KVM: Halt vcpu if page it tries to access is swapped out

If a guest accesses swapped out memory do not swap it in from vcpu thread
context. Schedule work to do swapping and put vcpu into halted state
instead.

Interrupts will still be delivered to the guest and if interrupt will
cause reschedule guest will continue to run another task.

[avi: remove call to get_user_pages_noio(), nacked by Linus; this
      makes everything synchrnous again]

Acked-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

4 files changed, 266 insertions, 11 deletions

diff --git a/virt/kvm/Kconfig b/virt/kvm/Kconfig
index 7f1178f6b839..f63ccb0a5982 100644
--- a/virt/kvm/Kconfig
+++ b/virt/kvm/Kconfig
@@ -15,3 +15,6 @@ config KVM_APIC_ARCHITECTURE
 
 config KVM_MMIO
        bool
+
+config KVM_ASYNC_PF
+       bool
diff --git a/virt/kvm/async_pf.c b/virt/kvm/async_pf.c
new file mode 100644
index 000000000000..857d63431cb7
--- /dev/null
+++ b/virt/kvm/async_pf.c
@@ -0,0 +1,190 @@
+/*
+ * kvm asynchronous fault support
+ *
+ * Copyright 2010 Red Hat, Inc.
+ *
+ * Author:
+ *      Gleb Natapov <gleb@redhat.com>
+ *
+ * This file is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software Foundation,
+ * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA.
+ */
+
+#include <linux/kvm_host.h>
+#include <linux/slab.h>
+#include <linux/module.h>
+#include <linux/mmu_context.h>
+
+#include "async_pf.h"
+#include <trace/events/kvm.h>
+
+static struct kmem_cache *async_pf_cache;
+
+int kvm_async_pf_init(void)
+{
+        async_pf_cache = KMEM_CACHE(kvm_async_pf, 0);
+
+        if (!async_pf_cache)
+                return -ENOMEM;
+
+        return 0;
+}
+
+void kvm_async_pf_deinit(void)
+{
+        if (async_pf_cache)
+                kmem_cache_destroy(async_pf_cache);
+        async_pf_cache = NULL;
+}
+
+void kvm_async_pf_vcpu_init(struct kvm_vcpu *vcpu)
+{
+        INIT_LIST_HEAD(&vcpu->async_pf.done);
+        INIT_LIST_HEAD(&vcpu->async_pf.queue);
+        spin_lock_init(&vcpu->async_pf.lock);
+}
+
+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;
+        struct kvm_vcpu *vcpu = apf->vcpu;
+        unsigned long addr = apf->addr;
+        gva_t gva = apf->gva;
+
+        might_sleep();
+
+        use_mm(mm);
+        down_read(&mm->mmap_sem);
+        get_user_pages(current, mm, addr, 1, 1, 0, &page, 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;
+        apf->done = true;
+        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);
+
+        if (waitqueue_active(&vcpu->wq))
+                wake_up_interruptible(&vcpu->wq);
+
+        mmdrop(mm);
+        kvm_put_kvm(vcpu->kvm);
+}
+
+void kvm_clear_async_pf_completion_queue(struct kvm_vcpu *vcpu)
+{
+        /* cancel outstanding work queue item */
+        while (!list_empty(&vcpu->async_pf.queue)) {
+                struct kvm_async_pf *work =
+                        list_entry(vcpu->async_pf.queue.next,
+                                   typeof(*work), queue);
+                cancel_work_sync(&work->work);
+                list_del(&work->queue);
+                if (!work->done) /* work was canceled */
+                        kmem_cache_free(async_pf_cache, work);
+        }
+
+        spin_lock(&vcpu->async_pf.lock);
+        while (!list_empty(&vcpu->async_pf.done)) {
+                struct kvm_async_pf *work =
+                        list_entry(vcpu->async_pf.done.next,
+                                   typeof(*work), link);
+                list_del(&work->link);
+                if (work->page)
+                        put_page(work->page);
+                kmem_cache_free(async_pf_cache, work);
+        }
+        spin_unlock(&vcpu->async_pf.lock);
+
+        vcpu->async_pf.queued = 0;
+}
+
+void kvm_check_async_pf_completion(struct kvm_vcpu *vcpu)
+{
+        struct kvm_async_pf *work;
+
+        if (list_empty_careful(&vcpu->async_pf.done))
+                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);
+
+        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);
+}
+
+int kvm_setup_async_pf(struct kvm_vcpu *vcpu, gva_t gva, gfn_t gfn,
+                       struct kvm_arch_async_pf *arch)
+{
+        struct kvm_async_pf *work;
+
+        if (vcpu->async_pf.queued >= ASYNC_PF_PER_VCPU)
+                return 0;
+
+        /* setup delayed work */
+
+        /*
+         * do alloc nowait since if we are going to sleep anyway we
+         * may as well sleep faulting in page
+         */
+        work = kmem_cache_zalloc(async_pf_cache, GFP_NOWAIT);
+        if (!work)
+                return 0;
+
+        work->page = NULL;
+        work->done = false;
+        work->vcpu = vcpu;
+        work->gva = gva;
+        work->addr = gfn_to_hva(vcpu->kvm, gfn);
+        work->arch = *arch;
+        work->mm = current->mm;
+        atomic_inc(&work->mm->mm_count);
+        kvm_get_kvm(work->vcpu->kvm);
+
+        /* this can't really happen otherwise gfn_to_pfn_async
+           would succeed */
+        if (unlikely(kvm_is_error_hva(work->addr)))
+                goto retry_sync;
+
+        INIT_WORK(&work->work, async_pf_execute);
+        if (!schedule_work(&work->work))
+                goto retry_sync;
+
+        list_add_tail(&work->queue, &vcpu->async_pf.queue);
+        vcpu->async_pf.queued++;
+        kvm_arch_async_page_not_present(vcpu, work);
+        return 1;
+retry_sync:
+        kvm_put_kvm(work->vcpu->kvm);
+        mmdrop(work->mm);
+        kmem_cache_free(async_pf_cache, work);
+        return 0;
+}
diff --git a/virt/kvm/async_pf.h b/virt/kvm/async_pf.h
new file mode 100644
index 000000000000..e7ef6447cb82
--- /dev/null
+++ b/virt/kvm/async_pf.h
@@ -0,0 +1,36 @@
+/*
+ * kvm asynchronous fault support
+ *
+ * Copyright 2010 Red Hat, Inc.
+ *
+ * Author:
+ *      Gleb Natapov <gleb@redhat.com>
+ *
+ * This file is free software; you can redistribute it and/or modify
+ * it under the terms of version 2 of the GNU General Public License
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software Foundation,
+ * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA.
+ */
+
+#ifndef __KVM_ASYNC_PF_H__
+#define __KVM_ASYNC_PF_H__
+
+#ifdef CONFIG_KVM_ASYNC_PF
+int kvm_async_pf_init(void);
+void kvm_async_pf_deinit(void);
+void kvm_async_pf_vcpu_init(struct kvm_vcpu *vcpu);
+#else
+#define kvm_async_pf_init() (0)
+#define kvm_async_pf_deinit() do{}while(0)
+#define kvm_async_pf_vcpu_init(C) do{}while(0)
+#endif
+
+#endif
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 5225052aebc1..75fd590c0214 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -55,6 +55,7 @@
 #include <asm-generic/bitops/le.h>
 
 #include "coalesced_mmio.h"
+#include "async_pf.h"
 
 #define CREATE_TRACE_POINTS
 #include <trace/events/kvm.h>
@@ -186,6 +187,7 @@ int kvm_vcpu_init(struct kvm_vcpu *vcpu, struct kvm *kvm, unsigned id)
         vcpu->kvm = kvm;
         vcpu->vcpu_id = id;
         init_waitqueue_head(&vcpu->wq);
+        kvm_async_pf_vcpu_init(vcpu);
 
         page = alloc_page(GFP_KERNEL | __GFP_ZERO);
         if (!page) {
@@ -946,15 +948,20 @@ unsigned long gfn_to_hva(struct kvm *kvm, gfn_t gfn)
 }
 EXPORT_SYMBOL_GPL(gfn_to_hva);
 
-static pfn_t hva_to_pfn(struct kvm *kvm, unsigned long addr, bool atomic)
+static pfn_t hva_to_pfn(struct kvm *kvm, unsigned long addr, bool atomic,
+                        bool *async)
 {
         struct page *page[1];
-        int npages;
+        int npages = 0;
         pfn_t pfn;
 
-        if (atomic)
+        /* we can do it either atomically or asynchronously, not both */
+        BUG_ON(atomic && async);
+
+        if (atomic || async)
                 npages = __get_user_pages_fast(addr, 1, 1, page);
-        else {
+
+        if (unlikely(npages != 1) && !atomic) {
                 might_sleep();
                 npages = get_user_pages_fast(addr, 1, 1, page);
         }
@@ -976,6 +983,9 @@ static pfn_t hva_to_pfn(struct kvm *kvm, unsigned long addr, bool atomic)
 
                 if (vma == NULL || addr < vma->vm_start ||
                     !(vma->vm_flags & VM_PFNMAP)) {
+                        if (async && !(vma->vm_flags & VM_PFNMAP) &&
+                            (vma->vm_flags & VM_WRITE))
+                                *async = true;
                         up_read(&current->mm->mmap_sem);
 return_fault_page:
                         get_page(fault_page);
@@ -993,32 +1003,41 @@ return_fault_page:
 
 pfn_t hva_to_pfn_atomic(struct kvm *kvm, unsigned long addr)
 {
-        return hva_to_pfn(kvm, addr, true);
+        return hva_to_pfn(kvm, addr, true, NULL);
 }
 EXPORT_SYMBOL_GPL(hva_to_pfn_atomic);
 
-static pfn_t __gfn_to_pfn(struct kvm *kvm, gfn_t gfn, bool atomic)
+static pfn_t __gfn_to_pfn(struct kvm *kvm, gfn_t gfn, bool atomic, bool *async)
 {
         unsigned long addr;
 
+        if (async)
+                *async = false;
+
         addr = gfn_to_hva(kvm, gfn);
         if (kvm_is_error_hva(addr)) {
                 get_page(bad_page);
                 return page_to_pfn(bad_page);
         }
 
-        return hva_to_pfn(kvm, addr, atomic);
+        return hva_to_pfn(kvm, addr, atomic, async);
 }
 
 pfn_t gfn_to_pfn_atomic(struct kvm *kvm, gfn_t gfn)
 {
-        return __gfn_to_pfn(kvm, gfn, true);
+        return __gfn_to_pfn(kvm, gfn, true, NULL);
 }
 EXPORT_SYMBOL_GPL(gfn_to_pfn_atomic);
 
+pfn_t gfn_to_pfn_async(struct kvm *kvm, gfn_t gfn, bool *async)
+{
+        return __gfn_to_pfn(kvm, gfn, false, async);
+}
+EXPORT_SYMBOL_GPL(gfn_to_pfn_async);
+
 pfn_t gfn_to_pfn(struct kvm *kvm, gfn_t gfn)
 {
-        return __gfn_to_pfn(kvm, gfn, false);
+        return __gfn_to_pfn(kvm, gfn, false, NULL);
 }
 EXPORT_SYMBOL_GPL(gfn_to_pfn);
 
@@ -1026,7 +1045,7 @@ pfn_t gfn_to_pfn_memslot(struct kvm *kvm,
                          struct kvm_memory_slot *slot, gfn_t gfn)
 {
         unsigned long addr = gfn_to_hva_memslot(slot, gfn);
-        return hva_to_pfn(kvm, addr, false);
+        return hva_to_pfn(kvm, addr, false, NULL);
 }
 
 int gfn_to_page_many_atomic(struct kvm *kvm, gfn_t gfn, struct page **pages,
@@ -2336,6 +2355,10 @@ int kvm_init(void *opaque, unsigned vcpu_size, unsigned vcpu_align,
                 goto out_free_5;
         }
 
+        r = kvm_async_pf_init();
+        if (r)
+                goto out_free;
+
         kvm_chardev_ops.owner = module;
         kvm_vm_fops.owner = module;
         kvm_vcpu_fops.owner = module;
@@ -2343,7 +2366,7 @@ int kvm_init(void *opaque, unsigned vcpu_size, unsigned vcpu_align,
         r = misc_register(&kvm_dev);
         if (r) {
                 printk(KERN_ERR "kvm: misc device register failed\n");
-                goto out_free;
+                goto out_unreg;
         }
 
         kvm_preempt_ops.sched_in = kvm_sched_in;
@@ -2353,6 +2376,8 @@ int kvm_init(void *opaque, unsigned vcpu_size, unsigned vcpu_align,
 
         return 0;
 
+out_unreg:
+        kvm_async_pf_deinit();
 out_free:
         kmem_cache_destroy(kvm_vcpu_cache);
 out_free_5:
@@ -2385,6 +2410,7 @@ void kvm_exit(void)
         kvm_exit_debug();
         misc_deregister(&kvm_dev);
         kmem_cache_destroy(kvm_vcpu_cache);
+        kvm_async_pf_deinit();
         sysdev_unregister(&kvm_sysdev);
         sysdev_class_unregister(&kvm_sysdev_class);
         unregister_reboot_notifier(&kvm_reboot_notifier);