Re: [PATCH] clk: nuvoton: ma35d1: Use clk_hw pointers as mux parents

[Miquel Raynal]

Hello,

> Instead of declaring the parents in the clk_parent_data structure, use
> the actual HW clocks to lookup the parents directly: parents are
> described by an array of indices into the controller's main clock table
> (like in other clock controller drivers), which the "new" mux helper now
> resolves.

Sashiko complained about this:

```
[Severity: High]
Could this lead to a kernel panic if a parent clock registration fails?

Looking deeper, from my point of view the issue is purely
theoretical. The only possibility is one of the early registration
returning -ENOMEM. Despite the fact that this is quite hard to actually
reach, if someone gets this kind of issue early at boot, they will
likely have bigger concerns than their peripheral clocks not being
reachable.
```

Brian, Shan-Chun, Stephen, what do you think? Is this worth another
version?

I can send a v2, or make a follow-up, or just ignore
it. ma35d1_reg_clk_pll() and ma35d1_reg_adc_clkdiv() are equally
"affected".

Thanks for the feedback,
Miquèl