Re: [PATCH] counter: intel-qep: Replace manual mutex logic with lock guards

["Maxwell Doose" <m32285159@gmail.com>]

Hi Joao, comments inline.

On Thu May 7, 2026 at 10:40 PM CDT, Joao Paulo Menezes Linaris wrote:
> Use scoped_guard() for handling mutex lock instead of locking and
> unlocking mutex explicitly. This improves readability by eliminating
> the need for gotos and by clearly indicating mutex will be locked only
> when execution is in guard scope.
>
> Signed-off-by: Joao Paulo Menezes Linaris <jplinaris@usp.br>
> Co-developed-by: Guilherme Dias <guilhermeabreu200105@usp.br>
> Signed-off-by: Guilherme Dias <guilhermeabreu200105@usp.br>
> ---
> v1 -> v2:
>  - fold the commit message properly
>  - replace scoped_guard() with guard()
>  - remove unnecessary comments
>

If this is the v2 then the subject should be:

[PATCH v2] counter: ...rest of subject here...

You can do this by running git format-patch -1 -v2.

> +	guard(mutex)(&qep->lock);
> +	if (qep->enabled)
> +		return -EBUSY;
>  

Would recommend putting blank lines between guard(mutex)() and other
code, Andy usually treats the guard()()s as separate things.

>
>  	pm_runtime_get_sync(qep->dev);
>  	intel_qep_writel(qep, INTEL_QEPMAX, max);
>  	pm_runtime_put(qep->dev);
>  
> -out:
> -	mutex_unlock(&qep->lock);
> -	return ret;
> +	return 0;
>  }
>  
>  static int intel_qep_enable_read(struct counter_device *counter,
> @@ -226,10 +221,10 @@ static int intel_qep_enable_write(struct counter_device *counter,
>  	u32 reg;
>  	bool changed;
>  
> -	mutex_lock(&qep->lock);
> +	guard(mutex)(&qep->lock);
>

Again (and so on and so forth).

best regards,
max