On Mon, Nov 11, 2019 at 5:58 PM Theodore Y. Ts'o [off-list ref] wrote:

On Mon, Nov 11, 2019 at 03:55:35PM +0100, Jann Horn wrote:

quoted

Not on Linux, but on OpenBSD, they do use MAP_STACK now AFAIK; this
was announced here:
<http://openbsd-archive.7691.n7.nabble.com/stack-register-checking-td338238.html>.
Basically they periodically check whether the userspace stack pointer
points into a MAP_STACK region, and if not, they kill the process. So
even if it's a no-op on Linux...

Hmm, is that something we should do in Linux?  Even if we only check
on syscall entry, which should be pretty inexpensive, it seems like it
would be very effective in protecting various ROP techniques.

I'm not a big fan, especially if that would only happen on syscall
entry; at the point where you have enough control to perform syscalls,
it probably isn't too difficult to move your ROP stack over to a
legitimate stack.