Describe the high level behaviors of CPU affinity for IO. E.g. when a guest issues IO on VP n, what VP does storvsp, the disk backends, and the device backend see? What about for IDE or NVMe frontend storage? When does this fall down, and what's the behavior?
Ideally this would come after a document describing the OpenHCL executor and CPU model.
Describe the high level behaviors of CPU affinity for IO. E.g. when a guest issues IO on VP n, what VP does storvsp, the disk backends, and the device backend see? What about for IDE or NVMe frontend storage? When does this fall down, and what's the behavior?
Ideally this would come after a document describing the OpenHCL executor and CPU model.