Re: [PATCH v2 1/2] mm/migrate: Fix read-only page got writable when recover pte

From: Peter Xu <peterx@redhat.com>
To: David Hildenbrand <david@redhat.com>
Cc: linux-kernel@vger.kernel.org, linux-mm@kvack.org,
	Andrea Arcangeli <aarcange@redhat.com>,
	Axel Rasmussen <axelrasmussen@google.com>,
	Ives van Hoorne <ives@codesandbox.io>,
	Nadav Amit <nadav.amit@gmail.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	Mike Rapoport <rppt@linux.vnet.ibm.com>,
	stable@vger.kernel.org
Subject: Re: [PATCH v2 1/2] mm/migrate: Fix read-only page got writable when recover pte
Date: Tue, 15 Nov 2022 13:11:11 -0500	[thread overview]
Message-ID: <Y3PWP9RPjg5FObMN@x1n> (raw)
In-Reply-To: <2ed12722-2359-cb07-53e7-566d959d311e@redhat.com>

On Tue, Nov 15, 2022 at 06:54:09PM +0100, David Hildenbrand wrote:
> No it isn't natural. But sneaking such a change into your fix seems wrong.
> Touching !uffd-wp code should be separate, if we want to do this at all (as
> we discussed, maybe the better/cleaner approach is to eliminate writable
> migration entries if possible).

If you see I made the subject as mm/migrate, because I took it as a
migration change, not uffd-wp specific.  I also added the Fixes to the uffd
patch because afaict that's the only thing that it affects, but I can't
tell either.

As I said, I'm trying to follow the rule where we have read entry I want to
make sure write bit removed.

I hope we have this patch land soon, because it affects users.  I'm not
against further rework on migration entries as I replied in the other
email, but for this patch I keep the same thoughts as when I posted.

Thanks,

-- 
Peter Xu