Re: [PATCH] mm/hugetlb: avoid corrupting page->mapping in hugetlb_mcopy_atomic_pte

[Peter Xu <peterx@redhat.com>]

On Fri, Jul 15, 2022 at 11:56:40AM +0800, Miaohe Lin wrote:
> On 2022/7/14 23:52, Peter Xu wrote:
> > On Thu, Jul 14, 2022 at 05:59:53PM +0800, Miaohe Lin wrote:
> >> On 2022/7/14 1:23, Andrew Morton wrote:
> >>> On Tue, 12 Jul 2022 21:05:42 +0800 Miaohe Lin <linmiaohe@huawei.com> 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.
> >>>
> >>> Well that sounds bad.  And theories on why this has gone unnoticed for
> >>> over a year?  I assume this doesn't have coverage in our selftests?
> >>
> >> As discussed in another thread, when minor fault handling is proposed, only
> >> VM_SHARED vma is expected to be supported. And the test case is also missing.
> > 
> > Yes, after this patch applied it'll be great to have the test case covering
> > private mappings too.
> > 
> > It's just that it'll be a bit more than setting test_uffdio_minor=1 for
> > "hugetlb" test.  In hugetlb_allocate_area() we'll need to setup the alias
> > too for !shared case, it'll be a bit challenging since currently we're
> > using anonymous hugetlb mappings for private tests, and I'm not sure
> > whether we'll need the hugetlb path back just like what we have with
> > "hugetlb_shared" tests.
> 
> I'm afraid not. When minor fault handling is proposed, only VM_SHARED vma is
> expected to be supported. It seems it's hard to image how one might benefit
> from using it with a private mapping. But I'm not sure as I'm still a layman
> in userfaultfd now. Any further suggestions?

IIUC so far we all think it's not required to limit it to shared mappings
only?  The effort is mostly the same.

My suggestion is above - we could enable the kselftest for it, but I don't
strongly ask for that too because I don't know any real use of it, it'll
still be good to have it though for completeness.  It's just that we may
need to change some code back in 9ae8f2b849f79 on using fd-based memory, or
I don't know how to create the alias mapping properly.

Thanks,

-- 
Peter Xu