[PATCH] regulator: palmas: Fix calculating selector in palmas_map_voltage_ldo

[PATCH] regulator: palmas: Fix calculating selector in palmas_map_voltage_ldo

[Axel Lin]

This patch fixes below issues when choosing selector:

1. Current code returns negative selector if min_uV < 900000 which is wrong.
   For example, it is possible to satisfy the request with selector = 1 if
   the requested min_uV is 850000.
2. Current code may select a voltage lower than requested min_uV.
   For example, if the requested min_uV is 945000, current code chooses
   selector = 1 which is lower than requested min_uV.
   DIV_ROUND_UP to avoid this case.

Signed-off-by: Axel Lin <axel.lin@gmail.com>
---
 drivers/regulator/palmas-regulator.c |    9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/drivers/regulator/palmas-regulator.c b/drivers/regulator/palmas-regulator.c
index 17d19fb..0fcf355 100644
--- a/drivers/regulator/palmas-regulator.c
+++ b/drivers/regulator/palmas-regulator.c
@@ -486,9 +486,12 @@ static int palmas_map_voltage_ldo(struct regulator_dev *rdev,
 {
 	int ret, voltage;
 
-	ret = ((min_uV - 900000) / 50000) + 1;
-	if (ret < 0)
-		return ret;
+	if (min_uV == 0)
+		return 0;
+
+	if (min_uV < 900000)
+		min_uV = 900000;
+	ret = DIV_ROUND_UP(min_uV - 900000, 50000) + 1;
 
 	/* Map back into a voltage to verify we're still in bounds */
 	voltage = palmas_list_voltage_ldo(rdev, ret);
-- 
1.7.9.5

Re: [PATCH] regulator: palmas: Fix calculating selector in palmas_map_voltage_ldo

[Graeme Gregory]

On 18/07/12 04:30, Axel Lin wrote:
> This patch fixes below issues when choosing selector:
>
> 1. Current code returns negative selector if min_uV < 900000 which is wrong.
>    For example, it is possible to satisfy the request with selector = 1 if
>    the requested min_uV is 850000.
Isnt this defending against bugs in the regulator core? I am normally
against repeating the same checks at multiple levels in a stack.
> 2. Current code may select a voltage lower than requested min_uV.
>    For example, if the requested min_uV is 945000, current code chooses
>    selector = 1 which is lower than requested min_uV.
>    DIV_ROUND_UP to avoid this case.
>
> Signed-off-by: Axel Lin <axel.lin@gmail.com>
> ---
>  drivers/regulator/palmas-regulator.c |    9 ++++++---
>  1 file changed, 6 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/regulator/palmas-regulator.c b/drivers/regulator/palmas-regulator.c
> index 17d19fb..0fcf355 100644
> --- a/drivers/regulator/palmas-regulator.c
> +++ b/drivers/regulator/palmas-regulator.c
> @@ -486,9 +486,12 @@ static int palmas_map_voltage_ldo(struct regulator_dev *rdev,
>  {
>  	int ret, voltage;
>  
> -	ret = ((min_uV - 900000) / 50000) + 1;
> -	if (ret < 0)
> -		return ret;
> +	if (min_uV == 0)
> +		return 0;
> +
> +	if (min_uV < 900000)
> +		min_uV = 900000;
> +	ret = DIV_ROUND_UP(min_uV - 900000, 50000) + 1;
>  
>  	/* Map back into a voltage to verify we're still in bounds */
>  	voltage = palmas_list_voltage_ldo(rdev, ret);

Re: [PATCH] regulator: palmas: Fix calculating selector in palmas_map_voltage_ldo

[Mark Brown]

[-- Attachment #1: Type: text/plain, Size: 845 bytes --]

On Wed, Jul 18, 2012 at 08:55:30AM +0100, Graeme Gregory wrote:
> On 18/07/12 04:30, Axel Lin wrote:
> > This patch fixes below issues when choosing selector:

> > 1. Current code returns negative selector if min_uV < 900000 which is wrong.
> >    For example, it is possible to satisfy the request with selector = 1 if
> >    the requested min_uV is 850000.

> Isnt this defending against bugs in the regulator core? I am normally
> against repeating the same checks at multiple levels in a stack.

This isn't a bug in the core, a negative selector is an error code (and
in general whatever selector is returned by the driver ought to agree
with selectors the driver actually supports).  It's probably at least a
questionable decision in the machine constraints to allow a request for
less than the regulator can deliver to get through though.

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