This has gotten a little messy and disjoint, so I'm definitely in favor of cleaning it up. HE> -#define CIM_RASD_TYPE_PROC 3 HE> -#define CIM_RASD_TYPE_MEM 4 HE> -#define CIM_RASD_TYPE_NET 10 HE> -#define CIM_RASD_TYPE_DISK 17 HE> +#define CIM_TYPE_ALL 0 HE> +#define CIM_RES_TYPE_PROC 3 HE> +#define CIM_RES_TYPE_MEM 4 HE> +#define CIM_RES_TYPE_NET 10 HE> +#define CIM_RES_TYPE_DISK 17 HE> +#define CIM_RES_TYPE_UNKNOWN 1000 HE> + HE> +#define CIM_RASD_TYPE_PROC CIM_RES_TYPE_PROC HE> +#define CIM_RASD_TYPE_MEM CIM_RES_TYPE_MEM HE> +#define CIM_RASD_TYPE_NET CIM_RES_TYPE_NET HE> +#define CIM_RASD_TYPE_DISK CIM_RES_TYPE_DISK HE> + HE> +#define CIM_POOL_TYPE_ALL CIM_TYPE_ALL HE> +#define CIM_POOL_TYPE_PROC CIM_RES_TYPE_PROC HE> +#define CIM_POOL_TYPE_MEM CIM_RES_TYPE_MEM HE> +#define CIM_POOL_TYPE_NET CIM_RES_TYPE_NET HE> +#define CIM_POOL_TYPE_DISK CIM_RES_TYPE_DISK HE> +#define CIM_POOL_TYPE_UNKNOWN CIM_RES_TYPE_UNKNOWN I know that in several places, we assume that a device type is the same as the CIM resource type. Above, you define CIM_RASD_TYPE_foo to CIM_RES_TYPE_foo, presumably to avoid having to touch all the places that use the former. I think to reduce ambiguity and to enforce the fact that these must remain the same, it would be better to just change CIM_RASD_TYPE_foo to CIM_RES_TYPE_foo everywhere. Thoughts? -- Dan Smith IBM Linux Technology Center Open Hypervisor Team email: danms@us.ibm.com