Re: [kvm-devel] [patch 2/2] KVM: fix kvm_vcpu_kick vs __vcpu_run race

[Avi K. <av...@qu...>]

Marcelo Tosatti wrote:
> On Fri, Apr 11, 2008 at 03:18:19PM +0300, Avi Kivity wrote:
>   
>> kvm_vcpu_kick() can be called from nonatomic contexts, so the vcpu->cpu 
>> == cpu check is dangerous (and will warn on preemptible kernels, no?)
>>     
>
> Doh, right. How's this.
>
> -----------
>
> KVM: fix kvm_vcpu_kick vs __vcpu_run race
>
> There is a window open between testing of pending IRQ's 
> and assignment of guest_mode in __vcpu_run.
>
> Injection of IRQ's can race with __vcpu_run as follows:
>
> CPU0                                CPU1
> kvm_x86_ops->run()
> vcpu->guest_mode = 0                SET_IRQ_LINE ioctl
> ..
> kvm_x86_ops->inject_pending_irq     
> kvm_cpu_has_interrupt()
>
>                                     apic_test_and_set_irr()
>                                     kvm_vcpu_kick
>                                     if (vcpu->guest_mode)
>                                         send_ipi()
>                                     
> vcpu->guest_mode = 1
>
> So move guest_mode=1 assignment before ->inject_pending_irq, and make
> sure that it won't reorder after it.
>
>
>   
Applied, but this

> @@ -3944,11 +3950,13 @@ static void vcpu_kick_intr(void *info)
>  void kvm_vcpu_kick(struct kvm_vcpu *vcpu)
>  {
>  	int ipi_pcpu = vcpu->cpu;
> +	int cpu = get_cpu();
>  
>  	if (waitqueue_active(&vcpu->wq)) {
>  		wake_up_interruptible(&vcpu->wq);
>  		++vcpu->stat.halt_wakeup;
>  	}
> -	if (vcpu->guest_mode)
> +	if (vcpu->guest_mode && vcpu->cpu != cpu)
>  		smp_call_function_single(ipi_pcpu, vcpu_kick_intr, vcpu, 0, 0);
> +	put_cpu();
>  }
>   

Looks like a no-op now, as vcpu_kick_intr() does nothing and 
smp_call_function_single() won't force an exit if vcpu->cpu == cpu, so I 
dropped this hunk.

-- 
error compiling committee.c: too many arguments to function