On Tue, Oct 26, 2010 at 04:14:24PM -0500, Scott Wood wrote:
> On Tue, 26 Oct 2010 22:34:51 +0200
> Gleb Natapov <g...@redhat.com> wrote:
>
> > On Tue, Oct 26, 2010 at 07:57:00PM +0000, Blue Swirl wrote:
> > > On Tue, Oct 26, 2010 at 7:35 PM, Gleb Natapov <g...@redhat.com> wrote:
> > > > But looking elsewhere I found some description of DTS. It is very
> > > > elaborate and looks like this:
> > > >
> > > > /p...@xxx {
> > > > plenty of info here
> > > > }
> > > >
> > > > The only example of /p...@xxx that I found is here
> > > > http://wiki.freebsd.org/FlattenedDeviceTree but not any spec about its
> > > > format.
> > >
> > > That's FDT, it's a bit different.
> > >
> > > There are some trees here:
> > > http://penguinppc.org/historical/dev-trees-html/trees-index.html
> > >
> > > For example dual G4 500 has several /p...@xyz nodes.
> > >
> > Yes, it has: /p...@f0000000 for instance. Now lets try to decipher
> > address f0000000 according to pci2_1.pdf below. It says:
> > The text representation of a PCI address is one of the following forms:
> > DD
> > DD,F
> > [n]i[t]DD,F,RR,NNNNNNNN
> > [n]m[t][p]DD,F,RR,NNNNNNNN
> > [n]x[p]DD,F,RR,NNNNNNNNNNNNNNNN
> > where:
> > DD is an ASCII hexadecimal number in the
> > range 0...1F
> > F is an ASCII numeral in the range 0...7
> > RR is an ASCII hexadecimal number in the
> > range 0...FF
> > NNNNNNNN is an ASCII hexadecimal number in the
> > range 0...FFFFFFFF
> > NNNNNNNNNNNNNNNN is an ASCII hexadecimal number in the
> > range 0...FFFFFFFFFFFFFFFF
> > [n] is the letter 'n', whose presence is
> > optional
> > [t] is the letter 't', whose presence is
> > optional
> > [p] is the letter 'p', whose presence is
> > optional
> > i is the letter 'i'
> > m is the letter 'm'
> > x is the letter 'x'
> > , is the character ',' (comma)
> >
> > Nothing resembles f0000000. There is also 2.2.1.1 Numerical Representation
> > but no luck there too. This number is illegal according to it.
>
> The encoding above applies to unit addresses inside the PCI bus node.
>
> The unit address of the PCI controller node itself is encoded according
> to its parent bus.
>
Ah so to talk to pci controller OS writes to 0xf0000000?
http://penguinppc.org/historical/dev-trees-html/g4_agp_500_1.html has 3
such pci definitions:
/p...@f0000000
/p...@f2000000
/p...@f4000000
If G4 500 has 3 PCI domains it all makes perfect sense then.
> > May be this is memory address PCI bar is mapped into?
>
> It should correspond to the node's reg property.
>
> What's in the reg property depends on the binding for that particular
> PCI controller.
>
> -Scott
--
Gleb.
