Open Firmware device tree gives no clue what device might decode at physical address 0x80000000 to 0x8008200 on a G4 New World Macintosh. The mmu has three adjacent Virtual=Real translations for that block. They are the only address translations reported between the top or physical dram at 20000000 and the start of the PCI bridges at f0000000. The cpu node has marked them unavailable.
I now update the question because I discover the KeyLargo southbridge chip accepts addresses in that range, but I still don't know what device is at that physical address.
I'm not a specialist in Macs, but the article powerbook G4 issues after 9.10 upgrade says:
dmesg output:
[ 35.304239] yenta_cardbus 0001:10:13.0: pcmcia: parent PCI bridge Memory window: 0x80000000 - 0xafffffff
So it might be the PCMCIA.
Why are there only 7 hex digits in that upper address you're interested in? Did you mean
0x80008200
or did you mean
0x80082000
?
If you meant the latter, that looks like a little over 512KB, which I believe was the size of the L2 cache in some models of the PPC 74xx (G4) processor, which could be a clue. See if you have an l2-cache
node somewhere in your Open Firmware device tree, and poke around there.
The unidentified physical addresses are resources accessed on or via the KeyLargo Mac/IO chip:
80000000 - 8007ffff gpio (includes internal hd and cdrom)
80080000 - 80080fff usb1
80081000 - 80081fff usb0
This may be verified at the Open Firmware prompt as follows (although it could differ in some details from one eMac to the next):
dev pci1 mem-addr-base .
dev usb1 hc-base .
dev usb0 hc-base .
User contributions licensed under CC BY-SA 3.0