RE: [PATCH net-next v4 5/5] net: wangxun: implement pci_error_handlers ops

quoted

+static void wx_io_resume(struct pci_dev *pdev)
+{
+	struct wx *wx = pci_get_drvdata(pdev);
+	struct net_device *netdev;
+	int err;
+
+	netdev = wx->netdev;
+	rtnl_lock();
+	if (netif_running(netdev)) {
+		err = netdev->netdev_ops->ndo_open(netdev);
+		if (err) {
+			wx_err(wx, "Failed to open netdev after
reset\n");

If you jump over the netif_device_attach(netdev);
I'm afraid the interface probably will be UP (IFF_UP set, because netif_running() returned true)
but permanently "detached" - the kernel's watchdog, TX queue wakeups,
and carrier-change notifications all check netif_device_present() and
will silently no-op.

If .ndo_open() fails during recovery, the device has not been successfully
reinitialized and therefore should remain detached. 
Calling netif_device_attach() unconditionally would expose an interface whose
queues, interrupts or other resources may not have been restored successfully.