wait_event_...usive_locked(9)

💤 wait_event_...usive_locked(9) — the glow-up ✨

okay so real talk, this is a page from the Kernel API Manual (yes, the actual kernel manual, we're fancy now 💅) and the star of the show is wait_event_interruptible_exclusive_locked. long name? yes. iconic behavior? also yes. let's get into it bestie 🫶

---

🎀 NAME

wait_event_interruptible_exclusive_locked — sleep exclusively until a condition gets true 🦋

like, it naps, but it naps with main character energy — nobody else on the list gets to wake up with it. periodt.

---

✨ SYNOPSIS

wait_event_interruptible_exclusive_locked (wq , condition );

that's it. two things in, vibes out 💖 (also yes, that spacing is exactly how it's written, don't come for me)

---

🌸 ARGUMENTS

• wq — the waitqueue to wait on 🛏️ like, this is the bed your process is gonna crash in while it waits.
• condition — a C expression for the event to wait for ⭐ this is the "am I awake yet??" check.

---

💜 DESCRIPTION

okay so here's the whole situation, no cap:

The process is put to sleep (TASK_INTERRUPTIBLE) until the condition evaluates to true or a signal is received. and the condition gets checked every single time the waitqueue wq is woken up 🔥 like, every poke = one vibe check.

⚠️ big one: it must be called with wq.lock being held. like, that's not optional, that's the whole contract bestie.

That spinlock gets unlocked while sleeping 😴 but the condition testing is done while the lock is held, and when this macro exits, the lock is held. so it comes back exactly how it left — icon behavior, honestly.

The lock is locked/unlocked using spin_lock/spin_unlock functions, and those must match the way they are locked/unlocked outside of this macro. 🧚 matchy matchy or it does not slay.

The process is put on the wait queue with a WQ_FLAG_EXCLUSIVE flag set — and the way that works is, when other processes are waiting, if this process is awakened, further processes are not considered. 🙈 like sorry girlies, this one's mine.

wake_up_locked has to be called after changing any variable that could change the result of the wait condition. 🫡 that's the assignment, don't skip it.

Return values (the receipts 💀):

• -ERESTARTSYS if it was interrupted by a signal 😭
• 0 if the condition evaluated to true ✅

---

🌈 SEE ALSO

the whole friend group 🫶 also from the kernel file ./include/linux/wait.h:

waitqueue_active(9), wq_has_single_sleeper(9), wq_has_sleeper(9), wake_up_pollfree(9), wait_event(9), wait_event_freezable(9), wait_event_timeout(9), wait_event_cmd(9), wait_event_interruptible(9), wait_event_interruptible_timeout(9), wait_event_hrtimeout(9), wait_event_interruptible_hrtimeout(9), wait_event_idle(9), wait_event_idle_exclusive(9), wait_event_idle_timeout(9), wait_event_idle_exclusive_timeout(9), wait_event_interruptible_locked(9), wait_event_interruptible_locked_irq(9), wait_event_interruptible_exclusive_locked_irq(9), wait_event_killable(9), wait_event_state(9), wait_event_killable_timeout(9), wait_event_lock_irq_cmd(9), wait_event_lock_irq(9), wait_event_interruptible_lock_irq_cmd(9), wait_event_interruptible_lock_irq(9), wait_event_interruptible_lock_irq_timeout(9)

---

mwah 💋 that's the page!! this one lives at ./include/linux and the version is dated October 2026 — so yeah, future girlies are still out here waiting on waitqueues. go forth and sleep exclusively, bestie 👑✨