Re: siginfo pid not populated from ptrace?

From: Kees Cook
Date: Thu Dec 06 2018 - 17:56:05 EST


On Thu, Dec 6, 2018 at 2:43 PM Eric W. Biederman <ebiederm@xxxxxxxxxxxx> wrote:
>
> Kees Cook <keescook@xxxxxxxxxxxx> writes:
> > What should we do for v4.20? I need to have the selftests actually
> > passing. :)
>
> For v4.20 we need to do one of two things.
> 1) Present a plausible case that someone will could care about,
> we document it in the commit we can perform my earlier partial revert.

If SIGSTOP si_pid can't be used to determine who sent the signal
reliably even before, then I'm guessing we'll never see a real-world
case where this matters.

> 2) Remove the sanity check seccomp_bpf.c
>
> I really just want to ensure we have clear reasoning here.

I'll remove it for now and add a link to this conversation, in case
anyone else goes looking.

-Kees

--
Kees Cook