Port-ofppc archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: Is OpenPIC mandatory for an ofppc-machine?
On Sat, Sep 6, 2014 at 11:40 AM, Frank Wille <frank%phoenix.owl.de@localhost>
wrote:
> Artyom Tarasenko wrote:
>
>> experimenting with Open Firmware for QEMU PReP target.
>> [...]
>> And here it hangs. Saw in some boot logs openpic initialization here.
>> QEMU uses an i8259, can this be the reason?
>
> Does QEMU PReP really emulate OpenFirmware correctly? There must be an
> OFW-tree entry with a name or device_type of "interrupt-controller".
> NetBSD/ofppc works with the Pegasos 2, which is an OpenFirmware machine
> with i8259 PIC. So theoretically it should work.
That's a good news, thank you. Here is my device tree:
ok show-devs
/dropin-fs
/flash@fff00000
/cpus
/pci
/mmu
/memory@0
/aliases
/options
/openprom
/chosen
/packages
/cpus/PowerPC,603@0
/pci/vga@2
/pci/ide@b,1
/pci/isa@b
/pci/ide@b,1/disk
/pci/isa@b/nvram@i74
/pci/isa@b/rtc@i74
/pci/isa@b/gpio@i3e0
/pci/isa@b/power@i380
/pci/isa@b/8042@i60
/pci/isa@b/fdc@i3f0
/pci/isa@b/parallel@i378
/pci/isa@b/serial@i2f8
/pci/isa@b/serial@i3f8
/pci/isa@b/configuration@i15c
/pci/isa@b/timer@i40
/pci/isa@b/interrupt-controller@i20
ok dev /pci/isa@b/interrupt-controller@i20
ok .properties
built-in
model WINB,83C553
reg 00000001 00000020 00000002
00000001 000000a0 00000002
00000001 000004d0 00000002
compatible chrp,iic
pci10ad,105-interrupt-controller
pci8086,484-interrupt-controller
pnpPNP,0
device_type interrupt-controller
name interrupt-controller
#address-cells 00000000
#interrupt-cells 00000002
interrupts 00000000
00000000
interrupt-controller
ok
Do you see anything suspicious?
> OFW PIC detection is in src/sys/arch/powerpc/pci/pci_machdep_ofw.c
> genofw_find_ofpics().
Looks like the code (
http://fxr.watson.org/fxr/source/arch/powerpc/pci/pci_machdep_ofw.c?v=NETBSD#L142
)
is expecting just one "compatible" propery. Or do I read it wrong?
Otoh the i8259 is the default. It seems that it should work without
the "interrupt-controller" node at all. I tried removing this node
from the device tree, but I see no visible change in the behavior.
Btw, the boot process seems to be much slower than for NetBSD/prep on
the same machine: writting output looks like an original 300 bod
terminal. Is it normal?
Btw2, we are not emulating OpenFirmware. It is the original
OpenFirmware from FirmWorks which I slightly adjusted for qemu -M
prep.
--
Regards,
Artyom Tarasenko
SPARC and PPC PReP under qemu blog: http://tyom.blogspot.com/search/label/qemu
Home |
Main Index |
Thread Index |
Old Index