Re: [RFC PATCH v4 11/36] i386/tdx: Initialize TDX before creating TD vcpus

[Xiaoyao Li]

On 5/23/2022 5:20 PM, Gerd Hoffmann wrote:
> > +int tdx_pre_create_vcpu(CPUState *cpu)
> > +{
> > +    MachineState *ms = MACHINE(qdev_get_machine());
> > +    X86CPU *x86cpu = X86_CPU(cpu);
> > +    CPUX86State *env = &x86cpu->env;
> > +    struct kvm_tdx_init_vm init_vm;
> > +    int r = 0;
> > +
> > +    qemu_mutex_lock(&tdx_guest->lock);
> > +    if (tdx_guest->initialized) {
> > +        goto out;
> > +    }
> > +
> > +    memset(&init_vm, 0, sizeof(init_vm));
> > +    init_vm.cpuid.nent = kvm_x86_arch_cpuid(env, init_vm.entries, 0);
> > +
> > +    init_vm.attributes = tdx_guest->attributes;
> > +    init_vm.max_vcpus = ms->smp.cpus;
> > +
> > +    r = tdx_vm_ioctl(KVM_TDX_INIT_VM, 0, &init_vm);
> > +    if (r < 0) {
> > +        error_report("KVM_TDX_INIT_VM failed %s", strerror(-r));
> > +        goto out;
> > +    }
> > +
> > +    tdx_guest->initialized = true;
> > +
> > +out:
> > +    qemu_mutex_unlock(&tdx_guest->lock);
> > +    return r;
> > +}
>
> Hmm, hooking *vm* initialization into *vcpu* creation looks wrong to me.

That's because for TDX, it has to do VM-scope (feature) initialization 
before creating vcpu. This is new to KVM and QEMU, that every feature is 
vcpu-scope and configured per-vcpu before.

To minimize the change to QEMU, we want to utilize @cpu and @cpu->env to 
grab the configuration info. That's why it goes this way.

Do you have any better idea on it?

> take care,
>    Gerd