2WHAT IS Flash-Friendly File System (F2FS)?
   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.
  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.
  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.
  21The file system formatting tool, "mkfs.f2fs", is available from the following
  22git tree:
  23>> git://
  25For reporting bugs and sending patches, please use the following mailing list:
  32Log-structured File System (LFS)
  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.
  44Wandering Tree Problem
  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.
  54[1] Bityutskiy, A. 2005. JFFS3 design issues.
  56Cleaning Overhead
  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.
  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.
  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.
  78Flash Awareness
  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
  84Wandering Tree Problem
  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.
  90Cleaning Overhead
  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
 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.
 106                       Default value for this option is on. So garbage
 107                       collection is on by default.
 108disable_roll_forward   Disable the roll-forward recovery routine
 109discard                Issue discard/TRIM commands when a segment is cleaned.
 110no_heap                Disable heap-style segment allocation which finds free
 111                       segments for data from the beginning of main area, while
 112                       for node from the end of main area.
 113nouser_xattr           Disable Extended User Attributes. Note: xattr is enabled
 114                       by default if CONFIG_F2FS_FS_XATTR is selected.
 115noacl                  Disable POSIX Access Control List. Note: acl is enabled
 116                       by default if CONFIG_F2FS_FS_POSIX_ACL is selected.
 117active_logs=%u         Support configuring the number of active logs. In the
 118                       current design, f2fs supports only 2, 4, and 6 logs.
 119                       Default number is 6.
 120disable_ext_identify   Disable the extension list configured by mkfs, so f2fs
 121                       does not aware of cold files such as media files.
 127/sys/kernel/debug/f2fs/ contains information about all the partitions mounted as
 128f2fs. Each file shows the whole f2fs information.
 130/sys/kernel/debug/f2fs/status includes:
 131 - major file system information managed by f2fs currently
 132 - average SIT information about whole segments
 133 - current memory footprint consumed by f2fs.
 1391. Download userland tools and compile them.
 1412. Skip, if f2fs was compiled statically inside kernel.
 142   Otherwise, insert the f2fs.ko module.
 143 # insmod f2fs.ko
 1453. Create a directory trying to mount
 146 # mkdir /mnt/f2fs
 1484. Format the block device, and then mount as f2fs
 149 # mkfs.f2fs -l label /dev/block_device
 150 # mount -t f2fs /dev/block_device /mnt/f2fs
 152Format options
 154-l [label]   : Give a volume label, up to 512 unicode name.
 155-a [0 or 1]  : Split start location of each area for heap-based allocation.
 156               1 is set by default, which performs this.
 157-o [int]     : Set overprovision ratio in percent over volume size.
 158               5 is set by default.
 159-s [int]     : Set the number of segments per section.
 160               1 is set by default.
 161-z [int]     : Set the number of sections per zone.
 162               1 is set by default.
 163-e [str]     : Set basic extension list. e.g. "mp3,gif,mov"
 164-t [0 or 1]  : Disable discard command or not.
 165               1 is set by default, which conducts discard.
 171On-disk Layout
 174F2FS divides the whole volume into a number of segments, each of which is fixed
 175to 2MB in size. A section is composed of consecutive segments, and a zone
 176consists of a set of sections. By default, section and zone sizes are set to one
 177segment size identically, but users can easily modify the sizes by mkfs.
 179F2FS splits the entire volume into six areas, and all the areas except superblock
 180consists of multiple segments as described below.
 182                                            align with the zone size <-|
 183                 |-> align with the segment size
 184     _________________________________________________________________________
 185    |            |            |   Segment   |    Node     |   Segment  |      |
 186    | Superblock | Checkpoint |    Info.    |   Address   |   Summary  | Main |
 187    |    (SB)    |   (CP)     | Table (SIT) | Table (NAT) | Area (SSA) |      |
 188    |____________|_____2______|______N______|______N______|______N_____|__N___|
 189                                                                       .      .
 190                                                             .                .
 191                                                 .                            .
 192                                    ._________________________________________.
 193                                    |_Segment_|_..._|_Segment_|_..._|_Segment_|
 194                                    .           .
 195                                    ._________._________
 196                                    |_section_|__...__|_
 197                                    .            .
 198                                    .________.
 199                                    |__zone__|
 201- Superblock (SB)
 202 : It is located at the beginning of the partition, and there exist two copies
 203   to avoid file system crash. It contains basic partition information and some
 204   default parameters of f2fs.
 206- Checkpoint (CP)
 207 : It contains file system information, bitmaps for valid NAT/SIT sets, orphan
 208   inode lists, and summary entries of current active segments.
 210- Segment Information Table (SIT)
 211 : It contains segment information such as valid block count and bitmap for the
 212   validity of all the blocks.
 214- Node Address Table (NAT)
 215 : It is composed of a block address table for all the node blocks stored in
 216   Main area.
 218- Segment Summary Area (SSA)
 219 : It contains summary entries which contains the owner information of all the
 220   data and node blocks stored in Main area.
 222- Main Area
 223 : It contains file and directory data including their indices.
 225In order to avoid misalignment between file system and flash-based storage, F2FS
 226aligns the start block address of CP with the segment size. Also, it aligns the
 227start block address of Main area with the zone size by reserving some segments
 228in SSA area.
 230Reference the following survey for additional technical details.
 233File System Metadata Structure
 236F2FS adopts the checkpointing scheme to maintain file system consistency. At
 237mount time, F2FS first tries to find the last valid checkpoint data by scanning
 238CP area. In order to reduce the scanning time, F2FS uses only two copies of CP.
 239One of them always indicates the last valid data, which is called as shadow copy
 240mechanism. In addition to CP, NAT and SIT also adopt the shadow copy mechanism.
 242For file system consistency, each CP points to which NAT and SIT copies are
 243valid, as shown as below.
 245  +--------+----------+---------+
 246  |   CP   |    SIT   |   NAT   |
 247  +--------+----------+---------+
 248  .         .          .          .
 249  .            .              .              .
 250  .               .                 .                 .
 251  +-------+-------+--------+--------+--------+--------+
 252  | CP #0 | CP #1 | SIT #0 | SIT #1 | NAT #0 | NAT #1 |
 253  +-------+-------+--------+--------+--------+--------+
 254     |             ^                          ^
 255     |             |                          |
 256     `----------------------------------------'
 258Index Structure
 261The key data structure to manage the data locations is a "node". Similar to
 262traditional file structures, F2FS has three types of node: inode, direct node,
 263indirect node. F2FS assigns 4KB to an inode block which contains 923 data block
 264indices, two direct node pointers, two indirect node pointers, and one double
 265indirect node pointer as described below. One direct node block contains 1018
 266data blocks, and one indirect node block contains also 1018 node blocks. Thus,
 267one inode block (i.e., a file) covers:
 269  4KB * (923 + 2 * 1018 + 2 * 1018 * 1018 + 1018 * 1018 * 1018) := 3.94TB.
 271   Inode block (4KB)
 272     |- data (923)
 273     |- direct node (2)
 274     |          `- data (1018)
 275     |- indirect node (2)
 276     |            `- direct node (1018)
 277     |                       `- data (1018)
 278     `- double indirect node (1)
 279                         `- indirect node (1018)
 280                                      `- direct node (1018)
 281                                                 `- data (1018)
 283Note that, all the node blocks are mapped by NAT which means the location of
 284each node is translated by the NAT table. In the consideration of the wandering
 285tree problem, F2FS is able to cut off the propagation of node updates caused by
 286leaf data writes.
 288Directory Structure
 291A directory entry occupies 11 bytes, which consists of the following attributes.
 293- hash          hash value of the file name
 294- ino           inode number
 295- len           the length of file name
 296- type          file type such as directory, symlink, etc
 298A dentry block consists of 214 dentry slots and file names. Therein a bitmap is
 299used to represent whether each dentry is valid or not. A dentry block occupies
 3004KB with the following composition.
 302  Dentry Block(4 K) = bitmap (27 bytes) + reserved (3 bytes) +
 303                      dentries(11 * 214 bytes) + file name (8 * 214 bytes)
 305                         [Bucket]
 306             +--------------------------------+
 307             |dentry block 1 | dentry block 2 |
 308             +--------------------------------+
 309             .               .
 310       .                             .
 311  .       [Dentry Block Structure: 4KB]       .
 312  +--------+----------+----------+------------+
 313  | bitmap | reserved | dentries | file names |
 314  +--------+----------+----------+------------+
 315  [Dentry Block: 4KB] .   .
 316                 .               .
 317            .                          .
 318            +------+------+-----+------+
 319            | hash | ino  | len | type |
 320            +------+------+-----+------+
 321            [Dentry Structure: 11 bytes]
 323F2FS implements multi-level hash tables for directory structure. Each level has
 324a hash table with dedicated number of hash buckets as shown below. Note that
 325"A(2B)" means a bucket includes 2 data blocks.
 328A : bucket
 329B : block
 333level #0   | A(2B)
 334           |
 335level #1   | A(2B) - A(2B)
 336           |
 337level #2   | A(2B) - A(2B) - A(2B) - A(2B)
 338     .     |   .       .       .       .
 339level #N/2 | A(2B) - A(2B) - A(2B) - A(2B) - A(2B) - ... - A(2B)
 340     .     |   .       .       .       .
 341level #N   | A(4B) - A(4B) - A(4B) - A(4B) - A(4B) - ... - A(4B)
 343The number of blocks and buckets are determined by,
 345                            ,- 2, if n < MAX_DIR_HASH_DEPTH / 2,
 346  # of blocks in level #n = |
 347                            `- 4, Otherwise
 349                             ,- 2^n, if n < MAX_DIR_HASH_DEPTH / 2,
 350  # of buckets in level #n = |
 351                             `- 2^((MAX_DIR_HASH_DEPTH / 2) - 1), Otherwise
 353When F2FS finds a file name in a directory, at first a hash value of the file
 354name is calculated. Then, F2FS scans the hash table in level #0 to find the
 355dentry consisting of the file name and its inode number. If not found, F2FS
 356scans the next hash table in level #1. In this way, F2FS scans hash tables in
 357each levels incrementally from 1 to N. In each levels F2FS needs to scan only
 358one bucket determined by the following equation, which shows O(log(# of files))
 361  bucket number to scan in level #n = (hash value) % (# of buckets in level #n)
 363In the case of file creation, F2FS finds empty consecutive slots that cover the
 364file name. F2FS searches the empty slots in the hash tables of whole levels from
 3651 to N in the same way as the lookup operation.
 367The following figure shows an example of two cases holding children.
 368       --------------> Dir <--------------
 369       |                                 |
 370    child                             child
 372    child - child                     [hole] - child
 374    child - child - child             [hole] - [hole] - child
 376   Case 1:                           Case 2:
 377   Number of children = 6,           Number of children = 3,
 378   File size = 7                     File size = 7
 380Default Block Allocation
 383At runtime, F2FS manages six active logs inside "Main" area: Hot/Warm/Cold node
 384and Hot/Warm/Cold data.
 386- Hot node      contains direct node blocks of directories.
 387- Warm node     contains direct node blocks except hot node blocks.
 388- Cold node     contains indirect node blocks
 389- Hot data      contains dentry blocks
 390- Warm data     contains data blocks except hot and cold data blocks
 391- Cold data     contains multimedia data or migrated data blocks
 393LFS has two schemes for free space management: threaded log and copy-and-compac-
 394tion. The copy-and-compaction scheme which is known as cleaning, is well-suited
 395for devices showing very good sequential write performance, since free segments
 396are served all the time for writing new data. However, it suffers from cleaning
 397overhead under high utilization. Contrarily, the threaded log scheme suffers
 398from random writes, but no cleaning process is needed. F2FS adopts a hybrid
 399scheme where the copy-and-compaction scheme is adopted by default, but the
 400policy is dynamically changed to the threaded log scheme according to the file
 401system status.
 403In order to align F2FS with underlying flash-based storage, F2FS allocates a
 404segment in a unit of section. F2FS expects that the section size would be the
 405same as the unit size of garbage collection in FTL. Furthermore, with respect
 406to the mapping granularity in FTL, F2FS allocates each section of the active
 407logs from different zones as much as possible, since FTL can write the data in
 408the active logs into one allocation unit according to its mapping granularity.
 410Cleaning process
 413F2FS does cleaning both on demand and in the background. On-demand cleaning is
 414triggered when there are not enough free segments to serve VFS calls. Background
 415cleaner is operated by a kernel thread, and triggers the cleaning job when the
 416system is idle.
 418F2FS supports two victim selection policies: greedy and cost-benefit algorithms.
 419In the greedy algorithm, F2FS selects a victim segment having the smallest number
 420of valid blocks. In the cost-benefit algorithm, F2FS selects a victim segment
 421according to the segment age and the number of valid blocks in order to address
 422log block thrashing problem in the greedy algorithm. F2FS adopts the greedy
 423algorithm for on-demand cleaner, while background cleaner adopts cost-benefit
 426In order to identify whether the data in the victim segment are valid or not,
 427F2FS manages a bitmap. Each bit represents the validity of a block, and the
 428bitmap is composed of a bit stream covering whole blocks in main area.
 429 kindly hosted by Redpill Linpro AS, provider of Linux consulting and operations services since 1995.