On 24.02.2012, at 11:59, Dinar Valeev
On Fri, Feb 24, 2012 at 10:11 AM, Andreas Färber
wrote: Am 24.02.2012 03:42, schrieb Larry Stotler:
On Thu, Feb 23, 2012 at 5:18 PM, Andreas Färber
wrote: After lots of Alt-E (don't show again), Alt-J (use this file) and Alt-N (don't use this file) I finally got the installation running through until it wants to reboot. After reboot the system is not bootable and drops back to System Management menu and the boot options there too do not recognize any bootable devices apart from CD-ROM and NICs. Previously there was AIX5 on the disk and it seemed it was completely repartitioned by the installer.
On your system, YaST probably just used the whole disk. I've never used any of the POWER systems so I can't really help.
SLED 11(SUSE Linux Enterprise Desktop) would likely have better support for your hardware. [...]
Our quest here is testing and fixing Factory, not running some old-but-working version. :) Kudos to Dinar for getting it this far already!
So I am mainly in need of pointers how to get some of these fixed.
I verified that using ext4 and always choosing Alt+J for the checksums did not help.
I notice that the CD that boots fine uses \suse\yaboot.ibm, whereas the FAT16 boot partition uses just a yaboot file. Are there differences between yaboot and yaboot.ibm maybe? I tried appending all kinds of things in OF such as :1,yaboot without success, and the SMS menu is able to boot fine from CD but not from disk. If SMS doesn't sees a disk to boot from, then you don't have bootloader installed to the disk.
The right debug would be after isntallation run lilo in verbose mode. So we can see if yaboot wrapped correctly.
I finally have G5 on my desk (ppc970 based). So soon I'll start to do tests on G5 as well besides pSeries tests.
And yes. Back to december I remember that bootloader installation was failed even on pSeries. So something is wrong with lilo or perl-Bootloader.
At least the Apple OF has support for ELF straight in the firmware, so one thing to try would be to completely circumvent yaboot and boot the kernel straight from there. I don't think you can use an initrd with this though, so you might want to compile your own kernel with everything set to =y that you need to boot. It's just an alternative to get the system booting at all and debug yaboot from there. You can also try to run it in qemu-system-ppc64 with -M pseries, which should behave very similar to the pseries OF. Just that you can there modify the image from the outside to debug. Alex -- To unsubscribe, e-mail: opensuse-ppc+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-ppc+owner@opensuse.org