virtio: set ISR on dataplane notifications

Dataplane has been omitting forever the step of setting ISR when
an interrupt is raised.  This caused little breakage, because the
specification actually says that ISR may not be updated in MSI mode.

Some versions of the Windows drivers however didn't clear MSI mode
correctly, and proceeded using polling mode (using ISR, not the used
ring index!) for crashdump and hibernation.  If it were just crashdump
and hibernation it would not be a big deal, but recent releases of
Windows do not really shut down, but rather log out and hibernate to
make the next startup faster.  Hence, this manifested as a more serious
hang during shutdown with e.g. Windows 8.1 and virtio-win 1.8.0 RPMs.
Newer versions fixed this, while older versions do not use MSI at all.

The failure has always been there for virtio dataplane, but it became
visible after commits 9ffe337 ("virtio-blk: always use dataplane path
if ioeventfd is active", 2016-10-30) and ad07cd6 ("virtio-scsi: always
use dataplane path if ioeventfd is active", 2016-10-30) made virtio-blk
and virtio-scsi always use the dataplane code under KVM.  The good news
therefore is that it was not a bug in the patches---they were doing
exactly what they were meant for, i.e. shake out remaining dataplane bugs.

The fix is not hard, so it's worth arranging for the broken drivers.
The virtio_should_notify+event_notifier_set pair that is common to
virtio-blk and virtio-scsi dataplane is replaced with a new public
function virtio_notify_irqfd that also sets ISR.  The irqfd emulation
code now need not set ISR anymore, so virtio_irq is removed.

Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Farhan Ali <alifm@linux.vnet.ibm.com>
Tested-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

5 files changed, 31 insertions, 20 deletions

diff --git a/hw/block/dataplane/virtio-blk.c b/hw/block/dataplane/virtio-blk.c
index 90ef557c8c..d1f9f63eaf 100644
--- a/hw/block/dataplane/virtio-blk.c
+++ b/hw/block/dataplane/virtio-blk.c
@@ -68,9 +68,7 @@ static void notify_guest_bh(void *opaque)
             unsigned i = j + ctzl(bits);
             VirtQueue *vq = virtio_get_queue(s->vdev, i);
 
-            if (virtio_should_notify(s->vdev, vq)) {
-                event_notifier_set(virtio_queue_get_guest_notifier(vq));
-            }
+            virtio_notify_irqfd(s->vdev, vq);
 
             bits &= bits - 1; /* clear right-most bit */
         }
diff --git a/hw/scsi/virtio-scsi-dataplane.c b/hw/scsi/virtio-scsi-dataplane.c
index f2ea29dbc3..6b8d0f0024 100644
--- a/hw/scsi/virtio-scsi-dataplane.c
+++ b/hw/scsi/virtio-scsi-dataplane.c
@@ -95,13 +95,6 @@ static int virtio_scsi_vring_init(VirtIOSCSI *s, VirtQueue *vq, int n,
     return 0;
 }
 
-void virtio_scsi_dataplane_notify(VirtIODevice *vdev, VirtIOSCSIReq *req)
-{
-    if (virtio_should_notify(vdev, req->vq)) {
-        event_notifier_set(virtio_queue_get_guest_notifier(req->vq));
-    }
-}
-
 /* assumes s->ctx held */
 static void virtio_scsi_clear_aio(VirtIOSCSI *s)
 {
diff --git a/hw/scsi/virtio-scsi.c b/hw/scsi/virtio-scsi.c
index 3e5ae6ac0f..10fd687193 100644
--- a/hw/scsi/virtio-scsi.c
+++ b/hw/scsi/virtio-scsi.c
@@ -69,7 +69,7 @@ static void virtio_scsi_complete_req(VirtIOSCSIReq *req)
     qemu_iovec_from_buf(&req->resp_iov, 0, &req->resp, req->resp_size);
     virtqueue_push(vq, &req->elem, req->qsgl.size + req->resp_iov.size);
     if (s->dataplane_started && !s->dataplane_fenced) {
-        virtio_scsi_dataplane_notify(vdev, req);
+        virtio_notify_irqfd(vdev, vq);
     } else {
         virtio_notify(vdev, vq);
     }
diff --git a/hw/virtio/trace-events b/hw/virtio/trace-events
index 8756cefa79..7b6f55e70e 100644
--- a/hw/virtio/trace-events
+++ b/hw/virtio/trace-events
@@ -5,7 +5,7 @@ virtqueue_fill(void *vq, const void *elem, unsigned int len, unsigned int idx) "
 virtqueue_flush(void *vq, unsigned int count) "vq %p count %u"
 virtqueue_pop(void *vq, void *elem, unsigned int in_num, unsigned int out_num) "vq %p elem %p in_num %u out_num %u"
 virtio_queue_notify(void *vdev, int n, void *vq) "vdev %p n %d vq %p"
-virtio_irq(void *vq) "vq %p"
+virtio_notify_irqfd(void *vdev, void *vq) "vdev %p vq %p"
 virtio_notify(void *vdev, void *vq) "vdev %p vq %p"
 virtio_set_status(void *vdev, uint8_t val) "vdev %p val %u"
 
diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
index 138a414cbe..1af2de2714 100644
--- a/hw/virtio/virtio.c
+++ b/hw/virtio/virtio.c
@@ -1330,13 +1330,6 @@ static void virtio_set_isr(VirtIODevice *vdev, int value)
     }
 }
 
-void virtio_irq(VirtQueue *vq)
-{
-    trace_virtio_irq(vq);
-    virtio_set_isr(vq->vdev, 0x1);
-    virtio_notify_vector(vq->vdev, vq->vector);
-}
-
 bool virtio_should_notify(VirtIODevice *vdev, VirtQueue *vq)
 {
     uint16_t old, new;
@@ -1360,6 +1353,33 @@ bool virtio_should_notify(VirtIODevice *vdev, VirtQueue *vq)
     return !v || vring_need_event(vring_get_used_event(vq), new, old);
 }
 
+void virtio_notify_irqfd(VirtIODevice *vdev, VirtQueue *vq)
+{
+    if (!virtio_should_notify(vdev, vq)) {
+        return;
+    }
+
+    trace_virtio_notify_irqfd(vdev, vq);
+
+    /*
+     * virtio spec 1.0 says ISR bit 0 should be ignored with MSI, but
+     * windows drivers included in virtio-win 1.8.0 (circa 2015) are
+     * incorrectly polling this bit during crashdump and hibernation
+     * in MSI mode, causing a hang if this bit is never updated.
+     * Recent releases of Windows do not really shut down, but rather
+     * log out and hibernate to make the next startup faster.  Hence,
+     * this manifested as a more serious hang during shutdown with
+     *
+     * Next driver release from 2016 fixed this problem, so working around it
+     * is not a must, but it's easy to do so let's do it here.
+     *
+     * Note: it's safe to update ISR from any thread as it was switched
+     * to an atomic operation.
+     */
+    virtio_set_isr(vq->vdev, 0x1);
+    event_notifier_set(&vq->guest_notifier);
+}
+
 void virtio_notify(VirtIODevice *vdev, VirtQueue *vq)
 {
     if (!virtio_should_notify(vdev, vq)) {
@@ -1994,7 +2014,7 @@ static void virtio_queue_guest_notifier_read(EventNotifier *n)
 {
     VirtQueue *vq = container_of(n, VirtQueue, guest_notifier);
     if (event_notifier_test_and_clear(n)) {
-        virtio_irq(vq);
+        virtio_notify_vector(vq->vdev, vq->vector);
     }
 }