Re: [PATCH] Input: sun4i-ts: add thermal driver dependency

[Dmitry Torokhov <dmitry.torokhov@gmail.com>]

On Thu, Jan 29, 2015 at 10:51:51AM -0800, Dmitry Torokhov wrote:
> On Thu, Jan 29, 2015 at 08:50:18AM +0100, Hans de Goede wrote:
> > Hi,
> > 
> > On 28-01-15 16:17, Arnd Bergmann wrote:
> > >The sun4i-ts driver has had a dependency on the thermal code
> > >since it was first merged,
> > 
> > The "since it was first merged bit" is not true, this is caused
> > by recent changes by ChenYu.
> > 
> > Other then that no objections from me against the proposed fix.
> > 
> > Regards,
> > 
> > Hans
> > 
> > 
> >  but this is not currently enforced
> > >in Kconfig, so in some randconfig builds we get
> > >
> > >drivers/built-in.o: In function `sun4i_ts_remove':
> > >:(.text+0x2376f4): undefined reference to `thermal_zone_of_sensor_unregister'
> > >drivers/built-in.o: In function `sun4i_ts_probe':
> > >:(.text+0x237a94): undefined reference to `thermal_zone_of_sensor_register'
> > >:(.text+0x237c00): undefined reference to `thermal_zone_of_sensor_unregister'
> > >
> > >We need the dependency on THERMAL in order to ensure that this
> > >driver becomes a loadable module if the thermal support itself
> > >is modular, while the dependency on THERMAL_OF is a runtime dependency
> > >and the driver will still build but not work if it is missing.
> > >
> > >Signed-off-by: Arnd Bergmann <arnd@arndb.de>
> > >Fixes: 6decea7c5438e2 ("Input: add driver for Allwinner sunxi SoC's rtp controller")
> > >
> > >diff --git a/drivers/input/touchscreen/Kconfig b/drivers/input/touchscreen/Kconfig
> > >index 58917525126e..e2447f0063b7 100644
> > >--- a/drivers/input/touchscreen/Kconfig
> > >+++ b/drivers/input/touchscreen/Kconfig
> > >@@ -943,6 +943,8 @@ config TOUCHSCREEN_SUN4I
> > >  	tristate "Allwinner sun4i resistive touchscreen controller support"
> > >  	depends on ARCH_SUNXI || COMPILE_TEST
> > >  	depends on HWMON
> > >+	depends on THERMAL
> > >+	depends on THERMAL_OF || COMPILE_TEST
> 
> THERMAL_OF does have proper stubs and I do not see why we need to force
> having TERMAL even if we don't do compile test. Should the dependency be:
> 
> 	depends on !THERMAL_OF || TERMAL

*ping*

Thanks.

-- 
Dmitry