Andrew Gallatin writes:
>
> When I installed darwinx86-801 on my ServerWorks based P4 server, I
> got messages like this:
>
> Jun 22 13:07:26 scream kernel[0]: i82557eeprom: checksum 7f67 incorrect
> Jun 22 13:07:26 scream kernel[0]: Intel82557: couldn't allocate eeprom objecti82557eeprom: checksum bac4 incorrect
> Jun 22 13:07:26 scream kernel[0]: Intel82557: couldn't allocate eeprom objecti82557eeprom: checksum ee19 incorrect
>
> I thought "no big deal", and installed a 3com 3c905 and was happy.
>
Actually, it gets worse. Darwin seems to be destroying my 3c905 as
well, And I don't have a card maching this one to copy the eeprom from.
Damn..
Apple3Com3C90xB: 3Com EtherLink 3C905B-TX Regs 0xfc5fcc00 IRQ 29
Apple3Com3C90xB: WARNING: EEPROM checksum mismatch
Apple3Com3C90xB: WARNING: EEPROM checksum mismatch
Apple3Com3C90xB: 3Com EtherLink 3C905B-TX Regs 0xfc5fcc00 IRQ 29
Apple3Com3C90xB: 3Com EtherLink 3C905B-TX Regs 0xfc5fcc00 IRQ 29
Apple3Com3C90xB: Ethernet address 25:01:00:00:00:00
Apple3Com3C90xB: Ethernet address 00:50:04:4b:25:01
Apple3Com3C90xB: Ethernet address 00:c9:00:50:04:4b
The really interesting thing is that it seems to see 2 "phantom" cards:
# ioreg | grep 3C
| | | +-o Apple3Com3C90xB <class Apple3Com3C90xB, !registered, !matched, active, busy 0, retain count 8>
| | | +-o Apple3Com3C90xB <class Apple3Com3C90xB, !registered, !matched, active, busy 0, retain count 8>
| | | +-o Apple3Com3C90xB <class Apple3Com3C90xB, !registered, !matched, active, busy 0, retain count 8>
I also note that my company's rev3 card (pci14c1,8043@3) is repeated 3 times, as is the display controller,
but our rev4 and rev5 cards and the GbE card are missing.
Linux and FreeBSD show the PCI bus as looking like this:
0000:00:00.0 Host bridge: ServerWorks GCNB-LE Host Bridge (rev 32)
0000:00:00.1 Host bridge: ServerWorks GCNB-LE Host Bridge
0000:00:03.0 Non-VGA unclassified device: MYRICOM Inc. Myrinet 2000 Scalable Cluster Interconnect (rev 03)
0000:00:05.0 Ethernet controller: 3Com Corporation 3c905B 100BaseTX [Cyclone] (rev 64)
0000:00:07.0 VGA compatible controller: ATI Technologies Inc Rage XL (rev 27)
0000:00:08.0 Ethernet controller: Intel Corp. 82557/8/9 [Ethernet Pro 100] (rev 10)
0000:00:0f.0 ISA bridge: ServerWorks CSB5 South Bridge (rev 93)
0000:00:0f.1 IDE interface: ServerWorks CSB5 IDE Controller (rev 93)
0000:00:0f.3 Host bridge: ServerWorks CSB5 LPC bridge
0000:00:10.0 Host bridge: ServerWorks CIOB-X2 PCI-X I/O Bridge (rev 03)
0000:00:10.2 Host bridge: ServerWorks CIOB-X2 PCI-X I/O Bridge (rev 03)
0000:01:03.0 Network controller: MYRICOM Inc. Myrinet 2000 Scalable Cluster Interconnect (rev 05)
0000:02:03.0 Network controller: MYRICOM Inc. Myrinet 2000 Scalable Cluster Interconnect (rev 04)
0000:02:07.0 Ethernet controller: Intel Corp. 82545EM Gigabit Ethernet Controller (Copper) (rev 01)
My conclusion is that there is something wrong with the PCI peer bus
walking code on this machine, and the root bus is being repeated 2
additional times, once for each PCI-X peer bus. This problem goes
away if I add platform=x86pc to the boot flags, but that seems to just
ignore the PCI-X buses, and the second HTT core of the CPU.
The whole point of using this machine was to be able to port some
drivers to x86 for the cards in the PCI-X slots. Is there any hope to
get Darwin to see the PCI-X buses? Maybe by loading corrected ACPI tables it likes better?
I've left full ioreg output at http://people.freebsd.org/~gallatin/ioreg.trinity.gz
in case anybody is interested. I've also left the asl, as dumped by FreeBSD, at
http://people.freebsd.org/~gallatin/trinity.asl
Drew
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Darwin-x86 mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/darwin-x86/email@hidden
This email sent to email@hidden
