1.. SPDX-License-Identifier: GPL-2.0 2 3================================= 4Flash-Friendly File System (F2FS) 5================================= 6 7Overview 8======== 9 10NAND flash memory-based storage devices, such as SSD, eMMC, and SD cards, have 11been equipped on a variety systems ranging from mobile to server systems. Since 12they are known to have different characteristics from the conventional rotating 13disks, a file system, an upper layer to the storage device, should adapt to the 14changes from the sketch in the design level. 15 16F2FS is a file system exploiting NAND flash memory-based storage devices, which 17is based on Log-structured File System (LFS). The design has been focused on 18addressing the fundamental issues in LFS, which are snowball effect of wandering 19tree and high cleaning overhead. 20 21Since a NAND flash memory-based storage device shows different characteristic 22according to its internal geometry or flash memory management scheme, namely FTL, 23F2FS and its tools support various parameters not only for configuring on-disk 24layout, but also for selecting allocation and cleaning algorithms. 25 26The following git tree provides the file system formatting tool (mkfs.f2fs), 27a consistency checking tool (fsck.f2fs), and a debugging tool (dump.f2fs). 28 29- git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs-tools.git 30 31For sending patches, please use the following mailing list: 32 33- linux-f2fs-devel@lists.sourceforge.net 34 35For reporting bugs, please use the following f2fs bug tracker link: 36 37- https://bugzilla.kernel.org/enter_bug.cgi?product=File%20System&component=f2fs 38 39Background and Design issues 40============================ 41 42Log-structured File System (LFS) 43-------------------------------- 44"A log-structured file system writes all modifications to disk sequentially in 45a log-like structure, thereby speeding up both file writing and crash recovery. 46The log is the only structure on disk; it contains indexing information so that 47files can be read back from the log efficiently. In order to maintain large free 48areas on disk for fast writing, we divide the log into segments and use a 49segment cleaner to compress the live information from heavily fragmented 50segments." from Rosenblum, M. and Ousterhout, J. K., 1992, "The design and 51implementation of a log-structured file system", ACM Trans. Computer Systems 5210, 1, 26–52. 53 54Wandering Tree Problem 55---------------------- 56In LFS, when a file data is updated and written to the end of log, its direct 57pointer block is updated due to the changed location. Then the indirect pointer 58block is also updated due to the direct pointer block update. In this manner, 59the upper index structures such as inode, inode map, and checkpoint block are 60also updated recursively. This problem is called as wandering tree problem [1], 61and in order to enhance the performance, it should eliminate or relax the update 62propagation as much as possible. 63 64[1] Bityutskiy, A. 2005. JFFS3 design issues. http://www.linux-mtd.infradead.org/ 65 66Cleaning Overhead 67----------------- 68Since LFS is based on out-of-place writes, it produces so many obsolete blocks 69scattered across the whole storage. In order to serve new empty log space, it 70needs to reclaim these obsolete blocks seamlessly to users. This job is called 71as a cleaning process. 72 73The process consists of three operations as follows. 74 751. A victim segment is selected through referencing segment usage table. 762. It loads parent index structures of all the data in the victim identified by 77 segment summary blocks. 783. It checks the cross-reference between the data and its parent index structure. 794. It moves valid data selectively. 80 81This cleaning job may cause unexpected long delays, so the most important goal 82is to hide the latencies to users. And also definitely, it should reduce the 83amount of valid data to be moved, and move them quickly as well. 84 85Key Features 86============ 87 88Flash Awareness 89--------------- 90- Enlarge the random write area for better performance, but provide the high 91 spatial locality 92- Align FS data structures to the operational units in FTL as best efforts 93 94Wandering Tree Problem 95---------------------- 96- Use a term, “node”, that represents inodes as well as various pointer blocks 97- Introduce Node Address Table (NAT) containing the locations of all the “node” 98 blocks; this will cut off the update propagation. 99 100Cleaning Overhead 101----------------- 102- Support a background cleaning process 103- Support greedy and cost-benefit algorithms for victim selection policies 104- Support multi-head logs for static/dynamic hot and cold data separation 105- Introduce adaptive logging for efficient block allocation 106 107Mount Options 108============= 109 110 111======================== ============================================================ 112background_gc=%s Turn on/off cleaning operations, namely garbage 113 collection, triggered in background when I/O subsystem is 114 idle. If background_gc=on, it will turn on the garbage 115 collection and if background_gc=off, garbage collection 116 will be turned off. If background_gc=sync, it will turn 117 on synchronous garbage collection running in background. 118 Default value for this option is on. So garbage 119 collection is on by default. 120gc_merge When background_gc is on, this option can be enabled to 121 let background GC thread to handle foreground GC requests, 122 it can eliminate the sluggish issue caused by slow foreground 123 GC operation when GC is triggered from a process with limited 124 I/O and CPU resources. 125nogc_merge Disable GC merge feature. 126disable_roll_forward Disable the roll-forward recovery routine 127norecovery Disable the roll-forward recovery routine, mounted read- 128 only (i.e., -o ro,disable_roll_forward) 129discard/nodiscard Enable/disable real-time discard in f2fs, if discard is 130 enabled, f2fs will issue discard/TRIM commands when a 131 segment is cleaned. 132heap/no_heap Deprecated. 133nouser_xattr Disable Extended User Attributes. Note: xattr is enabled 134 by default if CONFIG_F2FS_FS_XATTR is selected. 135noacl Disable POSIX Access Control List. Note: acl is enabled 136 by default if CONFIG_F2FS_FS_POSIX_ACL is selected. 137active_logs=%u Support configuring the number of active logs. In the 138 current design, f2fs supports only 2, 4, and 6 logs. 139 Default number is 6. 140disable_ext_identify Disable the extension list configured by mkfs, so f2fs 141 is not aware of cold files such as media files. 142inline_xattr Enable the inline xattrs feature. 143noinline_xattr Disable the inline xattrs feature. 144inline_xattr_size=%u Support configuring inline xattr size, it depends on 145 flexible inline xattr feature. 146inline_data Enable the inline data feature: Newly created small (<~3.4k) 147 files can be written into inode block. 148inline_dentry Enable the inline dir feature: data in newly created 149 directory entries can be written into inode block. The 150 space of inode block which is used to store inline 151 dentries is limited to ~3.4k. 152noinline_dentry Disable the inline dentry feature. 153flush_merge Merge concurrent cache_flush commands as much as possible 154 to eliminate redundant command issues. If the underlying 155 device handles the cache_flush command relatively slowly, 156 recommend to enable this option. 157nobarrier This option can be used if underlying storage guarantees 158 its cached data should be written to the novolatile area. 159 If this option is set, no cache_flush commands are issued 160 but f2fs still guarantees the write ordering of all the 161 data writes. 162barrier If this option is set, cache_flush commands are allowed to be 163 issued. 164fastboot This option is used when a system wants to reduce mount 165 time as much as possible, even though normal performance 166 can be sacrificed. 167extent_cache Enable an extent cache based on rb-tree, it can cache 168 as many as extent which map between contiguous logical 169 address and physical address per inode, resulting in 170 increasing the cache hit ratio. Set by default. 171noextent_cache Disable an extent cache based on rb-tree explicitly, see 172 the above extent_cache mount option. 173noinline_data Disable the inline data feature, inline data feature is 174 enabled by default. 175data_flush Enable data flushing before checkpoint in order to 176 persist data of regular and symlink. 177reserve_root=%d Support configuring reserved space which is used for 178 allocation from a privileged user with specified uid or 179 gid, unit: 4KB, the default limit is 12.5% of user blocks. 180reserve_node=%d Support configuring reserved nodes which are used for 181 allocation from a privileged user with specified uid or 182 gid, the default limit is 12.5% of all nodes. 183resuid=%d The user ID which may use the reserved blocks and nodes. 184resgid=%d The group ID which may use the reserved blocks and nodes. 185fault_injection=%d Enable fault injection in all supported types with 186 specified injection rate. 187fault_type=%d Support configuring fault injection type, should be 188 enabled with fault_injection option, fault type value 189 is shown below, it supports single or combined type. 190 191 .. code-block:: none 192 193 =========================== ========== 194 Type_Name Type_Value 195 =========================== ========== 196 FAULT_KMALLOC 0x00000001 197 FAULT_KVMALLOC 0x00000002 198 FAULT_PAGE_ALLOC 0x00000004 199 FAULT_PAGE_GET 0x00000008 200 FAULT_ALLOC_BIO 0x00000010 (obsolete) 201 FAULT_ALLOC_NID 0x00000020 202 FAULT_ORPHAN 0x00000040 203 FAULT_BLOCK 0x00000080 204 FAULT_DIR_DEPTH 0x00000100 205 FAULT_EVICT_INODE 0x00000200 206 FAULT_TRUNCATE 0x00000400 207 FAULT_READ_IO 0x00000800 208 FAULT_CHECKPOINT 0x00001000 209 FAULT_DISCARD 0x00002000 210 FAULT_WRITE_IO 0x00004000 211 FAULT_SLAB_ALLOC 0x00008000 212 FAULT_DQUOT_INIT 0x00010000 213 FAULT_LOCK_OP 0x00020000 214 FAULT_BLKADDR_VALIDITY 0x00040000 215 FAULT_BLKADDR_CONSISTENCE 0x00080000 216 FAULT_NO_SEGMENT 0x00100000 217 FAULT_INCONSISTENT_FOOTER 0x00200000 218 FAULT_TIMEOUT 0x00400000 (1000ms) 219 FAULT_VMALLOC 0x00800000 220 =========================== ========== 221mode=%s Control block allocation mode which supports "adaptive" 222 and "lfs". In "lfs" mode, there should be no random 223 writes towards main area. 224 "fragment:segment" and "fragment:block" are newly added here. 225 These are developer options for experiments to simulate filesystem 226 fragmentation/after-GC situation itself. The developers use these 227 modes to understand filesystem fragmentation/after-GC condition well, 228 and eventually get some insights to handle them better. 229 In "fragment:segment", f2fs allocates a new segment in random 230 position. With this, we can simulate the after-GC condition. 231 In "fragment:block", we can scatter block allocation with 232 "max_fragment_chunk" and "max_fragment_hole" sysfs nodes. 233 We added some randomness to both chunk and hole size to make 234 it close to realistic IO pattern. So, in this mode, f2fs will allocate 235 1..<max_fragment_chunk> blocks in a chunk and make a hole in the 236 length of 1..<max_fragment_hole> by turns. With this, the newly 237 allocated blocks will be scattered throughout the whole partition. 238 Note that "fragment:block" implicitly enables "fragment:segment" 239 option for more randomness. 240 Please, use these options for your experiments and we strongly 241 recommend to re-format the filesystem after using these options. 242usrquota Enable plain user disk quota accounting. 243grpquota Enable plain group disk quota accounting. 244prjquota Enable plain project quota accounting. 245usrjquota=<file> Appoint specified file and type during mount, so that quota 246grpjquota=<file> information can be properly updated during recovery flow, 247prjjquota=<file> <quota file>: must be in root directory; 248jqfmt=<quota type> <quota type>: [vfsold,vfsv0,vfsv1]. 249usrjquota= Turn off user journalled quota. 250grpjquota= Turn off group journalled quota. 251prjjquota= Turn off project journalled quota. 252quota Enable plain user disk quota accounting. 253noquota Disable all plain disk quota option. 254alloc_mode=%s Adjust block allocation policy, which supports "reuse" 255 and "default". 256fsync_mode=%s Control the policy of fsync. Currently supports "posix", 257 "strict", and "nobarrier". In "posix" mode, which is 258 default, fsync will follow POSIX semantics and does a 259 light operation to improve the filesystem performance. 260 In "strict" mode, fsync will be heavy and behaves in line 261 with xfs, ext4 and btrfs, where xfstest generic/342 will 262 pass, but the performance will regress. "nobarrier" is 263 based on "posix", but doesn't issue flush command for 264 non-atomic files likewise "nobarrier" mount option. 265test_dummy_encryption 266test_dummy_encryption=%s 267 Enable dummy encryption, which provides a fake fscrypt 268 context. The fake fscrypt context is used by xfstests. 269 The argument may be either "v1" or "v2", in order to 270 select the corresponding fscrypt policy version. 271checkpoint=%s[:%u[%]] Set to "disable" to turn off checkpointing. Set to "enable" 272 to re-enable checkpointing. Is enabled by default. While 273 disabled, any unmounting or unexpected shutdowns will cause 274 the filesystem contents to appear as they did when the 275 filesystem was mounted with that option. 276 While mounting with checkpoint=disable, the filesystem must 277 run garbage collection to ensure that all available space can 278 be used. If this takes too much time, the mount may return 279 EAGAIN. You may optionally add a value to indicate how much 280 of the disk you would be willing to temporarily give up to 281 avoid additional garbage collection. This can be given as a 282 number of blocks, or as a percent. For instance, mounting 283 with checkpoint=disable:100% would always succeed, but it may 284 hide up to all remaining free space. The actual space that 285 would be unusable can be viewed at /sys/fs/f2fs/<disk>/unusable 286 This space is reclaimed once checkpoint=enable. 287checkpoint_merge When checkpoint is enabled, this can be used to create a kernel 288 daemon and make it to merge concurrent checkpoint requests as 289 much as possible to eliminate redundant checkpoint issues. Plus, 290 we can eliminate the sluggish issue caused by slow checkpoint 291 operation when the checkpoint is done in a process context in 292 a cgroup having low i/o budget and cpu shares. To make this 293 do better, we set the default i/o priority of the kernel daemon 294 to "3", to give one higher priority than other kernel threads. 295 This is the same way to give a I/O priority to the jbd2 296 journaling thread of ext4 filesystem. 297nocheckpoint_merge Disable checkpoint merge feature. 298compress_algorithm=%s Control compress algorithm, currently f2fs supports "lzo", 299 "lz4", "zstd" and "lzo-rle" algorithm. 300compress_algorithm=%s:%d Control compress algorithm and its compress level, now, only 301 "lz4" and "zstd" support compress level config:: 302 303 ========= =========== 304 algorithm level range 305 ========= =========== 306 lz4 3 - 16 307 zstd 1 - 22 308 ========= =========== 309 310compress_log_size=%u Support configuring compress cluster size. The size will 311 be 4KB * (1 << %u). The default and minimum sizes are 16KB. 312compress_extension=%s Support adding specified extension, so that f2fs can enable 313 compression on those corresponding files, e.g. if all files 314 with '.ext' has high compression rate, we can set the '.ext' 315 on compression extension list and enable compression on 316 these file by default rather than to enable it via ioctl. 317 For other files, we can still enable compression via ioctl. 318 Note that, there is one reserved special extension '*', it 319 can be set to enable compression for all files. 320nocompress_extension=%s Support adding specified extension, so that f2fs can disable 321 compression on those corresponding files, just contrary to compression extension. 322 If you know exactly which files cannot be compressed, you can use this. 323 The same extension name can't appear in both compress and nocompress 324 extension at the same time. 325 If the compress extension specifies all files, the types specified by the 326 nocompress extension will be treated as special cases and will not be compressed. 327 Don't allow use '*' to specifie all file in nocompress extension. 328 After add nocompress_extension, the priority should be: 329 dir_flag < comp_extention,nocompress_extension < comp_file_flag,no_comp_file_flag. 330 See more in compression sections. 331 332compress_chksum Support verifying chksum of raw data in compressed cluster. 333compress_mode=%s Control file compression mode. This supports "fs" and "user" 334 modes. In "fs" mode (default), f2fs does automatic compression 335 on the compression enabled files. In "user" mode, f2fs disables 336 the automaic compression and gives the user discretion of 337 choosing the target file and the timing. The user can do manual 338 compression/decompression on the compression enabled files using 339 ioctls. 340compress_cache Support to use address space of a filesystem managed inode to 341 cache compressed block, in order to improve cache hit ratio of 342 random read. 343inlinecrypt When possible, encrypt/decrypt the contents of encrypted 344 files using the blk-crypto framework rather than 345 filesystem-layer encryption. This allows the use of 346 inline encryption hardware. The on-disk format is 347 unaffected. For more details, see 348 Documentation/block/inline-encryption.rst. 349atgc Enable age-threshold garbage collection, it provides high 350 effectiveness and efficiency on background GC. 351discard_unit=%s Control discard unit, the argument can be "block", "segment" 352 and "section", issued discard command's offset/size will be 353 aligned to the unit, by default, "discard_unit=block" is set, 354 so that small discard functionality is enabled. 355 For blkzoned device, "discard_unit=section" will be set by 356 default, it is helpful for large sized SMR or ZNS devices to 357 reduce memory cost by getting rid of fs metadata supports small 358 discard. 359memory=%s Control memory mode. This supports "normal" and "low" modes. 360 "low" mode is introduced to support low memory devices. 361 Because of the nature of low memory devices, in this mode, f2fs 362 will try to save memory sometimes by sacrificing performance. 363 "normal" mode is the default mode and same as before. 364age_extent_cache Enable an age extent cache based on rb-tree. It records 365 data block update frequency of the extent per inode, in 366 order to provide better temperature hints for data block 367 allocation. 368errors=%s Specify f2fs behavior on critical errors. This supports modes: 369 "panic", "continue" and "remount-ro", respectively, trigger 370 panic immediately, continue without doing anything, and remount 371 the partition in read-only mode. By default it uses "continue" 372 mode. 373 374 .. code-block:: none 375 376 ====================== =============== =============== ======== 377 mode continue remount-ro panic 378 ====================== =============== =============== ======== 379 access ops normal normal N/A 380 syscall errors -EIO -EROFS N/A 381 mount option rw ro N/A 382 pending dir write keep keep N/A 383 pending non-dir write drop keep N/A 384 pending node write drop keep N/A 385 pending meta write keep keep N/A 386 ====================== =============== =============== ======== 387nat_bits Enable nat_bits feature to enhance full/empty nat blocks access, 388 by default it's disabled. 389lookup_mode=%s Control the directory lookup behavior for casefolded 390 directories. This option has no effect on directories 391 that do not have the casefold feature enabled. 392 393 .. code-block:: none 394 395 ================== ======================================== 396 Value Description 397 ================== ======================================== 398 perf (Default) Enforces a hash-only lookup. 399 The linear search fallback is always 400 disabled, ignoring the on-disk flag. 401 compat Enables the linear search fallback for 402 compatibility with directory entries 403 created by older kernel that used a 404 different case-folding algorithm. 405 This mode ignores the on-disk flag. 406 auto F2FS determines the mode based on the 407 on-disk `SB_ENC_NO_COMPAT_FALLBACK_FL` 408 flag. 409 ================== ======================================== 410======================== ============================================================ 411 412Debugfs Entries 413=============== 414 415/sys/kernel/debug/f2fs/ contains information about all the partitions mounted as 416f2fs. Each file shows the whole f2fs information. 417 418/sys/kernel/debug/f2fs/status includes: 419 420 - major file system information managed by f2fs currently 421 - average SIT information about whole segments 422 - current memory footprint consumed by f2fs. 423 424Sysfs Entries 425============= 426 427Information about mounted f2fs file systems can be found in 428/sys/fs/f2fs. Each mounted filesystem will have a directory in 429/sys/fs/f2fs based on its device name (i.e., /sys/fs/f2fs/sda). 430The files in each per-device directory are shown in table below. 431 432Files in /sys/fs/f2fs/<devname> 433(see also Documentation/ABI/testing/sysfs-fs-f2fs) 434 435Usage 436===== 437 4381. Download userland tools and compile them. 439 4402. Skip, if f2fs was compiled statically inside kernel. 441 Otherwise, insert the f2fs.ko module:: 442 443 # insmod f2fs.ko 444 4453. Create a directory to use when mounting:: 446 447 # mkdir /mnt/f2fs 448 4494. Format the block device, and then mount as f2fs:: 450 451 # mkfs.f2fs -l label /dev/block_device 452 # mount -t f2fs /dev/block_device /mnt/f2fs 453 454mkfs.f2fs 455--------- 456The mkfs.f2fs is for the use of formatting a partition as the f2fs filesystem, 457which builds a basic on-disk layout. 458 459The quick options consist of: 460 461=============== =========================================================== 462``-l [label]`` Give a volume label, up to 512 unicode name. 463``-a [0 or 1]`` Split start location of each area for heap-based allocation. 464 465 1 is set by default, which performs this. 466``-o [int]`` Set overprovision ratio in percent over volume size. 467 468 5 is set by default. 469``-s [int]`` Set the number of segments per section. 470 471 1 is set by default. 472``-z [int]`` Set the number of sections per zone. 473 474 1 is set by default. 475``-e [str]`` Set basic extension list. e.g. "mp3,gif,mov" 476``-t [0 or 1]`` Disable discard command or not. 477 478 1 is set by default, which conducts discard. 479=============== =========================================================== 480 481Note: please refer to the manpage of mkfs.f2fs(8) to get full option list. 482 483fsck.f2fs 484--------- 485The fsck.f2fs is a tool to check the consistency of an f2fs-formatted 486partition, which examines whether the filesystem metadata and user-made data 487are cross-referenced correctly or not. 488Note that, initial version of the tool does not fix any inconsistency. 489 490The quick options consist of:: 491 492 -d debug level [default:0] 493 494Note: please refer to the manpage of fsck.f2fs(8) to get full option list. 495 496dump.f2fs 497--------- 498The dump.f2fs shows the information of specific inode and dumps SSA and SIT to 499file. Each file is dump_ssa and dump_sit. 500 501The dump.f2fs is used to debug on-disk data structures of the f2fs filesystem. 502It shows on-disk inode information recognized by a given inode number, and is 503able to dump all the SSA and SIT entries into predefined files, ./dump_ssa and 504./dump_sit respectively. 505 506The options consist of:: 507 508 -d debug level [default:0] 509 -i inode no (hex) 510 -s [SIT dump segno from #1~#2 (decimal), for all 0~-1] 511 -a [SSA dump segno from #1~#2 (decimal), for all 0~-1] 512 513Examples:: 514 515 # dump.f2fs -i [ino] /dev/sdx 516 # dump.f2fs -s 0~-1 /dev/sdx (SIT dump) 517 # dump.f2fs -a 0~-1 /dev/sdx (SSA dump) 518 519Note: please refer to the manpage of dump.f2fs(8) to get full option list. 520 521sload.f2fs 522---------- 523The sload.f2fs gives a way to insert files and directories in the existing disk 524image. This tool is useful when building f2fs images given compiled files. 525 526Note: please refer to the manpage of sload.f2fs(8) to get full option list. 527 528resize.f2fs 529----------- 530The resize.f2fs lets a user resize the f2fs-formatted disk image, while preserving 531all the files and directories stored in the image. 532 533Note: please refer to the manpage of resize.f2fs(8) to get full option list. 534 535defrag.f2fs 536----------- 537The defrag.f2fs can be used to defragment scattered written data as well as 538filesystem metadata across the disk. This can improve the write speed by giving 539more free consecutive space. 540 541Note: please refer to the manpage of defrag.f2fs(8) to get full option list. 542 543f2fs_io 544------- 545The f2fs_io is a simple tool to issue various filesystem APIs as well as 546f2fs-specific ones, which is very useful for QA tests. 547 548Note: please refer to the manpage of f2fs_io(8) to get full option list. 549 550Design 551====== 552 553On-disk Layout 554-------------- 555 556F2FS divides the whole volume into a number of segments, each of which is fixed 557to 2MB in size. A section is composed of consecutive segments, and a zone 558consists of a set of sections. By default, section and zone sizes are set to one 559segment size identically, but users can easily modify the sizes by mkfs. 560 561F2FS splits the entire volume into six areas, and all the areas except superblock 562consist of multiple segments as described below:: 563 564 align with the zone size <-| 565 |-> align with the segment size 566 _________________________________________________________________________ 567 | | | Segment | Node | Segment | | 568 | Superblock | Checkpoint | Info. | Address | Summary | Main | 569 | (SB) | (CP) | Table (SIT) | Table (NAT) | Area (SSA) | | 570 |____________|_____2______|______N______|______N______|______N_____|__N___| 571 . . 572 . . 573 . . 574 ._________________________________________. 575 |_Segment_|_..._|_Segment_|_..._|_Segment_| 576 . . 577 ._________._________ 578 |_section_|__...__|_ 579 . . 580 .________. 581 |__zone__| 582 583- Superblock (SB) 584 It is located at the beginning of the partition, and there exist two copies 585 to avoid file system crash. It contains basic partition information and some 586 default parameters of f2fs. 587 588- Checkpoint (CP) 589 It contains file system information, bitmaps for valid NAT/SIT sets, orphan 590 inode lists, and summary entries of current active segments. 591 592- Segment Information Table (SIT) 593 It contains segment information such as valid block count and bitmap for the 594 validity of all the blocks. 595 596- Node Address Table (NAT) 597 It is composed of a block address table for all the node blocks stored in 598 Main area. 599 600- Segment Summary Area (SSA) 601 It contains summary entries which contains the owner information of all the 602 data and node blocks stored in Main area. 603 604- Main Area 605 It contains file and directory data including their indices. 606 607In order to avoid misalignment between file system and flash-based storage, F2FS 608aligns the start block address of CP with the segment size. Also, it aligns the 609start block address of Main area with the zone size by reserving some segments 610in SSA area. 611 612Reference the following survey for additional technical details. 613https://wiki.linaro.org/WorkingGroups/Kernel/Projects/FlashCardSurvey 614 615File System Metadata Structure 616------------------------------ 617 618F2FS adopts the checkpointing scheme to maintain file system consistency. At 619mount time, F2FS first tries to find the last valid checkpoint data by scanning 620CP area. In order to reduce the scanning time, F2FS uses only two copies of CP. 621One of them always indicates the last valid data, which is called as shadow copy 622mechanism. In addition to CP, NAT and SIT also adopt the shadow copy mechanism. 623 624For file system consistency, each CP points to which NAT and SIT copies are 625valid, as shown as below:: 626 627 +--------+----------+---------+ 628 | CP | SIT | NAT | 629 +--------+----------+---------+ 630 . . . . 631 . . . . 632 . . . . 633 +-------+-------+--------+--------+--------+--------+ 634 | CP #0 | CP #1 | SIT #0 | SIT #1 | NAT #0 | NAT #1 | 635 +-------+-------+--------+--------+--------+--------+ 636 | ^ ^ 637 | | | 638 `----------------------------------------' 639 640Index Structure 641--------------- 642 643The key data structure to manage the data locations is a "node". Similar to 644traditional file structures, F2FS has three types of node: inode, direct node, 645indirect node. F2FS assigns 4KB to an inode block which contains 923 data block 646indices, two direct node pointers, two indirect node pointers, and one double 647indirect node pointer as described below. One direct node block contains 1018 648data blocks, and one indirect node block contains also 1018 node blocks. Thus, 649one inode block (i.e., a file) covers:: 650 651 4KB * (923 + 2 * 1018 + 2 * 1018 * 1018 + 1018 * 1018 * 1018) := 3.94TB. 652 653 Inode block (4KB) 654 |- data (923) 655 |- direct node (2) 656 | `- data (1018) 657 |- indirect node (2) 658 | `- direct node (1018) 659 | `- data (1018) 660 `- double indirect node (1) 661 `- indirect node (1018) 662 `- direct node (1018) 663 `- data (1018) 664 665Note that all the node blocks are mapped by NAT which means the location of 666each node is translated by the NAT table. In the consideration of the wandering 667tree problem, F2FS is able to cut off the propagation of node updates caused by 668leaf data writes. 669 670Directory Structure 671------------------- 672 673A directory entry occupies 11 bytes, which consists of the following attributes. 674 675- hash hash value of the file name 676- ino inode number 677- len the length of file name 678- type file type such as directory, symlink, etc 679 680A dentry block consists of 214 dentry slots and file names. Therein a bitmap is 681used to represent whether each dentry is valid or not. A dentry block occupies 6824KB with the following composition. 683 684:: 685 686 Dentry Block(4 K) = bitmap (27 bytes) + reserved (3 bytes) + 687 dentries(11 * 214 bytes) + file name (8 * 214 bytes) 688 689 [Bucket] 690 +--------------------------------+ 691 |dentry block 1 | dentry block 2 | 692 +--------------------------------+ 693 . . 694 . . 695 . [Dentry Block Structure: 4KB] . 696 +--------+----------+----------+------------+ 697 | bitmap | reserved | dentries | file names | 698 +--------+----------+----------+------------+ 699 [Dentry Block: 4KB] . . 700 . . 701 . . 702 +------+------+-----+------+ 703 | hash | ino | len | type | 704 +------+------+-----+------+ 705 [Dentry Structure: 11 bytes] 706 707F2FS implements multi-level hash tables for directory structure. Each level has 708a hash table with dedicated number of hash buckets as shown below. Note that 709"A(2B)" means a bucket includes 2 data blocks. 710 711:: 712 713 ---------------------- 714 A : bucket 715 B : block 716 N : MAX_DIR_HASH_DEPTH 717 ---------------------- 718 719 level #0 | A(2B) 720 | 721 level #1 | A(2B) - A(2B) 722 | 723 level #2 | A(2B) - A(2B) - A(2B) - A(2B) 724 . | . . . . 725 level #N/2 | A(2B) - A(2B) - A(2B) - A(2B) - A(2B) - ... - A(2B) 726 . | . . . . 727 level #N | A(4B) - A(4B) - A(4B) - A(4B) - A(4B) - ... - A(4B) 728 729The number of blocks and buckets are determined by:: 730 731 ,- 2, if n < MAX_DIR_HASH_DEPTH / 2, 732 # of blocks in level #n = | 733 `- 4, Otherwise 734 735 ,- 2^(n + dir_level), 736 | if n + dir_level < MAX_DIR_HASH_DEPTH / 2, 737 # of buckets in level #n = | 738 `- 2^((MAX_DIR_HASH_DEPTH / 2) - 1), 739 Otherwise 740 741When F2FS finds a file name in a directory, at first a hash value of the file 742name is calculated. Then, F2FS scans the hash table in level #0 to find the 743dentry consisting of the file name and its inode number. If not found, F2FS 744scans the next hash table in level #1. In this way, F2FS scans hash tables in 745each levels incrementally from 1 to N. In each level F2FS needs to scan only 746one bucket determined by the following equation, which shows O(log(# of files)) 747complexity:: 748 749 bucket number to scan in level #n = (hash value) % (# of buckets in level #n) 750 751In the case of file creation, F2FS finds empty consecutive slots that cover the 752file name. F2FS searches the empty slots in the hash tables of whole levels from 7531 to N in the same way as the lookup operation. 754 755The following figure shows an example of two cases holding children:: 756 757 --------------> Dir <-------------- 758 | | 759 child child 760 761 child - child [hole] - child 762 763 child - child - child [hole] - [hole] - child 764 765 Case 1: Case 2: 766 Number of children = 6, Number of children = 3, 767 File size = 7 File size = 7 768 769Default Block Allocation 770------------------------ 771 772At runtime, F2FS manages six active logs inside "Main" area: Hot/Warm/Cold node 773and Hot/Warm/Cold data. 774 775- Hot node contains direct node blocks of directories. 776- Warm node contains direct node blocks except hot node blocks. 777- Cold node contains indirect node blocks 778- Hot data contains dentry blocks 779- Warm data contains data blocks except hot and cold data blocks 780- Cold data contains multimedia data or migrated data blocks 781 782LFS has two schemes for free space management: threaded log and copy-and-compac- 783tion. The copy-and-compaction scheme which is known as cleaning, is well-suited 784for devices showing very good sequential write performance, since free segments 785are served all the time for writing new data. However, it suffers from cleaning 786overhead under high utilization. Contrarily, the threaded log scheme suffers 787from random writes, but no cleaning process is needed. F2FS adopts a hybrid 788scheme where the copy-and-compaction scheme is adopted by default, but the 789policy is dynamically changed to the threaded log scheme according to the file 790system status. 791 792In order to align F2FS with underlying flash-based storage, F2FS allocates a 793segment in a unit of section. F2FS expects that the section size would be the 794same as the unit size of garbage collection in FTL. Furthermore, with respect 795to the mapping granularity in FTL, F2FS allocates each section of the active 796logs from different zones as much as possible, since FTL can write the data in 797the active logs into one allocation unit according to its mapping granularity. 798 799Cleaning process 800---------------- 801 802F2FS does cleaning both on demand and in the background. On-demand cleaning is 803triggered when there are not enough free segments to serve VFS calls. Background 804cleaner is operated by a kernel thread, and triggers the cleaning job when the 805system is idle. 806 807F2FS supports two victim selection policies: greedy and cost-benefit algorithms. 808In the greedy algorithm, F2FS selects a victim segment having the smallest number 809of valid blocks. In the cost-benefit algorithm, F2FS selects a victim segment 810according to the segment age and the number of valid blocks in order to address 811log block thrashing problem in the greedy algorithm. F2FS adopts the greedy 812algorithm for on-demand cleaner, while background cleaner adopts cost-benefit 813algorithm. 814 815In order to identify whether the data in the victim segment are valid or not, 816F2FS manages a bitmap. Each bit represents the validity of a block, and the 817bitmap is composed of a bit stream covering whole blocks in main area. 818 819Write-hint Policy 820----------------- 821 822F2FS sets the whint all the time with the below policy. 823 824===================== ======================== =================== 825User F2FS Block 826===================== ======================== =================== 827N/A META WRITE_LIFE_NONE|REQ_META 828N/A HOT_NODE WRITE_LIFE_NONE 829N/A WARM_NODE WRITE_LIFE_MEDIUM 830N/A COLD_NODE WRITE_LIFE_LONG 831ioctl(COLD) COLD_DATA WRITE_LIFE_EXTREME 832extension list " " 833 834-- buffered io 835------------------------------------------------------------------ 836N/A COLD_DATA WRITE_LIFE_EXTREME 837N/A HOT_DATA WRITE_LIFE_SHORT 838N/A WARM_DATA WRITE_LIFE_NOT_SET 839 840-- direct io 841------------------------------------------------------------------ 842WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME 843WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT 844WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_NOT_SET 845WRITE_LIFE_NONE " WRITE_LIFE_NONE 846WRITE_LIFE_MEDIUM " WRITE_LIFE_MEDIUM 847WRITE_LIFE_LONG " WRITE_LIFE_LONG 848===================== ======================== =================== 849 850Fallocate(2) Policy 851------------------- 852 853The default policy follows the below POSIX rule. 854 855Allocating disk space 856 The default operation (i.e., mode is zero) of fallocate() allocates 857 the disk space within the range specified by offset and len. The 858 file size (as reported by stat(2)) will be changed if offset+len is 859 greater than the file size. Any subregion within the range specified 860 by offset and len that did not contain data before the call will be 861 initialized to zero. This default behavior closely resembles the 862 behavior of the posix_fallocate(3) library function, and is intended 863 as a method of optimally implementing that function. 864 865However, once F2FS receives ioctl(fd, F2FS_IOC_SET_PIN_FILE) in prior to 866fallocate(fd, DEFAULT_MODE), it allocates on-disk block addresses having 867zero or random data, which is useful to the below scenario where: 868 869 1. create(fd) 870 2. ioctl(fd, F2FS_IOC_SET_PIN_FILE) 871 3. fallocate(fd, 0, 0, size) 872 4. address = fibmap(fd, offset) 873 5. open(blkdev) 874 6. write(blkdev, address) 875 876Compression implementation 877-------------------------- 878 879- New term named cluster is defined as basic unit of compression, file can 880 be divided into multiple clusters logically. One cluster includes 4 << n 881 (n >= 0) logical pages, compression size is also cluster size, each of 882 cluster can be compressed or not. 883 884- In cluster metadata layout, one special block address is used to indicate 885 a cluster is a compressed one or normal one; for compressed cluster, following 886 metadata maps cluster to [1, 4 << n - 1] physical blocks, in where f2fs 887 stores data including compress header and compressed data. 888 889- In order to eliminate write amplification during overwrite, F2FS only 890 support compression on write-once file, data can be compressed only when 891 all logical blocks in cluster contain valid data and compress ratio of 892 cluster data is lower than specified threshold. 893 894- To enable compression on regular inode, there are four ways: 895 896 * chattr +c file 897 * chattr +c dir; touch dir/file 898 * mount w/ -o compress_extension=ext; touch file.ext 899 * mount w/ -o compress_extension=*; touch any_file 900 901- To disable compression on regular inode, there are two ways: 902 903 * chattr -c file 904 * mount w/ -o nocompress_extension=ext; touch file.ext 905 906- Priority in between FS_COMPR_FL, FS_NOCOMP_FS, extensions: 907 908 * compress_extension=so; nocompress_extension=zip; chattr +c dir; touch 909 dir/foo.so; touch dir/bar.zip; touch dir/baz.txt; then foo.so and baz.txt 910 should be compresse, bar.zip should be non-compressed. chattr +c dir/bar.zip 911 can enable compress on bar.zip. 912 * compress_extension=so; nocompress_extension=zip; chattr -c dir; touch 913 dir/foo.so; touch dir/bar.zip; touch dir/baz.txt; then foo.so should be 914 compresse, bar.zip and baz.txt should be non-compressed. 915 chattr+c dir/bar.zip; chattr+c dir/baz.txt; can enable compress on bar.zip 916 and baz.txt. 917 918- At this point, compression feature doesn't expose compressed space to user 919 directly in order to guarantee potential data updates later to the space. 920 Instead, the main goal is to reduce data writes to flash disk as much as 921 possible, resulting in extending disk life time as well as relaxing IO 922 congestion. Alternatively, we've added ioctl(F2FS_IOC_RELEASE_COMPRESS_BLOCKS) 923 interface to reclaim compressed space and show it to user after setting a 924 special flag to the inode. Once the compressed space is released, the flag 925 will block writing data to the file until either the compressed space is 926 reserved via ioctl(F2FS_IOC_RESERVE_COMPRESS_BLOCKS) or the file size is 927 truncated to zero. 928 929Compress metadata layout:: 930 931 [Dnode Structure] 932 +-----------------------------------------------+ 933 | cluster 1 | cluster 2 | ......... | cluster N | 934 +-----------------------------------------------+ 935 . . . . 936 . . . . 937 . Compressed Cluster . . Normal Cluster . 938 +----------+---------+---------+---------+ +---------+---------+---------+---------+ 939 |compr flag| block 1 | block 2 | block 3 | | block 1 | block 2 | block 3 | block 4 | 940 +----------+---------+---------+---------+ +---------+---------+---------+---------+ 941 . . 942 . . 943 . . 944 +-------------+-------------+----------+----------------------------+ 945 | data length | data chksum | reserved | compressed data | 946 +-------------+-------------+----------+----------------------------+ 947 948Compression mode 949-------------------------- 950 951f2fs supports "fs" and "user" compression modes with "compression_mode" mount option. 952With this option, f2fs provides a choice to select the way how to compress the 953compression enabled files (refer to "Compression implementation" section for how to 954enable compression on a regular inode). 955 9561) compress_mode=fs 957 958 This is the default option. f2fs does automatic compression in the writeback of the 959 compression enabled files. 960 9612) compress_mode=user 962 963 This disables the automatic compression and gives the user discretion of choosing the 964 target file and the timing. The user can do manual compression/decompression on the 965 compression enabled files using F2FS_IOC_DECOMPRESS_FILE and F2FS_IOC_COMPRESS_FILE 966 ioctls like the below. 967 968To decompress a file:: 969 970 fd = open(filename, O_WRONLY, 0); 971 ret = ioctl(fd, F2FS_IOC_DECOMPRESS_FILE); 972 973To compress a file:: 974 975 fd = open(filename, O_WRONLY, 0); 976 ret = ioctl(fd, F2FS_IOC_COMPRESS_FILE); 977 978NVMe Zoned Namespace devices 979---------------------------- 980 981- ZNS defines a per-zone capacity which can be equal or less than the 982 zone-size. Zone-capacity is the number of usable blocks in the zone. 983 F2FS checks if zone-capacity is less than zone-size, if it is, then any 984 segment which starts after the zone-capacity is marked as not-free in 985 the free segment bitmap at initial mount time. These segments are marked 986 as permanently used so they are not allocated for writes and 987 consequently are not needed to be garbage collected. In case the 988 zone-capacity is not aligned to default segment size(2MB), then a segment 989 can start before the zone-capacity and span across zone-capacity boundary. 990 Such spanning segments are also considered as usable segments. All blocks 991 past the zone-capacity are considered unusable in these segments. 992 993Device aliasing feature 994----------------------- 995 996f2fs can utilize a special file called a "device aliasing file." This file allows 997the entire storage device to be mapped with a single, large extent, not using 998the usual f2fs node structures. This mapped area is pinned and primarily intended 999for holding the space. 1000 1001Essentially, this mechanism allows a portion of the f2fs area to be temporarily 1002reserved and used by another filesystem or for different purposes. Once that 1003external usage is complete, the device aliasing file can be deleted, releasing 1004the reserved space back to F2FS for its own use. 1005 1006.. code-block:: 1007 1008 # ls /dev/vd* 1009 /dev/vdb (32GB) /dev/vdc (32GB) 1010 # mkfs.ext4 /dev/vdc 1011 # mkfs.f2fs -c /dev/vdc@vdc.file /dev/vdb 1012 # mount /dev/vdb /mnt/f2fs 1013 # ls -l /mnt/f2fs 1014 vdc.file 1015 # df -h 1016 /dev/vdb 64G 33G 32G 52% /mnt/f2fs 1017 1018 # mount -o loop /dev/vdc /mnt/ext4 1019 # df -h 1020 /dev/vdb 64G 33G 32G 52% /mnt/f2fs 1021 /dev/loop7 32G 24K 30G 1% /mnt/ext4 1022 # umount /mnt/ext4 1023 1024 # f2fs_io getflags /mnt/f2fs/vdc.file 1025 get a flag on /mnt/f2fs/vdc.file ret=0, flags=nocow(pinned),immutable 1026 # f2fs_io setflags noimmutable /mnt/f2fs/vdc.file 1027 get a flag on noimmutable ret=0, flags=800010 1028 set a flag on /mnt/f2fs/vdc.file ret=0, flags=noimmutable 1029 # rm /mnt/f2fs/vdc.file 1030 # df -h 1031 /dev/vdb 64G 753M 64G 2% /mnt/f2fs 1032 1033So, the key idea is, user can do any file operations on /dev/vdc, and 1034reclaim the space after the use, while the space is counted as /data. 1035That doesn't require modifying partition size and filesystem format. 1036