Re: [PATCH 1/5] KVM: simplify branch check in host poll code

[Zhenzhong Duan]

On 2019/11/2 5:03, Marcelo Tosatti wrote:
> On Sat, Oct 26, 2019 at 11:23:55AM +0800, Zhenzhong Duan wrote:
> > Remove redundant check.
> >
> > Signed-off-by: Zhenzhong Duan <zhenzhong.duan@xxxxxxxxxx>
> > ---
> >   virt/kvm/kvm_main.c | 9 ++++-----
> >   1 file changed, 4 insertions(+), 5 deletions(-)
> >
> > diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
> > index 67ef3f2..2ca2979 100644
> > --- a/virt/kvm/kvm_main.c
> > +++ b/virt/kvm/kvm_main.c
> > @@ -2366,13 +2366,12 @@ void kvm_vcpu_block(struct kvm_vcpu *vcpu)
> >   		} else if (halt_poll_ns) {
> >   			if (block_ns <= vcpu->halt_poll_ns)
> >   				;
> > -			/* we had a long block, shrink polling */
> > -			else if (vcpu->halt_poll_ns && block_ns > halt_poll_ns)
> > -				shrink_halt_poll_ns(vcpu);
> >   			/* we had a short halt and our poll time is too small */
> > -			else if (vcpu->halt_poll_ns < halt_poll_ns &&
> This is not a redundant check: it avoids from calling
> into grow_halt_poll_ns, which will do:
>
> 	1) Multiplication
> 	2) Cap that back to halt_poll_ns
> 	3) Invoke the trace_kvm_halt_poll_ns_grow tracepoint
> 	   (when in fact vcpu->halt_poll_ns did not grow).

In this branch, vcpu->halt_poll_ns < block_ns is true, and if block_ns < 
halt_poll_ns,

then vcpu->halt_poll_ns < halt_poll_ns is always true, isn't it?


I realized I ignored the situation that halt_poll_ns and 
halt_poll_ns_grow could be

updated at runtime, so pls ignore this patch, I'll fix it by following 
the guest haltpoll code.

Thanks

Zhenzhong