KVM: handle more completed apfs if possible
Xiao Guangrong [Tue, 2 Nov 2010 09:35:35 +0000 (17:35 +0800)]
If it's no need to inject async #PF to PV guest we can handle
more completed apfs at one time, so we can retry guest #PF
as early as possible

Signed-off-by: Xiao Guangrong <xiaoguangrong@cn.fujitsu.com>
Acked-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

virt/kvm/async_pf.c

index 60df9e0..100c66e 100644 (file)
@@ -124,24 +124,24 @@ void kvm_check_async_pf_completion(struct kvm_vcpu *vcpu)
 {
        struct kvm_async_pf *work;
 
-       if (list_empty_careful(&vcpu->async_pf.done) ||
-           !kvm_arch_can_inject_async_page_present(vcpu))
-               return;
-
-       spin_lock(&vcpu->async_pf.lock);
-       work = list_first_entry(&vcpu->async_pf.done, typeof(*work), link);
-       list_del(&work->link);
-       spin_unlock(&vcpu->async_pf.lock);
+       while (!list_empty_careful(&vcpu->async_pf.done) &&
+             kvm_arch_can_inject_async_page_present(vcpu)) {
+               spin_lock(&vcpu->async_pf.lock);
+               work = list_first_entry(&vcpu->async_pf.done, typeof(*work),
+                                             link);
+               list_del(&work->link);
+               spin_unlock(&vcpu->async_pf.lock);
 
-       if (work->page)
-               kvm_arch_async_page_ready(vcpu, work);
-       kvm_arch_async_page_present(vcpu, work);
+               if (work->page)
+                       kvm_arch_async_page_ready(vcpu, work);
+               kvm_arch_async_page_present(vcpu, work);
 
-       list_del(&work->queue);
-       vcpu->async_pf.queued--;
-       if (work->page)
-               put_page(work->page);
-       kmem_cache_free(async_pf_cache, work);
+               list_del(&work->queue);
+               vcpu->async_pf.queued--;
+               if (work->page)
+                       put_page(work->page);
+               kmem_cache_free(async_pf_cache, work);
+       }
 }
 
 int kvm_setup_async_pf(struct kvm_vcpu *vcpu, gva_t gva, gfn_t gfn,