[PATCH 0/2] regulator: core: always_on fixups

[PATCH 0/2] regulator: core: always_on fixups

[Markus Pargmann]

Hi,

The second patch adds a check for dummy regulators that are not always_on. The
first patch adds a always_on constraint to the dummy regulator.

Regards,

Markus


Markus Pargmann (2):
  regulator: dummy: Should be always-on
  regulator: core: Error on not always_on dummy regulators

 drivers/regulator/core.c  | 27 ++++++++++++++++++++++++---
 drivers/regulator/dummy.c |  6 +++++-
 2 files changed, 29 insertions(+), 4 deletions(-)

-- 
1.8.5.3

[PATCH 1/2] regulator: dummy: Should be always-on

[Markus Pargmann]

Regulator dummy does not have any enable operations. So it is always_on.

Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
---
 drivers/regulator/dummy.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/regulator/dummy.c b/drivers/regulator/dummy.c
index df9f425..2436db9 100644
--- a/drivers/regulator/dummy.c
+++ b/drivers/regulator/dummy.c
@@ -25,7 +25,11 @@
 
 struct regulator_dev *dummy_regulator_rdev;
 
-static struct regulator_init_data dummy_initdata;
+static struct regulator_init_data dummy_initdata = {
+	.constraints = {
+		.always_on = 1,
+	},
+};
 
 static struct regulator_ops dummy_ops;
 
-- 
1.8.5.3

Re: [PATCH 1/2] regulator: dummy: Should be always-on

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 159 bytes --]

On Tue, Feb 25, 2014 at 11:03:46AM +0100, Markus Pargmann wrote:

> Regulator dummy does not have any enable operations. So it is always_on.

Applied, thanks.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

[PATCH 2/2] regulator: core: Error on not always_on dummy regulators

[Markus Pargmann]

It does not make sense to have dummy/fixed regulators that are not
always_on. This patch adds a check in the set_machine_constraints
function for this case and returns -EINVAL if such a regulator is
registered.

This also removes the return -EINVAL from _regulator_do_enable() if the
regulator can not be enabled.

Signed-off-by: Markus Pargmann <mpa@pengutronix.de>
---
 drivers/regulator/core.c | 27 ++++++++++++++++++++++++---
 1 file changed, 24 insertions(+), 3 deletions(-)

diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c
index 6f71985..cb10faf 100644
--- a/drivers/regulator/core.c
+++ b/drivers/regulator/core.c
@@ -954,6 +954,7 @@ static int machine_constraints_current(struct regulator_dev *rdev,
 }
 
 static int _regulator_do_enable(struct regulator_dev *rdev);
+static bool regulator_can_enable(struct regulator_dev *rdev);
 
 /**
  * set_machine_constraints - sets regulator constraints
@@ -1012,12 +1013,19 @@ static int set_machine_constraints(struct regulator_dev *rdev,
 		}
 	}
 
+	if (!rdev->constraints->always_on && !regulator_can_enable(rdev))
+	{
+		rdev_err(rdev, "dummy regulator without always_on constraints is not valid\n");
+		ret = -EINVAL;
+		goto out;
+	}
+
 	/* If the constraints say the regulator should be on at this point
 	 * and we have control then make sure it is enabled.
 	 */
 	if (rdev->constraints->always_on || rdev->constraints->boot_on) {
 		ret = _regulator_do_enable(rdev);
-		if (ret < 0 && ret != -EINVAL) {
+		if (ret) {
 			rdev_err(rdev, "failed to enable\n");
 			goto out;
 		}
@@ -1739,6 +1747,21 @@ static int regulator_ena_gpio_ctrl(struct regulator_dev *rdev, bool enable)
 	return 0;
 }
 
+/*
+ * Returns true if the regulator can be enabled through ops->enable or
+ * enable gpios.
+ */
+static bool regulator_can_enable(struct regulator_dev *rdev)
+{
+	if (rdev->desc->ops->enable)
+		return true;
+
+	if (rdev->ena_pin)
+		return true;
+
+	return false;
+}
+
 static int _regulator_do_enable(struct regulator_dev *rdev)
 {
 	int ret, delay;
@@ -1763,8 +1786,6 @@ static int _regulator_do_enable(struct regulator_dev *rdev)
 		ret = rdev->desc->ops->enable(rdev);
 		if (ret < 0)
 			return ret;
-	} else {
-		return -EINVAL;
 	}
 
 	/* Allow the regulator to ramp; it would be useful to extend
-- 
1.8.5.3

Re: [PATCH 2/2] regulator: core: Error on not always_on dummy regulators

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 580 bytes --]

On Tue, Feb 25, 2014 at 11:03:47AM +0100, Markus Pargmann wrote:
> It does not make sense to have dummy/fixed regulators that are not
> always_on. This patch adds a check in the set_machine_constraints
> function for this case and returns -EINVAL if such a regulator is
> registered.

Such a regulator isn't going to be useful in a system but that doesn't
mean that they shouldn't be registerable since that allows people to
look at their state.  For example a device might have some GPIO managed
regulators where the GPIO is tied low if the regulator isn't in use in
the system.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]