On Mon, Jul 14, 2025 at 04:34:39PM +0530, Jagadeesh Kona wrote:
>
>
> On 7/9/2025 5:18 PM, Luca Weiss wrote:
> > Hi Jagadeesh,
> >
> > On Tue Jul 8, 2025 at 7:34 AM CEST, Jagadeesh Kona wrote:
> >>
> >>
> >> On 7/4/2025 12:46 PM, Luca Weiss wrote:
> >>> Add support to register the rcg dfs in qcom_cc_really_probe(). This
> >>> allows users to move the call from the probe function to static
> >>> properties.
> >>>
> >>> Signed-off-by: Luca Weiss <[email protected]>
> >>> ---
> >>> drivers/clk/qcom/common.c | 8 ++++++++
> >>> drivers/clk/qcom/common.h | 2 ++
> >>> 2 files changed, 10 insertions(+)
> >>> diff --git a/drivers/clk/qcom/common.h b/drivers/clk/qcom/common.h
> >>> index
> >>> 0f4b2d40c65cf94de694226f63ca30f4181d0ce5..dbe7ebe5b8710fb03c1671ac9022e608a6aad35f
> >>> 100644
> >>> --- a/drivers/clk/qcom/common.h
> >>> +++ b/drivers/clk/qcom/common.h
> >>> @@ -43,6 +43,8 @@ struct qcom_cc_desc {
> >>> size_t num_gdscs;
> >>> struct clk_hw **clk_hws;
> >>> size_t num_clk_hws;
> >>> + const struct clk_rcg_dfs_data *dfs_rcgs;
> >>> + size_t num_dfs_rcgs;
> >>
> >> Since all other probe configurations are added in qcom_cc_driver_data
> >> struct, it is better to
> >> add rcg dfs data also to qcom_cc_driver_data struct instead of
> >> qcom_cc_desc.
> >
> > Dmitry suggested to use struct qcom_cc_desc:
> > https://lore.kernel.org/linux-arm-msm/w4eujq72uqflqpsqshc7zhu6lkc7owufep2g2rjacvzgj44vmf@auonp4ugbgow/
> >
> > I personally don't have a preference.
> >
>
> Hi Luca,
>
> As other probe configurations like critical clock cbcrs & pll configurations
> are part of driver_data, it is better
> to include this DFS data also in driver_data.
>
> Dmitry, I hope it is fine to move the above DFS data to qcom_cc_driver_data
> struct instead of qcom_cc_desc.
Yes, of course.
--
With best wishes
Dmitry
