Greetings.
I seek some clarification.
minijail in general assumes the threat model where an executable is initially trusted but can be exploited and become untrusted. That is my understanding after seeing LD_PRELOAD being used.
However in the event that one wishes to assume the executable itself is of minimal trustworthiness, would it be correct to be using T=static and assume that you get the same level of security as other sandboxing applications such as bubblewrap,nsjail,etc?
Additionally, my understanding behind the rationale for using LD_PRELOAD is the ability to more accurately target system calls for seccomp (forbid syscalls used by a libc after they are used at the start) and not having to use ambient capabilities. Are there any other benefits that arise?
Thank you.