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 truned 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 Issue discard/TRIM commands when a segment is cleaned. 113no_heap Disable heap-style segment allocation which finds free 114 segments for data from the beginning of main area, while 115 for node from the end of main area. 116nouser_xattr Disable Extended User Attributes. Note: xattr is enabled 117 by default if CONFIG_F2FS_FS_XATTR is selected. 118noacl Disable POSIX Access Control List. Note: acl is enabled 119 by default if CONFIG_F2FS_FS_POSIX_ACL is selected. 120active_logs=%u Support configuring the number of active logs. In the 121 current design, f2fs supports only 2, 4, and 6 logs. 122 Default number is 6. 123disable_ext_identify Disable the extension list configured by mkfs, so f2fs 124 does not aware of cold files such as media files. 125inline_xattr Enable the inline xattrs feature. 126inline_data Enable the inline data feature: New created small(<~3.4k) 127 files can be written into inode block. 128inline_dentry Enable the inline dir feature: data in new created 129 directory entries can be written into inode block. The 130 space of inode block which is used to store inline 131 dentries is limited to ~3.4k. 132flush_merge Merge concurrent cache_flush commands as much as possible 133 to eliminate redundant command issues. If the underlying 134 device handles the cache_flush command relatively slowly, 135 recommend to enable this option. 136nobarrier This option can be used if underlying storage guarantees 137 its cached data should be written to the novolatile area. 138 If this option is set, no cache_flush commands are issued 139 but f2fs still guarantees the write ordering of all the 140 data writes. 141fastboot This option is used when a system wants to reduce mount 142 time as much as possible, even though normal performance 143 can be sacrificed. 144extent_cache Enable an extent cache based on rb-tree, it can cache 145 as many as extent which map between contiguous logical 146 address and physical address per inode, resulting in 147 increasing the cache hit ratio. Set by default. 148noextent_cache Diable an extent cache based on rb-tree explicitly, see 149 the above extent_cache mount option. 150noinline_data Disable the inline data feature, inline data feature is 151 enabled by default. 152 153================================================================================ 154DEBUGFS ENTRIES 155================================================================================ 156 157/sys/kernel/debug/f2fs/ contains information about all the partitions mounted as 158f2fs. Each file shows the whole f2fs information. 159 160/sys/kernel/debug/f2fs/status includes: 161 - major file system information managed by f2fs currently 162 - average SIT information about whole segments 163 - current memory footprint consumed by f2fs. 164 165================================================================================ 166SYSFS ENTRIES 167================================================================================ 168 169Information about mounted f2f2 file systems can be found in 170/sys/fs/f2fs. Each mounted filesystem will have a directory in 171/sys/fs/f2fs based on its device name (i.e., /sys/fs/f2fs/sda). 172The files in each per-device directory are shown in table below. 173 174Files in /sys/fs/f2fs/<devname> 175(see also Documentation/ABI/testing/sysfs-fs-f2fs) 176.............................................................................. 177 File Content 178 179 gc_max_sleep_time This tuning parameter controls the maximum sleep 180 time for the garbage collection thread. Time is 181 in milliseconds. 182 183 gc_min_sleep_time This tuning parameter controls the minimum sleep 184 time for the garbage collection thread. Time is 185 in milliseconds. 186 187 gc_no_gc_sleep_time This tuning parameter controls the default sleep 188 time for the garbage collection thread. Time is 189 in milliseconds. 190 191 gc_idle This parameter controls the selection of victim 192 policy for garbage collection. Setting gc_idle = 0 193 (default) will disable this option. Setting 194 gc_idle = 1 will select the Cost Benefit approach 195 & setting gc_idle = 2 will select the greedy aproach. 196 197 reclaim_segments This parameter controls the number of prefree 198 segments to be reclaimed. If the number of prefree 199 segments is larger than the number of segments 200 in the proportion to the percentage over total 201 volume size, f2fs tries to conduct checkpoint to 202 reclaim the prefree segments to free segments. 203 By default, 5% over total # of segments. 204 205 max_small_discards This parameter controls the number of discard 206 commands that consist small blocks less than 2MB. 207 The candidates to be discarded are cached until 208 checkpoint is triggered, and issued during the 209 checkpoint. By default, it is disabled with 0. 210 211 trim_sections This parameter controls the number of sections 212 to be trimmed out in batch mode when FITRIM 213 conducts. 32 sections is set by default. 214 215 ipu_policy This parameter controls the policy of in-place 216 updates in f2fs. There are five policies: 217 0x01: F2FS_IPU_FORCE, 0x02: F2FS_IPU_SSR, 218 0x04: F2FS_IPU_UTIL, 0x08: F2FS_IPU_SSR_UTIL, 219 0x10: F2FS_IPU_FSYNC. 220 221 min_ipu_util This parameter controls the threshold to trigger 222 in-place-updates. The number indicates percentage 223 of the filesystem utilization, and used by 224 F2FS_IPU_UTIL and F2FS_IPU_SSR_UTIL policies. 225 226 min_fsync_blocks This parameter controls the threshold to trigger 227 in-place-updates when F2FS_IPU_FSYNC mode is set. 228 The number indicates the number of dirty pages 229 when fsync needs to flush on its call path. If 230 the number is less than this value, it triggers 231 in-place-updates. 232 233 max_victim_search This parameter controls the number of trials to 234 find a victim segment when conducting SSR and 235 cleaning operations. The default value is 4096 236 which covers 8GB block address range. 237 238 dir_level This parameter controls the directory level to 239 support large directory. If a directory has a 240 number of files, it can reduce the file lookup 241 latency by increasing this dir_level value. 242 Otherwise, it needs to decrease this value to 243 reduce the space overhead. The default value is 0. 244 245 ram_thresh This parameter controls the memory footprint used 246 by free nids and cached nat entries. By default, 247 10 is set, which indicates 10 MB / 1 GB RAM. 248 249================================================================================ 250USAGE 251================================================================================ 252 2531. Download userland tools and compile them. 254 2552. Skip, if f2fs was compiled statically inside kernel. 256 Otherwise, insert the f2fs.ko module. 257 # insmod f2fs.ko 258 2593. Create a directory trying to mount 260 # mkdir /mnt/f2fs 261 2624. Format the block device, and then mount as f2fs 263 # mkfs.f2fs -l label /dev/block_device 264 # mount -t f2fs /dev/block_device /mnt/f2fs 265 266mkfs.f2fs 267--------- 268The mkfs.f2fs is for the use of formatting a partition as the f2fs filesystem, 269which builds a basic on-disk layout. 270 271The options consist of: 272-l [label] : Give a volume label, up to 512 unicode name. 273-a [0 or 1] : Split start location of each area for heap-based allocation. 274 1 is set by default, which performs this. 275-o [int] : Set overprovision ratio in percent over volume size. 276 5 is set by default. 277-s [int] : Set the number of segments per section. 278 1 is set by default. 279-z [int] : Set the number of sections per zone. 280 1 is set by default. 281-e [str] : Set basic extension list. e.g. "mp3,gif,mov" 282-t [0 or 1] : Disable discard command or not. 283 1 is set by default, which conducts discard. 284 285fsck.f2fs 286--------- 287The fsck.f2fs is a tool to check the consistency of an f2fs-formatted 288partition, which examines whether the filesystem metadata and user-made data 289are cross-referenced correctly or not. 290Note that, initial version of the tool does not fix any inconsistency. 291 292The options consist of: 293 -d debug level [default:0] 294 295dump.f2fs 296--------- 297The dump.f2fs shows the information of specific inode and dumps SSA and SIT to 298file. Each file is dump_ssa and dump_sit. 299 300The dump.f2fs is used to debug on-disk data structures of the f2fs filesystem. 301It shows on-disk inode information reconized by a given inode number, and is 302able to dump all the SSA and SIT entries into predefined files, ./dump_ssa and 303./dump_sit respectively. 304 305The options consist of: 306 -d debug level [default:0] 307 -i inode no (hex) 308 -s [SIT dump segno from #1~#2 (decimal), for all 0~-1] 309 -a [SSA dump segno from #1~#2 (decimal), for all 0~-1] 310 311Examples: 312# dump.f2fs -i [ino] /dev/sdx 313# dump.f2fs -s 0~-1 /dev/sdx (SIT dump) 314# dump.f2fs -a 0~-1 /dev/sdx (SSA dump) 315 316================================================================================ 317DESIGN 318================================================================================ 319 320On-disk Layout 321-------------- 322 323F2FS divides the whole volume into a number of segments, each of which is fixed 324to 2MB in size. A section is composed of consecutive segments, and a zone 325consists of a set of sections. By default, section and zone sizes are set to one 326segment size identically, but users can easily modify the sizes by mkfs. 327 328F2FS splits the entire volume into six areas, and all the areas except superblock 329consists of multiple segments as described below. 330 331 align with the zone size <-| 332 |-> align with the segment size 333 _________________________________________________________________________ 334 | | | Segment | Node | Segment | | 335 | Superblock | Checkpoint | Info. | Address | Summary | Main | 336 | (SB) | (CP) | Table (SIT) | Table (NAT) | Area (SSA) | | 337 |____________|_____2______|______N______|______N______|______N_____|__N___| 338 . . 339 . . 340 . . 341 ._________________________________________. 342 |_Segment_|_..._|_Segment_|_..._|_Segment_| 343 . . 344 ._________._________ 345 |_section_|__...__|_ 346 . . 347 .________. 348 |__zone__| 349 350- Superblock (SB) 351 : It is located at the beginning of the partition, and there exist two copies 352 to avoid file system crash. It contains basic partition information and some 353 default parameters of f2fs. 354 355- Checkpoint (CP) 356 : It contains file system information, bitmaps for valid NAT/SIT sets, orphan 357 inode lists, and summary entries of current active segments. 358 359- Segment Information Table (SIT) 360 : It contains segment information such as valid block count and bitmap for the 361 validity of all the blocks. 362 363- Node Address Table (NAT) 364 : It is composed of a block address table for all the node blocks stored in 365 Main area. 366 367- Segment Summary Area (SSA) 368 : It contains summary entries which contains the owner information of all the 369 data and node blocks stored in Main area. 370 371- Main Area 372 : It contains file and directory data including their indices. 373 374In order to avoid misalignment between file system and flash-based storage, F2FS 375aligns the start block address of CP with the segment size. Also, it aligns the 376start block address of Main area with the zone size by reserving some segments 377in SSA area. 378 379Reference the following survey for additional technical details. 380https://wiki.linaro.org/WorkingGroups/Kernel/Projects/FlashCardSurvey 381 382File System Metadata Structure 383------------------------------ 384 385F2FS adopts the checkpointing scheme to maintain file system consistency. At 386mount time, F2FS first tries to find the last valid checkpoint data by scanning 387CP area. In order to reduce the scanning time, F2FS uses only two copies of CP. 388One of them always indicates the last valid data, which is called as shadow copy 389mechanism. In addition to CP, NAT and SIT also adopt the shadow copy mechanism. 390 391For file system consistency, each CP points to which NAT and SIT copies are 392valid, as shown as below. 393 394 +--------+----------+---------+ 395 | CP | SIT | NAT | 396 +--------+----------+---------+ 397 . . . . 398 . . . . 399 . . . . 400 +-------+-------+--------+--------+--------+--------+ 401 | CP #0 | CP #1 | SIT #0 | SIT #1 | NAT #0 | NAT #1 | 402 +-------+-------+--------+--------+--------+--------+ 403 | ^ ^ 404 | | | 405 `----------------------------------------' 406 407Index Structure 408--------------- 409 410The key data structure to manage the data locations is a "node". Similar to 411traditional file structures, F2FS has three types of node: inode, direct node, 412indirect node. F2FS assigns 4KB to an inode block which contains 923 data block 413indices, two direct node pointers, two indirect node pointers, and one double 414indirect node pointer as described below. One direct node block contains 1018 415data blocks, and one indirect node block contains also 1018 node blocks. Thus, 416one inode block (i.e., a file) covers: 417 418 4KB * (923 + 2 * 1018 + 2 * 1018 * 1018 + 1018 * 1018 * 1018) := 3.94TB. 419 420 Inode block (4KB) 421 |- data (923) 422 |- direct node (2) 423 | `- data (1018) 424 |- indirect node (2) 425 | `- direct node (1018) 426 | `- data (1018) 427 `- double indirect node (1) 428 `- indirect node (1018) 429 `- direct node (1018) 430 `- data (1018) 431 432Note that, all the node blocks are mapped by NAT which means the location of 433each node is translated by the NAT table. In the consideration of the wandering 434tree problem, F2FS is able to cut off the propagation of node updates caused by 435leaf data writes. 436 437Directory Structure 438------------------- 439 440A directory entry occupies 11 bytes, which consists of the following attributes. 441 442- hash hash value of the file name 443- ino inode number 444- len the length of file name 445- type file type such as directory, symlink, etc 446 447A dentry block consists of 214 dentry slots and file names. Therein a bitmap is 448used to represent whether each dentry is valid or not. A dentry block occupies 4494KB with the following composition. 450 451 Dentry Block(4 K) = bitmap (27 bytes) + reserved (3 bytes) + 452 dentries(11 * 214 bytes) + file name (8 * 214 bytes) 453 454 [Bucket] 455 +--------------------------------+ 456 |dentry block 1 | dentry block 2 | 457 +--------------------------------+ 458 . . 459 . . 460 . [Dentry Block Structure: 4KB] . 461 +--------+----------+----------+------------+ 462 | bitmap | reserved | dentries | file names | 463 +--------+----------+----------+------------+ 464 [Dentry Block: 4KB] . . 465 . . 466 . . 467 +------+------+-----+------+ 468 | hash | ino | len | type | 469 +------+------+-----+------+ 470 [Dentry Structure: 11 bytes] 471 472F2FS implements multi-level hash tables for directory structure. Each level has 473a hash table with dedicated number of hash buckets as shown below. Note that 474"A(2B)" means a bucket includes 2 data blocks. 475 476---------------------- 477A : bucket 478B : block 479N : MAX_DIR_HASH_DEPTH 480---------------------- 481 482level #0 | A(2B) 483 | 484level #1 | A(2B) - A(2B) 485 | 486level #2 | A(2B) - A(2B) - A(2B) - A(2B) 487 . | . . . . 488level #N/2 | A(2B) - A(2B) - A(2B) - A(2B) - A(2B) - ... - A(2B) 489 . | . . . . 490level #N | A(4B) - A(4B) - A(4B) - A(4B) - A(4B) - ... - A(4B) 491 492The number of blocks and buckets are determined by, 493 494 ,- 2, if n < MAX_DIR_HASH_DEPTH / 2, 495 # of blocks in level #n = | 496 `- 4, Otherwise 497 498 ,- 2^(n + dir_level), 499 | if n + dir_level < MAX_DIR_HASH_DEPTH / 2, 500 # of buckets in level #n = | 501 `- 2^((MAX_DIR_HASH_DEPTH / 2) - 1), 502 Otherwise 503 504When F2FS finds a file name in a directory, at first a hash value of the file 505name is calculated. Then, F2FS scans the hash table in level #0 to find the 506dentry consisting of the file name and its inode number. If not found, F2FS 507scans the next hash table in level #1. In this way, F2FS scans hash tables in 508each levels incrementally from 1 to N. In each levels F2FS needs to scan only 509one bucket determined by the following equation, which shows O(log(# of files)) 510complexity. 511 512 bucket number to scan in level #n = (hash value) % (# of buckets in level #n) 513 514In the case of file creation, F2FS finds empty consecutive slots that cover the 515file name. F2FS searches the empty slots in the hash tables of whole levels from 5161 to N in the same way as the lookup operation. 517 518The following figure shows an example of two cases holding children. 519 --------------> Dir <-------------- 520 | | 521 child child 522 523 child - child [hole] - child 524 525 child - child - child [hole] - [hole] - child 526 527 Case 1: Case 2: 528 Number of children = 6, Number of children = 3, 529 File size = 7 File size = 7 530 531Default Block Allocation 532------------------------ 533 534At runtime, F2FS manages six active logs inside "Main" area: Hot/Warm/Cold node 535and Hot/Warm/Cold data. 536 537- Hot node contains direct node blocks of directories. 538- Warm node contains direct node blocks except hot node blocks. 539- Cold node contains indirect node blocks 540- Hot data contains dentry blocks 541- Warm data contains data blocks except hot and cold data blocks 542- Cold data contains multimedia data or migrated data blocks 543 544LFS has two schemes for free space management: threaded log and copy-and-compac- 545tion. The copy-and-compaction scheme which is known as cleaning, is well-suited 546for devices showing very good sequential write performance, since free segments 547are served all the time for writing new data. However, it suffers from cleaning 548overhead under high utilization. Contrarily, the threaded log scheme suffers 549from random writes, but no cleaning process is needed. F2FS adopts a hybrid 550scheme where the copy-and-compaction scheme is adopted by default, but the 551policy is dynamically changed to the threaded log scheme according to the file 552system status. 553 554In order to align F2FS with underlying flash-based storage, F2FS allocates a 555segment in a unit of section. F2FS expects that the section size would be the 556same as the unit size of garbage collection in FTL. Furthermore, with respect 557to the mapping granularity in FTL, F2FS allocates each section of the active 558logs from different zones as much as possible, since FTL can write the data in 559the active logs into one allocation unit according to its mapping granularity. 560 561Cleaning process 562---------------- 563 564F2FS does cleaning both on demand and in the background. On-demand cleaning is 565triggered when there are not enough free segments to serve VFS calls. Background 566cleaner is operated by a kernel thread, and triggers the cleaning job when the 567system is idle. 568 569F2FS supports two victim selection policies: greedy and cost-benefit algorithms. 570In the greedy algorithm, F2FS selects a victim segment having the smallest number 571of valid blocks. In the cost-benefit algorithm, F2FS selects a victim segment 572according to the segment age and the number of valid blocks in order to address 573log block thrashing problem in the greedy algorithm. F2FS adopts the greedy 574algorithm for on-demand cleaner, while background cleaner adopts cost-benefit 575algorithm. 576 577In order to identify whether the data in the victim segment are valid or not, 578F2FS manages a bitmap. Each bit represents the validity of a block, and the 579bitmap is composed of a bit stream covering whole blocks in main area. 580