[RFC PATCH v2 3/3] KVM: arm64: Optimise FPSIMD handling to reduce guest/host thrashing

[Dave Martin]

On Mon, Apr 09, 2018 at 11:48:18AM +0200, Christoffer Dall wrote:

[...]

> > diff --git a/arch/arm64/kvm/hyp/switch.c b/arch/arm64/kvm/hyp/switch.c
> > index 8605e04..797b259 100644
> > --- a/arch/arm64/kvm/hyp/switch.c
> > +++ b/arch/arm64/kvm/hyp/switch.c
> > @@ -27,6 +27,7 @@
> >  #include <asm/kvm_mmu.h>
> >  #include <asm/fpsimd.h>
> >  #include <asm/debug-monitors.h>
> > +#include <asm/thread_info.h>
> >  
> >  static bool __hyp_text __fpsimd_enabled_nvhe(void)
> >  {
> > @@ -47,24 +48,40 @@ bool __hyp_text __fpsimd_enabled(void)
> >  	return __fpsimd_is_enabled()();
> >  }
> >  
> > -static void __hyp_text __activate_traps_vhe(void)
> > +static bool update_fp_enabled(struct kvm_vcpu *vcpu)
> > +{
> > +	if (vcpu->arch.host_thread_info->flags & _TIF_FOREIGN_FPSTATE) {
> > +		vcpu->arch.host_fpsimd_state = NULL;
> 
> I can't see where host_fpsimd_state gets set to anything else than NULL,
> what am I missing?

It's set in kvm_arch_vcpu_load_fp().  Once we enter the run loop the
pointer can only be changed to NULL, because the host state only needs
to be saved once.

Cheers
---Dave