Re: [RFC][PATCH] exec: Use init rlimits for setuid exec

From: Linus Torvalds
Date: Fri Jul 07 2017 - 01:04:09 EST

On Thu, Jul 6, 2017 at 9:48 PM, Andy Lutomirski <luto@xxxxxxxxxx> wrote:
> How about a much simpler solution: don't read rlimit at all in
> copy_strings(), let alone try to enforce it.

People have historically relied on E2BIG and then splitting things
into multiple chunks (ie do the whole 'xargs' thing).

But I agree that *if* we abort cleanly with E2BIG, then it's fine if
we copy a bit too much first. It's just that we need to check early
enough that an E2BIG error is possible (ie not after the "point of no

And yes, I think we can get rid of the layout differences based on
rlimit. After all, if I read that right the thing currently clamps
that "gap" thing to 128MB anyway (and that's on the *low* side).