Home | Notifications | New Note | Local | Federated | Search | Logout

Note Detail

Dr. Brian Callahan@bcallah@bsd.network (2026-08-11 23:04:28)
A wild #blog post appears!

I discuss a new #research publication: an empirical evaluation of the #OpenBSD -fret-clean flag. We examine the history of the mitigation, measure its costs, and deliberate whether or not it is worth keeping.

Worth the read if you like security and/or compilers.

https://briancallahan.net/blog/20260811.html

#freebsd #netbsd #dragonflybsd #bsd #linux #unix #solaris #illumos #compiler #compilers #llvm #gcc #rop #cybersecurity #cybersec #cyber #security #infosec #informationsecurity
---Reply---
Sun Microdevil Pte Ltd@koakuma@uwu.social (2026-08-13 10:02:40)
@bcallah I can't help but keep wondering how many of those mitigations would be rendered unnecessary under a different execution model (e.g. "safe" language and/or "safe" substate like CHERI)

Feels like what they do is mainly try to contain attackers if they somehow managed to break outta the semantics of the PL's abstract machine, but if you have safer execution model then such breakages are less likely to happen in the first place, no?