Something I've noticed when closing #233.
Related: #115
For some reasons, if the VMM is at a higher priority than the VCPUs,
the VCPUs will never make meaningful progress when Linux tries to bring
up secondary CPUs. Causing the guest to hang.
My theory is that our hacky WFI implementation is causing problem, i.e.
the VMM is starving the VCPUs of execution time somehow. Because a WFI
fault will result in an immediate resumption.
This needs to be fixed properly.
Revert be86159 when closing.
Something I've noticed when closing #233.
Related: #115
For some reasons, if the VMM is at a higher priority than the VCPUs,
the VCPUs will never make meaningful progress when Linux tries to bring
up secondary CPUs. Causing the guest to hang.
My theory is that our hacky WFI implementation is causing problem, i.e.
the VMM is starving the VCPUs of execution time somehow. Because a WFI
fault will result in an immediate resumption.
This needs to be fixed properly.
Revert be86159 when closing.