On Thu, Aug 09, 2007 at 09:53:41PM +0200, Segher Boessenkool wrote:
> >>>> I haven't heard or thought of anything better either. Using
> >>>> "ranges"
> >>>> is conceptually wrong, even ignoring the technical problems that
> >>>> come
> >>>> with it.
> >>>
> >>> Why is "ranges" conceptually wrong?
> >>
> >> The flash partitions aren't separate devices sitting on a
> >> "flash bus", they are "sub-devices" of their parent.
> >
> > Well, yes, but nonetheless the partitions show up as part of the
> > overall physical address space. How do you encode that other than in
> > 'ranges'?
>
> All that address space shows up in the flash node already. To
> access the partitions you have to go via that "master" node
> anyway (some commands have to be sent to address 0 on the flash,
> or similar).
Hrm, I suppose. Although for read-only access that's not relevant.
> It is a very nice feature to not only be able to translate addresses
> "up" the device tree, but also "down".
I don't follow, sorry.
> Also, it mimics reality, just like a good OF citizen should:
> those partitions aren't actually devices at all, so they
> certainly shouldn't be assigned a part of the host address
> space.
Hrm. I'm not entirely convinced that the distinction between
"actually a device" and "not actually a device" is really as clear cut
as you imply.
> >>> To be honest this looks rather to me like another case where having
> >>> overlapping 'reg' and 'ranges' would actually make sense.
> >>
> >> It never makes sense. You should give the "master" device
> >> the full "reg" range it covers, and have it define its own
> >> address space; "sub-devices" can carve out their little hunk
> >> from that. You don't want more than one device owning the
> >> same address range in the same address space.
> >
> > Why not? After all, the physical address ranges of the flash
> > partitions really do overlap with that of the flash device as a whole.
>
> They don't overlap, a partition is a proper subset of the flash.
A proper subset is a form of overlapping (indeed cases of being a
proper subset form a proper subset of cases of overlapping, to be
gratuitously meta-set-theoretic).
> Which as usual is shown as "reg" in the child node and #a,#s in
> the parent node.
That in no way encodes that the child addresses are a subset of the
parent address space. Instead #a and #s establish a new, separate
address space for the children, and without 'ranges', there's no
information about any sort of connection, overlapping, proper-subset
or otherwise, with the parent address space.
--
David Gibson | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
http://www.ozlabs.org/~dgibson
_______________________________________________
Linuxppc-dev mailing list
Linuxppc-dev@ozlabs.org
https://ozlabs.org/mailman/listinfo/linuxppc-dev
