[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <43682e2c485a48deb8f40c9ee060d45c@analog.com>
Date: Thu, 16 Dec 2021 10:54:05 +0000
From: "Tanislav, Cosmin" <Cosmin.Tanislav@...log.com>
To: Kees Cook <keescook@...omium.org>,
Lars-Peter Clausen <lars@...afoo.de>
CC: "Hennerich, Michael" <Michael.Hennerich@...log.com>,
Jonathan Cameron <jic23@...nel.org>,
Linus Walleij <linus.walleij@...aro.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"linux-iio@...r.kernel.org" <linux-iio@...r.kernel.org>,
"linux-hardening@...r.kernel.org" <linux-hardening@...r.kernel.org>
Subject: RE: [PATCH] iio: addac: Do not reference negative array offsets
> -----Original Message-----
> From: Kees Cook <keescook@...omium.org>
> Sent: Thursday, December 16, 2021 1:23 AM
> To: Lars-Peter Clausen <lars@...afoo.de>
> Cc: Kees Cook <keescook@...omium.org>; Hennerich, Michael
> <Michael.Hennerich@...log.com>; Tanislav, Cosmin
> <Cosmin.Tanislav@...log.com>; Jonathan Cameron <jic23@...nel.org>;
> Linus Walleij <linus.walleij@...aro.org>; linux-kernel@...r.kernel.org; linux-
> iio@...r.kernel.org; linux-hardening@...r.kernel.org
> Subject: [PATCH] iio: addac: Do not reference negative array offsets
>
> [External]
>
> Instead of aiming rx_buf at an invalid array-boundary-crossing location,
> just skip the first assignment. Seen when building with -Warray-bounds:
>
> drivers/iio/addac/ad74413r.c: In function 'ad74413r_update_scan_mode':
> drivers/iio/addac/ad74413r.c:843:22: warning: array subscript -4 is below
> array bounds of 'u8[16]' { aka 'unsigned char[16]'} [-Warray-bounds]
> 843 | u8 *rx_buf = &st->adc_samples_buf.rx_buf[-1 *
> AD74413R_FRAME_SIZE];
> |
> ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> drivers/iio/addac/ad74413r.c:84:20: note: while referencing 'rx_buf'
> 84 | u8 rx_buf[AD74413R_FRAME_SIZE *
> AD74413R_CHANNEL_MAX];
> | ^~~~~~
>
> Fixes: fea251b6a5db ("iio: addac: add AD74413R driver")
> Signed-off-by: Kees Cook <keescook@...omium.org>
> ---
> drivers/iio/addac/ad74413r.c | 10 ++++++----
> 1 file changed, 6 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/iio/addac/ad74413r.c b/drivers/iio/addac/ad74413r.c
> index cbd9aa9b399a..b0a6d8ee5133 100644
> --- a/drivers/iio/addac/ad74413r.c
> +++ b/drivers/iio/addac/ad74413r.c
> @@ -840,7 +840,7 @@ static int ad74413r_update_scan_mode(struct iio_dev
> *indio_dev,
> {
> struct ad74413r_state *st = iio_priv(indio_dev);
> struct spi_transfer *xfer = st->adc_samples_xfer;
> - u8 *rx_buf = &st->adc_samples_buf.rx_buf[-1 *
> AD74413R_FRAME_SIZE];
> + u8 *rx_buf = st->adc_samples_buf.rx_buf;
> u8 *tx_buf = st->adc_samples_tx_buf;
> unsigned int channel;
> int ret;
> @@ -877,9 +877,8 @@ static int ad74413r_update_scan_mode(struct iio_dev
> *indio_dev,
> if (ret)
> goto out;
>
> - st->adc_active_channels++;
>
> - if (xfer == st->adc_samples_xfer)
> + if (xfer == st->adc_samples_xfer || st->adc_active_channels
> == 0)
You can probably keep only one of the checks. Both xfer and adc_active_channels
will be incremented after your changes anyway.
> xfer->rx_buf = NULL;
> else
> xfer->rx_buf = rx_buf;
> @@ -896,7 +895,10 @@ static int ad74413r_update_scan_mode(struct
> iio_dev *indio_dev,
>
> xfer++;
> tx_buf += AD74413R_FRAME_SIZE;
> - rx_buf += AD74413R_FRAME_SIZE;
> + if (st->adc_active_channels)
> + rx_buf += AD74413R_FRAME_SIZE;
> +
> + st->adc_active_channels++;
> }
>
> xfer->rx_buf = rx_buf;
> --
> 2.30.2
Powered by blists - more mailing lists