On Tue, 14 Feb 2023 06:50:44 -0800, Rodrigo Vivi wrote:
>
> > +static int
> > +hwm_power_max_write(struct hwm_drvdata *ddat, long val)
> +{
> > + struct i915_hwmon *hwmon = ddat->hwmon;
> > + u32 nval;
> > +
> > + /* Computation in 64-bits to avoid overflow. Round to nearest. */
> > + nval = DIV_ROUND_CLOSEST_ULL((u64)val << hwmon->scl_shift_power,
> > SF_POWER);
> > +
> > + hwm_locked_with_pm_intel_uncore_rmw(ddat, hwmon->rg.pkg_rapl_limit,
> > + PKG_PWR_LIM_1,
> > + REG_FIELD_PREP(PKG_PWR_LIM_1,
> > nval));
> > + return 0;
>
> Let's keep this function as void and the return 0 in the previous spot.
Hmm, see your point. Though there is an identical situation for
hwm_power_max_read read too (in hwm_power_read). Maybe I'll change it there
too in the same patch to keep things symmetrical and retain your R-b?
> With that change:
>
> Reviewed-by: Rodrigo Vivi <[email protected]>
Thanks.
