On Thu, Jul 6, 2017 at 10:45 PM, Kees Cook <keesc...@chromium.org> wrote: > On Thu, Jul 6, 2017 at 10:36 PM, Andy Lutomirski <l...@kernel.org> wrote: >> Aren't there real use cases that use many megs of arguments? > > They'd be relatively new since the args were pretty limited before. > I'd be curious to see them. > >> We could probably get away with saying max(rlimit(RLIMIT_STACK), 2MB) >> as long as we make sure later on that we don't screw up if we've >> overallocated? > > min, not max, but yeah. Here's part of what I have for get_arg_page(): > > rlim = current->signal->rlim; > - if (size > READ_ONCE(rlim[RLIMIT_STACK].rlim_cur) / 4) > + arg_stack = READ_ONCE(rlim[RLIMIT_STACK].rlim_cur); > + arg_stack = min_t(unsigned long, arg_stack, _STK_LIM) / 4; > + if (size > arg_stack) > goto fail;
I really did mean max, the idea being that, if we're going to increase rlim_cur, it's a bit odd to fail the exec if it would have worked under the higher value. That being said, I see no real exploit vector here if just rlimit(RLIMIT_STACK) is used. (Can you just use rlimit()? The open-coding seems entirely useless.) I thought of another approach, though: change the rlimit macros so that a secureexec program always gets at least 8MB stack. Might be less regression-prone.
