+Gregory On Mon, Jul 14, 2014 at 11:36 AM, Julius Werner <jwerner@xxxxxxxxxxxx> wrote: >> Nope - since CONFIG_USB_XHCI_MVEBU can be 'y' or 'm' we need the ifneq >> here (which matches against both) to ensure xhci-mvebu.o is built is >> part of xhci-plat-hcd.o. > > Oh... does it make sense to have it tristate at all, then? Looks like > was never really buildable as an independent module (and still won't > be after this patch), so I bet that was just a mistake when the > Kconfig for it was first written that should be fixed. Making it a bool wouldn't really be accurate either since it could still be built as part of the xhci-plat-hcd loadable module. Gregory, do you have a preference? I could also re-spin making xhci-mvebu and independently-loadable module. > (Ideally, xhci-plat.c should probably go away in favor of dwc3/host.c > and xhci-mvebu.c implementing the wrapper to xhci_init_driver() > directly, but that is a longer-term goal.) Well I don't think xhci-plat can ever go away (it's possible that there are XHCI platform devices that need no extra platform-specific setup), but yes, I agree that dwc3 and xhci-mvebu should be converted over to use xhci_init_driver(). -- To unsubscribe from this list: send the line "unsubscribe linux-tegra" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
