Messages in this thread | | | Date | Mon, 3 Jul 2023 17:31:38 +0300 | Subject | Re: [PATCH 4/5] soc: qcom: geni-se: Allow any combination of icc paths | From | Dmitry Baryshkov <> |
| |
On 03/07/2023 16:31, Konrad Dybcio wrote: > Not all SoCs provide all the usual paths. By the looks of it, at least > SM8150 and SM8250 don't have one that would resemble "qup-core". > > Check for the error that icc_get throws and assign a NULL value to each > path that can't be found to effectively allow any combination of icc paths > (which, like previously, includes no icc paths). The ICC APIs gracefully > handle a NULL path by exiting early. > > Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org> > --- > drivers/soc/qcom/qcom-geni-se.c | 9 +++++++-- > 1 file changed, 7 insertions(+), 2 deletions(-) > > diff --git a/drivers/soc/qcom/qcom-geni-se.c b/drivers/soc/qcom/qcom-geni-se.c > index ba788762835f..a5e2e8925c8e 100644 > --- a/drivers/soc/qcom/qcom-geni-se.c > +++ b/drivers/soc/qcom/qcom-geni-se.c > @@ -813,8 +813,13 @@ int geni_icc_get(struct geni_se *se, const char *icc_ddr) > continue; > > se->icc_paths[i].path = devm_of_icc_get(se->dev, icc_names[i]);
Would it make sense to add (devm_)of_icc_get_optional instead? I think we already have several usecases for such API call
For this patch:
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
> - if (IS_ERR(se->icc_paths[i].path)) > - goto err; > + if (IS_ERR(se->icc_paths[i].path)) { > + /* Not all SoCs implement all the paths */ > + if (PTR_ERR(se->icc_paths[i].path) == -ENODATA) > + se->icc_paths[i].path = NULL; > + else > + goto err; > + } > } > > return 0; >
-- With best wishes Dmitry
| |