Re: [PATCH 09/10] mm: Prevent madvise from changing shadow stack

From: Yu-cheng Yu
Date: Thu Jun 07 2018 - 17:21:26 EST


On Thu, 2018-06-07 at 14:09 -0700, Nadav Amit wrote:
> Yu-cheng Yu <yu-cheng.yu@xxxxxxxxx> wrote:
>
> > Signed-off-by: Yu-cheng Yu <yu-cheng.yu@xxxxxxxxx>
> > ---
> > mm/madvise.c | 9 +++++++++
> > 1 file changed, 9 insertions(+)
> >
> > diff --git a/mm/madvise.c b/mm/madvise.c
> > index 4d3c922ea1a1..2a6988badd6b 100644
> > --- a/mm/madvise.c
> > +++ b/mm/madvise.c
> > @@ -839,6 +839,14 @@ SYSCALL_DEFINE3(madvise, unsigned long, start, size_t, len_in, int, behavior)
> > if (vma && start > vma->vm_start)
> > prev = vma;
> >
> > + /*
> > + * Don't do anything on shadow stack.
> > + */
> > + if (vma->vm_flags & VM_SHSTK) {
> > + error = -EINVAL;
> > + goto out_no_plug;
> > + }
> > +
> > blk_start_plug(&plug);
> > for (;;) {
> > /* Still start < end. */
>
> What happens if the madvise() revolves multiple VMAs, the first one is not
> VM_SHSTK, but the another one is? Shouldnât the test be done inside the
> loop, potentially in madvise_vma() ?
>

I will fix it. Thanks!