Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20220712130542.18836-1-linmiaohe@huawei.com> <Ys2xyCUnqpJt0eIo@monkey>
 <Ys7ViSsOxLaQJIfy@xz-m1.local> <Ys7uXHdzzFezUub5@xz-m1.local>
 <CAJHvVcgqb6R6ePMbgmA8LpMpMgrGWie9ZCTuR4MA77GTvi4XAw@mail.gmail.com> <Ys9XAbuxkDUlnmv0@monkey>
In-Reply-To: <Ys9XAbuxkDUlnmv0@monkey>
From:   Axel Rasmussen <axelrasmussen@google.com>
Date:   Wed, 13 Jul 2022 17:20:07 -0700
Message-ID: <CAJHvVci9iij+eDV-EWDOtjmWFYo0H+1LkzKBp6=XOpwDA4Jh-g@mail.gmail.com>
Subject: Re: [PATCH] mm/hugetlb: avoid corrupting page->mapping in hugetlb_mcopy_atomic_pte
To:     Mike Kravetz <mike.kravetz@oracle.com>
Cc:     Peter Xu <peterx@redhat.com>, Miaohe Lin <linmiaohe@huawei.com>,
        Andrew Morton <akpm@linux-foundation.org>,
        Muchun Song <songmuchun@bytedance.com>,
        Linux MM <linux-mm@kvack.org>,
        LKML <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Wed, Jul 13, 2022 at 4:36 PM Mike Kravetz <mike.kravetz@oracle.com> wrote:
>
> On 07/13/22 15:46, Axel Rasmussen wrote:
> > I think there is a small mistake in this patch.
> >
> > Consider the non-minor-fault case. We have this block:
> >
> > /* Add shared, newly allocated pages to the page cache. */
> > if (vm_shared && !is_continue) {
> >         /* ... */
> > }
> >
> > In here, we've added the newly allocated page to the page cache, and
> > we've set this page_in_pagecache flag to true. But we *do not* setup
> > rmap for this page in this block. I think in this case, the patch will
> > cause us to do the wrong thing: we should hugepage_add_new_anon_rmap()
> > further down, but with this patch we dup instead.
>
> I am not sure I follow.  The patch from Miaohe Lin would not change any
> behavior in the 'if (vm_shared && !is_continue)' case.  In this case
> both vm_shared and page_in_pagecache are true.
>
> IIUC, the patch would address the case where !vm_shared && is_continue.

Ah, you're right, my interpretation of the various flags got mixed up
somewhere along the way.

page_in_pagecache is equivalent to vm_shared in this function,
*except* when we have is_continue. Given that, I think this patch is
correct in the vm_shared case (no behavior change). In case of
!vm_shared && is_continue, I agree the patch is a correction to the
previous behavior.

>
> On 07/12/22 21:05, Miaohe Lin wrote:
> > In MCOPY_ATOMIC_CONTINUE case with a non-shared VMA, pages in the page
> > cache are installed in the ptes. But hugepage_add_new_anon_rmap is called
> > for them mistakenly because they're not vm_shared. This will corrupt the
> > page->mapping used by page cache code.
> >
> > Fixes: f619147104c8 ("userfaultfd: add UFFDIO_CONTINUE ioctl")
> > Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
> > ---
> >  mm/hugetlb.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/mm/hugetlb.c b/mm/hugetlb.c
> > index 8d379e03f672..b232e1508e49 100644
> > --- a/mm/hugetlb.c
> > +++ b/mm/hugetlb.c
> > @@ -6038,7 +6038,7 @@ int hugetlb_mcopy_atomic_pte(struct mm_struct *dst_mm,
> >       if (!huge_pte_none_mostly(huge_ptep_get(dst_pte)))
> >               goto out_release_unlock;
> >
> > -     if (vm_shared) {
> > +     if (page_in_pagecache) {
> >               page_dup_file_rmap(page, true);
> >       } else {
> >               ClearHPageRestoreReserve(page);
>
> --
> Mike Kravetz