Re: [PATCH v3 2/3] hwmon: (isl28022) Fix integer overflow in power calculation on 32-bit

[sashiko-bot@kernel.org]

Sashiko AI review found 1 potential issue(s):
- [Low] The use of the `min()` macro with mixed types (`u64` and `long`) will cause build failures when this patch is backported to older stable kernels.
--

commit 7d032fe46f91e450d244ac30f5851808fee10667
Author: Sanman Pradhan <psanman@juniper.net>

hwmon: (isl28022) Fix integer overflow in power calculation on 32-bit

This commit fixes an integer overflow on 32-bit systems in the power
calculation of the isl28022 driver by switching to 64-bit arithmetic
and using div_u64().

> diff --git a/drivers/hwmon/isl28022.c b/drivers/hwmon/isl28022.c
> index c2e559dde63f6..c5a34ceedcdb2 100644
> --- a/drivers/hwmon/isl28022.c
> +++ b/drivers/hwmon/isl28022.c

[ ... ]

> @@ -185,8 +186,8 @@ static int isl28022_read_power(struct device *dev, u32 attr, long *val)
>  				  ISL28022_REG_POWER, &regval);
>  		if (err < 0)
>  			return err;
> -		*val = ((51200000L * ((long)data->gain)) /
> -			(long)data->shunt) * (long)regval;
> +		*val = min(div_u64(51200000ULL * data->gain * regval,
> +				   data->shunt), LONG_MAX);

Is there a risk of build failures when this patch is backported to older
stable kernels?

Since this patch is marked with "Cc: stable@vger.kernel.org", it will likely
be backported to older LTS kernels like 5.15 or 6.1. While modern kernels
have relaxed the type-checking in the min() macro to allow safely comparable
mixed types, older kernels strictly enforce identical types via pointer
comparison.

Comparing a u64 (from div_u64()) and a long (LONG_MAX) might trigger a
"comparison of distinct pointer types lacks a cast" compiler error during
backporting.

Would it be better to use min_t(u64, div_u64(...), LONG_MAX) to avoid this
issue on older trees?

>  		break;
>  	default:
>  		return -EOPNOTSUPP;
>

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260408163029.353777-1-sanman.pradhan@hpe.com?part=2