this post was submitted on 03 Dec 2023
693 points (95.2% liked)
linuxmemes
20686 readers
1287 users here now
I use Arch btw
Sister communities:
- LemmyMemes: Memes
- LemmyShitpost: Anything and everything goes.
- RISA: Star Trek memes and shitposts
Community rules
- Follow the site-wide rules and code of conduct
- Be civil
- Post Linux-related content
- No recent reposts
Please report posts and comments that break these rules!
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
I doubt it. It's the halting problem. There are perfectly legitimate uses for similar things that you can't detect if it'll halt or not prior to running it. Maybe they'd patch it to avoid this specific string, but you'd just have to make something that looks like it could do something but never halts.
That's why I run all my terminal commands through ChatGPT to verify they aren't some sort of fork bomb. My system is unusably slow, but it's AI protected, futuristic, and super practical.
Seems inefficient, one should just integrate ChatGPT into Bash to automatically check these things.
You said 'ls' but did you really mean 'ls -la'? Imma go ahead and just give you the output from 'cat /dev/urandom' anyway.
I said "ls" but I really meant "sl". I just wanted to watch that steam locomotive animation.
They could always do what Android does and give you a prompt to force close an app that hangs for too long, or have a default subprocess limit and an optional whitelist of programs that can have as many subprocesses as they want.
The thing about fork bombs that it's not particular process which takes up all the resources, they're all doing nothing in a minimal amount of space. You could say "ok this group of processes is using a lot of resources" and kill it but then you're probably going to take down the whole user session as the starting point is not trivial to establish. Though I guess you could just kill all shells connected to the fork morass, won't fix the general case but it's a start. OTOH I don't think kernel devs are keen on special-case solutions.
You don't really have to kill every process, limiting spawning of new usermode processes after a limit has been reached should be enough, combine that with a warning and always reserving enouh resources for the kernel and critically important processes to remain working and the user should have all the tools needed to find what is causing the issue and kill the responsible processes
While nobody really cares enough to fix these kinds of problems for your basic home computer, I think this problem is mostly solved for cloud/virtualization providers