On Mon, Oct 22, 2012 at 10:13:07AM -0400, Alan Stern wrote: > On Mon, 22 Oct 2012, Tony Prisk wrote: > > > I suspected this might be a problem when I first looked at the ehci > > code, and in working on the ehci-platform driver recently it has become > > apparent this is actually going to be a problem in the near future. > > > > Because of the way ehci-hcd includes the source for variations: > > ... > > #ifdef CONFIG_ARCH_VT8500 > > #include "ehci-vt8500.c" > > #define PLATFORM_DRIVER vt8500_ehci_driver > > #endif > > > > #ifdef CONFIG_PLAT_SPEAR > > #include "ehci-spear.c" > > #define PLATFORM_DRIVER spear_ehci_hcd_driver > > #endif > > ... > > > > You can't have multiple drivers selected without getting a > > PLATFORM_DRIVER redefined warning/error. > > > > With ARM moving to multiplatform, there is going to be a need for a > > kernel config to select all EHCI drivers that are needed. > > > > Does Florian's (or another) patchset address this problem?? > > Florian's patches will help, because they decrease the number of > platform drivers to consider. But they don't solve the problem. > > I have been working on a separate set of changes that _will_ fix the > problem. I'm waiting for Florian's patches to be accepted before I > send mine in. > > Alan Stern > > P.S.: The same problem applies to ohci-hcd and xhci-hcd (and even > uhci-hcd, if any ARM-based systems support it).
xhci-hcd has a single xhci-plat.c and I plan to nag as much as I can if anyone tries to add another xhci-$arm-platform-here.c :-) -- balbi
signature.asc
Description: Digital signature
