ib_unregiste...vice_and_put(9)

📦 version 7.3.1, in nixpkgs unstable

💅 ib_unregiste...vice_and_put(9) — bestie edition 🎀

so this is a page from the Kernel API Manual, and the star of the show is ib_unregister_device_and_put ✨ (yes the header abbreviates it as ib_unregiste...vice_and_put(9), that's just the title doing its little truncated thing, don't worry about it 💖)

🦋 NAME 🦋

ib_unregister_device_and_put — unregister a device while holding a 'get'. that's it. that's the whole vibe check 👑

✨ SYNOPSIS ✨

void ib_unregister_device_and_put (struct ib_device *ib_dev );

like, it returns void, so don't go looking for a return value, there isn't one, periodt 💀

🎀 ARGUMENTS 🎀

  • ib_dev — the device to unregister. the main character of this whole operation 💅

🌸 DESCRIPTION 🌸

okay so listen 🫶 — this function is literally the same as ib_unregister_device, except it also does an internal ib_device_put for you, and that put is supposed to match a 'get' that you (the caller) already grabbed. so you're not creating a get out of nowhere, you're squaring up a get you already had. no cap.

it is safe to call this routine concurrently from multiple threads while holding the 'get' 🔥 like, multiple threads can roll up at once, it's fine. and when the function returns, the device is fully unregistered. done. finished. ate and left no crumbs 💖

⚠️ drivers using this flow MUST use the driver_unregister callback to clean up their resources associated with the device and dealloc it. MUST. not "should", not "maybe if you feel like it" — MUST, bestie 😭🙈

🧚 SEE ALSO 🧚

the kernel file ./drivers/infiniband/core/device.c — that's where the magic lives 🦋

and the whole extended friend group, all of them, in order, no one left out 💜:

and that's the whole page, bestie 🌈✨ straight from ./drivers/infiniband/core, dated October 2026, in the Kernel API Manual — go forth and unregister responsibly 🫶💖🎀

📦 where this page comes from