On Fri, Jan 19, 2018 at 5:12 PM, David Windsor <dave@xxxxxxxxxxxx> wrote:
> I have some spare cycles; is there any more relevant information outside of this thread?
Awesome, thanks! Context is in the other commits, but mainly I want to
double-check that nothing breaks with these changes, and that all the
races for changing stack rlimits during exec are fixed. And then, just
a sanity-check that the design approach to attaching the stack limit
to the bprm isn't crazy. :)
-Kees
>>> [1] 04e35f4495dd ("exec: avoid RLIMIT_STACK races with prlimit()")
>>> [2] 779f4e1c6c7c ("Revert "exec: avoid RLIMIT_STACK races with prlimit()"")
>>> [3] to security@xxxxxxxxxx, "Subject: existing rlimit races?"
--
Kees Cook
Pixel Security
