Re: [PATCH] Add chipid to fc2580.c

[Oliver Schinagl <oliver+list@schinagl.nl>]

On 05-11-12 23:58, Antti Palosaari wrote:
> On 10/30/2012 11:24 AM, Oliver Schinagl wrote:
>> On 29-10-12 02:09, Antti Palosaari wrote:
>>> su 28.10.2012 22:07 Mauro Carvalho Chehab kirjoitti:
>>>> Em Mon, 22 Oct 2012 10:17:26 +0200
>>>> Oliver Schinagl <oliver+list@schinagl.nl> escreveu:
>>>>
>>>>> diff --git a/drivers/media/tuners/fc2580.c
>>>>> b/drivers/media/tuners/fc2580.c
>>>>> index aff39ae..102d942 100644
>>>>> I found a fellow Asus U3100+ user (mentioned him before with the
>>>>> firmware issue) that even when using the latest firmware, still see's
>>>>> 0xff as the chipID.
>>>> You missed to add a signed-off-by on your patch.
>>>>
>>>> Maybe it would make sense, in this case, to print some warning 
>>>> message,
>>>> as this could be due to a bug either at the hardware or at some place
>>>> at the driver, like the gpio config settings for this device.
>>>>
>>>> Anyway, Antti, your call.
>>> I am on holiday now and dont want to look much these things at the
>>> moment.
>>>
>>> Having 0x00 or 0xff as chip id is something very very stupid and not
>>> exits
>>> in real world. It is good indicator I2C operation was failing. Check
>>> GPIOs, see windows sniffs, add sleep, test if other I2C reads are 
>>> working
>>> later, etc. to find out more info and fix it properly. In worst case
>>> it is
>>> possible that I2C reads are not working at all...
>> This was a random report for someone who I assisted via e-mail to get
>> the latest git clone from antti's tree. Building, enabling debugging and
>> getting this information alone took a week. I don't think we have the
>> possibility to get a dump from anything. The stick has been working fine
>> from my understanding using the 0xff tunerID. How to handle support for
>> these 'bugged' tuners, I leave that up to you :)
>
> Honestly I don't want to add hack like that with this little 
> information. It must be found out if all I2C readings are failing, or 
> just the first one, or some other condition. Currently there is only 
> two register reads on that driver. Guess what happens if someone 
> enhances that driver so that one bit from certain register is 
> changed... Set register bit 7, current register value is 0x00. 
> Register value will be 0x7ff as read returns always 0xff :-(
true, do you have some test code, that could test this? or a test module 
that prints some debug information for this specific case? I could then 
have the user in question try it out and report his findings?

oliver
>
>>
>> AFTER your well deserved holiday. Enjoy and have a great time!
>>
>>>
>>>
>>>>>
>>>>> --- a/drivers/media/tuners/fc2580.c
>>>>> +++ b/drivers/media/tuners/fc2580.c
>>>>> @@ -497,6 +497,7 @@ struct dvb_frontend *fc2580_attach(struct
>>>>> dvb_frontend *fe,
>>>>>           switch (chip_id) {
>>>>>           case 0x56:
>>>>>           case 0x5a:
>>>>> +       case 0xff:
>>>>>                   break;
>>>>>           default:
>>>>>                   goto err;
>>>>>
>>>>> -- 
>>>>> To unsubscribe from this list: send the line "unsubscribe 
>>>>> linux-media"
>>>>> in
>>>>> the body of a message to majordomo@vger.kernel.org
>>>>> More majordomo info at http://vger.kernel.org/majordomo-info.html
>>>>
>>>> -- 
>>>> Regards,
>>>> Mauro
>>>>
>>>
>>> -- 
>>> To unsubscribe from this list: send the line "unsubscribe 
>>> linux-media" in
>>> the body of a message to majordomo@vger.kernel.org
>>> More majordomo info at http://vger.kernel.org/majordomo-info.html
>>
>
> regards
> Antti
>