Re: [PATCH 2/4] iio: magnetometer: ak8974: add_device_randomness(fw + serial)

[Jonathan Cameron <jic23@kernel.org>]

On Thu, 17 Aug 2017 15:56:11 +0200
Michał Mirosław <mirq-linux@rere.qmqm.pl> wrote:

> Mix device-specific data into randomness pool.
> 
> Signed-off-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>

Hmm. I wondered about the firmware version as a source of randomness, but
will leave it in here.

Applied to the togreg branch of iio.git and pushed out as testing.

Thanks,

Jonathan

> ---
>  drivers/iio/magnetometer/ak8974.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/drivers/iio/magnetometer/ak8974.c b/drivers/iio/magnetometer/ak8974.c
> index 76091da20a0c..ab204f8560e3 100644
> --- a/drivers/iio/magnetometer/ak8974.c
> +++ b/drivers/iio/magnetometer/ak8974.c
> @@ -20,6 +20,7 @@
>  #include <linux/mutex.h>
>  #include <linux/delay.h>
>  #include <linux/bitops.h>
> +#include <linux/random.h>
>  #include <linux/regmap.h>
>  #include <linux/regulator/consumer.h>
>  #include <linux/pm_runtime.h>
> @@ -466,10 +467,12 @@ static int ak8974_detect(struct ak8974 *ak8974)
>  		ret = regmap_read(ak8974->map, AMI305_VER, &fw);
>  		if (ret)
>  			return ret;
> +		add_device_randomness(&fw, sizeof(fw));

The firmware version seems rather to guessable to really count as randomness.
presumably it would constant for a particular batch of chips.


>  		fw &= 0x7f; /* only bits 0 thru 6 valid */
>  		ret = ak8974_get_u16_val(ak8974, AMI305_SN, &sn);
>  		if (ret)
>  			return ret;
> +		add_device_randomness(&sn, sizeof(sn));

This is better.  The serial number should be hard to guess on a particular
part.

Jonathan
>  		dev_info(&ak8974->i2c->dev,
>  			 "detected %s, FW ver %02x, S/N: %04x\n",
>  			 name, fw, sn);