regulator/max1586: fix V3 gain calculation integer overflow

On Thu, May 28, 2009 at 10:59 AM, Mark Brown <broonie@opensource.wolfsonmicro.com> wrote:
> On Thu, May 28, 2009 at 07:15:16AM +0200, Philipp Zabel wrote:
>> The V3 regulator can be configured with an external resistor
>> connected to the feedback pin (R24 in the data sheet) to
>> increase the voltage range.
>>
>> For example, hx4700 has R24 = 3.32 kOhm to achieve a maximum
>> V3 voltage of 1.55 V which is needed for 624 MHz CPU frequency.
>>
>> Signed-off-by: Philipp Zabel <philipp.zabel@gmail.com>
>
> Looks good.
>
> Acked-by: Mark Brown <broonie@opensource.wolfsonmicro.com>

Thanks, but it turns out I hit a 32 bit integer overflow in
the gain calculation. I'd like to mend that with the following
patch. Now max_uV could be increased up to 4.294 V, enough to
charge LiPo cells.

Signed-off-by: Philipp Zabel <philipp.zabel@gmail.com>
Acked-by: Robert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: Liam Girdwood <lrg@slimlogic.co.uk>

diff --git a/drivers/regulator/max1586.c b/drivers/regulator/max1586.c
index 92799f4..2c082d3 100644 (file)
--- a/drivers/regulator/max1586.c
+++ b/drivers/regulator/max1586.c
@@ -40,8 +40,8 @@ struct max1586_data {
        struct i2c_client *client;
 
        /* min/max V3 voltage */
-       int min_uV;
-       int max_uV;
+       unsigned int min_uV;
+       unsigned int max_uV;
 
        struct regulator_dev *rdev[0];
 };
@@ -199,8 +199,8 @@ static int max1586_pmic_probe(struct i2c_client *client,
                ret = -EINVAL;
                goto out_unmap;
        }
-       max1586->min_uV = MAX1586_V3_MIN_UV * pdata->v3_gain / 1000000;
-       max1586->max_uV = MAX1586_V3_MAX_UV * pdata->v3_gain / 1000000;
+       max1586->min_uV = MAX1586_V3_MIN_UV / 1000 * pdata->v3_gain / 1000;
+       max1586->max_uV = MAX1586_V3_MAX_UV / 1000 * pdata->v3_gain / 1000;
 
        rdev = max1586->rdev;
        for (i = 0; i < pdata->num_subdevs && i <= MAX1586_V6; i++) {