s390-virtio: Accommodate guests using virtqueues too early

Feature updates are not a synchronuous operation for the legacy
s390-virtio transport. This transport syncs the guest feature bits
(those from finalize) on the set_status hypercall. Before that qemu
thinks that features are zero, which means QEMU will misbehave, e.g.
it will not write the event index, even if the guest asks for it.

Let's detect the case where a kick happens before the driver is ready
and force sync the features.
With this workaround, it is now safe to switch to the common feature
bit handling code as used by all other transports.

Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>

1 files changed, 10 insertions, 0 deletions

diff --git a/hw/s390x/s390-virtio.c b/hw/s390x/s390-virtio.c
index 3a1b9ee2d0..59750dbfcd 100644
--- a/hw/s390x/s390-virtio.c
+++ b/hw/s390x/s390-virtio.c
@@ -77,6 +77,16 @@ static int s390_virtio_hcall_notify(const uint64_t *args)
     if (mem > ram_size) {
         VirtIOS390Device *dev = s390_virtio_bus_find_vring(s390_bus, mem, &i);
         if (dev) {
+            /*
+             * Older kernels will use the virtqueue before setting DRIVER_OK.
+             * In this case the feature bits are not yet up to date, meaning
+             * that several funny things can happen, e.g. the guest thinks
+             * EVENT_IDX is on and QEMU thinks it is off. Let's force a feature
+             * and status sync.
+             */
+            if (!(dev->vdev->status & VIRTIO_CONFIG_S_DRIVER_OK)) {
+                s390_virtio_device_update_status(dev);
+            }
             virtio_queue_notify(dev->vdev, i);
         } else {
             r = -EINVAL;