[lm-sensors] [PATCH v3] hwmon: twl4030: Driver for twl4030 madc module
guenter.roeck at ericsson.com
Thu Jan 6 22:55:03 CET 2011
On Thu, 2011-01-06 at 15:21 -0500, Mark Brown wrote:
> On Thu, Jan 06, 2011 at 07:04:30AM -0800, Guenter Roeck wrote:
> > On Thu, Jan 06, 2011 at 07:07:13AM -0500, Mark Brown wrote:
> > > Why? It's not like hwmon has an unreasonably large core or similar.
> > Because it creates an unnecessary dependency, and because it is not hwmon's
> > responsibility to provide infrastructure for other subsystems or drivers.
> hwmon isn't really doing anything, though. The *driver* is doing
> something but it doesn't really impact the core that much. Not that I'm
> particularly sold on putting the ADC core in here, but total NACK based
> on that alone seems rather harsh.
Possibly. However, I had suggested the following earlier (to the 1st
version of the patch):
> I commented on this a couple of times below - the driver mixes generic
> ADC reading functions with hwmon functionality. Generic ADC reading
> functionality should be moved into another driver, possibly to mfd.
Obviously that was ignored. Maybe someone is willing to listen this time
I won't let people break modularity just for convenience in a subsystem
I am responsible for. And forcing the hwmon subsystem, and with it a
specific hwmon driver, to exist just because the adc functionality it
provides is needed by some other (most likely unrelated) subsystem /
driver _does_ break modularity. Worse, it is completely unnecessary to
do so. Other twl4030 functionality was extracted into generic code.
twl-core.c, twl4030-codec.c, twl4030-irq.c, twl4030-power.c are all in
mfd. I fail to see the problem with mfd/twl4030-adc.c.
More information about the lm-sensors