On Wed,  3 Sep 2025 21:17:28 +0800
Zihuan Zhang [off-list ref] wrote:

Replace the manual cpufreq_cpu_put() with __free(put_cpufreq_policy)
annotation for policy references. This reduces the risk of reference
counting mistakes and aligns the code with the latest kernel style.

No functional change intended.

Signed-off-by: Zihuan Zhang <redacted>

This falls into the mess of mixing gotos with cleanup.h usage.

The guidance in cleanup.h IIRC say don't do this.  It isn't (I think) buggy here
but it does make things harder to reason about and generally removes
the point of doing __free.  So I think if you are going to do this one
you need to do it fully which is a little more complex.
Need to deal with parent_cpu_data which isn't that hard.

If you mix the two, Linus may get grumpy!

Return directly here (and after changes below, in all error paths.

 		}
 
 		parent_cpu_data = get_parent_cpu_data(p_data, policy);
-		if (parent_cpu_data) {
-			cpufreq_cpu_put(policy);
+		if (parent_cpu_data)
 			continue;

This is the first use of parent_cpu_data. If it's set at this point
we don't use it at all.  So step 1. Rename this to split this
use from the one that follows.

-		}
 
 		parent_cpu_data = kzalloc(sizeof(*parent_cpu_data),
 						GFP_KERNEL);

This one needs to be
		struct devfreq_cpu_data *parent_cpu_data __free(kfree) =
			kzalloc(sizeof(*parent_cpu_data), GFP_KERNEL);

		

then here we need to ensure we don't free parent_cpu_data. Hence

		list_add_tail(&(no_free_ptr(parent_cpu_data)->node,
			      &p_data->cpu_data_list);

That that point we have passed ownership of the data to the list.

And all this error block goes away.

-err_put_policy:
-	cpufreq_cpu_put(policy);
 err:
 
 	return ret;