Re: [PATCH 2/3] KVM: X86: implement the logic for spinlock optimization

[Paolo Bonzini <pbonzini@redhat.com>]

On 07/08/2017 14:28, Longpeng(Mike) wrote:
> * kvm_arch_spin_in_kernel() returns whether the vcpu (which exits due to
> spinlock) is CPL=0. It only be called by kvm_vcpu_on_spin(), and the
> input vcpu is 'me' which get a PAUSE exit now. *
> 
> I split kvm_arch_vcpu_in_kernel(in RFC) into two functions:
> kvm_arch_spin_in_kernel and kvm_arch_preempt_in_kernel
> 
> Because of KVM/VMX L1 never set CPU_BASED_PAUSE_EXITING and only set
> SECONDARY_EXEC_PAUSE_LOOP_EXITING if supported, so for L1:

I understand better now.  I think vmx.c should just return true from
vmx_spin_in_kernel.  However, kvm_arch_vcpu_spin_in_kernel is not
necessary.  Instead you should make "in_kern" an argument to
kvm_vcpu_on_spin (maybe renamed to "yield_to_kernel_mode_vcpu").

Then vmx.c can just call "kvm_vcpu_on_spin(vcpu, true)".

> 1. get a PAUSE exit with CPL=0 if PLE is supported
> 2. never get a PAUSE exit if don't support PLE
> 
> So, I think it can direct return true(CPL=0) if supports PLE.
> 
> But for nested KVM/VMX(I'm not familiar with nested), it could set
> CPU_BASED_PAUSE_EXITING, so I think get_cpl() is also needed.

If the nested hypervisor sets CPU_BASED_PAUSE_EXITING, a PAUSE vmexit
while running a nested guest would be reflected to the nested
hypervisor.  So you wouldn't get to handle_pause and thus to
kvm_vcpu_on_spin.

Thanks,

Paolo