Bug 116012

Summary: LTC18120 - can't install opensuse10 on legacy iseries viodasd
Product: [openSUSE] SUSE LINUX 10.0 Reporter: Mike Ranweiler <mranweil>
Component: InstallationAssignee: Steffen Winterfeldt <snwint>
Status: RESOLVED DUPLICATE QA Contact: Klaus Kämpf <kkaempf>
Severity: Major    
Priority: P5 - None CC: bugproxy, jsrain
Version: Beta 4   
Target Milestone: ---   
Hardware: PowerPC-64   
OS: All   
Whiteboard:
Found By: Third Party Developer/Partner Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---

Description Mike Ranweiler 2005-09-08 22:24:50 UTC
Problem description:  The viodasd module to access the virtual disk is not part
of the ISERIES64 installation kernel used on legacy (pre-power5) iSeries.  For
most people using a virtual disk this means you have nothing to install onto
(most iSeries people use viodasd or vscsi).

Vscsi is an option for people with v5r3 legacy iseries boxes, but generally
people use viodasd.  I think Olaf mentioned that he thought hwinfo wasn't
loading this.

Hardware Environment
    Machine type (p650, x235, SF2, etc.):  iSeries 830
    Cpu type (Power4, Power5, IA-64, etc.):  creaky old istar

Software Environment
    OpenSuSE 10 Beta4 (yes, rc1 is out and I want to try that, too, but that
isn't going to be setup internally until tomorrow - I did at least do a quick
bugzilla search for this).

Additional information:
I checked kernel-iseries64-2.6.13-3.ppc.rpm and VIODASD=y so I'm assuming this
is a problem in the installation system.

After I aborted out of YaST during the installation back to linuxrc, I see this
(ie, no viodasd):

> 3

Show Loaded Modules

VFAT
Generic Mid-Level SCSI Driver
SCSI Disk Driver
SCSI CD-ROM driver
iSeries virtual ethernet
SCSI Tape Driver
Device Mapper
RAID-0 (striping) mode
RAID-1 (mirroring) mode
RAID-4/RAID-5 xor
RAID-4/RAID-5 mode
RAID-6 mode

Kernel Modules (Hardware Drivers)

1) Load ppc_iseries64 Modules
2) Load Unknown Modules
3) Show Loaded Modules
4) Unload Modules
5) Add Driver Update
6) Show Driver Updates

> 1

Load ppc_iseries64 Modules

 1) ipr : ipr SCSI
 2) ibmvscsic : IBM Virtual SCSI
 3) iseries_veth : iSeries virtual ethernet
 4) pcnet32 : AMD PCI PCnet32 (PCI bus NE2100)
 5) acenic : Alteon AceNIC/3C985/NetGear GA620
 6) e100 : Intel PRO/100 / EtherExpr. PRO/100
 7) e1000 : Intel PRO/1000
 8) olympic : IBM Olympic chipset PCI Token Ring
 9) sr_mod : SCSI CD-ROM driver
10) lpfc : Emulex LP fibrechannel storage
11) dm-mod : Device Mapper
12) linear : Linear (append) mode
13) raid0 : RAID-0 (striping) mode
14) raid1 : RAID-1 (mirroring) mode
15) raid5 : RAID-4/RAID-5 mode
16) raid6 : RAID-6 mode
17) raid10 : RAID-10 mode
18) xor : RAID-4/RAID-5 xor
19) loop_fish2 : crypto loop dev
20) jfs : JFS
21) smbfs : SMB
22) vfat : VFAT
23) xfs : XFS

>
Comment 1 Steffen Winterfeldt 2005-09-09 09:41:56 UTC
VIODASD is compiled in (as you already noted) - so no module. 
Comment 2 Olaf Hering 2005-09-09 11:09:55 UTC
I assume that hwinfo is still broken. Can you boot with start_shell and check
dmesg and /var/log/boot.msg if the kernel does detect them? There is another bug
open for that, since a couple of weeks, 
Comment 3 Mike Ranweiler 2005-09-12 22:31:44 UTC
Want me to dup this to bug 94272?
Comment 4 LTC BugProxy 2005-09-12 22:36:57 UTC
changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|Novell116012-  viodasd      |Novell116012 - can't install
                   |module not part of ISERIES64|opensuse10 on legacy iseries
                   |installer                   |viodasd




------- Additional Comments From mranweil@us.ibm.com  2005-09-12 18:30 EDT -------
Saying there was no viodasd was a brain fart of colossal proportions, I'm glad
you knew what I meant.  I assume you meant this bug, Olaf:
https://bugzilla.novell.com/show_bug.cgi?id=94272
That's from June, so maybe not what you meant...

I re-tried this with SP1, same problem.  Can see this in the kernel messages:

RAMDISK driver initialized: 16 RAM disks of 123456K size 1024 blocksize
loop: loaded (max 8 devices)
viod: vers 1.64, hosting partition 0
viod: disk 0: 9221310 sectors (4502 MB) CHS=574/255/63 sector size 512
 iseries/vda: iseries/vda1 iseries/vda2 iseries/vda3
viocd: vers 1.06, hosting partition 0
Uniform CD-ROM driver Revision: 3.20
viocd: cd iseries/vcda is iSeries resource OPT01      type 6321, model 002
viocd: cd iseries/vcdb is iSeries resource OPTVRT02   type 632B, model 001
mice: PS/2 mouse device common for all mice
md: md driver 0.90.2 MAX_MD_DEVS=256, MD_SB_DISKS=27
md: bitmap version 3.38

From linuxrc system info -> harddisks/cdroms only shows the two cdroms, though.
 fdisk in rescue shows the disk.  hwinfo there the long listing shows this:
>> block.4: partition
----- /proc/partitions -----
     7     0      74255 loop0
   112     0    4610655 iseries/vda
   112     1      16033 iseries/vda1
   112     2     104422 iseries/vda2
   112     3    4490167 iseries/vda3
----- /proc/partitions end -----
disks:
  iseries/vda
partitions:
  iseries/vda1
  iseries/vda2
  iseries/vda3

But short doesn't see it:
Rescue:~ # hwinfo --short
cpu:
                       RS64-III (icestar), 500 MHz
network:
                       IBM Virtual Ethernet card 4
network interface:
  eth0                 Ethernet network interface
  lo                   Loopback network interface
cdrom:
  /dev/iseries/vcda    CD-ROM
  /dev/iseries/vcdb    CD-ROM
memory:
                       Main Memory
unknown:
                       DMA controller
                       PROM
Rescue:~ #  
Which looks similar to Olaf's bug 94272. 
Comment 5 Steffen Winterfeldt 2005-09-13 12:03:10 UTC

*** This bug has been marked as a duplicate of 94272 ***
Comment 6 LTC BugProxy 2006-03-08 19:14:59 UTC
changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |FIXEDAWAITINGTEST
         Resolution|                            |FIX_BY_DISTRO




------- Additional Comments From mranweil@us.ibm.com  2006-03-08 14:12 EDT -------
This is working now.