On Mon, May 13, 2013 at 5:08 AM, Johannes Berg <johannes@xxxxxxxxxxxxxxxx> wrote: > On Sat, 2013-05-11 at 11:54 -0700, Luis R. Rodriguez wrote: >> * For kernels >= 2.6.31: >> o use the standard >> /dev/rfkill >> /sys/class/rfkill > > I think this is wrong. It means that if you build rfkill (e.g. for 3.2) > then it would conflict with the kernel's rfkill. It would replace it unless you built it obj-y. What issues are there with replacing in your example 3.2's rfkill? > The thing is that *usually*, for >=2.6.31, you just want to use the > kernel's rfkill. So for that usage, we should backport the rfkill API, > like we usually backport things. On second thought I can see this being true given that we don't backport WiMax, UWB, GPS (although I think this is serial and FM (not used in the kernel) but the API has remained consistent too so don't see this breaking those drivers ? > However, NFC probably requires a backported rfkill, not just backporting > around rfkill. Indeed. > For this we do need it named rfkill_backport though, I > think. Using a separate namespace is fine too but that would then require also changing the module name which I just saw now were only doing before it seems on compat-drivers, this change got lost on backports somehow. > Maybe we can backport the NFC support in some other way? Using a different module and namespace is possible, we however lost the Makefile changes even for older kernels. Luis -- To unsubscribe from this list: send the line "unsubscribe backports" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
