/tech/ - Technology

> M1 Macs also use Pointer Authentication Codes (PAC) to ensure forward and backward-edge protection.
You might want to include this for that part: https://www.kernel.org/doc/html/latest/arm64/pointer-authentication.html

Nice arguments
Too bad I didn't read them

>>17436
it's not "arguments", it's exposition of data /g/tard
If you use linux, it's worth taking a look at. I use qubes, i'm not anti-linux or anything. Knowledge is power comrade

>>17437
"exposition of data" have some balls you STEMlord

bumppp

tanenbaum was right about microkernels. having separate containers and systems that don't rely on each other is the most important thing in security right now. QubesOS literally runs a bunch of VMs, sandboxing sensitive ones from the safer ones (AdminVM, VaultVM…). Of course that won't happen to Linux because it's already a huge bloated codebase with millions of lines of code, so the only way out is doing something closer to OpenBSD, QubesOS or MINIX, if you don't want to contribute to those directly for some reason

>>18186
also, using a memory safe language is a must. C is completely outdated for modern OS development

Linux was never designed to be a super duper tight-asshole security magic pill. If you want that, go deal with OpenBSD or whatever.
>>18187
C can be memory-safe, it just requires a skilled eye for vulnerabilities instead of relying on a run-time program to handle it for you.

>>18186
HURD is already a functional OS!! It's microkerneled and modular all the way down

Well I wouldn't call it catch up. Linux is still secure in a lot of ways that it's corporate competition lags behind in. So like the user is in a pickle rn of chosing between a bunch of OSes that are all vulnerable in some glaring, publicly known way.

I assume these are the case for linux-hardened too, or is this just for the vanilla kernel?