On Tue, Mar 10, 2009 at 2:29 PM, Anton Vorontsov <avoront...@ru.mvista.com> wrote: > On Tue, Mar 10, 2009 at 01:48:26PM -0600, Grant Likely wrote: >> Regardless, I >> think all the drivers should be using common code for obtaining the >> phy_device from the device tree. > > Not necessary `struct phy_device'. All we need is some common > routine for translating PHY's "mdio_node->full_name + phy id" to > phy_bus_id.
This only works if the network driver knows how the MDIO bus is named. The current code assumes that the MDIO bus name is the register address, but this is a driver implementation detail and MDIO bus drivers can deviate from this. What about MDIO busses that don't have a reg property? For example, fs_enet/mii-bitbang.c sets the bus name to "CPM2 Bitbanged MII". There is no one-size-fits-all way to figure out the phy bus id from the ethernet driver side of things. The sure and simple way to guarantee a match between the PHY device node and the phy_device is to use the PHY device_node pointer itself as the search key. I concede that my first attempt at this was overly complex, but I've reworked the code and I think it makes thinks considerably simpler. I'll post a new series tomorrow. I've got patches to make the ucc_eth, gianfar and fs_enet drivers to use the device_node method also. g. -- Grant Likely, B.Sc., P.Eng. Secret Lab Technologies Ltd. _______________________________________________ Linuxppc-dev mailing list Linuxppc-dev@ozlabs.org https://ozlabs.org/mailman/listinfo/linuxppc-dev