oom_badness(9)

okay bestie, welcome to oom_badness(9) from the Kernel API Manual — yes, the whole page is about one very specific OOM heuristic, and it is lowkey giving main character energy. 💅✨

💖 NAME 💖

oom_badness — the heuristic function that determines which candidate task to kill. No cap, that's the whole vibe. 🫶

✨ SYNOPSIS ✨

long oom_badness (struct task_struct *p , unsigned long totalpages );

🌸 ARGUMENTS 🌸

• p — the task_struct of which task we should calculate. Like, whose badness are we even computing, bestie? That's the task in question. 🎀
• totalpages — total present RAM allowed for page allocation. That's the memory ceiling in the mix. ✨

🎀 DESCRIPTION 🎀

okay so real talk: the heuristic for determining which task to kill is made to be as simple and predictable as possible. That's the whole design brief, and honestly? iconic. 💅

The goal is to return the highest value for the task consuming the most memory to avoid subsequent oom failures. So yeah — biggest memory goblin gets the highest value, and that's on purpose to dodge the next oom situation. 🔥

🦋 SEE ALSO 🦋

Kernel file ./mm/oom_kill.c is the source-y bestie, and then there's the whole friend group:

• oom_cpuset_eligible(9)
• dump_tasks(9)
• mark_oom_victim(9)
• exit_oom_victim(9)
• oom_killer_enable(9)
• oom_killer_disable(9)
• out_of_memory(9)

shoutout to the source/version ./mm, the date October 2026, and oom_badness(9) — that's the footer, bestie, and that's a wrap 💖✨🎀