oom_killer_disable(9)

πŸ“¦ version 7.3.1, in nixpkgs unstable

πŸ’– oom_killer_disable β€” the kernel manual, but make it bestie ✨

okay so this is oom_killer_disable(9) from the Kernel API Manual and honestly? it's giving "hold my hand while I do something scary" energy πŸ’… buckle up bestie, we're talking about the OOM killer.


🌸 NAME 🌸

oom_killer_disable β€” literally just disables the OOM killer. that's the whole vibe. no cap.


✨ SYNOPSIS ✨

bool oom_killer_disable (signed long timeout );

look at her πŸ‘‘ one tiny function, one big responsibility.


πŸŽ€ ARGUMENTS πŸŽ€

  • timeout β€” the maximum timeout to wait for oom victims, measured in jiffies. so it's like, the deadline. the "girl I can't wait forever" number ⏰

πŸ¦‹ DESCRIPTION πŸ¦‹

okay so the way this works: it forces all page allocations to fail rather than trigger the OOM killer. instead of the kernel going full main-character and killing something, everything just... fails. iconic, honestly.

it will block and wait until all OOM victims are killed or the given timeout expires. so she's patient, but only up to a point 🫢

⚠️ big warning, bestie, listen up: the function cannot be called when there are runnable user tasks because userspace would see unexpected allocation failures as a result. like, imagine being a user task just vibing and suddenly every allocation fails? embarrassing. πŸ’€
πŸ™ˆ also: any new usage of this function should be consulted with MM people. yes, actual humans. the MM people. go talk to them, iykyk. don't just rawdog it into your patch fr fr.

Returns true if successful and false if the OOM killer cannot be disabled. simple as that 🌈


⭐ SEE ALSO ⭐


and that's the page, bestie!! πŸ“ source is ./mm, dated October 2026, coming at you straight from the Kernel API Manual β€” slay, stay allocated, and please talk to the MM people before you ship it πŸ’–βœ¨πŸŽ€

πŸ“¦ where this page comes from