A guest running with KVM PR ends up irresponsive after migration most of the
time. This happens because the HPT allocated by QEMU is likely to have a
different address on the destination than it had on the source, but we push
the source address to KVM.
This series does a little cleanup and fixes the issue. I could successfully
test it with a nested setup (KVM PR running in KVM HV).
However, this isn't enough to fix migration when using KVM PR on baremetal...
CPUs seem to end up looping on H_CEDE in the guest. I can't figure out what's
happening... Any suggestion would be appreciated.
Cheers,
--
Greg
---
Greg Kurz (4):
spapr: only update SDR1 once per-cpu during CAS
spapr: introduce a helper to compute the address of the HPT
ppc: kvm: introduce a helper to update SDR1 for a single CPU
ppc: kvm: update HPT pointer in KVM PR after migration
hw/ppc/spapr.c | 15 +++++++++++++++
hw/ppc/spapr_cpu_core.c | 8 +++-----
hw/ppc/spapr_hcall.c | 20 +++++++++-----------
include/hw/ppc/spapr.h | 1 +
target/ppc/cpu.h | 1 +
target/ppc/kvm.c | 12 ++++++++----
target/ppc/kvm_ppc.h | 3 ++-
target/ppc/machine.c | 7 +++++++
8 files changed, 46 insertions(+), 21 deletions(-)
