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_BIO 0x000000010 (obsolete) 189 FAULT_ALLOC_NID 0x000000020 190 FAULT_ORPHAN 0x000000040 191 FAULT_BLOCK 0x000000080 192 FAULT_DIR_DEPTH 0x000000100 193 FAULT_EVICT_INODE 0x000000200 194 FAULT_TRUNCATE 0x000000400 195 FAULT_READ_IO 0x000000800 196 FAULT_CHECKPOINT 0x000001000 197 FAULT_DISCARD 0x000002000 198 FAULT_WRITE_IO 0x000004000 199 FAULT_SLAB_ALLOC 0x000008000 200 FAULT_DQUOT_INIT 0x000010000 201 FAULT_LOCK_OP 0x000020000 202 =================== =========== 203mode=%s Control block allocation mode which supports "adaptive" 204 and "lfs". In "lfs" mode, there should be no random 205 writes towards main area. 206 "fragment:segment" and "fragment:block" are newly added here. 207 These are developer options for experiments to simulate filesystem 208 fragmentation/after-GC situation itself. The developers use these 209 modes to understand filesystem fragmentation/after-GC condition well, 210 and eventually get some insights to handle them better. 211 In "fragment:segment", f2fs allocates a new segment in ramdom 212 position. With this, we can simulate the after-GC condition. 213 In "fragment:block", we can scatter block allocation with 214 "max_fragment_chunk" and "max_fragment_hole" sysfs nodes. 215 We added some randomness to both chunk and hole size to make 216 it close to realistic IO pattern. So, in this mode, f2fs will allocate 217 1..<max_fragment_chunk> blocks in a chunk and make a hole in the 218 length of 1..<max_fragment_hole> by turns. With this, the newly 219 allocated blocks will be scattered throughout the whole partition. 220 Note that "fragment:block" implicitly enables "fragment:segment" 221 option for more randomness. 222 Please, use these options for your experiments and we strongly 223 recommend to re-format the filesystem after using these options. 224io_bits=%u Set the bit size of write IO requests. It should be set 225 with "mode=lfs". 226usrquota Enable plain user disk quota accounting. 227grpquota Enable plain group disk quota accounting. 228prjquota Enable plain project quota accounting. 229usrjquota=<file> Appoint specified file and type during mount, so that quota 230grpjquota=<file> information can be properly updated during recovery flow, 231prjjquota=<file> <quota file>: must be in root directory; 232jqfmt=<quota type> <quota type>: [vfsold,vfsv0,vfsv1]. 233offusrjquota Turn off user journalled quota. 234offgrpjquota Turn off group journalled quota. 235offprjjquota Turn off project journalled quota. 236quota Enable plain user disk quota accounting. 237noquota Disable all plain disk quota option. 238alloc_mode=%s Adjust block allocation policy, which supports "reuse" 239 and "default". 240fsync_mode=%s Control the policy of fsync. Currently supports "posix", 241 "strict", and "nobarrier". In "posix" mode, which is 242 default, fsync will follow POSIX semantics and does a 243 light operation to improve the filesystem performance. 244 In "strict" mode, fsync will be heavy and behaves in line 245 with xfs, ext4 and btrfs, where xfstest generic/342 will 246 pass, but the performance will regress. "nobarrier" is 247 based on "posix", but doesn't issue flush command for 248 non-atomic files likewise "nobarrier" mount option. 249test_dummy_encryption 250test_dummy_encryption=%s 251 Enable dummy encryption, which provides a fake fscrypt 252 context. The fake fscrypt context is used by xfstests. 253 The argument may be either "v1" or "v2", in order to 254 select the corresponding fscrypt policy version. 255checkpoint=%s[:%u[%]] Set to "disable" to turn off checkpointing. Set to "enable" 256 to reenable checkpointing. Is enabled by default. While 257 disabled, any unmounting or unexpected shutdowns will cause 258 the filesystem contents to appear as they did when the 259 filesystem was mounted with that option. 260 While mounting with checkpoint=disabled, the filesystem must 261 run garbage collection to ensure that all available space can 262 be used. If this takes too much time, the mount may return 263 EAGAIN. You may optionally add a value to indicate how much 264 of the disk you would be willing to temporarily give up to 265 avoid additional garbage collection. This can be given as a 266 number of blocks, or as a percent. For instance, mounting 267 with checkpoint=disable:100% would always succeed, but it may 268 hide up to all remaining free space. The actual space that 269 would be unusable can be viewed at /sys/fs/f2fs/<disk>/unusable 270 This space is reclaimed once checkpoint=enable. 271checkpoint_merge When checkpoint is enabled, this can be used to create a kernel 272 daemon and make it to merge concurrent checkpoint requests as 273 much as possible to eliminate redundant checkpoint issues. Plus, 274 we can eliminate the sluggish issue caused by slow checkpoint 275 operation when the checkpoint is done in a process context in 276 a cgroup having low i/o budget and cpu shares. To make this 277 do better, we set the default i/o priority of the kernel daemon 278 to "3", to give one higher priority than other kernel threads. 279 This is the same way to give a I/O priority to the jbd2 280 journaling thread of ext4 filesystem. 281nocheckpoint_merge Disable checkpoint merge feature. 282compress_algorithm=%s Control compress algorithm, currently f2fs supports "lzo", 283 "lz4", "zstd" and "lzo-rle" algorithm. 284compress_algorithm=%s:%d Control compress algorithm and its compress level, now, only 285 "lz4" and "zstd" support compress level config. 286 algorithm level range 287 lz4 3 - 16 288 zstd 1 - 22 289compress_log_size=%u Support configuring compress cluster size, the size will 290 be 4KB * (1 << %u), 16KB is minimum size, also it's 291 default size. 292compress_extension=%s Support adding specified extension, so that f2fs can enable 293 compression on those corresponding files, e.g. if all files 294 with '.ext' has high compression rate, we can set the '.ext' 295 on compression extension list and enable compression on 296 these file by default rather than to enable it via ioctl. 297 For other files, we can still enable compression via ioctl. 298 Note that, there is one reserved special extension '*', it 299 can be set to enable compression for all files. 300nocompress_extension=%s Support adding specified extension, so that f2fs can disable 301 compression on those corresponding files, just contrary to compression extension. 302 If you know exactly which files cannot be compressed, you can use this. 303 The same extension name can't appear in both compress and nocompress 304 extension at the same time. 305 If the compress extension specifies all files, the types specified by the 306 nocompress extension will be treated as special cases and will not be compressed. 307 Don't allow use '*' to specifie all file in nocompress extension. 308 After add nocompress_extension, the priority should be: 309 dir_flag < comp_extention,nocompress_extension < comp_file_flag,no_comp_file_flag. 310 See more in compression sections. 311 312compress_chksum Support verifying chksum of raw data in compressed cluster. 313compress_mode=%s Control file compression mode. This supports "fs" and "user" 314 modes. In "fs" mode (default), f2fs does automatic compression 315 on the compression enabled files. In "user" mode, f2fs disables 316 the automaic compression and gives the user discretion of 317 choosing the target file and the timing. The user can do manual 318 compression/decompression on the compression enabled files using 319 ioctls. 320compress_cache Support to use address space of a filesystem managed inode to 321 cache compressed block, in order to improve cache hit ratio of 322 random read. 323inlinecrypt When possible, encrypt/decrypt the contents of encrypted 324 files using the blk-crypto framework rather than 325 filesystem-layer encryption. This allows the use of 326 inline encryption hardware. The on-disk format is 327 unaffected. For more details, see 328 Documentation/block/inline-encryption.rst. 329atgc Enable age-threshold garbage collection, it provides high 330 effectiveness and efficiency on background GC. 331discard_unit=%s Control discard unit, the argument can be "block", "segment" 332 and "section", issued discard command's offset/size will be 333 aligned to the unit, by default, "discard_unit=block" is set, 334 so that small discard functionality is enabled. 335 For blkzoned device, "discard_unit=section" will be set by 336 default, it is helpful for large sized SMR or ZNS devices to 337 reduce memory cost by getting rid of fs metadata supports small 338 discard. 339======================== ============================================================ 340 341Debugfs Entries 342=============== 343 344/sys/kernel/debug/f2fs/ contains information about all the partitions mounted as 345f2fs. Each file shows the whole f2fs information. 346 347/sys/kernel/debug/f2fs/status includes: 348 349 - major file system information managed by f2fs currently 350 - average SIT information about whole segments 351 - current memory footprint consumed by f2fs. 352 353Sysfs Entries 354============= 355 356Information about mounted f2fs file systems can be found in 357/sys/fs/f2fs. Each mounted filesystem will have a directory in 358/sys/fs/f2fs based on its device name (i.e., /sys/fs/f2fs/sda). 359The files in each per-device directory are shown in table below. 360 361Files in /sys/fs/f2fs/<devname> 362(see also Documentation/ABI/testing/sysfs-fs-f2fs) 363 364Usage 365===== 366 3671. Download userland tools and compile them. 368 3692. Skip, if f2fs was compiled statically inside kernel. 370 Otherwise, insert the f2fs.ko module:: 371 372 # insmod f2fs.ko 373 3743. Create a directory to use when mounting:: 375 376 # mkdir /mnt/f2fs 377 3784. Format the block device, and then mount as f2fs:: 379 380 # mkfs.f2fs -l label /dev/block_device 381 # mount -t f2fs /dev/block_device /mnt/f2fs 382 383mkfs.f2fs 384--------- 385The mkfs.f2fs is for the use of formatting a partition as the f2fs filesystem, 386which builds a basic on-disk layout. 387 388The quick options consist of: 389 390=============== =========================================================== 391``-l [label]`` Give a volume label, up to 512 unicode name. 392``-a [0 or 1]`` Split start location of each area for heap-based allocation. 393 394 1 is set by default, which performs this. 395``-o [int]`` Set overprovision ratio in percent over volume size. 396 397 5 is set by default. 398``-s [int]`` Set the number of segments per section. 399 400 1 is set by default. 401``-z [int]`` Set the number of sections per zone. 402 403 1 is set by default. 404``-e [str]`` Set basic extension list. e.g. "mp3,gif,mov" 405``-t [0 or 1]`` Disable discard command or not. 406 407 1 is set by default, which conducts discard. 408=============== =========================================================== 409 410Note: please refer to the manpage of mkfs.f2fs(8) to get full option list. 411 412fsck.f2fs 413--------- 414The fsck.f2fs is a tool to check the consistency of an f2fs-formatted 415partition, which examines whether the filesystem metadata and user-made data 416are cross-referenced correctly or not. 417Note that, initial version of the tool does not fix any inconsistency. 418 419The quick options consist of:: 420 421 -d debug level [default:0] 422 423Note: please refer to the manpage of fsck.f2fs(8) to get full option list. 424 425dump.f2fs 426--------- 427The dump.f2fs shows the information of specific inode and dumps SSA and SIT to 428file. Each file is dump_ssa and dump_sit. 429 430The dump.f2fs is used to debug on-disk data structures of the f2fs filesystem. 431It shows on-disk inode information recognized by a given inode number, and is 432able to dump all the SSA and SIT entries into predefined files, ./dump_ssa and 433./dump_sit respectively. 434 435The options consist of:: 436 437 -d debug level [default:0] 438 -i inode no (hex) 439 -s [SIT dump segno from #1~#2 (decimal), for all 0~-1] 440 -a [SSA dump segno from #1~#2 (decimal), for all 0~-1] 441 442Examples:: 443 444 # dump.f2fs -i [ino] /dev/sdx 445 # dump.f2fs -s 0~-1 /dev/sdx (SIT dump) 446 # dump.f2fs -a 0~-1 /dev/sdx (SSA dump) 447 448Note: please refer to the manpage of dump.f2fs(8) to get full option list. 449 450sload.f2fs 451---------- 452The sload.f2fs gives a way to insert files and directories in the exisiting disk 453image. This tool is useful when building f2fs images given compiled files. 454 455Note: please refer to the manpage of sload.f2fs(8) to get full option list. 456 457resize.f2fs 458----------- 459The resize.f2fs lets a user resize the f2fs-formatted disk image, while preserving 460all the files and directories stored in the image. 461 462Note: please refer to the manpage of resize.f2fs(8) to get full option list. 463 464defrag.f2fs 465----------- 466The defrag.f2fs can be used to defragment scattered written data as well as 467filesystem metadata across the disk. This can improve the write speed by giving 468more free consecutive space. 469 470Note: please refer to the manpage of defrag.f2fs(8) to get full option list. 471 472f2fs_io 473------- 474The f2fs_io is a simple tool to issue various filesystem APIs as well as 475f2fs-specific ones, which is very useful for QA tests. 476 477Note: please refer to the manpage of f2fs_io(8) to get full option list. 478 479Design 480====== 481 482On-disk Layout 483-------------- 484 485F2FS divides the whole volume into a number of segments, each of which is fixed 486to 2MB in size. A section is composed of consecutive segments, and a zone 487consists of a set of sections. By default, section and zone sizes are set to one 488segment size identically, but users can easily modify the sizes by mkfs. 489 490F2FS splits the entire volume into six areas, and all the areas except superblock 491consist of multiple segments as described below:: 492 493 align with the zone size <-| 494 |-> align with the segment size 495 _________________________________________________________________________ 496 | | | Segment | Node | Segment | | 497 | Superblock | Checkpoint | Info. | Address | Summary | Main | 498 | (SB) | (CP) | Table (SIT) | Table (NAT) | Area (SSA) | | 499 |____________|_____2______|______N______|______N______|______N_____|__N___| 500 . . 501 . . 502 . . 503 ._________________________________________. 504 |_Segment_|_..._|_Segment_|_..._|_Segment_| 505 . . 506 ._________._________ 507 |_section_|__...__|_ 508 . . 509 .________. 510 |__zone__| 511 512- Superblock (SB) 513 It is located at the beginning of the partition, and there exist two copies 514 to avoid file system crash. It contains basic partition information and some 515 default parameters of f2fs. 516 517- Checkpoint (CP) 518 It contains file system information, bitmaps for valid NAT/SIT sets, orphan 519 inode lists, and summary entries of current active segments. 520 521- Segment Information Table (SIT) 522 It contains segment information such as valid block count and bitmap for the 523 validity of all the blocks. 524 525- Node Address Table (NAT) 526 It is composed of a block address table for all the node blocks stored in 527 Main area. 528 529- Segment Summary Area (SSA) 530 It contains summary entries which contains the owner information of all the 531 data and node blocks stored in Main area. 532 533- Main Area 534 It contains file and directory data including their indices. 535 536In order to avoid misalignment between file system and flash-based storage, F2FS 537aligns the start block address of CP with the segment size. Also, it aligns the 538start block address of Main area with the zone size by reserving some segments 539in SSA area. 540 541Reference the following survey for additional technical details. 542https://wiki.linaro.org/WorkingGroups/Kernel/Projects/FlashCardSurvey 543 544File System Metadata Structure 545------------------------------ 546 547F2FS adopts the checkpointing scheme to maintain file system consistency. At 548mount time, F2FS first tries to find the last valid checkpoint data by scanning 549CP area. In order to reduce the scanning time, F2FS uses only two copies of CP. 550One of them always indicates the last valid data, which is called as shadow copy 551mechanism. In addition to CP, NAT and SIT also adopt the shadow copy mechanism. 552 553For file system consistency, each CP points to which NAT and SIT copies are 554valid, as shown as below:: 555 556 +--------+----------+---------+ 557 | CP | SIT | NAT | 558 +--------+----------+---------+ 559 . . . . 560 . . . . 561 . . . . 562 +-------+-------+--------+--------+--------+--------+ 563 | CP #0 | CP #1 | SIT #0 | SIT #1 | NAT #0 | NAT #1 | 564 +-------+-------+--------+--------+--------+--------+ 565 | ^ ^ 566 | | | 567 `----------------------------------------' 568 569Index Structure 570--------------- 571 572The key data structure to manage the data locations is a "node". Similar to 573traditional file structures, F2FS has three types of node: inode, direct node, 574indirect node. F2FS assigns 4KB to an inode block which contains 923 data block 575indices, two direct node pointers, two indirect node pointers, and one double 576indirect node pointer as described below. One direct node block contains 1018 577data blocks, and one indirect node block contains also 1018 node blocks. Thus, 578one inode block (i.e., a file) covers:: 579 580 4KB * (923 + 2 * 1018 + 2 * 1018 * 1018 + 1018 * 1018 * 1018) := 3.94TB. 581 582 Inode block (4KB) 583 |- data (923) 584 |- direct node (2) 585 | `- data (1018) 586 |- indirect node (2) 587 | `- direct node (1018) 588 | `- data (1018) 589 `- double indirect node (1) 590 `- indirect node (1018) 591 `- direct node (1018) 592 `- data (1018) 593 594Note that all the node blocks are mapped by NAT which means the location of 595each node is translated by the NAT table. In the consideration of the wandering 596tree problem, F2FS is able to cut off the propagation of node updates caused by 597leaf data writes. 598 599Directory Structure 600------------------- 601 602A directory entry occupies 11 bytes, which consists of the following attributes. 603 604- hash hash value of the file name 605- ino inode number 606- len the length of file name 607- type file type such as directory, symlink, etc 608 609A dentry block consists of 214 dentry slots and file names. Therein a bitmap is 610used to represent whether each dentry is valid or not. A dentry block occupies 6114KB with the following composition. 612 613:: 614 615 Dentry Block(4 K) = bitmap (27 bytes) + reserved (3 bytes) + 616 dentries(11 * 214 bytes) + file name (8 * 214 bytes) 617 618 [Bucket] 619 +--------------------------------+ 620 |dentry block 1 | dentry block 2 | 621 +--------------------------------+ 622 . . 623 . . 624 . [Dentry Block Structure: 4KB] . 625 +--------+----------+----------+------------+ 626 | bitmap | reserved | dentries | file names | 627 +--------+----------+----------+------------+ 628 [Dentry Block: 4KB] . . 629 . . 630 . . 631 +------+------+-----+------+ 632 | hash | ino | len | type | 633 +------+------+-----+------+ 634 [Dentry Structure: 11 bytes] 635 636F2FS implements multi-level hash tables for directory structure. Each level has 637a hash table with dedicated number of hash buckets as shown below. Note that 638"A(2B)" means a bucket includes 2 data blocks. 639 640:: 641 642 ---------------------- 643 A : bucket 644 B : block 645 N : MAX_DIR_HASH_DEPTH 646 ---------------------- 647 648 level #0 | A(2B) 649 | 650 level #1 | A(2B) - A(2B) 651 | 652 level #2 | A(2B) - A(2B) - A(2B) - A(2B) 653 . | . . . . 654 level #N/2 | A(2B) - A(2B) - A(2B) - A(2B) - A(2B) - ... - A(2B) 655 . | . . . . 656 level #N | A(4B) - A(4B) - A(4B) - A(4B) - A(4B) - ... - A(4B) 657 658The number of blocks and buckets are determined by:: 659 660 ,- 2, if n < MAX_DIR_HASH_DEPTH / 2, 661 # of blocks in level #n = | 662 `- 4, Otherwise 663 664 ,- 2^(n + dir_level), 665 | if n + dir_level < MAX_DIR_HASH_DEPTH / 2, 666 # of buckets in level #n = | 667 `- 2^((MAX_DIR_HASH_DEPTH / 2) - 1), 668 Otherwise 669 670When F2FS finds a file name in a directory, at first a hash value of the file 671name is calculated. Then, F2FS scans the hash table in level #0 to find the 672dentry consisting of the file name and its inode number. If not found, F2FS 673scans the next hash table in level #1. In this way, F2FS scans hash tables in 674each levels incrementally from 1 to N. In each level F2FS needs to scan only 675one bucket determined by the following equation, which shows O(log(# of files)) 676complexity:: 677 678 bucket number to scan in level #n = (hash value) % (# of buckets in level #n) 679 680In the case of file creation, F2FS finds empty consecutive slots that cover the 681file name. F2FS searches the empty slots in the hash tables of whole levels from 6821 to N in the same way as the lookup operation. 683 684The following figure shows an example of two cases holding children:: 685 686 --------------> Dir <-------------- 687 | | 688 child child 689 690 child - child [hole] - child 691 692 child - child - child [hole] - [hole] - child 693 694 Case 1: Case 2: 695 Number of children = 6, Number of children = 3, 696 File size = 7 File size = 7 697 698Default Block Allocation 699------------------------ 700 701At runtime, F2FS manages six active logs inside "Main" area: Hot/Warm/Cold node 702and Hot/Warm/Cold data. 703 704- Hot node contains direct node blocks of directories. 705- Warm node contains direct node blocks except hot node blocks. 706- Cold node contains indirect node blocks 707- Hot data contains dentry blocks 708- Warm data contains data blocks except hot and cold data blocks 709- Cold data contains multimedia data or migrated data blocks 710 711LFS has two schemes for free space management: threaded log and copy-and-compac- 712tion. The copy-and-compaction scheme which is known as cleaning, is well-suited 713for devices showing very good sequential write performance, since free segments 714are served all the time for writing new data. However, it suffers from cleaning 715overhead under high utilization. Contrarily, the threaded log scheme suffers 716from random writes, but no cleaning process is needed. F2FS adopts a hybrid 717scheme where the copy-and-compaction scheme is adopted by default, but the 718policy is dynamically changed to the threaded log scheme according to the file 719system status. 720 721In order to align F2FS with underlying flash-based storage, F2FS allocates a 722segment in a unit of section. F2FS expects that the section size would be the 723same as the unit size of garbage collection in FTL. Furthermore, with respect 724to the mapping granularity in FTL, F2FS allocates each section of the active 725logs from different zones as much as possible, since FTL can write the data in 726the active logs into one allocation unit according to its mapping granularity. 727 728Cleaning process 729---------------- 730 731F2FS does cleaning both on demand and in the background. On-demand cleaning is 732triggered when there are not enough free segments to serve VFS calls. Background 733cleaner is operated by a kernel thread, and triggers the cleaning job when the 734system is idle. 735 736F2FS supports two victim selection policies: greedy and cost-benefit algorithms. 737In the greedy algorithm, F2FS selects a victim segment having the smallest number 738of valid blocks. In the cost-benefit algorithm, F2FS selects a victim segment 739according to the segment age and the number of valid blocks in order to address 740log block thrashing problem in the greedy algorithm. F2FS adopts the greedy 741algorithm for on-demand cleaner, while background cleaner adopts cost-benefit 742algorithm. 743 744In order to identify whether the data in the victim segment are valid or not, 745F2FS manages a bitmap. Each bit represents the validity of a block, and the 746bitmap is composed of a bit stream covering whole blocks in main area. 747 748Fallocate(2) Policy 749------------------- 750 751The default policy follows the below POSIX rule. 752 753Allocating disk space 754 The default operation (i.e., mode is zero) of fallocate() allocates 755 the disk space within the range specified by offset and len. The 756 file size (as reported by stat(2)) will be changed if offset+len is 757 greater than the file size. Any subregion within the range specified 758 by offset and len that did not contain data before the call will be 759 initialized to zero. This default behavior closely resembles the 760 behavior of the posix_fallocate(3) library function, and is intended 761 as a method of optimally implementing that function. 762 763However, once F2FS receives ioctl(fd, F2FS_IOC_SET_PIN_FILE) in prior to 764fallocate(fd, DEFAULT_MODE), it allocates on-disk block addressess having 765zero or random data, which is useful to the below scenario where: 766 767 1. create(fd) 768 2. ioctl(fd, F2FS_IOC_SET_PIN_FILE) 769 3. fallocate(fd, 0, 0, size) 770 4. address = fibmap(fd, offset) 771 5. open(blkdev) 772 6. write(blkdev, address) 773 774Compression implementation 775-------------------------- 776 777- New term named cluster is defined as basic unit of compression, file can 778 be divided into multiple clusters logically. One cluster includes 4 << n 779 (n >= 0) logical pages, compression size is also cluster size, each of 780 cluster can be compressed or not. 781 782- In cluster metadata layout, one special block address is used to indicate 783 a cluster is a compressed one or normal one; for compressed cluster, following 784 metadata maps cluster to [1, 4 << n - 1] physical blocks, in where f2fs 785 stores data including compress header and compressed data. 786 787- In order to eliminate write amplification during overwrite, F2FS only 788 support compression on write-once file, data can be compressed only when 789 all logical blocks in cluster contain valid data and compress ratio of 790 cluster data is lower than specified threshold. 791 792- To enable compression on regular inode, there are four ways: 793 794 * chattr +c file 795 * chattr +c dir; touch dir/file 796 * mount w/ -o compress_extension=ext; touch file.ext 797 * mount w/ -o compress_extension=*; touch any_file 798 799- To disable compression on regular inode, there are two ways: 800 801 * chattr -c file 802 * mount w/ -o nocompress_extension=ext; touch file.ext 803 804- Priority in between FS_COMPR_FL, FS_NOCOMP_FS, extensions: 805 806 * compress_extension=so; nocompress_extension=zip; chattr +c dir; touch 807 dir/foo.so; touch dir/bar.zip; touch dir/baz.txt; then foo.so and baz.txt 808 should be compresse, bar.zip should be non-compressed. chattr +c dir/bar.zip 809 can enable compress on bar.zip. 810 * compress_extension=so; nocompress_extension=zip; chattr -c dir; touch 811 dir/foo.so; touch dir/bar.zip; touch dir/baz.txt; then foo.so should be 812 compresse, bar.zip and baz.txt should be non-compressed. 813 chattr+c dir/bar.zip; chattr+c dir/baz.txt; can enable compress on bar.zip 814 and baz.txt. 815 816- At this point, compression feature doesn't expose compressed space to user 817 directly in order to guarantee potential data updates later to the space. 818 Instead, the main goal is to reduce data writes to flash disk as much as 819 possible, resulting in extending disk life time as well as relaxing IO 820 congestion. Alternatively, we've added ioctl(F2FS_IOC_RELEASE_COMPRESS_BLOCKS) 821 interface to reclaim compressed space and show it to user after putting the 822 immutable bit. Immutable bit, after release, it doesn't allow writing/mmaping 823 on the file, until reserving compressed space via 824 ioctl(F2FS_IOC_RESERVE_COMPRESS_BLOCKS) or truncating filesize to zero. 825 826Compress metadata layout:: 827 828 [Dnode Structure] 829 +-----------------------------------------------+ 830 | cluster 1 | cluster 2 | ......... | cluster N | 831 +-----------------------------------------------+ 832 . . . . 833 . . . . 834 . Compressed Cluster . . Normal Cluster . 835 +----------+---------+---------+---------+ +---------+---------+---------+---------+ 836 |compr flag| block 1 | block 2 | block 3 | | block 1 | block 2 | block 3 | block 4 | 837 +----------+---------+---------+---------+ +---------+---------+---------+---------+ 838 . . 839 . . 840 . . 841 +-------------+-------------+----------+----------------------------+ 842 | data length | data chksum | reserved | compressed data | 843 +-------------+-------------+----------+----------------------------+ 844 845Compression mode 846-------------------------- 847 848f2fs supports "fs" and "user" compression modes with "compression_mode" mount option. 849With this option, f2fs provides a choice to select the way how to compress the 850compression enabled files (refer to "Compression implementation" section for how to 851enable compression on a regular inode). 852 8531) compress_mode=fs 854This is the default option. f2fs does automatic compression in the writeback of the 855compression enabled files. 856 8572) compress_mode=user 858This disables the automatic compression and gives the user discretion of choosing the 859target file and the timing. The user can do manual compression/decompression on the 860compression enabled files using F2FS_IOC_DECOMPRESS_FILE and F2FS_IOC_COMPRESS_FILE 861ioctls like the below. 862 863To decompress a file, 864 865fd = open(filename, O_WRONLY, 0); 866ret = ioctl(fd, F2FS_IOC_DECOMPRESS_FILE); 867 868To compress a file, 869 870fd = open(filename, O_WRONLY, 0); 871ret = ioctl(fd, F2FS_IOC_COMPRESS_FILE); 872 873NVMe Zoned Namespace devices 874---------------------------- 875 876- ZNS defines a per-zone capacity which can be equal or less than the 877 zone-size. Zone-capacity is the number of usable blocks in the zone. 878 F2FS checks if zone-capacity is less than zone-size, if it is, then any 879 segment which starts after the zone-capacity is marked as not-free in 880 the free segment bitmap at initial mount time. These segments are marked 881 as permanently used so they are not allocated for writes and 882 consequently are not needed to be garbage collected. In case the 883 zone-capacity is not aligned to default segment size(2MB), then a segment 884 can start before the zone-capacity and span across zone-capacity boundary. 885 Such spanning segments are also considered as usable segments. All blocks 886 past the zone-capacity are considered unusable in these segments. 887