Re: Patch to support drivers compiled into the kernel

[Dan Williams <dcbw redhat com>]

On Tue, 2009-03-03 at 20:37 +0100, Kay Sievers wrote:
> On Tue, Mar 3, 2009 at 20:32, Bill C Riemers <briemers redhat com> wrote:
> > Bill C Riemers wrote:
> >> OK.  If I need to go and patch coLinux, I will.   But what are the
> >> correct values?   There is no PCI/USB/PCMCMIA device.
> 
> Yeah, right. This is not handled.
> 
> > > This is
> >> virtual.   Do I need to invent a dummy driver and label it as the parent?
> 
> > BTW.  Does anyone know how network drivers are handled in other UML
> > implementations?
> >
> > http://en.wikipedia.org/wiki/User-mode_Linux
> 
> They do the same, I guess.
> 
> > Do the other implementations define a dummy bus, or do they fail with
> > NetworkManager as well?
> 
> Xen provides a "xen" bus, which HAL explicitly knows about. All other
> implementations, which do not provide qemu-like emulated hardware,
> will not work with HAL/NetworkManager, I guess.

So how do these virtual devices present themselves?  They've got to have
*some* piece of code in the kernel that knows how to talk to them, and
that piece of code is a driver.  That driver needs to get exposed
somehow through HAL in info.linux.driver, no?

Dan