compute_par...tive_cpumask(9)

📦 version 7.3.1, in nixpkgs unstable

💅 compute_partition_effective_cpumask(9) — from the Kernel API Manual, bestie 🎀

okay so picture this: it's a (9) manpage, it lives in the Kernel API Manual, and its whole personality is "figure out a partition's effective CPU mask." her name got a little chopped in the header (compute_par...tive_cpumask(9)) like she's texting from a bad signal, but she's iconic, we move. ✨

🌸 NAME 🌸

compute_partition_effective_cpumask — literally just computes effective_cpus for a partition. that's the whole gig, no cap 💖

🦋 SYNOPSIS 🦋

void compute_partition_effective_cpumask (struct cpuset *cs , struct cpumask *new_ecpus );

like yes, that's the signature, copy it exactly, do NOT freestyle it 🙈

🎀 ARGUMENTS 🎀

  • cs — the partition root cpuset. the main character of this whole operation 👑
  • new_ecpus — the effective_cpus that was previously computed, and which we're about to update. so she's not new-new, she's getting a glow-up ✨

💜 DESCRIPTION 💜

So here's the vibe: this function computes the effective_cpus of a partition root by scanning effective_xcpus of child partition roots and excluding their effective_xcpus. that's it, that's the math 🧮

She's got a side effect though — and it's a real one, so listen up: this has the side effect of invalidating valid child partition roots, if necessary. not me casually invalidating children 😭 but it's fine, it's on purpose.

Why is it fine? Because she's called from either cpuset_hotplug_update_tasks or update_cpumasks_hier, where parent and children are modified successively. so we don't need to call update_parent_effective_cpumask, and the child's effective_cpus will be updated in later iterations anyway. slay, deferred work, queen of not doing things twice 💅

🫶 little heads-up, bestie: rcu_read_lock is assumed to be held. assumed. like, she's trusting you. don't betray her. hold the lock fr fr.

🌈 SEE ALSO 🌈

Kernel file ./kernel/cgroup/cpuset.c, plus this absolutely stacked friend group 💖:

and that's the page!! 🍓 she ships from ./kernel/cgroup, dated October 2026, and she's filed under compute_par...tive_cpumask(9). mwah, go hold that rcu lock, ily 🧚💜

📦 where this page comes from