tda8290 regression fix

On 2012-09-19 18:45, Mauro Carvalho Chehab wrote:
> Em 19-09-2012 10:01, Anders Thomson escreveu:
> >  On 2012-09-18 11:53, Mauro Carvalho Chehab wrote:
> >>  Em 16-09-2012 05:48, Anders Thomson escreveu:
> >>  >   It doesn't make any difference though :-( I still have the layer of noise...
> >>
> >>  That's weird. Hmm... perhaps priv->cfg.config is being initialized
> >>  latter. Maybe you can then do, instead:
> >>
> >>                    return -EREMOTEIO;
> >>            }
> >>
> >>  +        priv->cfg.switch_addr = priv->i2c_props.addr;
> >>            if ((data == 0x83) || (data == 0x84)) {
> >>                    priv->ver |= TDA18271;
> >>                    tda829x_tda18271_config.config = priv->cfg.config;
> >>
> >>
> >  No dice:
> >   $ git diff | cat
> >  diff --git a/drivers/media/common/tuners/tda8290.c b/drivers/media/common/tuners/tda8290.c
> >  index 8c48521..16d7ff7 100644
> >  --- a/drivers/media/common/tuners/tda8290.c
> >  +++ b/drivers/media/common/tuners/tda8290.c
> >  @@ -627,6 +627,9 @@ static int tda829x_find_tuner(struct dvb_frontend *fe)
> >                  return -EREMOTEIO;
> >          }
> >
> >  +       tuner_info("ANDERS: old priv->cfg.switch_addr %x\n", priv->cfg.switch_addr);
> >  +       priv->cfg.switch_addr = priv->i2c_props.addr;
> >  +       tuner_info("ANDERS: new priv->cfg.switch_addr %x\n", priv->cfg.switch_addr);
> >          if ((data == 0x83) || (data == 0x84)) {
> >                  priv->ver |= TDA18271;
> >                  tda829x_tda18271_config.config = priv->cfg.config;
> >  @@ -640,7 +643,6 @@ static int tda829x_find_tuner(struct dvb_frontend *fe)
> >
> >                  dvb_attach(tda827x_attach, fe, priv->tda827x_addr,
> >                             priv->i2c_props.adap,&priv->cfg);
> >  -               priv->cfg.switch_addr = priv->i2c_props.addr;
> >          }
> >          if (fe->ops.tuner_ops.init)
> >                  fe->ops.tuner_ops.init(fe);
> >  anders@tv /usr/src/linux $ dmesg | grep ANDERS
> >  [    5.667022] tda829x 4-004b: ANDERS: old priv->cfg.switch_addr 0
> >  [    5.667025] tda829x 4-004b: ANDERS: new priv->cfg.switch_addr 4b
>
> switch_addr got properly filled here.
Well, it got filled with 4b, but that it not what is needed. Recall my 
own patch:
# cat /TV_CARD.diff
                 fe->ops.tuner_ops.init(fe);
It needs to be filled with 0xc2 / 2. I'm not sure where I got that 
expression from, but it is the sum of my efforts tracing code changes 
around 2.6.26.
> >
> >  Whereas to work, I need:
> >  anders@tv /usr/src/linux $ grep ANDERS /3.3.8-d.patched
> >  [    6.565254] tda829x 5-004b: ANDERS: setting switch_addr. was 0x00, new 0x4b
>
> What looks weird here is that the device number changed from 4 to 5.
I have a webcam permanently on USB, so it ends up as two v4l2 devices 
(with device enumeration changing between boots)
>
> Do you have more than one board on your machine?
>
> >  [    6.565265] tda829x 5-004b: ANDERS: new 0x61
>
>
> The 0x61 address should be filled already by the existing code, otherwise
> you wouldn't be able to switch from one channel to another one.
>
> If you're in doubt, you could add an extra printk at the initialization code,
> in order to see what's happening there.
Not sure I follow here. Which code should set the ox61 address? I'd be 
more than happy to add printks. Where? I recall getting lost in how this 
stuff uses the i2c code in the past.
--
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

tda8290 regression fix

Commit Message

Comments

Patch