writeback_inodes_sb(9)

πŸ“¦ version 7.3.1, in nixpkgs unstable

πŸ’– writeback_inodes_sb(9) πŸ’–

okay so real talk bestie, welcome to the writeback_inodes_sb(9) page β€” it lives in the Kernel API Manual and it's like, short but important, yk? grab a lil snack and let's get into it πŸ“


✨ NAME ✨

writeback_inodes_sb β€” writeback dirty inodes from given super_block πŸ’…


✨ SYNOPSIS ✨

void writeback_inodes_sb (struct super_block *sb , enum wb_reason reason
);

literally that's the whole signature. no cap, it's a void function so it doesn't hand anything back to you 🫢


✨ ARGUMENTS ✨

  • sb β€” the superblock. that's the one. the main character. πŸ‘‘
  • reason β€” the reason why some writeback work was initiated. like, why are we doing this, bestie? now it's on the record ✨

✨ DESCRIPTION ✨

So what it does: it starts writeback on some inodes on this super_block. that's the vibe. 🌈

And here's the part you gotta internalize, fr fr β€” no guarantees are made on how many (if any) will be written. could be a bunch, could be zero, could be one sad little inode. the universe decides. πŸ’€

Also, and this is the big one: this function does not wait for IO completion of submitted IO. so like, it kicks things off and then dips. it's giving "seen at 2:47 AM" energy. if you were expecting it to hang around until the IO is done, sorry bestie, that's not this function 😭

lowkey important: non-waiting = you can't assume anything is flushed to disk when it returns. that's the whole caveat, don't sleep on it πŸ™ˆ

✨ SEE ALSO ✨

Kernel file ./fs/fs-writeback.c β€” and honestly the whole squad shows up here, go say hi to all of them:


and that's the page!! πŸ¦‹ lil sign-off for you: this one came out of ./fs, dated October 2026, and it's filed under writeback_inodes_sb(9) β€” so if you ever forget where you read it, now you know πŸ’œβœ¨ periodt.

πŸ“¦ where this page comes from