tick_nohz_irq_exit(9)

📦 version 7.3.1, in nixpkgs unstable

💅 tick_nohz_irq_exit(9) — like, the Kernel API Manual, bestie ✨

okay so real talk: this is a page straight out of the Kernel API Manual, and it's about the function tick_nohz_irq_exit — the little "hey, we just left an IRQ" heads-up for the tick. short page, big vibes, let's go 🎀

🌸 NAME 🌸

tick_nohz_irq_exit — notify the tick about IRQ exit. that's it, that's the whole job 💖

✨ SYNOPSIS ✨

void tick_nohz_irq_exit (void );

🫶 ARGUMENTS 🫶

  • void — no arguments. none. zero. the function said "i need nothing from you" and honestly? iconic 🙈

🦋 DESCRIPTION 🦋

so here's the situation, bestie. a timer might have been added, modified, or deleted — and that could've happened either because of the current IRQ itself, or because of some other place that's using this IRQ as a notification. also possible: this IRQ updated the RCU callback list. point is, any of these events might mean the next tick needs a whole re-evaluation. and what happens next depends on the context, so let's break it down 👇

  1. if the CPU is idle and there's no resched pending, then you just proceed with idle time accounting. easy. the next tick gets re-evaluated on the next idle loop iteration. no drama, no cap 💅
  2. if the CPU is nohz_full:

2.1) if there's any tick dependency, restart the tick if it's stopped. like, wake it back up, it's needed 🔥

2.2) if there's no tick dependency, then (re-)evaluate the next tick and stop/update it accordingly. read that again: stop/update it accordingly — that's the move ✨

👑 SEE ALSO 👑

kernel file ./kernel/time/tick-sched.c, plus the whole squad:

ngl this one's short but it matters — miss the re-evaluation and the tick just sits there looking silly 😭 handle it, main character 🧚

aaand that's the page, bestie 🍓 it lives in ./kernel/time, dated October 2026, and yes — the header up top said tick_nohz_irq_exit(9) and the footer said tick_nohz_irq_exit(9) again, because it ate and left no crumbs 💜⭐🌈

📦 where this page comes from