Re: [PATCH] hwmon: lm63: Add support for the NatSemi LM96163.

[Thierry Reding]

* Jean Delvare wrote:
> Hi Thierry,
> 
> On Tue, 8 Nov 2011 08:30:52 +0100, Thierry Reding wrote:
> > * Guenter Roeck wrote:
> > > Hi,
> > > 
> > > On Mon, 2011-11-07 at 05:20 -0500, Thierry Reding wrote:
> > > > Signed-off-by: Thierry Reding <thierry.reding@xxxxxxxxxxxxxxxxx>
> > > > ---
> > > >  drivers/hwmon/lm63.c |    2 +-
> > > >  1 files changed, 1 insertions(+), 1 deletions(-)
> > > > 
> > > > diff --git a/drivers/hwmon/lm63.c b/drivers/hwmon/lm63.c
> > > > index 508cb29..94e373c 100644
> > > > --- a/drivers/hwmon/lm63.c
> > > > +++ b/drivers/hwmon/lm63.c
> > > > @@ -483,7 +483,7 @@ static int lm63_detect(struct i2c_client *new_client,
> > > >  		return -ENODEV;
> > > >  	}
> > > >  
> > > > -	if (chip_id == 0x41 && address == 0x4c)
> > > > +	if ((chip_id == 0x41 || chip_id == 0x49) && address == 0x4c)
> > > >  		strlcpy(info->type, "lm63", I2C_NAME_SIZE);
> > > >  	else if (chip_id == 0x51 && (address == 0x18 || address == 0x4e))
> > > >  		strlcpy(info->type, "lm64", I2C_NAME_SIZE);
> > > 
> > > Minor - the new chip should be mentioned in drivers/hwmon/Kconfig and in
> > > Documentation/hwmon/lm63.
> > 
> > I don't see how to list this variant meaningfully in the documentation.
> > Since, as you mention, no functionality specific to the LM96163 is added,
> > this patch really only adds support for one more lm63 chip.
> 
> The new chip name should at least be listed in
> Documentation/hwmon/lm63, drivers/hwmon/Kconfig.

Okay.

> > > My main concern is that LM96163 supports a high resolution PWM mode.
> > > Since that mode is much more convenient than the LM63-compatible mode, I
> > > think it is likely to be enabled in designs using LM96163, and it should
> > > be supported by the driver if we add support for LM96163.
> > 
> > Unfortunately our design doesn't use the PWM at all. It is only used for
> > temperature measurements, so I'm not able to test any of the PWM code.
> 
> You still need to ensure that nothing bad will happen if users load the
> lm63 driver with a LM96163 chip (if nothing else, by disabling the
> unimplemented/untested feature.)

Okay. I'll make sure to mention that in the documentation as well.

Thanks,
Thierry

Attachment: pgp00000.pgp
Description: PGP signature