Received-SPF: pass (google.com: domain of linux-kernel+bounces-212640-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Precedence: bulk
MIME-Version: 1.0
References: <20240612180554.1328409-1-joychakr@google.com> <f0dbf963-bfd9-4a0b-8284-d141999da184@moroto.mountain>
In-Reply-To: <f0dbf963-bfd9-4a0b-8284-d141999da184@moroto.mountain>
From: Joy Chakraborty <joychakr@google.com>
Date: Thu, 13 Jun 2024 12:07:46 +0530
Message-ID: <CAOSNQF1AiD5rcpJr=c8Dov=j-g4=xOZXViX+Xibu_kBA=2rzgA@mail.gmail.com>
Subject: Re: [PATCH] rtc: abx80x: Fix return value of nvmem callback on read
To: Dan Carpenter <dan.carpenter@linaro.org>
Cc: Sean Anderson <sean.anderson@seco.com>, 
	Alexandre Belloni <alexandre.belloni@bootlin.com>, 
	Srinivas Kandagatla <srinivas.kandagatla@linaro.org>, linux-rtc@vger.kernel.org, 
	linux-kernel@vger.kernel.org, stable@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Thu, Jun 13, 2024 at 11:48=E2=80=AFAM Dan Carpenter <dan.carpenter@linar=
o.org> wrote:
>
> On Wed, Jun 12, 2024 at 06:05:54PM +0000, Joy Chakraborty wrote:
> > Read callbacks registered with nvmem core expect 0 to be returned on
> > success and a negative value to be returned on failure.
> >
> > abx80x_nvmem_xfer() on read calls i2c_smbus_read_i2c_block_data() which
> > returns the number of bytes read on success as per its api description,
> > this return value is handled as an error and returned to nvmem even on
> > success.
> >
> > Fix to handle all possible values that would be returned by
> > i2c_smbus_read_i2c_block_data().
> >
> > Fixes: e90ff8ede777 ("rtc: abx80x: Add nvmem support")
> > Cc: stable@vger.kernel.org
> > Signed-off-by: Joy Chakraborty <joychakr@google.com>
> > ---
> >  drivers/rtc/rtc-abx80x.c | 9 ++++++++-
> >  1 file changed, 8 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/rtc/rtc-abx80x.c b/drivers/rtc/rtc-abx80x.c
> > index fde2b8054c2e..0f5847d1ca2a 100644
> > --- a/drivers/rtc/rtc-abx80x.c
> > +++ b/drivers/rtc/rtc-abx80x.c
> > @@ -711,9 +711,16 @@ static int abx80x_nvmem_xfer(struct abx80x_priv *p=
riv, unsigned int offset,
> >               else
> >                       ret =3D i2c_smbus_read_i2c_block_data(priv->clien=
t, reg,
> >                                                           len, val);
> > -             if (ret)
> > +             if (ret < 0)
> >                       return ret;
> >
> > +             if (!write) {
> > +                     if (ret)
> > +                             len =3D ret;
> > +                     else
> > +                             return -EIO;
> > +             }
>
> I guess this is the conservative approach.  Ie.  Don't break things
> which aren't already broken.  But I suspect the correct approach is to
> say:
>
>         if (ret !=3D len)
>                 return -EIO;
>
> Ah well.  Being conservative is good.  It probably doesn't ever happen
> in real life so it probably doesn't matter either way.
>
> I don't really like the if (write) follow by and if (!write)...  It
> would add more lines, but improve readability if we just duplicate the
> code a big:
>
>         if (write) {
>                 ret =3D write();
>                 if (ret)
>                         return ret;
>         } else {
>                 ret =3D read();
>                 if (ret <=3D 0)
>                         return ret ?: -EIO;
>                 len =3D ret;
>         }
>

Sure, I'll do this in a follow up patch.

Thanks
Joy
> regards,
> dan carpenter
>