On 22/04/2017 at 10:54:23 -0700, Moritz Fischer wrote: > On Thu, Apr 20, 2017 at 10:25 AM, Moritz Fischer <mdf@xxxxxxxxxx> wrote: > > On Thu, Apr 20, 2017 at 10:56:34AM -0500, Rob Herring wrote: > >> On Mon, Apr 17, 2017 at 03:40:10PM -0700, Moritz Fischer wrote: > >> > Introduce a device tree binding for specifying the trickle charger > >> > configuration for ds1374. This is based on the code for ds13390. > >> > > >> > Signed-off-by: Moritz Fischer <mdf@xxxxxxxxxx> > >> > --- > >> > .../devicetree/bindings/rtc/dallas,ds1374.txt | 18 ++++++++ > >> > drivers/rtc/rtc-ds1374.c | 54 ++++++++++++++++++++++ > >> > 2 files changed, 72 insertions(+) > >> > create mode 100644 Documentation/devicetree/bindings/rtc/dallas,ds1374.txt > >> > > >> > diff --git a/Documentation/devicetree/bindings/rtc/dallas,ds1374.txt b/Documentation/devicetree/bindings/rtc/dallas,ds1374.txt > >> > new file mode 100644 > >> > index 0000000..4cf5bd7 > >> > --- /dev/null > >> > +++ b/Documentation/devicetree/bindings/rtc/dallas,ds1374.txt > >> > @@ -0,0 +1,18 @@ > >> > +* Dallas DS1374 I2C Real-Time Clock / WDT > >> > >> Please remove from trivial-devices.txt, too. (which is moving in 4.12 > >> BTW) > > > > Ok, I'll redo this on top of b7e252fcddfa573bb1ee275b53bba6cef85671d4 > > (Documentation: devicetree: move trivial-devices out of I2C realm) then. > > Follow up question, right now one selects between WDT and ALARM mode with > a CONFIG_RTC_DRV_DS1374_WDT=y statically at compile time. > > I'd like to add a 'dallas,mode = <DS1374_WDT>;' or > dallas,enable-watchdog property > to the binding, same goes for the ability to remap the WDT reset output to the > interrupt pin (which is currently not supported, but my hardware needs this). > > Would be the right way to add the remapping something like > 'dallas,remap-reset-to-int' ? > > Ideas? This change would obviously break people's setups where they > select one or > the other behavior via the build time option. Is that acceptable seen > that relying on > build time CONFIG_FOO seems like a bad assumption to begin with? > > A bit of background: I currently started cleaning up a bunch of issues > in this driver, > like refactoring it to use the watchdog framework instead of open > coding everything, > make setting the timeout actually work (right now it the timeout to > tick conversion is > hosed). > I think I would do something with MFD as you were suggesting on IRC. You can have a look at the flexcom driver (atmel-flexcom.c) which is simply selecting a mode and then using whatever IP driver already existed (UART, I2C or SPI). I didn't have a close look but maybe that fits. Then, as you are concerned wtih backward compatibility, you could enforce the mode from the MFD driver, depending on the CONFIG_RTC_DRV_DS1374_WDT value. -- Alexandre Belloni, Free Electrons Embedded Linux and Kernel engineering http://free-electrons.com -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
