Re: [PATCH v5] mm/hwpoison: fix race between hugetlb free/demotion and memory_failure_hugetlb()

["HORIGUCHI NAOYA(堀口 直也)" <naoya.horiguchi@nec.com>]

On Mon, Mar 21, 2022 at 05:46:48PM -0700, Yang Shi wrote:
> On Thu, Mar 17, 2022 at 10:16 PM Naoya Horiguchi
> <naoya.horiguchi@linux.dev> wrote:
> >
> > From: Naoya Horiguchi <naoya.horiguchi@nec.com>
> >
> > There is a race condition between memory_failure_hugetlb() and hugetlb
> > free/demotion, which causes setting PageHWPoison flag on the wrong page.
> > The one simple result is that wrong processes can be killed, but another
> > (more serious) one is that the actual error is left unhandled, so no one
> > prevents later access to it, and that might lead to more serious results
> > like consuming corrupted data.
> >
> > Think about the below race window:
> >
> >   CPU 1                                   CPU 2
> >   memory_failure_hugetlb
> >   struct page *head = compound_head(p);
> >                                           hugetlb page might be freed to
> >                                           buddy, or even changed to another
> >                                           compound page.
> >
> >   get_hwpoison_page -- page is not what we want now...
> >
> > The compound_head is called outside hugetlb_lock, so the head is not
> > reliable.
> >
> > So set PageHWPoison flag after passing prechecks. And to detect
> > potential violation, this patch also introduces a new action type
> > MF_MSG_DIFFERENT_PAGE_SIZE.
> >
> > Reported-by: Mike Kravetz <mike.kravetz@oracle.com>
> > Signed-off-by: Naoya Horiguchi <naoya.horiguchi@nec.com>
> > Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
> > Cc: <stable@vger.kernel.org>
...
> > @@ -1547,21 +1545,31 @@ static int memory_failure_hugetlb(unsigned long pfn, int flags)
> >          * If this happens just bail out.
> >          */
> >         if (!PageHuge(p) || compound_head(p) != head) {
> > +               if (TestSetPageHWPoison(p))
> > +                       already_hwpoisoned = pfn;
> > +               else
> > +                       num_poisoned_pages_inc();
> >                 action_result(pfn, MF_MSG_DIFFERENT_PAGE_SIZE, MF_IGNORED);
> 
> The commit log says "this patch also introduces a new action type
> MF_MSG_DIFFERENT_PAGE_SIZE", but it is not defined in the patch and it
> is called here. Did I miss something?

Sorry, you're right.  MF_MSG_DIFFERENT_PAGE_SIZE is defined in the separate
patch in mmotm, and disappeared when rebasing (not intended).
I think of rebasing this to mainline again to apply cleanly to -stable,
expecting it to applied before other recent hwpoison patches.

Thanks,
Naoya Horiguchi