I wonder about the way this is split between the platform file and this "sysdev" file.
The platform file is so small (and similar to other platform files) that I wonder if we can move toward something aking to the 5200 simple platform, that is have one platform with a list of matches. Now, the sysdev file I'm not so happy about. sysdev isn't supposed to contain board code, it's supposed to contain drivers for low level system devices such as interrupt controllers, PCI bridges, etc... that are shared by multiple platforms. The code in there definitely looks like it should be done differently too. The IBM IIC driver should already have an of_platform_device created for it, so we can have here a generic driver for that chip that instanciates itself based on it. No need to create a platform device .. Or is the ibm iic host driver only in the form of a platform device nowadays ? That can be a problem as we end up with basically 2 devices for the same device node... Why wouldn't ibm iic be an of platform driver ? If it's for arch/ppc compatibility, we don't care.. You can then expose all i2c devices underneath it with generic code too too no ? I haven't followed the discussion about i2c bindings but it looks like you should be able to have some generic code that instanciate all i2c devices in the device tree based on nodes under a given controller... Ben. _______________________________________________ Linuxppc-dev mailing list Linuxppc-dev@ozlabs.org https://ozlabs.org/mailman/listinfo/linuxppc-dev
