Date:   Thu, 2 Feb 2023 15:04:46 -0500
From:   Peter Xu <peterx@redhat.com>
To:     Yang Shi <shy828301@gmail.com>
Cc:     David Stevens <stevensd@chromium.org>,
        David Hildenbrand <david@redhat.com>,
        "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>,
        linux-mm@kvack.org, Andrew Morton <akpm@linux-foundation.org>,
        linux-kernel@vger.kernel.org, Hugh Dickins <hughd@google.com>
Subject: Re: [PATCH] mm/khugepaged: skip shmem with armed userfaultfd
Message-ID: <Y9wXXoSY+QqoKMMx@x1n>
References: <20230201034137.2463113-1-stevensd@google.com>
 <CAHbLzkpbV2LOoTpWwSOS+UUsYJiZX4vO78jZSr6xmpAGNGoH5w@mail.gmail.com>
 <Y9rRCN9EfqzwYnDG@x1n>
 <CAHbLzkqyhS5thdVKa-jcS5iNUNxe95hagBncWaG=CQTh=LU70w@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <CAHbLzkqyhS5thdVKa-jcS5iNUNxe95hagBncWaG=CQTh=LU70w@mail.gmail.com>
Precedence: bulk

On Wed, Feb 01, 2023 at 03:57:33PM -0800, Yang Shi wrote:
> > There's another problem where the current vma may not have uffd armed,
> > khugepaged may think it has nothing to do with uffd and moved on with
> > collapsing, but actually it's armed in another vma of either the current mm
> > or just another mm's.
> 
> Out of curiosity, could you please elaborate how another vma armed
> with userfaultfd could have an impact on the vmas that are not armed?

It's e.g. when >1 vmas mapped to the same shmem file on the same range, one
registered with uffd missing, others not.  Then others can cause page cache
populated without generating message to the vma that got uffd missing mode
registered, so there'll be the same issue as when khugepaged accidentally
does thp collapsings.  Thanks,

-- 
Peter Xu