Re: PowerMac11,2 Dual Core G5 panic trap on -current

[Martin Pieuchot]

On 17/08/15(Mon) 11:11, Martin Pieuchot wrote:
> On 16/08/15(Sun) 17:37, Donovan Watteau wrote:
> > On Sat, 15 Aug 2015, Donovan Watteau wrote:
> > > Hi,
> > >
> > > I saw some new PowerMac11,2 support in plus58.html. It mentions a
> > > Quad G5, while I have a Dual Core G5 (and an NVIDIA...), so here's
> > > what happens.
> > >
> > > I'm booting cd58.iso, with "boot cd:,ofwboot /5.8/macppc/bsd.rd" on
> > > OpenFirmware.
> > >
> > > However, it panics quickly, and the screen gets a bit corrupted
> > > (I have pictures if it helps, this one is copied by hand).
> > >
> > > Copyright (c) 1982, 1986, 1989, 1991, 19[corrupted]
> > > The Regents of the University of[corrupted]
> > > Copyright (c) 1995-2015 OpenBSD. All rig[corrupted]org
> > >
> > > OpenBSD 5.8 (RAMDISK) #644: Thu Aug 6 1[corrupted]
> > > der...@macppc.openbsd.org:/usr/src/[corrupted]mpile/RAMDISK
> > > real mem = 3758096384 (3584MB)
> > > avail mem = 2019610624 (1926MB)
> > > mainbus0 at root: model PowerMac11,2
> > > cpu0 at mainbus0: 970MP (Revision 0x101)[corrupted] MHz
> > > cpu at mainbus0 not configured
> > > mem at mainbus0 not configured
> > > memc0 at mainbus0: u4 rev 0x42
> > > kiic0 at memc0 offset 0xf8001000
> > > iic0 at kiic0
> > > "ds1775" at iic0 addr 0x4a not configured
> > > [other corrupted line, but has similarities with the previous line]
> > > "max6690" at iic0 addr 0x4c not configured
> > > "max6690" at iic0 addr 0x4e not configured
> > > "dart" at memc0 offset 0xf8033000 not configured
> > > openpic0 at memc0 offset 0xf8040000: version 0x4614 feature 7b0302 BE
> > > mpcpcibr0 at mainbus0 pci: u4-pcie
> > > pci0 at mpcpcibr0 bus 0
> > > vgafb0 at pci0 dev 0 function 0 "NVIDIA GeForce 6600" rev 0xa2, mmio
> > > vgafb0: aperture needed
> > > wsdisplay0 at vgafb0 mux 1: console (std, vt100 emulation)
> > > ht0 at mainbus0: u3-ht, 10 devices
> > > pci1 at ht0 bus0
> > > ppb0 at pci1 dev 1 function 0 "ServerWorks HT-2000 PCIX" rev 0xa3
> > > pci2 at ppb0 bus 4
> > > ppb1 at pci1 dev 2 function 0 "ServerWorks HT-2000 PCIX" rev 0xa3
> > > pci3 at ppb1 bus 5
> > > bge0 at pci3 dev 4 function 0 "Broadcom BCM5780" rev 0x03, BCM5714 B3
> > > (0x8003)panic: trap type 400 at 7000000 (0+0x7000000) lr 333dd8
> > > halted
> > > ok
> > > 0>
> > >
> > > Then it goes back to OpenFirmware and I can't type anything.
> > >
> > > I have another macppc I can use to compile kernels, if that helps.
> > >
> > > ...
> >
> > I did some more testing, and it looks like this issue only shows
> > up with bsd.rd. bsd.mp works fine once the system is installed.
> > I had to disable bge(4) with "boot -c" in order to work around
> > this.
>
> Good to know, I'll see if I can reproduce and investigate.

I just committed a fix to ht(4) that seems to be related to the problem
you're seeing, I'd be interested to know if you still see it with the
r1.18 of /sys/arch/macppc/pci/ht.c.

[Donovan Watteau]

I've built an updated kernel (with r1.18 of /sys/arch/macppc/pci/ht.c),
but I'm afraid I'm still having the same issue.

The previous panic always happens on cold boots and on all subsequent
reboots, unless I reboot once with "boot -c" and "disable bge" in UKC.
Once I've booted with these settings, all the following reboots work.

That is, I always have to boot this way to avoid the panic:

OF: boot hd:,ofwboot
OF: boot -c
UKC> disable bge
UKC> quit
# reboot
OF: boot hd:,ofwboot

ddb(4) gives the usual "send ps and trace output", but unfortunately
I can't type anything there.