| Developer(s) | Nokia with help ofUniversity of Szeged |
|---|---|
| Full name | Unsorted Block Image File System |
| Introduced | 2008; 17 years ago (2008) withLinux kernel 2.6.27 |
| Structures | |
| Directory contents | B+ trees |
| Limits | |
| Allowed filename characters | Any byte except NUL and forward slash "/"[citation needed] |
| Features | |
| Forks | Yes |
| Attributes | Yes |
| File system permissions | Unix permissions |
| Transparent compression | Yes |
| Other | |
| Supported operating systems | Linux |
UBIFS (UBI File System, more fullyUnsorted Block Image File System) is aflash file system for unmanagedflash memory devices.[1]UBIFS works on top of anUBI (unsorted block image) layer,[2] which is itself on top of amemory technology device (MTD) layer.[3]The file system is developed by Nokia engineers with help of theUniversity of Szeged, Hungary. Development began in earnest in 2007, with the first stable release made toLinux kernel 2.6.27 in October 2008.[4]
Two major differences between UBIFS andJFFS2 are that UBIFS supports write caching,[5] and UBIFS errs on the pessimistic side of free space calculation.[6]UBIFS tends to perform better than JFFS2 for largeNAND flash memory devices.[7] This is a consequence of the UBIFS design goals:[8] faster mounting, quicker access to large files, and improved write speeds. UBIFS also preserves or improves uponJFFS2's on-the-fly compression, recoverability and power fail tolerance.[8] UBIFS's on-the-fly data compression allowszlib (deflate algorithm),LZO orZstandard.
UBIFS stores indexes in flash whereasJFFS2 stores filesystem indexes in memory.[9] This directly impacts the scalability ofJFFS2 as the tables must be rebuilt every time the volume is mounted. Also, the JFFS2 tables may consume enough system RAM that some images may be unusable.
UBI (Unsorted Block Images)[10] is an erase block management layer forflash memory devices. UBI serves two purposes, trackingNAND flash memory bad blocks and providingwear leveling. Wear leveling spreads the erases and writes across the entire flash device. UBI presents logical erase blocks to higher layers and maps these to physical erase blocks. UBI was written specifically for UBIFS so that UBIFS does not have to deal with wear leveling and bad blocks. However, UBI may also be useful withsquashfs and NAND flash;squashfs is not aware of NAND flash bad blocks.
UBI's documentation explains that it is not a completeflash translation layer (FTL). Although a FTL also handles bad blocks and wear leveling, the interface a FTL provides is a block device with small (typically 512 bytes) sectors that can be written completely independently. In contrast, UBI's interface directly exposes erase blocks and programmable pages (which are different sizes, and much larger than typical block device sectors), and filesystems that use UBI must be aware of the sizes and restrictions on how blocks must be erased before being written.
UBI is in some ways analogous to aLogical Volume Manager. In typical usage, rather than partitioning flash into fixed regions, a single UBI device spans the entire flash (except for perhaps a few pages in fixed locations reserved for the bootloader), and multiple volumes are created within the UBI device. This allows wear-leveling to be spread across the whole flash, even if some volumes are written more frequently than others. UBI volumes can bestatic (which contain a whole file or image written once and protected by CRC-32 by UBI) ordynamic (which contain a read-write filesystem that is responsible for its own data integrity). The only filesystem that directly supports UBI is UBIFS, but usinggluebi it's possible to emulate a MTD device, which can then be used to run other flash filesystems like JFFS2 and YAFFS, and usingubiblk it's possible to emulate block devices, which can run common filesystems like Ext4.
UBI was augmented in Linux 3.7 with fastmap support.[11][12] Fastmap maintains an on-disk version of information previously created in memory by scanning the entire flash device. The code falls back to the previous mechanism of a full scan on failures and older UBI systems will simply ignore the fastmap information.
It does not work on top of block devices. UBIFS was designed to work on top of raw flash.