Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <E9CF24C7-3080-4720-B540-BAF03068336B@gmail.com>
 <1E0741E0-2BD9-4FA3-BA41-4E83315A10A8@joelfernandes.org> <1AF98387-B78C-4556-BE2E-E8F88ADACF8A@gmail.com>
 <cc9b292c-99b1-bec9-ba8e-9c202b5835cd@joelfernandes.org> <962bb2b940e64e7da7b71d11b307defc@AcuMS.aculab.com>
In-Reply-To: <962bb2b940e64e7da7b71d11b307defc@AcuMS.aculab.com>
From:   Joel Fernandes <joel@joelfernandes.org>
Date:   Fri, 21 Jul 2023 13:14:40 -0400
Message-ID: <CAEXW_YS_raHUrvVAFPpnhL2PRH0hkcqT=1hD+gQOg_cMLkGrjQ@mail.gmail.com>
Subject: Re: Question about the barrier() in hlist_nulls_for_each_entry_rcu()
To:     David Laight <David.Laight@aculab.com>
Cc:     Alan Huang <mmpgouride@gmail.com>,
        Eric Dumazet <edumazet@google.com>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "netdev@vger.kernel.org" <netdev@vger.kernel.org>,
        "rcu@vger.kernel.org" <rcu@vger.kernel.org>,
        "Paul E. McKenney" <paulmck@kernel.org>,
        "roman.gushchin@linux.dev" <roman.gushchin@linux.dev>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Fri, Jul 21, 2023 at 11:59=E2=80=AFAM David Laight <David.Laight@aculab.=
com> wrote:
>
> ....
> > Right, it shouldn't need to cache. To Eric's point it might be risky to=
 remove
> > the barrier() and someone needs to explain that issue first (or IMO the=
re needs
> > to be another tangible reason like performance etc). Anyway, FWIW I wro=
te a
> > simple program and I am not seeing the head->first cached with the patt=
ern you
> > shared above:
> >
> > #include <stdlib.h>
> >
> > #define READ_ONCE(x) (*(volatile typeof(x) *)&(x))
> > #define barrier() __asm__ __volatile__("": : :"memory")
> >
> > typedef struct list_head {
> >      int first;
> >      struct list_head *next;
> > } list_head;
> >
> > int main() {
> >      list_head *head =3D (list_head *)malloc(sizeof(list_head));
> >      head->first =3D 1;
> >      head->next =3D 0;
> >
> >      READ_ONCE(head->first);
> >      barrier();
> >      READ_ONCE(head->first);
> >
> >      free(head);
> >      return 0;
> > }
>
> You probably need to try harder to generate the error.
> It probably has something to do code surrounding the
> sk_nulls_for_each_rcu() in the ca065d0c^ version of udp.c.
>
> That patch removes the retry loop - and probably breaks udp receive.
> The issue is that sockets can be moved between the 'hash2' chains
> (eg by connect()) without being freed.

I was just replying to Alan's question on the behavior of READ_ONCE()
since I myself recently got surprised by compiler optimizations
related to it. I haven't looked into the actual UDP code.

 - Joel