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;
MIME-Version: 1.0
References: <20190929164848.13930-1-yzhai003@ucr.edu>
In-Reply-To: <20190929164848.13930-1-yzhai003@ucr.edu>
From:   Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Date:   Mon, 30 Sep 2019 23:16:15 +0200
Message-ID: <CAFBinCCewgHnvL+HRXR8MBOT4fdnC-nxDAw8E5rGf+NZDsbaJg@mail.gmail.com>
Subject: Re: [PATCH] iio: adc: meson-saradc: Variables could be uninitalized
 if regmap_read() fails
To:     Yizhuo <yzhai003@ucr.edu>
Cc:     csong@cs.ucr.edu, zhiyunq@cs.ucr.edu,
        Jonathan Cameron <jic23@kernel.org>,
        Hartmut Knaack <knaack.h@gmx.de>,
        Lars-Peter Clausen <lars@metafoo.de>,
        Peter Meerwald-Stadler <pmeerw@pmeerw.net>,
        Kevin Hilman <khilman@baylibre.com>,
        Neil Armstrong <narmstrong@baylibre.com>,
        Thomas Gleixner <tglx@linutronix.de>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Nicholas Mc Guire <hofrat@osadl.org>,
        linux-iio@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
        linux-amlogic@lists.infradead.org, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Hi Yizhuo,

thank you for this patch

On Sun, Sep 29, 2019 at 6:48 PM Yizhuo <yzhai003@ucr.edu> wrote:
>
> Several functions in this file are trying to use regmap_read() to
> initialize the specific variable, however, if regmap_read() fails,
> the variable could be uninitialized but used directly, which is
> potentially unsafe. The return value of regmap_read() should be
> checked and handled.
the meson_saradc driver is using a MMIO regmap so read failures are
unlikely (unless there's a bug in the code somewhere)
did you find these issues with some static code analysis tool or did
you experience a real world problem?

> Signed-off-by: Yizhuo <yzhai003@ucr.edu>
> ---
>  drivers/iio/adc/meson_saradc.c | 30 ++++++++++++++++++++++++------
>  1 file changed, 24 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/iio/adc/meson_saradc.c b/drivers/iio/adc/meson_saradc.c
> index 7b28d045d271..4b6c2983ef39 100644
> --- a/drivers/iio/adc/meson_saradc.c
> +++ b/drivers/iio/adc/meson_saradc.c
> @@ -323,6 +323,7 @@ static int meson_sar_adc_wait_busy_clear(struct iio_dev *indio_dev)
>  {
>         struct meson_sar_adc_priv *priv = iio_priv(indio_dev);
>         int regval, timeout = 10000;
> +       int ret;
why not add "ret" to the variables in the line above?
so this could be shortened to (which is also consistent with the
coding style in meson_saradc):
  int ret, regval, timeout = 10000;

> @@ -506,7 +514,10 @@ static int meson_sar_adc_lock(struct iio_dev *indio_dev)
>                  */
>                 do {
>                         udelay(1);
> -                       regmap_read(priv->regmap, MESON_SAR_ADC_DELAY, &val);
> +                       ret = regmap_read(priv->regmap,
> +                                               MESON_SAR_ADC_DELAY, &val);
> +                       if (ret)
> +                               return ret;
this is a big problem because we're returning with &indio_dev->mlock held
see the "timeout < 0" block below

> @@ -771,7 +782,7 @@ static int meson_sar_adc_init(struct iio_dev *indio_dev)
>  {
>         struct meson_sar_adc_priv *priv = iio_priv(indio_dev);
>         int regval, i, ret;
> -
> +       int ret;
this removes an empty line between the variable declarations and the
first code (comment)
also "ret" is already defined in the line before
(I haven't compile-tested this myself yet but I'm surprised that this
doesn't give an error or at least a warning)

> @@ -1013,8 +1027,12 @@ static irqreturn_t meson_sar_adc_irq(int irq, void *data)
>         struct meson_sar_adc_priv *priv = iio_priv(indio_dev);
>         unsigned int cnt, threshold;
>         u32 regval;
> +       int ret;
> +
> +       ret = regmap_read(priv->regmap, MESON_SAR_ADC_REG0, &regval);
> +       if (ret)
> +               return ret;
this function doesn't return "int" but irqreturn_t
the only allowed values are: [0]


Martin


[0] https://elixir.bootlin.com/linux/v5.3/source/include/linux/irqreturn.h#L11