scx_cmask_ref_init(9)

📦 version 7.3.1, in nixpkgs unstable

💅 scx_cmask_ref_init — the Kernel API Manual, bestie edition ✨

Okay so picture this: you're in the Kernel API Manual, section 9, and today's main character is scx_cmask_ref_init. This is the page that explains it, and yes I'm gonna walk you through every bit of it 🫶

🌸 NAME 🌸

scx_cmask_ref_init — bind a scx_cmask_ref to a BPF-arena cmask 🎀

Like, it's the little gluing ceremony where your ref gets attached to a cmask living in the BPF arena. Iconic.

✨ SYNOPSIS ✨

int scx_cmask_ref_init (struct scx_sched *sch , const struct scx_cmask
*src , struct scx_cmask_ref *ref );

Three args in, one int out. She's a function, she returns int, that's the whole vibe 🔥

🦋 ARGUMENTS 🦋

  • sch — the scheduler whose arena is hosting src. Like, it's the venue and src is the guest of honour, you know? 👑
  • src — the BPF-supplied cmask, already rebased to its kernel address. Very demure, very rebased 💜
  • ref — the output ref. This is where the result goes. It's giving "you had one job and it's to be the out-param" 💀

🧚 DESCRIPTION 🧚

So what she actually does is take a snapshot of src's base, nr_cids, and alloc_words 🍓

And the reason the snapshot is literally necessary? Because BPF may mutate the live header asynchronously 😭 So you can't just keep staring at the original header and hoping it stays still — it won't, she moves. Snapshot it, freeze the moment, move on. Main character energy 🌈

Return values, no cap:

  • 0 on success 🎉
  • -EINVAL if the range is out of bounds, or if alloc_words doesn't cover it 🙈
Fr fr, check your bounds and check that alloc_words actually covers the range, or you're getting -EINVAL and you will know about it.

🎀 SEE ALSO 🎀

Kernel file ./kernel/sched/ext/cid.c — and the whole squad, all of them:

💖 That's the whole page, bestie. Source lives at ./kernel/sched/ext, dated October 2026 — so go forth, snapshot your headers, respect the async mutation, and slay ✨💅

📦 where this page comes from