linux-kernel - Re: [PATCH 1/6] iio: imu: adis16480: Use the default data ready pin configuration

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 20 Feb 2019 10:20:23 +0000
From:   Jonathan Cameron <jic23@...nel.org>
To:     Stefan Popa <stefan.popa@...log.com>
Cc:     <robh+dt@...nel.org>, <Michael.Hennerich@...log.com>,
        <knaack.h@....de>, <lars@...afoo.de>, <pmeerw@...erw.net>,
        <gregkh@...uxfoundation.org>, <linux-kernel@...r.kernel.org>,
        <linux-iio@...r.kernel.org>
Subject: Re: [PATCH 1/6] iio: imu: adis16480: Use the default data ready pin
 configuration

On Tue, 19 Feb 2019 19:12:13 +0200
Stefan Popa <stefan.popa@...log.com> wrote:

> The FNCTIO_CTRL register, Bits[3:0] provide three configuration options
> for the data ready function: on/off, polarity, and DIOx line. The
> factory default assigns DIO2 as a positive polarity, data ready signal.
> 
> The adis16480_enable_irq() function, overwrites this configuration when
> it enables/disables the data ready pin by only setting BIT[3].
> As a result, the data ready signal becomes DIO1 pin which is assigned as
> negative polarity.
> 
> This patch reads the FNCTIO_CTRL register and creates a mask, such that
> only data ready enable (BIT[3]) will be modified when
> adis16480_enable_irq function is called.

So this is potentially a problem.  As I read this, we just changed the default.
So a device that has been relying on this 'bug' for a long time will now
not work as it will be expecting the interrupt on the wrong physical pin.

So, whilst it might seem logical to let the device stay with it's default
we can't do it because the defacto Linux default is the other choice.

The delights of having to support old 'bugs' :)

Jonathan
> 
> Signed-off-by: Stefan Popa <stefan.popa@...log.com>
> ---
>  drivers/iio/imu/adis16480.c | 18 ++++++++++++++++--
>  1 file changed, 16 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/iio/imu/adis16480.c b/drivers/iio/imu/adis16480.c
> index a27fe20..d222188 100644
> --- a/drivers/iio/imu/adis16480.c
> +++ b/drivers/iio/imu/adis16480.c
> @@ -9,6 +9,7 @@
>   *
>   */
>  
> +#include <linux/bitfield.h>
>  #include <linux/interrupt.h>
>  #include <linux/delay.h>
>  #include <linux/mutex.h>
> @@ -107,6 +108,10 @@
>  #define ADIS16480_FIR_COEF_C(x)			ADIS16480_FIR_COEF(0x09, (x))
>  #define ADIS16480_FIR_COEF_D(x)			ADIS16480_FIR_COEF(0x0B, (x))
>  
> +/* ADIS16480_REG_FNCTIO_CTRL */
> +#define ADIS16480_DRDY_EN_MSK		BIT(3)
> +#define ADIS16480_DRDY_EN(x)		FIELD_PREP(ADIS16480_DRDY_EN_MSK, x)
> +
>  struct adis16480_chip_info {
>  	unsigned int num_channels;
>  	const struct iio_chan_spec *channels;
> @@ -741,8 +746,17 @@ static int adis16480_stop_device(struct iio_dev *indio_dev)
>  
>  static int adis16480_enable_irq(struct adis *adis, bool enable)
>  {
> -	return adis_write_reg_16(adis, ADIS16480_REG_FNCTIO_CTRL,
> -		enable ? BIT(3) : 0);
> +	uint16_t val;
> +	int ret;
> +
> +	ret = adis_read_reg_16(adis, ADIS16480_REG_FNCTIO_CTRL, &val);
> +	if (ret < 0)
> +		return ret;
> +
> +	val &= ~ADIS16480_DRDY_EN_MSK;
> +	val |= ADIS16480_DRDY_EN(enable);
> +
> +	return adis_write_reg_16(adis, ADIS16480_REG_FNCTIO_CTRL, val);
>  }
>  
>  static int adis16480_initial_setup(struct iio_dev *indio_dev)