On Mon, 2020-05-11 at 15:08 +0300, Luca Coelho wrote:
> On Mon, 2020-05-11 at 14:47 +0300, Kalle Valo wrote:
> > Luca Coelho <luca@xxxxxxxxx> writes:
> >
> > > From: Gil Adam <gil.adam@xxxxxxxxx>
> > >
> > > Evaluate the appropriate DSM from ACPI to enable 5.15,5.35 GHz
> > > bands in Indonesia. If enabled send LARI_CONFIG_CHANGE cmd to fw.
> > >
> > > Signed-off-by: Gil Adam <gil.adam@xxxxxxxxx>
> > > Signed-off-by: Luca Coelho <luciano.coelho@xxxxxxxxx>
> >
> > [...]
> >
> > > +static int iwl_mvm_eval_dsm_indonesia_5g2(struct iwl_mvm *mvm)
> > > +{
> > > + int ret = iwl_acpi_get_dsm_u8((&mvm->fwrt)->dev, 0,
> > > + DSM_FUNC_ENABLE_INDONESIA_5G2);
> > > +
> > > + IWL_DEBUG_RADIO(mvm,
> > > + "Evaluated DSM function ENABLE_INDONESIA_5G2, ret=%d\n",
> > > + ret);
> > > +
> > > + return ret == 1 ? 1 : 0;
> >
> > Ugh. If it's _really_ important not to return anything bigger than 1
> > maybe "!!ret"? Or maybe even just change the function to return boolean?
>
> Ugh, indeed. Guess who was not paying attention? Sorry.
Actually the idea was to turn errors (negative numbers) into 0. But in
any case, a boolean would be better here. This function was changed in
a later patch snf handles more values, though.
--
Luca.
