The pending nvmem patches have been applied to mtd/next. Some time ago I also sent a patch with the mtd-mac-address-increment-byte upstream but it got rejected since dts should not contain this kind of binding.
The patch directly implements the increment and increment-byte binding to the of_get_mac_address. So now the question, to clean things and follow a more standard framework, wouldn't be better to replace the patch generic pending/681 with a more generic implementation and convert all the dts to the standard mac-address binding and nvmem implementation? With a quick search, there are LOTS of dts that use these special bindings and it would take a lot of time to convert all the dts. (think the best approach would be to convert the dts one target at time) To improve this, I would propose to modify the 681 patch to only add the mtd-mac-address binding and the reading of the mac address from mtd and take my patch that implement the increment and increment-byte function to the general of_get_mac_address. This way the hard part of defining the nvmem cell can be postponed and the mtd-mac-address-increment-byte and mtd-mac-address-increment can simply be renamed to the new more generic mac-address-increment-byte and mac-address-increment binding with no regression since they implement the same thing in the same way. _______________________________________________ openwrt-devel mailing list [email protected] https://lists.openwrt.org/mailman/listinfo/openwrt-devel
