[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <1493119428.24567.190.camel@linux.intel.com>
Date: Tue, 25 Apr 2017 14:23:48 +0300
From: Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
To: Jan Kiszka <jan.kiszka@...mens.com>,
Peter Meerwald-Stadler <pmeerw@...erw.net>,
Mika Westerberg <mika.westerberg@...ux.intel.com>
Cc: Jonathan Cameron <jic23@...nel.org>, linux-iio@...r.kernel.org,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Sascha Weisenberger <sascha.weisenberger@...mens.com>
Subject: Re: [PATCH] iio: adc: Add support for TI ADC1x8s102
On Tue, 2017-04-25 at 11:32 +0200, Jan Kiszka wrote:
> On 2017-04-25 09:31, Peter Meerwald-Stadler wrote:
+Cc: Mika.
> > I think board-specific stuff should not go into the driver -> DT?
>
> Unfortunately, we only have half-baked ACPI on those boards.
That's the main problem and still no excuse for uglifying code.
> > > +
> > > +#include <linux/platform_data/adc1x8s102.h>
> >
> > who needs platform data these days :)
>
> Apparently, quite a few devices.
New drivers are not supposed to use platform data.
> But the situation here is:
> - Existing hardware has incomplete ACPI description that needs to be
> augmented by software.
Yes, and this software is called DSDT table in BIOS. Linux kernel has a
support for properly formed table already for few releases.
> - I'm not aware of a complete DT specification for this device. If
> there is any somewhere, I can happily include it.
Looking to proposed code there is only one property for it, if there is
an existing binding for the same property you may just simple re-use it.
> > > +
> > > +#define ADC1x8S102_GALILEO2_CS 8
> >
> > this board-specific detail shouldn't be here
>
> Where should it go then?
Obviously in (properly formed) ACPI.
> no board specific stuff here please
>
> Needed to make it work. If there is a better file to keep that, I'll
> move it.
Ideally you need BIOS fixed for that.
Otherwise you may do a separate code which would provide CS GPIO look up
table.
Mika, what do you think about fixing this in the C code for existing
devices?
--
Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
Intel Finland Oy
Powered by blists - more mailing lists