KVM: Drop FOLL_GET in GUP when doing async page fault

Page pinning is not mandatory in kvm async page fault processing since
after async page fault event is delivered to a guest it accesses page once
again and does its own GUP.  Drop the FOLL_GET flag in GUP in async_pf
code, and do some simplifying in check/clear processing.

Suggested-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Gu zheng <guz.fnst@cn.fujitsu.com>
Signed-off-by: chai wen <chaiw.fnst@cn.fujitsu.com>
Signed-off-by: Gleb Natapov <gleb@redhat.com>

arch/x86/kvm/x86.c

@@ -7298,7 +7298,7 @@ void kvm_arch_async_page_ready(struct kvm_vcpu *vcpu, struct kvm_async_pf *work)
 	int r;
 
 	if ((vcpu->arch.mmu.direct_map != work->arch.direct_map) ||
-	      is_error_page(work->page))
+	      work->wakeup_all)
 		return;
 
 	r = kvm_mmu_reload(vcpu);
@@ -7408,7 +7408,7 @@ void kvm_arch_async_page_present(struct kvm_vcpu *vcpu,
 	struct x86_exception fault;
 
 	trace_kvm_async_pf_ready(work->arch.token, work->gva);
-	if (is_error_page(work->page))
+	if (work->wakeup_all)
 		work->arch.token = ~0; /* broadcast wakeup */
 	else
 		kvm_del_async_pf_gfn(vcpu, work->arch.gfn);

include/linux/kvm_host.h

@@ -189,7 +189,7 @@ struct kvm_async_pf {
 	gva_t gva;
 	unsigned long addr;
 	struct kvm_arch_async_pf arch;
-	struct page *page;
+	bool   wakeup_all;
 };
 
 void kvm_clear_async_pf_completion_queue(struct kvm_vcpu *vcpu);

include/trace/events/kvm.h

@@ -296,23 +296,21 @@ DEFINE_EVENT(kvm_async_pf_nopresent_ready, kvm_async_pf_ready,
 
 TRACE_EVENT(
 	kvm_async_pf_completed,
-	TP_PROTO(unsigned long address, struct page *page, u64 gva),
-	TP_ARGS(address, page, gva),
+	TP_PROTO(unsigned long address, u64 gva),
+	TP_ARGS(address, gva),
 
 	TP_STRUCT__entry(
 		__field(unsigned long, address)
-		__field(pfn_t, pfn)
 		__field(u64, gva)
 		),
 
 	TP_fast_assign(
 		__entry->address = address;
-		__entry->pfn = page ? page_to_pfn(page) : 0;
 		__entry->gva = gva;
 		),
 
-	TP_printk("gva %#llx address %#lx pfn %#llx",  __entry->gva,
-		  __entry->address, __entry->pfn)
+	TP_printk("gva %#llx address %#lx",  __entry->gva,
+		  __entry->address)
 );
 
 #endif

virt/kvm/async_pf.c

@@ -56,7 +56,6 @@ void kvm_async_pf_vcpu_init(struct kvm_vcpu *vcpu)
 
 static void async_pf_execute(struct work_struct *work)
 {
-	struct page *page = NULL;
 	struct kvm_async_pf *apf =
 		container_of(work, struct kvm_async_pf, work);
 	struct mm_struct *mm = apf->mm;
@@ -68,21 +67,20 @@ static void async_pf_execute(struct work_struct *work)
 
 	use_mm(mm);
 	down_read(&mm->mmap_sem);
-	get_user_pages(current, mm, addr, 1, 1, 0, &page, NULL);
+	get_user_pages(current, mm, addr, 1, 1, 0, NULL, NULL);
 	up_read(&mm->mmap_sem);
 	unuse_mm(mm);
 
 	spin_lock(&vcpu->async_pf.lock);
 	list_add_tail(&apf->link, &vcpu->async_pf.done);
-	apf->page = page;
 	spin_unlock(&vcpu->async_pf.lock);
 
 	/*
 	 * apf may be freed by kvm_check_async_pf_completion() after
 	 * this point
 	 */
 
-	trace_kvm_async_pf_completed(addr, page, gva);
+	trace_kvm_async_pf_completed(addr, gva);
 
 	if (waitqueue_active(&vcpu->wq))
 		wake_up_interruptible(&vcpu->wq);
@@ -112,8 +110,6 @@ void kvm_clear_async_pf_completion_queue(struct kvm_vcpu *vcpu)
 			list_entry(vcpu->async_pf.done.next,
 				   typeof(*work), link);
 		list_del(&work->link);
-		if (!is_error_page(work->page))
-			kvm_release_page_clean(work->page);
 		kmem_cache_free(async_pf_cache, work);
 	}
 	spin_unlock(&vcpu->async_pf.lock);
@@ -133,14 +129,11 @@ void kvm_check_async_pf_completion(struct kvm_vcpu *vcpu)
 		list_del(&work->link);
 		spin_unlock(&vcpu->async_pf.lock);
 
-		if (work->page)
-			kvm_arch_async_page_ready(vcpu, work);
+		kvm_arch_async_page_ready(vcpu, work);
 		kvm_arch_async_page_present(vcpu, work);
 
 		list_del(&work->queue);
 		vcpu->async_pf.queued--;
-		if (!is_error_page(work->page))
-			kvm_release_page_clean(work->page);
 		kmem_cache_free(async_pf_cache, work);
 	}
 }
@@ -163,7 +156,7 @@ int kvm_setup_async_pf(struct kvm_vcpu *vcpu, gva_t gva, gfn_t gfn,
 	if (!work)
 		return 0;
 
-	work->page = NULL;
+	work->wakeup_all = false;
 	work->vcpu = vcpu;
 	work->gva = gva;
 	work->addr = gfn_to_hva(vcpu->kvm, gfn);
@@ -203,7 +196,7 @@ int kvm_async_pf_wakeup_all(struct kvm_vcpu *vcpu)
 	if (!work)
 		return -ENOMEM;
 
-	work->page = KVM_ERR_PTR_BAD_PAGE;
+	work->wakeup_all = true;
 	INIT_LIST_HEAD(&work->queue); /* for list_del to work */
 
 	spin_lock(&vcpu->async_pf.lock);