On Wed, Oct 29, 2025 at 11:00:18PM +0530, Shivaprasad G Bhat wrote:
> Hi Michal,
>
> On 10/21/25 4:37 PM, Michal Suchánek wrote:
> > Hello,
> >
> > I noticed removal of old pSeries revisions.
> >
> > FTR to boot Linux 3.0 I need pSeries-2.7 (already removed earlier).
> >
> > The thing that broke booting linux 3.0 for me is
> > 357d1e3bc7d2d80e5271bc4f3ac8537e30dc8046 spapr: Improved placement of
> > PCI host bridges in guest memory map
> >
> > I do not use Linux 3.0 anymore which is the reason I did not notice this
> > breakage due to old platform revision removal.
>
> I tried booting linux kernel 3.13.0-170-generic from ubuntu 14.04.6 LTS with
> the oldest supported machine pseries-5.0 as of now.
>
> It worked fine.
>
>
> qemu-system-ppc64 -machine pseries-5.0 -accel tcg -nographic -m size=12G
> -cpu power8 -smp 1 -drive file=/root/images/ubuntu16.04.qcow2,format=qcow2,if=none,id=drive-virtio-disk0
> -device virtio-blk-pci,drive=drive-virtio-disk0,id=virtio-disk0 -serial
> mon:stdio -kernel /root/images/vmlinux-3.13.0-170-generic -initrd
> /root/images/initrd.img-3.13.0-170-generic -append
> "BOOT_IMAGE=/boot/vmlinux-4.4.0-142-generic
> root=UUID=94fba90c-dbb0-4f8d-bc3e-acd5f2e54749 ro vt.handoff=7"
>
>
> shiva@ubuntu:~$ uname -a
> Linux ubuntu 3.13.0-170-generic #220-Ubuntu SMP Thu May 9 12:44:25 UTC 2019
> ppc64le ppc64le ppc64le GNU/Linux
> shiva@ubuntu:~$ cat /proc/cpuinfo
> processor : 0
> cpu : POWER8 (architected), altivec supported
> clock : 1000.000000MHz
> revision : 2.0 (pvr 004d 0200)
>
> timebase : 512000000
> platform : pSeries
> model : IBM pSeries (emulated by qemu)
> machine : CHRP IBM pSeries (emulated by qemu)
>
>
> Hope that helps.
How does that help?
Of course newer kernels do work.
However, distributions using Linux 3.0 don't anymore, at least that
particular Linux 3.0 build.
I am not particularly concerned, only noting that the qemu pSeries
emulation is not suitable for running very old distributions because
they are not compatible with newer qemu.
It can be a bug in the old kernel but it's not triggered by actual
hardware nor older qemu. Or it may bee a bug with that new qemu PCI bus
organization that is not triggered by newer kernel. I don't know and
since nobody cared enough to diagnose it so far we will probably never
know.
Thanks
Michal