Comment 6 for bug 1813165

I have bisected the kernel and ended with first bad commit:

commit 14c07ad89f4d728a468caaea6a769c018c2b8dd6
Author: Vitaly Kuznetsov <email address hidden>
Date: Mon Oct 8 21:28:08 2018 +0200

    x86/kvm/mmu: introduce guest_mmu

    When EPT is used for nested guest we need to re-init MMU as shadow
    EPT MMU (nested_ept_init_mmu_context() does that). When we return back
    from L2 to L1 kvm_mmu_reset_context() in nested_vmx_load_cr3() resets
    MMU back to normal TDP mode. Add a special 'guest_mmu' so we can use
    separate root caches; the improved hit rate is not very important for
    single vCPU performance, but it avoids contention on the mmu_lock for
    many vCPUs.

It seems to me this is some timing issue. If a L2 vm fails to start I can get it running with a few hard reboots.

Also changing the L2 config helps, -net none, -nodefaults and -vga cirrus let the vm start most of the time.