HI, On Tue, Aug 18, 2015 at 04:31:17PM +0300, Heikki Krogerus wrote: > On Mon, Aug 17, 2015 at 12:59:52PM -0500, Felipe Balbi wrote: > > On Mon, Aug 17, 2015 at 10:26:53AM -0700, David Cohen wrote: > > > On Mon, Aug 17, 2015 at 08:41:19AM -0500, Felipe Balbi wrote: > > > > On Mon, Aug 17, 2015 at 02:32:06PM +0300, Heikki Krogerus wrote: > > > > > Sharing the APCI companion with the core platforms device so > > > > ^^^^ > > > > ACPI :-) > > > > > > > > > it has access to all DWC3's resources. > > > > > > > > and which resources are these ? and why can't they be abstrated out in > > > > dwc3-pci ? > > > > > > I could think about USB phy devices on legacy BYT-CR products. But since > > > we got Heikki's ULPI bus we are able to move ACPI companion to their own > > > phy device on future. > > > > we need a better commit log explaining why this is really needed. > > OK. > > I'm also thinking about the properties with this. If we convert core.c > to use the unified device property interface (check the attached > diff), like we ultimately will need to do, the driver is then ready to > pick those properties also from ACPI tables. fair enough. > It does not actually have any functional affect on DT platforms. It > only makes it possible to read the property also from ACPI (and > later from "build-in" property), so I'll prepare a patch from that > diff and send it together with this one if you guys are OK with that? I'm fine with it. I would just turn device_read_property() into the default the pdata/DT the fallback. -- balbi
Attachment:
signature.asc
Description: Digital signature