re: bruteforcing, seems you'd need to do some stronger killing (panic_on_oops-like, or our GRKERNSEC_KERN_LOCKOUT), as BUG()-like behavior won't cut it in many instances for preventing that
Conversation
and re: MarkUs, some of their optimizations (e.g. early unmapping) look incompatible with the Linux kernel heap at least
1
Replying to
Yeah, MarkUs-like approach is currently only being discussed in the context of userspace MTE