SCSI and FC Transport: add netlink support for posting of transport events

include/scsi/scsi_netlink.h line 73

Quoted 13 lines · James Smart

include/scsi/scsi_netlink.h +13 -0

+/*
+ * Vendor ID:
+ *   If transports post vendor-unique events, they must pass a well-known
+ *   32-bit vendor identifier. This identifier consists of 8 bits indicating
+ *   the "type" of identifier contained, and 24 bits of id data.
+ *
+ *   Identifiers for each type:
+ *    PCI :  ID data is the 16 bit PCI Registered Vendor ID
+ */
+#define SCSI_NL_VID_ID_MASK			0x00FFFFFF
+#define SCSI_NL_VID_TYPE_MASK			0xFF000000
+#define SCSI_NL_VID_TYPE_PCI			0x01000000
+

Would it be possible to make this a little larger? For iscsi we wanted to use the MAC address. It can wait. I do not care if it goes in now or later. I just don't want to be the first one to break apps by changing the number and seeing how many checked version numbers :)

How big of a number do you need ? 48 bits ?
We can up to 64bits, but I'd reserve 8bits for a "type" field. (ugh, sounds like I'm redefining naming authorities...)

On a side thought - is the mac address really the right thing to use for a vendor id. Wouldn't you be extracting the vendor id from the mac address ?

Quoted 4 lines · James Smart

How big of a number do you need ? 48 bits ?
We can up to 64bits, but I'd reserve 8bits for a "type" field. (ugh, sounds like I'm redefining naming authorities...)

Quoted 4 lines

On a side thought - is the mac address really the right thing to use for a vendor id. Wouldn't you be extracting the vendor id from the mac address ?

I was looking for a persistent per hba value for some other long reason that we can work around in userspace so I take back that comment.

Are you only sending the vendor id to match some vendor specific userspace code with the event?

Currently, for iscsi we only send the
host no then match the driver, host and some driver or vendor specific code by using the host no and proc_name attr.

This is because for
software iscsi and iser we do not know the pci device that will be used. For software iscsi it could be a bonded device or multiple devices depending on tables getting updated. What would we use for the vendor id in this case?

Is there a null value or some special unused one we can abuse or why do you need the vendor id value if you can look that up in sysfs by just following the host no to the scsi host dir then going to the pci device?

Quoted 10 lines · Mike Christie

Quoted 8 lines · James Smart

How big of a number do you need ? 48 bits ?
We can up to 64bits, but I'd reserve 8bits for a "type" field. (ugh, sounds like I'm redefining naming authorities...)

On a side thought - is the mac address really the right thing to use for a vendor id. Wouldn't you be extracting the vendor id from the mac address ?

Quoted 2 lines

I was looking for a persistent per hba value for some other long reason that we can work around in userspace so I take back that comment.

Ok - does that mean you want more than 24 bits for a vendor id ? Seems reasonable...

Quoted 3 lines

Are you only sending the vendor id to match some vendor specific userspace code with the event?

I'm only passing vendor id on events that are vendor specific. Thus, match on vendor id to decode the vendor event. All other events are "well-known" by the transport, so vendor id is irrelevant. For well-known events, host no
should be all you need.

If you need to find the vendor matched to the host no, then you follow the other sysfs links appropriately (and should eventually resolve this from the general driver framework attributes).

The passing of the vendor id in the vendor messages is simply to make application code simpler - so it doesn't have to following all the sysfs links, or know the nuances of different i/o busses and their sysfs attributes if the vendor supports more than just pci.

Quoted 3 lines

Currently, for iscsi we only send the
host no then match the driver, host and some driver or vendor specific code by using the host no and proc_name attr.

Yep this is what I would expect - except that proc_name won't always be there (it's deprecated).

Quoted 6 lines

This is because for
software iscsi and iser we do not know the pci device that will be used. For software iscsi it could be a bonded device or multiple devices depending on tables getting updated. What would we use for the vendor id in this case?

Well - in this case, it's not a vendor-specific event being generated. It's an event by the iscsi/iser layer. This should fit the "well-known to the iscsi/iser transport" event definition and not require a vendor id.

Quoted 6 lines · James Smart

Quoted 5 lines · Mike Christie

Quoted 4 lines · James Smart

How big of a number do you need ? 48 bits ?
We can up to 64bits, but I'd reserve 8bits for a "type" field. (ugh, sounds like I'm redefining naming authorities...)

Quoted 8 lines

Quoted 6 lines

Quoted 4 lines

On a side thought - is the mac address really the right thing to use for a vendor id. Wouldn't you be extracting the vendor id from the mac address ?

I was looking for a persistent per hba value for some other long reason that we can work around in userspace so I take back that comment.

Ok - does that mean you want more than 24 bits for a vendor id ? Seems reasonable...

No. I meant to say below I do not need it. I thought that is what you were telling me :) Given the host no I can find the mac address or pci vendor or device id in sysfs

Quoted 20 lines

Quoted 2 lines

Are you only sending the vendor id to match some vendor specific userspace code with the event?

I'm only passing vendor id on events that are vendor specific. Thus, match on vendor id to decode the vendor event. All other events are "well-known" by the transport, so vendor id is irrelevant. For well-known events, host no
should be all you need.

If you need to find the vendor matched to the host no, then you follow the other sysfs links appropriately (and should eventually resolve this from the general driver framework attributes).

The passing of the vendor id in the vendor messages is simply to make application code simpler - so it doesn't have to following all the sysfs links, or know the nuances of different i/o busses and their sysfs attributes if the vendor supports more than just pci.

Quoted 3 lines

Currently, for iscsi we only send the
host no then match the driver, host and some driver or vendor specific code by using the host no and proc_name attr.

Quoted 2 lines

Yep this is what I would expect - except that proc_name won't always be there (it's deprecated).

Are you sure it is depreciated? I thought the proc name for proc was depreciated but then we added it to sysfs so we could figure out which module goes with which host or something didn't we?

It is not vendor specific but it is driver specific and so we need a way to route the message to the proper userspace handler for similar reasons you may need a vendor id? For example qlogic handling may be slightly different in userspace then broadcom. So what I am saying is I thought the driver and vendor ids are similar, but vendor id may not work for iscsi_tcp/iser?

Quoted 5 lines

This is because for
software iscsi and iser we do not know the pci device that will be used. For software iscsi it could be a bonded device or multiple devices depending on tables getting updated. What would we use for the vendor id in this case?

Well - in this case, it's not a vendor-specific event being generated. It's an event by the iscsi/iser layer. This should fit the "well-known to the iscsi/iser transport" event definition and not require a vendor id.

If we expected to have multiple iscsi/iser stacks present (vs the 1 and only
1 implementation we have for the other transports), then either that gets built into the event message (as a stack identifier) or you could munge the same thing into the vendor id (e.g. define a "type" that indicates s/w assigned - and the vendor id is the stack identifier (but this method seems hoaky)).

OK ... I looked through this, but I'm not at all keen on entangling SCSI with NET (i.e. requiring NET if SCSI is selected).

How about doing it this way?

James

---
Index: BUILD-2.6/drivers/scsi/Kconfig
===================================================================

drivers/scsi/Kconfig +6 -0

@@ -27,6 +27,11 @@ config SCSI
 	  However, do not compile this as a module if your root file system
 	  (the one containing the directory /) is located on a SCSI device.
 
+config SCSI_NETLINK
+	tristate
+	default	n
+	select NET
+
 config SCSI_PROC_FS
 	bool "legacy /proc/scsi/ support"
 	depends on SCSI && PROC_FS

@@ -222,6 +227,7 @@ config SCSI_SPI_ATTRS
 config SCSI_FC_ATTRS
 	tristate "FiberChannel Transport Attributes"
 	depends on SCSI
+	select SCSI_NETLINK
 	help
 	  If you wish to export transport-specific information about
 	  each attached FiberChannel device to sysfs, say Y.

Index: BUILD-2.6/drivers/scsi/Makefile
===================================================================

drivers/scsi/Makefile +1 -0

@@ -143,6 +143,7 @@ scsi_mod-y			+= scsi.o hosts.o scsi_ioct
 				   scsicam.o scsi_error.o scsi_lib.o \
 				   scsi_scan.o scsi_sysfs.o \
 				   scsi_devinfo.o
+scsi_mod-$(CONFIG_SCSI_NETLINK)	+= scsi_netlink.o
 scsi_mod-$(CONFIG_SYSCTL)	+= scsi_sysctl.o
 scsi_mod-$(CONFIG_SCSI_PROC_FS)	+= scsi_proc.o
 

Index: BUILD-2.6/drivers/scsi/scsi.c
===================================================================

drivers/scsi/scsi.c +3 -0

@@ -1115,6 +1115,8 @@ static int __init init_scsi(void)
 	for_each_possible_cpu(i)
 		INIT_LIST_HEAD(&per_cpu(scsi_done_q, i));
 
+	scsi_netlink_init();
+
 	printk(KERN_NOTICE "SCSI subsystem initialized\n");
 	return 0;
 

@@ -1135,6 +1137,7 @@ cleanup_queue:
 
 static void __exit exit_scsi(void)
 {
+	scsi_netlink_exit();
 	scsi_sysfs_unregister();
 	scsi_exit_sysctl();
 	scsi_exit_hosts();

Index: BUILD-2.6/drivers/scsi/scsi_netlink.c
===================================================================

drivers/scsi/scsi_netlink.c new file +199 -0

@@ -0,0 +1,199 @@
+/*
+ *  scsi_netlink.c  - SCSI Transport Netlink Interface
+ *
+ *  Copyright (C) 2006   James Smart, Emulex Corporation
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to the Free Software
+ *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+#include <linux/time.h>
+#include <linux/jiffies.h>
+#include <linux/security.h>
+#include <net/sock.h>
+#include <net/netlink.h>
+
+#include <scsi/scsi_netlink.h>
+#include "scsi_priv.h"
+
+struct sock *scsi_nl_sock = NULL;
+EXPORT_SYMBOL_GPL(scsi_nl_sock);
+
+
+/**
+ * scsi_nl_rcv_msg -
+ *    Receive message handler. Extracts message from a receive buffer.
+ *    Validates message header and calls appropriate transport message handler
+ *
+ * @skb:		socket receive buffer
+ *
+ **/
+static void
+scsi_nl_rcv_msg(struct sk_buff *skb)
+{
+	struct nlmsghdr *nlh;
+	struct scsi_nl_hdr *hdr;
+	uint32_t rlen;
+	int err;
+
+	while (skb->len >= NLMSG_SPACE(0)) {
+		err = 0;
+
+		nlh = (struct nlmsghdr *) skb->data;
+		if ((nlh->nlmsg_len < (sizeof(*nlh) + sizeof(*hdr))) ||
+		    (skb->len < nlh->nlmsg_len)) {
+			printk(KERN_WARNING "%s: discarding partial skb\n",
+				 __FUNCTION__);
+			return;
+		}
+
+		rlen = NLMSG_ALIGN(nlh->nlmsg_len);
+		if (rlen > skb->len)
+			rlen = skb->len;
+
+		if (nlh->nlmsg_type != SCSI_TRANSPORT_MSG) {
+			err = -EBADMSG;
+			goto next_msg;
+		}
+
+		hdr = NLMSG_DATA(nlh);
+		if ((hdr->version != SCSI_NL_VERSION) ||
+		    (hdr->magic != SCSI_NL_MAGIC)) {
+			err = -EPROTOTYPE;
+			goto next_msg;
+		}
+
+		if (security_netlink_recv(skb, CAP_SYS_ADMIN)) {
+			err = -EPERM;
+			goto next_msg;
+		}
+
+		if (nlh->nlmsg_len < (sizeof(*nlh) + hdr->msglen)) {
+			printk(KERN_WARNING "%s: discarding partial message\n",
+				 __FUNCTION__);
+			return;
+		}
+
+		/*
+		 * We currently don't support anyone sending us a message
+		 */
+
+next_msg:
+		if ((err) || (nlh->nlmsg_flags & NLM_F_ACK))
+			netlink_ack(skb, nlh, err);
+
+		skb_pull(skb, rlen);
+	}
+}
+
+
+/**
+ * scsi_nl_rcv_msg -
+ *    Receive handler for a socket. Extracts a received message buffer from
+ *    the socket, and starts message processing.
+ *
+ * @sk:		socket
+ * @len:	unused
+ *
+ **/
+static void
+scsi_nl_rcv(struct sock *sk, int len)
+{
+	struct sk_buff *skb;
+
+	while ((skb = skb_dequeue(&sk->sk_receive_queue))) {
+		scsi_nl_rcv_msg(skb);
+		kfree_skb(skb);
+	}
+}
+
+
+/**
+ * scsi_nl_rcv_event -
+ *    Event handler for a netlink socket.
+ *
+ * @this:		event notifier block
+ * @event:		event type
+ * @ptr:		event payload
+ *
+ **/
+static int
+scsi_nl_rcv_event(struct notifier_block *this, unsigned long event, void *ptr)
+{
+	struct netlink_notify *n = ptr;
+
+	if (n->protocol != NETLINK_SCSITRANSPORT)
+		return NOTIFY_DONE;
+
+	/*
+	 * Currently, we are not tracking PID's, etc. There is nothing
+	 * to handle.
+	 */
+
+	return NOTIFY_DONE;
+}
+
+static struct notifier_block scsi_netlink_notifier = {
+	.notifier_call  = scsi_nl_rcv_event,
+};
+
+
+/**
+ * scsi_netlink_init -
+ *    Called by SCSI subsystem to intialize the SCSI transport netlink
+ *    interface
+ *
+ **/
+void
+scsi_netlink_init(void)
+{
+	int error;
+
+	error = netlink_register_notifier(&scsi_netlink_notifier);
+	if (error) {
+		printk(KERN_ERR "%s: register of event handler failed - %d\n",
+				__FUNCTION__, error);
+		return;
+	}
+
+	scsi_nl_sock = netlink_kernel_create(NETLINK_SCSITRANSPORT,
+				SCSI_NL_GRP_CNT, scsi_nl_rcv, THIS_MODULE);
+	if (!scsi_nl_sock) {
+		printk(KERN_ERR "%s: register of recieve handler failed\n",
+				__FUNCTION__);
+		netlink_unregister_notifier(&scsi_netlink_notifier);
+	}
+
+	return;
+}
+
+
+/**
+ * scsi_netlink_exit -
+ *    Called by SCSI subsystem to disable the SCSI transport netlink
+ *    interface
+ *
+ **/
+void
+scsi_netlink_exit(void)
+{
+	if (scsi_nl_sock) {
+		sock_release(scsi_nl_sock->sk_socket);
+		netlink_unregister_notifier(&scsi_netlink_notifier);
+	}
+
+	return;
+}
+
+

Index: BUILD-2.6/drivers/scsi/scsi_priv.h
===================================================================

drivers/scsi/scsi_priv.h +11 -0

@@ -8,6 +8,7 @@ struct scsi_cmnd;
 struct scsi_device;
 struct scsi_host_template;
 struct Scsi_Host;
+struct scsi_nl_hdr;
 
 
 /*

@@ -110,6 +111,16 @@ extern void __scsi_remove_device(struct 
 
 extern struct bus_type scsi_bus_type;
 
+/* scsi_netlink.c */
+#ifdef CONFIG_SCSI_NETLINK
+extern void scsi_netlink_init(void);
+extern void scsi_netlink_exit(void);
+extern struct sock *scsi_nl_sock;
+#else
+static inline void scsi_netlink_init(void) {}
+static inline void scsi_netlink_exit(void) {}
+#endif
+
 /* 
  * internal scsi timeout functions: for use by mid-layer and transport
  * classes.

Index: BUILD-2.6/drivers/scsi/scsi_transport_fc.c
===================================================================

drivers/scsi/scsi_transport_fc.c +199 -1

@@ -32,6 +32,9 @@
 #include <scsi/scsi_transport.h>
 #include <scsi/scsi_transport_fc.h>
 #include <scsi/scsi_cmnd.h>
+#include <linux/netlink.h>
+#include <net/netlink.h>
+#include <scsi/scsi_netlink_fc.h>
 #include "scsi_priv.h"
 
 static int fc_queue_work(struct Scsi_Host *, struct work_struct *);

@@ -93,6 +96,29 @@ fc_enum_name_search(port_type, fc_port_t
 #define FC_PORTTYPE_MAX_NAMELEN		50
 
 
+/* Convert fc_host_event_code values to ascii string name */
+static const struct {
+	enum fc_host_event_code		value;
+	char				*name;
+} fc_host_event_code_names[] = {
+	{ FCH_EVT_LIP,			"lip" },
+	{ FCH_EVT_LINKUP,		"link_up" },
+	{ FCH_EVT_LINKDOWN,		"link_down" },
+	{ FCH_EVT_LIPRESET,		"lip_reset" },
+	{ FCH_EVT_RSCN,			"rscn" },
+	{ FCH_EVT_ADAPTER_CHANGE,	"adapter_chg" },
+	{ FCH_EVT_PORT_UNKNOWN,		"port_unknown" },
+	{ FCH_EVT_PORT_ONLINE,		"port_online" },
+	{ FCH_EVT_PORT_OFFLINE,		"port_offline" },
+	{ FCH_EVT_PORT_FABRIC,		"port_fabric" },
+	{ FCH_EVT_LINK_UNKNOWN,		"link_unknown" },
+	{ FCH_EVT_VENDOR_UNIQUE,	"vendor_unique" },
+};
+fc_enum_name_search(host_event_code, fc_host_event_code,
+		fc_host_event_code_names)
+#define FC_HOST_EVENT_CODE_MAX_NAMELEN	30
+
+
 /* Convert fc_port_state values to ascii string name */
 static struct {
 	enum fc_port_state	value;

@@ -377,10 +403,182 @@ MODULE_PARM_DESC(dev_loss_tmo,
 		 " exceeded, the scsi target is removed. Value should be"
 		 " between 1 and SCSI_DEVICE_BLOCK_MAX_TIMEOUT.");
 
+/**
+ * Netlink Infrastructure
+ **/
+
+static atomic_t fc_event_seq;
+
+/**
+ * fc_get_event_number - Obtain the next sequential FC event number
+ *
+ * Notes:
+ *   We could have inline'd this, but it would have required fc_event_seq to
+ *   be exposed. For now, live with the subroutine call.
+ *   Atomic used to avoid lock/unlock...
+ **/
+u32
+fc_get_event_number(void)
+{
+	return atomic_add_return(1, &fc_event_seq);
+}
+EXPORT_SYMBOL(fc_get_event_number);
+
+
+/**
+ * fc_host_post_event - called to post an even on an fc_host.
+ *
+ * @shost:		host the event occurred on
+ * @event_number:	fc event number obtained from get_fc_event_number()
+ * @event_code:		fc_host event being posted
+ * @event_data:		32bits of data for the event being posted
+ *
+ * Notes:
+ *	This routine assumes no locks are held on entry.
+ **/
+void
+fc_host_post_event(struct Scsi_Host *shost, u32 event_number,
+		enum fc_host_event_code event_code, u32 event_data)
+{
+	struct sk_buff *skb;
+	struct nlmsghdr	*nlh;
+	struct fc_nl_event *event;
+	const char *name;
+	u32 len, skblen;
+	int err;
+
+	if (!scsi_nl_sock) {
+		err = -ENOENT;
+		goto send_fail;
+	}
+
+	len = FC_NL_MSGALIGN(sizeof(*event));
+	skblen = NLMSG_SPACE(len);
+
+	skb = alloc_skb(skblen, GFP_KERNEL);
+	if (!skb) {
+		err = -ENOBUFS;
+		goto send_fail;
+	}
+
+	nlh = nlmsg_put(skb, 0, 0, SCSI_TRANSPORT_MSG,
+				skblen - sizeof(*nlh), 0);
+	if (!nlh) {
+		err = -ENOBUFS;
+		goto send_fail_skb;
+	}
+	event = NLMSG_DATA(nlh);
+
+	INIT_SCSI_NL_HDR(&event->snlh, SCSI_NL_TRANSPORT_FC,
+				FC_NL_ASYNC_EVENT, len);
+	event->seconds = get_seconds();
+	event->vendor_id = 0;
+	event->host_no = shost->host_no;
+	event->event_datalen = sizeof(u32);	/* bytes */
+	event->event_num = event_number;
+	event->event_code = event_code;
+	event->event_data = event_data;
+
+	err = nlmsg_multicast(scsi_nl_sock, skb, 0, SCSI_NL_GRP_FC_EVENTS);
+	if (err && (err != -ESRCH))	/* filter no recipient errors */
+		/* nlmsg_multicast already kfree_skb'd */
+		goto send_fail;
+
+	return;
+
+send_fail_skb:
+	kfree_skb(skb);
+send_fail:
+	name = get_fc_host_event_code_name(event_code);
+	printk(KERN_WARNING
+		"%s: Dropped Event : host %d %s data 0x%08x - err %d\n",
+		__FUNCTION__, shost->host_no,
+		(name) ? name : "<unknown>", event_data, err);
+	return;
+}
+EXPORT_SYMBOL(fc_host_post_event);
+
+
+/**
+ * fc_host_post_vendor_event - called to post a vendor unique event on
+ *                             a fc_host
+ *
+ * @shost:		host the event occurred on
+ * @event_number:	fc event number obtained from get_fc_event_number()
+ * @data_len:		amount, in bytes, of vendor unique data
+ * @data_buf:		pointer to vendor unique data
+ *
+ * Notes:
+ *	This routine assumes no locks are held on entry.
+ **/
+void
+fc_host_post_vendor_event(struct Scsi_Host *shost, u32 event_number,
+		u32 data_len, char * data_buf, u32 vendor_id)
+{
+	struct sk_buff *skb;
+	struct nlmsghdr	*nlh;
+	struct fc_nl_event *event;
+	u32 len, skblen;
+	int err;
+
+	if (!scsi_nl_sock) {
+		err = -ENOENT;
+		goto send_vendor_fail;
+	}
+
+	len = FC_NL_MSGALIGN(sizeof(*event) + data_len);
+	skblen = NLMSG_SPACE(len);
+
+	skb = alloc_skb(skblen, GFP_KERNEL);
+	if (!skb) {
+		err = -ENOBUFS;
+		goto send_vendor_fail;
+	}
+
+	nlh = nlmsg_put(skb, 0, 0, SCSI_TRANSPORT_MSG,
+				skblen - sizeof(*nlh), 0);
+	if (!nlh) {
+		err = -ENOBUFS;
+		goto send_vendor_fail_skb;
+	}
+	event = NLMSG_DATA(nlh);
+
+	INIT_SCSI_NL_HDR(&event->snlh, SCSI_NL_TRANSPORT_FC,
+				FC_NL_ASYNC_EVENT, len);
+	event->seconds = get_seconds();
+	event->vendor_id = vendor_id;
+	event->host_no = shost->host_no;
+	event->event_datalen = data_len;	/* bytes */
+	event->event_num = event_number;
+	event->event_code = FCH_EVT_VENDOR_UNIQUE;
+	memcpy(&event->event_data, data_buf, data_len);
+
+	err = nlmsg_multicast(scsi_nl_sock, skb, 0, SCSI_NL_GRP_FC_EVENTS);
+	if (err && (err != -ESRCH))	/* filter no recipient errors */
+		/* nlmsg_multicast already kfree_skb'd */
+		goto send_vendor_fail;
+
+	return;
+
+send_vendor_fail_skb:
+	kfree_skb(skb);
+send_vendor_fail:
+	printk(KERN_WARNING
+		"%s: Dropped Event : host %d vendor_unique - err %d\n",
+		__FUNCTION__, shost->host_no, err);
+	return;
+}
+EXPORT_SYMBOL(fc_host_post_vendor_event);
+
+
 
 static __init int fc_transport_init(void)
 {
-	int error = transport_class_register(&fc_host_class);
+	int error;
+
+	atomic_set(&fc_event_seq, 0);
+
+	error = transport_class_register(&fc_host_class);
 	if (error)
 		return error;
 	error = transport_class_register(&fc_rport_class);

Index: BUILD-2.6/include/linux/netlink.h
===================================================================

include/linux/netlink.h +2 -0

@@ -21,6 +21,8 @@
 #define NETLINK_DNRTMSG		14	/* DECnet routing messages */
 #define NETLINK_KOBJECT_UEVENT	15	/* Kernel messages to userspace */
 #define NETLINK_GENERIC		16
+/* leave room for NETLINK_DM (DM Events) */
+#define NETLINK_SCSITRANSPORT	18	/* SCSI Transports */
 
 #define MAX_LINKS 32		
 

Index: BUILD-2.6/include/scsi/scsi_netlink_fc.h
===================================================================

include/scsi/scsi_netlink_fc.h new file +71 -0

@@ -0,0 +1,71 @@
+/*
+ *  FC Transport Netlink Interface
+ *
+ *  Copyright (C) 2006   James Smart, Emulex Corporation
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to the Free Software
+ *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+#ifndef SCSI_NETLINK_FC_H
+#define SCSI_NETLINK_FC_H
+
+#include <scsi/scsi_netlink.h>
+
+/*
+ * This file intended to be included by both kernel and user space
+ */
+
+/*
+ * FC Transport Message Types
+ */
+	/* kernel -> user */
+#define FC_NL_ASYNC_EVENT			0x0100
+	/* user -> kernel */
+/* none */
+
+
+/*
+ * Message Structures :
+ */
+
+/* macro to round up message lengths to 8byte boundary */
+#define FC_NL_MSGALIGN(len)		(((len) + 7) & ~7)
+
+
+/*
+ * FC Transport Broadcast Event Message :
+ *   FC_NL_ASYNC_EVENT
+ *
+ * Note: if Vendor Unique message, &event_data will be  start of
+ * 	 vendor unique payload, and the length of the payload is
+ *       per event_datalen
+ *
+ * Note: When specifying vendor_id, be sure to read the Vendor Type and ID
+ *   formatting requirements specified in scsi_netlink.h
+ */
+struct fc_nl_event {
+	struct scsi_nl_hdr snlh;		/* must be 1st element ! */
+	uint64_t seconds;
+	uint32_t vendor_id;
+	uint16_t host_no;
+	uint16_t event_datalen;
+	uint32_t event_num;
+	uint32_t event_code;
+	uint32_t event_data;
+} __attribute__((aligned(sizeof(uint64_t))));
+
+
+#endif /* SCSI_NETLINK_FC_H */
+

Index: BUILD-2.6/include/scsi/scsi_netlink.h
===================================================================

include/scsi/scsi_netlink.h new file +86 -0

@@ -0,0 +1,86 @@
+/*
+ *  SCSI Transport Netlink Interface
+ *    Used for the posting of outbound SCSI transport events
+ *
+ *  Copyright (C) 2006   James Smart, Emulex Corporation
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License, or
+ *  (at your option) any later version.
+ *
+ *  This program is distributed in the hope that it will be useful,
+ *  but WITHOUT ANY WARRANTY; without even the implied warranty of
+ *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *  GNU General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, write to the Free Software
+ *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+#ifndef SCSI_NETLINK_H
+#define SCSI_NETLINK_H
+
+/*
+ * This file intended to be included by both kernel and user space
+ */
+
+/* Single Netlink Message type to send all SCSI Transport messages */
+#define SCSI_TRANSPORT_MSG		NLMSG_MIN_TYPE + 1
+
+/* SCSI Transport Broadcast Groups */
+	/* leaving groups 0 and 1 unassigned */
+#define SCSI_NL_GRP_FC_EVENTS		(1<<2)		/* Group 2 */
+#define SCSI_NL_GRP_CNT			3
+
+
+/* SCSI_TRANSPORT_MSG event message header */
+struct scsi_nl_hdr {
+	uint8_t version;
+	uint8_t transport;
+	uint16_t magic;
+	uint16_t msgtype;
+	uint16_t msglen;
+} __attribute__((aligned(sizeof(uint64_t))));
+
+/* scsi_nl_hdr->version value */
+#define SCSI_NL_VERSION				1
+
+/* scsi_nl_hdr->magic value */
+#define SCSI_NL_MAGIC				0xA1B2
+
+/* scsi_nl_hdr->transport value */
+#define SCSI_NL_TRANSPORT			0
+#define SCSI_NL_TRANSPORT_FC			1
+#define SCSI_NL_MAX_TRANSPORTS			2
+
+/* scsi_nl_hdr->msgtype values are defined in each transport */
+
+
+/*
+ * Vendor ID:
+ *   If transports post vendor-unique events, they must pass a well-known
+ *   32-bit vendor identifier. This identifier consists of 8 bits indicating
+ *   the "type" of identifier contained, and 24 bits of id data.
+ *
+ *   Identifiers for each type:
+ *    PCI :  ID data is the 16 bit PCI Registered Vendor ID
+ */
+#define SCSI_NL_VID_ID_MASK			0x00FFFFFF
+#define SCSI_NL_VID_TYPE_MASK			0xFF000000
+#define SCSI_NL_VID_TYPE_PCI			0x01000000
+
+
+#define INIT_SCSI_NL_HDR(hdr, t, mtype, mlen)			\
+	{							\
+	(hdr)->version = SCSI_NL_VERSION;			\
+	(hdr)->transport = t;					\
+	(hdr)->magic = SCSI_NL_MAGIC;				\
+	(hdr)->msgtype = mtype;					\
+	(hdr)->msglen = mlen;					\
+	}
+
+
+#endif /* SCSI_NETLINK_H */
+

Index: BUILD-2.6/include/scsi/scsi_transport_fc.h
===================================================================

include/scsi/scsi_transport_fc.h +34 -0

@@ -29,6 +29,7 @@
 
 #include <linux/sched.h>
 #include <scsi/scsi.h>
+#include <scsi/scsi_netlink.h>
 
 struct scsi_transport_template;
 

@@ -284,6 +285,30 @@ struct fc_host_statistics {
 
 
 /*
+ * FC Event Codes - Polled and Async, following FC HBAAPI v2.0 guidelines
+ */
+
+/*
+ * fc_host_event_code: If you alter this, you also need to alter
+ * scsi_transport_fc.c (for the ascii descriptions).
+ */
+enum fc_host_event_code  {
+	FCH_EVT_LIP			= 0x1,
+	FCH_EVT_LINKUP			= 0x2,
+	FCH_EVT_LINKDOWN		= 0x3,
+	FCH_EVT_LIPRESET		= 0x4,
+	FCH_EVT_RSCN			= 0x5,
+	FCH_EVT_ADAPTER_CHANGE		= 0x103,
+	FCH_EVT_PORT_UNKNOWN		= 0x200,
+	FCH_EVT_PORT_OFFLINE		= 0x201,
+	FCH_EVT_PORT_ONLINE		= 0x202,
+	FCH_EVT_PORT_FABRIC		= 0x204,
+	FCH_EVT_LINK_UNKNOWN		= 0x500,
+	FCH_EVT_VENDOR_UNIQUE		= 0xffff,
+};
+
+
+/*
  * FC Local Port (Host) Attributes
  *
  * Attributes are based on HBAAPI V2.0 definitions.

@@ -526,5 +551,14 @@ struct fc_rport *fc_remote_port_add(stru
 void fc_remote_port_delete(struct fc_rport  *rport);
 void fc_remote_port_rolechg(struct fc_rport  *rport, u32 roles);
 int scsi_is_fc_rport(const struct device *);
+u32 fc_get_event_number(void);
+void fc_host_post_event(struct Scsi_Host *shost, u32 event_number,
+		enum fc_host_event_code event_code, u32 event_data);
+void fc_host_post_vendor_event(struct Scsi_Host *shost, u32 event_number,
+		u32 data_len, char * data_buf, u32 vendor_id);
+	/* Note: when specifying vendor_id to fc_host_post_vendor_event()
+	 *   be sure to read the Vendor Type and ID formatting requirements
+	 *   specified in scsi_netlink.h
+	 */
 
 #endif /* SCSI_TRANSPORT_FC_H */