Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:2 as permitted sender) client-ip=2620:137:e000::3:2;
MIME-Version: 1.0
References: <20231119194740.94101-1-ryncsn@gmail.com> <20231119194740.94101-2-ryncsn@gmail.com>
 <ZVp2UE+MqkifJG4B@casper.infradead.org> <CAF8kJuOGUQNyZesw5BY3dfwBNhL6_RG_=tXc1ZQ4VRGRtGPuMw@mail.gmail.com>
In-Reply-To: <CAF8kJuOGUQNyZesw5BY3dfwBNhL6_RG_=tXc1ZQ4VRGRtGPuMw@mail.gmail.com>
From:   Kairui Song <ryncsn@gmail.com>
Date:   Mon, 20 Nov 2023 19:14:59 +0800
Message-ID: <CAMgjq7CO_9tXfwd2at-tT+O_JU-NGoG=9OWcqwkc9M3iL1Ga1A@mail.gmail.com>
Subject: Re: [PATCH 01/24] mm/swap: fix a potential undefined behavior issue
To:     Chris Li <chrisl@kernel.org>
Cc:     Matthew Wilcox <willy@infradead.org>,
        linux-mm <linux-mm@kvack.org>,
        Andrew Morton <akpm@linux-foundation.org>,
        "Huang, Ying" <ying.huang@intel.com>,
        David Hildenbrand <david@redhat.com>,
        Hugh Dickins <hughd@google.com>,
        Johannes Weiner <hannes@cmpxchg.org>,
        Michal Hocko <mhocko@suse.com>,
        LKML <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

Chris Li <chrisl@kernel.org> =E4=BA=8E2023=E5=B9=B411=E6=9C=8820=E6=97=A5=
=E5=91=A8=E4=B8=80 11:36=E5=86=99=E9=81=93=EF=BC=9A
>
> Hi Kairui,
>
> On Sun, Nov 19, 2023 at 12:55=E2=80=AFPM Matthew Wilcox <willy@infradead.=
org> wrote:
> >
> > On Mon, Nov 20, 2023 at 03:47:17AM +0800, Kairui Song wrote:
> > > From: Kairui Song <kasong@tencent.com>
> > >
> > > When folio is NULL, taking the address of its struct member is an
> > > undefined behavior, the UB is caused by applying -> operator
>
> I think dereferencing the NULL pointer is undefined behavior. There is
> no dereferencing here. It is just pointer arithmetic of NULL pointers,
> which is adding offset of page to the NULL pointer, you got NULL.
>
> > > won't lead to a real issue, still better to fix it, also makes the
> > > code less error-prone, when folio is NULL, page is also NULL,
> > > instead of a meanless offset value.
>
> I consider your reasoning is invalid. NULL pointer arithmetic should
> be legal. This patch is not needed.
>
> Chris

Hi, Chris and Matthew.

Thanks for the comments.

Right, it's just a language syntax level thing, since "->" have a
higher priority, so in the syntax level it is doing a member access
first, then take the address. By C definition  member access should
not happen if the object is invalid (NULL). Only a hypothesis problem
on paper...

This is indeed not needed since in reality it's just pointer
arithmetic. I'm OK dropping this.