vhost: fix log ctx signalling
authorMichael S. Tsirkin <mst@redhat.com>
Wed, 22 Sep 2010 10:31:53 +0000 (12:31 +0200)
committerMichael S. Tsirkin <mst@redhat.com>
Wed, 22 Sep 2010 14:21:33 +0000 (16:21 +0200)
The log eventfd signalling got put in dead code.
We didn't notice because qemu currently does polling
instead of eventfd select.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
drivers/vhost/vhost.c

index c579dcc9200ccabe0f1bcf79afed59577e4a6f38..dd3d6f7406f80092cdd8843acdd06c313e9cf164 100644 (file)
@@ -858,11 +858,12 @@ int vhost_log_write(struct vhost_virtqueue *vq, struct vhost_log *log,
                if (r < 0)
                        return r;
                len -= l;
-               if (!len)
+               if (!len) {
+                       if (vq->log_ctx)
+                               eventfd_signal(vq->log_ctx, 1);
                        return 0;
+               }
        }
-       if (vq->log_ctx)
-               eventfd_signal(vq->log_ctx, 1);
        /* Length written exceeds what we have stored. This is a bug. */
        BUG();
        return 0;