truncate_inode_pages(9)

💅 truncate_inode_pages — the Kernel API Manual page, bestie ✨

okay so this is truncate_inode_pages(9) from the Kernel API Manual, and honestly? she's a baddie with a very specific job. let's get into it 🎀

---

✨ NAME ✨

truncate_inode_pages — truncate all the pages from an offset 🫶

(like, emphasis on all, the manpage literally put asterisks around it, she's not being subtle)

---

🌸 SYNOPSIS 🌸

void truncate_inode_pages (struct address_space *mapping , loff_t lstart
);

void, no return value, no notes, no nothing — she does the thing and dips 💅

---

🦋 ARGUMENTS 🦋

• mapping — the mapping you wanna truncate. that's the one, bestie 🎯
• lstart — the offset you're truncating from. everything from here on out? gone ✂️

---

💖 DESCRIPTION 💖

okay so the whole vibe here is: this gets called under — and is serialised by — inode->i_rwsem and mapping->invalidate_lock. so like, both locks, at the same time, no cap. 🔒🔒 iykyk.

---

👑 NOTE 👑

⚠️ bestie listen, this is the important bit ⚠️

when this function returns, there can still be a page in the process of deletion (it's inside __filemap_remove_folio) hanging out in the specified range. so mapping->nrpages can be non-zero when this function returns — even after truncating the whole entire mapping 😭

like she said she's done but one page is still in the group chat leaving. fr fr. not a bug, just the vibe 🌈

---

🔥 SEE ALSO 🔥

Kernel file ./mm/truncate.c and her whole friend group:

• folio_invalidate(9)
• mapping_evict_folio(9)
• truncate_inode_pages_range(9)
• truncate_inode_pages_final(9)
• mapping_try_invalidate(9)
• invalidate_mapping_pages(9)
• invalidate_inode_pages2_range(9)
• invalidate_inode_pages2(9)
• truncate_pagecache(9)
• truncate_setsize(9)
• pagecache_isize_extended(9)
• truncate_pagecache_range(9)

---

and that's the page, periodt 💜

signing off from the ./mm directory, October 2026 edition of this very iconic Kernel API Manual entry — mwah, ily, go read truncate_inode_pages_range(9) next 🧚‍♀️⭐