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. 113gc_merge When background_gc is on, this option can be enabled to 114 let background GC thread to handle foreground GC requests, 115 it can eliminate the sluggish issue caused by slow foreground 116 GC operation when GC is triggered from a process with limited 117 I/O and CPU resources. 118nogc_merge Disable GC merge feature. 119disable_roll_forward Disable the roll-forward recovery routine 120norecovery Disable the roll-forward recovery routine, mounted read- 121 only (i.e., -o ro,disable_roll_forward) 122discard/nodiscard Enable/disable real-time discard in f2fs, if discard is 123 enabled, f2fs will issue discard/TRIM commands when a 124 segment is cleaned. 125no_heap Disable heap-style segment allocation which finds free 126 segments for data from the beginning of main area, while 127 for node from the end of main area. 128nouser_xattr Disable Extended User Attributes. Note: xattr is enabled 129 by default if CONFIG_F2FS_FS_XATTR is selected. 130noacl Disable POSIX Access Control List. Note: acl is enabled 131 by default if CONFIG_F2FS_FS_POSIX_ACL is selected. 132active_logs=%u Support configuring the number of active logs. In the 133 current design, f2fs supports only 2, 4, and 6 logs. 134 Default number is 6. 135disable_ext_identify Disable the extension list configured by mkfs, so f2fs 136 is not aware of cold files such as media files. 137inline_xattr Enable the inline xattrs feature. 138noinline_xattr Disable the inline xattrs feature. 139inline_xattr_size=%u Support configuring inline xattr size, it depends on 140 flexible inline xattr feature. 141inline_data Enable the inline data feature: Newly created small (<~3.4k) 142 files can be written into inode block. 143inline_dentry Enable the inline dir feature: data in newly created 144 directory entries can be written into inode block. The 145 space of inode block which is used to store inline 146 dentries is limited to ~3.4k. 147noinline_dentry Disable the inline dentry feature. 148flush_merge Merge concurrent cache_flush commands as much as possible 149 to eliminate redundant command issues. If the underlying 150 device handles the cache_flush command relatively slowly, 151 recommend to enable this option. 152nobarrier This option can be used if underlying storage guarantees 153 its cached data should be written to the novolatile area. 154 If this option is set, no cache_flush commands are issued 155 but f2fs still guarantees the write ordering of all the 156 data writes. 157fastboot This option is used when a system wants to reduce mount 158 time as much as possible, even though normal performance 159 can be sacrificed. 160extent_cache Enable an extent cache based on rb-tree, it can cache 161 as many as extent which map between contiguous logical 162 address and physical address per inode, resulting in 163 increasing the cache hit ratio. Set by default. 164noextent_cache Disable an extent cache based on rb-tree explicitly, see 165 the above extent_cache mount option. 166noinline_data Disable the inline data feature, inline data feature is 167 enabled by default. 168data_flush Enable data flushing before checkpoint in order to 169 persist data of regular and symlink. 170reserve_root=%d Support configuring reserved space which is used for 171 allocation from a privileged user with specified uid or 172 gid, unit: 4KB, the default limit is 0.2% of user blocks. 173resuid=%d The user ID which may use the reserved blocks. 174resgid=%d The group ID which may use the reserved blocks. 175fault_injection=%d Enable fault injection in all supported types with 176 specified injection rate. 177fault_type=%d Support configuring fault injection type, should be 178 enabled with fault_injection option, fault type value 179 is shown below, it supports single or combined type. 180 181 =================== =========== 182 Type_Name Type_Value 183 =================== =========== 184 FAULT_KMALLOC 0x000000001 185 FAULT_KVMALLOC 0x000000002 186 FAULT_PAGE_ALLOC 0x000000004 187 FAULT_PAGE_GET 0x000000008 188 FAULT_ALLOC_NID 0x000000020 189 FAULT_ORPHAN 0x000000040 190 FAULT_BLOCK 0x000000080 191 FAULT_DIR_DEPTH 0x000000100 192 FAULT_EVICT_INODE 0x000000200 193 FAULT_TRUNCATE 0x000000400 194 FAULT_READ_IO 0x000000800 195 FAULT_CHECKPOINT 0x000001000 196 FAULT_DISCARD 0x000002000 197 FAULT_WRITE_IO 0x000004000 198 =================== =========== 199mode=%s Control block allocation mode which supports "adaptive" 200 and "lfs". In "lfs" mode, there should be no random 201 writes towards main area. 202io_bits=%u Set the bit size of write IO requests. It should be set 203 with "mode=lfs". 204usrquota Enable plain user disk quota accounting. 205grpquota Enable plain group disk quota accounting. 206prjquota Enable plain project quota accounting. 207usrjquota=<file> Appoint specified file and type during mount, so that quota 208grpjquota=<file> information can be properly updated during recovery flow, 209prjjquota=<file> <quota file>: must be in root directory; 210jqfmt=<quota type> <quota type>: [vfsold,vfsv0,vfsv1]. 211offusrjquota Turn off user journalled quota. 212offgrpjquota Turn off group journalled quota. 213offprjjquota Turn off project journalled quota. 214quota Enable plain user disk quota accounting. 215noquota Disable all plain disk quota option. 216whint_mode=%s Control which write hints are passed down to block 217 layer. This supports "off", "user-based", and 218 "fs-based". In "off" mode (default), f2fs does not pass 219 down hints. In "user-based" mode, f2fs tries to pass 220 down hints given by users. And in "fs-based" mode, f2fs 221 passes down hints with its policy. 222alloc_mode=%s Adjust block allocation policy, which supports "reuse" 223 and "default". 224fsync_mode=%s Control the policy of fsync. Currently supports "posix", 225 "strict", and "nobarrier". In "posix" mode, which is 226 default, fsync will follow POSIX semantics and does a 227 light operation to improve the filesystem performance. 228 In "strict" mode, fsync will be heavy and behaves in line 229 with xfs, ext4 and btrfs, where xfstest generic/342 will 230 pass, but the performance will regress. "nobarrier" is 231 based on "posix", but doesn't issue flush command for 232 non-atomic files likewise "nobarrier" mount option. 233test_dummy_encryption 234test_dummy_encryption=%s 235 Enable dummy encryption, which provides a fake fscrypt 236 context. The fake fscrypt context is used by xfstests. 237 The argument may be either "v1" or "v2", in order to 238 select the corresponding fscrypt policy version. 239checkpoint=%s[:%u[%]] Set to "disable" to turn off checkpointing. Set to "enable" 240 to reenable checkpointing. Is enabled by default. While 241 disabled, any unmounting or unexpected shutdowns will cause 242 the filesystem contents to appear as they did when the 243 filesystem was mounted with that option. 244 While mounting with checkpoint=disabled, the filesystem must 245 run garbage collection to ensure that all available space can 246 be used. If this takes too much time, the mount may return 247 EAGAIN. You may optionally add a value to indicate how much 248 of the disk you would be willing to temporarily give up to 249 avoid additional garbage collection. This can be given as a 250 number of blocks, or as a percent. For instance, mounting 251 with checkpoint=disable:100% would always succeed, but it may 252 hide up to all remaining free space. The actual space that 253 would be unusable can be viewed at /sys/fs/f2fs/<disk>/unusable 254 This space is reclaimed once checkpoint=enable. 255checkpoint_merge When checkpoint is enabled, this can be used to create a kernel 256 daemon and make it to merge concurrent checkpoint requests as 257 much as possible to eliminate redundant checkpoint issues. Plus, 258 we can eliminate the sluggish issue caused by slow checkpoint 259 operation when the checkpoint is done in a process context in 260 a cgroup having low i/o budget and cpu shares. To make this 261 do better, we set the default i/o priority of the kernel daemon 262 to "3", to give one higher priority than other kernel threads. 263 This is the same way to give a I/O priority to the jbd2 264 journaling thread of ext4 filesystem. 265nocheckpoint_merge Disable checkpoint merge feature. 266compress_algorithm=%s Control compress algorithm, currently f2fs supports "lzo", 267 "lz4", "zstd" and "lzo-rle" algorithm. 268compress_algorithm=%s:%d Control compress algorithm and its compress level, now, only 269 "lz4" and "zstd" support compress level config. 270 algorithm level range 271 lz4 3 - 16 272 zstd 1 - 22 273compress_log_size=%u Support configuring compress cluster size, the size will 274 be 4KB * (1 << %u), 16KB is minimum size, also it's 275 default size. 276compress_extension=%s Support adding specified extension, so that f2fs can enable 277 compression on those corresponding files, e.g. if all files 278 with '.ext' has high compression rate, we can set the '.ext' 279 on compression extension list and enable compression on 280 these file by default rather than to enable it via ioctl. 281 For other files, we can still enable compression via ioctl. 282 Note that, there is one reserved special extension '*', it 283 can be set to enable compression for all files. 284compress_chksum Support verifying chksum of raw data in compressed cluster. 285compress_mode=%s Control file compression mode. This supports "fs" and "user" 286 modes. In "fs" mode (default), f2fs does automatic compression 287 on the compression enabled files. In "user" mode, f2fs disables 288 the automaic compression and gives the user discretion of 289 choosing the target file and the timing. The user can do manual 290 compression/decompression on the compression enabled files using 291 ioctls. 292compress_cache Support to use address space of a filesystem managed inode to 293 cache compressed block, in order to improve cache hit ratio of 294 random read. 295inlinecrypt When possible, encrypt/decrypt the contents of encrypted 296 files using the blk-crypto framework rather than 297 filesystem-layer encryption. This allows the use of 298 inline encryption hardware. The on-disk format is 299 unaffected. For more details, see 300 Documentation/block/inline-encryption.rst. 301atgc Enable age-threshold garbage collection, it provides high 302 effectiveness and efficiency on background GC. 303memory=%s Control memory mode. This supports "normal" and "low" modes. 304 "low" mode is introduced to support low memory devices. 305 Because of the nature of low memory devices, in this mode, f2fs 306 will try to save memory sometimes by sacrificing performance. 307 "normal" mode is the default mode and same as before. 308age_extent_cache Enable an age extent cache based on rb-tree. It records 309 data block update frequency of the extent per inode, in 310 order to provide better temperature hints for data block 311 allocation. 312======================== ============================================================ 313 314Debugfs Entries 315=============== 316 317/sys/kernel/debug/f2fs/ contains information about all the partitions mounted as 318f2fs. Each file shows the whole f2fs information. 319 320/sys/kernel/debug/f2fs/status includes: 321 322 - major file system information managed by f2fs currently 323 - average SIT information about whole segments 324 - current memory footprint consumed by f2fs. 325 326Sysfs Entries 327============= 328 329Information about mounted f2fs file systems can be found in 330/sys/fs/f2fs. Each mounted filesystem will have a directory in 331/sys/fs/f2fs based on its device name (i.e., /sys/fs/f2fs/sda). 332The files in each per-device directory are shown in table below. 333 334Files in /sys/fs/f2fs/<devname> 335(see also Documentation/ABI/testing/sysfs-fs-f2fs) 336 337Usage 338===== 339 3401. Download userland tools and compile them. 341 3422. Skip, if f2fs was compiled statically inside kernel. 343 Otherwise, insert the f2fs.ko module:: 344 345 # insmod f2fs.ko 346 3473. Create a directory to use when mounting:: 348 349 # mkdir /mnt/f2fs 350 3514. Format the block device, and then mount as f2fs:: 352 353 # mkfs.f2fs -l label /dev/block_device 354 # mount -t f2fs /dev/block_device /mnt/f2fs 355 356mkfs.f2fs 357--------- 358The mkfs.f2fs is for the use of formatting a partition as the f2fs filesystem, 359which builds a basic on-disk layout. 360 361The quick options consist of: 362 363=============== =========================================================== 364``-l [label]`` Give a volume label, up to 512 unicode name. 365``-a [0 or 1]`` Split start location of each area for heap-based allocation. 366 367 1 is set by default, which performs this. 368``-o [int]`` Set overprovision ratio in percent over volume size. 369 370 5 is set by default. 371``-s [int]`` Set the number of segments per section. 372 373 1 is set by default. 374``-z [int]`` Set the number of sections per zone. 375 376 1 is set by default. 377``-e [str]`` Set basic extension list. e.g. "mp3,gif,mov" 378``-t [0 or 1]`` Disable discard command or not. 379 380 1 is set by default, which conducts discard. 381=============== =========================================================== 382 383Note: please refer to the manpage of mkfs.f2fs(8) to get full option list. 384 385fsck.f2fs 386--------- 387The fsck.f2fs is a tool to check the consistency of an f2fs-formatted 388partition, which examines whether the filesystem metadata and user-made data 389are cross-referenced correctly or not. 390Note that, initial version of the tool does not fix any inconsistency. 391 392The quick options consist of:: 393 394 -d debug level [default:0] 395 396Note: please refer to the manpage of fsck.f2fs(8) to get full option list. 397 398dump.f2fs 399--------- 400The dump.f2fs shows the information of specific inode and dumps SSA and SIT to 401file. Each file is dump_ssa and dump_sit. 402 403The dump.f2fs is used to debug on-disk data structures of the f2fs filesystem. 404It shows on-disk inode information recognized by a given inode number, and is 405able to dump all the SSA and SIT entries into predefined files, ./dump_ssa and 406./dump_sit respectively. 407 408The options consist of:: 409 410 -d debug level [default:0] 411 -i inode no (hex) 412 -s [SIT dump segno from #1~#2 (decimal), for all 0~-1] 413 -a [SSA dump segno from #1~#2 (decimal), for all 0~-1] 414 415Examples:: 416 417 # dump.f2fs -i [ino] /dev/sdx 418 # dump.f2fs -s 0~-1 /dev/sdx (SIT dump) 419 # dump.f2fs -a 0~-1 /dev/sdx (SSA dump) 420 421Note: please refer to the manpage of dump.f2fs(8) to get full option list. 422 423sload.f2fs 424---------- 425The sload.f2fs gives a way to insert files and directories in the exisiting disk 426image. This tool is useful when building f2fs images given compiled files. 427 428Note: please refer to the manpage of sload.f2fs(8) to get full option list. 429 430resize.f2fs 431----------- 432The resize.f2fs lets a user resize the f2fs-formatted disk image, while preserving 433all the files and directories stored in the image. 434 435Note: please refer to the manpage of resize.f2fs(8) to get full option list. 436 437defrag.f2fs 438----------- 439The defrag.f2fs can be used to defragment scattered written data as well as 440filesystem metadata across the disk. This can improve the write speed by giving 441more free consecutive space. 442 443Note: please refer to the manpage of defrag.f2fs(8) to get full option list. 444 445f2fs_io 446------- 447The f2fs_io is a simple tool to issue various filesystem APIs as well as 448f2fs-specific ones, which is very useful for QA tests. 449 450Note: please refer to the manpage of f2fs_io(8) to get full option list. 451 452Design 453====== 454 455On-disk Layout 456-------------- 457 458F2FS divides the whole volume into a number of segments, each of which is fixed 459to 2MB in size. A section is composed of consecutive segments, and a zone 460consists of a set of sections. By default, section and zone sizes are set to one 461segment size identically, but users can easily modify the sizes by mkfs. 462 463F2FS splits the entire volume into six areas, and all the areas except superblock 464consist of multiple segments as described below:: 465 466 align with the zone size <-| 467 |-> align with the segment size 468 _________________________________________________________________________ 469 | | | Segment | Node | Segment | | 470 | Superblock | Checkpoint | Info. | Address | Summary | Main | 471 | (SB) | (CP) | Table (SIT) | Table (NAT) | Area (SSA) | | 472 |____________|_____2______|______N______|______N______|______N_____|__N___| 473 . . 474 . . 475 . . 476 ._________________________________________. 477 |_Segment_|_..._|_Segment_|_..._|_Segment_| 478 . . 479 ._________._________ 480 |_section_|__...__|_ 481 . . 482 .________. 483 |__zone__| 484 485- Superblock (SB) 486 It is located at the beginning of the partition, and there exist two copies 487 to avoid file system crash. It contains basic partition information and some 488 default parameters of f2fs. 489 490- Checkpoint (CP) 491 It contains file system information, bitmaps for valid NAT/SIT sets, orphan 492 inode lists, and summary entries of current active segments. 493 494- Segment Information Table (SIT) 495 It contains segment information such as valid block count and bitmap for the 496 validity of all the blocks. 497 498- Node Address Table (NAT) 499 It is composed of a block address table for all the node blocks stored in 500 Main area. 501 502- Segment Summary Area (SSA) 503 It contains summary entries which contains the owner information of all the 504 data and node blocks stored in Main area. 505 506- Main Area 507 It contains file and directory data including their indices. 508 509In order to avoid misalignment between file system and flash-based storage, F2FS 510aligns the start block address of CP with the segment size. Also, it aligns the 511start block address of Main area with the zone size by reserving some segments 512in SSA area. 513 514Reference the following survey for additional technical details. 515https://wiki.linaro.org/WorkingGroups/Kernel/Projects/FlashCardSurvey 516 517File System Metadata Structure 518------------------------------ 519 520F2FS adopts the checkpointing scheme to maintain file system consistency. At 521mount time, F2FS first tries to find the last valid checkpoint data by scanning 522CP area. In order to reduce the scanning time, F2FS uses only two copies of CP. 523One of them always indicates the last valid data, which is called as shadow copy 524mechanism. In addition to CP, NAT and SIT also adopt the shadow copy mechanism. 525 526For file system consistency, each CP points to which NAT and SIT copies are 527valid, as shown as below:: 528 529 +--------+----------+---------+ 530 | CP | SIT | NAT | 531 +--------+----------+---------+ 532 . . . . 533 . . . . 534 . . . . 535 +-------+-------+--------+--------+--------+--------+ 536 | CP #0 | CP #1 | SIT #0 | SIT #1 | NAT #0 | NAT #1 | 537 +-------+-------+--------+--------+--------+--------+ 538 | ^ ^ 539 | | | 540 `----------------------------------------' 541 542Index Structure 543--------------- 544 545The key data structure to manage the data locations is a "node". Similar to 546traditional file structures, F2FS has three types of node: inode, direct node, 547indirect node. F2FS assigns 4KB to an inode block which contains 923 data block 548indices, two direct node pointers, two indirect node pointers, and one double 549indirect node pointer as described below. One direct node block contains 1018 550data blocks, and one indirect node block contains also 1018 node blocks. Thus, 551one inode block (i.e., a file) covers:: 552 553 4KB * (923 + 2 * 1018 + 2 * 1018 * 1018 + 1018 * 1018 * 1018) := 3.94TB. 554 555 Inode block (4KB) 556 |- data (923) 557 |- direct node (2) 558 | `- data (1018) 559 |- indirect node (2) 560 | `- direct node (1018) 561 | `- data (1018) 562 `- double indirect node (1) 563 `- indirect node (1018) 564 `- direct node (1018) 565 `- data (1018) 566 567Note that all the node blocks are mapped by NAT which means the location of 568each node is translated by the NAT table. In the consideration of the wandering 569tree problem, F2FS is able to cut off the propagation of node updates caused by 570leaf data writes. 571 572Directory Structure 573------------------- 574 575A directory entry occupies 11 bytes, which consists of the following attributes. 576 577- hash hash value of the file name 578- ino inode number 579- len the length of file name 580- type file type such as directory, symlink, etc 581 582A dentry block consists of 214 dentry slots and file names. Therein a bitmap is 583used to represent whether each dentry is valid or not. A dentry block occupies 5844KB with the following composition. 585 586:: 587 588 Dentry Block(4 K) = bitmap (27 bytes) + reserved (3 bytes) + 589 dentries(11 * 214 bytes) + file name (8 * 214 bytes) 590 591 [Bucket] 592 +--------------------------------+ 593 |dentry block 1 | dentry block 2 | 594 +--------------------------------+ 595 . . 596 . . 597 . [Dentry Block Structure: 4KB] . 598 +--------+----------+----------+------------+ 599 | bitmap | reserved | dentries | file names | 600 +--------+----------+----------+------------+ 601 [Dentry Block: 4KB] . . 602 . . 603 . . 604 +------+------+-----+------+ 605 | hash | ino | len | type | 606 +------+------+-----+------+ 607 [Dentry Structure: 11 bytes] 608 609F2FS implements multi-level hash tables for directory structure. Each level has 610a hash table with dedicated number of hash buckets as shown below. Note that 611"A(2B)" means a bucket includes 2 data blocks. 612 613:: 614 615 ---------------------- 616 A : bucket 617 B : block 618 N : MAX_DIR_HASH_DEPTH 619 ---------------------- 620 621 level #0 | A(2B) 622 | 623 level #1 | A(2B) - A(2B) 624 | 625 level #2 | A(2B) - A(2B) - A(2B) - A(2B) 626 . | . . . . 627 level #N/2 | A(2B) - A(2B) - A(2B) - A(2B) - A(2B) - ... - A(2B) 628 . | . . . . 629 level #N | A(4B) - A(4B) - A(4B) - A(4B) - A(4B) - ... - A(4B) 630 631The number of blocks and buckets are determined by:: 632 633 ,- 2, if n < MAX_DIR_HASH_DEPTH / 2, 634 # of blocks in level #n = | 635 `- 4, Otherwise 636 637 ,- 2^(n + dir_level), 638 | if n + dir_level < MAX_DIR_HASH_DEPTH / 2, 639 # of buckets in level #n = | 640 `- 2^((MAX_DIR_HASH_DEPTH / 2) - 1), 641 Otherwise 642 643When F2FS finds a file name in a directory, at first a hash value of the file 644name is calculated. Then, F2FS scans the hash table in level #0 to find the 645dentry consisting of the file name and its inode number. If not found, F2FS 646scans the next hash table in level #1. In this way, F2FS scans hash tables in 647each levels incrementally from 1 to N. In each level F2FS needs to scan only 648one bucket determined by the following equation, which shows O(log(# of files)) 649complexity:: 650 651 bucket number to scan in level #n = (hash value) % (# of buckets in level #n) 652 653In the case of file creation, F2FS finds empty consecutive slots that cover the 654file name. F2FS searches the empty slots in the hash tables of whole levels from 6551 to N in the same way as the lookup operation. 656 657The following figure shows an example of two cases holding children:: 658 659 --------------> Dir <-------------- 660 | | 661 child child 662 663 child - child [hole] - child 664 665 child - child - child [hole] - [hole] - child 666 667 Case 1: Case 2: 668 Number of children = 6, Number of children = 3, 669 File size = 7 File size = 7 670 671Default Block Allocation 672------------------------ 673 674At runtime, F2FS manages six active logs inside "Main" area: Hot/Warm/Cold node 675and Hot/Warm/Cold data. 676 677- Hot node contains direct node blocks of directories. 678- Warm node contains direct node blocks except hot node blocks. 679- Cold node contains indirect node blocks 680- Hot data contains dentry blocks 681- Warm data contains data blocks except hot and cold data blocks 682- Cold data contains multimedia data or migrated data blocks 683 684LFS has two schemes for free space management: threaded log and copy-and-compac- 685tion. The copy-and-compaction scheme which is known as cleaning, is well-suited 686for devices showing very good sequential write performance, since free segments 687are served all the time for writing new data. However, it suffers from cleaning 688overhead under high utilization. Contrarily, the threaded log scheme suffers 689from random writes, but no cleaning process is needed. F2FS adopts a hybrid 690scheme where the copy-and-compaction scheme is adopted by default, but the 691policy is dynamically changed to the threaded log scheme according to the file 692system status. 693 694In order to align F2FS with underlying flash-based storage, F2FS allocates a 695segment in a unit of section. F2FS expects that the section size would be the 696same as the unit size of garbage collection in FTL. Furthermore, with respect 697to the mapping granularity in FTL, F2FS allocates each section of the active 698logs from different zones as much as possible, since FTL can write the data in 699the active logs into one allocation unit according to its mapping granularity. 700 701Cleaning process 702---------------- 703 704F2FS does cleaning both on demand and in the background. On-demand cleaning is 705triggered when there are not enough free segments to serve VFS calls. Background 706cleaner is operated by a kernel thread, and triggers the cleaning job when the 707system is idle. 708 709F2FS supports two victim selection policies: greedy and cost-benefit algorithms. 710In the greedy algorithm, F2FS selects a victim segment having the smallest number 711of valid blocks. In the cost-benefit algorithm, F2FS selects a victim segment 712according to the segment age and the number of valid blocks in order to address 713log block thrashing problem in the greedy algorithm. F2FS adopts the greedy 714algorithm for on-demand cleaner, while background cleaner adopts cost-benefit 715algorithm. 716 717In order to identify whether the data in the victim segment are valid or not, 718F2FS manages a bitmap. Each bit represents the validity of a block, and the 719bitmap is composed of a bit stream covering whole blocks in main area. 720 721Write-hint Policy 722----------------- 723 7241) whint_mode=off. F2FS only passes down WRITE_LIFE_NOT_SET. 725 7262) whint_mode=user-based. F2FS tries to pass down hints given by 727users. 728 729===================== ======================== =================== 730User F2FS Block 731===================== ======================== =================== 732N/A META WRITE_LIFE_NOT_SET 733N/A HOT_NODE " 734N/A WARM_NODE " 735N/A COLD_NODE " 736ioctl(COLD) COLD_DATA WRITE_LIFE_EXTREME 737extension list " " 738 739-- buffered io 740WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME 741WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT 742WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_NOT_SET 743WRITE_LIFE_NONE " " 744WRITE_LIFE_MEDIUM " " 745WRITE_LIFE_LONG " " 746 747-- direct io 748WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME 749WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT 750WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_NOT_SET 751WRITE_LIFE_NONE " WRITE_LIFE_NONE 752WRITE_LIFE_MEDIUM " WRITE_LIFE_MEDIUM 753WRITE_LIFE_LONG " WRITE_LIFE_LONG 754===================== ======================== =================== 755 7563) whint_mode=fs-based. F2FS passes down hints with its policy. 757 758===================== ======================== =================== 759User F2FS Block 760===================== ======================== =================== 761N/A META WRITE_LIFE_MEDIUM; 762N/A HOT_NODE WRITE_LIFE_NOT_SET 763N/A WARM_NODE " 764N/A COLD_NODE WRITE_LIFE_NONE 765ioctl(COLD) COLD_DATA WRITE_LIFE_EXTREME 766extension list " " 767 768-- buffered io 769WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME 770WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT 771WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_LONG 772WRITE_LIFE_NONE " " 773WRITE_LIFE_MEDIUM " " 774WRITE_LIFE_LONG " " 775 776-- direct io 777WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME 778WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT 779WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_NOT_SET 780WRITE_LIFE_NONE " WRITE_LIFE_NONE 781WRITE_LIFE_MEDIUM " WRITE_LIFE_MEDIUM 782WRITE_LIFE_LONG " WRITE_LIFE_LONG 783===================== ======================== =================== 784 785Fallocate(2) Policy 786------------------- 787 788The default policy follows the below POSIX rule. 789 790Allocating disk space 791 The default operation (i.e., mode is zero) of fallocate() allocates 792 the disk space within the range specified by offset and len. The 793 file size (as reported by stat(2)) will be changed if offset+len is 794 greater than the file size. Any subregion within the range specified 795 by offset and len that did not contain data before the call will be 796 initialized to zero. This default behavior closely resembles the 797 behavior of the posix_fallocate(3) library function, and is intended 798 as a method of optimally implementing that function. 799 800However, once F2FS receives ioctl(fd, F2FS_IOC_SET_PIN_FILE) in prior to 801fallocate(fd, DEFAULT_MODE), it allocates on-disk block addressess having 802zero or random data, which is useful to the below scenario where: 803 804 1. create(fd) 805 2. ioctl(fd, F2FS_IOC_SET_PIN_FILE) 806 3. fallocate(fd, 0, 0, size) 807 4. address = fibmap(fd, offset) 808 5. open(blkdev) 809 6. write(blkdev, address) 810 811Compression implementation 812-------------------------- 813 814- New term named cluster is defined as basic unit of compression, file can 815 be divided into multiple clusters logically. One cluster includes 4 << n 816 (n >= 0) logical pages, compression size is also cluster size, each of 817 cluster can be compressed or not. 818 819- In cluster metadata layout, one special block address is used to indicate 820 a cluster is a compressed one or normal one; for compressed cluster, following 821 metadata maps cluster to [1, 4 << n - 1] physical blocks, in where f2fs 822 stores data including compress header and compressed data. 823 824- In order to eliminate write amplification during overwrite, F2FS only 825 support compression on write-once file, data can be compressed only when 826 all logical blocks in cluster contain valid data and compress ratio of 827 cluster data is lower than specified threshold. 828 829- To enable compression on regular inode, there are three ways: 830 831 * chattr +c file 832 * chattr +c dir; touch dir/file 833 * mount w/ -o compress_extension=ext; touch file.ext 834 * mount w/ -o compress_extension=*; touch any_file 835 836- At this point, compression feature doesn't expose compressed space to user 837 directly in order to guarantee potential data updates later to the space. 838 Instead, the main goal is to reduce data writes to flash disk as much as 839 possible, resulting in extending disk life time as well as relaxing IO 840 congestion. Alternatively, we've added ioctl interface to reclaim compressed 841 space and show it to user after putting the immutable bit. 842 843Compress metadata layout:: 844 845 [Dnode Structure] 846 +-----------------------------------------------+ 847 | cluster 1 | cluster 2 | ......... | cluster N | 848 +-----------------------------------------------+ 849 . . . . 850 . . . . 851 . Compressed Cluster . . Normal Cluster . 852 +----------+---------+---------+---------+ +---------+---------+---------+---------+ 853 |compr flag| block 1 | block 2 | block 3 | | block 1 | block 2 | block 3 | block 4 | 854 +----------+---------+---------+---------+ +---------+---------+---------+---------+ 855 . . 856 . . 857 . . 858 +-------------+-------------+----------+----------------------------+ 859 | data length | data chksum | reserved | compressed data | 860 +-------------+-------------+----------+----------------------------+ 861 862Compression mode 863-------------------------- 864 865f2fs supports "fs" and "user" compression modes with "compression_mode" mount option. 866With this option, f2fs provides a choice to select the way how to compress the 867compression enabled files (refer to "Compression implementation" section for how to 868enable compression on a regular inode). 869 8701) compress_mode=fs 871This is the default option. f2fs does automatic compression in the writeback of the 872compression enabled files. 873 8742) compress_mode=user 875This disables the automatic compression and gives the user discretion of choosing the 876target file and the timing. The user can do manual compression/decompression on the 877compression enabled files using F2FS_IOC_DECOMPRESS_FILE and F2FS_IOC_COMPRESS_FILE 878ioctls like the below. 879 880To decompress a file, 881 882fd = open(filename, O_WRONLY, 0); 883ret = ioctl(fd, F2FS_IOC_DECOMPRESS_FILE); 884 885To compress a file, 886 887fd = open(filename, O_WRONLY, 0); 888ret = ioctl(fd, F2FS_IOC_COMPRESS_FILE); 889 890NVMe Zoned Namespace devices 891---------------------------- 892 893- ZNS defines a per-zone capacity which can be equal or less than the 894 zone-size. Zone-capacity is the number of usable blocks in the zone. 895 F2FS checks if zone-capacity is less than zone-size, if it is, then any 896 segment which starts after the zone-capacity is marked as not-free in 897 the free segment bitmap at initial mount time. These segments are marked 898 as permanently used so they are not allocated for writes and 899 consequently are not needed to be garbage collected. In case the 900 zone-capacity is not aligned to default segment size(2MB), then a segment 901 can start before the zone-capacity and span across zone-capacity boundary. 902 Such spanning segments are also considered as usable segments. All blocks 903 past the zone-capacity are considered unusable in these segments. 904