gfs2_file_write_iter(9)

💖 NAME 💖

okay so welcome to the gfs2_file_write_iter(9) page from the Kernel API Manual, bestie ✨ — gfs2_file_write_iter is literally here to perform a write to a file. iconic behavior, no cap 💅

---

✨ SYNOPSIS ✨

ssize_t gfs2_file_write_iter (struct kiocb *iocb , struct iov_iter *from
);

---

🫶 ARGUMENTS 🫶

• iocb — the io context. that's the context/vibe of this I/O moment, iykyk 🦋
• from — the data to write. yes bestie, this is the data going in 🎀

---

🎀 DESCRIPTION 🎀

real talk: we have to do a lock/unlock right here to refresh the inode size for O_APPEND writes. if we don't, we can land up writing at the wrong offset 😭💀

there is still a race — yes, even with this — but provided the app is using its own file locking, this will make O_APPEND work as expected. periodt 🔥

bestie please note: the race is still there. this is not a magic “no more races” button, it's a “makes O_APPEND behave as expected when the app does its own file locking” situation 🫶

---

🌈 SEE ALSO 🌈

Kernel file ./fs/gfs2/file.c is the home base, and the rest of the squad is:

• gfs2_llseek(9)
• gfs2_readdir(9)
• do_gfs2_set_flags(9)
• gfs2_size_hint(9)
• gfs2_allocate_folio_backing(9)
• gfs2_page_mkwrite(9)
• gfs2_mmap(9)
• gfs2_open_common(9)
• gfs2_open(9)
• gfs2_release(9)
• gfs2_fsync(9)
• calc_max_reserv(9)
• gfs2_lock(9)
• gfs2_flock(9)

---

mwah 💋 — this little page lives in ./fs/gfs2 and it's dated October 2026, slay 🎀✨