Re: [PATCH -next] usb: add usb.h stubs for CONFIG_USB not enabled

[Greg Kroah-Hartman <gregkh@linuxfoundation.org>]

On Sat, May 19, 2012 at 08:48:52AM -0700, Randy Dunlap wrote:
> On 05/19/2012 07:30 AM, Alan Stern wrote:
> 
> > On Fri, 18 May 2012, Randy Dunlap wrote:
> > 
> >> From: Randy Dunlap <rdunlap@xenotime.net>
> >>
> >> Add stubs for some usb core functions when CONFIG_USB
> >> is not enabled.  This fixes these comedi build errors:
> >>
> >> ERROR: "usb_deregister" [drivers/staging/comedi/comedi.ko] undefined!
> >> ERROR: "usb_register_driver" [drivers/staging/comedi/comedi.ko] undefined!
> >>
> >> Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
> >> ---
> >> I am only adding the stubs that are needed for comedi to build.
> >> I expect that more stubs will be needed as more build errors are
> >> found.  It seems to me that linux/usb.h might need more (major)
> >> moving of lines to support more stubs in the future.
> > 
> > This doesn't seem like the right thing to do.  If a driver needs to 
> > register with the USB core then that driver should depend on 
> > CONFIG_USB.
> > 
> > On the other hand, if the driver is capable of being built without USB 
> > support then it should use the necessary preprocessor tests to protect 
> > calls to any USB routines.
> 
> 
> Yeah, my first patch (not on linux-usb list) to staging/comedi/drivers.c
> just surrounded the usb helper functions with IS_ENABLED(CONFIG_USB).
> I'm fine with that patch, but Greg said that he wanted stubs...

In looking at this a bit more, I think we should go back to the "fix up
comedi" patch, as it is less intrusive, and as we haven't needed these
stubs for the past 10+ years, we probably shouldn't add them now :)

So, do you happen to have that old patch around, or do you want me to
make those changes?

thanks,

greg k-h