On 06/21, Linus Torvalds wrote: > > On Wed, Jun 21, 2017 at 1:27 PM, Oleg Nesterov <[email protected]> wrote: > > > > Now __do_page_fault() tries to expand the stack itself, and this check > > fails. > > But we want that check to trigger and cause the access to fail. > Accessing the stack below the stack pointer is wrong.
I understand. My point is that this check was invalidated by stack-guard-page a long ago, and this means that we add the user-visible change now. > Do you have a pointer to the report for this regression? I must have missed > it. See http://marc.info/?t=149794523000001&r=1&w=2 this thread is a bit confusing, the most relevant emails are http://marc.info/?l=linux-kernel&m=149806256621440&w=2 (with test case) http://marc.info/?l=linux-kernel&m=149806452322233&w=2 Actually, I got another (redhat internal) bug report after this discussion, and at first glance this is the same thing. Just in case, I agree that mmap(MAP_GROWSDOWN) is mostly useless, perhaps we do not even care. But still this is regression. Oleg.
