On Sun, Mar 04, 2012 at 04:50:52PM +0000, Arnd Bergmann wrote: > On Sunday 04 March 2012, Michael Walle wrote: > > > @@ -294,11 +295,17 @@ static int __exit mv_rtc_remove(struct > > > platform_device *pdev) return 0; > > > } > > > > > > +static struct of_device_id rtc_mv_of_match_table[] = { > > > + { .compatible = "marvell,rtc", }, > > > + {} > > > +}; > > > + > > Isn't here an ifdef CONFIG_OF missing to prevent compiler warnings? > > > > Yes, I guess so. I keep thinking that an __unused attribute would > actually be more helpful here because it lets the compiler automatically > discard the symbol when it's not used, but we still get potential > compiler errors and warnings if something inside of the unused data > is broken.
Hmmm, in the process of responding, I looked up Grant's original email [1] to show justification. However, rereading it shows I was overzealous in my desire to remove #ifdef's. :-P Grant only said adding of_match_ptr() would remove the #else portion of my original code: #ifdef CONFIG_OF ... #else #define rtc_mv_of_match_table = NULL; #endif I'll restore the #ifdef/#endif across the kirkwood code I've changed so far, without the #else. __unused shouldn't be necessary, then. Thanks for the review, Michael. thx, Jason. [1] http://www.spinics.net/lists/arm-kernel/msg162595.html _______________________________________________ devicetree-discuss mailing list devicetree-discuss@lists.ozlabs.org https://lists.ozlabs.org/listinfo/devicetree-discuss