nfs_start_io_read(9)

💖 nfs_start_io_read — it's giving "this file is doing buffered reads now" ✨

So bestie, this one's from the Kernel API Manual and it's lowkey iconic — it's the little function that tells the kernel "hey, a buffered read is about to happen, pls handle the locking for me" 🫶

---

✨ SYNOPSIS ✨

int nfs_start_io_read (struct inode *inode );

---

🌸 ARGUMENTS 🌸

• inode — the file inode, aka the main character of this whole operation 👑

---

🦋 DESCRIPTION 🦋

Okay so what this does: it declares that a buffered read operation is about to start, and on top of that it makes sure we block all direct I/O 💅

On exit, the function makes sure that the NFS_INO_ODIRECT flag is unset, and it holds a shared lock on inode->i_rwsem to ensure that the flag can't be changed. Fr fr, that's the whole safety mechanism.

In practice, this means buffered read operations are allowed to execute in parallel — thanks to the shared lock, bestie, it's teamwork ✨ — whereas direct I/O operations need to wait to grab an exclusive lock in order to set NFS_INO_ODIRECT. So they're literally queuing up 🙈

Heads up: buffered writes and truncates both take a write lock on inode->i_rwsem, meaning that those are serialised with respect to the reads. Not parallel. Serialised. Periodt. 💀

---

👑 SEE ALSO 👑

• Kernel file ./fs/nfs/io.c
• nfs_end_io_read(9)
• nfs_start_io_write(9)
• nfs_end_io_write(9)
• nfs_start_io_direct(9)
• nfs_start_io_direct_nowait(9)
• nfs_end_io_direct(9)

---

That's the whole vibe check, bestie — go read those siblings next, iykyk 🌈💜

Source: ./fs/nfs · Date: October 2026 · Kernel API Manual · nfs_start_io_read(9) — mwah, hope this slayed for you 🎀🦋💖