KVM Forum 2021

In virtualized environments, oversubscribing virtual CPUs (vCPUs) on physical CPUs (pCPUs) is common to utilize CPU resources efficiently. Unfortunately, excessive vCPU spinning, which occurs when a vCPU is waiting in a spin loop for an event from a descheduled vCPU, causes serious performance degradation. VM-agnostic KVM tries to prevent excessive vCPU spinning by rescheduling vCPUs when an excessive spin is detected by pause-loop-exiting. This talk explains how the KVM vCPU scheduler fails to avoid excessive vCPU spinning in many opportunities. Three problems have been identified: 1) scheduler mismatch, 2) lost opportunity, and 3) overboost. The first problem comes from the mismatch between the KVM vCPU scheduler and the Linux scheduler. The second and third problems come from an inefficient algorithm for choosing the next candidate vCPU to be scheduled. Simple modifications gracefully resolve the problems and the performance improves by up to 80 %. Results imply the VM-agnostic hypervisor can resolve excessive vCPU spinning more gracefully than previously believed.