1================================================================================ 2WHAT IS Flash-Friendly File System (F2FS)? 3================================================================================ 4 5NAND flash memory-based storage devices, such as SSD, eMMC, and SD cards, have 6been equipped on a variety systems ranging from mobile to server systems. Since 7they are known to have different characteristics from the conventional rotating 8disks, a file system, an upper layer to the storage device, should adapt to the 9changes from the sketch in the design level. 10 11F2FS is a file system exploiting NAND flash memory-based storage devices, which 12is based on Log-structured File System (LFS). The design has been focused on 13addressing the fundamental issues in LFS, which are snowball effect of wandering 14tree and high cleaning overhead. 15 16Since a NAND flash memory-based storage device shows different characteristic 17according to its internal geometry or flash memory management scheme, namely FTL, 18F2FS and its tools support various parameters not only for configuring on-disk 19layout, but also for selecting allocation and cleaning algorithms. 20 21The following git tree provides the file system formatting tool (mkfs.f2fs), 22a consistency checking tool (fsck.f2fs), and a debugging tool (dump.f2fs). 23>> git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs-tools.git 24 25For reporting bugs and sending patches, please use the following mailing list: 26>> linux-f2fs-devel@lists.sourceforge.net 27 28================================================================================ 29BACKGROUND AND DESIGN ISSUES 30================================================================================ 31 32Log-structured File System (LFS) 33-------------------------------- 34"A log-structured file system writes all modifications to disk sequentially in 35a log-like structure, thereby speeding up both file writing and crash recovery. 36The log is the only structure on disk; it contains indexing information so that 37files can be read back from the log efficiently. In order to maintain large free 38areas on disk for fast writing, we divide the log into segments and use a 39segment cleaner to compress the live information from heavily fragmented 40segments." from Rosenblum, M. and Ousterhout, J. K., 1992, "The design and 41implementation of a log-structured file system", ACM Trans. Computer Systems 4210, 1, 26–52. 43 44Wandering Tree Problem 45---------------------- 46In LFS, when a file data is updated and written to the end of log, its direct 47pointer block is updated due to the changed location. Then the indirect pointer 48block is also updated due to the direct pointer block update. In this manner, 49the upper index structures such as inode, inode map, and checkpoint block are 50also updated recursively. This problem is called as wandering tree problem [1], 51and in order to enhance the performance, it should eliminate or relax the update 52propagation as much as possible. 53 54[1] Bityutskiy, A. 2005. JFFS3 design issues. http://www.linux-mtd.infradead.org/ 55 56Cleaning Overhead 57----------------- 58Since LFS is based on out-of-place writes, it produces so many obsolete blocks 59scattered across the whole storage. In order to serve new empty log space, it 60needs to reclaim these obsolete blocks seamlessly to users. This job is called 61as a cleaning process. 62 63The process consists of three operations as follows. 641. A victim segment is selected through referencing segment usage table. 652. It loads parent index structures of all the data in the victim identified by 66 segment summary blocks. 673. It checks the cross-reference between the data and its parent index structure. 684. It moves valid data selectively. 69 70This cleaning job may cause unexpected long delays, so the most important goal 71is to hide the latencies to users. And also definitely, it should reduce the 72amount of valid data to be moved, and move them quickly as well. 73 74================================================================================ 75KEY FEATURES 76================================================================================ 77 78Flash Awareness 79--------------- 80- Enlarge the random write area for better performance, but provide the high 81 spatial locality 82- Align FS data structures to the operational units in FTL as best efforts 83 84Wandering Tree Problem 85---------------------- 86- Use a term, “node”, that represents inodes as well as various pointer blocks 87- Introduce Node Address Table (NAT) containing the locations of all the “node” 88 blocks; this will cut off the update propagation. 89 90Cleaning Overhead 91----------------- 92- Support a background cleaning process 93- Support greedy and cost-benefit algorithms for victim selection policies 94- Support multi-head logs for static/dynamic hot and cold data separation 95- Introduce adaptive logging for efficient block allocation 96 97================================================================================ 98MOUNT OPTIONS 99================================================================================ 100 101background_gc=%s Turn on/off cleaning operations, namely garbage 102 collection, triggered in background when I/O subsystem is 103 idle. If background_gc=on, it will turn on the garbage 104 collection and if background_gc=off, garbage collection 105 will be turned off. If background_gc=sync, it will turn 106 on synchronous garbage collection running in background. 107 Default value for this option is on. So garbage 108 collection is on by default. 109disable_roll_forward Disable the roll-forward recovery routine 110norecovery Disable the roll-forward recovery routine, mounted read- 111 only (i.e., -o ro,disable_roll_forward) 112discard/nodiscard Enable/disable real-time discard in f2fs, if discard is 113 enabled, f2fs will issue discard/TRIM commands when a 114 segment is cleaned. 115no_heap Disable heap-style segment allocation which finds free 116 segments for data from the beginning of main area, while 117 for node from the end of main area. 118nouser_xattr Disable Extended User Attributes. Note: xattr is enabled 119 by default if CONFIG_F2FS_FS_XATTR is selected. 120noacl Disable POSIX Access Control List. Note: acl is enabled 121 by default if CONFIG_F2FS_FS_POSIX_ACL is selected. 122active_logs=%u Support configuring the number of active logs. In the 123 current design, f2fs supports only 2, 4, and 6 logs. 124 Default number is 6. 125disable_ext_identify Disable the extension list configured by mkfs, so f2fs 126 does not aware of cold files such as media files. 127inline_xattr Enable the inline xattrs feature. 128noinline_xattr Disable the inline xattrs feature. 129inline_xattr_size=%u Support configuring inline xattr size, it depends on 130 flexible inline xattr feature. 131inline_data Enable the inline data feature: New created small(<~3.4k) 132 files can be written into inode block. 133inline_dentry Enable the inline dir feature: data in new created 134 directory entries can be written into inode block. The 135 space of inode block which is used to store inline 136 dentries is limited to ~3.4k. 137noinline_dentry Disable the inline dentry feature. 138flush_merge Merge concurrent cache_flush commands as much as possible 139 to eliminate redundant command issues. If the underlying 140 device handles the cache_flush command relatively slowly, 141 recommend to enable this option. 142nobarrier This option can be used if underlying storage guarantees 143 its cached data should be written to the novolatile area. 144 If this option is set, no cache_flush commands are issued 145 but f2fs still guarantees the write ordering of all the 146 data writes. 147fastboot This option is used when a system wants to reduce mount 148 time as much as possible, even though normal performance 149 can be sacrificed. 150extent_cache Enable an extent cache based on rb-tree, it can cache 151 as many as extent which map between contiguous logical 152 address and physical address per inode, resulting in 153 increasing the cache hit ratio. Set by default. 154noextent_cache Disable an extent cache based on rb-tree explicitly, see 155 the above extent_cache mount option. 156noinline_data Disable the inline data feature, inline data feature is 157 enabled by default. 158data_flush Enable data flushing before checkpoint in order to 159 persist data of regular and symlink. 160reserve_root=%d Support configuring reserved space which is used for 161 allocation from a privileged user with specified uid or 162 gid, unit: 4KB, the default limit is 0.2% of user blocks. 163resuid=%d The user ID which may use the reserved blocks. 164resgid=%d The group ID which may use the reserved blocks. 165fault_injection=%d Enable fault injection in all supported types with 166 specified injection rate. 167fault_type=%d Support configuring fault injection type, should be 168 enabled with fault_injection option, fault type value 169 is shown below, it supports single or combined type. 170 Type_Name Type_Value 171 FAULT_KMALLOC 0x000000001 172 FAULT_KVMALLOC 0x000000002 173 FAULT_PAGE_ALLOC 0x000000004 174 FAULT_PAGE_GET 0x000000008 175 FAULT_ALLOC_BIO 0x000000010 176 FAULT_ALLOC_NID 0x000000020 177 FAULT_ORPHAN 0x000000040 178 FAULT_BLOCK 0x000000080 179 FAULT_DIR_DEPTH 0x000000100 180 FAULT_EVICT_INODE 0x000000200 181 FAULT_TRUNCATE 0x000000400 182 FAULT_READ_IO 0x000000800 183 FAULT_CHECKPOINT 0x000001000 184 FAULT_DISCARD 0x000002000 185 FAULT_WRITE_IO 0x000004000 186mode=%s Control block allocation mode which supports "adaptive" 187 and "lfs". In "lfs" mode, there should be no random 188 writes towards main area. 189io_bits=%u Set the bit size of write IO requests. It should be set 190 with "mode=lfs". 191usrquota Enable plain user disk quota accounting. 192grpquota Enable plain group disk quota accounting. 193prjquota Enable plain project quota accounting. 194usrjquota=<file> Appoint specified file and type during mount, so that quota 195grpjquota=<file> information can be properly updated during recovery flow, 196prjjquota=<file> <quota file>: must be in root directory; 197jqfmt=<quota type> <quota type>: [vfsold,vfsv0,vfsv1]. 198offusrjquota Turn off user journelled quota. 199offgrpjquota Turn off group journelled quota. 200offprjjquota Turn off project journelled quota. 201quota Enable plain user disk quota accounting. 202noquota Disable all plain disk quota option. 203whint_mode=%s Control which write hints are passed down to block 204 layer. This supports "off", "user-based", and 205 "fs-based". In "off" mode (default), f2fs does not pass 206 down hints. In "user-based" mode, f2fs tries to pass 207 down hints given by users. And in "fs-based" mode, f2fs 208 passes down hints with its policy. 209alloc_mode=%s Adjust block allocation policy, which supports "reuse" 210 and "default". 211fsync_mode=%s Control the policy of fsync. Currently supports "posix", 212 "strict", and "nobarrier". In "posix" mode, which is 213 default, fsync will follow POSIX semantics and does a 214 light operation to improve the filesystem performance. 215 In "strict" mode, fsync will be heavy and behaves in line 216 with xfs, ext4 and btrfs, where xfstest generic/342 will 217 pass, but the performance will regress. "nobarrier" is 218 based on "posix", but doesn't issue flush command for 219 non-atomic files likewise "nobarrier" mount option. 220test_dummy_encryption Enable dummy encryption, which provides a fake fscrypt 221 context. The fake fscrypt context is used by xfstests. 222checkpoint=%s[:%u[%]] Set to "disable" to turn off checkpointing. Set to "enable" 223 to reenable checkpointing. Is enabled by default. While 224 disabled, any unmounting or unexpected shutdowns will cause 225 the filesystem contents to appear as they did when the 226 filesystem was mounted with that option. 227 While mounting with checkpoint=disabled, the filesystem must 228 run garbage collection to ensure that all available space can 229 be used. If this takes too much time, the mount may return 230 EAGAIN. You may optionally add a value to indicate how much 231 of the disk you would be willing to temporarily give up to 232 avoid additional garbage collection. This can be given as a 233 number of blocks, or as a percent. For instance, mounting 234 with checkpoint=disable:100% would always succeed, but it may 235 hide up to all remaining free space. The actual space that 236 would be unusable can be viewed at /sys/fs/f2fs/<disk>/unusable 237 This space is reclaimed once checkpoint=enable. 238 239================================================================================ 240DEBUGFS ENTRIES 241================================================================================ 242 243/sys/kernel/debug/f2fs/ contains information about all the partitions mounted as 244f2fs. Each file shows the whole f2fs information. 245 246/sys/kernel/debug/f2fs/status includes: 247 - major file system information managed by f2fs currently 248 - average SIT information about whole segments 249 - current memory footprint consumed by f2fs. 250 251================================================================================ 252SYSFS ENTRIES 253================================================================================ 254 255Information about mounted f2fs file systems can be found in 256/sys/fs/f2fs. Each mounted filesystem will have a directory in 257/sys/fs/f2fs based on its device name (i.e., /sys/fs/f2fs/sda). 258The files in each per-device directory are shown in table below. 259 260Files in /sys/fs/f2fs/<devname> 261(see also Documentation/ABI/testing/sysfs-fs-f2fs) 262.............................................................................. 263 File Content 264 265 gc_urgent_sleep_time This parameter controls sleep time for gc_urgent. 266 500 ms is set by default. See above gc_urgent. 267 268 gc_min_sleep_time This tuning parameter controls the minimum sleep 269 time for the garbage collection thread. Time is 270 in milliseconds. 271 272 gc_max_sleep_time This tuning parameter controls the maximum sleep 273 time for the garbage collection thread. Time is 274 in milliseconds. 275 276 gc_no_gc_sleep_time This tuning parameter controls the default sleep 277 time for the garbage collection thread. Time is 278 in milliseconds. 279 280 gc_idle This parameter controls the selection of victim 281 policy for garbage collection. Setting gc_idle = 0 282 (default) will disable this option. Setting 283 gc_idle = 1 will select the Cost Benefit approach 284 & setting gc_idle = 2 will select the greedy approach. 285 286 gc_urgent This parameter controls triggering background GCs 287 urgently or not. Setting gc_urgent = 0 [default] 288 makes back to default behavior, while if it is set 289 to 1, background thread starts to do GC by given 290 gc_urgent_sleep_time interval. 291 292 reclaim_segments This parameter controls the number of prefree 293 segments to be reclaimed. If the number of prefree 294 segments is larger than the number of segments 295 in the proportion to the percentage over total 296 volume size, f2fs tries to conduct checkpoint to 297 reclaim the prefree segments to free segments. 298 By default, 5% over total # of segments. 299 300 main_blkaddr This value gives the first block address of 301 MAIN area in the partition. 302 303 max_small_discards This parameter controls the number of discard 304 commands that consist small blocks less than 2MB. 305 The candidates to be discarded are cached until 306 checkpoint is triggered, and issued during the 307 checkpoint. By default, it is disabled with 0. 308 309 discard_granularity This parameter controls the granularity of discard 310 command size. It will issue discard commands iif 311 the size is larger than given granularity. Its 312 unit size is 4KB, and 4 (=16KB) is set by default. 313 The maximum value is 128 (=512KB). 314 315 reserved_blocks This parameter indicates the number of blocks that 316 f2fs reserves internally for root. 317 318 batched_trim_sections This parameter controls the number of sections 319 to be trimmed out in batch mode when FITRIM 320 conducts. 32 sections is set by default. 321 322 ipu_policy This parameter controls the policy of in-place 323 updates in f2fs. There are five policies: 324 0x01: F2FS_IPU_FORCE, 0x02: F2FS_IPU_SSR, 325 0x04: F2FS_IPU_UTIL, 0x08: F2FS_IPU_SSR_UTIL, 326 0x10: F2FS_IPU_FSYNC. 327 328 min_ipu_util This parameter controls the threshold to trigger 329 in-place-updates. The number indicates percentage 330 of the filesystem utilization, and used by 331 F2FS_IPU_UTIL and F2FS_IPU_SSR_UTIL policies. 332 333 min_fsync_blocks This parameter controls the threshold to trigger 334 in-place-updates when F2FS_IPU_FSYNC mode is set. 335 The number indicates the number of dirty pages 336 when fsync needs to flush on its call path. If 337 the number is less than this value, it triggers 338 in-place-updates. 339 340 min_seq_blocks This parameter controls the threshold to serialize 341 write IOs issued by multiple threads in parallel. 342 343 min_hot_blocks This parameter controls the threshold to allocate 344 a hot data log for pending data blocks to write. 345 346 min_ssr_sections This parameter adds the threshold when deciding 347 SSR block allocation. If this is large, SSR mode 348 will be enabled early. 349 350 ram_thresh This parameter controls the memory footprint used 351 by free nids and cached nat entries. By default, 352 1 is set, which indicates 10 MB / 1 GB RAM. 353 354 ra_nid_pages When building free nids, F2FS reads NAT blocks 355 ahead for speed up. Default is 0. 356 357 dirty_nats_ratio Given dirty ratio of cached nat entries, F2FS 358 determines flushing them in background. 359 360 max_victim_search This parameter controls the number of trials to 361 find a victim segment when conducting SSR and 362 cleaning operations. The default value is 4096 363 which covers 8GB block address range. 364 365 migration_granularity For large-sized sections, F2FS can stop GC given 366 this granularity instead of reclaiming entire 367 section. 368 369 dir_level This parameter controls the directory level to 370 support large directory. If a directory has a 371 number of files, it can reduce the file lookup 372 latency by increasing this dir_level value. 373 Otherwise, it needs to decrease this value to 374 reduce the space overhead. The default value is 0. 375 376 cp_interval F2FS tries to do checkpoint periodically, 60 secs 377 by default. 378 379 idle_interval F2FS detects system is idle, if there's no F2FS 380 operations during given interval, 5 secs by 381 default. 382 383 discard_idle_interval F2FS detects the discard thread is idle, given 384 time interval. Default is 5 secs. 385 386 gc_idle_interval F2FS detects the GC thread is idle, given time 387 interval. Default is 5 secs. 388 389 umount_discard_timeout When unmounting the disk, F2FS waits for finishing 390 queued discard commands which can take huge time. 391 This gives time out for it, 5 secs by default. 392 393 iostat_enable This controls to enable/disable iostat in F2FS. 394 395 readdir_ra This enables/disabled readahead of inode blocks 396 in readdir, and default is enabled. 397 398 gc_pin_file_thresh This indicates how many GC can be failed for the 399 pinned file. If it exceeds this, F2FS doesn't 400 guarantee its pinning state. 2048 trials is set 401 by default. 402 403 extension_list This enables to change extension_list for hot/cold 404 files in runtime. 405 406 inject_rate This controls injection rate of arbitrary faults. 407 408 inject_type This controls injection type of arbitrary faults. 409 410 dirty_segments This shows # of dirty segments. 411 412 lifetime_write_kbytes This shows # of data written to the disk. 413 414 features This shows current features enabled on F2FS. 415 416 current_reserved_blocks This shows # of blocks currently reserved. 417 418 unusable If checkpoint=disable, this shows the number of 419 blocks that are unusable. 420 If checkpoint=enable it shows the number of blocks 421 that would be unusable if checkpoint=disable were 422 to be set. 423 424encoding This shows the encoding used for casefolding. 425 If casefolding is not enabled, returns (none) 426 427================================================================================ 428USAGE 429================================================================================ 430 4311. Download userland tools and compile them. 432 4332. Skip, if f2fs was compiled statically inside kernel. 434 Otherwise, insert the f2fs.ko module. 435 # insmod f2fs.ko 436 4373. Create a directory trying to mount 438 # mkdir /mnt/f2fs 439 4404. Format the block device, and then mount as f2fs 441 # mkfs.f2fs -l label /dev/block_device 442 # mount -t f2fs /dev/block_device /mnt/f2fs 443 444mkfs.f2fs 445--------- 446The mkfs.f2fs is for the use of formatting a partition as the f2fs filesystem, 447which builds a basic on-disk layout. 448 449The options consist of: 450-l [label] : Give a volume label, up to 512 unicode name. 451-a [0 or 1] : Split start location of each area for heap-based allocation. 452 1 is set by default, which performs this. 453-o [int] : Set overprovision ratio in percent over volume size. 454 5 is set by default. 455-s [int] : Set the number of segments per section. 456 1 is set by default. 457-z [int] : Set the number of sections per zone. 458 1 is set by default. 459-e [str] : Set basic extension list. e.g. "mp3,gif,mov" 460-t [0 or 1] : Disable discard command or not. 461 1 is set by default, which conducts discard. 462 463fsck.f2fs 464--------- 465The fsck.f2fs is a tool to check the consistency of an f2fs-formatted 466partition, which examines whether the filesystem metadata and user-made data 467are cross-referenced correctly or not. 468Note that, initial version of the tool does not fix any inconsistency. 469 470The options consist of: 471 -d debug level [default:0] 472 473dump.f2fs 474--------- 475The dump.f2fs shows the information of specific inode and dumps SSA and SIT to 476file. Each file is dump_ssa and dump_sit. 477 478The dump.f2fs is used to debug on-disk data structures of the f2fs filesystem. 479It shows on-disk inode information recognized by a given inode number, and is 480able to dump all the SSA and SIT entries into predefined files, ./dump_ssa and 481./dump_sit respectively. 482 483The options consist of: 484 -d debug level [default:0] 485 -i inode no (hex) 486 -s [SIT dump segno from #1~#2 (decimal), for all 0~-1] 487 -a [SSA dump segno from #1~#2 (decimal), for all 0~-1] 488 489Examples: 490# dump.f2fs -i [ino] /dev/sdx 491# dump.f2fs -s 0~-1 /dev/sdx (SIT dump) 492# dump.f2fs -a 0~-1 /dev/sdx (SSA dump) 493 494================================================================================ 495DESIGN 496================================================================================ 497 498On-disk Layout 499-------------- 500 501F2FS divides the whole volume into a number of segments, each of which is fixed 502to 2MB in size. A section is composed of consecutive segments, and a zone 503consists of a set of sections. By default, section and zone sizes are set to one 504segment size identically, but users can easily modify the sizes by mkfs. 505 506F2FS splits the entire volume into six areas, and all the areas except superblock 507consists of multiple segments as described below. 508 509 align with the zone size <-| 510 |-> align with the segment size 511 _________________________________________________________________________ 512 | | | Segment | Node | Segment | | 513 | Superblock | Checkpoint | Info. | Address | Summary | Main | 514 | (SB) | (CP) | Table (SIT) | Table (NAT) | Area (SSA) | | 515 |____________|_____2______|______N______|______N______|______N_____|__N___| 516 . . 517 . . 518 . . 519 ._________________________________________. 520 |_Segment_|_..._|_Segment_|_..._|_Segment_| 521 . . 522 ._________._________ 523 |_section_|__...__|_ 524 . . 525 .________. 526 |__zone__| 527 528- Superblock (SB) 529 : It is located at the beginning of the partition, and there exist two copies 530 to avoid file system crash. It contains basic partition information and some 531 default parameters of f2fs. 532 533- Checkpoint (CP) 534 : It contains file system information, bitmaps for valid NAT/SIT sets, orphan 535 inode lists, and summary entries of current active segments. 536 537- Segment Information Table (SIT) 538 : It contains segment information such as valid block count and bitmap for the 539 validity of all the blocks. 540 541- Node Address Table (NAT) 542 : It is composed of a block address table for all the node blocks stored in 543 Main area. 544 545- Segment Summary Area (SSA) 546 : It contains summary entries which contains the owner information of all the 547 data and node blocks stored in Main area. 548 549- Main Area 550 : It contains file and directory data including their indices. 551 552In order to avoid misalignment between file system and flash-based storage, F2FS 553aligns the start block address of CP with the segment size. Also, it aligns the 554start block address of Main area with the zone size by reserving some segments 555in SSA area. 556 557Reference the following survey for additional technical details. 558https://wiki.linaro.org/WorkingGroups/Kernel/Projects/FlashCardSurvey 559 560File System Metadata Structure 561------------------------------ 562 563F2FS adopts the checkpointing scheme to maintain file system consistency. At 564mount time, F2FS first tries to find the last valid checkpoint data by scanning 565CP area. In order to reduce the scanning time, F2FS uses only two copies of CP. 566One of them always indicates the last valid data, which is called as shadow copy 567mechanism. In addition to CP, NAT and SIT also adopt the shadow copy mechanism. 568 569For file system consistency, each CP points to which NAT and SIT copies are 570valid, as shown as below. 571 572 +--------+----------+---------+ 573 | CP | SIT | NAT | 574 +--------+----------+---------+ 575 . . . . 576 . . . . 577 . . . . 578 +-------+-------+--------+--------+--------+--------+ 579 | CP #0 | CP #1 | SIT #0 | SIT #1 | NAT #0 | NAT #1 | 580 +-------+-------+--------+--------+--------+--------+ 581 | ^ ^ 582 | | | 583 `----------------------------------------' 584 585Index Structure 586--------------- 587 588The key data structure to manage the data locations is a "node". Similar to 589traditional file structures, F2FS has three types of node: inode, direct node, 590indirect node. F2FS assigns 4KB to an inode block which contains 923 data block 591indices, two direct node pointers, two indirect node pointers, and one double 592indirect node pointer as described below. One direct node block contains 1018 593data blocks, and one indirect node block contains also 1018 node blocks. Thus, 594one inode block (i.e., a file) covers: 595 596 4KB * (923 + 2 * 1018 + 2 * 1018 * 1018 + 1018 * 1018 * 1018) := 3.94TB. 597 598 Inode block (4KB) 599 |- data (923) 600 |- direct node (2) 601 | `- data (1018) 602 |- indirect node (2) 603 | `- direct node (1018) 604 | `- data (1018) 605 `- double indirect node (1) 606 `- indirect node (1018) 607 `- direct node (1018) 608 `- data (1018) 609 610Note that, all the node blocks are mapped by NAT which means the location of 611each node is translated by the NAT table. In the consideration of the wandering 612tree problem, F2FS is able to cut off the propagation of node updates caused by 613leaf data writes. 614 615Directory Structure 616------------------- 617 618A directory entry occupies 11 bytes, which consists of the following attributes. 619 620- hash hash value of the file name 621- ino inode number 622- len the length of file name 623- type file type such as directory, symlink, etc 624 625A dentry block consists of 214 dentry slots and file names. Therein a bitmap is 626used to represent whether each dentry is valid or not. A dentry block occupies 6274KB with the following composition. 628 629 Dentry Block(4 K) = bitmap (27 bytes) + reserved (3 bytes) + 630 dentries(11 * 214 bytes) + file name (8 * 214 bytes) 631 632 [Bucket] 633 +--------------------------------+ 634 |dentry block 1 | dentry block 2 | 635 +--------------------------------+ 636 . . 637 . . 638 . [Dentry Block Structure: 4KB] . 639 +--------+----------+----------+------------+ 640 | bitmap | reserved | dentries | file names | 641 +--------+----------+----------+------------+ 642 [Dentry Block: 4KB] . . 643 . . 644 . . 645 +------+------+-----+------+ 646 | hash | ino | len | type | 647 +------+------+-----+------+ 648 [Dentry Structure: 11 bytes] 649 650F2FS implements multi-level hash tables for directory structure. Each level has 651a hash table with dedicated number of hash buckets as shown below. Note that 652"A(2B)" means a bucket includes 2 data blocks. 653 654---------------------- 655A : bucket 656B : block 657N : MAX_DIR_HASH_DEPTH 658---------------------- 659 660level #0 | A(2B) 661 | 662level #1 | A(2B) - A(2B) 663 | 664level #2 | A(2B) - A(2B) - A(2B) - A(2B) 665 . | . . . . 666level #N/2 | A(2B) - A(2B) - A(2B) - A(2B) - A(2B) - ... - A(2B) 667 . | . . . . 668level #N | A(4B) - A(4B) - A(4B) - A(4B) - A(4B) - ... - A(4B) 669 670The number of blocks and buckets are determined by, 671 672 ,- 2, if n < MAX_DIR_HASH_DEPTH / 2, 673 # of blocks in level #n = | 674 `- 4, Otherwise 675 676 ,- 2^(n + dir_level), 677 | if n + dir_level < MAX_DIR_HASH_DEPTH / 2, 678 # of buckets in level #n = | 679 `- 2^((MAX_DIR_HASH_DEPTH / 2) - 1), 680 Otherwise 681 682When F2FS finds a file name in a directory, at first a hash value of the file 683name is calculated. Then, F2FS scans the hash table in level #0 to find the 684dentry consisting of the file name and its inode number. If not found, F2FS 685scans the next hash table in level #1. In this way, F2FS scans hash tables in 686each levels incrementally from 1 to N. In each levels F2FS needs to scan only 687one bucket determined by the following equation, which shows O(log(# of files)) 688complexity. 689 690 bucket number to scan in level #n = (hash value) % (# of buckets in level #n) 691 692In the case of file creation, F2FS finds empty consecutive slots that cover the 693file name. F2FS searches the empty slots in the hash tables of whole levels from 6941 to N in the same way as the lookup operation. 695 696The following figure shows an example of two cases holding children. 697 --------------> Dir <-------------- 698 | | 699 child child 700 701 child - child [hole] - child 702 703 child - child - child [hole] - [hole] - child 704 705 Case 1: Case 2: 706 Number of children = 6, Number of children = 3, 707 File size = 7 File size = 7 708 709Default Block Allocation 710------------------------ 711 712At runtime, F2FS manages six active logs inside "Main" area: Hot/Warm/Cold node 713and Hot/Warm/Cold data. 714 715- Hot node contains direct node blocks of directories. 716- Warm node contains direct node blocks except hot node blocks. 717- Cold node contains indirect node blocks 718- Hot data contains dentry blocks 719- Warm data contains data blocks except hot and cold data blocks 720- Cold data contains multimedia data or migrated data blocks 721 722LFS has two schemes for free space management: threaded log and copy-and-compac- 723tion. The copy-and-compaction scheme which is known as cleaning, is well-suited 724for devices showing very good sequential write performance, since free segments 725are served all the time for writing new data. However, it suffers from cleaning 726overhead under high utilization. Contrarily, the threaded log scheme suffers 727from random writes, but no cleaning process is needed. F2FS adopts a hybrid 728scheme where the copy-and-compaction scheme is adopted by default, but the 729policy is dynamically changed to the threaded log scheme according to the file 730system status. 731 732In order to align F2FS with underlying flash-based storage, F2FS allocates a 733segment in a unit of section. F2FS expects that the section size would be the 734same as the unit size of garbage collection in FTL. Furthermore, with respect 735to the mapping granularity in FTL, F2FS allocates each section of the active 736logs from different zones as much as possible, since FTL can write the data in 737the active logs into one allocation unit according to its mapping granularity. 738 739Cleaning process 740---------------- 741 742F2FS does cleaning both on demand and in the background. On-demand cleaning is 743triggered when there are not enough free segments to serve VFS calls. Background 744cleaner is operated by a kernel thread, and triggers the cleaning job when the 745system is idle. 746 747F2FS supports two victim selection policies: greedy and cost-benefit algorithms. 748In the greedy algorithm, F2FS selects a victim segment having the smallest number 749of valid blocks. In the cost-benefit algorithm, F2FS selects a victim segment 750according to the segment age and the number of valid blocks in order to address 751log block thrashing problem in the greedy algorithm. F2FS adopts the greedy 752algorithm for on-demand cleaner, while background cleaner adopts cost-benefit 753algorithm. 754 755In order to identify whether the data in the victim segment are valid or not, 756F2FS manages a bitmap. Each bit represents the validity of a block, and the 757bitmap is composed of a bit stream covering whole blocks in main area. 758 759Write-hint Policy 760----------------- 761 7621) whint_mode=off. F2FS only passes down WRITE_LIFE_NOT_SET. 763 7642) whint_mode=user-based. F2FS tries to pass down hints given by 765users. 766 767User F2FS Block 768---- ---- ----- 769 META WRITE_LIFE_NOT_SET 770 HOT_NODE " 771 WARM_NODE " 772 COLD_NODE " 773*ioctl(COLD) COLD_DATA WRITE_LIFE_EXTREME 774*extension list " " 775 776-- buffered 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 " " 781WRITE_LIFE_MEDIUM " " 782WRITE_LIFE_LONG " " 783 784-- direct io 785WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME 786WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT 787WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_NOT_SET 788WRITE_LIFE_NONE " WRITE_LIFE_NONE 789WRITE_LIFE_MEDIUM " WRITE_LIFE_MEDIUM 790WRITE_LIFE_LONG " WRITE_LIFE_LONG 791 7923) whint_mode=fs-based. F2FS passes down hints with its policy. 793 794User F2FS Block 795---- ---- ----- 796 META WRITE_LIFE_MEDIUM; 797 HOT_NODE WRITE_LIFE_NOT_SET 798 WARM_NODE " 799 COLD_NODE WRITE_LIFE_NONE 800ioctl(COLD) COLD_DATA WRITE_LIFE_EXTREME 801extension list " " 802 803-- buffered io 804WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME 805WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT 806WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_LONG 807WRITE_LIFE_NONE " " 808WRITE_LIFE_MEDIUM " " 809WRITE_LIFE_LONG " " 810 811-- direct io 812WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME 813WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT 814WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_NOT_SET 815WRITE_LIFE_NONE " WRITE_LIFE_NONE 816WRITE_LIFE_MEDIUM " WRITE_LIFE_MEDIUM 817WRITE_LIFE_LONG " WRITE_LIFE_LONG 818 819Fallocate(2) Policy 820------------------- 821 822The default policy follows the below posix rule. 823 824Allocating disk space 825 The default operation (i.e., mode is zero) of fallocate() allocates 826 the disk space within the range specified by offset and len. The 827 file size (as reported by stat(2)) will be changed if offset+len is 828 greater than the file size. Any subregion within the range specified 829 by offset and len that did not contain data before the call will be 830 initialized to zero. This default behavior closely resembles the 831 behavior of the posix_fallocate(3) library function, and is intended 832 as a method of optimally implementing that function. 833 834However, once F2FS receives ioctl(fd, F2FS_IOC_SET_PIN_FILE) in prior to 835fallocate(fd, DEFAULT_MODE), it allocates on-disk blocks addressess having 836zero or random data, which is useful to the below scenario where: 837 1. create(fd) 838 2. ioctl(fd, F2FS_IOC_SET_PIN_FILE) 839 3. fallocate(fd, 0, 0, size) 840 4. address = fibmap(fd, offset) 841 5. open(blkdev) 842 6. write(blkdev, address) 843