inode_newsize_ok(9)

🌸 NAME 🌸

inode_newsize_ok(9) — bestie, this is the Kernel API Manual entry for the function that asks: may this inode be truncated to a given size? ✨

💅 SYNOPSIS 💅

int inode_newsize_ok (const struct inode *inode , loff_t offset );

🎀 ARGUMENTS 🎀

• inode — the inode that's getting truncated, aka the main character here. 💖
• offset — the new size you want to assign to that inode. Not a suggestion, bestie.

🦋 DESCRIPTION 🦋

Okay so first rule: inode_newsize_ok must be called with i_rwsem held exclusively. Like, not optional, not “oops I forgot” — hold it exclusively. 💅

Then it checks filesystem limits and ulimits to make sure the new inode size is within limits. It will also send SIGXFSZ when necessary. Lowkey important. 👑

⚠️ If failure is returned, the caller must not proceed with the inode size change. Periodt.

Also: inode must be a file (not a directory), with appropriate permissions to allow truncate. And because this is giving “read the room”: inode_newsize_ok does NOT check these conditions. That's on the caller, bestie — no cap. 😭

🔥 RETURN 🔥

• 0 on success — slay. ✨
• -ve errno on failure — not slay, handle it. 💀

🌈 SEE ALSO 🌈

Kernel file ./fs/attr.c: setattr_should_drop_sgid(9), setattr_should_drop_suidgid(9), chown_ok(9), chgrp_ok(9), setattr_prepare(9), setattr_copy_mgtime(9), setattr_copy(9), notify_change(9).

mwah 💋 — source/version ./fs, date October 2026, page inode_newsize_ok(9). Byeee 🫶