1.. SPDX-License-Identifier: GPL-2.0 2 3========================================== 4WHAT IS Flash-Friendly File System (F2FS)? 5========================================== 6 7NAND flash memory-based storage devices, such as SSD, eMMC, and SD cards, have 8been equipped on a variety systems ranging from mobile to server systems. Since 9they are known to have different characteristics from the conventional rotating 10disks, a file system, an upper layer to the storage device, should adapt to the 11changes from the sketch in the design level. 12 13F2FS is a file system exploiting NAND flash memory-based storage devices, which 14is based on Log-structured File System (LFS). The design has been focused on 15addressing the fundamental issues in LFS, which are snowball effect of wandering 16tree and high cleaning overhead. 17 18Since a NAND flash memory-based storage device shows different characteristic 19according to its internal geometry or flash memory management scheme, namely FTL, 20F2FS and its tools support various parameters not only for configuring on-disk 21layout, but also for selecting allocation and cleaning algorithms. 22 23The following git tree provides the file system formatting tool (mkfs.f2fs), 24a consistency checking tool (fsck.f2fs), and a debugging tool (dump.f2fs). 25 26- git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs-tools.git 27 28For reporting bugs and sending patches, please use the following mailing list: 29 30- linux-f2fs-devel@lists.sourceforge.net 31 32Background and Design issues 33============================ 34 35Log-structured File System (LFS) 36-------------------------------- 37"A log-structured file system writes all modifications to disk sequentially in 38a log-like structure, thereby speeding up both file writing and crash recovery. 39The log is the only structure on disk; it contains indexing information so that 40files can be read back from the log efficiently. In order to maintain large free 41areas on disk for fast writing, we divide the log into segments and use a 42segment cleaner to compress the live information from heavily fragmented 43segments." from Rosenblum, M. and Ousterhout, J. K., 1992, "The design and 44implementation of a log-structured file system", ACM Trans. Computer Systems 4510, 1, 26–52. 46 47Wandering Tree Problem 48---------------------- 49In LFS, when a file data is updated and written to the end of log, its direct 50pointer block is updated due to the changed location. Then the indirect pointer 51block is also updated due to the direct pointer block update. In this manner, 52the upper index structures such as inode, inode map, and checkpoint block are 53also updated recursively. This problem is called as wandering tree problem [1], 54and in order to enhance the performance, it should eliminate or relax the update 55propagation as much as possible. 56 57[1] Bityutskiy, A. 2005. JFFS3 design issues. http://www.linux-mtd.infradead.org/ 58 59Cleaning Overhead 60----------------- 61Since LFS is based on out-of-place writes, it produces so many obsolete blocks 62scattered across the whole storage. In order to serve new empty log space, it 63needs to reclaim these obsolete blocks seamlessly to users. This job is called 64as a cleaning process. 65 66The process consists of three operations as follows. 67 681. A victim segment is selected through referencing segment usage table. 692. It loads parent index structures of all the data in the victim identified by 70 segment summary blocks. 713. It checks the cross-reference between the data and its parent index structure. 724. It moves valid data selectively. 73 74This cleaning job may cause unexpected long delays, so the most important goal 75is to hide the latencies to users. And also definitely, it should reduce the 76amount of valid data to be moved, and move them quickly as well. 77 78Key Features 79============ 80 81Flash Awareness 82--------------- 83- Enlarge the random write area for better performance, but provide the high 84 spatial locality 85- Align FS data structures to the operational units in FTL as best efforts 86 87Wandering Tree Problem 88---------------------- 89- Use a term, “node”, that represents inodes as well as various pointer blocks 90- Introduce Node Address Table (NAT) containing the locations of all the “node” 91 blocks; this will cut off the update propagation. 92 93Cleaning Overhead 94----------------- 95- Support a background cleaning process 96- Support greedy and cost-benefit algorithms for victim selection policies 97- Support multi-head logs for static/dynamic hot and cold data separation 98- Introduce adaptive logging for efficient block allocation 99 100Mount Options 101============= 102 103 104======================== ============================================================ 105background_gc=%s Turn on/off cleaning operations, namely garbage 106 collection, triggered in background when I/O subsystem is 107 idle. If background_gc=on, it will turn on the garbage 108 collection and if background_gc=off, garbage collection 109 will be turned off. If background_gc=sync, it will turn 110 on synchronous garbage collection running in background. 111 Default value for this option is on. So garbage 112 collection is on by default. 113disable_roll_forward Disable the roll-forward recovery routine 114norecovery Disable the roll-forward recovery routine, mounted read- 115 only (i.e., -o ro,disable_roll_forward) 116discard/nodiscard Enable/disable real-time discard in f2fs, if discard is 117 enabled, f2fs will issue discard/TRIM commands when a 118 segment is cleaned. 119no_heap Disable heap-style segment allocation which finds free 120 segments for data from the beginning of main area, while 121 for node from the end of main area. 122nouser_xattr Disable Extended User Attributes. Note: xattr is enabled 123 by default if CONFIG_F2FS_FS_XATTR is selected. 124noacl Disable POSIX Access Control List. Note: acl is enabled 125 by default if CONFIG_F2FS_FS_POSIX_ACL is selected. 126active_logs=%u Support configuring the number of active logs. In the 127 current design, f2fs supports only 2, 4, and 6 logs. 128 Default number is 6. 129disable_ext_identify Disable the extension list configured by mkfs, so f2fs 130 is not aware of cold files such as media files. 131inline_xattr Enable the inline xattrs feature. 132noinline_xattr Disable the inline xattrs feature. 133inline_xattr_size=%u Support configuring inline xattr size, it depends on 134 flexible inline xattr feature. 135inline_data Enable the inline data feature: Newly created small (<~3.4k) 136 files can be written into inode block. 137inline_dentry Enable the inline dir feature: data in newly created 138 directory entries can be written into inode block. The 139 space of inode block which is used to store inline 140 dentries is limited to ~3.4k. 141noinline_dentry Disable the inline dentry feature. 142flush_merge Merge concurrent cache_flush commands as much as possible 143 to eliminate redundant command issues. If the underlying 144 device handles the cache_flush command relatively slowly, 145 recommend to enable this option. 146nobarrier This option can be used if underlying storage guarantees 147 its cached data should be written to the novolatile area. 148 If this option is set, no cache_flush commands are issued 149 but f2fs still guarantees the write ordering of all the 150 data writes. 151fastboot This option is used when a system wants to reduce mount 152 time as much as possible, even though normal performance 153 can be sacrificed. 154extent_cache Enable an extent cache based on rb-tree, it can cache 155 as many as extent which map between contiguous logical 156 address and physical address per inode, resulting in 157 increasing the cache hit ratio. Set by default. 158noextent_cache Disable an extent cache based on rb-tree explicitly, see 159 the above extent_cache mount option. 160noinline_data Disable the inline data feature, inline data feature is 161 enabled by default. 162data_flush Enable data flushing before checkpoint in order to 163 persist data of regular and symlink. 164reserve_root=%d Support configuring reserved space which is used for 165 allocation from a privileged user with specified uid or 166 gid, unit: 4KB, the default limit is 0.2% of user blocks. 167resuid=%d The user ID which may use the reserved blocks. 168resgid=%d The group ID which may use the reserved blocks. 169fault_injection=%d Enable fault injection in all supported types with 170 specified injection rate. 171fault_type=%d Support configuring fault injection type, should be 172 enabled with fault_injection option, fault type value 173 is shown below, it supports single or combined type. 174 175 =================== =========== 176 Type_Name Type_Value 177 =================== =========== 178 FAULT_KMALLOC 0x000000001 179 FAULT_KVMALLOC 0x000000002 180 FAULT_PAGE_ALLOC 0x000000004 181 FAULT_PAGE_GET 0x000000008 182 FAULT_ALLOC_NID 0x000000020 183 FAULT_ORPHAN 0x000000040 184 FAULT_BLOCK 0x000000080 185 FAULT_DIR_DEPTH 0x000000100 186 FAULT_EVICT_INODE 0x000000200 187 FAULT_TRUNCATE 0x000000400 188 FAULT_READ_IO 0x000000800 189 FAULT_CHECKPOINT 0x000001000 190 FAULT_DISCARD 0x000002000 191 FAULT_WRITE_IO 0x000004000 192 =================== =========== 193mode=%s Control block allocation mode which supports "adaptive" 194 and "lfs". In "lfs" mode, there should be no random 195 writes towards main area. 196io_bits=%u Set the bit size of write IO requests. It should be set 197 with "mode=lfs". 198usrquota Enable plain user disk quota accounting. 199grpquota Enable plain group disk quota accounting. 200prjquota Enable plain project quota accounting. 201usrjquota=<file> Appoint specified file and type during mount, so that quota 202grpjquota=<file> information can be properly updated during recovery flow, 203prjjquota=<file> <quota file>: must be in root directory; 204jqfmt=<quota type> <quota type>: [vfsold,vfsv0,vfsv1]. 205offusrjquota Turn off user journalled quota. 206offgrpjquota Turn off group journalled quota. 207offprjjquota Turn off project journalled quota. 208quota Enable plain user disk quota accounting. 209noquota Disable all plain disk quota option. 210whint_mode=%s Control which write hints are passed down to block 211 layer. This supports "off", "user-based", and 212 "fs-based". In "off" mode (default), f2fs does not pass 213 down hints. In "user-based" mode, f2fs tries to pass 214 down hints given by users. And in "fs-based" mode, f2fs 215 passes down hints with its policy. 216alloc_mode=%s Adjust block allocation policy, which supports "reuse" 217 and "default". 218fsync_mode=%s Control the policy of fsync. Currently supports "posix", 219 "strict", and "nobarrier". In "posix" mode, which is 220 default, fsync will follow POSIX semantics and does a 221 light operation to improve the filesystem performance. 222 In "strict" mode, fsync will be heavy and behaves in line 223 with xfs, ext4 and btrfs, where xfstest generic/342 will 224 pass, but the performance will regress. "nobarrier" is 225 based on "posix", but doesn't issue flush command for 226 non-atomic files likewise "nobarrier" mount option. 227test_dummy_encryption 228test_dummy_encryption=%s 229 Enable dummy encryption, which provides a fake fscrypt 230 context. The fake fscrypt context is used by xfstests. 231 The argument may be either "v1" or "v2", in order to 232 select the corresponding fscrypt policy version. 233checkpoint=%s[:%u[%]] Set to "disable" to turn off checkpointing. Set to "enable" 234 to reenable checkpointing. Is enabled by default. While 235 disabled, any unmounting or unexpected shutdowns will cause 236 the filesystem contents to appear as they did when the 237 filesystem was mounted with that option. 238 While mounting with checkpoint=disabled, the filesystem must 239 run garbage collection to ensure that all available space can 240 be used. If this takes too much time, the mount may return 241 EAGAIN. You may optionally add a value to indicate how much 242 of the disk you would be willing to temporarily give up to 243 avoid additional garbage collection. This can be given as a 244 number of blocks, or as a percent. For instance, mounting 245 with checkpoint=disable:100% would always succeed, but it may 246 hide up to all remaining free space. The actual space that 247 would be unusable can be viewed at /sys/fs/f2fs/<disk>/unusable 248 This space is reclaimed once checkpoint=enable. 249checkpoint_merge When checkpoint is enabled, this can be used to create a kernel 250 daemon and make it to merge concurrent checkpoint requests as 251 much as possible to eliminate redundant checkpoint issues. Plus, 252 we can eliminate the sluggish issue caused by slow checkpoint 253 operation when the checkpoint is done in a process context in 254 a cgroup having low i/o budget and cpu shares. To make this 255 do better, we set the default i/o priority of the kernel daemon 256 to "3", to give one higher priority than other kernel threads. 257 This is the same way to give a I/O priority to the jbd2 258 journaling thread of ext4 filesystem. 259nocheckpoint_merge Disable checkpoint merge feature. 260compress_algorithm=%s Control compress algorithm, currently f2fs supports "lzo", 261 "lz4", "zstd" and "lzo-rle" algorithm. 262compress_algorithm=%s:%d Control compress algorithm and its compress level, now, only 263 "lz4" and "zstd" support compress level config. 264 algorithm level range 265 lz4 3 - 16 266 zstd 1 - 22 267compress_log_size=%u Support configuring compress cluster size, the size will 268 be 4KB * (1 << %u), 16KB is minimum size, also it's 269 default size. 270compress_extension=%s Support adding specified extension, so that f2fs can enable 271 compression on those corresponding files, e.g. if all files 272 with '.ext' has high compression rate, we can set the '.ext' 273 on compression extension list and enable compression on 274 these file by default rather than to enable it via ioctl. 275 For other files, we can still enable compression via ioctl. 276 Note that, there is one reserved special extension '*', it 277 can be set to enable compression for all files. 278compress_chksum Support verifying chksum of raw data in compressed cluster. 279compress_mode=%s Control file compression mode. This supports "fs" and "user" 280 modes. In "fs" mode (default), f2fs does automatic compression 281 on the compression enabled files. In "user" mode, f2fs disables 282 the automaic compression and gives the user discretion of 283 choosing the target file and the timing. The user can do manual 284 compression/decompression on the compression enabled files using 285 ioctls. 286inlinecrypt When possible, encrypt/decrypt the contents of encrypted 287 files using the blk-crypto framework rather than 288 filesystem-layer encryption. This allows the use of 289 inline encryption hardware. The on-disk format is 290 unaffected. For more details, see 291 Documentation/block/inline-encryption.rst. 292atgc Enable age-threshold garbage collection, it provides high 293 effectiveness and efficiency on background GC. 294======================== ============================================================ 295 296Debugfs Entries 297=============== 298 299/sys/kernel/debug/f2fs/ contains information about all the partitions mounted as 300f2fs. Each file shows the whole f2fs information. 301 302/sys/kernel/debug/f2fs/status includes: 303 304 - major file system information managed by f2fs currently 305 - average SIT information about whole segments 306 - current memory footprint consumed by f2fs. 307 308Sysfs Entries 309============= 310 311Information about mounted f2fs file systems can be found in 312/sys/fs/f2fs. Each mounted filesystem will have a directory in 313/sys/fs/f2fs based on its device name (i.e., /sys/fs/f2fs/sda). 314The files in each per-device directory are shown in table below. 315 316Files in /sys/fs/f2fs/<devname> 317(see also Documentation/ABI/testing/sysfs-fs-f2fs) 318 319Usage 320===== 321 3221. Download userland tools and compile them. 323 3242. Skip, if f2fs was compiled statically inside kernel. 325 Otherwise, insert the f2fs.ko module:: 326 327 # insmod f2fs.ko 328 3293. Create a directory to use when mounting:: 330 331 # mkdir /mnt/f2fs 332 3334. Format the block device, and then mount as f2fs:: 334 335 # mkfs.f2fs -l label /dev/block_device 336 # mount -t f2fs /dev/block_device /mnt/f2fs 337 338mkfs.f2fs 339--------- 340The mkfs.f2fs is for the use of formatting a partition as the f2fs filesystem, 341which builds a basic on-disk layout. 342 343The quick options consist of: 344 345=============== =========================================================== 346``-l [label]`` Give a volume label, up to 512 unicode name. 347``-a [0 or 1]`` Split start location of each area for heap-based allocation. 348 349 1 is set by default, which performs this. 350``-o [int]`` Set overprovision ratio in percent over volume size. 351 352 5 is set by default. 353``-s [int]`` Set the number of segments per section. 354 355 1 is set by default. 356``-z [int]`` Set the number of sections per zone. 357 358 1 is set by default. 359``-e [str]`` Set basic extension list. e.g. "mp3,gif,mov" 360``-t [0 or 1]`` Disable discard command or not. 361 362 1 is set by default, which conducts discard. 363=============== =========================================================== 364 365Note: please refer to the manpage of mkfs.f2fs(8) to get full option list. 366 367fsck.f2fs 368--------- 369The fsck.f2fs is a tool to check the consistency of an f2fs-formatted 370partition, which examines whether the filesystem metadata and user-made data 371are cross-referenced correctly or not. 372Note that, initial version of the tool does not fix any inconsistency. 373 374The quick options consist of:: 375 376 -d debug level [default:0] 377 378Note: please refer to the manpage of fsck.f2fs(8) to get full option list. 379 380dump.f2fs 381--------- 382The dump.f2fs shows the information of specific inode and dumps SSA and SIT to 383file. Each file is dump_ssa and dump_sit. 384 385The dump.f2fs is used to debug on-disk data structures of the f2fs filesystem. 386It shows on-disk inode information recognized by a given inode number, and is 387able to dump all the SSA and SIT entries into predefined files, ./dump_ssa and 388./dump_sit respectively. 389 390The options consist of:: 391 392 -d debug level [default:0] 393 -i inode no (hex) 394 -s [SIT dump segno from #1~#2 (decimal), for all 0~-1] 395 -a [SSA dump segno from #1~#2 (decimal), for all 0~-1] 396 397Examples:: 398 399 # dump.f2fs -i [ino] /dev/sdx 400 # dump.f2fs -s 0~-1 /dev/sdx (SIT dump) 401 # dump.f2fs -a 0~-1 /dev/sdx (SSA dump) 402 403Note: please refer to the manpage of dump.f2fs(8) to get full option list. 404 405sload.f2fs 406---------- 407The sload.f2fs gives a way to insert files and directories in the exisiting disk 408image. This tool is useful when building f2fs images given compiled files. 409 410Note: please refer to the manpage of sload.f2fs(8) to get full option list. 411 412resize.f2fs 413----------- 414The resize.f2fs lets a user resize the f2fs-formatted disk image, while preserving 415all the files and directories stored in the image. 416 417Note: please refer to the manpage of resize.f2fs(8) to get full option list. 418 419defrag.f2fs 420----------- 421The defrag.f2fs can be used to defragment scattered written data as well as 422filesystem metadata across the disk. This can improve the write speed by giving 423more free consecutive space. 424 425Note: please refer to the manpage of defrag.f2fs(8) to get full option list. 426 427f2fs_io 428------- 429The f2fs_io is a simple tool to issue various filesystem APIs as well as 430f2fs-specific ones, which is very useful for QA tests. 431 432Note: please refer to the manpage of f2fs_io(8) to get full option list. 433 434Design 435====== 436 437On-disk Layout 438-------------- 439 440F2FS divides the whole volume into a number of segments, each of which is fixed 441to 2MB in size. A section is composed of consecutive segments, and a zone 442consists of a set of sections. By default, section and zone sizes are set to one 443segment size identically, but users can easily modify the sizes by mkfs. 444 445F2FS splits the entire volume into six areas, and all the areas except superblock 446consist of multiple segments as described below:: 447 448 align with the zone size <-| 449 |-> align with the segment size 450 _________________________________________________________________________ 451 | | | Segment | Node | Segment | | 452 | Superblock | Checkpoint | Info. | Address | Summary | Main | 453 | (SB) | (CP) | Table (SIT) | Table (NAT) | Area (SSA) | | 454 |____________|_____2______|______N______|______N______|______N_____|__N___| 455 . . 456 . . 457 . . 458 ._________________________________________. 459 |_Segment_|_..._|_Segment_|_..._|_Segment_| 460 . . 461 ._________._________ 462 |_section_|__...__|_ 463 . . 464 .________. 465 |__zone__| 466 467- Superblock (SB) 468 It is located at the beginning of the partition, and there exist two copies 469 to avoid file system crash. It contains basic partition information and some 470 default parameters of f2fs. 471 472- Checkpoint (CP) 473 It contains file system information, bitmaps for valid NAT/SIT sets, orphan 474 inode lists, and summary entries of current active segments. 475 476- Segment Information Table (SIT) 477 It contains segment information such as valid block count and bitmap for the 478 validity of all the blocks. 479 480- Node Address Table (NAT) 481 It is composed of a block address table for all the node blocks stored in 482 Main area. 483 484- Segment Summary Area (SSA) 485 It contains summary entries which contains the owner information of all the 486 data and node blocks stored in Main area. 487 488- Main Area 489 It contains file and directory data including their indices. 490 491In order to avoid misalignment between file system and flash-based storage, F2FS 492aligns the start block address of CP with the segment size. Also, it aligns the 493start block address of Main area with the zone size by reserving some segments 494in SSA area. 495 496Reference the following survey for additional technical details. 497https://wiki.linaro.org/WorkingGroups/Kernel/Projects/FlashCardSurvey 498 499File System Metadata Structure 500------------------------------ 501 502F2FS adopts the checkpointing scheme to maintain file system consistency. At 503mount time, F2FS first tries to find the last valid checkpoint data by scanning 504CP area. In order to reduce the scanning time, F2FS uses only two copies of CP. 505One of them always indicates the last valid data, which is called as shadow copy 506mechanism. In addition to CP, NAT and SIT also adopt the shadow copy mechanism. 507 508For file system consistency, each CP points to which NAT and SIT copies are 509valid, as shown as below:: 510 511 +--------+----------+---------+ 512 | CP | SIT | NAT | 513 +--------+----------+---------+ 514 . . . . 515 . . . . 516 . . . . 517 +-------+-------+--------+--------+--------+--------+ 518 | CP #0 | CP #1 | SIT #0 | SIT #1 | NAT #0 | NAT #1 | 519 +-------+-------+--------+--------+--------+--------+ 520 | ^ ^ 521 | | | 522 `----------------------------------------' 523 524Index Structure 525--------------- 526 527The key data structure to manage the data locations is a "node". Similar to 528traditional file structures, F2FS has three types of node: inode, direct node, 529indirect node. F2FS assigns 4KB to an inode block which contains 923 data block 530indices, two direct node pointers, two indirect node pointers, and one double 531indirect node pointer as described below. One direct node block contains 1018 532data blocks, and one indirect node block contains also 1018 node blocks. Thus, 533one inode block (i.e., a file) covers:: 534 535 4KB * (923 + 2 * 1018 + 2 * 1018 * 1018 + 1018 * 1018 * 1018) := 3.94TB. 536 537 Inode block (4KB) 538 |- data (923) 539 |- direct node (2) 540 | `- data (1018) 541 |- indirect node (2) 542 | `- direct node (1018) 543 | `- data (1018) 544 `- double indirect node (1) 545 `- indirect node (1018) 546 `- direct node (1018) 547 `- data (1018) 548 549Note that all the node blocks are mapped by NAT which means the location of 550each node is translated by the NAT table. In the consideration of the wandering 551tree problem, F2FS is able to cut off the propagation of node updates caused by 552leaf data writes. 553 554Directory Structure 555------------------- 556 557A directory entry occupies 11 bytes, which consists of the following attributes. 558 559- hash hash value of the file name 560- ino inode number 561- len the length of file name 562- type file type such as directory, symlink, etc 563 564A dentry block consists of 214 dentry slots and file names. Therein a bitmap is 565used to represent whether each dentry is valid or not. A dentry block occupies 5664KB with the following composition. 567 568:: 569 570 Dentry Block(4 K) = bitmap (27 bytes) + reserved (3 bytes) + 571 dentries(11 * 214 bytes) + file name (8 * 214 bytes) 572 573 [Bucket] 574 +--------------------------------+ 575 |dentry block 1 | dentry block 2 | 576 +--------------------------------+ 577 . . 578 . . 579 . [Dentry Block Structure: 4KB] . 580 +--------+----------+----------+------------+ 581 | bitmap | reserved | dentries | file names | 582 +--------+----------+----------+------------+ 583 [Dentry Block: 4KB] . . 584 . . 585 . . 586 +------+------+-----+------+ 587 | hash | ino | len | type | 588 +------+------+-----+------+ 589 [Dentry Structure: 11 bytes] 590 591F2FS implements multi-level hash tables for directory structure. Each level has 592a hash table with dedicated number of hash buckets as shown below. Note that 593"A(2B)" means a bucket includes 2 data blocks. 594 595:: 596 597 ---------------------- 598 A : bucket 599 B : block 600 N : MAX_DIR_HASH_DEPTH 601 ---------------------- 602 603 level #0 | A(2B) 604 | 605 level #1 | A(2B) - A(2B) 606 | 607 level #2 | A(2B) - A(2B) - A(2B) - A(2B) 608 . | . . . . 609 level #N/2 | A(2B) - A(2B) - A(2B) - A(2B) - A(2B) - ... - A(2B) 610 . | . . . . 611 level #N | A(4B) - A(4B) - A(4B) - A(4B) - A(4B) - ... - A(4B) 612 613The number of blocks and buckets are determined by:: 614 615 ,- 2, if n < MAX_DIR_HASH_DEPTH / 2, 616 # of blocks in level #n = | 617 `- 4, Otherwise 618 619 ,- 2^(n + dir_level), 620 | if n + dir_level < MAX_DIR_HASH_DEPTH / 2, 621 # of buckets in level #n = | 622 `- 2^((MAX_DIR_HASH_DEPTH / 2) - 1), 623 Otherwise 624 625When F2FS finds a file name in a directory, at first a hash value of the file 626name is calculated. Then, F2FS scans the hash table in level #0 to find the 627dentry consisting of the file name and its inode number. If not found, F2FS 628scans the next hash table in level #1. In this way, F2FS scans hash tables in 629each levels incrementally from 1 to N. In each level F2FS needs to scan only 630one bucket determined by the following equation, which shows O(log(# of files)) 631complexity:: 632 633 bucket number to scan in level #n = (hash value) % (# of buckets in level #n) 634 635In the case of file creation, F2FS finds empty consecutive slots that cover the 636file name. F2FS searches the empty slots in the hash tables of whole levels from 6371 to N in the same way as the lookup operation. 638 639The following figure shows an example of two cases holding children:: 640 641 --------------> Dir <-------------- 642 | | 643 child child 644 645 child - child [hole] - child 646 647 child - child - child [hole] - [hole] - child 648 649 Case 1: Case 2: 650 Number of children = 6, Number of children = 3, 651 File size = 7 File size = 7 652 653Default Block Allocation 654------------------------ 655 656At runtime, F2FS manages six active logs inside "Main" area: Hot/Warm/Cold node 657and Hot/Warm/Cold data. 658 659- Hot node contains direct node blocks of directories. 660- Warm node contains direct node blocks except hot node blocks. 661- Cold node contains indirect node blocks 662- Hot data contains dentry blocks 663- Warm data contains data blocks except hot and cold data blocks 664- Cold data contains multimedia data or migrated data blocks 665 666LFS has two schemes for free space management: threaded log and copy-and-compac- 667tion. The copy-and-compaction scheme which is known as cleaning, is well-suited 668for devices showing very good sequential write performance, since free segments 669are served all the time for writing new data. However, it suffers from cleaning 670overhead under high utilization. Contrarily, the threaded log scheme suffers 671from random writes, but no cleaning process is needed. F2FS adopts a hybrid 672scheme where the copy-and-compaction scheme is adopted by default, but the 673policy is dynamically changed to the threaded log scheme according to the file 674system status. 675 676In order to align F2FS with underlying flash-based storage, F2FS allocates a 677segment in a unit of section. F2FS expects that the section size would be the 678same as the unit size of garbage collection in FTL. Furthermore, with respect 679to the mapping granularity in FTL, F2FS allocates each section of the active 680logs from different zones as much as possible, since FTL can write the data in 681the active logs into one allocation unit according to its mapping granularity. 682 683Cleaning process 684---------------- 685 686F2FS does cleaning both on demand and in the background. On-demand cleaning is 687triggered when there are not enough free segments to serve VFS calls. Background 688cleaner is operated by a kernel thread, and triggers the cleaning job when the 689system is idle. 690 691F2FS supports two victim selection policies: greedy and cost-benefit algorithms. 692In the greedy algorithm, F2FS selects a victim segment having the smallest number 693of valid blocks. In the cost-benefit algorithm, F2FS selects a victim segment 694according to the segment age and the number of valid blocks in order to address 695log block thrashing problem in the greedy algorithm. F2FS adopts the greedy 696algorithm for on-demand cleaner, while background cleaner adopts cost-benefit 697algorithm. 698 699In order to identify whether the data in the victim segment are valid or not, 700F2FS manages a bitmap. Each bit represents the validity of a block, and the 701bitmap is composed of a bit stream covering whole blocks in main area. 702 703Write-hint Policy 704----------------- 705 7061) whint_mode=off. F2FS only passes down WRITE_LIFE_NOT_SET. 707 7082) whint_mode=user-based. F2FS tries to pass down hints given by 709users. 710 711===================== ======================== =================== 712User F2FS Block 713===================== ======================== =================== 714 META WRITE_LIFE_NOT_SET 715 HOT_NODE " 716 WARM_NODE " 717 COLD_NODE " 718ioctl(COLD) COLD_DATA WRITE_LIFE_EXTREME 719extension list " " 720 721-- buffered io 722WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME 723WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT 724WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_NOT_SET 725WRITE_LIFE_NONE " " 726WRITE_LIFE_MEDIUM " " 727WRITE_LIFE_LONG " " 728 729-- direct io 730WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME 731WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT 732WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_NOT_SET 733WRITE_LIFE_NONE " WRITE_LIFE_NONE 734WRITE_LIFE_MEDIUM " WRITE_LIFE_MEDIUM 735WRITE_LIFE_LONG " WRITE_LIFE_LONG 736===================== ======================== =================== 737 7383) whint_mode=fs-based. F2FS passes down hints with its policy. 739 740===================== ======================== =================== 741User F2FS Block 742===================== ======================== =================== 743 META WRITE_LIFE_MEDIUM; 744 HOT_NODE WRITE_LIFE_NOT_SET 745 WARM_NODE " 746 COLD_NODE WRITE_LIFE_NONE 747ioctl(COLD) COLD_DATA WRITE_LIFE_EXTREME 748extension list " " 749 750-- buffered io 751WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME 752WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT 753WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_LONG 754WRITE_LIFE_NONE " " 755WRITE_LIFE_MEDIUM " " 756WRITE_LIFE_LONG " " 757 758-- direct io 759WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME 760WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT 761WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_NOT_SET 762WRITE_LIFE_NONE " WRITE_LIFE_NONE 763WRITE_LIFE_MEDIUM " WRITE_LIFE_MEDIUM 764WRITE_LIFE_LONG " WRITE_LIFE_LONG 765===================== ======================== =================== 766 767Fallocate(2) Policy 768------------------- 769 770The default policy follows the below POSIX rule. 771 772Allocating disk space 773 The default operation (i.e., mode is zero) of fallocate() allocates 774 the disk space within the range specified by offset and len. The 775 file size (as reported by stat(2)) will be changed if offset+len is 776 greater than the file size. Any subregion within the range specified 777 by offset and len that did not contain data before the call will be 778 initialized to zero. This default behavior closely resembles the 779 behavior of the posix_fallocate(3) library function, and is intended 780 as a method of optimally implementing that function. 781 782However, once F2FS receives ioctl(fd, F2FS_IOC_SET_PIN_FILE) in prior to 783fallocate(fd, DEFAULT_MODE), it allocates on-disk block addressess having 784zero or random data, which is useful to the below scenario where: 785 786 1. create(fd) 787 2. ioctl(fd, F2FS_IOC_SET_PIN_FILE) 788 3. fallocate(fd, 0, 0, size) 789 4. address = fibmap(fd, offset) 790 5. open(blkdev) 791 6. write(blkdev, address) 792 793Compression implementation 794-------------------------- 795 796- New term named cluster is defined as basic unit of compression, file can 797 be divided into multiple clusters logically. One cluster includes 4 << n 798 (n >= 0) logical pages, compression size is also cluster size, each of 799 cluster can be compressed or not. 800 801- In cluster metadata layout, one special block address is used to indicate 802 a cluster is a compressed one or normal one; for compressed cluster, following 803 metadata maps cluster to [1, 4 << n - 1] physical blocks, in where f2fs 804 stores data including compress header and compressed data. 805 806- In order to eliminate write amplification during overwrite, F2FS only 807 support compression on write-once file, data can be compressed only when 808 all logical blocks in cluster contain valid data and compress ratio of 809 cluster data is lower than specified threshold. 810 811- To enable compression on regular inode, there are three ways: 812 813 * chattr +c file 814 * chattr +c dir; touch dir/file 815 * mount w/ -o compress_extension=ext; touch file.ext 816 817Compress metadata layout:: 818 819 [Dnode Structure] 820 +-----------------------------------------------+ 821 | cluster 1 | cluster 2 | ......... | cluster N | 822 +-----------------------------------------------+ 823 . . . . 824 . . . . 825 . Compressed Cluster . . Normal Cluster . 826 +----------+---------+---------+---------+ +---------+---------+---------+---------+ 827 |compr flag| block 1 | block 2 | block 3 | | block 1 | block 2 | block 3 | block 4 | 828 +----------+---------+---------+---------+ +---------+---------+---------+---------+ 829 . . 830 . . 831 . . 832 +-------------+-------------+----------+----------------------------+ 833 | data length | data chksum | reserved | compressed data | 834 +-------------+-------------+----------+----------------------------+ 835 836Compression mode 837-------------------------- 838 839f2fs supports "fs" and "user" compression modes with "compression_mode" mount option. 840With this option, f2fs provides a choice to select the way how to compress the 841compression enabled files (refer to "Compression implementation" section for how to 842enable compression on a regular inode). 843 8441) compress_mode=fs 845This is the default option. f2fs does automatic compression in the writeback of the 846compression enabled files. 847 8482) compress_mode=user 849This disables the automatic compression and gives the user discretion of choosing the 850target file and the timing. The user can do manual compression/decompression on the 851compression enabled files using F2FS_IOC_DECOMPRESS_FILE and F2FS_IOC_COMPRESS_FILE 852ioctls like the below. 853 854To decompress a file, 855 856fd = open(filename, O_WRONLY, 0); 857ret = ioctl(fd, F2FS_IOC_DECOMPRESS_FILE); 858 859To compress a file, 860 861fd = open(filename, O_WRONLY, 0); 862ret = ioctl(fd, F2FS_IOC_COMPRESS_FILE); 863 864NVMe Zoned Namespace devices 865---------------------------- 866 867- ZNS defines a per-zone capacity which can be equal or less than the 868 zone-size. Zone-capacity is the number of usable blocks in the zone. 869 F2FS checks if zone-capacity is less than zone-size, if it is, then any 870 segment which starts after the zone-capacity is marked as not-free in 871 the free segment bitmap at initial mount time. These segments are marked 872 as permanently used so they are not allocated for writes and 873 consequently are not needed to be garbage collected. In case the 874 zone-capacity is not aligned to default segment size(2MB), then a segment 875 can start before the zone-capacity and span across zone-capacity boundary. 876 Such spanning segments are also considered as usable segments. All blocks 877 past the zone-capacity are considered unusable in these segments. 878