Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Content-Type: text/plain;
        charset=utf-8
Mime-Version: 1.0 (Mac OS X Mail 13.0 \(3608.60.0.2.5\))
Subject: Re: [PATCH -next] fork: annotate a data race in vm_area_dup()
From:   Qian Cai <cai@lca.pw>
In-Reply-To: <20200217223138.doaph66iwprbwhw5@box>
Date:   Mon, 17 Feb 2020 22:59:47 -0500
Cc:     Andrew Morton <akpm@linux-foundation.org>, elver@google.com,
        linux-mm@kvack.org, linux-kernel@vger.kernel.org
Content-Transfer-Encoding: quoted-printable
Message-Id: <EAD6E54D-8A57-4494-94F2-2EEEC3265560@lca.pw>
References: <20200217223138.doaph66iwprbwhw5@box>
To:     "Kirill A. Shutemov" <kirill@shutemov.name>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


> On Feb 17, 2020, at 5:31 PM, Kirill A. Shutemov <kirill@shutemov.name> =
wrote:
>=20
> I'm confused. AFAICS both sides hold mmap_sem on write:
>=20
> - vm_mmap_pgoff() takes mmap_sem for the write on the write side
>=20
> - do_mprotect_pkey() takes mmap_sem for the write on the read side
>=20
>=20
> What do I miss?

Ah, good catch. I missed the locking for the read there. This is =
interesting because Marco
did confirmed that the concurrency could happen,

https://lore.kernel.org/lkml/20191025173511.181416-1-elver@google.com/

If that means KCSAN is not at fault, then I could think of two things,

1) someone downgrades the lock.

I don=E2=80=99t think that a case here. Only __do_munmap() will do that =
but I did not see how
it will affect us here.

2) the reader and writer are two different processes.

So, they held a different mmap_sem, but I can=E2=80=99t see how could =
two processes shared
the same vm_area_struct. Also, file->f_mapping->i_mmap was also stored =
in the
writer, but I can=E2=80=99t see how it was also loaded in the reader.

Any ideas?=