littlefs

Protect against power loss while distributing flash erases

File operations provide strong copy-on-write guarantees so an interrupted operation can fall back to the last known good state. Small metadata logs and copy-on-write file structures share a block allocator that provides dynamic wear leveling across storage.

Sources: [1]

Fits embedded devices that store settings or logs in external flash

It fits devices that may lose power unexpectedly and need files and directories within limited RAM and ROM. Because users provide block-device operations and geometry through lfs_config, it can be integrated with custom boards and RTOS storage layers.

Sources: [1]

Verify real-device synchronization and the maintenance scope of version 2

Integrity depends on user-supplied read, program, erase, and sync functions plus correct flash geometry. If the device caches writes, sync must flush data to storage and guarantee that subsequent reads fetch the stored data.

The v2.11.3 release describes littlefs2 as feature-frozen and littlefs3 as in progress and unstable. It also fixes corruption involving multiple open write handles, so avoid freezing an old revision and test power interruption and concurrent-handle behavior on the real device.

Sources: [1][2]