diff mbox

[v4,05/10] power: bq24257: Add SW-based approach for Power Good determination

Message ID 1442339914-25843-6-git-send-email-dannenberg@ti.com (mailing list archive)
State Not Applicable, archived
Headers show

Commit Message

Andreas Dannenberg Sept. 15, 2015, 5:58 p.m. UTC
A software-based approach for determining the charger's input voltage
"Power Good" state is introduced for devices like the bq24250 which
don't have a dedicated hardware pin for that purpose. This SW-based
approach is also used for other devices (with dedicated PG pin) as a
fall back solution if that pin is not configured to be used through
"pg-gpios".

Signed-off-by: Andreas Dannenberg <dannenberg@ti.com>
---
 drivers/power/bq24257_charger.c | 49 ++++++++++++++++++++++++++++++++++++-----
 1 file changed, 43 insertions(+), 6 deletions(-)

Comments

Krzysztof Kozlowski Sept. 16, 2015, 5:41 a.m. UTC | #1
On 16.09.2015 02:58, Andreas Dannenberg wrote:
> A software-based approach for determining the charger's input voltage
> "Power Good" state is introduced for devices like the bq24250 which
> don't have a dedicated hardware pin for that purpose. This SW-based
> approach is also used for other devices (with dedicated PG pin) as a
> fall back solution if that pin is not configured to be used through
> "pg-gpios".
> 
> Signed-off-by: Andreas Dannenberg <dannenberg@ti.com>
> ---
>  drivers/power/bq24257_charger.c | 49 ++++++++++++++++++++++++++++++++++++-----
>  1 file changed, 43 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/power/bq24257_charger.c b/drivers/power/bq24257_charger.c
> index 55e4ee4..d7488cf 100644
> --- a/drivers/power/bq24257_charger.c
> +++ b/drivers/power/bq24257_charger.c
> @@ -103,6 +103,7 @@ struct bq24257_device {
>  	struct mutex lock; /* protect state data */
>  
>  	bool in_ilimit_autoset_disable;
> +	bool pg_gpio_disable;
>  };
>  
>  static bool bq24257_is_volatile_reg(struct device *dev, unsigned int reg)
> @@ -356,7 +357,26 @@ static int bq24257_get_chip_state(struct bq24257_device *bq,
>  
>  	state->fault = ret;
>  
> -	state->power_good = !gpiod_get_value_cansleep(bq->pg);
> +	if (bq->pg_gpio_disable)
> +		/*
> +		 * If we have a chip without a dedicated power-good GPIO or
> +		 * some other explicit bit that would provide this information
> +		 * assume the power is good if there is no supply related
> +		 * fault - and not good otherwise. There is a possibility for
> +		 * other errors to mask that power in fact is not good but this
> +		 * is probably the best we can do here.
> +		 */
> +		switch (state->fault) {
> +		case FAULT_INPUT_OVP:
> +		case FAULT_INPUT_UVLO:
> +		case FAULT_INPUT_LDO_LOW:
> +			state->power_good = false;
> +			break;
> +		default:
> +			state->power_good = true;
> +		}
> +	else
> +		state->power_good = !gpiod_get_value_cansleep(bq->pg);
>  
>  	return 0;
>  }
> @@ -676,7 +696,7 @@ static int bq24257_pg_gpio_probe(struct bq24257_device *bq)
>  {
>  	bq->pg = devm_gpiod_get_index(bq->dev, BQ24257_PG_GPIO, 0, GPIOD_IN);
>  	if (IS_ERR(bq->pg)) {
> -		dev_err(bq->dev, "could not probe PG pin\n");
> +		dev_info(bq->dev, "could not probe PG pin\n");

I think if pg-gpio is provided (e.g. by DTS) but it is invalid (return
value != ENOENT) then it is an error you could print. The driver will
fallback to the software method but still user/developer may want to
notice the error (e.g. error in DTS).

Anyway it is up to you, rest looks good:

Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>

Best regards,
Krzysztof

--
To unsubscribe from this list: send the line "unsubscribe linux-pm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Andreas Dannenberg Sept. 18, 2015, 9:28 p.m. UTC | #2
On Wed, Sep 16, 2015 at 02:41:13PM +0900, Krzysztof Kozlowski wrote:
> On 16.09.2015 02:58, Andreas Dannenberg wrote:
> > @@ -676,7 +696,7 @@ static int bq24257_pg_gpio_probe(struct bq24257_device *bq)
> >  {
> >  	bq->pg = devm_gpiod_get_index(bq->dev, BQ24257_PG_GPIO, 0, GPIOD_IN);
> >  	if (IS_ERR(bq->pg)) {
> > -		dev_err(bq->dev, "could not probe PG pin\n");
> > +		dev_info(bq->dev, "could not probe PG pin\n");
> 
> I think if pg-gpio is provided (e.g. by DTS) but it is invalid (return
> value != ENOENT) then it is an error you could print. The driver will
> fallback to the software method but still user/developer may want to
> notice the error (e.g. error in DTS).
> 
> Anyway it is up to you, rest looks good:

Krzysztof,
I looked at this closer and all the error scenarios I could come up with
to test regarding the pin definition in DT (wrong pin numbers, missing
definition altogether) all generated an -ENOENT coming out of
devm_gpiod_get_index() with no differentiation between the cases. I'm
sure there is some way to accomplish that but for the time being I left
the refinement between dev_err() and dev_info() alone.

Regards,

--
Andreas Dannenberg
Texas Instruments Inc

--
To unsubscribe from this list: send the line "unsubscribe linux-pm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/drivers/power/bq24257_charger.c b/drivers/power/bq24257_charger.c
index 55e4ee4..d7488cf 100644
--- a/drivers/power/bq24257_charger.c
+++ b/drivers/power/bq24257_charger.c
@@ -103,6 +103,7 @@  struct bq24257_device {
 	struct mutex lock; /* protect state data */
 
 	bool in_ilimit_autoset_disable;
+	bool pg_gpio_disable;
 };
 
 static bool bq24257_is_volatile_reg(struct device *dev, unsigned int reg)
@@ -356,7 +357,26 @@  static int bq24257_get_chip_state(struct bq24257_device *bq,
 
 	state->fault = ret;
 
-	state->power_good = !gpiod_get_value_cansleep(bq->pg);
+	if (bq->pg_gpio_disable)
+		/*
+		 * If we have a chip without a dedicated power-good GPIO or
+		 * some other explicit bit that would provide this information
+		 * assume the power is good if there is no supply related
+		 * fault - and not good otherwise. There is a possibility for
+		 * other errors to mask that power in fact is not good but this
+		 * is probably the best we can do here.
+		 */
+		switch (state->fault) {
+		case FAULT_INPUT_OVP:
+		case FAULT_INPUT_UVLO:
+		case FAULT_INPUT_LDO_LOW:
+			state->power_good = false;
+			break;
+		default:
+			state->power_good = true;
+		}
+	else
+		state->power_good = !gpiod_get_value_cansleep(bq->pg);
 
 	return 0;
 }
@@ -676,7 +696,7 @@  static int bq24257_pg_gpio_probe(struct bq24257_device *bq)
 {
 	bq->pg = devm_gpiod_get_index(bq->dev, BQ24257_PG_GPIO, 0, GPIOD_IN);
 	if (IS_ERR(bq->pg)) {
-		dev_err(bq->dev, "could not probe PG pin\n");
+		dev_info(bq->dev, "could not probe PG pin\n");
 		return PTR_ERR(bq->pg);
 	}
 
@@ -810,10 +830,27 @@  static int bq24257_probe(struct i2c_client *client,
 		INIT_DELAYED_WORK(&bq->iilimit_setup_work,
 				  bq24257_iilimit_setup_work);
 
-	/* we can only check Power Good status by probing the PG pin */
-	ret = bq24257_pg_gpio_probe(bq);
-	if (ret < 0)
-		return ret;
+	/*
+	 * The BQ24250 doesn't have a dedicated Power Good (PG) pin so we
+	 * explicitly disable this feature for this device and instead use
+	 * a SW-based approach to determine the PG state.
+	 */
+	if (bq->chip == BQ24250)
+		bq->pg_gpio_disable = true;
+
+	/*
+	 * For devices that do have a dedicated PG pin go ahead and probe it,
+	 * using the SW-based approach as a fall back solution. Note that the
+	 * use of the dedicated pin is preferred.
+	 */
+	if (!bq->pg_gpio_disable) {
+		ret = bq24257_pg_gpio_probe(bq);
+		if (ret < 0) {
+			dev_info(bq->dev,
+				 "using SW-based power-good detection\n");
+			bq->pg_gpio_disable = true;
+		}
+	}
 
 	/* reset all registers to defaults */
 	ret = bq24257_field_write(bq, F_RESET, 1);