On Wed, Mar 25, 2015 at 14:34:04 +0800, Luyao Huang wrote:

When we set a host not exist nodemask in memory device and then start the vm, qemu will report error.

# virsh start test3 error: Failed to start domain test3 error: internal error: process exited while connecting to monitor: 2015-03-25T01:12:17.205913Z qemu-kvm: -object memory-backend-ram,id=memdimm0 ,size=536870912,host-nodes=1-3,policy=bind: cannot bind memory to host NUMA nodes: Invalid argument

We have some function to check this, and add a check when build memory cmd line will report error more early and clearly. And the check will be done when we start a vm have memory device and hotplug a memory device. The error will be:

# virsh start test3 error: Failed to start domain test3 error: internal error: NUMA node 1 is unavailable

The error message is not ideal. I'll tweak the error code in virNumaNodesetIsAvailable() to VIR_ERR_CONFIG_UNSUPPORTED.

Signed-off-by: Luyao Huang <lhuang@redhat.com> --- src/qemu/qemu_command.c | 2 ++ 1 file changed, 2 insertions(+)

diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index 99a19d6..04c8df7 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -4757,6 +4757,8 @@ qemuBuildMemoryBackendStr(unsigned long long size, }

if (nodemask) { + if (!virNumaNodesetIsAvailable(nodemask)) + goto cleanup; if (virJSONValueObjectAdd(props, "m:host-nodes", nodemask, "S:policy", qemuNumaPolicyTypeToString(mode),

ACK. I'll tweak the commit message though. Peter