On Mon, Dec 09, 2013 at 10:59:38AM +0800, Chen-Yu Tsai wrote: > On Sat, Dec 7, 2013 at 7:06 PM, Tomasz Figa <tomasz.figa@xxxxxxxxx> wrote: > > On Saturday 07 of December 2013 11:07:37 maxime.ripard wrote: > >> On Sat, Dec 07, 2013 at 09:23:27AM +0800, Chen-Yu Tsai wrote: > >> > Point taken. The current implementation will override settings passed from > >> > platform data. ORing the two would be better. > >> > >> Platform_data and DT-based configuration are pretty unlikely to be > >> used together, so ORing it doesn't have much sense. > > > > In fact, the recommended way is to always use platform data alone if it is > > present or try to parse DT otherwise, so no mixing of data from these two > > sources should be done. > > Would binding platform data with compatibles, as I did so in this patch > series, be a bad idea then? What I meant was that you'll either be probed will pdev->dev.of_node or pdev->dev.platform_data filled, but not both at the same time, so ORing it isn't really sensible. But I don't see why you couldn't reuse the stmmac_platform_data structure in your patch. Maxime -- Maxime Ripard, Free Electrons Embedded Linux, Kernel and Android engineering http://free-electrons.com
Attachment:
signature.asc
Description: Digital signature