Handling interrupt routing cleanly...

[Eric W. Biederman]

Tarl Neustaedter <Tarl.Neustaedter at sun.com> writes:

> Eric W. Biederman wrote:
> > The only interesting thing I have is that in when using
> > APICs interrupts are routed differently than when using
> > the legacy x86 PIC.
> 
> Ah. That I missed. I'd never thought of interrupts being
> routed differently depending on something else. Assuming
> Linux is only going to use APICs, I wouldn't bother
> describing the legacy PIC, unless I'm missing something
> about interrupt configuration requiring that information
> anyway.

The APICs are not always present, and UP kernels tend not to use
them.  For the Opteron in 64bit mode you could probably not
worry about it though.

It has come up too many times in the past that I need to support
both that ignoring the legacy PIC in the design is just silly.

We currently can describe both the tables are just a hack and
don't adjust when things like pci bus numbers change, or
hypertransport unitids change.

Eric