KVM: Failed to create TCE64 table for liobn 0x80000001
Description of problem:
When rebooting the system we hit :
   ```
   KVM: Failed to create TCE64 table for liobn 0x80000001
   qemu-system-ppc64: ../system/memory.c:2666: memory_region_add_subregion_common: Assertion `!subregion->container' failed.
   Aborted (core dumped)
   ```
Steps to reproduce:
1. Start the machine
2. Reboot it
 
   ```
   curl -LO https://cloud.centos.org/centos/10-stream/ppc64le/images/CentOS-Stream-GenericCloud-10-20250512.0.ppc64le.qcow2
   export LIBGUESTFS_BACKEND=direct
   virt-customize -v -a CentOS-Stream-GenericCloud-10-20250512.0.ppc64le.qcow2 --root-password password:centos
   qemu-system-ppc64 --enable-kvm -m 4096 -smp 8 -hda CentOS-Stream-GenericCloud-10-20250512.0.ppc64le.qcow2 -vga none -nographic -device qemu-xhci
   # once logged into it
   systemctl reboot
   [...]
   KVM: Failed to create TCE64 table for liobn 0x80000001
   qemu-system-ppc64: ../system/memory.c:2666: memory_region_add_subregion_common: Assertion `!subregion->container' failed.
   Aborted (core dumped)
   ```
Additional information:
The issue was already reported on ML https://lists.nongnu.org/archive/html/qemu-devel/2025-03/msg05137.html  
I also hit that issue while building a CoreOS CentOS Stream 10 image https://github.com/openshift/os/issues/1818. I was able to validate that the commit https://github.com/torvalds/linux/commit/6aa989ab2bd0d37540c812b4270006ff794662e7 introduced the bug.