[linux-dvb] Hauppauge WinTV regreession from 2.6.24 to 2.6.25

[hermann pitton]

Am Sonntag, den 27.04.2008, 23:18 +0200 schrieb Hartmut Hackmann:
> Hi, Hermann, Mauro
> 
> hermann pitton schrieb:
> > Hi,
> > 
> > Am Samstag, den 26.04.2008, 20:19 -0300 schrieb Mauro Carvalho Chehab:
> >> On Sun, 27 Apr 2008 00:10:21 +0200
> >> hermann pitton <hermann-pitton at arcor.de> wrote:
> >>> Cool stuff!
> >>>
> >>> Works immediately for all tuners again. Analog TV, radio and DVB-T on
> >>> that machine is tested.
> >>>
> >>> Reviewed-by: Hermann Pitton <hermann-pitton at arcor.de>
> >> Thanks. I'll add it to the patch.
> >>
> >>> Maybe Hartmut can help too, but I will test also on the triple stuff and
> >>> the FMD1216ME/I MK3 hybrid tomorrow.
> >> Thanks.
> >>
> >> It would be helpful if tda9887 conf could also be validated. I didn't touch at
> >> the logic, but I saw some weird things:
> >>
> >> For example, SAA7134_BOARD_PHILIPS_EUROPA defines this:
> >> 	.tda9887_conf   = TDA9887_PRESENT | TDA9887_PORT1_ACTIVE
> >>
> >> And SAA7134_BOARD_PHILIPS_SNAKE keep the default values.
> >>
> >> However, there's an autodetection code that changes from EUROPA to SNAKE,
> >> without cleaning tda9887_conf:
> >>
> >>         case SAA7134_BOARD_PHILIPS_EUROPA:
> >>                 if (dev->autodetected && (dev->eedata[0x41] == 0x1c)) {
> >>                         /* Reconfigure board as Snake reference design */
> >>                         dev->board = SAA7134_BOARD_PHILIPS_SNAKE;
> >>                         dev->tuner_type = saa7134_boards[dev->board].tuner_type;
> >>                         printk(KERN_INFO "%s: Reconfigured board as %s\n",
> >>                                 dev->name, saa7134_boards[dev->board].name);
> >>                         break;
> >>
> >> I'm not sure if .tda9887_conf is missing at SNAKE board entry, or if the above
> >> code should be doing, instead:
> >>
> >> 	dev->tda9887_conf = saa7134_boards[dev->board].tda9887_conf;
> >>
> >> If the right thing to do is to initialize SNAKE with the same tda9887
> >> parameters as EUROPE, the better would be to add the .tda9887_conf to SNAKE
> >> entry.
> >>
> >> Cheers,
> >> Mauro
> > 
> > Hartmut has the board and knows better, but it looks like it only has
> > DVB-S and external analog video inputs. There is TUNER_ABSENT set, no
> > analog tuner, no tda9887 and also no DVB-T, but it unfortunately shares
> > the subsystem with the Philips Europa.
> > 
> Hermann is right, SNAKE has no analog tuner. These boards indeed share the same PCI ID,
> This code fragment reads the tuner ID from the eeprom to find out which board is there.
> 
> > I notice some unwanted behavior when testing md7134 FMD1216ME hybrid
> > boards.
> >
> Aha! I modified my board that it no longer runs with the current driver. But i observed
> something similar
> 
> > Unchanged is that the tda9887 is not up for analog after boot.
> > Previously one did reload "tuner" just once and was done.
> > 
> <snip>
> Don't have the time today, but lets roll back history: Not absolutely sure but if
> i remember correcly, the initialization sequence can be critical with hybrid tuners /
> NIM modules. The tda9887 may only be visible on I2C after a certain bit in the MOPLL
> is set (in byte4?)
> 
> Best regards
>    Hartmut

Hi Hartmut,

I remember this exactly, getting older anyway, but you had no chance
coming in with that on a quickly changing target, finally all of us shot
down as a crowd of lamers, a cheap target for kernel masterminds,
claiming we don't had anything in the right place ...

Let them pay for it now :)

Cheers,
Hermann