On Sun, 08 Aug 2010 22:34:44 -0700 (PDT) David Miller <davem@xxxxxxxxxxxxx> wrote: > From: Andres Salomon <dilinger@xxxxxxxxxx> > Date: Mon, 9 Aug 2010 01:32:45 -0400 > > > On Sun, 8 Aug 2010 23:12:21 -0600 > > Grant Likely <grant.likely@xxxxxxxxxxxx> wrote: > > > >> If you have to explicitly cast these function pointers, then you're > >> doing it wrong. :-) Listen to and fix the compiler complaint > >> here. > >> > > > > Hm, can you please expand on that? The reason it's necessary to > > cast is because sparc's prom_* functions are using ints instead of > > phandles. I don't understand why casting is the wrong thing here. > > > > I could write some 1-line wrapper functions that simply call prom_* > > rather than casting, I suppose. > > There's no reason the Sparc interfaces can't take phandles since they > are just ints, and that would make the call signatures of these > functions compatible with the types used by the other architectures. You're saying to convert the sparc code to use phandles? Doing so has caused phandles to show up in asm/openprom.h (linux_dev_v2_funcs's v2_inst2pkg returns a phandle, and linux_nodeops use phandles). This is fine, except asm/openprom.h is an exported kernel header, and linux/of.h is not. I'm tempted to break the phandle/device_node/property stuff out into a separate header (that gets exported), and include that from linux/of.h and asm/openprom.h. -- To unsubscribe from this list: send the line "unsubscribe sparclinux" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
