Re: [RFC PATCH v2 3/4] hwmon: adc128d818: Trivial code style fixup

[Guenter Roeck]

On 12/29/2016 12:30 PM, Alexander Koch wrote:
> On 12/29/2016 08:46 PM, Guenter Roeck wrote:
> > On Thu, Dec 29, 2016 at 07:22:12PM +0100, Alexander Koch wrote:
> > > On 12/26/2016 11:47 AM, Guenter Roeck wrote:
> > > > On 12/23/2016 02:12 PM, Alexander Koch wrote:
> > > > > Replace sysfs symbolic file permissions, e.g. 'S_IRUGO', by octal
> > > > > permissions. This fixes checkpatch.pl warnings.
> > > > >
> > > > > Signed-off-by: Alexander Koch <mail@xxxxxxxxxxxxxxxxx>
> > > > Please do not bother with those warnings and ignore checkpatch.
> > > > We are in the process of doing an automated conversion.
> > > Okay, so I shall create v3 of the patchset, without these changes. I've
> > > found a typo in the 4th patch so I thought about going v3 anyways.
> > >
> > > I've tested operation modes 1-3 on real hardware today and found no
> > > issues so far, so I hope v3 will have good chances of getting accepted.
> > Pretty much. One request, though: If there is no configuration data
> > from DT, I would like the driver to read the mode from the chip - if for
> > nothing else, this will let me test all modes, but it also supports the
> > case where the chip is configured by BIOS/ROMMON.
>
> Aye, will add this as new fourth patch then. I assume the chip reset in
> adc128_init_client() clears the chip mode as well, so I will read it in
> the operation mode block in adc128_probe().
>
> Just out of interest: how does this help you test the modes? Do you
> configure the chip externally and test it on a platform without
> devicetree support?

I use the i2c test driver in the kernel and a module test script, on x86.
See https://github.com/groeck/module-tests. That only works if the driver
does not overwrite the register values on probe.

Not really sure if resetting the chip during probe is such a good idea
in the first place. It is quite unusual.

Guenter