| blob | e5edd29687b50a74f8946ca215f639bb71b434fa |
1 ================================================================================
2 WHAT IS Flash-Friendly File System (F2FS)?
3 ================================================================================
5 NAND flash memory-based storage devices, such as SSD, eMMC, and SD cards, have
6 been equipped on a variety systems ranging from mobile to server systems. Since
7 they are known to have different characteristics from the conventional rotating
8 disks, a file system, an upper layer to the storage device, should adapt to the
9 changes from the sketch in the design level.
11 F2FS is a file system exploiting NAND flash memory-based storage devices, which
12 is based on Log-structured File System (LFS). The design has been focused on
13 addressing the fundamental issues in LFS, which are snowball effect of wandering
14 tree and high cleaning overhead.
16 Since a NAND flash memory-based storage device shows different characteristic
17 according to its internal geometry or flash memory management scheme, namely FTL,
18 F2FS and its tools support various parameters not only for configuring on-disk
19 layout, but also for selecting allocation and cleaning algorithms.
21 The following git tree provides the file system formatting tool (mkfs.f2fs),
22 a 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
25 For reporting bugs and sending patches, please use the following mailing list:
26 >> linux-f2fs-devel@lists.sourceforge.net
28 ================================================================================
29 BACKGROUND AND DESIGN ISSUES
30 ================================================================================
32 Log-structured File System (LFS)
33 --------------------------------
34 "A log-structured file system writes all modifications to disk sequentially in
35 a log-like structure, thereby speeding up both file writing and crash recovery.
36 The log is the only structure on disk; it contains indexing information so that
37 files can be read back from the log efficiently. In order to maintain large free
38 areas on disk for fast writing, we divide the log into segments and use a
39 segment cleaner to compress the live information from heavily fragmented
40 segments." from Rosenblum, M. and Ousterhout, J. K., 1992, "The design and
41 implementation of a log-structured file system", ACM Trans. Computer Systems
42 10, 1, 26–52.
44 Wandering Tree Problem
45 ----------------------
46 In LFS, when a file data is updated and written to the end of log, its direct
47 pointer block is updated due to the changed location. Then the indirect pointer
48 block is also updated due to the direct pointer block update. In this manner,
49 the upper index structures such as inode, inode map, and checkpoint block are
50 also updated recursively. This problem is called as wandering tree problem [1],
51 and in order to enhance the performance, it should eliminate or relax the update
52 propagation as much as possible.
54 [1] Bityutskiy, A. 2005. JFFS3 design issues. http://www.linux-mtd.infradead.org/
56 Cleaning Overhead
57 -----------------
58 Since LFS is based on out-of-place writes, it produces so many obsolete blocks
59 scattered across the whole storage. In order to serve new empty log space, it
60 needs to reclaim these obsolete blocks seamlessly to users. This job is called
61 as a cleaning process.
63 The process consists of three operations as follows.
64 1. A victim segment is selected through referencing segment usage table.
65 2. It loads parent index structures of all the data in the victim identified by
66 segment summary blocks.
67 3. It checks the cross-reference between the data and its parent index structure.
68 4. It moves valid data selectively.
70 This cleaning job may cause unexpected long delays, so the most important goal
71 is to hide the latencies to users. And also definitely, it should reduce the
72 amount of valid data to be moved, and move them quickly as well.
74 ================================================================================
75 KEY FEATURES
76 ================================================================================
78 Flash 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
84 Wandering 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.
90 Cleaning 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
97 ================================================================================
98 MOUNT OPTIONS
99 ================================================================================
101 background_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.
109 disable_roll_forward Disable the roll-forward recovery routine
110 norecovery Disable the roll-forward recovery routine, mounted read-
111 only (i.e., -o ro,disable_roll_forward)
112 discard/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.
115 no_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.
118 nouser_xattr Disable Extended User Attributes. Note: xattr is enabled
119 by default if CONFIG_F2FS_FS_XATTR is selected.
120 noacl Disable POSIX Access Control List. Note: acl is enabled
121 by default if CONFIG_F2FS_FS_POSIX_ACL is selected.
122 active_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.
125 disable_ext_identify Disable the extension list configured by mkfs, so f2fs
126 does not aware of cold files such as media files.
127 inline_xattr Enable the inline xattrs feature.
128 noinline_xattr Disable the inline xattrs feature.
129 inline_data Enable the inline data feature: New created small(<~3.4k)
130 files can be written into inode block.
131 inline_dentry Enable the inline dir feature: data in new created
132 directory entries can be written into inode block. The
133 space of inode block which is used to store inline
134 dentries is limited to ~3.4k.
135 noinline_dentry Disable the inline dentry feature.
136 flush_merge Merge concurrent cache_flush commands as much as possible
137 to eliminate redundant command issues. If the underlying
138 device handles the cache_flush command relatively slowly,
139 recommend to enable this option.
140 nobarrier This option can be used if underlying storage guarantees
141 its cached data should be written to the novolatile area.
142 If this option is set, no cache_flush commands are issued
143 but f2fs still guarantees the write ordering of all the
144 data writes.
145 fastboot This option is used when a system wants to reduce mount
146 time as much as possible, even though normal performance
147 can be sacrificed.
148 extent_cache Enable an extent cache based on rb-tree, it can cache
149 as many as extent which map between contiguous logical
150 address and physical address per inode, resulting in
151 increasing the cache hit ratio. Set by default.
152 noextent_cache Disable an extent cache based on rb-tree explicitly, see
153 the above extent_cache mount option.
154 noinline_data Disable the inline data feature, inline data feature is
155 enabled by default.
156 data_flush Enable data flushing before checkpoint in order to
157 persist data of regular and symlink.
158 fault_injection=%d Enable fault injection in all supported types with
159 specified injection rate.
160 fault_type=%d Support configuring fault injection type, should be
161 enabled with fault_injection option, fault type value
162 is shown below, it supports single or combined type.
163 Type_Name Type_Value
164 FAULT_KMALLOC 0x000000001
165 FAULT_KVMALLOC 0x000000002
166 FAULT_PAGE_ALLOC 0x000000004
167 FAULT_PAGE_GET 0x000000008
168 FAULT_ALLOC_BIO 0x000000010
169 FAULT_ALLOC_NID 0x000000020
170 FAULT_ORPHAN 0x000000040
171 FAULT_BLOCK 0x000000080
172 FAULT_DIR_DEPTH 0x000000100
173 FAULT_EVICT_INODE 0x000000200
174 FAULT_TRUNCATE 0x000000400
175 FAULT_IO 0x000000800
176 FAULT_CHECKPOINT 0x000001000
177 FAULT_DISCARD 0x000002000
178 mode=%s Control block allocation mode which supports "adaptive"
179 and "lfs". In "lfs" mode, there should be no random
180 writes towards main area.
181 io_bits=%u Set the bit size of write IO requests. It should be set
182 with "mode=lfs".
183 usrquota Enable plain user disk quota accounting.
184 grpquota Enable plain group disk quota accounting.
185 prjquota Enable plain project quota accounting.
186 usrjquota=<file> Appoint specified file and type during mount, so that quota
187 grpjquota=<file> information can be properly updated during recovery flow,
188 prjjquota=<file> <quota file>: must be in root directory;
189 jqfmt=<quota type> <quota type>: [vfsold,vfsv0,vfsv1].
190 offusrjquota Turn off user journelled quota.
191 offgrpjquota Turn off group journelled quota.
192 offprjjquota Turn off project journelled quota.
193 quota Enable plain user disk quota accounting.
194 noquota Disable all plain disk quota option.
195 whint_mode=%s Control which write hints are passed down to block
196 layer. This supports "off", "user-based", and
197 "fs-based". In "off" mode (default), f2fs does not pass
198 down hints. In "user-based" mode, f2fs tries to pass
199 down hints given by users. And in "fs-based" mode, f2fs
200 passes down hints with its policy.
201 alloc_mode=%s Adjust block allocation policy, which supports "reuse"
202 and "default".
203 fsync_mode=%s Control the policy of fsync. Currently supports "posix",
204 "strict", and "nobarrier". In "posix" mode, which is
205 default, fsync will follow POSIX semantics and does a
206 light operation to improve the filesystem performance.
207 In "strict" mode, fsync will be heavy and behaves in line
208 with xfs, ext4 and btrfs, where xfstest generic/342 will
209 pass, but the performance will regress. "nobarrier" is
210 based on "posix", but doesn't issue flush command for
211 non-atomic files likewise "nobarrier" mount option.
212 test_dummy_encryption Enable dummy encryption, which provides a fake fscrypt
213 context. The fake fscrypt context is used by xfstests.
215 ================================================================================
216 DEBUGFS ENTRIES
217 ================================================================================
219 /sys/kernel/debug/f2fs/ contains information about all the partitions mounted as
220 f2fs. Each file shows the whole f2fs information.
222 /sys/kernel/debug/f2fs/status includes:
223 - major file system information managed by f2fs currently
224 - average SIT information about whole segments
225 - current memory footprint consumed by f2fs.
227 ================================================================================
228 SYSFS ENTRIES
229 ================================================================================
231 Information about mounted f2fs file systems can be found in
232 /sys/fs/f2fs. Each mounted filesystem will have a directory in
233 /sys/fs/f2fs based on its device name (i.e., /sys/fs/f2fs/sda).
234 The files in each per-device directory are shown in table below.
236 Files in /sys/fs/f2fs/<devname>
237 (see also Documentation/ABI/testing/sysfs-fs-f2fs)
238 ..............................................................................
239 File Content
241 gc_max_sleep_time This tuning parameter controls the maximum sleep
242 time for the garbage collection thread. Time is
243 in milliseconds.
245 gc_min_sleep_time This tuning parameter controls the minimum sleep
246 time for the garbage collection thread. Time is
247 in milliseconds.
249 gc_no_gc_sleep_time This tuning parameter controls the default sleep
250 time for the garbage collection thread. Time is
251 in milliseconds.
253 gc_idle This parameter controls the selection of victim
254 policy for garbage collection. Setting gc_idle = 0
255 (default) will disable this option. Setting
256 gc_idle = 1 will select the Cost Benefit approach
257 & setting gc_idle = 2 will select the greedy approach.
259 gc_urgent This parameter controls triggering background GCs
260 urgently or not. Setting gc_urgent = 0 [default]
261 makes back to default behavior, while if it is set
262 to 1, background thread starts to do GC by given
263 gc_urgent_sleep_time interval.
265 gc_urgent_sleep_time This parameter controls sleep time for gc_urgent.
266 500 ms is set by default. See above gc_urgent.
268 reclaim_segments This parameter controls the number of prefree
269 segments to be reclaimed. If the number of prefree
270 segments is larger than the number of segments
271 in the proportion to the percentage over total
272 volume size, f2fs tries to conduct checkpoint to
273 reclaim the prefree segments to free segments.
274 By default, 5% over total # of segments.
276 max_small_discards This parameter controls the number of discard
277 commands that consist small blocks less than 2MB.
278 The candidates to be discarded are cached until
279 checkpoint is triggered, and issued during the
280 checkpoint. By default, it is disabled with 0.
282 trim_sections This parameter controls the number of sections
283 to be trimmed out in batch mode when FITRIM
284 conducts. 32 sections is set by default.
286 ipu_policy This parameter controls the policy of in-place
287 updates in f2fs. There are five policies:
288 0x01: F2FS_IPU_FORCE, 0x02: F2FS_IPU_SSR,
289 0x04: F2FS_IPU_UTIL, 0x08: F2FS_IPU_SSR_UTIL,
290 0x10: F2FS_IPU_FSYNC.
292 min_ipu_util This parameter controls the threshold to trigger
293 in-place-updates. The number indicates percentage
294 of the filesystem utilization, and used by
295 F2FS_IPU_UTIL and F2FS_IPU_SSR_UTIL policies.
297 min_fsync_blocks This parameter controls the threshold to trigger
298 in-place-updates when F2FS_IPU_FSYNC mode is set.
299 The number indicates the number of dirty pages
300 when fsync needs to flush on its call path. If
301 the number is less than this value, it triggers
302 in-place-updates.
304 max_victim_search This parameter controls the number of trials to
305 find a victim segment when conducting SSR and
306 cleaning operations. The default value is 4096
307 which covers 8GB block address range.
309 dir_level This parameter controls the directory level to
310 support large directory. If a directory has a
311 number of files, it can reduce the file lookup
312 latency by increasing this dir_level value.
313 Otherwise, it needs to decrease this value to
314 reduce the space overhead. The default value is 0.
316 ram_thresh This parameter controls the memory footprint used
317 by free nids and cached nat entries. By default,
318 10 is set, which indicates 10 MB / 1 GB RAM.
320 ================================================================================
321 USAGE
322 ================================================================================
324 1. Download userland tools and compile them.
326 2. Skip, if f2fs was compiled statically inside kernel.
327 Otherwise, insert the f2fs.ko module.
328 # insmod f2fs.ko
330 3. Create a directory trying to mount
331 # mkdir /mnt/f2fs
333 4. Format the block device, and then mount as f2fs
334 # mkfs.f2fs -l label /dev/block_device
335 # mount -t f2fs /dev/block_device /mnt/f2fs
337 mkfs.f2fs
338 ---------
339 The mkfs.f2fs is for the use of formatting a partition as the f2fs filesystem,
340 which builds a basic on-disk layout.
342 The options consist of:
343 -l [label] : Give a volume label, up to 512 unicode name.
344 -a [0 or 1] : Split start location of each area for heap-based allocation.
345 1 is set by default, which performs this.
346 -o [int] : Set overprovision ratio in percent over volume size.
347 5 is set by default.
348 -s [int] : Set the number of segments per section.
349 1 is set by default.
350 -z [int] : Set the number of sections per zone.
351 1 is set by default.
352 -e [str] : Set basic extension list. e.g. "mp3,gif,mov"
353 -t [0 or 1] : Disable discard command or not.
354 1 is set by default, which conducts discard.
356 fsck.f2fs
357 ---------
358 The fsck.f2fs is a tool to check the consistency of an f2fs-formatted
359 partition, which examines whether the filesystem metadata and user-made data
360 are cross-referenced correctly or not.
361 Note that, initial version of the tool does not fix any inconsistency.
363 The options consist of:
364 -d debug level [default:0]
366 dump.f2fs
367 ---------
368 The dump.f2fs shows the information of specific inode and dumps SSA and SIT to
369 file. Each file is dump_ssa and dump_sit.
371 The dump.f2fs is used to debug on-disk data structures of the f2fs filesystem.
372 It shows on-disk inode information recognized by a given inode number, and is
373 able to dump all the SSA and SIT entries into predefined files, ./dump_ssa and
374 ./dump_sit respectively.
376 The options consist of:
377 -d debug level [default:0]
378 -i inode no (hex)
379 -s [SIT dump segno from #1~#2 (decimal), for all 0~-1]
380 -a [SSA dump segno from #1~#2 (decimal), for all 0~-1]
382 Examples:
383 # dump.f2fs -i [ino] /dev/sdx
384 # dump.f2fs -s 0~-1 /dev/sdx (SIT dump)
385 # dump.f2fs -a 0~-1 /dev/sdx (SSA dump)
387 ================================================================================
388 DESIGN
389 ================================================================================
391 On-disk Layout
392 --------------
394 F2FS divides the whole volume into a number of segments, each of which is fixed
395 to 2MB in size. A section is composed of consecutive segments, and a zone
396 consists of a set of sections. By default, section and zone sizes are set to one
397 segment size identically, but users can easily modify the sizes by mkfs.
399 F2FS splits the entire volume into six areas, and all the areas except superblock
400 consists of multiple segments as described below.
402 align with the zone size <-|
403 |-> align with the segment size
404 _________________________________________________________________________
405 | | | Segment | Node | Segment | |
406 | Superblock | Checkpoint | Info. | Address | Summary | Main |
407 | (SB) | (CP) | Table (SIT) | Table (NAT) | Area (SSA) | |
408 |____________|_____2______|______N______|______N______|______N_____|__N___|
409 . .
410 . .
411 . .
412 ._________________________________________.
413 |_Segment_|_..._|_Segment_|_..._|_Segment_|
414 . .
415 ._________._________
416 |_section_|__...__|_
417 . .
418 .________.
419 |__zone__|
421 - Superblock (SB)
422 : It is located at the beginning of the partition, and there exist two copies
423 to avoid file system crash. It contains basic partition information and some
424 default parameters of f2fs.
426 - Checkpoint (CP)
427 : It contains file system information, bitmaps for valid NAT/SIT sets, orphan
428 inode lists, and summary entries of current active segments.
430 - Segment Information Table (SIT)
431 : It contains segment information such as valid block count and bitmap for the
432 validity of all the blocks.
434 - Node Address Table (NAT)
435 : It is composed of a block address table for all the node blocks stored in
436 Main area.
438 - Segment Summary Area (SSA)
439 : It contains summary entries which contains the owner information of all the
440 data and node blocks stored in Main area.
442 - Main Area
443 : It contains file and directory data including their indices.
445 In order to avoid misalignment between file system and flash-based storage, F2FS
446 aligns the start block address of CP with the segment size. Also, it aligns the
447 start block address of Main area with the zone size by reserving some segments
448 in SSA area.
450 Reference the following survey for additional technical details.
451 https://wiki.linaro.org/WorkingGroups/Kernel/Projects/FlashCardSurvey
453 File System Metadata Structure
454 ------------------------------
456 F2FS adopts the checkpointing scheme to maintain file system consistency. At
457 mount time, F2FS first tries to find the last valid checkpoint data by scanning
458 CP area. In order to reduce the scanning time, F2FS uses only two copies of CP.
459 One of them always indicates the last valid data, which is called as shadow copy
460 mechanism. In addition to CP, NAT and SIT also adopt the shadow copy mechanism.
462 For file system consistency, each CP points to which NAT and SIT copies are
463 valid, as shown as below.
465 +--------+----------+---------+
466 | CP | SIT | NAT |
467 +--------+----------+---------+
468 . . . .
469 . . . .
470 . . . .
471 +-------+-------+--------+--------+--------+--------+
472 | CP #0 | CP #1 | SIT #0 | SIT #1 | NAT #0 | NAT #1 |
473 +-------+-------+--------+--------+--------+--------+
474 | ^ ^
475 | | |
476 `----------------------------------------'
478 Index Structure
479 ---------------
481 The key data structure to manage the data locations is a "node". Similar to
482 traditional file structures, F2FS has three types of node: inode, direct node,
483 indirect node. F2FS assigns 4KB to an inode block which contains 923 data block
484 indices, two direct node pointers, two indirect node pointers, and one double
485 indirect node pointer as described below. One direct node block contains 1018
486 data blocks, and one indirect node block contains also 1018 node blocks. Thus,
487 one inode block (i.e., a file) covers:
489 4KB * (923 + 2 * 1018 + 2 * 1018 * 1018 + 1018 * 1018 * 1018) := 3.94TB.
491 Inode block (4KB)
492 |- data (923)
493 |- direct node (2)
494 | `- data (1018)
495 |- indirect node (2)
496 | `- direct node (1018)
497 | `- data (1018)
498 `- double indirect node (1)
499 `- indirect node (1018)
500 `- direct node (1018)
501 `- data (1018)
503 Note that, all the node blocks are mapped by NAT which means the location of
504 each node is translated by the NAT table. In the consideration of the wandering
505 tree problem, F2FS is able to cut off the propagation of node updates caused by
506 leaf data writes.
508 Directory Structure
509 -------------------
511 A directory entry occupies 11 bytes, which consists of the following attributes.
513 - hash hash value of the file name
514 - ino inode number
515 - len the length of file name
516 - type file type such as directory, symlink, etc
518 A dentry block consists of 214 dentry slots and file names. Therein a bitmap is
519 used to represent whether each dentry is valid or not. A dentry block occupies
520 4KB with the following composition.
522 Dentry Block(4 K) = bitmap (27 bytes) + reserved (3 bytes) +
523 dentries(11 * 214 bytes) + file name (8 * 214 bytes)
525 [Bucket]
526 +--------------------------------+
527 |dentry block 1 | dentry block 2 |
528 +--------------------------------+
529 . .
530 . .
531 . [Dentry Block Structure: 4KB] .
532 +--------+----------+----------+------------+
533 | bitmap | reserved | dentries | file names |
534 +--------+----------+----------+------------+
535 [Dentry Block: 4KB] . .
536 . .
537 . .
538 +------+------+-----+------+
539 | hash | ino | len | type |
540 +------+------+-----+------+
541 [Dentry Structure: 11 bytes]
543 F2FS implements multi-level hash tables for directory structure. Each level has
544 a hash table with dedicated number of hash buckets as shown below. Note that
545 "A(2B)" means a bucket includes 2 data blocks.
547 ----------------------
548 A : bucket
549 B : block
550 N : MAX_DIR_HASH_DEPTH
551 ----------------------
553 level #0 | A(2B)
554 |
555 level #1 | A(2B) - A(2B)
556 |
557 level #2 | A(2B) - A(2B) - A(2B) - A(2B)
558 . | . . . .
559 level #N/2 | A(2B) - A(2B) - A(2B) - A(2B) - A(2B) - ... - A(2B)
560 . | . . . .
561 level #N | A(4B) - A(4B) - A(4B) - A(4B) - A(4B) - ... - A(4B)
563 The number of blocks and buckets are determined by,
565 ,- 2, if n < MAX_DIR_HASH_DEPTH / 2,
566 # of blocks in level #n = |
567 `- 4, Otherwise
569 ,- 2^(n + dir_level),
570 | if n + dir_level < MAX_DIR_HASH_DEPTH / 2,
571 # of buckets in level #n = |
572 `- 2^((MAX_DIR_HASH_DEPTH / 2) - 1),
573 Otherwise
575 When F2FS finds a file name in a directory, at first a hash value of the file
576 name is calculated. Then, F2FS scans the hash table in level #0 to find the
577 dentry consisting of the file name and its inode number. If not found, F2FS
578 scans the next hash table in level #1. In this way, F2FS scans hash tables in
579 each levels incrementally from 1 to N. In each levels F2FS needs to scan only
580 one bucket determined by the following equation, which shows O(log(# of files))
581 complexity.
583 bucket number to scan in level #n = (hash value) % (# of buckets in level #n)
585 In the case of file creation, F2FS finds empty consecutive slots that cover the
586 file name. F2FS searches the empty slots in the hash tables of whole levels from
587 1 to N in the same way as the lookup operation.
589 The following figure shows an example of two cases holding children.
590 --------------> Dir <--------------
591 | |
592 child child
594 child - child [hole] - child
596 child - child - child [hole] - [hole] - child
598 Case 1: Case 2:
599 Number of children = 6, Number of children = 3,
600 File size = 7 File size = 7
602 Default Block Allocation
603 ------------------------
605 At runtime, F2FS manages six active logs inside "Main" area: Hot/Warm/Cold node
606 and Hot/Warm/Cold data.
608 - Hot node contains direct node blocks of directories.
609 - Warm node contains direct node blocks except hot node blocks.
610 - Cold node contains indirect node blocks
611 - Hot data contains dentry blocks
612 - Warm data contains data blocks except hot and cold data blocks
613 - Cold data contains multimedia data or migrated data blocks
615 LFS has two schemes for free space management: threaded log and copy-and-compac-
616 tion. The copy-and-compaction scheme which is known as cleaning, is well-suited
617 for devices showing very good sequential write performance, since free segments
618 are served all the time for writing new data. However, it suffers from cleaning
619 overhead under high utilization. Contrarily, the threaded log scheme suffers
620 from random writes, but no cleaning process is needed. F2FS adopts a hybrid
621 scheme where the copy-and-compaction scheme is adopted by default, but the
622 policy is dynamically changed to the threaded log scheme according to the file
623 system status.
625 In order to align F2FS with underlying flash-based storage, F2FS allocates a
626 segment in a unit of section. F2FS expects that the section size would be the
627 same as the unit size of garbage collection in FTL. Furthermore, with respect
628 to the mapping granularity in FTL, F2FS allocates each section of the active
629 logs from different zones as much as possible, since FTL can write the data in
630 the active logs into one allocation unit according to its mapping granularity.
632 Cleaning process
633 ----------------
635 F2FS does cleaning both on demand and in the background. On-demand cleaning is
636 triggered when there are not enough free segments to serve VFS calls. Background
637 cleaner is operated by a kernel thread, and triggers the cleaning job when the
638 system is idle.
640 F2FS supports two victim selection policies: greedy and cost-benefit algorithms.
641 In the greedy algorithm, F2FS selects a victim segment having the smallest number
642 of valid blocks. In the cost-benefit algorithm, F2FS selects a victim segment
643 according to the segment age and the number of valid blocks in order to address
644 log block thrashing problem in the greedy algorithm. F2FS adopts the greedy
645 algorithm for on-demand cleaner, while background cleaner adopts cost-benefit
646 algorithm.
648 In order to identify whether the data in the victim segment are valid or not,
649 F2FS manages a bitmap. Each bit represents the validity of a block, and the
650 bitmap is composed of a bit stream covering whole blocks in main area.
652 Write-hint Policy
653 -----------------
655 1) whint_mode=off. F2FS only passes down WRITE_LIFE_NOT_SET.
657 2) whint_mode=user-based. F2FS tries to pass down hints given by
658 users.
660 User F2FS Block
661 ---- ---- -----
662 META WRITE_LIFE_NOT_SET
663 HOT_NODE "
664 WARM_NODE "
665 COLD_NODE "
666 *ioctl(COLD) COLD_DATA WRITE_LIFE_EXTREME
667 *extension list " "
669 -- buffered io
670 WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME
671 WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT
672 WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_NOT_SET
673 WRITE_LIFE_NONE " "
674 WRITE_LIFE_MEDIUM " "
675 WRITE_LIFE_LONG " "
677 -- direct io
678 WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME
679 WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT
680 WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_NOT_SET
681 WRITE_LIFE_NONE " WRITE_LIFE_NONE
682 WRITE_LIFE_MEDIUM " WRITE_LIFE_MEDIUM
683 WRITE_LIFE_LONG " WRITE_LIFE_LONG
685 3) whint_mode=fs-based. F2FS passes down hints with its policy.
687 User F2FS Block
688 ---- ---- -----
689 META WRITE_LIFE_MEDIUM;
690 HOT_NODE WRITE_LIFE_NOT_SET
691 WARM_NODE "
692 COLD_NODE WRITE_LIFE_NONE
693 ioctl(COLD) COLD_DATA WRITE_LIFE_EXTREME
694 extension list " "
696 -- buffered io
697 WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME
698 WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT
699 WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_LONG
700 WRITE_LIFE_NONE " "
701 WRITE_LIFE_MEDIUM " "
702 WRITE_LIFE_LONG " "
704 -- direct io
705 WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME
706 WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT
707 WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_NOT_SET
708 WRITE_LIFE_NONE " WRITE_LIFE_NONE
709 WRITE_LIFE_MEDIUM " WRITE_LIFE_MEDIUM
710 WRITE_LIFE_LONG " WRITE_LIFE_LONG