Re: [PATCH v4 1/4] x86/kvm: Add "nopvspin" parameter to disable PV spinlocks

[Zhenzhong Duan]

On 2019/10/7 22:46, Boris Ostrovsky wrote:
> On 10/6/19 3:49 AM, Zhenzhong Duan wrote:
> > On 2019/10/4 22:52, Boris Ostrovsky wrote:
> >
> > > On 10/3/19 10:02 AM, Zhenzhong Duan wrote:
> > > >  Â void __init kvm_spinlock_init(void)
> > > >  Â {
> > > > -ÂÂÂ /* Does host kernel support KVM_FEATURE_PV_UNHALT? */
> > > > -ÂÂÂ if (!kvm_para_has_feature(KVM_FEATURE_PV_UNHALT))
> > > > -ÂÂÂÂÂÂÂ return;
> > > > -
> > > > -ÂÂÂ if (kvm_para_has_hint(KVM_HINTS_REALTIME))
> > > > +ÂÂÂ /*
> > > > +ÂÂÂÂ * Don't use the pvqspinlock code if no KVM_FEATURE_PV_UNHALT
> > > > feature
> > > > +ÂÂÂÂ * support, or there is REALTIME hints or only 1 vCPU.
> > > > +ÂÂÂÂ */
> > > > +ÂÂÂ if (!kvm_para_has_feature(KVM_FEATURE_PV_UNHALT) ||
> > > > +ÂÂÂÂÂÂÂ kvm_para_has_hint(KVM_HINTS_REALTIME) ||
> > > > +ÂÂÂÂÂÂÂ num_possible_cpus() == 1) {
> > > > +ÂÂÂÂÂÂÂ pr_info("PV spinlocks disabled\n");
> > > >  ÂÂÂÂÂÂÂÂÂ return;
> > > > +ÂÂÂ }
> > > >  Â -ÂÂÂ /* Don't use the pvqspinlock code if there is only 1 vCPU. */
> > > > -ÂÂÂ if (num_possible_cpus() == 1)
> > > > +ÂÂÂ if (nopvspin) {
> > > > +ÂÂÂÂÂÂÂ pr_info("PV spinlocks disabled forced by \"nopvspin\"
> > > > parameter.\n");
> > > > +ÂÂÂÂÂÂÂ static_branch_disable(&virt_spin_lock_key);
> > > Would it make sense to bring here the other site where the key is
> > > disabled (in kvm_smp_prepare_cpus())?
> > Thanks for point out, I'll do it. Just not clear if I should do that
> > in a separate patch,
> > there is a history about that code:
> >
> > Its original place was here and then moved to kvm_smp_prepare_cpus()
> > by below commit:
> > 34226b6b ("KVM: X86: Fix setup the virt_spin_lock_key before static
> > key get initialized")
> > which fixed jump_label_init() calling late issue.
> >
> > Then 8990cac6 ("x86/jump_label: Initialize static branching early")
> > move jump_label_init()
> > early, so commit 34226b6b could be reverted.
> Which is similar to what you did earlier for Xen.

You remember that, ok, I'll do the same for KVM.

Thanks

Zhenzhong