refcount_inc_not_zero(9)

📦 version 7.3.1, in nixpkgs unstable

💖 refcount_inc_not_zero — kernel API manual, bestie edition 💖

okay so this one is a <pink>kernel API manual</pink> page, and honestly? it's a little baddie. let's get into it 🫶

✨ NAME ✨

refcount_inc_not_zero — increment a refcount unless it is 0 💅

like, it bumps the counter only if the counter isn't already sitting at zero. no zero, no bump. periodt.

🌸 SYNOPSIS 🌸

bool refcount_inc_not_zero (refcount_t *r );

🎀 ARGUMENTS 🎀

  • r — the refcount to increment. that's the one, that's the girl 💖

🦋 DESCRIPTION 🦋

so it's similar to atomic_inc_not_zero, but plot twist: it will saturate at REFCOUNT_SATURATED and WARN. so if you keep pushing it past the limit, it doesn't just keep going — it maxes out and then yells at you in the logs 😭 iconic behaviour ngl.

it provides no memory ordering. like, none. zero. the assumption is that the caller has already guaranteed the object memory is stable (RCU, etc.) — so that part is on you, bestie 🙈. it does provide a control dependency and thereby orders future stores. see the comment on top 👀 (iykyk).

no cap: if you're reaching for this expecting it to do memory-ordering work for you, it is not doing that. caller guarantees stability of the object memory. that's the deal.

🔥 RETURN 🔥

  • true if the increment was successful ⭐
  • false otherwise 💀

👑 SEE ALSO 👑

kernel file ./include/linux/refcount.h — and the whole refcount friend group, all together now:

🌈 that's the whole page, bestie, we ate and left no crumbs. this little number lives in ./include/linux, and it's dated October 2026 — so yes, it's from the future, main character energy fr fr 🧚💜

📦 where this page comes from