Andrew Morton wrote:

On Fri, 22 Aug 2008 11:39:08 -0700
Karen Xie [off-list ref] wrote:

  

quoted

[PATCH 2/4 2.6.28] cxgb3 - handle ARP replies for private iSCSI IP address

From: Karen Xie <redacted>

stg can be tricky ...
[adding @adapter in cxgb3_arp_process doxygen header]
The accelerated iSCSI traffic uses a private IP address unknown to the OS.
The driver has to reply to ARP requests dedicated to the private IP address.

Signed-off-by: Divy Le Ray <redacted>
---

 drivers/net/cxgb3/sge.c |   71 ++++++++++++++++++++++++++++++++++++++++++++---
 1 files changed, 66 insertions(+), 5 deletions(-)

diff --git a/drivers/net/cxgb3/sge.c b/drivers/net/cxgb3/sge.c
index 1b0861d..d2a9285 100644
--- a/drivers/net/cxgb3/sge.c
+++ b/drivers/net/cxgb3/sge.c

@@ -36,6 +36,7 @@
 #include <linux/ip.h>
 #include <linux/tcp.h>
 #include <linux/dma-mapping.h>
+#include <net/arp.h>
 #include "common.h"
 #include "regs.h"
 #include "sge_defs.h"

@@ -1859,6 +1860,54 @@ static void restart_tx(struct sge_qset *qs)
 }
 
 /**
+ *	cxgb3_arp_process - process an ARP request probing a private IP address
+ *	@adapter: the adapter
+ *	@skb: the skbuff containing the ARP request
+ *
+ *	Check if the ARP request is probing the private IP address
+ *	dedicated to iSCSI, generate an ARP reply if so.
+ */
+static void cxgb3_arp_process(struct adapter *adapter, struct sk_buff *skb)
+{
+	struct net_device *dev = skb->dev;
+	struct port_info *pi;
+	struct arphdr *arp;
+	unsigned char *arp_ptr;
+	unsigned char *sha;
+	u32 sip, tip;
+
+	if (!dev)
+		return;
    

Can this happen?

  

quoted

+	skb_reset_network_header(skb);
+	arp = arp_hdr(skb);
+
+	if (arp->ar_op != htons(ARPOP_REQUEST))
+		return;
+
+	arp_ptr = (unsigned char *)(arp + 1);
+	sha = arp_ptr;
+	arp_ptr += dev->addr_len;
+	memcpy(&sip, arp_ptr, 4);
+	arp_ptr += 4;
+	arp_ptr += dev->addr_len;
+	memcpy(&tip, arp_ptr, 4);
    

Should arp_hdr_len() be used here?

  

quoted

+	pi = netdev_priv(dev);
+	if (ntohl(tip) != pi->iscsi_ipaddr)
+		return;
+
+	arp_send(ARPOP_REPLY, ETH_P_ARP, sip, dev, tip, sha,
+		 dev->dev_addr, sha);
+
+}
+
+static inline int is_arp(struct sk_buff *skb)
+{
+	return skb->protocol == htons(ETH_P_ARP);
    

Other net code uses __constant_htons() for this.

(Dunno why - if it makes a difference, htons() is broken?)


  

Doesn't this avoid complaints from sparse endian checking?