Re: [PATCH 2/2] cpufreq: sun50i: replace of_node_put() with automatic cleanup handler

From: Viresh Kumar
Date: Mon May 20 2024 - 05:32:10 EST


On 20-05-24, 11:28, Markus Elfring wrote:
>
> > > > +++ b/drivers/cpufreq/sun50i-cpufreq-nvmem.c
> > > > @@ -131,14 +131,14 @@ static const struct of_device_id cpu_opp_match_list[] = {
> > > > static bool dt_has_supported_hw(void)
> > > > {
> > > > bool has_opp_supported_hw = false;
> > > > - struct device_node *np;
> > > > struct device *cpu_dev;
> > > >
> > > > cpu_dev = get_cpu_device(0);
> > > > if (!cpu_dev)
> > > > return false;
> > > >
> > > > - np = dev_pm_opp_of_get_opp_desc_node(cpu_dev);
> > > > + struct device_node *np __free(device_node) =
> > > > + dev_pm_opp_of_get_opp_desc_node(cpu_dev);
> >
> > Won't that result in build warning, mixed code and definitions now ?
>
> I suggest to take another look at a corresponding information source.
>
> [PATCH v3 04/57] kbuild: Drop -Wdeclaration-after-statement
> https://lore.kernel.org/all/20230612093537.693926033@xxxxxxxxxxxxx/

Ah, I wasn't aware of this one.

> See also:
> https://gcc.gnu.org/onlinedocs/gcc-13.2.0/gcc/Warning-Options.html#index-Wdeclaration-after-statement
>
>
> Would you like to stress a scope reduction for the affected local variable
> by adding any curly brackets?

No, it looks fine.

--
viresh