All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH v1] regulator: gpio: check return value of of_get_named_gpio
@ 2016-03-10 10:04 Mihai Mihalache
  2016-03-12  6:23 ` Mark Brown
  0 siblings, 1 reply; 2+ messages in thread
From: Mihai Mihalache @ 2016-03-10 10:04 UTC (permalink / raw)
  To: linux-kernel, broonie, lgirdwood
  Cc: mihai.d.mihalache, hans.holmberg, yousaf.kaukab

At boot time the regulator driver can be initialized before the
gpio, in which case the call to of_get_named_gpio will return
EPROBE_DEFER. This value is silently passed to regulator_register
which will return success, although the gpio is not registered
(regulator_ena_gpio_request not called) as the value passed is
detected as invalid. The gpio_regulator_probe will therefore
succeed win no gpio requested.

Signed-off-by: Mihai Mihalache <mihai.d.mihalache@intel.com>
---

History:
	V1:
	- Fix comment from Mark Brown

 drivers/regulator/gpio-regulator.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/regulator/gpio-regulator.c b/drivers/regulator/gpio-regulator.c
index 464018d..9108c57 100644
--- a/drivers/regulator/gpio-regulator.c
+++ b/drivers/regulator/gpio-regulator.c
@@ -162,6 +162,8 @@ of_get_gpio_regulator_config(struct device *dev, struct device_node *np,
 	of_property_read_u32(np, "startup-delay-us", &config->startup_delay);
 
 	config->enable_gpio = of_get_named_gpio(np, "enable-gpio", 0);
+	if (IS_ERR_VALUE(config->enable_gpio))
+		return ERR_PTR(config->enable_gpio);
 
 	/* Fetch GPIOs. - optional property*/
 	ret = of_gpio_count(np);
-- 
2.1.4

^ permalink raw reply related	[flat|nested] 2+ messages in thread

* Re: [PATCH v1] regulator: gpio: check return value of of_get_named_gpio
  2016-03-10 10:04 [PATCH v1] regulator: gpio: check return value of of_get_named_gpio Mihai Mihalache
@ 2016-03-12  6:23 ` Mark Brown
  0 siblings, 0 replies; 2+ messages in thread
From: Mark Brown @ 2016-03-12  6:23 UTC (permalink / raw)
  To: Mihai Mihalache; +Cc: linux-kernel, lgirdwood, hans.holmberg, yousaf.kaukab

[-- Attachment #1: Type: text/plain, Size: 542 bytes --]

On Thu, Mar 10, 2016 at 02:04:59AM -0800, Mihai Mihalache wrote:

>  	config->enable_gpio = of_get_named_gpio(np, "enable-gpio", 0);
> +	if (IS_ERR_VALUE(config->enable_gpio))
> +		return ERR_PTR(config->enable_gpio);

This is an incompatible behaviour change.  enable-gpio is an optional
property, we check later if we got a valid GPIO and happily continue if
we didn't.  The driver should instead explicitly check for -EPROBE_DEFER
and handle that case separately, allowing the device to probe in cases
where the GPIO is just not provided.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 473 bytes --]

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2016-03-13  8:35 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-03-10 10:04 [PATCH v1] regulator: gpio: check return value of of_get_named_gpio Mihai Mihalache
2016-03-12  6:23 ` Mark Brown

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.