>> Think of PCI super-IO devices, for example.
>
> BTW have you ever seen such a chip? I never had...
On PowerMacs, for example, Apple's mac-io family of controllers are big
PCI devices containing at least an interrupt controller, a bunch of DMA
channels, a floppy controller, a VIA, a SCSI controller, a pair of SCCs
and an ethernet controller ;-)
Some of those pieces can be shared between different versions of the
chip, are handled by different drivers, etc... They are almost completely
independant.
This case is correctly handled by the hierarchical nature of Open
Firmware's device tree, and this requires the ability to use sub-regions
made from the original PCI region (in this case, one big i/o mapped
region in the PCI config space of the chip).
Note that it may be interesting to setup a way to automatically build all
the region structures from the OF device tree on PowerMacs and CHRP
machines, and then fixup the result with missing PCI cards (I heard that
some rare cards won't be visible to OF at all, which looks stranges and
contradicts OF theorical behaviour, but...)
Note that this may already be done, I didn't check vger 2.3 for a while...
--
E-Mail: <mailto:bh40@calva.net>
BenH. Web : <http://calvaweb.calvacom.fr/bh40/>
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/