"Ronald G. Minnich" rminnich@lanl.gov writes:
On 14 Feb 2003, Eric W. Biederman wrote:
So in the worst case you should be able to lie and put in a pseudo device, or two that happen to be called at the proper times.
pseudo-device per chip type?
Or a generic pseudo-device? or ...
In most cases I think just putting it in the table of devices so we find the device even if the default pci_scan does not report it, should be enough.
For the weird cases. I am content that I can put a dummy device at an appropriate place in the list of devices, and have it do something. That is not a case I plan on using.
If we can get the abstraction so it actually models what is going on. We are in much better shape. And that is what I intend to concentrate on.
So what I am looking at:
The current pci bus scan starts with a root pci bus device. And then finds the devices on that bus, and the recursively scans the sub busses.
And the following things are possible.
We can have multiple top level devices. We can have force a device onto a given pci bus.
I want to make this process device driven, and get the configuration to just decorate the device tree with information such as irq routing, and default device settings (like the baud rate).
Irq routing currently does interesting things if you plug in a pci card with a bridge chip on it, I believe all of our current tables have the potential to hand out the wrong information as the bus numbers change.
Eric