Re: Fixed (Was:Re: saa7134/2.6.26 regression, noisy output)

["tomlohave@gmail.com" <tomlohave@gmail.com>]

hermann pitton a écrit :
> Hi,
>
> Am Sonntag, den 17.05.2009, 15:52 +0200 schrieb tomlohave@gmail.com:
>   
>> hermann pitton a écrit :
>>     
>>> Hi Anders,
>>>
>>> Am Freitag, den 15.05.2009, 11:18 +0200 schrieb Anders Eriksson:
>>>   
>>>       
>>>> Success!
>>>>
>>>> I've tracked down the offending change. switch_addr takes on the wrong value
>>>> and setting the LNA fails. Here's a i2c dump:
>>>>
>>>> saa7133[0]: i2c eeprom e0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
>>>> saa7133[0]: i2c eeprom f0: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
>>>> saa7133[0]: i2c xfer: < 20 ERROR: NO_DEVICE
>>>> saa7133[0]: i2c xfer: < 84 ERROR: NO_DEVICE
>>>> saa7133[0]: i2c xfer: < 86 ERROR: NO_DEVICE
>>>> saa7133[0]: i2c xfer: < 94 ERROR: NO_DEVICE
>>>> saa7133[0]: i2c xfer: < 96 >
>>>> saa7133[0]: i2c xfer: < 96 00 >
>>>> saa7133[0]: i2c xfer: < 97 =01 =01 =00 =11 =01 =04 =01 =85 >
>>>> saa7133[0]: i2c xfer: < 96 1f >
>>>> saa7133[0]: i2c xfer: < 97 =89 >
>>>> tda8290_probe: tda8290 detected @ 1-004b
>>>> tuner' 1-004b: tda829x detected
>>>> tuner' 1-004b: Setting mode_mask to 0x0e
>>>> tuner' 1-004b: chip found @ 0x96 (saa7133[0])
>>>> tuner' 1-004b: tuner 0x4b: Tuner type absent
>>>> tuner' i2c attach [addr=0x4b,client=(tuner unset)]
>>>> tuner' 1-004b: Calling set_type_addr for type=54, addr=0xff, mode=0x04, config=0x01
>>>> tuner' 1-004b: set addr for type -1
>>>> tuner' 1-004b: defining GPIO callback
>>>> saa7133[0]: i2c xfer: < 96 1f >
>>>> saa7133[0]: i2c xfer: < 97 =89 >
>>>> tda8290_probe: tda8290 detected @ 1-004b
>>>> saa7133[0]: i2c xfer: < 96 2f >
>>>> saa7133[0]: i2c xfer: < 97 =00 >
>>>> saa7133[0]: i2c xfer: < 96 21 c0 >
>>>> saa7133[0]: i2c xfer: < c1 ERROR: NO_DEVICE
>>>> saa7133[0]: i2c xfer: < c3 =88 >
>>>> saa7133[0]: i2c xfer: < c5 ERROR: NO_DEVICE
>>>> saa7133[0]: i2c xfer: < c7 ERROR: NO_DEVICE
>>>> saa7133[0]: i2c xfer: < 96 21 00 >
>>>> tda829x 1-004b: setting tuner address to 61
>>>> saa7133[0]: i2c xfer: < 96 21 c0 >
>>>> saa7133[0]: i2c xfer: < c3 =08 >
>>>> tda827x: tda827x_attach:
>>>> tda827x: type set to Philips TDA827X
>>>> saa7133[0]: i2c xfer: < c3 =08 >
>>>> tda827x: tda827xa tuner found
>>>> tda827x: tda827x_init:
>>>> tda827x: tda827xa_sleep:
>>>> saa7133[0]: i2c xfer: < c2 30 90 >
>>>> saa7133[0]: i2c xfer: < 96 21 00 >
>>>> tda829x 1-004b: type set to tda8290+75a
>>>> saa7133[0]: i2c xfer: < 96 21 c0 >
>>>> saa7133[0]: i2c xfer: < c2 00 00 00 00 dc 05 8b 0c 04 20 ff 00 00 4b >
>>>> saa7133[0]: i2c xfer: < 96 21 00 >
>>>> saa7133[0]: i2c xfer: < 96 20 01 >
>>>> saa7133[0]: i2c xfer: < 96 30 6f >
>>>> tuner' 1-004b: type set to tda8290+75a
>>>> tuner' 1-004b: tv freq set to 400.00
>>>> tda829x 1-004b: setting tda829x to system xx
>>>> tda829x 1-004b: tda827xa config is 0x01
>>>> saa7133[0]: i2c xfer: < 96 01 00 >
>>>> saa7133[0]: i2c xfer: < 96 02 00 >
>>>> saa7133[0]: i2c xfer: < 96 00 00 >
>>>> saa7133[0]: i2c xfer: < 96 01 90 >
>>>> saa7133[0]: i2c xfer: < 96 28 14 >
>>>> saa7133[0]: i2c xfer: < 96 0f 88 >
>>>> saa7133[0]: i2c xfer: < 96 05 04 >
>>>> saa7133[0]: i2c xfer: < 96 0d 47 >
>>>> saa7133[0]: i2c xfer: < 96 21 c0 >
>>>> tda827x: setting tda827x to system xx
>>>> tda827x: setting LNA to high gain
>>>> saa7133[0]: i2c xfer: < 96 22 00 >
>>>>                         ^ This address is c2 in all kernels <= 5823b3a63c7661272ea7fef7635955e2a50d17eb
>>>>
>>>>
>>>> saa7133[0]: i2c xfer: < c2 00 32 f8 00 16 3b bb 1c 04 20 00 >
>>>> saa7133[0]: i2c xfer: < c2 90 ff e0 00 99 >
>>>> saa7133[0]: i2c xfer: < c2 a0 c0 >
>>>> saa7133[0]: i2c xfer: < c2 30 10 >
>>>> saa7133[0]: i2c xfer: < c3 =49 =a4 >
>>>> tda827x: AGC2 gain is: 10
>>>>                        ^ The gain reported on good kernels is 3 
>>>>
>>>> Looking at the source, the switch_addr to use in the later kernels is somehow 
>>>> autodetected. How that's done, I've yet to fully understand, but somehow it 
>>>> comes up with the wrong address.
>>>>
>>>> This patch (which obviously needs improvement) hardwires the address back to 
>>>> its original value, and works for 2.6.30-rc5.
>>>>
>>>> diff --git a/drivers/media/common/tuners/tda8290.c b/drivers/media/common/tuners/tda8290.c
>>>> index 064d14c..498cc7b 100644
>>>> --- a/drivers/media/common/tuners/tda8290.c
>>>> +++ b/drivers/media/common/tuners/tda8290.c
>>>> @@ -635,7 +635,11 @@ static int tda829x_find_tuner(struct dvb_frontend *fe)
>>>>  
>>>>                 dvb_attach(tda827x_attach, fe, priv->tda827x_addr,
>>>>                            priv->i2c_props.adap, &priv->cfg);
>>>> +               tuner_info("ANDERS: setting switch_addr. was 0x%02x, new 0x%02x\n",priv->cfg.switch_addr,priv->i2c_props.addr);
>>>>                 priv->cfg.switch_addr = priv->i2c_props.addr;
>>>> +               priv->cfg.switch_addr = 0xc2 / 2;
>>>> +               tuner_info("ANDERS: new 0x%02x\n",priv->cfg.switch_addr);
>>>> +
>>>>         }
>>>>         if (fe->ops.tuner_ops.init)
>>>>                 fe->ops.tuner_ops.init(fe);
>>>>
>>>>
>>>> Could you please help me out and shed some light on what the proper fix is for 
>>>> setting switch_addr? 
>>>>
>>>> Thanks,
>>>> /Anders
>>>>
>>>>     
>>>>         
>>> thanks a lot for all your time and energy you did spend on this.
>>>
>>> I suggest we start collecting photographs of different LNA circuits on
>>> the wiki.
>>>
>>> For now, Tom offered his support already off list, I think we should
>>> start about the question, if that early Hauppauge HVR 1110 has such an
>>> LNA type one at all, since this caused to not look at it further, as it
>>> seemed to be without problems.
>>>
>>> Tom, I know you carefully worked on it, but can you reassure that this
>>> LNA config one is really needed on your device?
>>>
>>>   
>>>       
>> Hello list,
>> you are talking about tuner_config = 1 for the hvr 1110, right ?
>> Changing this option doesn't affect the qualitie of the signal on tv
>> see http://forum.ubuntu-fr.org/viewtopic.php?pid=1472261
>> it 's an "old" discussion in french.
>>     
>
> hmm, strange, since we got a different report too.
>
>   
>> This option, as far as i remenber, was not provided by me ...
>>     
>
> Thomas, thanks for reply. Looking closer now you are right.
> It came not in over you, but Mike picked it up from Benoit.
>
> hg export 6746
> # HG changeset patch
> # User Michael Krufky <mkrufky@linuxtv.org>
> # Date 1197003604 18000
> # Node ID f43a48e8bf53bf6a83953e15631fd1f5049a34f0
> # Parent  dd0e14aaa659375bfa4356d85ff7bf6308b9baa8
> saa7134-dvb: fix tuning for WinTV HVR-1110
>
> From: Benoit Istin <beistin@gmail.com>
>
> There are several months my hvr1110 stop working.
> This is very simple to fix, for my card revision at least, by setting a
> missing field to the hauppauge_hvr_1110_config.
>   
Hello

I see,
what i don't remember is, when searching for good parameters for this 
card (1110), AGC and Co was not necessary...

correct me if i'm wrong :

patch from Anders impacts cards with .tuner_config=1
what i can do :

step 1 :
see if we really need .tuner_config = 1  on  hvr_1110_config otherwise 
change to .tuner_config = 0

step 2 :
if needed, apply the patch from Anders and look if it's  better or not 
both on analogic and dvb

step 3 : report this results


others ideas ?

PS : i need times because my multimedia box is on production and i 
prefer test this on another pc, you know : why change when all is good ?

> Signed-off-by: Benoit Istin <beistin@gmail.com>
> Signed-off-by: Michael Krufky <mkrufky@linuxtv.org>
>
> diff -r dd0e14aaa659 -r f43a48e8bf53 linux/drivers/media/video/saa7134/saa7134-dvb.c
> --- a/linux/drivers/media/video/saa7134/saa7134-dvb.c   Thu Dec 06 22:33:08 2007 -0500
> +++ b/linux/drivers/media/video/saa7134/saa7134-dvb.c   Fri Dec 07 00:00:04 2007 -0500
> @@ -662,6 +662,7 @@
>         .if_freq       = TDA10046_FREQ_045,
>         .i2c_gate      = 0x4b,
>         .tuner_address = 0x61,
> +       .tuner_config  = 1,
>         .request_firmware = philips_tda1004x_request_firmware
>  };
>
> I must have noticed that it is still missing for analog then.
> http://linuxtv.org/hg/v4l-dvb/annotate/b227949c41ad/linux/drivers/media/video/saa7134/saa7134-cards.c
>
>   
>> anyway with tuner debug=1 and .tuner_config=1 , i have no line with AGC 
>> or LNA on dmesg
>>     
>
> Anders already gave the debug options he does set.
>
>   
yes an i have many lines with AGC on my logs
>> I have somme glitchs with hvr1110 on dvb (not analogic tv) and many for 
>> one particular station call M6 (and i'm not the only one user, see 
>> previous post on ubuntu-fr.org, with short or long distance from tv 
>> relay) . Bug on 310i means potentially bug on hvr1110 as configuration 
>> on hvr 1110 was made from 310i
>>
>> and sorry for bad english
>>     
>
> sorry for mine too.
>   
better than me :)
> On some cards LNA config 1 seems to make a difference and on others not
> or is it only because already broken? Maybe we have different card
> versions too, but you provided initial support for it and your report of
> course still counts.
>
> Under normal conditions you should expect that such cards are working
> without any glitches. If your reception overall is on some limit, it is
> somehow normal that one transponder starts to fail at first more
> visible, if you have some glitches on the others already.
>
> The sensitivity is not fully linear over the whole spectrum. I see this
> on other tuners too, but the flaw there starts on different frequencies.
>
> The antenna has of course an important role too and all sorts of RF
> amplifiers involved. The other case is also often reported from other
> cards already. It might happen, if close to the transmitter, that one
> might need an attenuator.
>
> Maybe try Anders' patch on recent.
>
> Cheers,
> Hermann
>
>
>
>   

Cheers,

Thomas