Em 17-12-2011 19:54, Antti Palosaari escreveu:
> On 12/17/2011 10:57 PM, linu...@stefanringel.de wrote:
>> From: Stefan Ringel<linu...@stefanringel.de>
>>
>> Signed-off-by: Stefan Ringel<linu...@stefanringel.de>
>> ---
>>   drivers/media/common/tuners/mt2063.c |   24 ++++++++++++++++++------
>>   1 files changed, 18 insertions(+), 6 deletions(-)
>>
>> diff --git a/drivers/media/common/tuners/mt2063.c 
>> b/drivers/media/common/tuners/mt2063.c
>> index 6743ffe..5b4b1ec 100644
>> --- a/drivers/media/common/tuners/mt2063.c
>> +++ b/drivers/media/common/tuners/mt2063.c
>> @@ -2211,18 +2211,29 @@ static int mt2063_get_frequency(struct dvb_frontend 
>> *fe, u32 *freq)
>>       if (!state->init)
>>           return -ENODEV;
>>
>> -    /*
>> -     * FIXME: This is an API abuse at DRX-K driver: in order to allow
>> -     * tda18271 to change the IF based on the standard, it uses this
>> -     * callback as "get_if_frequency".
>> -     */
>> -    *freq = state->reference * 1000;
>> +    *freq = state->frequency;
>>
>>       dprintk(1, "frequency: %d\n", *freq);
>>
>>       return 0;
>>   }
>>
>> +static int mt2063_get_if_frequency(struct dvb_frontend *fe, u32 *freq)
>> +{
>> +    struct mt2063_state *state = fe->tuner_priv;
>> +
>> +    dprintk(2, "\n");
>> +
>> +    if (!state->init)
>> +        return -ENODEV;
>> +
>> +    *freq = state->AS_Data.f_out;
>> +
>> +    dprintk(1, "if frequency: %d\n", *freq);
>> +
>> +    return 0;
>> +}
>> +
>>   static int mt2063_get_bandwidth(struct dvb_frontend *fe, u32 *bw)
>>   {
>>       struct mt2063_state *state = fe->tuner_priv;
>> @@ -2253,6 +2264,7 @@ static struct dvb_tuner_ops mt2063_ops = {
>>       .set_analog_params = mt2063_set_analog_params,
>>       .set_params    = mt2063_set_params,
>>       .get_frequency = mt2063_get_frequency,
>> +    .get_if_frequency = mt2063_get_if_frequency,
>>       .get_bandwidth = mt2063_get_bandwidth,
>>       .release = mt2063_release,
>>   };
> 
> Without looking the code itself that seems like a little bit suspicious. As a 
> understand .get_frequency was abused as .get_if_frequency and you changed it 
> correctly now. Anyhow, earlier is was returning IF as (state->reference * 
> 1000) and now it does (state->AS_Data.f_out). Is that correct? And if yes, 
> why there is two variables in state having IF ?

Antti,

Stefan based his patch on my az6007 development tree:
        
http://git.linuxtv.org/mchehab/experimental.git/shortlog/refs/heads/az6007-2

As there is a driver there for mt2063.

Regards,
Mauro.
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to