On Mon, Jan 08, 2024 at 20:26:02 +0100, Andrea Bolognani wrote:
The default number of CPU clusters is 1, and values other than that one are currently rejected by all hypervisor drivers.
Signed-off-by: Andrea Bolognani <abologna@redhat.com> --- src/bhyve/bhyve_command.c | 5 +++++ src/conf/cpu_conf.c | 16 +++++++++++++++- src/conf/cpu_conf.h | 1 + src/conf/domain_conf.c | 1 + src/conf/schemas/cputypes.rng | 5 +++++ src/cpu/cpu.c | 1 + src/libxl/libxl_capabilities.c | 1 + src/qemu/qemu_command.c | 5 +++++ src/vmx/vmx.c | 7 +++++++ .../x86_64-host+guest,model486-result.xml | 2 +- .../x86_64-host+guest,models-result.xml | 2 +- tests/cputestdata/x86_64-host+guest-result.xml | 2 +- tests/cputestdata/x86_64-host+guest.xml | 2 +- .../x86_64-host+host-model-nofallback.xml | 2 +- ...host-Haswell-noTSX+Haswell,haswell-result.xml | 2 +- ...aswell-noTSX+Haswell-noTSX,haswell-result.xml | 2 +- ...4-host-Haswell-noTSX+Haswell-noTSX-result.xml | 2 +- .../x86_64-host-worse+guest-result.xml | 2 +- .../ppc64-modern-bulk-result-conf.xml | 2 +- .../ppc64-modern-bulk-result-live.xml | 2 +- .../ppc64-modern-individual-result-conf.xml | 2 +- .../ppc64-modern-individual-result-live.xml | 2 +- .../x86-modern-bulk-result-conf.xml | 2 +- .../x86-modern-bulk-result-live.xml | 2 +- .../x86-modern-individual-add-result-conf.xml | 2 +- .../x86-modern-individual-add-result-live.xml | 2 +- ...e-timeout+graphics-spice-timeout-password.xml | 2 +- .../qemuhotplug-graphics-spice-timeout.xml | 2 +- .../fd-memory-no-numa-topology.xml | 2 +- .../qemuxml2argvdata/fd-memory-numa-topology.xml | 2 +- .../fd-memory-numa-topology2.xml | 2 +- .../fd-memory-numa-topology3.xml | 2 +- tests/qemuxml2argvdata/hugepages-nvdimm.xml | 2 +- .../memfd-memory-default-hugepage.xml | 2 +- tests/qemuxml2argvdata/memfd-memory-numa.xml | 2 +- .../memory-hotplug-nvdimm-access.xml | 2 +- .../memory-hotplug-nvdimm-align.xml | 2 +- .../memory-hotplug-nvdimm-label.xml | 2 +- .../memory-hotplug-nvdimm-pmem.xml | 2 +- .../memory-hotplug-nvdimm-readonly.xml | 2 +- tests/qemuxml2argvdata/memory-hotplug-nvdimm.xml | 2 +- .../memory-hotplug-virtio-mem.xml | 2 +- .../memory-hotplug-virtio-pmem.xml | 2 +- .../cpu-numa-disjoint.x86_64-latest.xml | 2 +- .../cpu-numa-disordered.x86_64-latest.xml | 2 +- .../cpu-numa-memshared.x86_64-latest.xml | 2 +- .../cpu-numa-no-memory-element.x86_64-latest.xml | 2 +- .../cpu-numa1.x86_64-latest.xml | 2 +- .../cpu-numa2.x86_64-latest.xml | 2 +- .../memory-hotplug-dimm-addr.x86_64-latest.xml | 2 +- .../memory-hotplug-dimm.x86_64-latest.xml | 2 +- .../memory-hotplug-multiple.x86_64-latest.xml | 2 +- ...plug-nvdimm-ppc64-abi-update.ppc64-latest.xml | 2 +- .../memory-hotplug-nvdimm-ppc64.ppc64-latest.xml | 2 +- .../memory-hotplug.x86_64-latest.xml | 2 +- ...mad-auto-memory-vcpu-cpuset.x86_64-latest.xml | 2 +- ...cpu-no-cpuset-and-placement.x86_64-latest.xml | 2 +- ...numad-auto-vcpu-no-numatune.x86_64-latest.xml | 2 +- ...mad-static-vcpu-no-numatune.x86_64-latest.xml | 2 +- .../pci-expander-bus.x86_64-latest.xml | 2 +- .../pcie-expander-bus.x86_64-latest.xml | 2 +- .../pseries-phb-numa-node.ppc64-latest.xml | 2 +- tests/vmx2xmldata/esx-in-the-wild-10.xml | 2 +- tests/vmx2xmldata/esx-in-the-wild-8.xml | 2 +- tests/vmx2xmldata/esx-in-the-wild-9.xml | 2 +- 65 files changed, 97 insertions(+), 57 deletions(-)
[...]
diff --git a/src/conf/cpu_conf.h b/src/conf/cpu_conf.h index 3e4c53675c..2694022fed 100644 --- a/src/conf/cpu_conf.h +++ b/src/conf/cpu_conf.h @@ -148,6 +148,7 @@ struct _virCPUDef { unsigned int microcodeVersion; unsigned int sockets; unsigned int dies; + unsigned int clusters; unsigned int cores; unsigned int threads; unsigned int sigFamily;
git grep 'cpu->dies' shows that in src/libxl/libxl_capabilities.c libxlCapsInitCPU 'dies' is assigned to '1' but not in this patch. With the logic to format the cpu element this could cause problems. Reviewed-by: Peter Krempa <pkrempa@redhat.com>