Documentation/ABI/testing/sysfs-fs-f2fs

   1 What:           /sys/fs/f2fs/<disk>/gc_max_sleep_time
   2 Date:           July 2013
   3 Contact:        "Namjae Jeon" <namjae.jeon@samsung.com>
   4 Description:    Controls the maximum sleep time for gc_thread. Time
   5                 is in milliseconds.
   6
   7 What:           /sys/fs/f2fs/<disk>/gc_min_sleep_time
   8 Date:           July 2013
   9 Contact:        "Namjae Jeon" <namjae.jeon@samsung.com>
  10 Description:    Controls the minimum sleep time for gc_thread. Time
  11                 is in milliseconds.
  12
  13 What:           /sys/fs/f2fs/<disk>/gc_no_gc_sleep_time
  14 Date:           July 2013
  15 Contact:        "Namjae Jeon" <namjae.jeon@samsung.com>
  16 Description:    Controls the default sleep time for gc_thread. Time
  17                 is in milliseconds.
  18
  19 What:           /sys/fs/f2fs/<disk>/gc_idle
  20 Date:           July 2013
  21 Contact:        "Namjae Jeon" <namjae.jeon@samsung.com>
  22 Description:    Controls the victim selection policy for garbage collection.
  23                 Setting gc_idle = 0(default) will disable this option. Setting:
  24
  25                 ===========  ===============================================
  26                 gc_idle = 1  will select the Cost Benefit approach & setting
  27                 gc_idle = 2  will select the greedy approach & setting
  28                 gc_idle = 3  will select the age-threshold based approach.
  29                 ===========  ===============================================
  30
  31 What:           /sys/fs/f2fs/<disk>/reclaim_segments
  32 Date:           October 2013
  33 Contact:        "Jaegeuk Kim" <jaegeuk.kim@samsung.com>
  34 Description:    This parameter controls the number of prefree segments to be
  35                 reclaimed. If the number of prefree segments is larger than
  36                 the number of segments in the proportion to the percentage
  37                 over total volume size, f2fs tries to conduct checkpoint to
  38                 reclaim the prefree segments to free segments.
  39                 By default, 5% over total # of segments.
  40
  41 What:           /sys/fs/f2fs/<disk>/main_blkaddr
  42 Date:           November 2019
  43 Contact:        "Ramon Pantin" <pantin@google.com>
  44 Description:    Shows first block address of MAIN area.
  45
  46 What:           /sys/fs/f2fs/<disk>/ipu_policy
  47 Date:           November 2013
  48 Contact:        "Jaegeuk Kim" <jaegeuk.kim@samsung.com>
  49 Description:    Controls the in-place-update policy.
  50                 updates in f2fs. User can set:
  51
  52                 ===== =============== ===================================================
  53                 value policy          description
  54                 0x00  DISABLE         disable IPU(=default option in LFS mode)
  55                 0x01  FORCE           all the time
  56                 0x02  SSR             if SSR mode is activated
  57                 0x04  UTIL            if FS utilization is over threshold
  58                 0x08  SSR_UTIL        if SSR mode is activated and FS utilization is over
  59                                       threshold
  60                 0x10  FSYNC           activated in fsync path only for high performance
  61                                       flash storages. IPU will be triggered only if the
  62                                       # of dirty pages over min_fsync_blocks.
  63                                       (=default option)
  64                 0x20  ASYNC           do IPU given by asynchronous write requests
  65                 0x40  NOCACHE         disable IPU bio cache
  66                 0x80  HONOR_OPU_WRITE use OPU write prior to IPU write if inode has
  67                                       FI_OPU_WRITE flag
  68                 ===== =============== ===================================================
  69
  70                 Refer segment.h for details.
  71
  72 What:           /sys/fs/f2fs/<disk>/min_ipu_util
  73 Date:           November 2013
  74 Contact:        "Jaegeuk Kim" <jaegeuk.kim@samsung.com>
  75 Description:    Controls the FS utilization condition for the in-place-update
  76                 policies. It is used by F2FS_IPU_UTIL and F2FS_IPU_SSR_UTIL policies.
  77
  78 What:           /sys/fs/f2fs/<disk>/min_fsync_blocks
  79 Date:           September 2014
  80 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
  81 Description:    Controls the dirty page count condition for the in-place-update
  82                 policies.
  83
  84 What:           /sys/fs/f2fs/<disk>/min_seq_blocks
  85 Date:           August 2018
  86 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
  87 Description:    Controls the dirty page count condition for batched sequential
  88                 writes in writepages.
  89
  90 What:           /sys/fs/f2fs/<disk>/min_hot_blocks
  91 Date:           March 2017
  92 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
  93 Description:    Controls the dirty page count condition for redefining hot data.
  94
  95 What:           /sys/fs/f2fs/<disk>/min_ssr_sections
  96 Date:           October 2017
  97 Contact:        "Chao Yu" <yuchao0@huawei.com>
  98 Description:    Controls the free section threshold to trigger SSR allocation.
  99                 If this is large, SSR mode will be enabled early.
 100
 101 What:           /sys/fs/f2fs/<disk>/max_small_discards
 102 Date:           November 2013
 103 Contact:        "Jaegeuk Kim" <jaegeuk.kim@samsung.com>
 104 Description:    Controls the issue rate of discard commands that consist of small
 105                 blocks less than 2MB. The candidates to be discarded are cached during
 106                 checkpoint, and issued by issue_discard thread after checkpoint.
 107                 It is enabled by default.
 108
 109 What:           /sys/fs/f2fs/<disk>/max_ordered_discard
 110 Date:           October 2022
 111 Contact:        "Yangtao Li" <frank.li@vivo.com>
 112 Description:    Controls the maximum ordered discard, the unit size is one block(4KB).
 113                 Set it to 16 by default.
 114
 115 What:           /sys/fs/f2fs/<disk>/max_discard_request
 116 Date:           December 2021
 117 Contact:        "Konstantin Vyshetsky" <vkon@google.com>
 118 Description:    Controls the number of discards a thread will issue at a time.
 119                 Higher number will allow the discard thread to finish its work
 120                 faster, at the cost of higher latency for incoming I/O.
 121
 122 What:           /sys/fs/f2fs/<disk>/min_discard_issue_time
 123 Date:           December 2021
 124 Contact:        "Konstantin Vyshetsky" <vkon@google.com>
 125 Description:    Controls the interval the discard thread will wait between
 126                 issuing discard requests when there are discards to be issued and
 127                 no I/O aware interruptions occur.
 128
 129 What:           /sys/fs/f2fs/<disk>/mid_discard_issue_time
 130 Date:           December 2021
 131 Contact:        "Konstantin Vyshetsky" <vkon@google.com>
 132 Description:    Controls the interval the discard thread will wait between
 133                 issuing discard requests when there are discards to be issued and
 134                 an I/O aware interruption occurs.
 135
 136 What:           /sys/fs/f2fs/<disk>/max_discard_issue_time
 137 Date:           December 2021
 138 Contact:        "Konstantin Vyshetsky" <vkon@google.com>
 139 Description:    Controls the interval the discard thread will wait when there are
 140                 no discard operations to be issued.
 141
 142 What:           /sys/fs/f2fs/<disk>/discard_granularity
 143 Date:           July 2017
 144 Contact:        "Chao Yu" <yuchao0@huawei.com>
 145 Description:    Controls discard granularity of inner discard thread. Inner thread
 146                 will not issue discards with size that is smaller than granularity.
 147                 The unit size is one block(4KB), now only support configuring
 148                 in range of [1, 512]. Default value is 16.
 149                 For small devices, default value is 1.
 150
 151 What:           /sys/fs/f2fs/<disk>/umount_discard_timeout
 152 Date:           January 2019
 153 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
 154 Description:    Set timeout to issue discard commands during umount.
 155                 Default: 5 secs
 156
 157 What:           /sys/fs/f2fs/<disk>/pending_discard
 158 Date:           November 2021
 159 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
 160 Description:    Shows the number of pending discard commands in the queue.
 161
 162 What:           /sys/fs/f2fs/<disk>/max_victim_search
 163 Date:           January 2014
 164 Contact:        "Jaegeuk Kim" <jaegeuk.kim@samsung.com>
 165 Description:    Controls the number of trials to find a victim segment
 166                 when conducting SSR and cleaning operations. The default value
 167                 is 4096 which covers 8GB block address range.
 168
 169 What:           /sys/fs/f2fs/<disk>/migration_granularity
 170 Date:           October 2018
 171 Contact:        "Chao Yu" <yuchao0@huawei.com>
 172 Description:    Controls migration granularity of garbage collection on large
 173                 section, it can let GC move partial segment{s} of one section
 174                 in one GC cycle, so that dispersing heavy overhead GC to
 175                 multiple lightweight one.
 176
 177 What:           /sys/fs/f2fs/<disk>/dir_level
 178 Date:           March 2014
 179 Contact:        "Jaegeuk Kim" <jaegeuk.kim@samsung.com>
 180 Description:    Controls the directory level for large directory. If a
 181                 directory has a number of files, it can reduce the file lookup
 182                 latency by increasing this dir_level value. Otherwise, it
 183                 needs to decrease this value to reduce the space overhead.
 184                 The default value is 0.
 185
 186 What:           /sys/fs/f2fs/<disk>/ram_thresh
 187 Date:           March 2014
 188 Contact:        "Jaegeuk Kim" <jaegeuk.kim@samsung.com>
 189 Description:    Controls the memory footprint used by free nids and cached
 190                 nat entries. By default, 1 is set, which indicates
 191                 10 MB / 1 GB RAM.
 192
 193 What:           /sys/fs/f2fs/<disk>/cp_interval
 194 Date:           October 2015
 195 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
 196 Description:    Controls the checkpoint timing, set to 60 seconds by default.
 197
 198 What:           /sys/fs/f2fs/<disk>/idle_interval
 199 Date:           January 2016
 200 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
 201 Description:    Controls the idle timing of system, if there is no FS operation
 202                 during given interval.
 203                 Set to 5 seconds by default.
 204
 205 What:           /sys/fs/f2fs/<disk>/discard_idle_interval
 206 Date:           September 2018
 207 Contact:        "Chao Yu" <yuchao0@huawei.com>
 208 Contact:        "Sahitya Tummala" <quic_stummala@quicinc.com>
 209 Description:    Controls the idle timing of discard thread given
 210                 this time interval.
 211                 Default is 5 secs.
 212
 213 What:           /sys/fs/f2fs/<disk>/gc_idle_interval
 214 Date:           September 2018
 215 Contact:        "Chao Yu" <yuchao0@huawei.com>
 216 Contact:        "Sahitya Tummala" <quic_stummala@quicinc.com>
 217 Description:    Controls the idle timing for gc path. Set to 5 seconds by default.
 218
 219 What:           /sys/fs/f2fs/<disk>/iostat_enable
 220 Date:           August 2017
 221 Contact:        "Chao Yu" <yuchao0@huawei.com>
 222 Description:    Controls to enable/disable IO stat.
 223
 224 What:           /sys/fs/f2fs/<disk>/ra_nid_pages
 225 Date:           October 2015
 226 Contact:        "Chao Yu" <chao2.yu@samsung.com>
 227 Description:    Controls the count of nid pages to be readaheaded.
 228                 When building free nids, F2FS reads NAT blocks ahead for
 229                 speed up. Default is 0.
 230
 231 What:           /sys/fs/f2fs/<disk>/dirty_nats_ratio
 232 Date:           January 2016
 233 Contact:        "Chao Yu" <chao2.yu@samsung.com>
 234 Description:    Controls dirty nat entries ratio threshold, if current
 235                 ratio exceeds configured threshold, checkpoint will
 236                 be triggered for flushing dirty nat entries.
 237
 238 What:           /sys/fs/f2fs/<disk>/lifetime_write_kbytes
 239 Date:           January 2016
 240 Contact:        "Shuoran Liu" <liushuoran@huawei.com>
 241 Description:    Shows total written kbytes issued to disk.
 242
 243 What:           /sys/fs/f2fs/<disk>/features
 244 Date:           July 2017
 245 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
 246 Description:    <deprecated: should use /sys/fs/f2fs/<disk>/feature_list/>
 247                 Shows all enabled features in current device.
 248                 Supported features:
 249                 encryption, blkzoned, extra_attr, projquota, inode_checksum,
 250                 flexible_inline_xattr, quota_ino, inode_crtime, lost_found,
 251                 verity, sb_checksum, casefold, readonly, compression, pin_file.
 252
 253 What:           /sys/fs/f2fs/<disk>/feature_list/
 254 Date:           June 2021
 255 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
 256 Description:    Expand /sys/fs/f2fs/<disk>/features to meet sysfs rule.
 257                 Supported on-disk features:
 258                 encryption, block_zoned (aka blkzoned), extra_attr,
 259                 project_quota (aka projquota), inode_checksum,
 260                 flexible_inline_xattr, quota_ino, inode_crtime, lost_found,
 261                 verity, sb_checksum, casefold, readonly, compression.
 262                 Note that, pin_file is moved into /sys/fs/f2fs/features/.
 263
 264 What:           /sys/fs/f2fs/features/
 265 Date:           July 2017
 266 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
 267 Description:    Shows all enabled kernel features.
 268                 Supported features:
 269                 encryption, block_zoned, extra_attr, project_quota,
 270                 inode_checksum, flexible_inline_xattr, quota_ino,
 271                 inode_crtime, lost_found, verity, sb_checksum,
 272                 casefold, readonly, compression, test_dummy_encryption_v2,
 273                 atomic_write, pin_file, encrypted_casefold.
 274
 275 What:           /sys/fs/f2fs/<disk>/inject_rate
 276 Date:           May 2016
 277 Contact:        "Sheng Yong" <shengyong1@huawei.com>
 278 Description:    Controls the injection rate of arbitrary faults.
 279
 280 What:           /sys/fs/f2fs/<disk>/inject_type
 281 Date:           May 2016
 282 Contact:        "Sheng Yong" <shengyong1@huawei.com>
 283 Description:    Controls the injection type of arbitrary faults.
 284
 285 What:           /sys/fs/f2fs/<disk>/dirty_segments
 286 Date:           October 2017
 287 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
 288 Description:    Shows the number of dirty segments.
 289
 290 What:           /sys/fs/f2fs/<disk>/reserved_blocks
 291 Date:           June 2017
 292 Contact:        "Chao Yu" <yuchao0@huawei.com>
 293 Description:    Controls target reserved blocks in system, the threshold
 294                 is soft, it could exceed current available user space.
 295
 296 What:           /sys/fs/f2fs/<disk>/current_reserved_blocks
 297 Date:           October 2017
 298 Contact:        "Yunlong Song" <yunlong.song@huawei.com>
 299 Contact:        "Chao Yu" <yuchao0@huawei.com>
 300 Description:    Shows current reserved blocks in system, it may be temporarily
 301                 smaller than target_reserved_blocks, but will gradually
 302                 increase to target_reserved_blocks when more free blocks are
 303                 freed by user later.
 304
 305 What:           /sys/fs/f2fs/<disk>/gc_urgent
 306 Date:           August 2017
 307 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
 308 Description:    Do background GC aggressively when set. Set to 0 by default.
 309                 gc urgent high(1): does GC forcibly in a period of given
 310                 gc_urgent_sleep_time and ignores I/O idling check. uses greedy
 311                 GC approach and turns SSR mode on.
 312                 gc urgent low(2): lowers the bar of checking I/O idling in
 313                 order to process outstanding discard commands and GC a
 314                 little bit aggressively. always uses cost benefit GC approach,
 315                 and will override age-threshold GC approach if ATGC is enabled
 316                 at the same time.
 317                 gc urgent mid(3): does GC forcibly in a period of given
 318                 gc_urgent_sleep_time and executes a mid level of I/O idling check.
 319                 always uses cost benefit GC approach, and will override
 320                 age-threshold GC approach if ATGC is enabled at the same time.
 321
 322 What:           /sys/fs/f2fs/<disk>/gc_urgent_sleep_time
 323 Date:           August 2017
 324 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
 325 Description:    Controls sleep time of GC urgent mode. Set to 500ms by default.
 326
 327 What:           /sys/fs/f2fs/<disk>/readdir_ra
 328 Date:           November 2017
 329 Contact:        "Sheng Yong" <shengyong1@huawei.com>
 330 Description:    Controls readahead inode block in readdir. Enabled by default.
 331
 332 What:           /sys/fs/f2fs/<disk>/gc_pin_file_thresh
 333 Date:           January 2018
 334 Contact:        Jaegeuk Kim <jaegeuk@kernel.org>
 335 Description:    This indicates how many GC can be failed for the pinned
 336                 file. If it exceeds this, F2FS doesn't guarantee its pinning
 337                 state. 2048 trials is set by default, and 65535 as maximum.
 338
 339 What:           /sys/fs/f2fs/<disk>/extension_list
 340 Date:           February 2018
 341 Contact:        "Chao Yu" <yuchao0@huawei.com>
 342 Description:    Used to control configure extension list:
 343                 - Query: cat /sys/fs/f2fs/<disk>/extension_list
 344                 - Add: echo '[h/c]extension' > /sys/fs/f2fs/<disk>/extension_list
 345                 - Del: echo '[h/c]!extension' > /sys/fs/f2fs/<disk>/extension_list
 346                 - [h] means add/del hot file extension
 347                 - [c] means add/del cold file extension
 348
 349 What:           /sys/fs/f2fs/<disk>/unusable
 350 Date            April 2019
 351 Contact:        "Daniel Rosenberg" <drosen@google.com>
 352 Description:    If checkpoint=disable, it displays the number of blocks that
 353                 are unusable.
 354                 If checkpoint=enable it displays the number of blocks that
 355                 would be unusable if checkpoint=disable were to be set.
 356
 357 What:           /sys/fs/f2fs/<disk>/encoding
 358 Date            July 2019
 359 Contact:        "Daniel Rosenberg" <drosen@google.com>
 360 Description:    Displays name and version of the encoding set for the filesystem.
 361                 If no encoding is set, displays (none)
 362
 363 What:           /sys/fs/f2fs/<disk>/free_segments
 364 Date:           September 2019
 365 Contact:        "Hridya Valsaraju" <hridya@google.com>
 366 Description:    Number of free segments in disk.
 367
 368 What:           /sys/fs/f2fs/<disk>/cp_foreground_calls
 369 Date:           September 2019
 370 Contact:        "Hridya Valsaraju" <hridya@google.com>
 371 Description:    Number of checkpoint operations performed on demand. Available when
 372                 CONFIG_F2FS_STAT_FS=y.
 373
 374 What:           /sys/fs/f2fs/<disk>/cp_background_calls
 375 Date:           September 2019
 376 Contact:        "Hridya Valsaraju" <hridya@google.com>
 377 Description:    Number of checkpoint operations performed in the background to
 378                 free segments. Available when CONFIG_F2FS_STAT_FS=y.
 379
 380 What:           /sys/fs/f2fs/<disk>/gc_foreground_calls
 381 Date:           September 2019
 382 Contact:        "Hridya Valsaraju" <hridya@google.com>
 383 Description:    Number of garbage collection operations performed on demand.
 384                 Available when CONFIG_F2FS_STAT_FS=y.
 385
 386 What:           /sys/fs/f2fs/<disk>/gc_background_calls
 387 Date:           September 2019
 388 Contact:        "Hridya Valsaraju" <hridya@google.com>
 389 Description:    Number of garbage collection operations triggered in background.
 390                 Available when CONFIG_F2FS_STAT_FS=y.
 391
 392 What:           /sys/fs/f2fs/<disk>/moved_blocks_foreground
 393 Date:           September 2019
 394 Contact:        "Hridya Valsaraju" <hridya@google.com>
 395 Description:    Number of blocks moved by garbage collection in foreground.
 396                 Available when CONFIG_F2FS_STAT_FS=y.
 397
 398 What:           /sys/fs/f2fs/<disk>/moved_blocks_background
 399 Date:           September 2019
 400 Contact:        "Hridya Valsaraju" <hridya@google.com>
 401 Description:    Number of blocks moved by garbage collection in background.
 402                 Available when CONFIG_F2FS_STAT_FS=y.
 403
 404 What:           /sys/fs/f2fs/<disk>/avg_vblocks
 405 Date:           September 2019
 406 Contact:        "Hridya Valsaraju" <hridya@google.com>
 407 Description:    Average number of valid blocks.
 408                 Available when CONFIG_F2FS_STAT_FS=y.
 409
 410 What:           /sys/fs/f2fs/<disk>/mounted_time_sec
 411 Date:           February 2020
 412 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
 413 Description:    Show the mounted time in secs of this partition.
 414
 415 What:           /sys/fs/f2fs/<disk>/data_io_flag
 416 Date:           April 2020
 417 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
 418 Description:    Give a way to attach REQ_META|FUA to data writes
 419                 given temperature-based bits. Now the bits indicate:
 420
 421                 +-------------------+-------------------+
 422                 |      REQ_META     |      REQ_FUA      |
 423                 +------+------+-----+------+------+-----+
 424                 |    5 |    4 |   3 |    2 |    1 |   0 |
 425                 +------+------+-----+------+------+-----+
 426                 | Cold | Warm | Hot | Cold | Warm | Hot |
 427                 +------+------+-----+------+------+-----+
 428
 429 What:           /sys/fs/f2fs/<disk>/node_io_flag
 430 Date:           June 2020
 431 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
 432 Description:    Give a way to attach REQ_META|FUA to node writes
 433                 given temperature-based bits. Now the bits indicate:
 434
 435                 +-------------------+-------------------+
 436                 |      REQ_META     |      REQ_FUA      |
 437                 +------+------+-----+------+------+-----+
 438                 |    5 |    4 |   3 |    2 |    1 |   0 |
 439                 +------+------+-----+------+------+-----+
 440                 | Cold | Warm | Hot | Cold | Warm | Hot |
 441                 +------+------+-----+------+------+-----+
 442
 443 What:           /sys/fs/f2fs/<disk>/iostat_period_ms
 444 Date:           April 2020
 445 Contact:        "Daeho Jeong" <daehojeong@google.com>
 446 Description:    Give a way to change iostat_period time. 3secs by default.
 447                 The new iostat trace gives stats gap given the period.
 448 What:           /sys/fs/f2fs/<disk>/max_io_bytes
 449 Date:           December 2020
 450 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
 451 Description:    This gives a control to limit the bio size in f2fs.
 452                 Default is zero, which will follow underlying block layer limit,
 453                 whereas, if it has a certain bytes value, f2fs won't submit a
 454                 bio larger than that size.
 455
 456 What:           /sys/fs/f2fs/<disk>/stat/sb_status
 457 Date:           December 2020
 458 Contact:        "Chao Yu" <yuchao0@huawei.com>
 459 Description:    Show status of f2fs superblock in real time.
 460
 461                 ====== ===================== =================================
 462                 value  sb status macro       description
 463                 0x1    SBI_IS_DIRTY          dirty flag for checkpoint
 464                 0x2    SBI_IS_CLOSE          specify unmounting
 465                 0x4    SBI_NEED_FSCK         need fsck.f2fs to fix
 466                 0x8    SBI_POR_DOING         recovery is doing or not
 467                 0x10   SBI_NEED_SB_WRITE     need to recover superblock
 468                 0x20   SBI_NEED_CP           need to checkpoint
 469                 0x40   SBI_IS_SHUTDOWN       shutdown by ioctl
 470                 0x80   SBI_IS_RECOVERED      recovered orphan/data
 471                 0x100  SBI_CP_DISABLED       CP was disabled last mount
 472                 0x200  SBI_CP_DISABLED_QUICK CP was disabled quickly
 473                 0x400  SBI_QUOTA_NEED_FLUSH  need to flush quota info in CP
 474                 0x800  SBI_QUOTA_SKIP_FLUSH  skip flushing quota in current CP
 475                 0x1000 SBI_QUOTA_NEED_REPAIR quota file may be corrupted
 476                 0x2000 SBI_IS_RESIZEFS       resizefs is in process
 477                 0x4000 SBI_IS_FREEZING       freefs is in process
 478                 ====== ===================== =================================
 479
 480 What:           /sys/fs/f2fs/<disk>/stat/cp_status
 481 Date:           September 2022
 482 Contact:        "Chao Yu" <chao.yu@oppo.com>
 483 Description:    Show status of f2fs checkpoint in real time.
 484
 485                 =============================== ==============================
 486                 cp flag                         value
 487                 CP_UMOUNT_FLAG                  0x00000001
 488                 CP_ORPHAN_PRESENT_FLAG          0x00000002
 489                 CP_COMPACT_SUM_FLAG             0x00000004
 490                 CP_ERROR_FLAG                   0x00000008
 491                 CP_FSCK_FLAG                    0x00000010
 492                 CP_FASTBOOT_FLAG                0x00000020
 493                 CP_CRC_RECOVERY_FLAG            0x00000040
 494                 CP_NAT_BITS_FLAG                0x00000080
 495                 CP_TRIMMED_FLAG                 0x00000100
 496                 CP_NOCRC_RECOVERY_FLAG          0x00000200
 497                 CP_LARGE_NAT_BITMAP_FLAG        0x00000400
 498                 CP_QUOTA_NEED_FSCK_FLAG         0x00000800
 499                 CP_DISABLED_FLAG                0x00001000
 500                 CP_DISABLED_QUICK_FLAG          0x00002000
 501                 CP_RESIZEFS_FLAG                0x00004000
 502                 =============================== ==============================
 503
 504 What:           /sys/fs/f2fs/<disk>/stat/issued_discard
 505 Date:           December 2023
 506 Contact:        "Zhiguo Niu" <zhiguo.niu@unisoc.com>
 507 Description:    Shows the number of issued discard.
 508
 509 What:           /sys/fs/f2fs/<disk>/stat/queued_discard
 510 Date:           December 2023
 511 Contact:        "Zhiguo Niu" <zhiguo.niu@unisoc.com>
 512 Description:    Shows the number of queued discard.
 513
 514 What:           /sys/fs/f2fs/<disk>/stat/undiscard_blks
 515 Date:           December 2023
 516 Contact:        "Zhiguo Niu" <zhiguo.niu@unisoc.com>
 517 Description:    Shows the total number of undiscard blocks.
 518
 519 What:           /sys/fs/f2fs/<disk>/ckpt_thread_ioprio
 520 Date:           January 2021
 521 Contact:        "Daeho Jeong" <daehojeong@google.com>
 522 Description:    Give a way to change checkpoint merge daemon's io priority.
 523                 Its default value is "be,3", which means "BE" I/O class and
 524                 I/O priority "3". We can select the class between "rt" and "be",
 525                 and set the I/O priority within valid range of it. "," delimiter
 526                 is necessary in between I/O class and priority number.
 527
 528 What:           /sys/fs/f2fs/<disk>/ovp_segments
 529 Date:           March 2021
 530 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
 531 Description:    Shows the number of overprovision segments.
 532
 533 What:           /sys/fs/f2fs/<disk>/compr_written_block
 534 Date:           March 2021
 535 Contact:        "Daeho Jeong" <daehojeong@google.com>
 536 Description:    Show the block count written after compression since mount. Note
 537                 that when the compressed blocks are deleted, this count doesn't
 538                 decrease. If you write "0" here, you can initialize
 539                 compr_written_block and compr_saved_block to "0".
 540
 541 What:           /sys/fs/f2fs/<disk>/compr_saved_block
 542 Date:           March 2021
 543 Contact:        "Daeho Jeong" <daehojeong@google.com>
 544 Description:    Show the saved block count with compression since mount. Note
 545                 that when the compressed blocks are deleted, this count doesn't
 546                 decrease. If you write "0" here, you can initialize
 547                 compr_written_block and compr_saved_block to "0".
 548
 549 What:           /sys/fs/f2fs/<disk>/compr_new_inode
 550 Date:           March 2021
 551 Contact:        "Daeho Jeong" <daehojeong@google.com>
 552 Description:    Show the count of inode newly enabled for compression since mount.
 553                 Note that when the compression is disabled for the files, this count
 554                 doesn't decrease. If you write "0" here, you can initialize
 555                 compr_new_inode to "0".
 556
 557 What:           /sys/fs/f2fs/<disk>/atgc_candidate_ratio
 558 Date:           May 2021
 559 Contact:        "Chao Yu" <yuchao0@huawei.com>
 560 Description:    When ATGC is on, it controls candidate ratio in order to limit total
 561                 number of potential victim in all candidates, the value should be in
 562                 range of [0, 100], by default it was initialized as 20(%).
 563
 564 What:           /sys/fs/f2fs/<disk>/atgc_candidate_count
 565 Date:           May 2021
 566 Contact:        "Chao Yu" <yuchao0@huawei.com>
 567 Description:    When ATGC is on, it controls candidate count in order to limit total
 568                 number of potential victim in all candidates, by default it was
 569                 initialized as 10 (sections).
 570
 571 What:           /sys/fs/f2fs/<disk>/atgc_age_weight
 572 Date:           May 2021
 573 Contact:        "Chao Yu" <yuchao0@huawei.com>
 574 Description:    When ATGC is on, it controls age weight to balance weight proportion
 575                 in between aging and valid blocks, the value should be in range of
 576                 [0, 100], by default it was initialized as 60(%).
 577
 578 What:           /sys/fs/f2fs/<disk>/atgc_age_threshold
 579 Date:           May 2021
 580 Contact:        "Chao Yu" <yuchao0@huawei.com>
 581 Description:    When ATGC is on, it controls age threshold to bypass GCing young
 582                 candidates whose age is not beyond the threshold, by default it was
 583                 initialized as 604800 seconds (equals to 7 days).
 584
 585 What:           /sys/fs/f2fs/<disk>/atgc_enabled
 586 Date:           Feb 2024
 587 Contact:        "Jinbao Liu" <liujinbao1@xiaomi.com>
 588 Description:    It represents whether ATGC is on or off. The value is 1 which
 589                indicates that ATGC is on, and 0 indicates that it is off.
 590
 591 What:           /sys/fs/f2fs/<disk>/gc_reclaimed_segments
 592 Date:           July 2021
 593 Contact:        "Daeho Jeong" <daehojeong@google.com>
 594 Description:    Show how many segments have been reclaimed by GC during a specific
 595                 GC mode (0: GC normal, 1: GC idle CB, 2: GC idle greedy,
 596                 3: GC idle AT, 4: GC urgent high, 5: GC urgent low 6: GC urgent mid)
 597                 You can re-initialize this value to "0".
 598
 599 What:           /sys/fs/f2fs/<disk>/gc_segment_mode
 600 Date:           July 2021
 601 Contact:        "Daeho Jeong" <daehojeong@google.com>
 602 Description:    You can control for which gc mode the "gc_reclaimed_segments" node shows.
 603                 Refer to the description of the modes in "gc_reclaimed_segments".
 604
 605 What:           /sys/fs/f2fs/<disk>/seq_file_ra_mul
 606 Date:           July 2021
 607 Contact:        "Daeho Jeong" <daehojeong@google.com>
 608 Description:    You can control the multiplier value of bdi device readahead window size
 609                 between 2 (default) and 256 for POSIX_FADV_SEQUENTIAL advise option.
 610
 611 What:           /sys/fs/f2fs/<disk>/max_fragment_chunk
 612 Date:           August 2021
 613 Contact:        "Daeho Jeong" <daehojeong@google.com>
 614 Description:    With "mode=fragment:block" mount options, we can scatter block allocation.
 615                 f2fs will allocate 1..<max_fragment_chunk> blocks in a chunk and make a hole
 616                 in the length of 1..<max_fragment_hole> by turns. This value can be set
 617                 between 1..512 and the default value is 4.
 618
 619 What:           /sys/fs/f2fs/<disk>/max_fragment_hole
 620 Date:           August 2021
 621 Contact:        "Daeho Jeong" <daehojeong@google.com>
 622 Description:    With "mode=fragment:block" mount options, we can scatter block allocation.
 623                 f2fs will allocate 1..<max_fragment_chunk> blocks in a chunk and make a hole
 624                 in the length of 1..<max_fragment_hole> by turns. This value can be set
 625                 between 1..512 and the default value is 4.
 626
 627 What:           /sys/fs/f2fs/<disk>/gc_remaining_trials
 628 Date:           October 2022
 629 Contact:        "Yangtao Li" <frank.li@vivo.com>
 630 Description:    You can set the trial count limit for GC urgent and idle mode with this value.
 631                 If GC thread gets to the limit, the mode will turn back to GC normal mode.
 632                 By default, the value is zero, which means there is no limit like before.
 633
 634 What:           /sys/fs/f2fs/<disk>/max_roll_forward_node_blocks
 635 Date:           January 2022
 636 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
 637 Description:    Controls max # of node block writes to be used for roll forward
 638                 recovery. This can limit the roll forward recovery time.
 639
 640 What:           /sys/fs/f2fs/<disk>/unusable_blocks_per_sec
 641 Date:           June 2022
 642 Contact:        "Jaegeuk Kim" <jaegeuk@kernel.org>
 643 Description:    Shows the number of unusable blocks in a section which was defined by
 644                 the zone capacity reported by underlying zoned device.
 645
 646 What:           /sys/fs/f2fs/<disk>/current_atomic_write
 647 Date:           July 2022
 648 Contact:        "Daeho Jeong" <daehojeong@google.com>
 649 Description:    Show the total current atomic write block count, which is not committed yet.
 650                 This is a read-only entry.
 651
 652 What:           /sys/fs/f2fs/<disk>/peak_atomic_write
 653 Date:           July 2022
 654 Contact:        "Daeho Jeong" <daehojeong@google.com>
 655 Description:    Show the peak value of total current atomic write block count after boot.
 656                 If you write "0" here, you can initialize to "0".
 657
 658 What:           /sys/fs/f2fs/<disk>/committed_atomic_block
 659 Date:           July 2022
 660 Contact:        "Daeho Jeong" <daehojeong@google.com>
 661 Description:    Show the accumulated total committed atomic write block count after boot.
 662                 If you write "0" here, you can initialize to "0".
 663
 664 What:           /sys/fs/f2fs/<disk>/revoked_atomic_block
 665 Date:           July 2022
 666 Contact:        "Daeho Jeong" <daehojeong@google.com>
 667 Description:    Show the accumulated total revoked atomic write block count after boot.
 668                 If you write "0" here, you can initialize to "0".
 669
 670 What:           /sys/fs/f2fs/<disk>/gc_mode
 671 Date:           October 2022
 672 Contact:        "Yangtao Li" <frank.li@vivo.com>
 673 Description:    Show the current gc_mode as a string.
 674                 This is a read-only entry.
 675
 676 What:           /sys/fs/f2fs/<disk>/discard_urgent_util
 677 Date:           November 2022
 678 Contact:        "Yangtao Li" <frank.li@vivo.com>
 679 Description:    When space utilization exceeds this, do background DISCARD aggressively.
 680                 Does DISCARD forcibly in a period of given min_discard_issue_time when the number
 681                 of discards is not 0 and set discard granularity to 1.
 682                 Default: 80
 683
 684 What:           /sys/fs/f2fs/<disk>/hot_data_age_threshold
 685 Date:           November 2022
 686 Contact:        "Ping Xiong" <xiongping1@xiaomi.com>
 687 Description:    When DATA SEPARATION is on, it controls the age threshold to indicate
 688                 the data blocks as hot. By default it was initialized as 262144 blocks
 689                 (equals to 1GB).
 690
 691 What:           /sys/fs/f2fs/<disk>/warm_data_age_threshold
 692 Date:           November 2022
 693 Contact:        "Ping Xiong" <xiongping1@xiaomi.com>
 694 Description:    When DATA SEPARATION is on, it controls the age threshold to indicate
 695                 the data blocks as warm. By default it was initialized as 2621440 blocks
 696                 (equals to 10GB).
 697
 698 What:           /sys/fs/f2fs/<disk>/fault_rate
 699 Date:           May 2016
 700 Contact:        "Sheng Yong" <shengyong@oppo.com>
 701 Contact:        "Chao Yu" <chao@kernel.org>
 702 Description:    Enable fault injection in all supported types with
 703                 specified injection rate.
 704
 705 What:           /sys/fs/f2fs/<disk>/fault_type
 706 Date:           May 2016
 707 Contact:        "Sheng Yong" <shengyong@oppo.com>
 708 Contact:        "Chao Yu" <chao@kernel.org>
 709 Description:    Support configuring fault injection type, should be
 710                 enabled with fault_injection option, fault type value
 711                 is shown below, it supports single or combined type.
 712
 713                 ===========================      ===========
 714                 Type_Name                        Type_Value
 715                 ===========================      ===========
 716                 FAULT_KMALLOC                    0x000000001
 717                 FAULT_KVMALLOC                   0x000000002
 718                 FAULT_PAGE_ALLOC                 0x000000004
 719                 FAULT_PAGE_GET                   0x000000008
 720                 FAULT_ALLOC_BIO                  0x000000010 (obsolete)
 721                 FAULT_ALLOC_NID                  0x000000020
 722                 FAULT_ORPHAN                     0x000000040
 723                 FAULT_BLOCK                      0x000000080
 724                 FAULT_DIR_DEPTH                  0x000000100
 725                 FAULT_EVICT_INODE                0x000000200
 726                 FAULT_TRUNCATE                   0x000000400
 727                 FAULT_READ_IO                    0x000000800
 728                 FAULT_CHECKPOINT                 0x000001000
 729                 FAULT_DISCARD                    0x000002000
 730                 FAULT_WRITE_IO                   0x000004000
 731                 FAULT_SLAB_ALLOC                 0x000008000
 732                 FAULT_DQUOT_INIT                 0x000010000
 733                 FAULT_LOCK_OP                    0x000020000
 734                 FAULT_BLKADDR_VALIDITY           0x000040000
 735                 FAULT_BLKADDR_CONSISTENCE        0x000080000
 736                 FAULT_NO_SEGMENT                 0x000100000
 737                 ===========================      ===========
 738
 739 What:           /sys/fs/f2fs/<disk>/discard_io_aware_gran
 740 Date:           January 2023
 741 Contact:        "Yangtao Li" <frank.li@vivo.com>
 742 Description:    Controls background discard granularity of inner discard thread
 743                 when is not in idle. Inner thread will not issue discards with size that
 744                 is smaller than granularity. The unit size is one block(4KB), now only
 745                 support configuring in range of [0, 512].
 746                 Default: 512
 747
 748 What:           /sys/fs/f2fs/<disk>/last_age_weight
 749 Date:           January 2023
 750 Contact:        "Ping Xiong" <xiongping1@xiaomi.com>
 751 Description:    When DATA SEPARATION is on, it controls the weight of last data block age.
 752
 753 What:           /sys/fs/f2fs/<disk>/compress_watermark
 754 Date:           February 2023
 755 Contact:        "Yangtao Li" <frank.li@vivo.com>
 756 Description:    When compress cache is on, it controls free memory watermark
 757                 in order to limit caching compress page. If free memory is lower
 758                 than watermark, then deny caching compress page. The value should be in
 759                 range of (0, 100], by default it was initialized as 20(%).
 760
 761 What:           /sys/fs/f2fs/<disk>/compress_percent
 762 Date:           February 2023
 763 Contact:        "Yangtao Li" <frank.li@vivo.com>
 764 Description:    When compress cache is on, it controls cached page
 765                 percent(compress pages / free_ram) in order to limit caching compress page.
 766                 If cached page percent exceed threshold, then deny caching compress page.
 767                 The value should be in range of (0, 100], by default it was initialized
 768                 as 20(%).
 769
 770 What:           /sys/fs/f2fs/<disk>/discard_io_aware
 771 Date:           November 2023
 772 Contact:        "Chao Yu" <chao@kernel.org>
 773 Description:    It controls to enable/disable IO aware feature for background discard.
 774                 By default, the value is 1 which indicates IO aware is on.
 775
 776 What:           /sys/fs/f2fs/<disk>/blkzone_alloc_policy
 777 Date:           July 2024
 778 Contact:        "Yuanhong Liao" <liaoyuanhong@vivo.com>
 779 Description:    The zone UFS we are currently using consists of two parts:
 780                 conventional zones and sequential zones. It can be used to control which part
 781                 to prioritize for writes, with a default value of 0.
 782
 783                 ========================  =========================================
 784                 value                                     description
 785                 blkzone_alloc_policy = 0  Prioritize writing to sequential zones
 786                 blkzone_alloc_policy = 1  Only allow writing to sequential zones
 787                 blkzone_alloc_policy = 2  Prioritize writing to conventional zones
 788                 ========================  =========================================
 789
 790 What:           /sys/fs/f2fs/<disk>/migration_window_granularity
 791 Date:           September 2024
 792 Contact:        "Daeho Jeong" <daehojeong@google.com>
 793 Description:    Controls migration window granularity of garbage collection on large
 794                 section. it can control the scanning window granularity for GC migration
 795                 in a unit of segment, while migration_granularity controls the number
 796                 of segments which can be migrated at the same turn.
 797
 798 What:           /sys/fs/f2fs/<disk>/reserved_segments
 799 Date:           September 2024
 800 Contact:        "Daeho Jeong" <daehojeong@google.com>
 801 Description:    In order to fine tune GC behavior, we can control the number of
 802                 reserved segments.
 803
 804 What:           /sys/fs/f2fs/<disk>/gc_no_zoned_gc_percent
 805 Date:           September 2024
 806 Contact:        "Daeho Jeong" <daehojeong@google.com>
 807 Description:    If the percentage of free sections over total sections is above this
 808                 number, F2FS do not garbage collection for zoned devices through the
 809                 background GC thread. the default number is "60".
 810
 811 What:           /sys/fs/f2fs/<disk>/gc_boost_zoned_gc_percent
 812 Date:           September 2024
 813 Contact:        "Daeho Jeong" <daehojeong@google.com>
 814 Description:    If the percentage of free sections over total sections is under this
 815                 number, F2FS boosts garbage collection for zoned devices through the
 816                 background GC thread. the default number is "25".
 817
 818 What:           /sys/fs/f2fs/<disk>/gc_valid_thresh_ratio
 819 Date:           September 2024
 820 Contact:        "Daeho Jeong" <daehojeong@google.com>
 821 Description:    It controls the valid block ratio threshold not to trigger excessive GC
 822                 for zoned deivces. The initial value of it is 95(%). F2FS will stop the
 823                 background GC thread from intiating GC for sections having valid blocks
 824                 exceeding the ratio.
 825
 826 What:           /sys/fs/f2fs/<disk>/max_read_extent_count
 827 Date:           November 2024
 828 Contact:        "Chao Yu" <chao@kernel.org>
 829 Description:    It controls max read extent count for per-inode, the value of threshold
 830                 is 10240 by default.