We are running a survey to help us improve the experience for all of our members. If you see the survey appear, please take the time to tell us about your experience if you can.
Dear all,
Consider a scenario where a bare minimum hardware-assisted hypervisor is hosting 2 simple guest OSes on a dual-core Cortex-A15. I want to accurately measure the time it takes to switch from the guest OS to the hypervisor and then from the hypervisor to the other guest OS. Could you please advise me on a reliable and robust way to measure accurately the aforedescribed world switch time. Please share your experience.
Many thanks.
You could use the generic timer (ARM Cortex-A15 MPCore Processor Technical Reference Manual: Chapter 9. Generic Timer).
Something else to consider is what you class as the switch having finished. For example, the Hypervisor might use "lazy" context switching for the VFP registers. Does that still count as finished?