[linux-dvb] make dvb-pll probe for device presence

[hermann pitton]

Am Dienstag, den 20.06.2006, 23:39 +0200 schrieb Johannes Stezenbach:
> On Tue, Jun 20, 2006, Andrew de Quincey wrote:
> > Hi - heres that idea we talked about a week or so back - it makes dvb-pll 
> > check for an i2c device's presence before going any further.
> > 
> > wdyt? (this isn't a proper committable patch yet BTW; no signed off by etc)
> 
> > diff -r 676364f7622d linux/drivers/media/dvb/frontends/dvb-pll.c
> > --- a/linux/drivers/media/dvb/frontends/dvb-pll.c	Tue Jun 20 00:30:57 2006 -0300
> > +++ b/linux/drivers/media/dvb/frontends/dvb-pll.c	Tue Jun 20 09:28:48 2006 +0100
> > @@ -615,6 +615,20 @@ int dvb_pll_attach(struct dvb_frontend *
> >  {
> >  	struct dvb_pll_priv *priv = NULL;
> >  
> > +	if (fe->ops.i2c_gate_ctrl)
> > +		fe->ops.i2c_gate_ctrl(fe, 1);
> > +	if (i2c_smbus_xfer(i2c, pll_addr, 0, 0, 0, I2C_SMBUS_QUICK, NULL) < 0)
> > +		return -1;
> 
> Don't use smbus functions on i2c bus. (And if you do anyway,
> use I2C_SMBUS_WRITE/I2C_SMBUS_READ for readability).
> 
> Some i2c controllers (like flexcop) don't support 0 byte messages.
> 
> > +	/* prevent 24RF08 corruption */
> 
> That's what you get for doing random i2c bus writes...
> (But I guess no dvb/i2c card has such an EEPROM, AFAIK they are
> only used in IBM Thinkpads on the system monitoring smbus.)

Johannes, Andrew,

now at least I understand a little what was going on during the
FE_REFACTORING and dvb-pll once.

Thanks,
Hermann