_NOSYNC)
goto next;
rc = llapi_layout_comp_id_get(layout, &id);
if (rc < 0) {
fprintf(stderr,
"%s: Failed to get component id.\n",
progname);
goto error;
}
if (flags & LCME_FL_STALE) {
rc = -ESTALE;
fprintf(stderr,
"%s: EC component 0x%x is stale.\n",
progname, id);
goto error;
}
if (i >= ecs_size) {
fprintf(stderr, "%s: ec array is too small.\n",
progname);
rc = -EINVAL;
goto error;
}
ecs[i] = id;
i++;
next:
rc = llapi_layout_comp_use(layout,
LLAPI_LAYOUT_COMP_USE_NEXT);
if (rc < 0) {
fprintf(stderr,
"%s: move to the next layout component: %s.\n",
progname, strerror(errno));
goto error;
}
}
error:
return rc < 0 ? rc : i;
}
/**
* lfs_mirror_verify_file() - Verify a mirrored file.
* @fname: Mirrored file name.
* @mirror_ids: Specified mirror ids to be verified.
* @ids_nr: Number of specified mirror ids.
* @verbose: Verbose mode.
* @stale: %True if stale marking is requested
*
* This function verifies that each SYNC mirror of a mirrored file
* specified by @fname contains exactly the same data.
*
* If @mirror_ids is specified, then the function will verify the
* mirrors specified by @mirror_ids contain exactly the same data.
*
* If @verbose is specified, then the function will print where the
* differences are if the data do not match. Otherwise, it will
* just return an error in that case.
*
* Return: 0 on success or a negative error code on failure.
*/
static inline
int lfs_mirror_verify_file(const char *fname, __u16 *mirror_ids, int ids_nr,
int verbose, int stale)
{
struct verify_chunk chunks_array[1024] = { };
__u32 ec_array[1024] = { };
struct llapi_layout *layout = NULL;
struct stat stbuf;
uint32_t flr_state;
int fd;
int chunk_count;
int ec_count;
int idx = 0;
int rc = 0;
int rc1 = 0;
int rc2 = 0;
if (stat(fname, &stbuf) < 0) {
fprintf(stderr, "%s: cannot stat file '%s': %s.\n",
progname, fname, strerror(errno));
rc = -errno;
goto error;
}
if (!S_ISREG(stbuf.st_mode)) {
fprintf(stderr, "%s: '%s' is not a regular file.\n",
progname, fname);
rc = -EINVAL;
goto error;
}
if (stbuf.st_size == 0) {
if (verbose)
fprintf(stdout, "%s: '%s' file size is 0.\n",
progname, fname);
rc = 0;
goto error;
}
/* Allow mirror verify even without the key on encrypted files */
fd = open(fname, O_DIRECT | O_RDONLY | O_CIPHERTEXT);
if (fd < 0) {
fprintf(stderr, "%s: cannot open '%s': %s.\n",
progname, fname, strerror(errno));
rc = -errno;
goto error;
}
rc = llapi_lease_acquire(fd, LL_LEASE_RDLCK);
if (rc < 0) {
fprintf(stderr, "%s: '%s' llapi_lease_acquire failed: %s.\n",
progname, fname, strerror(errno));
goto close_fd;
}
layout = llapi_layout_get_by_fd(fd, 0);
if (!layout) {
fprintf(stderr, "%s: '%s' llapi_layout_get_by_fd failed: %s.\n",
progname, fname, strerror(errno));
rc = -errno;
llapi_lease_release(fd);
goto close_fd;
}
rc = llapi_layout_flags_get(layout, &flr_state);
if (rc < 0) {
fprintf(stderr, "%s: '%s' llapi_layout_flags_get failed: %s.\n",
progname, fname, strerror(errno));
rc = -errno;
goto free_layout;
}
flr_state &= LCM_FL_FLR_MASK;
switch (flr_state) {
case LCM_FL_NONE:
rc = -EINVAL;
fprintf(stderr, "%s: '%s' file state error: %s.\n",
progname, fname, llapi_layout_flags_string(flr_state));
goto free_layout;
default:
break;
}
/* find out mirror chunks to be verified */
chunk_count = lfs_mirror_prepare_chunk(layout, chunks_array,
ARRAY_SIZE(chunks_array),
ids_nr > 0);
if (chunk_count < 0) {
rc = chunk_count;
goto free_layout;
}
if (ids_nr > 0)
/* filter specified mirror ids */
filter_mirror_id(chunks_array, chunk_count, mirror_ids, ids_nr);
if (verbose > 2)
print_chunks(fname, chunks_array, chunk_count);
if (ids_nr <= 0) {
/* Check for and display stale mirrors */
rc1 = print_stale_mirrors(fd, layout, stbuf.st_size, verbose);
if (rc1 > 0) {
rc2 = -ESTALE;
} else if (rc1 < 0) {
rc2 = rc1;
}
}
for (idx = 0; idx < chunk_count; idx++) {
if (chunks_array[idx].chunk.e_start >= stbuf.st_size) {
if (verbose)
fprintf(stdout,
"%s: '%s' chunk "DEXT" exceeds file size %#llx: skipped\n",
progname, fname,
PEXT(&chunks_array[idx].chunk),
(unsigned long long)stbuf.st_size);
break;
}
if (chunks_array[idx].mirror_count == 0) {
fprintf(stderr,
"%s: '%s' chunk "DEXT" is invalid in all of the mirrors: ",
progname, fname,
PEXT(&chunks_array[idx].chunk));
if (verbose) {
fprintf(stderr, "skipped\n");
continue;
}
rc = -EINVAL;
fprintf(stderr, "failed\n");
goto free_layout;
}
if (chunks_array[idx].mirror_count == 1) {
if (verbose)
fprintf(stdout,
"%s: '%s' chunk "DEXT" is only valid in mirror %u: skipped\n",
progname, fname,
PEXT(&chunks_array[idx].chunk),
chunks_array[idx].mirror_id[0]);
continue;
}
rc = llapi_lease_check(fd);
if (rc != LL_LEASE_RDLCK) {
fprintf(stderr, "%s: '%s' lost lease lock.\n",
progname, fname);
goto free_layout;
}
/* verify one chunk */
rc1 = lfs_mirror_verify_chunk(fd, stbuf.st_size,
&chunks_array[idx], verbose,
layout);
if (rc1 < 0) {
rc2 = rc1;
if (!verbose && !stale) {
rc = rc1;
goto free_layout;
}
}
}
/* If stale marking is requested and there were verification errors,
* try to automatically mark mirrors as stale
*/
if (stale && rc2 < 0) {
int mark_rc = auto_mark_stale_mirrors(fd, layout, fname,
verbose);
if (mark_rc < 0) {
fprintf(stderr,
"%s: failed to auto-mark stale mirrors in '%s': %s\n",
progname, fname, strerror(-mark_rc));
} else {
rc2 = mark_rc;
}
}
if (rc2 < 0) {
rc = rc2;
goto free_layout;
}
/* find ec components to be verified */
ec_count = lfs_mirror_prepare_ec(layout, ec_array,
ARRAY_SIZE(ec_array));
if (ec_count < 0) {
rc = ec_count;
goto free_layout;
}
rc = llapi_ec_verify_comps(fd, layout, ec_array, ec_count);
free_layout:
llapi_layout_free(layout);
llapi_lease_release(fd);
close_fd:
close(fd);
error:
return rc;
}
/**
* lfs_mirror_verify() - Parse and execute lfs mirror verify command.
* @argc: The count of lfs mirror verify command line arguments.
* @argv: Array of strings for lfs mirror verify command line arguments.
*
* This function parses lfs mirror verify command and verifies the
* specified mirrored file(s).
*
* Return: 0 on success or a negative error code on failure.
*/
static inline int lfs_mirror_verify(int argc, char **argv)
{
__u16 mirror_ids[LUSTRE_MIRROR_COUNT_MAX] = { 0 };
int ids_nr = 0;
int c;
int stale = 0;
int verbose = 0;
int rc = 0;
int rc1 = 0;
char cmd[PATH_MAX];
struct option long_opts[] = {
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'o', .name = "only", .has_arg = required_argument },
{ .val = 's', .name = "stale", .has_arg = no_argument},
{ .val = 'v', .name = "verbose", .has_arg = no_argument },
{ .name = NULL } };
snprintf(cmd, sizeof(cmd), "%s %s", progname, argv[0]);
progname = cmd;
while ((c = getopt_long(argc, argv, "ho:sv", long_opts, NULL)) >= 0) {
switch (c) {
case 'o':
rc = parse_mirror_ids(mirror_ids,
ARRAY_SIZE(mirror_ids),
optarg);
if (rc < 0) {
fprintf(stderr,
"%s: bad mirror ids '%s'.\n",
progname, optarg);
goto error;
}
ids_nr = rc;
if (ids_nr < 2) {
fprintf(stderr,
"%s: at least 2 mirror ids needed with '--only' option.\n",
progname);
rc = CMD_HELP;
goto error;
}
break;
case 's':
stale = 1;
break;
case 'v':
verbose++;
break;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
rc = CMD_HELP;
goto error;
}
}
if (argc == optind) {
fprintf(stderr, "%s: no file name given.\n", progname);
rc = CMD_HELP;
goto error;
}
if (ids_nr > 0 && argc > optind + 1) {
fprintf(stderr,
"%s: '--only' cannot be used upon multiple files.\n",
progname);
rc = CMD_HELP;
goto error;
}
if (ids_nr > 0) {
rc = verify_mirror_ids(argv[optind], mirror_ids, ids_nr);
if (rc < 0)
goto error;
}
rc = 0;
for (; optind < argc; optind++) {
rc1 = lfs_mirror_verify_file(argv[optind], mirror_ids, ids_nr,
verbose, stale);
if (rc1 < 0)
rc = rc1;
}
error:
return rc;
}
static void lustre_som_swab(struct lustre_som_attrs *attrs)
{
#if __BYTE_ORDER == __BIG_ENDIAN
__swab16s(&attrs->lsa_valid);
__swab64s(&attrs->lsa_size);
__swab64s(&attrs->lsa_blocks);
#endif
}
enum lfs_som_type {
LFS_SOM_SIZE = 0x1,
LFS_SOM_BLOCKS = 0x2,
LFS_SOM_FLAGS = 0x4,
LFS_SOM_ATTR_ALL = LFS_SOM_SIZE | LFS_SOM_BLOCKS |
LFS_SOM_FLAGS,
};
static int lfs_getsom(int argc, char **argv)
{
const char *path;
struct lustre_som_attrs *attrs;
char buf[sizeof(*attrs) + 64];
enum lfs_som_type type = LFS_SOM_ATTR_ALL;
int rc = 0, c;
while ((c = getopt(argc, argv, "bfhs")) != -1) {
switch (c) {
case 'b':
type = LFS_SOM_BLOCKS;
break;
case 'f':
type = LFS_SOM_FLAGS;
break;
case 's':
type = LFS_SOM_SIZE;
break;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
return CMD_HELP;
}
}
argc -= optind;
argv += optind;
if (argc != 1) {
fprintf(stderr, "%s: %s\n",
progname, argc == 0 ? "miss file target" :
"input more than 2 files");
return CMD_HELP;
}
path = argv[0];
attrs = (void *)buf;
rc = lgetxattr(path, "trusted.som", attrs, sizeof(buf));
if (rc < 0) {
rc = -errno;
fprintf(stderr, "%s failed to get som xattr: %s (%d)\n",
argv[0], strerror(errno), errno);
return rc;
}
lustre_som_swab(attrs);
switch (type) {
case LFS_SOM_ATTR_ALL:
printf("file: %s size: %llu blocks: %llu flags: %x\n",
path, (unsigned long long)attrs->lsa_size,
(unsigned long long)attrs->lsa_blocks,
attrs->lsa_valid);
break;
case LFS_SOM_SIZE:
printf("%llu\n", (unsigned long long)attrs->lsa_size);
break;
case LFS_SOM_BLOCKS:
printf("%llu\n", (unsigned long long)attrs->lsa_blocks);
break;
case LFS_SOM_FLAGS:
printf("%x\n", attrs->lsa_valid);
break;
default:
fprintf(stderr, "%s: unknown option\n", progname);
return CMD_HELP;
}
return 0;
}
static int lfs_pcc_attach(int argc, char **argv)
{
struct option long_opts[] = {
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'i', .name = "id", .has_arg = required_argument },
{ .val = 'r', .name = "readonly", .has_arg = no_argument },
{ .val = 'w', .name = "write", .has_arg = no_argument },
{ .name = NULL } };
int c;
int rc = 0;
__u32 attach_id = 0;
const char *path;
char *end;
char fullpath[PATH_MAX];
enum lu_pcc_type type = LU_PCC_READONLY;
optind = 0;
while ((c = getopt_long(argc, argv, "hi:rw",
long_opts, NULL)) != -1) {
switch (c) {
case 'i':
errno = 0;
attach_id = strtoul(optarg, &end, 0);
if (errno != 0 || *end != '\0' ||
attach_id == 0 || attach_id > UINT32_MAX) {
fprintf(stderr,
"error: %s: bad archive ID '%s'\n",
progname, optarg);
return CMD_HELP;
}
break;
case 'r':
type = LU_PCC_READONLY;
break;
case 'w':
type = LU_PCC_READWRITE;
break;
case '?':
return CMD_HELP;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
return CMD_HELP;
}
}
if (argc <= optind) {
fprintf(stderr, "%s: must specify one or more file names\n",
argv[0]);
return CMD_HELP;
}
while (optind < argc) {
int rc2;
path = argv[optind++];
if (!realpath(path, fullpath)) {
fprintf(stderr, "%s: could not find path '%s': %s\n",
argv[0], path, strerror(errno));
if (rc == 0)
rc = -EINVAL;
continue;
}
rc2 = llapi_pcc_attach(fullpath, attach_id, type);
if (rc2 < 0) {
fprintf(stderr,
"%s: cannot attach '%s' to PCC with attach ID '%u': %s\n",
argv[0], path, attach_id, strerror(-rc2));
if (rc == 0)
rc = rc2;
}
}
return rc;
}
static int lfs_pcc_attach_fid(int argc, char **argv)
{
struct option long_opts[] = {
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'i', .name = "id", .has_arg = required_argument },
{ .val = 'r', .name = "readonly", .has_arg = no_argument },
{ .val = 'w', .name = "write", .has_arg = no_argument },
{ .val = 'm', .name = "mnt", .has_arg = required_argument },
{ .name = NULL } };
char short_opts[] = "i:m:rw";
int c;
int rc = 0;
__u32 attach_id = 0;
char *end;
const char *mntpath = NULL;
const char *fidstr;
enum lu_pcc_type type = LU_PCC_READONLY;
optind = 0;
while ((c = getopt_long(argc, argv, short_opts,
long_opts, NULL)) != -1) {
switch (c) {
case 'i':
errno = 0;
attach_id = strtoul(optarg, &end, 0);
if (errno != 0 || *end != '\0' ||
attach_id > UINT32_MAX) {
fprintf(stderr,
"error: %s: bad attach ID '%s'\n",
argv[0], optarg);
return CMD_HELP;
}
break;
case 'r':
type = LU_PCC_READONLY;
break;
case 'w':
type = LU_PCC_READWRITE;
break;
case 'm':
mntpath = optarg;
break;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
return CMD_HELP;
}
}
if (!mntpath) {
fprintf(stderr, "%s: must specify Lustre mount point\n",
argv[0]);
return CMD_HELP;
}
if (argc <= optind) {
fprintf(stderr, "%s: must specify one or more fids\n", argv[0]);
return CMD_HELP;
}
while (optind < argc) {
int rc2;
fidstr = argv[optind++];
rc2 = llapi_pcc_attach_fid_str(mntpath, fidstr,
attach_id, type);
if (rc2 < 0) {
fprintf(stderr,
"%s: cannot attach '%s' on '%s' to PCC with attach ID '%u': %s\n",
argv[0], fidstr, mntpath, attach_id,
strerror(-rc2));
}
if (rc == 0 && rc2 < 0)
rc = rc2;
}
return rc;
}
static int lfs_pcc_detach(int argc, char **argv)
{
struct option long_opts[] = {
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'k', .name = "keep", .has_arg = no_argument },
{ .name = NULL } };
char short_opts[] = "hk";
int c;
int rc = 0;
const char *path;
char fullpath[PATH_MAX];
__u32 detach_flags = PCC_DETACH_FL_UNCACHE;
optind = 0;
while ((c = getopt_long(argc, argv, short_opts,
long_opts, NULL)) != -1) {
switch (c) {
case 'k':
detach_flags = PCC_DETACH_FL_NONE;
break;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
return CMD_HELP;
}
}
if (argc <= optind) {
fprintf(stderr, "%s: must specify one or more FIDs\n", argv[0]);
return CMD_HELP;
}
while (optind < argc) {
int rc2;
path = argv[optind++];
if (!realpath(path, fullpath)) {
fprintf(stderr, "%s: could not find path '%s': %s\n",
argv[0], path, strerror(errno));
if (rc == 0)
rc = -EINVAL;
continue;
}
rc2 = llapi_pcc_detach_file(fullpath, detach_flags);
if (rc2 < 0 && rc2 != -ENOENT) {
rc2 = -errno;
fprintf(stderr, "%s: cannot detach '%s' from PCC: %s\n",
argv[0], path, strerror(errno));
if (rc == 0)
rc = rc2;
}
}
return rc;
}
static int lfs_pcc_detach_fid(int argc, char **argv)
{
struct option long_opts[] = {
{ .val = 'h', .name = "help", .has_arg = no_argument },
{ .val = 'k', .name = "keep", .has_arg = no_argument },
{ .val = 'm', .name = "mnt", .has_arg = required_argument },
{ .name = NULL } };
char short_opts[] = "hkm:";
const char *mntpath = NULL;
const char *fidstr;
__u32 detach_flags = PCC_DETACH_FL_UNCACHE;
int rc = 0;
int c;
int dirfd;
optind = 0;
while ((c = getopt_long(argc, argv, short_opts,
long_opts, NULL)) != -1) {
switch (c) {
case 'k':
detach_flags = PCC_DETACH_FL_NONE;
break;
case 'm':
mntpath = optarg;
break;
case '?':
return CMD_HELP;
default:
fprintf(stderr, "%s: unrecognized option '%s'\n",
progname, argv[optind - 1]);
fallthrough;
case 'h':
return CMD_HELP;
}
}
/* for backward compatibility, allow mntpath without -m */
if (!mntpath) {
if (argc <= optind) {
fprintf(stderr, "%s: must specify Lustre mount point\n",
argv[0]);
return CMD_HELP;
}
mntpath = argv[optind++];
}
if (argc <= optind) {
fprintf(stderr, "%s: must specify one or more FIDs\n", argv[0]);
return CMD_HELP;
}
dirfd = open(mntpath, O_RDONLY | O_DIRECTORY);
if (dirfd < 0) {
rc = -errno;
fprintf(stderr, "%s: cannot open directory '%s': %s",
argv[0], mntpath, strerror(errno));
return rc;
}
while (optind < argc) {
struct lu_fid fid;
int rc2;
fidstr = argv[optind++];
rc2 = llapi_fid_parse(fidstr, &fid, NULL);
if (rc2) {
fprintf(stderr, "%s: '%s' is not a valid FID\n",
argv[0], fidstr);
if (rc == 0)
rc = rc2;
continue;
}
rc2 = llapi_pcc_detach_at(dirfd, &fid, detach_flags);
if (rc2 < 0) {
fprintf(stderr,
"%s: cannot detach '%s' on '%s' from PCC: %s\n",
argv[0], fidstr, mntpath, strerror(-rc2));
if (rc == 0)
rc = rc2;
}
}
close(dirfd);
return rc;
}
static int lfs_pcc_state(int argc, char **argv)
{
int rc = 0;
const char *path;
struct lu_pcc_state state;
optind = 1;
if (argc <= 1) {
fprintf(stderr, "%s: must specify one or more file names\n",
progname);
return CMD_HELP;
}
while (optind < argc) {
int rc2;
path = argv[optind++];
rc2 = llapi_pcc_state_get(path, &state);
if (rc2 < 0) {
if (rc == 0)
rc = rc2;
fprintf(stderr,
"%s: cannot get PCC state of '%s': %s\n",
argv[0], path, strerror(-rc2));
continue;
}
printf("file: %s", path);
printf(", type: %s", pcc_type2string(state.pccs_type));
if (state.pccs_type == LU_PCC_NONE &&
state.pccs_open_count == 0) {
if (state.pccs_flags & PCC_STATE_FL_ATTACHING)
printf(", flags: attaching");
printf("\n");
continue;
}
printf(", PCC_file: %s", state.pccs_path);
if (state.pccs_flags & PCC_STATE_FL_UNLINKED)
printf(" (unlinked)");
printf(", open_count: %u", state.pccs_open_count);
printf(", flags: %x", state.pccs_flags);
printf("\n");
}
return rc;
}
static int lfs_pcc_delete(int argc, char **argv)
{
int rc = 0;
const char *path;
optind = 1;
if (argc <= 1) {
fprintf(stderr, "%s: must specify one or more file names\n",
argv[0]);
return CMD_HELP;
}
while (optind < argc) {
int rc2;
path = argv[optind++];
rc2 = mirror_split(path, 0, NULL,
MF_DESTROY | MF_FOREIGN, NULL);
if (rc2 < 0) {
if (rc == 0)
rc = rc2;
fprintf(stderr,
"%s: failed to delete PCC for '%s': %s\n",
argv[0], path, strerror(-rc2));
continue;
}
}
return rc;
}
static int lfs_pcc_pin(int argc, char **argv)
{
int rc = 0, c;
const char *path;
char *end;
char fullpath[PATH_MAX];
__u32 id = 0;
struct option long_opts[] = {
{ .val = 'i', .name = "id", .has_arg = required_argument },
{ .name = NULL } };
optind = 0;
while ((c = getopt_long(argc, argv, "i:",
long_opts, NULL)) != -1) {
switch (c) {
case 'i':
errno = 0;
id = strtoul(optarg, &end, 0);
if (errno != 0 || *end != '\0' ||
id == 0 || id >= UINT32_MAX) {
fprintf(stderr,
"error: %s: bad attach ID '%s'\n",
argv[0], optarg);
return CMD_HELP;
}
break;
case '?':
return CMD_HELP;
default:
fprintf(stderr, "%s: option '%s' unrecognized\n",
argv[0], argv[optind - 1]);
return CMD_HELP;
}
}
/* check parameters */
if (argc <= 1) {
fprintf(stderr, "%s: must specify one or more file names\n",
argv[0]);
return CMD_HELP;
}
while (optind < argc) {
int rc2;
path = argv[optind++];
if (!realpath(path, fullpath)) {
fprintf(stderr, "%s: could not find path '%s': %s\n",
argv[0], path, strerror(errno));
if (rc == 0)
rc = -EINVAL;
continue;
}
if (id == 0) {
rc2 = llapi_pcc_backend_id_get(fullpath,
LU_PCC_READONLY, &id);
if (rc2 < 0) {
fprintf(stderr,
"%s: failed to get id for '%s': %s\n",
argv[0], path, strerror(-rc2));
if (rc == 0)
rc = rc2;
continue;
}
}
rc2 = llapi_pcc_pin_file(fullpath, id);
if (rc2 < 0) {
fprintf(stderr, "%s: cannot pin '%s' for PCC: %s\n",
argv[0], path, strerror(-rc2));
if (rc == 0)
rc = rc2;
}
}
return rc;
}
static int lfs_pcc_unpin(int argc, char **argv)
{
int rc = 0, c;
const char *path;
char *end;
char fullpath[PATH_MAX];
__u32 id = 0;
struct option long_opts[] = {
{ .val = 'i', .name = "id", .has_arg = required_argument },
{ .name = NULL } };
optind = 0;
while ((c = getopt_long(argc, argv, "i:",
long_opts, NULL)) != -1) {
switch (c) {
case 'i':
errno = 0;
id = strtoul(optarg, &end, 0);
if (errno != 0 || *end != '\0' ||
id == 0 || id > UINT32_MAX) {
fprintf(stderr,
"error: %s: bad attach ID '%s'\n",
argv[0], optarg);
return CMD_HELP;
}
break;
case '?':
return CMD_HELP;
default:
fprintf(stderr, "%s: option '%s' unrecognized\n",
argv[0], argv[optind - 1]);
return CMD_HELP;
}
}
/* check parameters */
if (argc <= 1) {
fprintf(stderr, "%s: must specify one or more file names\n",
argv[0]);
return CMD_HELP;
}
while (optind < argc) {
int rc2;
path = argv[optind++];
if (!realpath(path, fullpath)) {
fprintf(stderr, "%s: could not find path '%s': %s\n",
argv[0], path, strerror(errno));
if (rc == 0)
rc = -EINVAL;
continue;
}
if (id == 0) {
rc2 = llapi_pcc_backend_id_get(fullpath,
LU_PCC_READONLY, &id);
if (rc2 < 0) {
fprintf(stderr,
"%s: failed to get id for '%s': %s\n",
argv[0], path, strerror(-rc2));
if (rc == 0)
rc = rc2;
continue;
}
}
rc2 = llapi_pcc_unpin_file(fullpath, id);
if (rc2 < 0) {
fprintf(stderr, "%s: cannot unpin '%s' for PCC: %s\n",
argv[0], path, strerror(-rc2));
if (rc == 0)
rc = rc2;
}
}
return rc;
}
int main(int argc, char **argv)
{
int rc;
/* Ensure that liblustreapi constructor has run */
if (!llapi_liblustreapi_initialized())
fprintf(stderr, "liblustreapi was not properly initialized\n");
setlinebuf(stdout);
opterr = 0;
progname = program_invocation_short_name; /* Used in error messages */
llapi_set_command_name(argv[1]);
rc = cfs_parser(argc, argv, cmdlist);
llapi_clear_command_name();
return rc < 0 ? -rc : rc;
}
