Re: [Xen-devel] [PATCH 3/4] VMX: Add posted interrupt supporting

["Zhang, Yang Z" <yang.z.zhang@xxxxxxxxx>]

Keir Fraser wrote on 2013-04-09:
> On 09/04/2013 07:01, "Yang Zhang" <yang.z.zhang@xxxxxxxxx> wrote:
> 
>> From: Yang Zhang <yang.z.zhang@xxxxxxxxx>
>> 
>> Add the supporting of using posted interrupt to deliver interrupt.
>> 
>> Signed-off-by: Yang Zhang <yang.z.zhang@xxxxxxxxx>
>> Reviewed-by: Jun Nakajima <jun.nakajima@xxxxxxxxx>
>> ---
> 
> ...
> 
>> +static void __vmx_deliver_posted_interrupt(struct vcpu *v)
>> +{
>> +    bool_t running;
>> +
>> +    running = v->is_running;
>> +    vcpu_unblock(v);
>> +    if ( running && (in_irq() || (v != current)) )
>> +    {
>> +        unsigned int cpu = v->processor;
>> +
>> +        if ( !test_and_set_bit(VCPU_KICK_SOFTIRQ, &softirq_pending(cpu))
>> +                && (cpu != smp_processor_id()) )
>> +            send_IPI_mask(cpumask_of(cpu),
> POSTED_INTERRUPT_VECTOR);
> 
> I don't think you need to tickle VCPU_KICK_SOFTIRQ here? You aren't
> synchronising with vmx_intr_assist() here, only notifying the processor
> itself of pending virtual interrupts. All that requires is an IPI of
> POSTED_INTERRUPT_VECTOR in some cases.
> 
> I suggest:
>     if ( running )
>     {
>         unsigned int cpu = v->processor;
>         if ( cpu != smp_processor_id() )
>             send_IPI_mask(...);
>     }
> That would just work, right? And avoids needing to duplicate some of the
> trickier logic in vcpu_kick().
If posted interrupt arrived after vmx_intr_assit() and before vmentry, then the 
interrupt cannot be inject to guest in time.
vmx_asm_do_vmentry:
        call vmx_intr_assist
        call nvmx_switch_guest
        ASSERT_NOT_IN_ATOMIC
                              <---------if posted interrupt arrived here and we 
don't set softirq, then the interrupt will be delay until next vmentry.
        GET_CURRENT(%rbx)
        cli

        mov  VCPU_processor(%rbx),%eax
        shl  $IRQSTAT_shift,%eax
        lea  irq_stat+IRQSTAT_softirq_pending(%rip),%rdx
        cmpl $0,(%rdx,%rax,1)
        jnz  .Lvmx_process_softirqs

Actually, the posted interrupt is same with a normal vcpu kick IPI, except that 
if cpu received it in non-root mode, it will not cause vmext. So we must follow 
the vcpu_kick logic.

> 
> A code comment to say that this is a simplified form of vcpu_kick() is
> probably worthwhile too. There are useful code comments in vcpu_kick()
> regarding why things are ordered as they are.


Best regards,
Yang



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel