[Bug 277382] Wrong (or at least strange) stack growth for amd64 pthread stacks

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=277382

--- Comment #4 from Paul Floyd <pjfloyd@wanadoo.fr> ---
The mmap and mprotect are done in libth thr_stack.c

265                  /* Map the stack and guard page together, and split guard
266                     page from allocated space: */
267                  if ((stackaddr = mmap(stackaddr, stacksize + guardsize,
268                       _rtld_get_stack_prot(), MAP_STACK,
269                       -1, 0)) != MAP_FAILED &&
270                      (guardsize == 0 ||
271                       mprotect(stackaddr, guardsize, PROT_NONE) == 0)) {
272                          stackaddr += guardsize;

From what I see that can the default can be overridden  with
_thr_attr_setguardsize but otherwise it is set to getpagesize().

Maybe libthr should check security.bsd.stack_guard_page and if it is 1 or more
not add the extra guard page.

-- 
You are receiving this mail because:
You are the assignee for the bug.