Re: [PATCH v3 12/14] gpio: regmap: Add optional runtime PM support

[Jonathan Cameron <jic23@kernel.org>]

On Fri, 14 Aug 2026 11:34:13 +0300
Andy Shevchenko <andriy.shevchenko@intel.com> wrote:

> On Thu, Aug 13, 2026 at 03:57:05PM +0200, Janani Sunil wrote:
> > Some gpio-regmap consumers share their regmap with a parent device that
> > may be runtime suspended. GPIO register accesses must resume that device
> > first.
> > 
> > Add an optional pm_dev field and acquire it before register translation
> > or access. Release it using runtime autosuspend after each operation.
> > Keep the device active across the complete direction-output sequence and
> > propagate failure when setting the initial output value.  
> 
> ...
> 
> >  static int gpio_regmap_get(struct gpio_chip *chip, unsigned int offset)
> >  {
> >  	struct gpio_regmap *gpio = gpiochip_get_data(chip);  
> 
> >  
> > -	ret = gpio->reg_mask_xlate(gpio, base, offset, &reg, &mask);
> > +	ret = gpio_regmap_runtime_get(gpio);
> >  	if (ret)
> >  		return ret;
> >  
> > +	ret = gpio->reg_mask_xlate(gpio, base, offset, &reg, &mask);
> > +	if (ret)
> > +		goto out_pm_put;
> > +
> >  	/* ensure we don't spoil any register cache with pin input values */
> >  	if (gpio->reg_dat_base == gpio->reg_set_base) {
> >  		ret = regmap_read_bypassed(gpio->regmap, reg, &val);
> >  		if (ret)
> > -			return ret;
> > +			goto out_pm_put;
> >  
> > -		return !!(val & mask);
> > +		ret = !!(val & mask);
> > +	} else {
> > +		ret = regmap_test_bits(gpio->regmap, reg, mask);
> >  	}
> >  
> > -	return regmap_test_bits(gpio->regmap, reg, mask);
> > +out_pm_put:
> > +	gpio_regmap_runtime_put(gpio);
> > +	return ret;
> >  }  
> 
> Instead of adding ugly goto:s, I would rather define the guard and acquire and
> use in the same way as other PM_RUNTIME_ACQUIRE*() work.
> 
Excellent point Andy.

I was thinking exactly this when looking at v2 just now (and feeling guilty for
being late to the discussion!)

Would need a local GPIO specific variant but that is easy to do - just copy
the way PM_RUNTIME_ACQUIRE*() works.

Given how much it will reduce the code churn + result in a more elegant result
for me this is a necessary change rather than a nice to have.

Thanks,

Jonathan