KR> + if (((dev->type == CIM_RES_TYPE_PROC) ||
KR> + (dev->type == CIM_RES_TYPE_MEM)) && (STREQ(prefix,
"KVM"))) {
KR> + cu_statusf(_BROKER, &s,
KR> + CMPI_RC_ERR_FAILED,
KR> + "Running guest cannot modify resource type %"
PRIu16,
KR> + dev->type);
KR> + goto out;
KR> + }
Hmm, this means we need to update the provider code when KVM gains
specific features like that. Is there some way we can just ignore
failures to dynamically add/rem/mod resources, or perhaps ignore
ENOTSUPPORTED or something?
--
Dan Smith
IBM Linux Technology Center
Open Hypervisor Team
email: danms(a)us.ibm.com