[linux-dvb] Kworld DVB-T 210 - dvb tuning problem

[timf]

hermann pitton wrote:
> Hi Hartmut,
>
> Am Freitag, den 11.04.2008, 23:23 +0200 schrieb Hartmut Hackmann:
>   
>> Hi, Hermann
>>
>> hermann pitton schrieb:
>>     
>>> Am Freitag, den 11.04.2008, 00:08 +0200 schrieb Hartmut Hackmann:
>>>       
>>>> HI, Tim
>>>>
>>>> timf schrieb:
>>>>         
>>>>> Hi Hartmut,
>>>>> OK, found some more spare time, but very, very frustrated!
>>>>>
>>>>> 1) Tried ubuntu 7.04, 7.10, 8.04
>>>>>     Tried with just modules that exist in kernel (no v4l-dvb)
>>>>>    Tried v4l-dvb from June 2007 and tried current v4l-dvb
>>>>>    Tried with/without Hartmut patch - changeset 7376    49ba58715fe0
>>>>>    Tried with .gpio_config   = TDA10046_GP11_I, or .gpio_config   = 
>>>>> TDA10046_GP01_I,
>>>>>    Tried using configs in saa7134-dvb.c matching tiger, tiger_s, 
>>>>> pinnacle 310i, twinhan 3056
>>>>>
>>>>>     # Australia / Perth (Roleystone transmitter)
>>>>>     # T freq bw fec_hi fec_lo mod transmission-mode guard-interval hierarchy
>>>>>     # SBS
>>>>>     T 704500000 7MHz 2/3 NONE QAM64 8k 1/8 NONE
>>>>>     # ABC
>>>>>     T 725500000 7MHz 3/4 NONE QAM64 8k 1/16 NONE
>>>>>     # Seven
>>>>>     T 746500000 7MHz 2/3 NONE QAM64 8k 1/16 NONE
>>>>>     # Nine
>>>>>     T 767500000 7MHz 3/4 NONE QAM64 8k 1/16 NONE
>>>>>     # Ten
>>>>>     T 788500000 7MHz 3/4 NONE QAM64 8k 1/16 NONE
>>>>>
>>>>> 2) I have these saa7134 cards:
>>>>>     - pinnacle 310i
>>>>>     - kworld 210
>>>>>
>>>>>     This cx88 card:
>>>>>     - dvico DVB-T Pro hybrid (analog tv not work)
>>>>>
>>>>> -   problem only occurs with kworld 210 in linux (works fine in WinXP)
>>>>>
>>>>> 3) In WinXP, all channels, both analog tv and dvb-t found
>>>>>
>>>>> 4) In linux, if start dvb-t first, never scans SBS - dmesg1
>>>>>
>>>>> 5) In linux, if start analog tv first, stop, then start dvb-t, scan 
>>>>> finds SBS - dmesg2
>>>>>
>>>>>           
>>>> a) The pinnacle 310i finds everything?
>>>>     It has the same chipset, but an almost perfectly handled tuner chip...
>>>>     This means that your initial config file is ok...
>>>> b) Does this mean that in case 4, all other channels are found?
>>>> c) Case 5: This finds everything?
>>>> d) What happens if you use the scan data of the pinnacle card?
>>>>     Does it tune SBS? Does it just take more time to stabilize?
>>>>     This can be understood.
>>>> e) Just to be sure: did you clarify the open point with .antenna_switch
>>>>     (i think so)
>>>> f) the kernel logs are as expected.
>>>> <snip>
>>>>
>>>>
>>>>         
>>>>> 6) Herman mentioned something called a "mode-switch" in the archives, 
>>>>> but not any description.
>>>>>           
>>>> I guess he meant the switching between analog, radio and dvb-t. This is the
>>>> GPIO handling and card depending.
>>>>         
>>> Tim must have it from when I mentioned the special case of card=87 and
>>> 94.
>>>
>>>       
>>>>> I tried to find some data sheets for tda8275 tda8290 but only found the 
>>>>> publicity pdf file from Phillips,
>>>>> so at least I can see they go together, so I presume this "mode-switch" 
>>>>> is coded into those modules.
>>>>> But those modules work for all other cards, so now I'm lost again.
>>>>>
>>>>> What else should I try?
>>>>>
>>>>>           
>>>> If my assumptions above are wrong, there is one other chance:
>>>> Recently i saw another card that does the (unusual) mode switching
>>>> like card 87. So to be sure, you might try to force this card type (be
>>>> aware of the antenna inputs, if in doubt, try both.
>>>>
>>>> Best regards
>>>>    Hartmut
>>>>
>>>>         
>>> For the Medion8800 Quad and CTX948 also showing this issue, needs to
>>> tune analog first to have good recepton on DVB-T, they are a little
>>> weaker on analog than other cards, but after that on DVB-T, they are as
>>> good than known good others.
>>>
>>> Cheers,
>>> Hermann
>>>
>>>       
>> Is this problem still there with the recent v4l-dvb code?
>> I found a problem with the GPIO initialization and fixed
>> this in patch 49ba58715fe0 3 weeks ago (The gpiomask was
>> not set until analog tuning occured).
>>
>> Best regards
>>    Hartmut
>>     
>
> argh, my DVB-T signal can not be split much, so I left that other
> machine with DVB-S, cable-tv and the empress almost untouched during the
> last weeks and without DVB-T.
>
> diff -r 1e295a94038e -r 49ba58715fe0 linux/drivers/media/video/saa7134/saa7134-core.c
> --- a/linux/drivers/media/video/saa7134/saa7134-core.c  Mon Mar 03 22:55:05 2008 +0100
> +++ b/linux/drivers/media/video/saa7134/saa7134-core.c  Sun Mar 16 23:49:43 2008 +0100
> @@ -960,6 +960,7 @@ static int __devinit saa7134_initdev(str
>         struct saa7134_dev *dev;
>         struct saa7134_mpeg_ops *mops;
>         int err;
> +       int mask;
>
>         dev = kzalloc(sizeof(*dev),GFP_KERNEL);
>         if (NULL == dev)
> @@ -1157,6 +1158,11 @@ static int __devinit saa7134_initdev(str
>         if (TUNER_ABSENT != dev->tuner_type)
>                 saa7134_i2c_call_clients(dev, TUNER_SET_STANDBY, NULL);
>
> +       if (card(dev).gpiomask != 0) {
> +               mask = card(dev).gpiomask;
> +               saa_andorl(SAA7134_GPIO_GPMODE0 >> 2,   mask, mask);
> +               saa_andorl(SAA7134_GPIO_GPSTATUS0 >> 2, mask, 0);
> +       }
>         return 0;
>
>   fail4:
>
> It is fixed!
>
> Thanks,
> Hermann
>
>
>   
Hi Hartmut and Hermann,

Yes, this patch has already been implemented in my saa7134-core.c, only 
differences are line numbers:

line 963      int mask;

line 1164    if (card(dev).gpiomask != 0) {
                            mask = card(dev).gpiomask;
                            saa_andorl(SAA7134_GPIO_GPMODE0 >> 2,   
mask, mask);
                            saa_andorl(SAA7134_GPIO_GPSTATUS0 >> 2, 
mask, 0);


Regards,
Tim