IBM OpenPower 720 ipr driver woes

IBM OpenPower 720 ipr driver woes

[Robert Knight]

Sometime after kernel 3.3.4, Fedora 17 running on an IBM OpenPower 720 
stopped being able to read the SCSI disks with an error concerning a 
return code -1.  In the updates since, it has not resumed working.

Does the ipr driver no longer support this (admittedly ancient) hardware?

Re: IBM OpenPower 720 ipr driver woes

[Brian King]

On 05/29/2013 10:15 AM, Robert Knight wrote:
> Sometime after kernel 3.3.4, Fedora 17 running on an IBM OpenPower
> 720 stopped being able to read the SCSI disks with an error
> concerning a return code -1.  In the updates since, it has not
> resumed working.
> 
> Does the ipr driver no longer support this (admittedly ancient)
> hardware? 

Do you have a boot log you can share?

Thanks,

Brian

-- 
Brian King
Power Linux I/O
IBM Linux Technology Center

Re: IBM OpenPower 720 ipr driver woes

[Robert Knight]

On 5/29/2013 2:11 PM, Brian King wrote:
> On 05/29/2013 10:15 AM, Robert Knight wrote:
>> Sometime after kernel 3.3.4, Fedora 17 running on an IBM OpenPower
>> 720 stopped being able to read the SCSI disks with an error
>> concerning a return code -1.  In the updates since, it has not
>> resumed working.
>>
>> Does the ipr driver no longer support this (admittedly ancient)
>> hardware?
> Do you have a boot log you can share?
>
> Thanks,
>
> Brian
>
Apologies for the slowness of my response.  I had wiped the system, 
Fedora 17 won't install without grub2 tweaks and I had to discover that.

I have copied and pasted the HMC console window:

> Boo
> ting `Fedora (3.8.11-100.fc17.ppc64)'
>
> Loading Fedora (3.8.11-100.fc17.ppc64)
> Loading initial ramdisk ...
> OF stdout device is: /vdevice/vty@30000000
> Preparing to boot Linux version 3.8.11-100.fc17.ppc64 
> (mockbuild@ppc-builder7.qa
> .fedoraproject.org) (gcc version 4.7.2 20120921 (Red Hat 4.7.2-2) 
> (GCC) ) #1 SMP
>  Wed May 1 13:51:54 MST 2013
> Detected machine type: 0000000000000101
> Max number of cores passed to firmware: 512 (NR_CPUS = 1024)
> Calling ibm,client-architecture-support... not implemented
> command line: BOOT_IMAGE=/vmlinuz-3.8.11-100.fc17.ppc64 
> root=/dev/mapper/vg_narv
> i-lv_root ro rd.md=0 rd.dm=0 KEYTABLE=us rd.lvm.lv=vg_narvi/lv_swap 
> rd.luks=0 rd
> .lvm.lv=vg_narvi/lv_root SYSFONT=True LANG=en_US.UTF-8 rhgb quiet
> memory layout at init:
>   memory_limit : 0000000000000000 (16 MB aligned)
>   alloc_bottom : 0000000004800000
>   alloc_top    : 0000000008000000
>   alloc_top_hi : 00000000ee000000
>   rmo_top      : 0000000008000000
>   ram_top      : 00000000ee000000
> instantiating rtas at 0x00000000077b0000... done
> Querying for OPAL presence... not there.
> boot cpu hw idx 0
> starting cpu hw idx 2... done
> copying OF device tree...
> Building dt strings...
> Building dt structure...
> Device tree strings 0x0000000004810000 -> 0x000000000481123d
> Device tree struct  0x0000000004820000 -> 0x0000000004840000
> Calling quiesce...
> returning from prom_init
> [    1.376359] ehci-pci 0000:c8:01.2: can't setup
> [    1.376749] ehci-pci 0000:c8:01.2: init 0000:c8:01.2 fail, -19
> [   11.388273] ohci_hcd 0000:c8:01.0: USB HC takeover failed! 
> (BIOS/SMM bug)
> [   11.388287] ohci_hcd 0000:c8:01.0: can't setup
> [   11.388686] ohci_hcd 0000:c8:01.0: init 0000:c8:01.0 fail, -16
> [   21.408214] ohci_hcd 0000:c8:01.1: USB HC takeover failed! 
> (BIOS/SMM bug)
> [   21.408228] ohci_hcd 0000:c8:01.1: can't setup
> [   21.408596] ohci_hcd 0000:c8:01.1: init 0000:c8:01.1 fail, -16
> Cannot open font file True
> Cannot open font file True
> [   21.949512] ipr 0001:d0:01.0: Write of cache line size failed
> [   22.032546] pata_pdc2027x: Invalid PLL input clock 0kHz, give up!
>
>
> dracut Warning: Unable to process initqueue
> Cannot open font file True
>
>
> dracut Warning: Unable to process initqueue
> dracut Warning: /dev/mapper/vg_narvi-lv_root does not exist
> dracut Warning: /dev/vg_narvi/lv_root does not exist
> dracut Warning: /dev/vg_narvi/lv_swap does not exist
>
> Dropping to debug shell.
>
> dracut:/#

Re: IBM OpenPower 720 ipr driver woes

[Benjamin Herrenschmidt]

On Mon, 2013-06-03 at 17:20 -0400, Robert Knight wrote:
> On 5/29/2013 2:11 PM, Brian King wrote:
> > On 05/29/2013 10:15 AM, Robert Knight wrote:
> >> Sometime after kernel 3.3.4, Fedora 17 running on an IBM OpenPower
> >> 720 stopped being able to read the SCSI disks with an error
> >> concerning a return code -1.  In the updates since, it has not
> >> resumed working.
> >>
> >> Does the ipr driver no longer support this (admittedly ancient)
> >> hardware?
> > Do you have a boot log you can share?
> >
> > Thanks,
> >
> > Brian
> >
> Apologies for the slowness of my response.  I had wiped the system, 
> Fedora 17 won't install without grub2 tweaks and I had to discover that.
> 
> I have copied and pasted the HMC console window:

Nothing about IPR here but the USB messages seem to indicate that
something's busted in general with PCI (ie, bus going EEH or something).

Not sure what's up as we don't see a complete boot log. However I don't
think fc17 got much love, have you tried 18 ? We've been running it on
a number of machines reasonably successfully.

Cheers,
Ben.

> > Boo
> > ting `Fedora (3.8.11-100.fc17.ppc64)'
> >
> > Loading Fedora (3.8.11-100.fc17.ppc64)
> > Loading initial ramdisk ...
> > OF stdout device is: /vdevice/vty@30000000
> > Preparing to boot Linux version 3.8.11-100.fc17.ppc64 
> > (mockbuild@ppc-builder7.qa
> > .fedoraproject.org) (gcc version 4.7.2 20120921 (Red Hat 4.7.2-2) 
> > (GCC) ) #1 SMP
> >  Wed May 1 13:51:54 MST 2013
> > Detected machine type: 0000000000000101
> > Max number of cores passed to firmware: 512 (NR_CPUS = 1024)
> > Calling ibm,client-architecture-support... not implemented
> > command line: BOOT_IMAGE=/vmlinuz-3.8.11-100.fc17.ppc64 
> > root=/dev/mapper/vg_narv
> > i-lv_root ro rd.md=0 rd.dm=0 KEYTABLE=us rd.lvm.lv=vg_narvi/lv_swap 
> > rd.luks=0 rd
> > .lvm.lv=vg_narvi/lv_root SYSFONT=True LANG=en_US.UTF-8 rhgb quiet
> > memory layout at init:
> >   memory_limit : 0000000000000000 (16 MB aligned)
> >   alloc_bottom : 0000000004800000
> >   alloc_top    : 0000000008000000
> >   alloc_top_hi : 00000000ee000000
> >   rmo_top      : 0000000008000000
> >   ram_top      : 00000000ee000000
> > instantiating rtas at 0x00000000077b0000... done
> > Querying for OPAL presence... not there.
> > boot cpu hw idx 0
> > starting cpu hw idx 2... done
> > copying OF device tree...
> > Building dt strings...
> > Building dt structure...
> > Device tree strings 0x0000000004810000 -> 0x000000000481123d
> > Device tree struct  0x0000000004820000 -> 0x0000000004840000
> > Calling quiesce...
> > returning from prom_init
> > [    1.376359] ehci-pci 0000:c8:01.2: can't setup
> > [    1.376749] ehci-pci 0000:c8:01.2: init 0000:c8:01.2 fail, -19
> > [   11.388273] ohci_hcd 0000:c8:01.0: USB HC takeover failed! 
> > (BIOS/SMM bug)
> > [   11.388287] ohci_hcd 0000:c8:01.0: can't setup
> > [   11.388686] ohci_hcd 0000:c8:01.0: init 0000:c8:01.0 fail, -16
> > [   21.408214] ohci_hcd 0000:c8:01.1: USB HC takeover failed! 
> > (BIOS/SMM bug)
> > [   21.408228] ohci_hcd 0000:c8:01.1: can't setup
> > [   21.408596] ohci_hcd 0000:c8:01.1: init 0000:c8:01.1 fail, -16
> > Cannot open font file True
> > Cannot open font file True
> > [   21.949512] ipr 0001:d0:01.0: Write of cache line size failed
> > [   22.032546] pata_pdc2027x: Invalid PLL input clock 0kHz, give up!
> >
> >
> > dracut Warning: Unable to process initqueue
> > Cannot open font file True
> >
> >
> > dracut Warning: Unable to process initqueue
> > dracut Warning: /dev/mapper/vg_narvi-lv_root does not exist
> > dracut Warning: /dev/vg_narvi/lv_root does not exist
> > dracut Warning: /dev/vg_narvi/lv_swap does not exist
> >
> > Dropping to debug shell.
> >
> > dracut:/#
> 
> 
> _______________________________________________
> Linuxppc-dev mailing list
> Linuxppc-dev@lists.ozlabs.org
> https://lists.ozlabs.org/listinfo/linuxppc-dev

Re: IBM OpenPower 720 ipr driver woes

[Tony Breeds]

[-- Attachment #1: Type: text/plain, Size: 422 bytes --]

On Mon, Jun 03, 2013 at 05:20:12PM -0400, Robert Knight wrote:

> >Device tree struct  0x0000000004820000 -> 0x0000000004840000
> >Calling quiesce...
> >returning from prom_init
> >[    1.376359] ehci-pci 0000:c8:01.2: can't setup

Can you try adding "debug" to the kernel commandline.  We're missing a
great chunk of detail. If you're starting from scratch either try F18 or
the F19 Beta (if you're brave :))

Yours Tony

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]

Re: IBM OpenPower 720 ipr driver woes

[Robert Knight]

[-- Attachment #1: Type: text/plain, Size: 1064 bytes --]

On 6/3/2013 8:01 PM, Tony Breeds wrote:
> On Mon, Jun 03, 2013 at 05:20:12PM -0400, Robert Knight wrote:
>
>>> Device tree struct  0x0000000004820000 -> 0x0000000004840000
>>> Calling quiesce...
>>> returning from prom_init
>>> [    1.376359] ehci-pci 0000:c8:01.2: can't setup
> Can you try adding "debug" to the kernel commandline.  We're missing a
> great chunk of detail. If you're starting from scratch either try F18 or
> the F19 Beta (if you're brave :))
>
> Yours Tony
So, two points.  Since I have no live disks, I can't copy the dmesg 
output in dracut onto a disk and just sent it to you -- the only way 
that I know to get it is cut and paste from a HMC console window into a 
vi buffer.

I removed rhgb and quiet in addition to adding debug to the kernel 
command line in the hope of providing you with information.

My second point is that this failure to connect to the disks happens in 
F18 in pretty much the same way -- it always ends with "pata_pdc2027x 
... give up!".

Here we go again.

Thank you for being willing to try to help remotely.



[-- Attachment #2: boot.log2 --]
[-- Type: text/plain, Size: 14232 bytes --]

[    0.000000] Allocated 917504 bytes for 1024 pacas at c0000000076d0000
[    0.000000] Using pSeries machine description
[    0.000000] Page orders: linear mapping = 24, virtual = 12, io = 12, vmemmap
= 24
[    0.000000] Found initrd at 0xc000000003800000:0xc0000000047f9800
[    0.000000] bootconsole [udbg0] enabled
[    0.000000] Partition configured for 4 cpus.
[    0.000000] CPU maps initialized for 2 threads per core
[    0.000000]  (thread shift is 1)
[    0.000000] Freed 851968 bytes for unused pacas
[    0.000000] Starting Linux PPC64 #1 SMP Wed May 1 13:51:54 MST 2013
[    0.000000] -----------------------------------------------------
[    0.000000] ppc64_pft_size                = 0x1a
[    0.000000] physicalMemorySize            = 0xee000000
[    0.000000] htab_hash_mask                = 0x7ffff
[    0.000000] -----------------------------------------------------
[    0.000000] Initializing cgroup subsys cpuset
[    0.000000] Initializing cgroup subsys cpu
[    0.000000] Linux version 3.8.11-100.fc17.ppc64 (mockbuild@ppc-builder7.qa.fe
doraproject.org) (gcc version 4.7.2 20120921 (Red Hat 4.7.2-2) (GCC) ) #1 SMP We
d May 1 13:51:54 MST 2013
[    0.000000] [boot]0012 Setup Arch
[    0.000000] Node 0 Memory: 0x0-0xee000000
[    0.000000] PCI host bridge /pci@800000020000002  ranges:
[    0.000000]   IO 0x000003fe00200000..0x000003fe002fffff -> 0x0000000000000000
[    0.000000]  MEM 0x0000040080000000..0x00000400bfffffff -> 0x00000000c0000000

[    0.000000] PCI host bridge /pci@800000020000003  ranges:
[    0.000000]   IO 0x000003fe00700000..0x000003fe007fffff -> 0x0000000000000000
[    0.000000]  MEM 0x00000401c0000000..0x00000401ffffffff -> 0x00000000c0000000

[    0.000000] PPC64 nvram contains 7168 bytes
[    0.000000] Zone ranges:
[    0.000000]   DMA      [mem 0x00000000-0xedffffff]
[    0.000000]   Normal   empty
[    0.000000] Movable zone start for each node
[    0.000000] Early memory node ranges
[    0.000000]   node   0: [mem 0x00000000-0xedffffff]
[    0.000000] On node 0 totalpages: 60928
[    0.000000]   DMA zone: 53 pages used for memmap
[    0.000000]   DMA zone: 0 pages reserved
[    0.000000]   DMA zone: 60875 pages, LIFO batch:1
[    0.000000] [boot]0015 Setup Done
[    0.000000] PERCPU: Embedded 2 pages/cpu @c000000001800000 s87168 r0 d43904 u
262144
[    0.000000] pcpu-alloc: s87168 r0 d43904 u262144 alloc=1*1048576             [    0.000000] pcpu-alloc: [0] 0 1 2 3
[    0.000000] Built 1 zonelists in Node order, mobility grouping on.  Total pag
es: 60875
[    0.000000] Policy zone: DMA
[    0.000000] Kernel command line: BOOT_IMAGE=/vmlinuz-3.8.11-100.fc17.ppc64 ro
ot=/dev/mapper/vg_narvi-lv_root ro rd.md=0 rd.dm=0 KEYTABLE=us rd.lvm.lv=vg_narv
i/lv_swap rd.luks=0 rd.lvm.lv=vg_narvi/lv_root SYSFONT=True LANG=en_US.UTF-8 deb
ug
[    0.000000] PID hash table entries: 4096 (order: -1, 32768 bytes)
[    0.000000] freeing bootmem node 0
[    0.000000] Memory: 3766784k/3899392k available (18560k kernel code, 132608k
reserved, 1984k data, 2785k bss, 6912k init)
[    0.000000] SLUB: Genslabs=19, HWalign=128, Order=0-3, MinObjects=0, CPUs=4,
Nodes=256
[    0.000000] Hierarchical RCU implementation.
[    0.000000]  RCU restricting CPUs from NR_CPUS=1024 to nr_cpu_ids=4.
[    0.000000] NR_IRQS:512 nr_irqs:512 16
[    0.000000] pic: no ISA interrupt controller
[    0.000000] time_init: decrementer frequency = 188.043000 MHz
[    0.000000] time_init: processor frequency   = 1504.344000 MHz
[    0.000000] clocksource: timebase mult[5516408] shift[24] registered
[    0.000000] clockevent: decrementer mult[30239607] shift[32] cpu[0]
[    0.000000] Console: colour dummy device 80x25
                                                                                [    0.000000] console [hvc0] enabled, bootconsole disabled
[    0.000000] allocated 974848 bytes of page_cgroup
[    0.000000] please try 'cgroup_disable=memory' option if you don't want memor
y cgroups
[    0.010396] pid_max: default: 32768 minimum: 301
[    0.010532] Security Framework initialized
[    0.010555] SELinux:  Initializing.
[    0.010575] SELinux:  Starting in permissive mode
[    0.011123] Dentry cache hash table entries: 524288 (order: 6, 4194304 bytes)
[    0.014271] Inode-cache hash table entries: 262144 (order: 5, 2097152 bytes)
[    0.015829] Mount-cache hash table entries: 4096
[    0.025291] Initializing cgroup subsys cpuacct
[    0.025301] Initializing cgroup subsys memory
[    0.025488] Initializing cgroup subsys devices
[    0.025496] Initializing cgroup subsys freezer
[    0.025504] Initializing cgroup subsys net_cls
[    0.025513] Initializing cgroup subsys blkio
[    0.025520] Initializing cgroup subsys perf_event
[    0.025561] ftrace: allocating 22464 entries in 9 pages
[    0.042773] EEH: pSeries platform initialized
[    0.042787] POWER5 performance monitor hardware support registered
[    0.046436] Brought up 4 CPUs
[    0.046450] Node 0 CPUs: 0-3                                                 [    0.048151] devtmpfs: initialized
[    0.084223] EEH: devices created
[    0.086942] atomic64 test passed
[    0.087097] NET: Registered protocol family 16
[    0.087144] pseries_eeh_init: RTAS service <ibm,get-config-addr-info2> and <i
bm,get-config-addr-info> invalid
[    0.087155] eeh_init: Failed to call platform init function (-22)
[    0.087364] IBM eBus Device Driver
[    0.091962] PCI: Probing PCI hardware
[    0.092055] PCI host bridge to bus 0000:00
[    0.092068] pci_bus 0000:00: root bus resource [io  0x10000-0x10ffff] (bus ad
dress [0x0000-0xfffff])
[    0.092083] pci_bus 0000:00: root bus resource [mem 0x40080000000-0x400bfffff
ff] (bus address [0xc0000000-0xffffffff])
[    0.092097] pci_bus 0000:00: root bus resource [bus 00-ff]
[    0.101004] IOMMU table initialized, virtual merging enabled
[    0.105916] pci 0000:c8:01.0: supports D1 D2
[    0.105929] pci 0000:c8:01.0: PME# supported from D0 D1 D2 D3hot
[    0.108376] pci 0000:c8:01.1: supports D1 D2
[    0.108386] pci 0000:c8:01.1: PME# supported from D0 D1 D2 D3hot
[    0.110812] pci 0000:c8:01.2: supports D1 D2
[    0.110822] pci 0000:c8:01.2: PME# supported from D0 D1 D2 D3hot
[    0.114267] pci 0000:cc:01.0: supports D1                                    [    0.118840] pci 0000:c0:01.0: PME# supported from D0 D3hot D3cold
[    0.123005] pci 0000:c0:01.1: PME# supported from D0 D3hot D3cold
[    0.128128] pci 0000:d0:01.0: PME# supported from D0 D3hot D3cold
[    0.132394] pci 0000:d0:01.1: PME# supported from D0 D3hot D3cold
[    0.135218] PCI host bridge to bus 0001:00
[    0.135230] pci_bus 0001:00: root bus resource [io  0x120000-0x21ffff] (bus a
ddress [0x0000-0xfffff])
[    0.135245] pci_bus 0001:00: root bus resource [mem 0x401c0000000-0x401ffffff
ff] (bus address [0xc0000000-0xffffffff])
[    0.135259] pci_bus 0001:00: root bus resource [bus 00-ff]
[    0.149012] pci 0001:c8:01.0: PME# supported from D0 D3hot D3cold
[    0.153406] pci 0001:c8:01.1: PME# supported from D0 D3hot D3cold
[    0.165162] PCI: Probing PCI hardware done
[    0.165193] opal: Node not found
[    0.167689] bio: create slab <bio-0> at 0
[    0.167992] vgaarb: loaded
[    0.168146] SCSI subsystem initialized
[    0.168224] libata version 3.00 loaded.
[    0.168317] usbcore: registered new interface driver usbfs
[    0.168345] usbcore: registered new interface driver hub
[    0.168407] usbcore: registered new device driver usb
[    0.168766] NetLabel: Initializing
[    0.168774] NetLabel:  domain hash size = 128                                [    0.168780] NetLabel:  protocols = UNLABELED CIPSOv4
[    0.168811] NetLabel:  unlabeled traffic allowed by default
[    0.169041] Switching to clocksource timebase
[    0.192701] NET: Registered protocol family 2
[    0.193026] TCP established hash table entries: 32768 (order: 3, 524288 bytes
)
[    0.193445] TCP bind hash table entries: 32768 (order: 3, 524288 bytes)
[    0.193804] TCP: Hash tables configured (established 32768 bind 32768)
[    0.193846] TCP: reno registered
[    0.193859] UDP hash table entries: 2048 (order: 0, 65536 bytes)
[    0.193923] UDP-Lite hash table entries: 2048 (order: 0, 65536 bytes)
[    0.194112] NET: Registered protocol family 1
[    0.309531] pci 0000:c8:01.2: enabling device (0140 -> 0142)
[    0.310778] PCI: CLS 128 bytes, default 128
[    0.310876] Unpacking initramfs...
[    1.101738] Freeing initrd memory: 16384k freed
[    1.102578] RTAS daemon started
[    1.103547] RTAS: event: 14, Type: Platform Error, Severity: 2
[    1.103949] pseries_idle_driver registered
[    1.104295] audit: initializing netlink socket (disabled)
[    1.104326] type=2000 audit(1370322269.090:1): initialized
[    1.364540] HugeTLB registered 16 MB page size, pre-allocated 0 pages
[    1.368592] VFS: Disk quotas dquot_6.5.2
[    1.368747] Dquot-cache hash table entries: 8192 (order 0, 65536 bytes)
[    1.369961] msgmni has been set to 7516
[    1.370178] SELinux:  Registering netfilter hooks
[    1.371311] alg: No test for stdrng (krng)
[    1.371335] NET: Registered protocol family 38
[    1.371440] Block layer SCSI generic (bsg) driver version 0.4 loaded (major 2
52)
[    1.371502] io scheduler noop registered
[    1.371510] io scheduler deadline registered
[    1.371536] io scheduler cfq registered (default)
[    1.371729] pci_hotplug: PCI Hot Plug PCI Core version: 0.5
[    1.372753] Serial: 8250/16550 driver, 4 ports, IRQ sharing enabled
[    1.373487] Linux agpgart interface v0.103
[    1.375442] loop: module loaded
[    1.375525] Uniform Multi-Platform E-IDE driver
[    1.375593] ide-gd driver 1.18
[    1.375904] libphy: Fixed MDIO Bus: probed
[    1.376023] ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver
[    1.376046] ehci-pci: EHCI PCI platform driver
[    1.376238] ehci-pci 0000:c8:01.2: enabling device (0140 -> 0142)
[    1.376650] ehci-pci 0000:c8:01.2: EHCI Host Controller
[    1.376755] ehci-pci 0000:c8:01.2: new USB bus registered, assigned bus numbe
r 1                                                                             [    1.377184] ehci-pci 0000:c8:01.2: can't setup
[    1.377197] ehci-pci 0000:c8:01.2: USB bus 1 deregistered
[    1.377567] ehci-pci 0000:c8:01.2: init 0000:c8:01.2 fail, -19
[    1.377598] ohci_hcd: USB 1.1 'Open' Host Controller (OHCI) Driver
[    1.389160] ohci_hcd 0000:c8:01.0: Refused to change power state, currently i
n D3
[    1.389322] ohci_hcd 0000:c8:01.0: OHCI Host Controller
[    1.389429] ohci_hcd 0000:c8:01.0: new USB bus registered, assigned bus numbe
r 1
[   11.389147] ohci_hcd 0000:c8:01.0: USB HC takeover failed!  (BIOS/SMM bug)
[   11.389161] ohci_hcd 0000:c8:01.0: can't setup
[   11.389172] ohci_hcd 0000:c8:01.0: USB bus 1 deregistered
[   11.389572] ohci_hcd 0000:c8:01.0: init 0000:c8:01.0 fail, -16
[   11.389586] ohci_hcd: probe of 0000:c8:01.0 failed with error -16
[   11.409162] ohci_hcd 0000:c8:01.1: Refused to change power state, currently i
n D3
[   11.409334] ohci_hcd 0000:c8:01.1: OHCI Host Controller
[   11.409562] ohci_hcd 0000:c8:01.1: new USB bus registered, assigned bus numbe
r 1
[   21.409088] ohci_hcd 0000:c8:01.1: USB HC takeover failed!  (BIOS/SMM bug)
[   21.409102] ohci_hcd 0000:c8:01.1: can't setup
[   21.409113] ohci_hcd 0000:c8:01.1: USB bus 1 deregistered
[   21.409495] ohci_hcd 0000:c8:01.1: init 0000:c8:01.1 fail, -16               [   21.409508] ohci_hcd: probe of 0000:c8:01.1 failed with error -16
[   21.409604] uhci_hcd: USB Universal Host Controller Interface driver
[   21.410032] usbcore: registered new interface driver usbserial
[   21.410069] usbcore: registered new interface driver usbserial_generic
[   21.410091] usbserial: USB Serial support registered for generic
[   21.410402] mousedev: PS/2 mouse device common for all mice
[   21.410711] rtc-generic rtc-generic: rtc core: registered rtc-generic as rtc0
[   21.410969] device-mapper: uevent: version 1.0.3
[   21.411142] device-mapper: ioctl: 4.23.1-ioctl (2012-12-18) initialised: dm-d
evel@redhat.com
[   21.411279] cpuidle: using governor ladder
[   21.411423] cpuidle: using governor menu
[   21.411733] usbcore: registered new interface driver usbhid
[   21.411741] usbhid: USB HID core driver
[   21.411820] drop_monitor: Initializing network drop monitor service
[   21.411990] ip_tables: (C) 2000-2006 Netfilter Core Team
[   21.412026] TCP: cubic registered
[   21.412034] Initializing XFRM netlink socket
[   21.412297] NET: Registered protocol family 10
[   21.412687] mip6: Mobile IPv6
[   21.412697] NET: Registered protocol family 17
[   21.412859] Running MSI bitmap self-tests ...
[   21.416401] PM: Hibernation image not present or could not be loaded.        [   21.416428] registered taskstats version 1
[   21.424596] rtc-generic rtc-generic: setting system clock to 2013-06-04 05:04
:50 UTC (1370322290)
[   21.428021] Freeing unused kernel memory: 6912k freed
[   21.567622] dracut: dracut-018-105.git20120927.fc17
[   21.680580] udevd[112]: starting version 182
[   21.791462] dracut: Starting plymouth daemon
[   21.926482] pata_pdc2027x 0000:cc:01.0: version 1.0
[   21.926719] pata_pdc2027x 0000:cc:01.0: enabling device (0000 -> 0003)
[   21.938847] ipr: IBM Power RAID SCSI Device Driver version: 2.5.4 (July 11, 2
012)
[   21.939061] ipr 0001:d0:01.0: Found IOA with IRQ: 151
[   21.942276] ipr 0001:d0:01.0: Write of cache line size failed
[   21.942584] ipr: probe of 0001:d0:01.0 failed with error -5
[   22.027478] pata_pdc2027x 0000:cc:01.0: PLL input clock 0 kHz
[   22.027493] pata_pdc2027x: Invalid PLL input clock 0kHz, give up!
[   22.029939] scsi1 : pata_pdc2027x
[   22.030669] scsi2 : pata_pdc2027x
[   22.030809] ata1: PATA max UDMA/133 mmio m16384@0x400b0000000 cmd 0x400b00017
c0 irq 118
[   22.030824] ata2: PATA max UDMA/133 mmio m16384@0x400b0000000 cmd 0x400b00015
c0 irq 118
[   22.189135] pata_pdc2027x: 40-conductor cable detected on port 0             [   22.349130] pata_pdc2027x: 40-conductor cable detected on port 1
[   43.699176] dracut Warning: Unable to process initqueue
[   43.704038] dracut Warning: /dev/mapper/vg_narvi-lv_root does not exist
[   43.704543] dracut Warning: /dev/vg_narvi/lv_root does not exist
[   43.705036] dracut Warning: /dev/vg_narvi/lv_swap does not exist

Re: IBM OpenPower 720 ipr driver woes

[Tony Breeds]

[-- Attachment #1: Type: text/plain, Size: 1419 bytes --]

On Mon, Jun 03, 2013 at 09:40:52PM -0400, Robert Knight wrote:
> On 6/3/2013 8:01 PM, Tony Breeds wrote:
> >On Mon, Jun 03, 2013 at 05:20:12PM -0400, Robert Knight wrote:
> >
> >>>Device tree struct  0x0000000004820000 -> 0x0000000004840000
> >>>Calling quiesce...
> >>>returning from prom_init
> >>>[    1.376359] ehci-pci 0000:c8:01.2: can't setup
> >Can you try adding "debug" to the kernel commandline.  We're missing a
> >great chunk of detail. If you're starting from scratch either try F18 or
> >the F19 Beta (if you're brave :))
> >
> >Yours Tony
> So, two points.  Since I have no live disks, I can't copy the dmesg
> output in dracut onto a disk and just sent it to you -- the only way
> that I know to get it is cut and paste from a HMC console window
> into a vi buffer.

Okay, when I've been in that situation, I run "script" locally and then
conenct to the HMC console.

When you're done you can exit script adn then a file called typescript
will exist in the directory you ran script in.

It's a less overhead way of doign what you're already doing.
 [    0.087097] NET: Registered protocol family 16

> [    0.087144] pseries_eeh_init: RTAS service <ibm,get-config-addr-info2> and <i
> bm,get-config-addr-info> invalid
> [    0.087155] eeh_init: Failed to call platform init function (-22)

Hmm this seems pretty strange to me.  Gavin are these RTAS tokens
supported on older power5 boxes?


Yours Tony

[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]

Re: IBM OpenPower 720 ipr driver woes

[Gavin Shan]

[-- Attachment #1: Type: text/plain, Size: 1738 bytes --]

On Tue, Jun 04, 2013 at 01:16:52PM +1000, Tony Breeds wrote:
>On Mon, Jun 03, 2013 at 09:40:52PM -0400, Robert Knight wrote:
>> On 6/3/2013 8:01 PM, Tony Breeds wrote:
>> >On Mon, Jun 03, 2013 at 05:20:12PM -0400, Robert Knight wrote:
>> >
>> >>>Device tree struct  0x0000000004820000 -> 0x0000000004840000
>> >>>Calling quiesce...
>> >>>returning from prom_init
>> >>>[    1.376359] ehci-pci 0000:c8:01.2: can't setup
>> >Can you try adding "debug" to the kernel commandline.  We're missing a
>> >great chunk of detail. If you're starting from scratch either try F18 or
>> >the F19 Beta (if you're brave :))
>> >
>> >Yours Tony
>> So, two points.  Since I have no live disks, I can't copy the dmesg
>> output in dracut onto a disk and just sent it to you -- the only way
>> that I know to get it is cut and paste from a HMC console window
>> into a vi buffer.
>
>Okay, when I've been in that situation, I run "script" locally and then
>conenct to the HMC console.
>
>When you're done you can exit script adn then a file called typescript
>will exist in the directory you ran script in.
>
>It's a less overhead way of doign what you're already doing.
> [    0.087097] NET: Registered protocol family 16
>
>> [    0.087144] pseries_eeh_init: RTAS service <ibm,get-config-addr-info2> and <i
>> bm,get-config-addr-info> invalid
>> [    0.087155] eeh_init: Failed to call platform init function (-22)
>
>Hmm this seems pretty strange to me.  Gavin are these RTAS tokens
>supported on older power5 boxes?
>

Yes, Tony. "ibm,get-config-addr-info" should be supported on Power5 box.
Newer PowerBox (e.g. P7) should support "ibm,get-config-addr-info2"

Please have a try on the attached patch, which is based on mainline (3.10).

Thanks,
Gavin



[-- Attachment #2: 0001-powerpc-eeh-Don-t-check-RTAS-token-to-get-PE-addr.patch --]
[-- Type: text/x-diff, Size: 2348 bytes --]

>From 04771628d53e1e6883063ed21bd6825ee9680366 Mon Sep 17 00:00:00 2001
From: Gavin Shan <shangw@linux.vnet.ibm.com>
Date: Tue, 4 Jun 2013 11:47:59 +0800
Subject: [PATCH] powerpc/eeh: Don't check RTAS token to get PE addr

RTAS token "ibm,get-config-addr-info" or ibm,get-config-addr-info2"
are used to retrieve the PE address according to PCI address, which
made up of domain/bus/slot/function. If we don't have those 2 tokens,
the domain/bus/slot/function would be used as the address for EEH
RTAS operations. Some older f/w might not have those 2 tokens and
that blocks the EEH functionality to be initialized.

The patch skips the check on those 2 tokens so we can bring up EEH
functionality successfully. And domain/bus/slot/function will be
used as address for EEH RTAS operations.

Signed-off-by: Gavin Shan <shangw@linux.vnet.ibm.com>
---
 arch/powerpc/platforms/pseries/eeh_pseries.c |   12 +++++-------
 1 files changed, 5 insertions(+), 7 deletions(-)

diff --git a/arch/powerpc/platforms/pseries/eeh_pseries.c b/arch/powerpc/platforms/pseries/eeh_pseries.c
index 19506f9..b456b15 100644
--- a/arch/powerpc/platforms/pseries/eeh_pseries.c
+++ b/arch/powerpc/platforms/pseries/eeh_pseries.c
@@ -83,7 +83,11 @@ static int pseries_eeh_init(void)
 	ibm_configure_pe		= rtas_token("ibm,configure-pe");
 	ibm_configure_bridge		= rtas_token("ibm,configure-bridge");
 
-	/* necessary sanity check */
+	/*
+	 * Necessary sanity check. We needn't check "get-config-addr-info"
+	 * and its variant since the old firmware probably support address
+	 * of domain/bus/slot/function for EEH RTAS operations.
+	 */
 	if (ibm_set_eeh_option == RTAS_UNKNOWN_SERVICE) {
 		pr_warning("%s: RTAS service <ibm,set-eeh-option> invalid\n",
 			__func__);
@@ -102,12 +106,6 @@ static int pseries_eeh_init(void)
 		pr_warning("%s: RTAS service <ibm,slot-error-detail> invalid\n",
 			__func__);
 		return -EINVAL;
-	} else if (ibm_get_config_addr_info2 == RTAS_UNKNOWN_SERVICE &&
-		   ibm_get_config_addr_info == RTAS_UNKNOWN_SERVICE) {
-		pr_warning("%s: RTAS service <ibm,get-config-addr-info2> and "
-			"<ibm,get-config-addr-info> invalid\n",
-			__func__);
-		return -EINVAL;
 	} else if (ibm_configure_pe == RTAS_UNKNOWN_SERVICE &&
 		   ibm_configure_bridge == RTAS_UNKNOWN_SERVICE) {
 		pr_warning("%s: RTAS service <ibm,configure-pe> and "
-- 
1.7.5.4

Re: IBM OpenPower 720 ipr driver woes

[Brian King]

On 06/03/2013 08:40 PM, Robert Knight wrote:
> [   21.938847] ipr: IBM Power RAID SCSI Device Driver version: 2.5.4 (July 11, 2
> 012)
> [   21.939061] ipr 0001:d0:01.0: Found IOA with IRQ: 151
> [   21.942276] ipr 0001:d0:01.0: Write of cache line size failed
> [   21.942584] ipr: probe of 0001:d0:01.0 failed with error -5

This looks to be the reason the ipr driver is failing to load. Not sure why this
is failing, I don't think this code has changed for a very long time. Arguably,
this shouldn't be a fatal error for the driver. Wendy - do you have access to
one of these systems so we could take a look at this? 

-Brian

-- 
Brian King
Power Linux I/O
IBM Linux Technology Center

Re: IBM OpenPower 720 ipr driver woes

[Robert Knight]

On 6/4/2013 6:16 PM, Brian King wrote:
> Wendy - do you have access to
> one of these systems so we could take a look at this?
>
I can provide access if there's not another available.

Regards,
Robert

Re: IBM OpenPower 720 ipr driver woes

[Robert Knight]

On 6/3/2013 11:52 PM, Gavin Shan wrote:
> On Tue, Jun 04, 2013 at 01:16:52PM +1000, Tony Breeds wrote:
>> On Mon, Jun 03, 2013 at 09:40:52PM -0400, Robert Knight wrote:
>>> On 6/3/2013 8:01 PM, Tony Breeds wrote:
>>>> On Mon, Jun 03, 2013 at 05:20:12PM -0400, Robert Knight wrote:
>>>>
>>>>>> Device tree struct  0x0000000004820000 -> 0x0000000004840000
>>>>>> Calling quiesce...
>>>>>> returning from prom_init
>>>>>> [    1.376359] ehci-pci 0000:c8:01.2: can't setup
>>>> Can you try adding "debug" to the kernel commandline.  We're missing a
>>>> great chunk of detail. If you're starting from scratch either try F18 or
>>>> the F19 Beta (if you're brave :))
>>>>
>>>> Yours Tony
>>> So, two points.  Since I have no live disks, I can't copy the dmesg
>>> output in dracut onto a disk and just sent it to you -- the only way
>>> that I know to get it is cut and paste from a HMC console window
>>> into a vi buffer.
>> Okay, when I've been in that situation, I run "script" locally and then
>> conenct to the HMC console.
>>
>> When you're done you can exit script adn then a file called typescript
>> will exist in the directory you ran script in.
>>
>> It's a less overhead way of doign what you're already doing.
>> [    0.087097] NET: Registered protocol family 16
>>
>>> [    0.087144] pseries_eeh_init: RTAS service <ibm,get-config-addr-info2> and <i
>>> bm,get-config-addr-info> invalid
>>> [    0.087155] eeh_init: Failed to call platform init function (-22)
>> Hmm this seems pretty strange to me.  Gavin are these RTAS tokens
>> supported on older power5 boxes?
>>
> Yes, Tony. "ibm,get-config-addr-info" should be supported on Power5 box.
> Newer PowerBox (e.g. P7) should support "ibm,get-config-addr-info2"
>
> Please have a try on the attached patch, which is based on mainline (3.10).
>
> Thanks,
> Gavin
>
>
The system boots with that patch.  I applied it to kernel-3.8.11-100.

Re: IBM OpenPower 720 ipr driver woes

[wenxiong]

Quoting Robert Knight <knight@princeton.edu>:

> Sometime after kernel 3.3.4, Fedora 17 running on an IBM OpenPower  
> 720 stopped being able to read the SCSI disks with an error  
> concerning a return code -1.  In the updates since, it has not  
> resumed working.
>
Can you tell me which kernel you saw this issue(I assumed that 3.3.4  
works fine).
Also can you check your system firmware for me?

Thanks,
Wendy

Re: IBM OpenPower 720 ipr driver woes

[Robert Knight]

On 6/5/2013 6:01 PM, wenxiong@linux.vnet.ibm.com wrote:
>
> Quoting Robert Knight <knight@princeton.edu>:
>
>> Sometime after kernel 3.3.4, Fedora 17 running on an IBM OpenPower 
>> 720 stopped being able to read the SCSI disks with an error 
>> concerning a return code -1.  In the updates since, it has not 
>> resumed working.
>>
> Can you tell me which kernel you saw this issue(I assumed that 3.3.4 
> works fine).
kernel 3.8.11-100, for example
> Also can you check your system firmware for me?
I don't know how to do that.  lsvpd?

Re: IBM OpenPower 720 ipr driver woes

[Brian King]

On 06/05/2013 04:14 PM, Robert Knight wrote:
> On 6/3/2013 11:52 PM, Gavin Shan wrote:
>> On Tue, Jun 04, 2013 at 01:16:52PM +1000, Tony Breeds wrote:
>>> On Mon, Jun 03, 2013 at 09:40:52PM -0400, Robert Knight wrote:
>>>> On 6/3/2013 8:01 PM, Tony Breeds wrote:
>>>>> On Mon, Jun 03, 2013 at 05:20:12PM -0400, Robert Knight wrote:
>>>>>
>>>>>>> Device tree struct  0x0000000004820000 -> 0x0000000004840000
>>>>>>> Calling quiesce...
>>>>>>> returning from prom_init
>>>>>>> [    1.376359] ehci-pci 0000:c8:01.2: can't setup
>>>>> Can you try adding "debug" to the kernel commandline.  We're missing a
>>>>> great chunk of detail. If you're starting from scratch either try F18 or
>>>>> the F19 Beta (if you're brave :))
>>>>>
>>>>> Yours Tony
>>>> So, two points.  Since I have no live disks, I can't copy the dmesg
>>>> output in dracut onto a disk and just sent it to you -- the only way
>>>> that I know to get it is cut and paste from a HMC console window
>>>> into a vi buffer.
>>> Okay, when I've been in that situation, I run "script" locally and then
>>> conenct to the HMC console.
>>>
>>> When you're done you can exit script adn then a file called typescript
>>> will exist in the directory you ran script in.
>>>
>>> It's a less overhead way of doign what you're already doing.
>>> [    0.087097] NET: Registered protocol family 16
>>>
>>>> [    0.087144] pseries_eeh_init: RTAS service <ibm,get-config-addr-info2> and <i
>>>> bm,get-config-addr-info> invalid
>>>> [    0.087155] eeh_init: Failed to call platform init function (-22)
>>> Hmm this seems pretty strange to me.  Gavin are these RTAS tokens
>>> supported on older power5 boxes?
>>>
>> Yes, Tony. "ibm,get-config-addr-info" should be supported on Power5 box.
>> Newer PowerBox (e.g. P7) should support "ibm,get-config-addr-info2"
>>
>> Please have a try on the attached patch, which is based on mainline (3.10).
>>
>> Thanks,
>> Gavin
>>
>>
> The system boots with that patch.  I applied it to kernel-3.8.11-100.

Does that patch resolve all your issues, or are there still issues with ipr remaining
after applying the patch?

Thanks,

Brian

-- 
Brian King
Power Linux I/O
IBM Linux Technology Center

Re: IBM OpenPower 720 ipr driver woes

[Robert Knight]

On 06/06/2013 07:32 AM, Brian King wrote:
> On 06/05/2013 04:14 PM, Robert Knight wrote:
>> On 6/3/2013 11:52 PM, Gavin Shan wrote:
>>> On Tue, Jun 04, 2013 at 01:16:52PM +1000, Tony Breeds wrote:
>>>> On Mon, Jun 03, 2013 at 09:40:52PM -0400, Robert Knight wrote:
>>>>> On 6/3/2013 8:01 PM, Tony Breeds wrote:
>>>>>> On Mon, Jun 03, 2013 at 05:20:12PM -0400, Robert Knight wrote:
>>>>>>
>>>>>>>> Device tree struct  0x0000000004820000 -> 0x0000000004840000
>>>>>>>> Calling quiesce...
>>>>>>>> returning from prom_init
>>>>>>>> [    1.376359] ehci-pci 0000:c8:01.2: can't setup
>>>>>> Can you try adding "debug" to the kernel commandline.  We're missing a
>>>>>> great chunk of detail. If you're starting from scratch either try F18 or
>>>>>> the F19 Beta (if you're brave :))
>>>>>>
>>>>>> Yours Tony
>>>>> So, two points.  Since I have no live disks, I can't copy the dmesg
>>>>> output in dracut onto a disk and just sent it to you -- the only way
>>>>> that I know to get it is cut and paste from a HMC console window
>>>>> into a vi buffer.
>>>> Okay, when I've been in that situation, I run "script" locally and then
>>>> conenct to the HMC console.
>>>>
>>>> When you're done you can exit script adn then a file called typescript
>>>> will exist in the directory you ran script in.
>>>>
>>>> It's a less overhead way of doign what you're already doing.
>>>> [    0.087097] NET: Registered protocol family 16
>>>>
>>>>> [    0.087144] pseries_eeh_init: RTAS service <ibm,get-config-addr-info2> and <i
>>>>> bm,get-config-addr-info> invalid
>>>>> [    0.087155] eeh_init: Failed to call platform init function (-22)
>>>> Hmm this seems pretty strange to me.  Gavin are these RTAS tokens
>>>> supported on older power5 boxes?
>>>>
>>> Yes, Tony. "ibm,get-config-addr-info" should be supported on Power5 box.
>>> Newer PowerBox (e.g. P7) should support "ibm,get-config-addr-info2"
>>>
>>> Please have a try on the attached patch, which is based on mainline (3.10).
>>>
>>> Thanks,
>>> Gavin
>>>
>>>
>> The system boots with that patch.  I applied it to kernel-3.8.11-100.
> Does that patch resolve all your issues, or are there still issues with ipr remaining
> after applying the patch?
>
> Thanks,
>
> Brian
>
Yes.  I've started rebuilding the kernel and I'm up to the module 
building part, so I'd say it is solid.  Will this patch make it into 
some version of the kernel?

What was killing me was that it would not complete boot.  It now does.  
I see:

[   11.934481] scsi 0:0:15:0: Resetting device
[   11.934813] ipr 0001:d0:01.0: Adapter being reset as a result of 
error recovery.

on each boot.  It does not appear to affect operation.

Thank you and the rest of the team for your rapid and helpful responses.

Best regards,
Robert

Re: IBM OpenPower 720 ipr driver woes

[Gavin Shan]

On Thu, Jun 06, 2013 at 08:39:45AM -0400, Robert Knight wrote:
>On 06/06/2013 07:32 AM, Brian King wrote:
>>On 06/05/2013 04:14 PM, Robert Knight wrote:
>>>On 6/3/2013 11:52 PM, Gavin Shan wrote:
>>>>On Tue, Jun 04, 2013 at 01:16:52PM +1000, Tony Breeds wrote:
>>>>>On Mon, Jun 03, 2013 at 09:40:52PM -0400, Robert Knight wrote:
>>>>>>On 6/3/2013 8:01 PM, Tony Breeds wrote:
>>>>>>>On Mon, Jun 03, 2013 at 05:20:12PM -0400, Robert Knight wrote:

.../...

>Yes.  I've started rebuilding the kernel and I'm up to the module
>building part, so I'd say it is solid.  Will this patch make it into
>some version of the kernel?
>

The patch is being pushed to mainline or linux-next, and backported
to stable-kernel (v3.4+)


>What was killing me was that it would not complete boot.  It now
>does.  I see:
>
>[   11.934481] scsi 0:0:15:0: Resetting device
>[   11.934813] ipr 0001:d0:01.0: Adapter being reset as a result of
>error recovery.
>
>on each boot.  It does not appear to affect operation.
>
>Thank you and the rest of the team for your rapid and helpful responses.
>

Thanks,
Gavin

Re: IBM OpenPower 720 ipr driver woes

[Robert Knight]

On 06/06/2013 08:24 PM, Gavin Shan wrote:
> On Thu, Jun 06, 2013 at 08:39:45AM -0400, Robert Knight wrote:
>> On 06/06/2013 07:32 AM, Brian King wrote:
>>> On 06/05/2013 04:14 PM, Robert Knight wrote:
>>>> On 6/3/2013 11:52 PM, Gavin Shan wrote:
>>>>> On Tue, Jun 04, 2013 at 01:16:52PM +1000, Tony Breeds wrote:
>>>>>> On Mon, Jun 03, 2013 at 09:40:52PM -0400, Robert Knight wrote:
>>>>>>> On 6/3/2013 8:01 PM, Tony Breeds wrote:
>>>>>>>> On Mon, Jun 03, 2013 at 05:20:12PM -0400, Robert Knight wrote:
> .../...
>
>> Yes.  I've started rebuilding the kernel and I'm up to the module
>> building part, so I'd say it is solid.  Will this patch make it into
>> some version of the kernel?
>>
> The patch is being pushed to mainline or linux-next, and backported
> to stable-kernel (v3.4+)
>
>
>> What was killing me was that it would not complete boot.  It now
>> does.  I see:
>>
>> [   11.934481] scsi 0:0:15:0: Resetting device
>> [   11.934813] ipr 0001:d0:01.0: Adapter being reset as a result of
>> error recovery.
>>
>> on each boot.  It does not appear to affect operation.
>>
>> Thank you and the rest of the team for your rapid and helpful responses.
>>
> Thanks,
> Gavin
>
Well, it's four months later and I'm trying to get Fedora 20 Alpha to 
install on that same machine.  It appears to still have the same 
problem.  Did that patch ever make it into the mainline?

Strangely, the kernel from the installer (using DVD image) does NOT have 
the problem, only the installed system.

Regards,
Robert

Re: IBM OpenPower 720 ipr driver woes

[Robert Knight]

On 10/17/2013 10:57 AM, Robert Knight wrote:
> Well, it's four months later and I'm trying to get Fedora 20 Alpha to 
> install on that same machine.  It appears to still have the same 
> problem.  Did that patch ever make it into the mainline?
>
> Strangely, the kernel from the installer (using DVD image) does NOT 
> have the problem, only the installed system.
Just to follow up on this.

The problem does not appear to be the kernel.  Installing on one disk 
succeeds in producing a bootable system every time.  Installing on all 
four disks (two 73 GB disks and two 146 GB disks) never succeeds.

Although there are a storm of ipr driver messages as before, that is 
likely not the problem.

I'm going to continue on #fedora-ppc and the corresponding mailing list.