Re: [PATCH 6/6] OPP: Use mutex locking guards

[David Lechner]

On 4/24/25 5:36 AM, Viresh Kumar wrote:
> Use mutex locking guard in the OPP core.
> 
> No intentional functional impact.

There is an unintentional functional change here.

> 
> Signed-off-by: Viresh Kumar <viresh.kumar@xxxxxxxxxx>
> ---

...

> @@ -2660,17 +2632,16 @@ struct dev_pm_opp *dev_pm_opp_xlate_required_opp(struct opp_table *src_table,
>  		return ERR_PTR(-EBUSY);
>  
>  	for (i = 0; i < src_table->required_opp_count; i++) {
> -		if (src_table->required_opp_tables[i] == dst_table) {
> -			mutex_lock(&src_table->lock);
> +		if (src_table->required_opp_tables[i] != dst_table)
> +			continue;
>  
> +		scoped_guard(mutex, &src_table->lock) {
>  			list_for_each_entry(opp, &src_table->opp_list, node) {
>  				if (opp == src_opp) {
>  					dest_opp = dev_pm_opp_get(opp->required_opps[i]);
>  					break;
>  				}
>  			}
> -
> -			mutex_unlock(&src_table->lock);
>  			break;
>  		}
>  	}

scoped_guard() is implemented as a for loop. So now this break statement
breaks out out the scoped_guard() and not out of the outer for loop. Now
the outer loop always iterates to completion.

Assuming each item in src_table->required_opp_tables is unique, this doesn't
look like a bug. Rather it is just doing unnecessary extra iterations.

To preserve the logic, the break should be moved out of the scoped_guard().