Received-SPF: pass (google.com: domain of linux-bluetooth+bounces-5089-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Precedence: bulk
MIME-Version: 1.0
References: <20240530150057.444585-1-hadess@hadess.net> <20240530150057.444585-10-hadess@hadess.net>
 <8c979897beeb8316616bb7c15f60093be177577c.camel@iki.fi> <727c93065d8d1e56e9fcd5b5bed724d63dfbe871.camel@hadess.net>
In-Reply-To: <727c93065d8d1e56e9fcd5b5bed724d63dfbe871.camel@hadess.net>
From: Luiz Augusto von Dentz <luiz.dentz@gmail.com>
Date: Mon, 3 Jun 2024 15:19:25 -0400
Message-ID: <CABBYNZJB5kYZd5=jEz6sC5TOe9EMphMDBkKgHGjpTf_AH8fXdQ@mail.gmail.com>
Subject: Re: [BlueZ 9/9] avdtp: Fix manipulating struct as an array
To: Bastien Nocera <hadess@hadess.net>
Cc: Pauli Virtanen <pav@iki.fi>, linux-bluetooth@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Hi Bastien,

On Fri, May 31, 2024 at 11:51=E2=80=AFAM Bastien Nocera <hadess@hadess.net>=
 wrote:
>
> On Thu, 2024-05-30 at 19:57 +0300, Pauli Virtanen wrote:
> > Hi,
> >
> > to, 2024-05-30 kello 16:58 +0200, Bastien Nocera kirjoitti:
> > > Don't manipulate the "req" structs as if they were flat arrays,
> > > static
> > > analysis and humans are both equally confused by this kind of
> > > usage.
> >
> > struct start_req {
> >       union {
> >               struct seid required[1];
> >               struct seid seids[0];
> >       };
> > } __attribute__ ((packed));
> >
> > and access only via req->seids?
>
> That's a good idea, I'll give it a try.
>
> > <snip>
> > > +#define
> > > REQ_GET_NTH_SEID(x)                                         \
> > > +   static struct seid
> > > *                                           \
> > > +   x##_req_get_nth_seid(struct x##_req *req, int count, int
> > > i)  \
> > > +   {
> > >     \
> > > +           if (count =3D=3D 0 || i >=3D
> > > count)                              \
> > > +                   return
> > > NULL;                                       \
> > > +           if (i =3D=3D
> > > 1)                                          \
> > > +                   return &req-
> > > >first_seid;                        \
> > > +           return &req-
> > > >other_seids[i];                            \
> >
> > (i =3D=3D 0) and [i - 1]?
>
> *facepalm*
>
> Yes, this will need a v2, thanks for spotting that.

Ive applied the set except for this one, please resend once you are
done with the suggested changes.

--=20
Luiz Augusto von Dentz