NetBSD 4.0.1 on 7300 with G3 card

[Georg Schwarz <georg.schwarz%freenet.de@localhost>]

I've installed NetBSD 4.0.1 on an Apple PowerMac 7300 with Sonnet  
Crescendo 500 MHz G3 CPU card.
Installation via Open Firmware 1.0.5 using boot floppies was almost  
straightforward thanks to the excellent documentation on www.NetBSD.org.
The only serious issue I ran into was the bug described in
http://mail-index.netbsd.org/port-macppc/2008/12/01/msg000486.html

A few questions remain:

- the only media selection option for the mc ethernet adapter seems  
to be "manual". This works fine for me. I wonder however how one  
could select AAUI vs. 10baseT.
- why is there no DHCP support as with the i386 boot floppies?
- likewise, there does not seem to be support for keyboard layouts  
other than US English. Is this an Open Firmware issue? Manually  
selecting wscons (with the finished installation) seems to work fine.
- The Sonnect Crescendo CPU seems to work fine. Now I am wondering  
how to enable the L2 cache.

cpu0 at mainbus0: 750 (Revision 3.0), ID 0 (primary)
cpu0: HID0 8290c0a4<EMCP,ECLK,DOZE,DPM,ICE,DCE,SGE,BTIC,BHT>,  
powersave: 1
cpu0: 500.00 MHz L2 cache present but not enabled

I came across http://mail-index.netbsd.org/port-macppc/ 
2001/01/05/0000.html and http://mail-index.netbsd.org/port-macppc/ 
2001/01/05/0007.html
How do I figure out which L2RAM option is appropriate?

According to the sticker on the card I have a 500 MHz G3 with 1 MB  
L2. I assume the L2 is running at only half the 500 MHz, or is it  
really using the full 500 MHz?
I think it is this card:
http://www.sonnettech.com/product/crescendo_pci.html

There's another strange thing. I've added a UF200 PCI firewire/USB  
combo card. This is how it is detected:

fwohci0 at pci1 dev 0 function 0: Lucent Technologies FW322/323 IEEE  
1394 Host Controller (rev. 0x04)
fwohci0: interrupting at irq 25
fwohci0: OHCI version 1.0 (ROM=1)
fwohci0: No. of Isochronous channels is 8.
fwohci0: EUI64 00:60:1d:00:00:00:00:dd
fwohci0: Phy 1394a available S400, 3 ports.
fwohci0: Link S400, max_rec 2048 bytes.
ieee1394if0 at fwohci0: IEEE1394 bus
fwip0 at ieee1394if0: IP over IEEE1394
fwohci0: Initiate bus reset
ohci0 at pci1 dev 9 function 0: Lucent Technologies Sierra Wireless  
AirCard 580 (rev. 0x10)
ohci0: interrupting at irq 25
ohci0: OHCI version 1.0, legacy support
usb0 at ohci0: USB revision 1.0
uhub0 at usb0
uhub0: Lucent Technolo OHCI root hub, class 9/0, rev 1.00/1.00, addr 1
uhub0: 2 ports with 2 removable, self powered

why on earth is NetBSD talking about a Lucent Sierra Wireless mobile  
adapter??? Is this a PCI ID mismatch in the kernel?

In src/sys/dev/pci/pcidevs I find:

vendor LUCENT           0x11c1  Lucent Technologies

product LUCENT EVDO             0x5802  Sierra Wireless AirCard 580

However, http://www.pcidatabase.com/vendor_details.php?id=979 lists  
0x5802 as

Chip Number:    USS-312MC
Chip Description:       2-port PCI-to-USB OpenHCI Host Ctrlr


which makes much more sense.
Maybe someone might want to check other OS's kernel sources.
Is the Sierra Wireless AirCard 580 a PCI device?
It is also listed in dev/usb/usbdevs
I wonder whether this is correct or whether this is the same mistake.
According to http://shopping.msn.com/specs/sierra-wireless- 
aircard-580-network-adapter/itemid696159879/?itemtext=itemname:sierra- 
wireless-aircard-580-network-adapter
that card is a cardbus device. I don't see any PCI or USB stuff  
involved.
However, it seems to involve some USB stuff (comes with its own USB  
chip?) according to
http://www.linuxsir.org/bbs/archive/index.php/t-119912.html




--
Georg Schwarz    http://home.pages.de/~schwarz/
georg.schwarz%freenet.de@localhost  +49 151 11559652