ieee1394: eth1394: correct return codes in hard_start_xmit
Stefan Richter [Mon, 2 Apr 2007 00:24:27 +0000 (02:24 +0200)]
This patch actually doesn't change anything because there was always 0
== NETDEV_TX_OK returned before.

TODO: Return NETDEV_TX_BUSY in error case and test in different error
conditions.

Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>

drivers/ieee1394/eth1394.c

index b5cd107..1c26322 100644 (file)
@@ -1527,7 +1527,6 @@ static int ether1394_tx(struct sk_buff *skb, struct net_device *dev)
        unsigned long flags;
        nodeid_t dest_node;
        eth1394_tx_type tx_type;
-       int ret = 0;
        unsigned int tx_len;
        unsigned int max_payload;
        u16 dg_size;
@@ -1537,26 +1536,20 @@ static int ether1394_tx(struct sk_buff *skb, struct net_device *dev)
        struct eth1394_node_info *node_info = NULL;
 
        ptask = kmem_cache_alloc(packet_task_cache, GFP_ATOMIC);
-       if (ptask == NULL) {
-               ret = -ENOMEM;
+       if (ptask == NULL)
                goto fail;
-       }
 
        /* XXX Ignore this for now. Noticed that when MacOSX is the IRM,
         * it does not set our validity bit. We need to compensate for
         * that somewhere else, but not in eth1394. */
 #if 0
-       if ((priv->host->csr.broadcast_channel & 0xc0000000) != 0xc0000000) {
-               ret = -EAGAIN;
+       if ((priv->host->csr.broadcast_channel & 0xc0000000) != 0xc0000000)
                goto fail;
-       }
 #endif
 
        skb = skb_share_check(skb, GFP_ATOMIC);
-       if (!skb) {
-               ret = -ENOMEM;
+       if (!skb)
                goto fail;
-       }
 
        /* Get rid of the fake eth1394 header, but save a pointer */
        eth = (struct eth1394hdr *)skb->data;
@@ -1583,16 +1576,13 @@ static int ether1394_tx(struct sk_buff *skb, struct net_device *dev)
 
                node = eth1394_find_node_guid(&priv->ip_node_list,
                                              be64_to_cpu(guid));
-               if (!node) {
-                       ret = -EAGAIN;
+               if (!node)
                        goto fail;
-               }
+
                node_info =
                    (struct eth1394_node_info *)node->ud->device.driver_data;
-               if (node_info->fifo == CSR1212_INVALID_ADDR_SPACE) {
-                       ret = -EAGAIN;
+               if (node_info->fifo == CSR1212_INVALID_ADDR_SPACE)
                        goto fail;
-               }
 
                dest_node = node->ud->ne->nodeid;
                max_payload = node_info->maxpayload;
@@ -1639,7 +1629,7 @@ static int ether1394_tx(struct sk_buff *skb, struct net_device *dev)
                goto fail;
 
        netif_wake_queue(dev);
-       return 0;
+       return NETDEV_TX_OK;
 fail:
        if (ptask)
                kmem_cache_free(packet_task_cache, ptask);
@@ -1655,7 +1645,15 @@ fail:
        if (netif_queue_stopped(dev))
                netif_wake_queue(dev);
 
-       return 0;  /* returning non-zero causes serious problems */
+       /*
+        * FIXME: According to a patch from 2003-02-26, "returning non-zero
+        * causes serious problems" here, allegedly.  Before that patch,
+        * -ERRNO was returned which is not appropriate under Linux 2.6.
+        * Perhaps more needs to be done?  Stop the queue in serious
+        * conditions and restart it elsewhere?
+        */
+       /* return NETDEV_TX_BUSY; */
+       return NETDEV_TX_OK;
 }
 
 static void ether1394_get_drvinfo(struct net_device *dev,