[PATCH] mt9v032: Correct the logic for the auto-exposure setting

[PATCH] mt9v032: Correct the logic for the auto-exposure setting

[Kartik Mohta]

The driver uses the ctrl value passed in as a bool to determine whether
to enable auto-exposure, but the auto-exposure setting is defined as an
enum where AUTO has a value of 0 and MANUAL has a value of 1. This leads
to a reversed logic where if you send in AUTO, it actually sets manual
exposure and vice-versa.

Signed-off-by: Kartik Mohta <kartikmohta@gmail.com>
---
 drivers/media/video/mt9v032.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/drivers/media/video/mt9v032.c b/drivers/media/video/mt9v032.c
index 75e253a..8ea8737 100644
--- a/drivers/media/video/mt9v032.c
+++ b/drivers/media/video/mt9v032.c
@@ -470,6 +470,7 @@ static int mt9v032_s_ctrl(struct v4l2_ctrl *ctrl)
 			container_of(ctrl->handler, struct mt9v032, ctrls);
 	struct i2c_client *client = v4l2_get_subdevdata(&mt9v032->subdev);
 	u16 data;
+	int aec_value;
 
 	switch (ctrl->id) {
 	case V4L2_CID_AUTOGAIN:
@@ -480,8 +481,13 @@ static int mt9v032_s_ctrl(struct v4l2_ctrl *ctrl)
 		return mt9v032_write(client, MT9V032_ANALOG_GAIN, ctrl->val);
 
 	case V4L2_CID_EXPOSURE_AUTO:
+		if(ctrl->val == V4L2_EXPOSURE_MANUAL)
+			aec_value = 0;
+		else
+			aec_value = 1;
+
 		return mt9v032_update_aec_agc(mt9v032, MT9V032_AEC_ENABLE,
-					      ctrl->val);
+					      aec_value);
 
 	case V4L2_CID_EXPOSURE:
 		return mt9v032_write(client, MT9V032_TOTAL_SHUTTER_WIDTH,
-- 
1.7.10.1

Re: [PATCH] mt9v032: Correct the logic for the auto-exposure setting

[Laurent Pinchart]

Hi Kartik,

Thank you for the patch.

On Wednesday 02 May 2012 18:19:08 Kartik Mohta wrote:
> The driver uses the ctrl value passed in as a bool to determine whether
> to enable auto-exposure, but the auto-exposure setting is defined as an
> enum where AUTO has a value of 0 and MANUAL has a value of 1. This leads
> to a reversed logic where if you send in AUTO, it actually sets manual
> exposure and vice-versa.
> 
> Signed-off-by: Kartik Mohta <kartikmohta@gmail.com>
> ---
>  drivers/media/video/mt9v032.c |    8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/media/video/mt9v032.c b/drivers/media/video/mt9v032.c
> index 75e253a..8ea8737 100644
> --- a/drivers/media/video/mt9v032.c
> +++ b/drivers/media/video/mt9v032.c
> @@ -470,6 +470,7 @@ static int mt9v032_s_ctrl(struct v4l2_ctrl *ctrl)
>  			container_of(ctrl->handler, struct mt9v032, ctrls);
>  	struct i2c_client *client = v4l2_get_subdevdata(&mt9v032->subdev);
>  	u16 data;
> +	int aec_value;
> 
>  	switch (ctrl->id) {
>  	case V4L2_CID_AUTOGAIN:
> @@ -480,8 +481,13 @@ static int mt9v032_s_ctrl(struct v4l2_ctrl *ctrl)
>  		return mt9v032_write(client, MT9V032_ANALOG_GAIN, ctrl->val);
> 
>  	case V4L2_CID_EXPOSURE_AUTO:
> +		if(ctrl->val == V4L2_EXPOSURE_MANUAL)
> +			aec_value = 0;
> +		else
> +			aec_value = 1;
> +
>  		return mt9v032_update_aec_agc(mt9v032, MT9V032_AEC_ENABLE,
> -					      ctrl->val);
> +					      aec_value);

What about just

		return mt9v032_update_aec_agc(mt9v032, MT9V032_AEC_ENABLE,
					      !ctrl->val);

If you're fine with that change I'll modify the patch accordingly, there's no 
need to resubmit (I'll of course keep the patch attribution).

> 
>  	case V4L2_CID_EXPOSURE:
>  		return mt9v032_write(client, MT9V032_TOTAL_SHUTTER_WIDTH,

-- 
Regards,

Laurent Pinchart

Re: [PATCH] mt9v032: Correct the logic for the auto-exposure setting

[Kartik Mohta]

Hi,

On Tue, May 8, 2012 at 7:12 AM, Laurent Pinchart
<laurent.pinchart@ideasonboard.com> wrote:
> Hi Kartik,
>
> Thank you for the patch.
>
> On Wednesday 02 May 2012 18:19:08 Kartik Mohta wrote:
>> The driver uses the ctrl value passed in as a bool to determine whether
>> to enable auto-exposure, but the auto-exposure setting is defined as an
>> enum where AUTO has a value of 0 and MANUAL has a value of 1. This leads
>> to a reversed logic where if you send in AUTO, it actually sets manual
>> exposure and vice-versa.
>>
>> Signed-off-by: Kartik Mohta <kartikmohta@gmail.com>
>> ---
>>  drivers/media/video/mt9v032.c |    8 +++++++-
>>  1 file changed, 7 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/media/video/mt9v032.c b/drivers/media/video/mt9v032.c
>> index 75e253a..8ea8737 100644
>> --- a/drivers/media/video/mt9v032.c
>> +++ b/drivers/media/video/mt9v032.c
>> @@ -470,6 +470,7 @@ static int mt9v032_s_ctrl(struct v4l2_ctrl *ctrl)
>>                       container_of(ctrl->handler, struct mt9v032, ctrls);
>>       struct i2c_client *client = v4l2_get_subdevdata(&mt9v032->subdev);
>>       u16 data;
>> +     int aec_value;
>>
>>       switch (ctrl->id) {
>>       case V4L2_CID_AUTOGAIN:
>> @@ -480,8 +481,13 @@ static int mt9v032_s_ctrl(struct v4l2_ctrl *ctrl)
>>               return mt9v032_write(client, MT9V032_ANALOG_GAIN, ctrl->val);
>>
>>       case V4L2_CID_EXPOSURE_AUTO:
>> +             if(ctrl->val == V4L2_EXPOSURE_MANUAL)
>> +                     aec_value = 0;
>> +             else
>> +                     aec_value = 1;
>> +
>>               return mt9v032_update_aec_agc(mt9v032, MT9V032_AEC_ENABLE,
>> -                                           ctrl->val);
>> +                                           aec_value);
>
> What about just
>
>                return mt9v032_update_aec_agc(mt9v032, MT9V032_AEC_ENABLE,
>                                              !ctrl->val);
>
> If you're fine with that change I'll modify the patch accordingly, there's no
> need to resubmit (I'll of course keep the patch attribution).
>

That should work since the only supported exposure modes are auto and
manual with enum values 0 and 1 respectively, but then aren't you
depending on the values of the enum to not change in the future? Also
the change gives an impression that the value is a bool which it is
not. If that is fine, you can change it.

-- 
Kartik

Re: [PATCH] mt9v032: Correct the logic for the auto-exposure setting

[Laurent Pinchart]

Hi Kartik,

On Tuesday 08 May 2012 11:51:29 Kartik Mohta wrote:
> On Tue, May 8, 2012 at 7:12 AM, Laurent Pinchart wrote:
> > On Wednesday 02 May 2012 18:19:08 Kartik Mohta wrote:
> >> The driver uses the ctrl value passed in as a bool to determine whether
> >> to enable auto-exposure, but the auto-exposure setting is defined as an
> >> enum where AUTO has a value of 0 and MANUAL has a value of 1. This leads
> >> to a reversed logic where if you send in AUTO, it actually sets manual
> >> exposure and vice-versa.
> >> 
> >> Signed-off-by: Kartik Mohta <kartikmohta@gmail.com>
> >> ---
> >>  drivers/media/video/mt9v032.c |    8 +++++++-
> >>  1 file changed, 7 insertions(+), 1 deletion(-)
> >> 
> >> diff --git a/drivers/media/video/mt9v032.c
> >> b/drivers/media/video/mt9v032.c
> >> index 75e253a..8ea8737 100644
> >> --- a/drivers/media/video/mt9v032.c
> >> +++ b/drivers/media/video/mt9v032.c
> >> @@ -470,6 +470,7 @@ static int mt9v032_s_ctrl(struct v4l2_ctrl *ctrl)
> >>                       container_of(ctrl->handler, struct mt9v032, ctrls);
> >>       struct i2c_client *client = v4l2_get_subdevdata(&mt9v032->subdev);
> >>       u16 data;
> >> +     int aec_value;
> >> 
> >>       switch (ctrl->id) {
> >>       case V4L2_CID_AUTOGAIN:
> >> @@ -480,8 +481,13 @@ static int mt9v032_s_ctrl(struct v4l2_ctrl *ctrl)
> >>               return mt9v032_write(client, MT9V032_ANALOG_GAIN,
> >> ctrl->val);
> >> 
> >>       case V4L2_CID_EXPOSURE_AUTO:
> >> +             if(ctrl->val == V4L2_EXPOSURE_MANUAL)
> >> +                     aec_value = 0;
> >> +             else
> >> +                     aec_value = 1;
> >> +
> >>               return mt9v032_update_aec_agc(mt9v032, MT9V032_AEC_ENABLE,
> >> -                                           ctrl->val);
> >> +                                           aec_value);
> > 
> > What about just
> > 
> >                return mt9v032_update_aec_agc(mt9v032, MT9V032_AEC_ENABLE,
> >                                              !ctrl->val);
> > 
> > If you're fine with that change I'll modify the patch accordingly, there's
> > no need to resubmit (I'll of course keep the patch attribution).
> 
> That should work since the only supported exposure modes are auto and manual
> with enum values 0 and 1 respectively, but then aren't you depending on the
> values of the enum to not change in the future?

The values are part of the V4L2 public API so they can't change.

> Also the change gives an impression that the value is a bool which it is
> not. If that is fine, you can change it.

OK thank you.

-- 
Regards,

Laurent Pinchart