From: Wanpeng Li Date: Tue, 18 May 2021 12:00:32 +0000 (-0700) Subject: KVM: X86: Bail out of direct yield in case of under-committed scenarios X-Git-Url: http://git.maquefel.me/?a=commitdiff_plain;h=72b268a8e9307a1757f61af080e990b5baa11d2a;p=linux.git KVM: X86: Bail out of direct yield in case of under-committed scenarios In case of under-committed scenarios, vCPUs can be scheduled easily; kvm_vcpu_yield_to adds extra overhead, and it is also common to see when vcpu->ready is true but yield later failing due to p->state is TASK_RUNNING. Let's bail out in such scenarios by checking the length of current cpu runqueue, which can be treated as a hint of under-committed instead of guarantee of accuracy. 30%+ of directed-yield attempts can now avoid the expensive lookups in kvm_sched_yield() in an under-committed scenario. Signed-off-by: Wanpeng Li Message-Id: <1621339235-11131-2-git-send-email-wanpengli@tencent.com> Signed-off-by: Paolo Bonzini --- diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 9b6bca6169291..dfb7c320581fc 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -8360,6 +8360,9 @@ static void kvm_sched_yield(struct kvm_vcpu *vcpu, unsigned long dest_id) vcpu->stat.directed_yield_attempted++; + if (single_task_running()) + goto no_yield; + rcu_read_lock(); map = rcu_dereference(vcpu->kvm->arch.apic_map);