tjh.dev
Linux kernel mirror (for testing)
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel
os
linux
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.
160fault_injection=%d Enable fault injection in all supported types with
161 specified injection rate.
162fault_type=%d Support configuring fault injection type, should be
163 enabled with fault_injection option, fault type value
164 is shown below, it supports single or combined type.
165 Type_Name Type_Value
166 FAULT_KMALLOC 0x000000001
167 FAULT_KVMALLOC 0x000000002
168 FAULT_PAGE_ALLOC 0x000000004
169 FAULT_PAGE_GET 0x000000008
170 FAULT_ALLOC_BIO 0x000000010
171 FAULT_ALLOC_NID 0x000000020
172 FAULT_ORPHAN 0x000000040
173 FAULT_BLOCK 0x000000080
174 FAULT_DIR_DEPTH 0x000000100
175 FAULT_EVICT_INODE 0x000000200
176 FAULT_TRUNCATE 0x000000400
177 FAULT_READ_IO 0x000000800
178 FAULT_CHECKPOINT 0x000001000
179 FAULT_DISCARD 0x000002000
180 FAULT_WRITE_IO 0x000004000
181mode=%s Control block allocation mode which supports "adaptive"
182 and "lfs". In "lfs" mode, there should be no random
183 writes towards main area.
184io_bits=%u Set the bit size of write IO requests. It should be set
185 with "mode=lfs".
186usrquota Enable plain user disk quota accounting.
187grpquota Enable plain group disk quota accounting.
188prjquota Enable plain project quota accounting.
189usrjquota=<file> Appoint specified file and type during mount, so that quota
190grpjquota=<file> information can be properly updated during recovery flow,
191prjjquota=<file> <quota file>: must be in root directory;
192jqfmt=<quota type> <quota type>: [vfsold,vfsv0,vfsv1].
193offusrjquota Turn off user journelled quota.
194offgrpjquota Turn off group journelled quota.
195offprjjquota Turn off project journelled quota.
196quota Enable plain user disk quota accounting.
197noquota Disable all plain disk quota option.
198whint_mode=%s Control which write hints are passed down to block
199 layer. This supports "off", "user-based", and
200 "fs-based". In "off" mode (default), f2fs does not pass
201 down hints. In "user-based" mode, f2fs tries to pass
202 down hints given by users. And in "fs-based" mode, f2fs
203 passes down hints with its policy.
204alloc_mode=%s Adjust block allocation policy, which supports "reuse"
205 and "default".
206fsync_mode=%s Control the policy of fsync. Currently supports "posix",
207 "strict", and "nobarrier". In "posix" mode, which is
208 default, fsync will follow POSIX semantics and does a
209 light operation to improve the filesystem performance.
210 In "strict" mode, fsync will be heavy and behaves in line
211 with xfs, ext4 and btrfs, where xfstest generic/342 will
212 pass, but the performance will regress. "nobarrier" is
213 based on "posix", but doesn't issue flush command for
214 non-atomic files likewise "nobarrier" mount option.
215test_dummy_encryption Enable dummy encryption, which provides a fake fscrypt
216 context. The fake fscrypt context is used by xfstests.
217checkpoint=%s Set to "disable" to turn off checkpointing. Set to "enable"
218 to reenable checkpointing. Is enabled by default. While
219 disabled, any unmounting or unexpected shutdowns will cause
220 the filesystem contents to appear as they did when the
221 filesystem was mounted with that option.
222
223================================================================================
224DEBUGFS ENTRIES
225================================================================================
226
227/sys/kernel/debug/f2fs/ contains information about all the partitions mounted as
228f2fs. Each file shows the whole f2fs information.
229
230/sys/kernel/debug/f2fs/status includes:
231 - major file system information managed by f2fs currently
232 - average SIT information about whole segments
233 - current memory footprint consumed by f2fs.
234
235================================================================================
236SYSFS ENTRIES
237================================================================================
238
239Information about mounted f2fs file systems can be found in
240/sys/fs/f2fs. Each mounted filesystem will have a directory in
241/sys/fs/f2fs based on its device name (i.e., /sys/fs/f2fs/sda).
242The files in each per-device directory are shown in table below.
243
244Files in /sys/fs/f2fs/<devname>
245(see also Documentation/ABI/testing/sysfs-fs-f2fs)
246..............................................................................
247 File Content
248
249 gc_max_sleep_time This tuning parameter controls the maximum sleep
250 time for the garbage collection thread. Time is
251 in milliseconds.
252
253 gc_min_sleep_time This tuning parameter controls the minimum sleep
254 time for the garbage collection thread. Time is
255 in milliseconds.
256
257 gc_no_gc_sleep_time This tuning parameter controls the default sleep
258 time for the garbage collection thread. Time is
259 in milliseconds.
260
261 gc_idle This parameter controls the selection of victim
262 policy for garbage collection. Setting gc_idle = 0
263 (default) will disable this option. Setting
264 gc_idle = 1 will select the Cost Benefit approach
265 & setting gc_idle = 2 will select the greedy approach.
266
267 gc_urgent This parameter controls triggering background GCs
268 urgently or not. Setting gc_urgent = 0 [default]
269 makes back to default behavior, while if it is set
270 to 1, background thread starts to do GC by given
271 gc_urgent_sleep_time interval.
272
273 gc_urgent_sleep_time This parameter controls sleep time for gc_urgent.
274 500 ms is set by default. See above gc_urgent.
275
276 reclaim_segments This parameter controls the number of prefree
277 segments to be reclaimed. If the number of prefree
278 segments is larger than the number of segments
279 in the proportion to the percentage over total
280 volume size, f2fs tries to conduct checkpoint to
281 reclaim the prefree segments to free segments.
282 By default, 5% over total # of segments.
283
284 max_small_discards This parameter controls the number of discard
285 commands that consist small blocks less than 2MB.
286 The candidates to be discarded are cached until
287 checkpoint is triggered, and issued during the
288 checkpoint. By default, it is disabled with 0.
289
290 trim_sections This parameter controls the number of sections
291 to be trimmed out in batch mode when FITRIM
292 conducts. 32 sections is set by default.
293
294 ipu_policy This parameter controls the policy of in-place
295 updates in f2fs. There are five policies:
296 0x01: F2FS_IPU_FORCE, 0x02: F2FS_IPU_SSR,
297 0x04: F2FS_IPU_UTIL, 0x08: F2FS_IPU_SSR_UTIL,
298 0x10: F2FS_IPU_FSYNC.
299
300 min_ipu_util This parameter controls the threshold to trigger
301 in-place-updates. The number indicates percentage
302 of the filesystem utilization, and used by
303 F2FS_IPU_UTIL and F2FS_IPU_SSR_UTIL policies.
304
305 min_fsync_blocks This parameter controls the threshold to trigger
306 in-place-updates when F2FS_IPU_FSYNC mode is set.
307 The number indicates the number of dirty pages
308 when fsync needs to flush on its call path. If
309 the number is less than this value, it triggers
310 in-place-updates.
311
312 max_victim_search This parameter controls the number of trials to
313 find a victim segment when conducting SSR and
314 cleaning operations. The default value is 4096
315 which covers 8GB block address range.
316
317 dir_level This parameter controls the directory level to
318 support large directory. If a directory has a
319 number of files, it can reduce the file lookup
320 latency by increasing this dir_level value.
321 Otherwise, it needs to decrease this value to
322 reduce the space overhead. The default value is 0.
323
324 ram_thresh This parameter controls the memory footprint used
325 by free nids and cached nat entries. By default,
326 10 is set, which indicates 10 MB / 1 GB RAM.
327
328================================================================================
329USAGE
330================================================================================
331
3321. Download userland tools and compile them.
333
3342. Skip, if f2fs was compiled statically inside kernel.
335 Otherwise, insert the f2fs.ko module.
336 # insmod f2fs.ko
337
3383. Create a directory trying to mount
339 # mkdir /mnt/f2fs
340
3414. Format the block device, and then mount as f2fs
342 # mkfs.f2fs -l label /dev/block_device
343 # mount -t f2fs /dev/block_device /mnt/f2fs
344
345mkfs.f2fs
346---------
347The mkfs.f2fs is for the use of formatting a partition as the f2fs filesystem,
348which builds a basic on-disk layout.
349
350The options consist of:
351-l [label] : Give a volume label, up to 512 unicode name.
352-a [0 or 1] : Split start location of each area for heap-based allocation.
353 1 is set by default, which performs this.
354-o [int] : Set overprovision ratio in percent over volume size.
355 5 is set by default.
356-s [int] : Set the number of segments per section.
357 1 is set by default.
358-z [int] : Set the number of sections per zone.
359 1 is set by default.
360-e [str] : Set basic extension list. e.g. "mp3,gif,mov"
361-t [0 or 1] : Disable discard command or not.
362 1 is set by default, which conducts discard.
363
364fsck.f2fs
365---------
366The fsck.f2fs is a tool to check the consistency of an f2fs-formatted
367partition, which examines whether the filesystem metadata and user-made data
368are cross-referenced correctly or not.
369Note that, initial version of the tool does not fix any inconsistency.
370
371The options consist of:
372 -d debug level [default:0]
373
374dump.f2fs
375---------
376The dump.f2fs shows the information of specific inode and dumps SSA and SIT to
377file. Each file is dump_ssa and dump_sit.
378
379The dump.f2fs is used to debug on-disk data structures of the f2fs filesystem.
380It shows on-disk inode information recognized by a given inode number, and is
381able to dump all the SSA and SIT entries into predefined files, ./dump_ssa and
382./dump_sit respectively.
383
384The options consist of:
385 -d debug level [default:0]
386 -i inode no (hex)
387 -s [SIT dump segno from #1~#2 (decimal), for all 0~-1]
388 -a [SSA dump segno from #1~#2 (decimal), for all 0~-1]
389
390Examples:
391# dump.f2fs -i [ino] /dev/sdx
392# dump.f2fs -s 0~-1 /dev/sdx (SIT dump)
393# dump.f2fs -a 0~-1 /dev/sdx (SSA dump)
394
395================================================================================
396DESIGN
397================================================================================
398
399On-disk Layout
400--------------
401
402F2FS divides the whole volume into a number of segments, each of which is fixed
403to 2MB in size. A section is composed of consecutive segments, and a zone
404consists of a set of sections. By default, section and zone sizes are set to one
405segment size identically, but users can easily modify the sizes by mkfs.
406
407F2FS splits the entire volume into six areas, and all the areas except superblock
408consists of multiple segments as described below.
409
410 align with the zone size <-|
411 |-> align with the segment size
412 _________________________________________________________________________
413 | | | Segment | Node | Segment | |
414 | Superblock | Checkpoint | Info. | Address | Summary | Main |
415 | (SB) | (CP) | Table (SIT) | Table (NAT) | Area (SSA) | |
416 |____________|_____2______|______N______|______N______|______N_____|__N___|
417 . .
418 . .
419 . .
420 ._________________________________________.
421 |_Segment_|_..._|_Segment_|_..._|_Segment_|
422 . .
423 ._________._________
424 |_section_|__...__|_
425 . .
426 .________.
427 |__zone__|
428
429- Superblock (SB)
430 : It is located at the beginning of the partition, and there exist two copies
431 to avoid file system crash. It contains basic partition information and some
432 default parameters of f2fs.
433
434- Checkpoint (CP)
435 : It contains file system information, bitmaps for valid NAT/SIT sets, orphan
436 inode lists, and summary entries of current active segments.
437
438- Segment Information Table (SIT)
439 : It contains segment information such as valid block count and bitmap for the
440 validity of all the blocks.
441
442- Node Address Table (NAT)
443 : It is composed of a block address table for all the node blocks stored in
444 Main area.
445
446- Segment Summary Area (SSA)
447 : It contains summary entries which contains the owner information of all the
448 data and node blocks stored in Main area.
449
450- Main Area
451 : It contains file and directory data including their indices.
452
453In order to avoid misalignment between file system and flash-based storage, F2FS
454aligns the start block address of CP with the segment size. Also, it aligns the
455start block address of Main area with the zone size by reserving some segments
456in SSA area.
457
458Reference the following survey for additional technical details.
459https://wiki.linaro.org/WorkingGroups/Kernel/Projects/FlashCardSurvey
460
461File System Metadata Structure
462------------------------------
463
464F2FS adopts the checkpointing scheme to maintain file system consistency. At
465mount time, F2FS first tries to find the last valid checkpoint data by scanning
466CP area. In order to reduce the scanning time, F2FS uses only two copies of CP.
467One of them always indicates the last valid data, which is called as shadow copy
468mechanism. In addition to CP, NAT and SIT also adopt the shadow copy mechanism.
469
470For file system consistency, each CP points to which NAT and SIT copies are
471valid, as shown as below.
472
473 +--------+----------+---------+
474 | CP | SIT | NAT |
475 +--------+----------+---------+
476 . . . .
477 . . . .
478 . . . .
479 +-------+-------+--------+--------+--------+--------+
480 | CP #0 | CP #1 | SIT #0 | SIT #1 | NAT #0 | NAT #1 |
481 +-------+-------+--------+--------+--------+--------+
482 | ^ ^
483 | | |
484 `----------------------------------------'
485
486Index Structure
487---------------
488
489The key data structure to manage the data locations is a "node". Similar to
490traditional file structures, F2FS has three types of node: inode, direct node,
491indirect node. F2FS assigns 4KB to an inode block which contains 923 data block
492indices, two direct node pointers, two indirect node pointers, and one double
493indirect node pointer as described below. One direct node block contains 1018
494data blocks, and one indirect node block contains also 1018 node blocks. Thus,
495one inode block (i.e., a file) covers:
496
497 4KB * (923 + 2 * 1018 + 2 * 1018 * 1018 + 1018 * 1018 * 1018) := 3.94TB.
498
499 Inode block (4KB)
500 |- data (923)
501 |- direct node (2)
502 | `- data (1018)
503 |- indirect node (2)
504 | `- direct node (1018)
505 | `- data (1018)
506 `- double indirect node (1)
507 `- indirect node (1018)
508 `- direct node (1018)
509 `- data (1018)
510
511Note that, all the node blocks are mapped by NAT which means the location of
512each node is translated by the NAT table. In the consideration of the wandering
513tree problem, F2FS is able to cut off the propagation of node updates caused by
514leaf data writes.
515
516Directory Structure
517-------------------
518
519A directory entry occupies 11 bytes, which consists of the following attributes.
520
521- hash hash value of the file name
522- ino inode number
523- len the length of file name
524- type file type such as directory, symlink, etc
525
526A dentry block consists of 214 dentry slots and file names. Therein a bitmap is
527used to represent whether each dentry is valid or not. A dentry block occupies
5284KB with the following composition.
529
530 Dentry Block(4 K) = bitmap (27 bytes) + reserved (3 bytes) +
531 dentries(11 * 214 bytes) + file name (8 * 214 bytes)
532
533 [Bucket]
534 +--------------------------------+
535 |dentry block 1 | dentry block 2 |
536 +--------------------------------+
537 . .
538 . .
539 . [Dentry Block Structure: 4KB] .
540 +--------+----------+----------+------------+
541 | bitmap | reserved | dentries | file names |
542 +--------+----------+----------+------------+
543 [Dentry Block: 4KB] . .
544 . .
545 . .
546 +------+------+-----+------+
547 | hash | ino | len | type |
548 +------+------+-----+------+
549 [Dentry Structure: 11 bytes]
550
551F2FS implements multi-level hash tables for directory structure. Each level has
552a hash table with dedicated number of hash buckets as shown below. Note that
553"A(2B)" means a bucket includes 2 data blocks.
554
555----------------------
556A : bucket
557B : block
558N : MAX_DIR_HASH_DEPTH
559----------------------
560
561level #0 | A(2B)
562 |
563level #1 | A(2B) - A(2B)
564 |
565level #2 | A(2B) - A(2B) - A(2B) - A(2B)
566 . | . . . .
567level #N/2 | A(2B) - A(2B) - A(2B) - A(2B) - A(2B) - ... - A(2B)
568 . | . . . .
569level #N | A(4B) - A(4B) - A(4B) - A(4B) - A(4B) - ... - A(4B)
570
571The number of blocks and buckets are determined by,
572
573 ,- 2, if n < MAX_DIR_HASH_DEPTH / 2,
574 # of blocks in level #n = |
575 `- 4, Otherwise
576
577 ,- 2^(n + dir_level),
578 | if n + dir_level < MAX_DIR_HASH_DEPTH / 2,
579 # of buckets in level #n = |
580 `- 2^((MAX_DIR_HASH_DEPTH / 2) - 1),
581 Otherwise
582
583When F2FS finds a file name in a directory, at first a hash value of the file
584name is calculated. Then, F2FS scans the hash table in level #0 to find the
585dentry consisting of the file name and its inode number. If not found, F2FS
586scans the next hash table in level #1. In this way, F2FS scans hash tables in
587each levels incrementally from 1 to N. In each levels F2FS needs to scan only
588one bucket determined by the following equation, which shows O(log(# of files))
589complexity.
590
591 bucket number to scan in level #n = (hash value) % (# of buckets in level #n)
592
593In the case of file creation, F2FS finds empty consecutive slots that cover the
594file name. F2FS searches the empty slots in the hash tables of whole levels from
5951 to N in the same way as the lookup operation.
596
597The following figure shows an example of two cases holding children.
598 --------------> Dir <--------------
599 | |
600 child child
601
602 child - child [hole] - child
603
604 child - child - child [hole] - [hole] - child
605
606 Case 1: Case 2:
607 Number of children = 6, Number of children = 3,
608 File size = 7 File size = 7
609
610Default Block Allocation
611------------------------
612
613At runtime, F2FS manages six active logs inside "Main" area: Hot/Warm/Cold node
614and Hot/Warm/Cold data.
615
616- Hot node contains direct node blocks of directories.
617- Warm node contains direct node blocks except hot node blocks.
618- Cold node contains indirect node blocks
619- Hot data contains dentry blocks
620- Warm data contains data blocks except hot and cold data blocks
621- Cold data contains multimedia data or migrated data blocks
622
623LFS has two schemes for free space management: threaded log and copy-and-compac-
624tion. The copy-and-compaction scheme which is known as cleaning, is well-suited
625for devices showing very good sequential write performance, since free segments
626are served all the time for writing new data. However, it suffers from cleaning
627overhead under high utilization. Contrarily, the threaded log scheme suffers
628from random writes, but no cleaning process is needed. F2FS adopts a hybrid
629scheme where the copy-and-compaction scheme is adopted by default, but the
630policy is dynamically changed to the threaded log scheme according to the file
631system status.
632
633In order to align F2FS with underlying flash-based storage, F2FS allocates a
634segment in a unit of section. F2FS expects that the section size would be the
635same as the unit size of garbage collection in FTL. Furthermore, with respect
636to the mapping granularity in FTL, F2FS allocates each section of the active
637logs from different zones as much as possible, since FTL can write the data in
638the active logs into one allocation unit according to its mapping granularity.
639
640Cleaning process
641----------------
642
643F2FS does cleaning both on demand and in the background. On-demand cleaning is
644triggered when there are not enough free segments to serve VFS calls. Background
645cleaner is operated by a kernel thread, and triggers the cleaning job when the
646system is idle.
647
648F2FS supports two victim selection policies: greedy and cost-benefit algorithms.
649In the greedy algorithm, F2FS selects a victim segment having the smallest number
650of valid blocks. In the cost-benefit algorithm, F2FS selects a victim segment
651according to the segment age and the number of valid blocks in order to address
652log block thrashing problem in the greedy algorithm. F2FS adopts the greedy
653algorithm for on-demand cleaner, while background cleaner adopts cost-benefit
654algorithm.
655
656In order to identify whether the data in the victim segment are valid or not,
657F2FS manages a bitmap. Each bit represents the validity of a block, and the
658bitmap is composed of a bit stream covering whole blocks in main area.
659
660Write-hint Policy
661-----------------
662
6631) whint_mode=off. F2FS only passes down WRITE_LIFE_NOT_SET.
664
6652) whint_mode=user-based. F2FS tries to pass down hints given by
666users.
667
668User F2FS Block
669---- ---- -----
670 META WRITE_LIFE_NOT_SET
671 HOT_NODE "
672 WARM_NODE "
673 COLD_NODE "
674*ioctl(COLD) COLD_DATA WRITE_LIFE_EXTREME
675*extension list " "
676
677-- buffered io
678WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME
679WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT
680WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_NOT_SET
681WRITE_LIFE_NONE " "
682WRITE_LIFE_MEDIUM " "
683WRITE_LIFE_LONG " "
684
685-- direct io
686WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME
687WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT
688WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_NOT_SET
689WRITE_LIFE_NONE " WRITE_LIFE_NONE
690WRITE_LIFE_MEDIUM " WRITE_LIFE_MEDIUM
691WRITE_LIFE_LONG " WRITE_LIFE_LONG
692
6933) whint_mode=fs-based. F2FS passes down hints with its policy.
694
695User F2FS Block
696---- ---- -----
697 META WRITE_LIFE_MEDIUM;
698 HOT_NODE WRITE_LIFE_NOT_SET
699 WARM_NODE "
700 COLD_NODE WRITE_LIFE_NONE
701ioctl(COLD) COLD_DATA WRITE_LIFE_EXTREME
702extension list " "
703
704-- buffered io
705WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME
706WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT
707WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_LONG
708WRITE_LIFE_NONE " "
709WRITE_LIFE_MEDIUM " "
710WRITE_LIFE_LONG " "
711
712-- direct io
713WRITE_LIFE_EXTREME COLD_DATA WRITE_LIFE_EXTREME
714WRITE_LIFE_SHORT HOT_DATA WRITE_LIFE_SHORT
715WRITE_LIFE_NOT_SET WARM_DATA WRITE_LIFE_NOT_SET
716WRITE_LIFE_NONE " WRITE_LIFE_NONE
717WRITE_LIFE_MEDIUM " WRITE_LIFE_MEDIUM
718WRITE_LIFE_LONG " WRITE_LIFE_LONG