Re: [PATCH 3/5] KVM: add KVM_USER_EXIT vm ioctl for userspace exit

[Paolo Bonzini <pbonzini@redhat.com>]

On 05/08/2015 15:34, Radim Krčmář wrote:
> 2015-08-05 15:29+0200, Paolo Bonzini:
>> On 05/08/2015 15:21, Radim Krčmář wrote:
>>> +	kvm_for_each_vcpu(idx, vcpu, kvm)
>>> +		if (vcpu->vcpu_id == vcpu_id) {
>>> +			kvm_make_request(KVM_REQ_EXIT, vcpu);
>>> +			kvm_vcpu_kick(vcpu);
>>> +
>>> +			return 0;
>>> +		}
>>> +
>>
>> Why not a vcpu ioctl?  kvm_for_each_vcpu can become pretty expensive if
>> you have a few dozen VCPUs.
> 
> Yeah, it will be slow, the only bright side is low frequency of calls.
> 
> vcpu ioctl should only be issued by the vcpu thread so it would
> significantly limit use.

That's a general limitation, but you can lift it for particular ioctls.

See in particular this:

#if defined(CONFIG_S390) || defined(CONFIG_PPC) || defined(CONFIG_MIPS)
        /*
         * Special cases: vcpu ioctls that are asynchronous to vcpu execution,
         * so vcpu_load() would break it.
         */
        if (ioctl == KVM_S390_INTERRUPT || ioctl == KVM_S390_IRQ || ioctl == KVM_INTERRUPT)
                return kvm_arch_vcpu_ioctl(filp, ioctl, arg);
#endif

You can add an "if (ioctl == KVM_USER_EXIT)" before.

Paolo