Hi Greg, Arnd, On Fri, 10 May 2024 12:57:24 +0200 "Arnd Bergmann" <arnd@xxxxxxxx> wrote: > On Fri, May 10, 2024, at 12:54, Luca Ceresoli wrote: > > On Fri, 10 May 2024 12:24:06 +0200 "Arnd Bergmann" <arnd@xxxxxxxx> wrote: > >> On Fri, May 10, 2024, at 09:55, Greg Kroah-Hartman wrote: > >> > On Fri, May 10, 2024 at 09:10:41AM +0200, Luca Ceresoli wrote: > >> >> > >> >> +config GE_SUNH_CONNECTOR > >> >> + tristate "GE SUNH hotplug add-on connector" > >> >> + depends on OF > >> >> + select OF_OVERLAY > >> >> + select FW_LOADER > >> >> + select NVMEM > >> >> + select DRM_HOTPLUG_BRIDGE > >> > > >> > Can these be depends instead of select? 'select' causes dependencies > >> > that are hard, if not almost impossible, to detect at times why > >> > something is being enabled. > >> > >> I think FW_LOADER needs to be 'select' since it is normally > >> a hidden symbol and gets selected by its users, all the other > >> ones should be 'depends on'. > > > > I see, makes sense. > > > > And as you pointed that out, I realize perhaps DRM_HOTPLUG_BRIDGE could > > become a hidden symbol as it's not expected to be used alone. > > It's slightly easier to keep it as a visible symbol > with 'depends on' though, since otherwise you have to > add 'depends on' statments for anything that DRM_HOTPLUG_BRIDGE > in turn depends on, most notably DRM itself. I see, sure. Thanks both, changes applied locally. Luca -- Luca Ceresoli, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
