Re: [PATCH 4/7] phy: qcom-mipi-csi2: Add sa8775p D-PHY support

[Bryan O'Donoghue <bod@kernel.org>]

On 17/10/2026 00:54, Anusha Arun Nandi wrote:
> The timer-clock lookup in phy_qcom_mipi_csi2_probe() and the OPP
> clkname passed to devm_pm_opp_set_clkname() were both hardcoded
> string literals sized to x1e80100 ("timer"/"core"). Add
> timer_clk_name/opp_clk_name to struct mipi_csi2phy_soc_cfg so each
> SoC can declare its own clock-name strings; x1e80100's soc_cfg now
> declares "timer"/"core" explicitly, with no behaviour change.
> SA8775P reuses the same "timer"/"core" names as x1e80100 for this
> pair of clocks.

LoL I had that in the last design and review said "hard code it" happy 
to agree with myself in the past, however this change should be 
standalone and precede the Lemans work.

> SA8775P's CSIPHY has a single power-domain, the Titan Top GDSC,
> matching the legacy in-tree CAMSS driver's sa8775p_resources.pd_name
> = "top" and the existing lemans.dtsi camss node. Like x1e80100's own
> "top" domain, it is not OPP-scaled, only kept enabled.
> 
> Signed-off-by: Anusha Arun Nandi<anusha.nandi@oss.qualcomm.com>
> ---
>   .../qualcomm/phy-qcom-mipi-csi2-3ph-dphy.c    | 112 ++++++++++++++++++
>   .../phy/qualcomm/phy-qcom-mipi-csi2-core.c    |   5 +-
>   drivers/phy/qualcomm/phy-qcom-mipi-csi2.h     |   4 +
>   3 files changed, 119 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/phy/qualcomm/phy-qcom-mipi-csi2-3ph-dphy.c b/drivers/phy/qualcomm/phy-qcom-mipi-csi2-3ph-dphy.c
> index 966d79c98f9b..85e89235f7c9 100644
> --- a/drivers/phy/qualcomm/phy-qcom-mipi-csi2-3ph-dphy.c
> +++ b/drivers/phy/qualcomm/phy-qcom-mipi-csi2-3ph-dphy.c
> @@ -380,6 +380,118 @@ const struct mipi_csi2phy_soc_cfg mipi_csi2_dphy_4nm_x1e = {
>   	.num_supplies = ARRAY_SIZE(x1e_supplies),
>   	.clk_names = (const char **)x1e_clks,
>   	.num_clk = ARRAY_SIZE(x1e_clks),
> +	.timer_clk_name = "timer",
> +	.opp_clk_name = "core",
>   	.genpds = x1e_genpds,
>   	.num_genpds = ARRAY_SIZE(x1e_genpds),
>   };

Yes but split it up - "remove hard coded timer and opp names" - actually 
you want to make the OPP optional no ? In which case you'd expect to see 
a patch to make OPPs optional - instead of folding that into a 
functionality drop.

Because say we want to revert or re-order the addition of Lemans in the 
git log - we then have the problem of mixed functionality.

Or say I want to cherry-pick a backport - I don't necessarily want to 
include lots of SoC specific stuff that doesn't affect my part.

=> More granular patches.

---
bod