On Fri, Jul 17, 2015 at 06:58:32PM +0900, Krzysztof Kozlowski wrote: > 2015-07-17 18:43 GMT+09:00 Mark Brown <[email protected]>:
> > The driver core apparently now does this and we shouldn't be doing it > > statically so we get all this code churn. I'm not sure I understand the > > win TBH. > If you are referring to my recent patchset it was about platform and > i2c drivers where appropriate register functions do it (e.g. > i2c_register_driver()). But regulator_register passes owner from > regulator_desc: It's not just your patch set, it's the latest thing in code churn. > rdev->owner = regulator_desc->owner; > and then calls device_register(). I think the owner for regulator > device won't be set. Oh, right yes. Picked the wrong bit of code when merging a code motion patch.
signature.asc
Description: Digital signature
