linux-kernel - Re: [PATCH] iio: fix scale application in iio_convert_raw_to_processed

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] [day] [month] [year] [list]

Message-ID: <20240803163058.369cb426@jic23-huawei>
Date: Sat, 3 Aug 2024 16:30:58 +0100
From: Jonathan Cameron <jic23@...nel.org>
To: Matteo Martelli <matteomartelli3@...il.com>
Cc: Lars-Peter Clausen <lars@...afoo.de>, linux-iio@...r.kernel.org,
 linux-kernel@...r.kernel.org
Subject: Re: [PATCH] iio: fix scale application in
 iio_convert_raw_to_processed_unlocked

On Tue, 30 Jul 2024 10:11:53 +0200
Matteo Martelli <matteomartelli3@...il.com> wrote:

> When the scale_type is IIO_VAL_INT_PLUS_MICRO or IIO_VAL_INT_PLUS_NANO
> the scale passed as argument is only applied to the fractional part of
> the value. Fix it by also multiplying the integer part by the scale
> provided.
> 
> Fixes: 48e44ce0f881 ("iio:inkern: Add function to read the processed value")
> Signed-off-by: Matteo Martelli <matteomartelli3@...il.com>
Ouch.  I only vaguely recall the history of this, but I suspect that
extra 'scale' was an after thought and we messed it up.

Anyhow, applied to the fixes-togreg branch of iio.git and marked for
stable.

Jonathan

> ---
> I found out that after iio_hwmon switched to
> iio_read_channel_processed_scale() to convert the power channel value
> from milli-Watts to micro-Watts [1], lm-sensors started to show
> unexpected values for the power channel of the pac1921 iio driver I was
> testing. It looks the issue relies in the
> iio_convert_raw_to_processed_unlocked() function that only applies the
> given scale to the fractional part if the channel scale type is
> IIO_VAL_INT_PLUS_MICRO or IIO_VAL_INT_PLUS_NANO.
> 
> For example with a raw power value of 71, a power scale type of
> IIO_VAL_INT_PLUS_NANO and power scale value of 9.775200000 (mW) the
> processed power value would be ~694 mW but when scaled to uW by the
> iio_hwmon calling iio_read_channel_processed_scale() with a scale of
> 1000, the processed power would wrongly result to ~55678 uW (~55mW) =
> 71*9 + 71*775200000*1000/1000000000. This because the scale of 1000 is
> only applied to the fractional part of the power value. Instead it
> should be 71*9*1000 + 71*775200000*1000/1000000000 = 694039 uW.
> 
> [1]: https://lore.kernel.org/linux-hwmon/20240620212005.821805-1-sean.anderson@linux.dev/
> ---
>  drivers/iio/inkern.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/iio/inkern.c b/drivers/iio/inkern.c
> index 9f484c94bc6e..151099be2863 100644
> --- a/drivers/iio/inkern.c
> +++ b/drivers/iio/inkern.c
> @@ -647,17 +647,17 @@ static int iio_convert_raw_to_processed_unlocked(struct iio_channel *chan,
>  		break;
>  	case IIO_VAL_INT_PLUS_MICRO:
>  		if (scale_val2 < 0)
> -			*processed = -raw64 * scale_val;
> +			*processed = -raw64 * scale_val * scale;
>  		else
> -			*processed = raw64 * scale_val;
> +			*processed = raw64 * scale_val * scale;
>  		*processed += div_s64(raw64 * (s64)scale_val2 * scale,
>  				      1000000LL);
>  		break;
>  	case IIO_VAL_INT_PLUS_NANO:
>  		if (scale_val2 < 0)
> -			*processed = -raw64 * scale_val;
> +			*processed = -raw64 * scale_val * scale;
>  		else
> -			*processed = raw64 * scale_val;
> +			*processed = raw64 * scale_val * scale;
>  		*processed += div_s64(raw64 * (s64)scale_val2 * scale,
>  				      1000000000LL);
>  		break;
> 
> ---
> base-commit: 1ebab783647a9e3bf357002d5c4ff060c8474a0a
> change-id: 20240729-iio-fix-scale-287b7630374e
> 
> Best regards,