On Tue, Sep 15, 2020 at 10:07 AM Paul Cercueil <paul@xxxxxxxxxxxxxxx> wrote: > > > > Le mar. 15 sept. 2020 à 10:03, Rob Herring <robh@xxxxxxxxxx> a écrit : > > On Tue, Sep 15, 2020 at 4:07 AM Paul Cercueil <paul@xxxxxxxxxxxxxxx> > > wrote: > >> > >> Hi Rob, > >> > >> Le lun. 14 sept. 2020 à 16:12, Rob Herring <robh@xxxxxxxxxx> a > >> écrit : > >> > On Fri, Sep 04, 2020 at 03:11:52PM +0200, Paul Cercueil wrote: > >> >> CONFIG_OF is selected by CONFIG_MACH_INGENIC, therefore we don't > >> >> need to > >> >> handle the case where Device Tree is not supported. > >> > > >> > What about COMPILE_TEST? If not supported, why not? > >> > >> What about it? It will still compile fine with COMPILE_TEST. > > > > CONFIG_OF could be disabled in that case, so the above reasoning > > doesn't hold. > > > > CONFIG_OF can be disabled in that case, correct, but why should we > care? The driver will still compile fine. Indeed, because jz4780_i2c_of_matches isn't within a CONFIG_OF ifdef as is sometimes done and is when you need of_match_ptr(). IMO, the commit msg should have something like "The driver is only used with CONFIG_OF enabled, so of_match_ptr() is not necessary. jz4780_i2c_of_matches is always defined." Rob
