patch-2.4.7 linux/drivers/ieee1394/sbp2.c

Next file: linux/drivers/ieee1394/sbp2.h
Previous file: linux/drivers/ieee1394/raw1394.c
Back to the patch index
Back to the overall index

diff -u --recursive --new-file v2.4.6/linux/drivers/ieee1394/sbp2.c linux/drivers/ieee1394/sbp2.c
@@ -0,0 +1,3588 @@
+/*
+ * sbp2.c - SBP-2 protocol driver for IEEE-1394
+ *
+ * Copyright (C) 2000 James Goodwin, Filanet Corporation (www.filanet.com)
+ * jamesg@filanet.com
+ *
+ * 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.
+ */
+
+/*
+ * Brief Description:
+ *
+ * This driver implements the Serial Bus Protocol 2 (SBP-2) over IEEE-1394
+ * under Linux. The SBP-2 driver is implemented as an IEEE-1394 high-level
+ * driver. It also registers as a SCSI lower-level driver in order to accept
+ * SCSI commands for transport using SBP-2.
+ *
+ * Driver Loading:
+ *
+ * Currently, the SBP-2 driver is supported only as a module. Because the 
+ * Linux SCSI stack is not Plug-N-Play aware, module load order is 
+ * important. Assuming the SCSI core drivers are either built into the 
+ * kernel or already loaded as modules, you should load the IEEE-1394 modules 
+ * in the following order:
+ *
+ * 	ieee1394 (e.g. insmod ieee1394)
+ *	ohci1394 (e.g. insmod ohci1394)
+ *	sbp2 (e.g. insmod sbp2)
+ *
+ * The SBP-2 driver will attempt to discover any attached SBP-2 devices when first
+ * loaded, or after any IEEE-1394 bus reset (e.g. a hot-plug). It will then print 
+ * out a debug message indicating if it was able to discover a SBP-2 device.
+ *
+ * Currently, the SBP-2 driver will catch any attached SBP-2 devices during the
+ * initial scsi bus scan (when the driver is first loaded). To add or remove
+ * SBP-2 devices after this initial scan (i.e. if you plug-in or un-plug a 
+ * device after the SBP-2 driver is loaded), you must either use the scsi procfs
+ * add-single-device, remove-single-device, or a shell script such as 
+ * rescan-scsi-bus.sh.
+ *
+ * The easiest way to add/detect new SBP-2 devices is to run the shell script
+ * rescan-scsi-bus.sh (or re-load the SBP-2 driver). This script may be 
+ * found at:
+ * http://www.garloff.de/kurt/linux/rescan-scsi-bus.sh
+ *
+ * As an alternative, you may manually add/remove SBP-2 devices via the procfs with
+ * add-single-device <h> <b> <t> <l> or remove-single-device <h> <b> <t> <l>, where:
+ *	<h> = host (starting at zero for first SCSI adapter)
+ *	<b> = bus (normally zero)
+ *	<t> = target (starting at zero for first SBP-2 device)
+ *	<l> = lun (normally zero)
+ *
+ * e.g. To manually add/detect a new SBP-2 device
+ *	echo "scsi add-single-device 0 0 0 0" > /proc/scsi/scsi
+ *
+ * e.g. To manually remove a SBP-2 device after it's been unplugged
+ *	echo "scsi remove-single-device 0 0 0 0" > /proc/scsi/scsi
+ *
+ * e.g. To check to see which SBP-2/SCSI devices are currently registered
+ * 	cat /proc/scsi/scsi
+ *
+ * After scanning for new SCSI devices (above), you may access any attached 
+ * SBP-2 storage devices as if they were SCSI devices (e.g. mount /dev/sda1, 
+ * fdisk, mkfs, etc.).
+ *
+ *
+ * Module Load Options:
+ *
+ * The SBP-2 driver now has a number of module load parameters available for use
+ * in debugging/testing. Following are the valid parameters 
+ *	
+ * no_bus_scan - Skip the initial scsi bus scan during module load
+ * (1 = skip bus scan, 0 = perform bus scan, default = 0)
+ *
+ * mode_sense_hack - Emulate mode sense for devices like 1394 memory stick readers
+ * (1 = emulate/fake mode sense, 0 = do not emulate/fake mode sense, default = 0)  
+ *
+ * max_speed - Force max speed allowed
+ * (0 = 100mb, 1 = 200mb, 2 = 400mb, default = auto configure)
+ *
+ * serialize_io - Force scsi stack to send down one command at a time, for debugging
+ * (1 = serialize all I/O, 0 = do not serialize I/O, default = 1) 
+ *
+ * no_large_packets - Force scsi stack to limit max packet size sent down, for debugging
+ * (1 = limit max transfer size, 0 = do not limit max packet size, default = 0)
+ *
+ * (e.g. insmod sbp2 no_bus_scan=1)
+ *
+ *
+ * Current Support:
+ *
+ * The SBP-2 driver is still in an early state, but supports a variety of devices.
+ * I have read/written many gigabytes of data from/to SBP-2 drives, and have seen 
+ * performance of more than 16 MBytes/s on individual drives (limit of the media 
+ * transfer rate).
+ *
+ * Following are the devices that have been tested successfully:
+ *
+ *	- Western Digital IEEE-1394 hard drives
+ *	- Maxtor IEEE-1394 hard drives
+ *	- VST (SmartDisk) IEEE-1394 hard drives and Zip drives (several flavors)
+ *	- LaCie IEEE-1394 hard drives (several flavors)
+ *	- QPS IEEE-1394 CD-RW/DVD drives and hard drives
+ *	- BusLink IEEE-1394 hard drives
+ *	- Iomega IEEE-1394 Zip/Jazz drives
+ *	- ClubMac IEEE-1394 hard drives
+ *	- FirePower IEEE-1394 hard drives
+ *	- EzQuest IEEE-1394 hard drives and CD-RW drives
+ *	- Castlewood/ADS IEEE-1394 ORB drives
+ *	- Evergreen IEEE-1394 hard drives and CD-RW drives
+ *	- Addonics IEEE-1394 CD-RW drives
+ *	- Bellstor IEEE-1394 hard drives and CD-RW drives
+ *	- APDrives IEEE-1394 hard drives
+ *	- Fujitsu IEEE-1394 MO drives
+ *	- Sony IEEE-1394 CD-RW drives
+ *	- Epson IEEE-1394 scanner
+ *	- ADS IEEE-1394 memory stick and compact flash readers 
+ *	  (e.g. "insmod sbp2 mode_sense_hack=1" for mem stick and flash readers))
+ *	- SBP-2 bridge-based devices (LSI, Oxford Semiconductor, Indigita bridges)
+ *	- Various other standard IEEE-1394 hard drives and enclosures
+ *
+ *
+ * Performance Issues:
+ *
+ *	- Make sure you are "not" running fat/fat32 on your attached SBP-2 drives. You'll
+ *	  get much better performance formatting the drive ext2 (but you will lose the
+ *	  ability to easily move the drive between Windows/Linux).
+ *
+ *
+ * Current Issues:
+ *
+ *	- Currently, all I/O from the scsi stack is serialized by default, as there
+ *	  are some stress issues under investigation with deserialized I/O. To enable
+ *	  deserialized I/O for testing, do "insmod sbp2 serialize_io=0"
+ *
+ *	- Hot-Plugging: Need to add procfs support and integration with linux
+ *	  hot-plug support (http://linux-hotplug.sourceforge.net) for auto-mounting 
+ *	  of drives.
+ *
+ *	- Error Handling: SCSI aborts and bus reset requests are handled somewhat
+ *	  but the code needs additional debugging.
+ *
+ *	- IEEE-1394 Bus Management: There is currently little bus management
+ *	  in the core IEEE-1394 stack. Because of this, the SBP-2 driver handles
+ *	  detection of SBP-2 devices itself. This should be moved to the core
+ *	  stack.
+ *
+ *	- The SBP-2 driver is currently only supported as a module. It would not take
+ *	  much work to allow it to be compiled into the kernel, but you'd have to 
+ *	  add some init code to the kernel to support this... and modules are much
+ *	  more flexible anyway.   ;-)
+ *
+ *	- Workaround for PPC pismo firewire chipset (enable SBP2_PPC_PISMO_WORKAROUND
+ *	  define below).
+ *
+ *
+ * Core IEEE-1394 Stack Changes:
+ *
+ *	- The IEEE-1394 core stack guid code attempts to read the node unique id from
+ *	  each attached device after a bus reset. It currently uses a block read
+ *	  request to do this, which "upsets" certain not-well-behaved devices, such as
+ *	  some drives from QPS. If you have trouble with your IEEE-1394 storage 
+ *	  device being detected after loading sbp2, try commenting out the 
+ *	  init_ieee1394_guid() and cleanup_ieee1394_guid() lines at the bottom of 
+ *	  ieee1394_core.c (and rebuild ieee1394.o). 
+ *
+ *	- In ohci1394.h, remove the IEEE1394_USE_BOTTOM_HALVES #define, and rebuild. 
+ *	  This will give you around 30% to 40% performance increase.
+ *
+ *
+ * History:
+ *
+ *	07/25/00 - Initial revision (JSG)
+ *	08/11/00 - Following changes/bug fixes were made (JSG):
+ *		   * Bug fix to SCSI procfs code (still needs to be synched with 2.4 kernel).
+ *		   * Bug fix where request sense commands were actually sent on the bus.
+ *		   * Changed bus reset/abort code to deal with devices that spin up quite
+ *		     slowly (which result in SCSI time-outs).
+ *		   * "More" properly pull information from device's config rom, for enumeration
+ *		     of SBP-2 devices, and determining SBP-2 register offsets.
+ *		   * Change Simplified Direct Access Device type to Direct Access Device type in
+ *		     returned inquiry data, in order to make the SCSI stack happy.
+ *		   * Modified driver to register with the SCSI stack "before" enumerating any attached
+ *		     SBP-2 devices. This means that you'll have to use procfs scsi-add-device or 
+ *		     some sort of script to discover new SBP-2 devices.
+ *		   * Minor re-write of some code and other minor changes.
+ *	08/28/00 - Following changes/bug fixes were made (JSG):
+ *		   * Bug fixes to scatter/gather support (case of one s/g element)
+ *		   * Updated direction table for scsi commands (mostly DVD commands)
+ *		   * Retries when trying to detect SBP-2 devices (for slow devices)
+ *		   * Slightly better error handling (previously none) when commands time-out.
+ *		   * Misc. other bug fixes and code reorganization.
+ *	09/13/00 - Following changes/bug fixes were made (JSG)
+ *		   * Moved detection/enumeration code to a kernel thread which is woken up when IEEE-1394
+ *		     bus resets occur.
+ *		   * Added code to handle bus resets and hot-plugging while devices are mounted, but full
+ *		     hot-plug support is not quite there yet.
+ *		   * Now use speed map to determine speed and max payload sizes for ORBs
+ *		   * Clean-up of code and reorganization 
+ *	09/19/00 - Added better hot-plug support and other minor changes (JSG)
+ *	10/15/00 - Fixes for latest 2.4.0 test kernel, minor fix for hot-plug race. (JSG)
+ *	12/03/00 - Created pool of request packet structures for use in sending out sbp2 command
+ *		   and agent reset requests. This removes the kmallocs/kfrees in the critical I/O paths,
+ *		   and also deals with some subtle race conditions related to allocating and freeing
+ *		   packets. (JSG)
+ *      12/09/00 - Improved the sbp2 device detection by actually reading the root and unit 
+ *		   directory (khk@khk.net)
+ *	12/23/00 - Following changes/enhancements were made (JSG)
+ *		   * Only do SCSI to RBC command conversion for Direct Access and Simplified
+ *		     Direct Access Devices (this is pulled from the config rom root directory).
+ *		     This is needed because doing the conversion for all device types broke the
+ *		     Epson scanner. Still looking for a better way of determining when to convert
+ *		     commands (for RBC devices). Thanks to khk for helping on this!
+ *		   * Added ability to "emulate" physical dma support, for host adapters such as TILynx.
+ *		   * Determine max payload and speed by also looking at the host adapter's max_rec field.
+ *	01/19/01 - Added checks to sbp2 login and made the login time-out longer. Also fixed a compile 
+ *		   problem for 2.4.0. (JSG)
+ *	01/24/01 - Fixed problem when individual s/g elements are 64KB or larger. Needed to break
+ *		   up these larger elements, since the sbp2 page table element size is only 16 bits. (JSG)
+ *	01/29/01 - Minor byteswap fix for login response (used for reconnect and log out).
+ *	03/07/01 - Following changes/enhancements were made (JSG)
+ *		   * Changes to allow us to catch the initial scsi bus scan (for detecting sbp2
+ *		     devices when first loading sbp2.o). To disable this, un-define 
+ *		     SBP2_SUPPORT_INITIAL_BUS_SCAN.
+ *		   * Temporary fix to deal with many sbp2 devices that do not support individual
+ *		     transfers of greater than 128KB in size. 
+ *		   * Mode sense conversion from 6 byte to 10 byte versions for CDRW/DVD devices. (Mark Burton)
+ *		   * Define allowing support for goofy sbp2 devices that do not support mode
+ *		     sense command at all, allowing them to be mounted rw (such as 1394 memory
+ *		     stick and compact flash readers). Define SBP2_MODE_SENSE_WRITE_PROTECT_HACK
+ *		     if you need this fix.
+ *	03/29/01 - Major performance enhancements and misc. other changes. Thanks to Daniel Berlin for many of
+ *		   changes and suggestions for change:
+ *		   * Now use sbp2 doorbell and link commands on the fly (instead of serializing requests)
+ *		   * Removed all bit fields in an attempt to run on PPC machines (still needs a little more work)
+ *		   * Added large request break-up/linking support for sbp2 chipsets that do not support transfers 
+ *		     greater than 128KB in size.
+ *		   * Bumped up max commands per lun to two, and max total outstanding commands to eight.
+ *	04/03/01 - Minor clean-up. Write orb pointer directly if no outstanding commands (saves one 1394 bus
+ *		   transaction). Added module load options (bus scan, mode sense hack, max speed, serialize_io,
+ *		   no_large_transfers). Better bus reset handling while I/O pending. Set serialize_io to 1 by 
+ *		   default (debugging of deserialized I/O in progress).
+ *	04/04/01 - Added workaround for PPC Pismo firewire chipset. See #define below. (Daniel Berlin)
+ *	04/20/01 - Minor clean-up. Allocate more orb structures when running with sbp2 target chipsets with
+ *		   128KB max transfer limit.
+ *	06/16/01 - Converted DMA interfaces to pci_dma - Ben Collins
+ *							 <bcollins@debian.org
+ */
+    
+/*
+ * Includes
+ */
+
+#include <linux/config.h>
+#include <linux/kernel.h>
+#include <linux/list.h>
+#include <linux/string.h>
+#include <linux/slab.h>
+#include <linux/fs.h>
+#include <linux/poll.h>
+#include <linux/module.h>
+#include <linux/types.h>
+#include <linux/delay.h>
+#include <linux/sched.h>
+#include <linux/proc_fs.h>
+#include <linux/blk.h>
+#include <linux/smp_lock.h>
+#include <asm/current.h>
+#include <asm/uaccess.h>
+#include <asm/io.h>
+#include <asm/byteorder.h>
+#include <asm/system.h>
+#include <asm/io.h>
+#include <asm/scatterlist.h>
+
+#include "ieee1394.h"
+#include "ieee1394_types.h"
+#include "ieee1394_core.h"
+#include "hosts.h"
+#include "highlevel.h"
+#include "ieee1394_transactions.h"
+#include "../scsi/scsi.h"
+#include "../scsi/hosts.h"
+#include "../scsi/sd.h"
+#include "sbp2.h"
+
+/*
+ * PPC firewire Pismo chipset workaround!!!
+ *
+ * This is a workaround for a bug in the firewire pismo chipset. For some odd reason the status
+ * fifo address hi/lo must be byteswapped and the response address byteswapped, but no other
+ * parts of the structure. Apple's drivers seem to specifically check for the pismo and do
+ * the same workaround for sbp2. (Daniel Berlin)
+ *
+ * Please enable the following define if you're running on the PPC Pismo chipset.
+ */
+
+#ifdef CONFIG_IEEE1394_SBP2_PISMO
+#define SBP2_NEED_LOGIN_DESCRIPTOR_WORKAROUND
+#endif
+
+/*
+ * Module load parameter definitions
+ */
+
+/*
+ * Normally the sbp2 driver tries to catch the initial scsi bus scan to pick up any 
+ * attached sbp2 devices. Setting no_bus_scan to 1 tells the sbp2 driver not to catch
+ * this initial scsi bus scan on module load. You can always either add or remove devices 
+ * later through the rescan-scsi-bus.sh script or scsi procfs.
+ */
+MODULE_PARM(no_bus_scan,"i");
+MODULE_PARM_DESC(no_bus_scan, "Skip the initial scsi bus scan during module load");
+static int no_bus_scan = 0;
+
+/*
+ * Set mode_sense_hack to 1 if you have some sort of unusual sbp2 device, like a 1394 memory 
+ * stick reader, compact flash reader, or MO drive that does not support mode sense. Allows
+ * you to mount the media rw instead of ro.
+ */
+MODULE_PARM(mode_sense_hack,"i");
+MODULE_PARM_DESC(mode_sense_hack, "Emulate mode sense for devices like 1394 memory stick readers");
+static int mode_sense_hack = 0;
+
+/*
+ * Change max_speed on module load if you have a bad IEEE-1394 controller that has trouble running
+ * 2KB packets at 400mb.
+ *
+ * NOTE: On certain OHCI parts I have seen short packets on async transmit (probably 
+ * due to PCI latency/throughput issues with the part). You can bump down the speed if
+ * you are running into problems.
+ *
+ * Valid values:
+ * max_speed = 2 (default: max speed 400mb)
+ * max_speed = 1 (max speed 200mb)
+ * max_speed = 0 (max speed 100mb)
+ */
+MODULE_PARM(max_speed,"i");
+MODULE_PARM_DESC(max_speed, "Force down max speed (2 = 400mb default, 1 = 200mb, 0 = 100mb)");
+static int max_speed = SPEED_S400;
+
+/*
+ * Set serialize_io to 1 if you'd like only one scsi command sent down to us at a time (debugging).
+ */
+MODULE_PARM(serialize_io,"i");
+MODULE_PARM_DESC(serialize_io, "Serialize all I/O coming down from the scsi drivers (debugging)");
+static int serialize_io = 1;	/* serialize I/O until stress issues are resolved */
+
+/*
+ * Set no_large_packets to 1 if you'd like to limit the size of requests sent down to us (normally
+ * the sbp2 driver will break up any requests to any individual devices with 128KB transfer size limits).
+ * Sets max s/g list elements to 0x1f in size and disables s/g clustering.
+ */
+MODULE_PARM(no_large_packets,"i");
+MODULE_PARM_DESC(no_large_packets, "Do not allow large transfers from scsi drivers (debugging)");
+static int no_large_packets = 0;
+ 
+/*
+ * Debug levels, configured via kernel config.
+ */
+
+#ifdef CONFIG_IEEE1394_SBP2_DEBUG_ORBS
+#define SBP2_ORB_DEBUG(fmt, args...)	HPSB_ERR("sbp2("__FUNCTION__"): "fmt, ## args)
+u32 global_outstanding_command_orbs = 0;
+#define outstanding_orb_incr global_outstanding_command_orbs++
+#define outstanding_orb_decr global_outstanding_command_orbs--
+#else
+#define SBP2_ORB_DEBUG(fmt, args...)
+#define outstanding_orb_incr
+#define outstanding_orb_decr
+#endif
+
+#ifdef CONFIG_IEEE1394_SBP2_DEBUG_DMA
+#define SBP2_DMA_ALLOC(fmt, args...) \
+	HPSB_ERR("sbp2("__FUNCTION__")alloc(%d): "fmt, \
+		 ++global_outstanding_dmas, ## args)
+#define SBP2_DMA_FREE(fmt, args...) \
+	HPSB_ERR("sbp2("__FUNCTION__")free(%d): "fmt, \
+		 --global_outstanding_dmas, ## args)
+u32 global_outstanding_dmas = 0;
+#else
+#define SBP2_DMA_ALLOC(fmt, args...)
+#define SBP2_DMA_FREE(fmt, args...)
+#endif
+
+#if CONFIG_IEEE1394_SBP2_DEBUG >= 2
+#define SBP2_DEBUG(fmt, args...)	HPSB_ERR(fmt, ## args)	
+#define SBP2_INFO(fmt, args...)		HPSB_ERR(fmt, ## args)	
+#define SBP2_NOTICE(fmt, args...)	HPSB_ERR(fmt, ## args)	
+#define SBP2_WARN(fmt, args...)		HPSB_ERR(fmt, ## args)
+#elif CONFIG_IEEE1394_SBP2_DEBUG == 1
+#define SBP2_DEBUG(fmt, args...)	HPSB_DEBUG(fmt, ## args)
+#define SBP2_INFO(fmt, args...)		HPSB_INFO(fmt, ## args)
+#define SBP2_NOTICE(fmt, args...)	HPSB_NOTICE(fmt, ## args)
+#define SBP2_WARN(fmt, args...)		HPSB_WARN(fmt, ## args)
+#else 
+#define SBP2_DEBUG(fmt, args...)	
+#define SBP2_INFO(fmt, args...)	
+#define SBP2_NOTICE(fmt, args...)	
+#define SBP2_WARN(fmt, args...)	
+#endif
+
+#define SBP2_ERR(fmt, args...)		HPSB_ERR(fmt, ## args)
+
+/*
+ * Spinlock debugging stuff. I'm playing it safe until the driver has been debugged on SMP. (JSG)
+ */
+/* #define SBP2_USE_REAL_SPINLOCKS */
+#ifdef SBP2_USE_REAL_SPINLOCKS
+#define sbp2_spin_lock(lock, flags)	spin_lock_irqsave(lock, flags)	
+#define sbp2_spin_unlock(lock, flags)	spin_unlock_irqrestore(lock, flags);
+#else
+#define sbp2_spin_lock(lock, flags)	do {save_flags(flags); cli();} while (0)	
+#define sbp2_spin_unlock(lock, flags)	do {restore_flags(flags);} while (0)
+#endif
+
+/*
+ * Globals
+ */
+
+Scsi_Host_Template *global_scsi_tpnt = NULL;
+
+LIST_HEAD(sbp2_host_info_list);
+static int sbp2_host_count = 0;
+spinlock_t sbp2_host_info_lock = SPIN_LOCK_UNLOCKED;
+
+static struct hpsb_highlevel *sbp2_hl_handle = NULL;
+
+static struct hpsb_highlevel_ops sbp2_hl_ops = {
+	sbp2_add_host,
+	sbp2_remove_host,
+	sbp2_host_reset,
+	NULL,
+	NULL
+};
+
+static struct hpsb_address_ops sbp2_ops = {
+	write: sbp2_handle_status_write,
+};
+
+#if 0
+static struct hpsb_address_ops sbp2_physdma_ops = {
+	read: sbp2_handle_physdma_read,
+	write: sbp2_handle_physdma_write,
+};
+#endif
+
+/**************************************
+ * General utility functions
+ **************************************/
+
+
+#ifndef __BIG_ENDIAN
+/*
+ * Converts a buffer from be32 to cpu byte ordering. Length is in bytes.
+ */
+static __inline__ void sbp2util_be32_to_cpu_buffer(void *buffer, int length)
+{
+	u32 *temp = buffer;
+
+	for (length = (length >> 2); length--; )
+		temp[length] = be32_to_cpu(temp[length]);
+
+	return;
+}
+
+/*
+ * Converts a buffer from cpu to be32 byte ordering. Length is in bytes.
+ */
+static __inline__ void sbp2util_cpu_to_be32_buffer(void *buffer, int length)
+{
+	u32 *temp = buffer;
+
+	for (length = (length >> 2); length--; )
+		temp[length] = cpu_to_be32(temp[length]);
+
+	return;
+}
+#else /* BIG_ENDIAN */
+/* Why waste the cpu cycles? */
+#define sbp2util_be32_to_cpu_buffer(x,y)
+#define sbp2util_cpu_to_be32_buffer(x,y)
+#endif
+
+/*
+ * This function does quadlet sized reads (used by detection code)
+ */
+static int sbp2util_read_quadlet(struct sbp2scsi_host_info *hi, nodeid_t node, u64 addr,
+				 quadlet_t *buffer)
+{
+	int retval = 0;
+	int retry_count = 3;
+
+	/*
+	 * Retry a couple times if needed (for slow devices)
+	 */
+	do {
+
+		retval = hpsb_read(hi->host, node, addr, buffer, 4);
+
+		if (retval) {
+			SBP2_DEBUG("sbp2: sbp2util_read_quadlet data packet error");
+			current->state = TASK_INTERRUPTIBLE;
+			schedule_timeout(HZ/50);	/* 20ms delay */
+		}
+
+		retry_count--;
+
+	} while (retval && retry_count);
+
+	return(retval);
+}
+
+/*
+ * This function returns the address of the unit directory.
+ */
+static int sbp2util_unit_directory(struct sbp2scsi_host_info *hi, nodeid_t node_id, u64 *unit_directory_addr)
+{
+	quadlet_t root_directory_length, current_quadlet;
+	u64 current_addr;
+	int length, i;
+
+	/*
+	 * First, read the first quadlet of the root directory to determine its size
+	 */
+	if (sbp2util_read_quadlet(hi, LOCAL_BUS | node_id, CONFIG_ROM_ROOT_DIR_BASE, 
+				  &root_directory_length)) {
+		SBP2_DEBUG("sbp2: Error reading root directory length - bad status");
+		return(-EIO);   
+	}
+
+	current_addr = CONFIG_ROM_ROOT_DIR_BASE;
+	length = be32_to_cpu(root_directory_length) >> 16;
+
+	/*
+	 * Step through the root directory and look for the "Unit_Directory entry", which
+	 * contains the offset to the unit directory.
+	 */
+	for (i=0; i < length; i++) {
+
+		current_addr += 4;
+
+		if (sbp2util_read_quadlet(hi, LOCAL_BUS | node_id, current_addr, &current_quadlet)) {
+			SBP2_DEBUG("sbp2: Error reading at address 0x%08x%08x - bad status", 
+				   (unsigned int) ((current_addr) >> 32), (unsigned int) ((current_addr) & 0xffffffff));
+			return(-EIO);   
+		}
+
+		/*  
+		 * Check for unit directory offset tag 
+		 */
+		if ((be32_to_cpu(current_quadlet) >> 24) == SBP2_UNIT_DIRECTORY_OFFSET_KEY) {
+			*unit_directory_addr = current_addr + 4 * ((be32_to_cpu(current_quadlet) & 0xffffff));
+			SBP2_DEBUG("sbp2: unit_directory_addr = %lu", *unit_directory_addr);
+		}
+	}
+
+	return(0);
+}
+
+/*
+ * This function is called to initially create a packet pool for use in sbp2 I/O requests.
+ * This packet pool is used when sending out sbp2 command and agent reset requests, and 
+ * allows us to remove all kmallocs/kfrees from the critical I/O paths.
+ */
+static int sbp2util_create_request_packet_pool(struct sbp2scsi_host_info *hi)
+{
+	struct hpsb_packet *packet;
+	int i;
+	unsigned long flags;
+
+	/*
+	 * Create SBP2_MAX_REQUEST_PACKETS number of request packets.
+	 */
+	sbp2_spin_lock(&hi->sbp2_request_packet_lock, flags);
+	for (i=0; i<SBP2_MAX_REQUEST_PACKETS; i++) {
+
+		/*
+		 * Max payload of 8 bytes since the sbp2 command request uses a payload of 
+		 * 8 bytes, and agent reset is a quadlet write request. Bump this up if we
+		 * plan on using this pool for other stuff.
+		 */
+		packet = alloc_hpsb_packet(8);
+
+		if (!packet) {
+			SBP2_ERR("sbp2: sbp2util_create_request_packet_pool - packet allocation failed!");
+			sbp2_spin_unlock(&hi->sbp2_request_packet_lock, flags);
+			return(-ENOMEM);
+		}
+
+		/* 
+		 * Put these request packets into a free list
+		 */
+		INIT_LIST_HEAD(&hi->request_packet[i].list);
+		hi->request_packet[i].packet = packet;
+		list_add_tail(&hi->request_packet[i].list, &hi->sbp2_req_free);
+
+	}
+	sbp2_spin_unlock(&hi->sbp2_request_packet_lock, flags);
+
+	return(0);
+}
+
+/*
+ * This function is called to remove the packet pool. It is called when the sbp2 driver is unloaded.
+ */
+static void sbp2util_remove_request_packet_pool(struct sbp2scsi_host_info *hi)
+{
+	struct list_head *lh;
+	struct sbp2_request_packet *request_packet;
+	unsigned long flags;
+
+	/* 
+	 * Go through free list releasing packets
+	 */
+	sbp2_spin_lock(&hi->sbp2_request_packet_lock, flags);
+	while (!list_empty(&hi->sbp2_req_free)) {
+
+		lh = hi->sbp2_req_free.next;
+		list_del(lh);
+
+		request_packet = list_entry(lh, struct sbp2_request_packet, list);
+
+		/*
+		 * Free the hpsb packets that we allocated for the pool
+		 */
+		if (request_packet) {
+			free_hpsb_packet(request_packet->packet);
+		}
+
+	}
+	sbp2_spin_unlock(&hi->sbp2_request_packet_lock, flags);
+
+	return;
+}
+
+/*
+ * This function is called to retrieve a block write packet from our packet pool. This function is
+ * used in place of calling alloc_hpsb_packet (which costs us three kmallocs). Instead we 
+ * just pull out a free request packet and re-initialize values in it. I'm sure this can still
+ * stand some more optimization. 
+ */
+static struct sbp2_request_packet *sbp2util_allocate_write_request_packet(struct sbp2scsi_host_info *hi,
+									  nodeid_t node, u64 addr,
+									  size_t data_size,
+									  quadlet_t data) {
+	struct list_head *lh;
+	struct sbp2_request_packet *request_packet = NULL;
+	struct hpsb_packet *packet;
+	unsigned long flags;
+
+	sbp2_spin_lock(&hi->sbp2_request_packet_lock, flags);
+	if (!list_empty(&hi->sbp2_req_free)) {
+
+		/*
+		 * Pull out a free request packet
+		 */
+		lh = hi->sbp2_req_free.next;
+		list_del(lh);
+
+		request_packet = list_entry(lh, struct sbp2_request_packet, list);
+		packet = request_packet->packet;
+
+		/*
+		 * Initialize the packet (this is really initialization the core 1394 stack should do,
+		 * but I'm doing it myself to avoid the overhead).
+		 */
+		packet->data_size = data_size;
+		INIT_LIST_HEAD(&packet->list);
+		sema_init(&packet->state_change, 0);
+		packet->state = unused;
+		packet->generation = get_hpsb_generation();
+		packet->data_be = 1;
+
+		packet->host = hi->host;
+		packet->tlabel = get_tlabel(hi->host, node, 1);
+		packet->node_id = node;
+
+		if (!data_size) {
+			fill_async_writequad(packet, addr, data);
+		} else {
+			fill_async_writeblock(packet, addr, data_size);         
+		}
+
+		/*
+		 * Set up a task queue completion routine, which returns the packet to the free list
+		 * and releases the tlabel
+		 */
+		request_packet->tq.routine = (void (*)(void*))sbp2util_free_request_packet;
+		request_packet->tq.data = request_packet;
+		request_packet->hi_context = hi;
+		queue_task(&request_packet->tq, &packet->complete_tq);
+
+		/*
+		 * Now, put the packet on the in-use list
+		 */
+		list_add_tail(&request_packet->list, &hi->sbp2_req_inuse);
+
+	} else {
+		SBP2_ERR("sbp2: sbp2util_allocate_request_packet - no packets available!");
+	}
+	sbp2_spin_unlock(&hi->sbp2_request_packet_lock, flags);
+
+	return(request_packet);
+}
+
+/*
+ * This function is called to return a packet to our packet pool. It is also called as a 
+ * completion routine when a request packet is completed.
+ */
+static void sbp2util_free_request_packet(struct sbp2_request_packet *request_packet)
+{
+	unsigned long flags;
+	struct sbp2scsi_host_info *hi = request_packet->hi_context;
+
+	/*
+	 * Free the tlabel, and return the packet to the free pool
+	 */
+	sbp2_spin_lock(&hi->sbp2_request_packet_lock, flags);
+	free_tlabel(hi->host, LOCAL_BUS | request_packet->packet->node_id,
+		    request_packet->packet->tlabel);
+	list_del(&request_packet->list);
+	list_add_tail(&request_packet->list, &hi->sbp2_req_free);
+	sbp2_spin_unlock(&hi->sbp2_request_packet_lock, flags);
+
+	return;
+}
+
+/*
+ * This function is called to create a pool of command orbs used for command processing. It is called
+ * when a new sbp2 device is detected.
+ */
+static int sbp2util_create_command_orb_pool(struct scsi_id_instance_data *scsi_id,
+					    struct sbp2scsi_host_info *hi)
+{
+	int i;
+	unsigned long flags;
+	struct sbp2_command_info *command;
+        
+	sbp2_spin_lock(&scsi_id->sbp2_command_orb_lock, flags);
+	for (i = 0; i < scsi_id->sbp2_total_command_orbs; i++) {
+		command = (struct sbp2_command_info *)
+		    kmalloc(sizeof(struct sbp2_command_info), GFP_KERNEL);
+		if (!command) {
+			sbp2_spin_unlock(&scsi_id->sbp2_command_orb_lock, flags);
+			return(-ENOMEM);
+		}
+		memset(command, '\0', sizeof(struct sbp2_command_info));
+		command->command_orb_dma =
+			pci_map_single (hi->host->pdev, &command->command_orb,
+					sizeof(struct sbp2_command_orb),
+					PCI_DMA_BIDIRECTIONAL);
+		SBP2_DMA_ALLOC("single command orb DMA");
+		command->sge_dma =
+			pci_map_single (hi->host->pdev, &command->scatter_gather_element,
+					sizeof(command->scatter_gather_element),
+					PCI_DMA_BIDIRECTIONAL);
+		SBP2_DMA_ALLOC("scatter_gather_element");
+		INIT_LIST_HEAD(&command->list);
+		list_add_tail(&command->list, &scsi_id->sbp2_command_orb_completed);
+	}
+	sbp2_spin_unlock(&scsi_id->sbp2_command_orb_lock, flags);
+	return 0;
+}
+
+/*
+ * This function is called to delete a pool of command orbs.
+ */
+static void sbp2util_remove_command_orb_pool(struct scsi_id_instance_data *scsi_id,
+					     struct sbp2scsi_host_info *hi)
+{
+	struct list_head *lh;
+	struct sbp2_command_info *command;
+	unsigned long flags;
+        
+	sbp2_spin_lock(&scsi_id->sbp2_command_orb_lock, flags);
+	if (!list_empty(&scsi_id->sbp2_command_orb_completed)) {
+		list_for_each(lh, &scsi_id->sbp2_command_orb_completed) {
+			command = list_entry(lh, struct sbp2_command_info, list);
+
+			/* Release our generic DMA's */
+			pci_unmap_single(hi->host->pdev, command->command_orb_dma,
+					 sizeof(struct sbp2_command_orb),
+					 PCI_DMA_BIDIRECTIONAL);
+			SBP2_DMA_FREE("single command orb DMA");
+			pci_unmap_single(hi->host->pdev, command->sge_dma,
+					 sizeof(command->scatter_gather_element),
+					 PCI_DMA_BIDIRECTIONAL);
+			SBP2_DMA_FREE("scatter_gather_element");
+
+			kfree(command);
+		}
+	}
+	sbp2_spin_unlock(&scsi_id->sbp2_command_orb_lock, flags);
+	return;
+}
+
+/* 
+ * This functions finds the sbp2_command for a given outstanding
+ * command orb. Only looks at the inuse list.
+ */
+static struct sbp2_command_info *sbp2util_find_command_for_orb(struct scsi_id_instance_data *scsi_id, dma_addr_t orb)
+{
+	struct list_head *lh;
+	struct sbp2_command_info *command;
+	unsigned long flags;
+
+	sbp2_spin_lock(&scsi_id->sbp2_command_orb_lock, flags);
+	if (!list_empty(&scsi_id->sbp2_command_orb_inuse)) {
+		list_for_each(lh, &scsi_id->sbp2_command_orb_inuse) {
+			command = list_entry(lh, struct sbp2_command_info, list);
+			if (command->command_orb_dma == orb) {
+				sbp2_spin_unlock(&scsi_id->sbp2_command_orb_lock, flags);
+				return (command);
+			}
+		}
+	}
+	sbp2_spin_unlock(&scsi_id->sbp2_command_orb_lock, flags);
+
+	SBP2_ORB_DEBUG("could not match command orb %x", (unsigned int)orb);
+
+	return(NULL);
+}
+
+/* 
+ * This functions finds the sbp2_command for a given outstanding SCpnt. Only looks at the inuse list 
+ */
+static struct sbp2_command_info *sbp2util_find_command_for_SCpnt(struct scsi_id_instance_data *scsi_id, void *SCpnt)
+{
+	struct list_head *lh;
+	struct sbp2_command_info *command;
+	unsigned long flags;
+
+	sbp2_spin_lock(&scsi_id->sbp2_command_orb_lock, flags);
+	if (!list_empty(&scsi_id->sbp2_command_orb_inuse)) {
+		list_for_each(lh, &scsi_id->sbp2_command_orb_inuse) {
+			command = list_entry(lh, struct sbp2_command_info, list);
+			if (command->Current_SCpnt == SCpnt) {
+				sbp2_spin_unlock(&scsi_id->sbp2_command_orb_lock, flags);
+				return (command);
+			}
+		}
+	}
+	sbp2_spin_unlock(&scsi_id->sbp2_command_orb_lock, flags);
+	return(NULL);
+}
+
+/*
+ * This function allocates a command orb used to send a scsi command.
+ */
+static struct sbp2_command_info *sbp2util_allocate_command_orb(struct scsi_id_instance_data *scsi_id, 
+							       Scsi_Cmnd *Current_SCpnt, 
+							       void (*Current_done)(Scsi_Cmnd *),
+							       struct sbp2scsi_host_info *hi)
+{
+	struct list_head *lh;
+	struct sbp2_command_info *command = NULL;
+	unsigned long flags;
+
+	sbp2_spin_lock(&scsi_id->sbp2_command_orb_lock, flags);
+	if (!list_empty(&scsi_id->sbp2_command_orb_completed)) {
+		lh = scsi_id->sbp2_command_orb_completed.next;
+		list_del(lh);
+		command = list_entry(lh, struct sbp2_command_info, list);
+		command->Current_done = Current_done;
+		command->Current_SCpnt = Current_SCpnt;
+		command->linked = 0;
+		list_add_tail(&command->list, &scsi_id->sbp2_command_orb_inuse);
+	} else {
+		SBP2_ERR("sbp2: sbp2util_allocate_command_orb - No orbs available!");
+	}
+	sbp2_spin_unlock(&scsi_id->sbp2_command_orb_lock, flags);
+	return (command);
+}
+
+/* Free our DMA's */
+static void sbp2util_free_command_dma(struct sbp2_command_info *command)
+{
+	struct sbp2scsi_host_info *hi;
+	
+	hi = (struct sbp2scsi_host_info *) command->Current_SCpnt->host->hostdata[0];
+
+	if (hi == NULL) {
+		printk(KERN_ERR __FUNCTION__": hi == NULL\n");
+		return;
+	}
+
+	if (command->cmd_dma) {
+		pci_unmap_single(hi->host->pdev, command->cmd_dma,
+				 command->dma_size, command->dma_dir);
+		SBP2_DMA_FREE("single bulk");
+		command->cmd_dma = 0;
+	}
+
+	if (command->sge_buffer) {
+		pci_unmap_sg(hi->host->pdev, command->sge_buffer,
+			     command->dma_size, command->dma_dir);
+		SBP2_DMA_FREE("scatter list");
+		command->sge_buffer = NULL;
+	}
+}
+
+/*
+ * This function moves a command to the completed orb list.
+ */
+static void sbp2util_mark_command_completed(struct scsi_id_instance_data *scsi_id, struct sbp2_command_info *command)
+{
+	unsigned long flags;
+
+	sbp2_spin_lock(&scsi_id->sbp2_command_orb_lock, flags);
+	list_del(&command->list);
+	sbp2util_free_command_dma(command);
+	list_add_tail(&command->list, &scsi_id->sbp2_command_orb_completed);
+	sbp2_spin_unlock(&scsi_id->sbp2_command_orb_lock, flags);
+}
+
+/*********************************************
+ * IEEE-1394 core driver stack related section
+ *********************************************/
+
+/*
+ * This function is called at SCSI init in order to register our driver with the
+ * IEEE-1394 stack
+ */
+int sbp2_init(void)
+{
+	SBP2_DEBUG("sbp2: sbp2_init");
+
+	/*
+	 * Register our high level driver with 1394 stack
+	 */
+	sbp2_hl_handle = hpsb_register_highlevel(SBP2_DEVICE_NAME, &sbp2_hl_ops);
+
+	if (sbp2_hl_handle == NULL) {
+		SBP2_ERR("sbp2: sbp2 failed to register with ieee1394 highlevel");
+		return(-ENOMEM);
+	}
+
+	/*
+	 * Register our sbp2 status address space...
+	 */
+	hpsb_register_addrspace(sbp2_hl_handle, &sbp2_ops, SBP2_STATUS_FIFO_ADDRESS,
+				SBP2_STATUS_FIFO_ADDRESS + sizeof(struct sbp2_status_block));
+
+	/*
+	 * Register physical dma address space... used for
+	 * adapters not supporting hardware phys dma.
+	 *
+	 * XXX: Disabled for now.
+	 */
+	/* hpsb_register_addrspace(sbp2_hl_handle, &sbp2_physdma_ops,
+				   0x0ULL, 0xfffffffcULL); */
+
+	return(0);
+}
+
+/*
+ * This function is called from cleanup module, or during shut-down, in order to 
+ * unregister our driver
+ */
+void sbp2_cleanup(void)
+{
+	SBP2_DEBUG("sbp2: sbp2_cleanup");
+
+	if (sbp2_hl_handle) {
+		hpsb_unregister_highlevel(sbp2_hl_handle);
+		sbp2_hl_handle = NULL;
+	}
+	return;
+}
+
+/*
+ * This function is called after registering our operations in sbp2_init. We go ahead and
+ * allocate some memory for our host info structure, and init some structures.
+ */
+static void sbp2_add_host(struct hpsb_host *host)
+{
+	struct sbp2scsi_host_info *hi;
+	unsigned int flags;
+
+	SBP2_DEBUG("sbp2: sbp2_add_host");
+
+	/*
+	 * Allocate some memory for our host info structure
+	 */
+	hi = (struct sbp2scsi_host_info *)kmalloc(sizeof(struct sbp2scsi_host_info), GFP_KERNEL);
+
+	if (hi != NULL) {
+
+		/*
+		 * Initialize some host stuff
+		 */
+		memset(hi, 0, sizeof(struct sbp2scsi_host_info));
+		INIT_LIST_HEAD(&hi->list);
+		INIT_LIST_HEAD(&hi->sbp2_req_inuse);
+		INIT_LIST_HEAD(&hi->sbp2_req_free);
+		hi->host = host;
+		hi->sbp2_command_lock = SPIN_LOCK_UNLOCKED;
+		hi->sbp2_request_packet_lock = SPIN_LOCK_UNLOCKED;
+
+		/*
+		 * Create our request packet pool (pool of packets for use in I/O)
+		 */
+		if (sbp2util_create_request_packet_pool(hi)) {
+			SBP2_ERR("sbp2: sbp2util_create_request_packet_pool failed!");
+			return;
+		}
+
+		sbp2_spin_lock(&sbp2_host_info_lock, flags);
+		list_add_tail(&hi->list, &sbp2_host_info_list);
+		sbp2_host_count++;
+		sbp2_spin_lock(&sbp2_host_info_lock, flags);
+
+		/*
+		 * Initialize us to bus reset in progress
+		 */
+		hi->bus_reset_in_progress = 1;
+
+		/*
+		 * Register our host with the SCSI stack. 
+		 */
+		sbp2scsi_register_scsi_host(hi);
+
+		/*
+		 * Start our kernel thread to deal with sbp2 device detection
+		 */
+		init_waitqueue_head(&hi->sbp2_detection_wait);
+		hi->sbp2_detection_pid = 0;
+		hi->sbp2_detection_pid = kernel_thread(sbp2_detection_thread, hi, CLONE_FS | CLONE_FILES | CLONE_SIGHAND);
+
+	}
+
+	return;
+}
+
+/*
+ * This fuction returns a host info structure from the host structure, in case we have multiple hosts
+ */
+static struct sbp2scsi_host_info *sbp2_find_host_info(struct hpsb_host *host) {
+	struct list_head *lh;
+	struct sbp2scsi_host_info *hi;
+
+	lh = sbp2_host_info_list.next;
+	while (lh != &sbp2_host_info_list) {
+		hi = list_entry(lh, struct sbp2scsi_host_info, list);
+		if (hi->host == host) {
+			return hi;
+		}
+		lh = lh->next;
+	}
+
+	return(NULL);
+}
+
+/*
+ * This function is called when the host is removed
+ */
+static void sbp2_remove_host(struct hpsb_host *host)
+{
+	struct sbp2scsi_host_info *hi;
+	int i;
+	unsigned int flags;
+
+	SBP2_DEBUG("sbp2: sbp2_remove_host");
+
+	sbp2_spin_lock(&sbp2_host_info_lock, flags);
+	hi = sbp2_find_host_info(host);
+
+	if (hi != NULL) {
+
+		/*
+		 * Need to remove any attached SBP-2 devices. Also make sure to logout of all devices.
+		 */
+		for (i=0; i<SBP2SCSI_MAX_SCSI_IDS; i++) {
+			if (hi->scsi_id[i]) {
+				sbp2_logout_device(hi, hi->scsi_id[i]);
+				hi->scsi_id[i]->validated = 0;
+			}
+		}
+
+		sbp2_remove_unvalidated_devices(hi);
+
+		list_del(&hi->list);
+		sbp2_host_count--;
+	}
+	sbp2_spin_unlock(&sbp2_host_info_lock, flags);
+
+	if (hi == NULL) {
+		SBP2_ERR("sbp2: attempt to remove unknown host %p", host);
+		return;
+	}
+
+	/*
+	 * Remove the packet pool (release the packets)
+	 */
+	sbp2util_remove_request_packet_pool(hi);
+
+	/* 
+	 * Kill our detection thread 
+	 */
+	if (hi->sbp2_detection_pid >= 0) {
+		kill_proc(hi->sbp2_detection_pid, SIGINT, 1);
+	}
+
+	/*
+	 * Give the detection thread a little time to exit
+	 */
+	current->state = TASK_INTERRUPTIBLE;
+	schedule_timeout(HZ);	/* 1 second delay */
+
+	kfree(hi);
+	hi = NULL;
+
+	return;
+}
+
+/*
+ * This is our sbp2 detection thread. It is signalled when bus resets occur
+ * so that we can find and initialize any sbp2 devices. 
+ */
+static int sbp2_detection_thread(void *__hi)
+{
+	struct sbp2scsi_host_info *hi = (struct sbp2scsi_host_info *)__hi;
+
+	SBP2_DEBUG("sbp2: sbp2_detection_thread");
+
+	lock_kernel();
+
+	/*
+	 * This thread doesn't need any user-level access,
+	 * so get rid of all our resources
+	 */
+	daemonize();
+
+	/* 
+	 * Set-up a nice name
+	 */
+	strcpy(current->comm, "sbp2");
+
+	unlock_kernel();
+        
+	while ((!signal_pending(current)) && hi) {
+
+		/*
+		 * Process our bus reset now
+		 */
+		if (hi) {
+			MOD_INC_USE_COUNT;
+			sbp2_bus_reset_handler(hi);
+			MOD_DEC_USE_COUNT;
+		}
+
+		/*
+		 * Sleep until next bus reset
+		 */
+		if (hi) {
+			interruptible_sleep_on(&hi->sbp2_detection_wait);
+		}
+	}
+
+	return(0);
+}
+
+/*
+ * This function is where we first pull the node unique ids, and then allocate memory and register
+ * a SBP-2 device
+ */
+static int sbp2_start_device(struct sbp2scsi_host_info *hi, int node_id)
+{
+	quadlet_t node_unique_id_lo, node_unique_id_hi;
+	u64 node_unique_id;
+	struct scsi_id_instance_data *scsi_id = NULL;
+	int i;
+
+	SBP2_DEBUG("sbp2: sbp2_start_device");
+
+	/*
+	 * Let's read the node unique id off of the device (using two quadlet reads for hi and lo)
+	 */
+	if (sbp2util_read_quadlet(hi, LOCAL_BUS | node_id, CONFIG_ROM_NODE_UNIQUE_ID_HI_ADDRESS, 
+				  &node_unique_id_hi)) {
+		SBP2_DEBUG("sbp2: Error reading node unique id - bad status");
+		return(-EIO);
+	}
+
+	if (sbp2util_read_quadlet(hi, LOCAL_BUS | node_id, CONFIG_ROM_NODE_UNIQUE_ID_LO_ADDRESS, 
+				  &node_unique_id_lo)) {
+		SBP2_DEBUG("sbp2: Error reading node unique id - bad status");
+		return(-EIO);
+	}
+
+	/*
+	 * Spit out the node unique ids we got
+	 */
+	SBP2_DEBUG("sbp2: Node %x, node unique id hi = %x", (LOCAL_BUS | node_id), (unsigned int) node_unique_id_hi);
+	SBP2_DEBUG("sbp2: Node %x, node unique id lo = %x", (LOCAL_BUS | node_id), (unsigned int) node_unique_id_lo);
+
+	node_unique_id = (((u64)node_unique_id_hi) << 32) | ((u64)node_unique_id_lo);
+
+	/*
+	 * First, we need to find out whether this is a "new" SBP-2 device plugged in, or one that already
+	 * exists and is initialized. We do this by looping through our scsi id instance data structures
+	 * looking for matching node unique ids.
+	 */
+	for (i=0; i<SBP2SCSI_MAX_SCSI_IDS; i++) {
+
+		if (hi->scsi_id[i]) {
+
+			if (hi->scsi_id[i]->node_unique_id == node_unique_id) {
+
+				/*
+				 * Update our node id
+				 */
+				hi->scsi_id[i]->node_id = node_id;
+
+				/*
+				 * Mark the device as validated, since it still exists on the bus
+				 */
+				hi->scsi_id[i]->validated = 1;
+				SBP2_DEBUG("sbp2: SBP-2 device re-validated, SCSI ID = %x", (unsigned int) i);
+
+				/*
+				 * Reconnect to the sbp-2 device
+				 */
+				if (sbp2_reconnect_device(hi, hi->scsi_id[i])) {
+
+					/*
+					 * Ok, reconnect has failed. Perhaps we didn't reconnect fast enough. Try
+					 * doing a regular login.
+					 */
+					if (sbp2_login_device(hi, hi->scsi_id[i])) {
+
+						/*
+						 * Login failed too... so, just mark him as unvalidated, so that he gets cleaned up
+						 * later
+						 */
+						SBP2_ERR("sbp2: sbp2_reconnect_device failed!");
+						hi->scsi_id[i]->validated = 0;
+					}
+				}
+
+				if (hi->scsi_id[i]->validated) {
+
+					/*
+					 * Set max retries to something large on the device
+					 */
+					sbp2_set_busy_timeout(hi, hi->scsi_id[i]);
+
+					/*
+					 * Do a SBP-2 fetch agent reset
+					 */
+					sbp2_agent_reset(hi, hi->scsi_id[i], 0);
+
+					/*
+					 * Get the max speed and packet size that we can use
+					 */
+					sbp2_max_speed_and_size(hi, hi->scsi_id[i]);
+
+				}
+
+				/*
+				 * Nothing more to do, since we found the device
+				 */
+				return(0);
+
+			}
+		}
+	}
+
+	/*
+	 * This really is a "new" device plugged in. Let's allocate memory for our scsi id instance data
+	 */
+	scsi_id = (struct scsi_id_instance_data *)kmalloc(sizeof(struct scsi_id_instance_data),
+							  GFP_KERNEL);
+	if (!scsi_id)
+		goto alloc_fail_first;
+	memset(scsi_id, 0, sizeof(struct scsi_id_instance_data));
+
+	/* Login FIFO DMA */
+	scsi_id->login_response =
+		pci_alloc_consistent(hi->host->pdev, sizeof(struct sbp2_login_response),
+				     &scsi_id->login_response_dma);
+	if (!scsi_id->login_response)
+		goto alloc_fail;
+	SBP2_DMA_ALLOC("consistent DMA region for login FIFO");
+
+	/* Reconnect ORB DMA */
+	scsi_id->reconnect_orb =
+		pci_alloc_consistent(hi->host->pdev, sizeof(struct sbp2_reconnect_orb),
+				     &scsi_id->reconnect_orb_dma);
+	if (!scsi_id->reconnect_orb)
+		goto alloc_fail;
+	SBP2_DMA_ALLOC("consistent DMA region for reconnect ORB");
+
+	/* Logout ORB DMA */
+	scsi_id->logout_orb =
+		pci_alloc_consistent(hi->host->pdev, sizeof(struct sbp2_logout_orb),
+				     &scsi_id->logout_orb_dma);
+	if (!scsi_id->logout_orb)
+		goto alloc_fail;
+	SBP2_DMA_ALLOC("consistent DMA region for logout ORB");
+
+	/* Login ORB DMA */
+	scsi_id->login_orb =
+		pci_alloc_consistent(hi->host->pdev, sizeof(struct sbp2_login_orb),
+				     &scsi_id->login_orb_dma);
+	if (scsi_id->login_orb == NULL) {
+alloc_fail:
+		if (scsi_id->logout_orb) {
+			pci_free_consistent(hi->host->pdev,
+					sizeof(struct sbp2_logout_orb),
+					scsi_id->logout_orb,
+					scsi_id->logout_orb_dma);
+			SBP2_DMA_FREE("logout ORB DMA");
+		}
+
+		if (scsi_id->reconnect_orb) {
+			pci_free_consistent(hi->host->pdev,
+					sizeof(struct sbp2_reconnect_orb),
+					scsi_id->reconnect_orb,
+					scsi_id->reconnect_orb_dma);
+			SBP2_DMA_FREE("reconnect ORB DMA");
+		}
+
+		if (scsi_id->login_response) {
+			pci_free_consistent(hi->host->pdev,
+					sizeof(struct sbp2_login_response),
+					scsi_id->login_response,
+					scsi_id->login_response_dma);
+			SBP2_DMA_FREE("login FIFO DMA");
+		}
+
+		kfree(scsi_id);
+alloc_fail_first:
+		SBP2_ERR ("sbp2: Could not allocate memory for scsi_id");
+		return(-ENOMEM);
+	}
+	SBP2_DMA_ALLOC("consistent DMA region for login ORB");
+
+	/*
+	 * Initialize some of the fields in this structure
+	 */
+	scsi_id->node_id = node_id;
+	scsi_id->node_unique_id = node_unique_id;
+	scsi_id->validated = 1;
+	scsi_id->speed_code = SPEED_S100;
+	scsi_id->max_payload_size = MAX_PAYLOAD_S100;
+
+	init_waitqueue_head(&scsi_id->sbp2_login_wait);
+
+	/* 
+	 * Initialize structures needed for the command orb pool.
+	 */
+	INIT_LIST_HEAD(&scsi_id->sbp2_command_orb_inuse);
+	INIT_LIST_HEAD(&scsi_id->sbp2_command_orb_completed);
+	scsi_id->sbp2_command_orb_lock = SPIN_LOCK_UNLOCKED;
+	scsi_id->sbp2_total_command_orbs = 0;
+
+	/*
+	 * Make sure that we've gotten ahold of the sbp2 management agent address. Also figure out the
+	 * command set being used (SCSI or RBC).
+	 */
+	if (sbp2_parse_unit_directory(hi, scsi_id)) {
+		SBP2_ERR("sbp2: Error while parsing sbp2 unit directory");
+		hi->scsi_id[i]->validated = 0;
+		return(-EIO);
+	}
+
+	scsi_id->sbp2_total_command_orbs = SBP2_MAX_COMMAND_ORBS;
+
+	/* 
+	 * Knock the total command orbs down if we are serializing I/O
+	 */
+	if (serialize_io) {
+		scsi_id->sbp2_total_command_orbs = 2;	/* one extra for good measure */
+	}
+
+	/*
+	 * Allocate some extra command orb structures for devices with 128KB limit
+	 */
+	if (scsi_id->sbp2_firmware_revision == SBP2_128KB_BROKEN_FIRMWARE) {
+		scsi_id->sbp2_total_command_orbs *= 4;
+	} 
+
+	/*
+	 * Create our command orb pool
+	 */
+	if (sbp2util_create_command_orb_pool(scsi_id, hi)) {
+		SBP2_ERR("sbp2: sbp2util_create_command_orb_pool failed!");
+		hi->scsi_id[i]->validated = 0;
+		return (-ENOMEM);
+	}
+
+	/*
+	 * Find an empty spot to stick our scsi id instance data. 
+	 */
+	for (i=0; i<SBP2SCSI_MAX_SCSI_IDS; i++) {
+		if (!hi->scsi_id[i]) {
+			hi->scsi_id[i] = scsi_id;
+			SBP2_DEBUG("sbp2: New SBP-2 device inserted, SCSI ID = %x", (unsigned int) i);
+			break;
+		}
+	}
+
+	/*
+	 * Make sure we are not out of space
+	 */
+	if (i >= SBP2SCSI_MAX_SCSI_IDS) {
+		SBP2_ERR("sbp2: No slots left for SBP-2 device");
+		hi->scsi_id[i]->validated = 0;
+		return(-EBUSY);
+	}
+
+	/*
+	 * Login to the sbp-2 device
+	 */
+	if (sbp2_login_device(hi, hi->scsi_id[i])) {
+
+		/*
+		 * Login failed... so, just mark him as unvalidated, so that he gets cleaned up later
+		 */
+		SBP2_ERR("sbp2: sbp2_login_device failed");
+		hi->scsi_id[i]->validated = 0;
+	}
+
+	if (hi->scsi_id[i]->validated) {
+
+		/*
+		 * Set max retries to something large on the device
+		 */
+		sbp2_set_busy_timeout(hi, hi->scsi_id[i]);
+
+		/*
+		 * Do a SBP-2 fetch agent reset
+		 */
+		sbp2_agent_reset(hi, hi->scsi_id[i], 0);
+
+		/*
+		 * Get the max speed and packet size that we can use
+		 */
+		sbp2_max_speed_and_size(hi, hi->scsi_id[i]);
+
+	}
+
+	return(0);
+}
+
+/*
+ * This function trys to determine if a device is a valid SBP-2 device
+ */
+static int sbp2_check_device(struct sbp2scsi_host_info *hi, int node_id)
+{
+	quadlet_t unit_spec_id_data = 0, unit_sw_ver_data = 0;
+	quadlet_t unit_directory_length, current_quadlet;
+	u64 unit_directory_addr, current_addr;
+	unsigned int i, length;
+
+	SBP2_DEBUG("sbp2: sbp2_check_device");
+
+	/*
+	 * Let's try and read the unit spec id and unit sw ver to determine if this is an SBP2 device...
+	 */
+
+	if (sbp2util_unit_directory(hi, LOCAL_BUS | node_id, &unit_directory_addr)) {
+		SBP2_DEBUG("sbp2: Error reading unit directory address - bad status");
+		return(-EIO);   
+	}
+
+	/*
+	 * Read the size of the unit directory
+	 */
+	if (sbp2util_read_quadlet(hi, LOCAL_BUS | node_id, unit_directory_addr, 
+				  &unit_directory_length)) {
+		SBP2_DEBUG("sbp2: Error reading root directory length - bad status");
+		return(-EIO);   
+	}
+
+	current_addr = unit_directory_addr;
+	length = be32_to_cpu(unit_directory_length) >> 16;
+
+	/*
+	 * Now, step through the unit directory and look for the unit_spec_ID and the unit_sw_version
+	 */
+	for (i=0; i < length; i++) {
+
+		current_addr += 4;
+
+		if (sbp2util_read_quadlet(hi, LOCAL_BUS | node_id, current_addr, &current_quadlet)) {
+			SBP2_DEBUG("sbp2: Error reading at address 0x%08x%08x - bad status", 
+				   (unsigned int) ((current_addr) >> 32), (unsigned int) ((current_addr) & 0xffffffff));
+			return(-EIO);   
+		}
+
+		/* 
+		 * Check for unit_spec_ID tag 
+		 */
+		if ((be32_to_cpu(current_quadlet) >> 24) == SBP2_UNIT_SPEC_ID_KEY) {
+			unit_spec_id_data = current_quadlet;
+			SBP2_DEBUG("sbp2: Node %x, unit spec id = %x", (LOCAL_BUS | node_id), 
+				   (unsigned int) be32_to_cpu(unit_spec_id_data));
+		}
+
+		/* 
+		 * Check for unit_sw_version tag 
+		 */
+		if ((be32_to_cpu(current_quadlet) >> 24) == SBP2_UNIT_SW_VERSION_KEY) {
+			unit_sw_ver_data = current_quadlet;
+			SBP2_DEBUG("sbp2: Node %x, unit sw version = %x", (LOCAL_BUS | node_id), 
+				   (unsigned int) be32_to_cpu(unit_sw_ver_data));
+		}
+	}
+
+	/*
+	 * Validate unit spec id and unit sw ver to see if this is an SBP-2 device
+	 */
+	if ((be32_to_cpu(unit_spec_id_data) != SBP2_UNIT_SPEC_ID_ENTRY) ||
+	    (be32_to_cpu(unit_sw_ver_data) != SBP2_SW_VERSION_ENTRY)) {
+
+		/*
+		 * Not an sbp2 device
+		 */
+		return(-ENXIO);
+	}
+
+	/*
+	 * This device is a valid SBP-2 device
+	 */
+	SBP2_INFO("sbp2: Node 0x%04x, Found SBP-2 device", (LOCAL_BUS | node_id));
+	return(0);
+}
+
+/*
+ * This function removes (cleans-up after) any unvalidated sbp2 devices
+ */
+static void sbp2_remove_unvalidated_devices(struct sbp2scsi_host_info *hi)
+{
+	int i;
+
+	/*
+	 * Loop through and free any unvalidated scsi id instance data structures
+	 */
+	for (i=0; i<SBP2SCSI_MAX_SCSI_IDS; i++) {
+		if (hi->scsi_id[i]) {
+			if (!hi->scsi_id[i]->validated) {
+
+				/*
+				 * Complete any pending commands with selection timeout
+				 */
+				sbp2scsi_complete_all_commands(hi, hi->scsi_id[i], DID_NO_CONNECT);
+       			
+				/* 
+				 * Clean up any other structures
+				 */
+				if (hi->scsi_id[i]->sbp2_total_command_orbs) {
+					sbp2util_remove_command_orb_pool(hi->scsi_id[i], hi);
+				}
+				if (hi->scsi_id[i]->login_response) {
+					pci_free_consistent(hi->host->pdev,
+							    sizeof(struct sbp2_login_response),
+							    hi->scsi_id[i]->login_response,
+							    hi->scsi_id[i]->login_response_dma);
+					SBP2_DMA_FREE("single login FIFO");
+				}
+
+				if (hi->scsi_id[i]->login_orb) {
+					pci_free_consistent(hi->host->pdev,
+							    sizeof(struct sbp2_login_orb),
+							    hi->scsi_id[i]->login_orb,
+							    hi->scsi_id[i]->login_orb_dma);
+					SBP2_DMA_FREE("single login ORB");
+				}
+
+				if (hi->scsi_id[i]->reconnect_orb) {
+					pci_free_consistent(hi->host->pdev,
+							    sizeof(struct sbp2_reconnect_orb),
+							    hi->scsi_id[i]->reconnect_orb,
+							    hi->scsi_id[i]->reconnect_orb_dma);
+					SBP2_DMA_FREE("single reconnect orb");
+				}
+
+				if (hi->scsi_id[i]->logout_orb) {
+					pci_free_consistent(hi->host->pdev,
+							    sizeof(struct sbp2_logout_orb),
+							    hi->scsi_id[i]->logout_orb,
+							    hi->scsi_id[i]->reconnect_orb_dma);
+					SBP2_DMA_FREE("single logout orb");
+				}
+
+				kfree(hi->scsi_id[i]);
+				hi->scsi_id[i] = NULL;
+				SBP2_DEBUG("sbp2: Unvalidated SBP-2 device removed, SCSI ID = %x", (unsigned int) i);
+			}
+		}
+	}
+
+	return;
+}
+
+/*
+ * This function is our reset handler. It is run out of a thread, since we get 
+ * notified of a bus reset from a bh (or interrupt).
+ */
+static void sbp2_bus_reset_handler(void *context)
+{
+	struct sbp2scsi_host_info *hi = context;
+	quadlet_t signature_data;
+	int i;
+	unsigned long flags;
+	struct scsi_id_instance_data *scsi_id;
+
+	SBP2_DEBUG("sbp2: sbp2_bus_reset_handler");
+
+	/*
+	 * TODO. Check and keep track of generation number of all requests, in case a
+	 * bus reset occurs while trying to find and login to SBP-2 devices.
+	 */
+
+	/*
+	 * First thing to do. Invalidate all SBP-2 devices. This is needed so that
+	 * we stop sending down I/O requests to the device, and also so that we can
+	 * figure out which devices have disappeared after a bus reset.
+	 */
+	for (i=0; i<SBP2SCSI_MAX_SCSI_IDS; i++) {
+		if (hi->scsi_id[i]) {
+			hi->scsi_id[i]->validated = 0;
+		}
+	}
+
+	/*
+	 * Give the sbp2 devices a little time to recover after the bus reset
+	 */
+	current->state = TASK_INTERRUPTIBLE;
+	schedule_timeout(HZ/2);		/* 1/2 second delay */
+
+	/*
+	 * Spit out what we know from the host
+	 */
+	SBP2_DEBUG("host: node_count = %x", (unsigned int) hi->host->node_count);
+	SBP2_DEBUG("host: selfid_count = %x", (unsigned int) hi->host->selfid_count);
+	SBP2_DEBUG("host: node_id = %x", (unsigned int) hi->host->node_id);
+	SBP2_DEBUG("host: irm_id = %x", (unsigned int) hi->host->irm_id);
+	SBP2_DEBUG("host: busmgr_id = %x", (unsigned int) hi->host->busmgr_id);
+	SBP2_DEBUG("host: is_root = %x", (unsigned int) hi->host->is_root);
+	SBP2_DEBUG("host: is_cycmst = %x", (unsigned int) hi->host->is_cycmst);
+	SBP2_DEBUG("host: is_irm = %x", (unsigned int) hi->host->is_irm);
+	SBP2_DEBUG("host: is_busmgr = %x", (unsigned int) hi->host->is_busmgr);
+
+	/*
+	 * Let's try and figure out which devices out there are SBP-2 devices! Loop through all 
+	 * nodes out there.
+	 */
+	for (i=0; i<hi->host->node_count; i++) {
+
+		/*
+		 * Don't read from ourselves!
+		 */
+		if (i != ((hi->host->node_id) & NODE_MASK)) {
+
+			/*
+			 * Try and send a request for a config rom signature. This is expected to fail for
+			 * some nodes, as they might be repeater phys or not be initialized.
+			 */
+			if (!sbp2util_read_quadlet(hi, LOCAL_BUS | i, CONFIG_ROM_SIGNATURE_ADDRESS, &signature_data)) {
+
+				if (be32_to_cpu(signature_data) == IEEE1394_CONFIG_ROM_SIGNATURE) {
+
+					/*
+					 * Hey, we've got a valid responding IEEE1394 node. Need to now see if it's an SBP-2 device
+					 */
+					if (!sbp2_check_device(hi, i)) {
+
+						/*
+						 * Found an SBP-2 device. Now, actually start the device.
+						 */
+						sbp2_start_device(hi, i);
+					}
+				}
+			}
+		}
+	}
+
+	/*
+	 * This code needs protection
+	 */
+	sbp2_spin_lock(&hi->sbp2_command_lock, flags);
+
+	/*
+	 * Ok, we've discovered and re-validated all SBP-2 devices out there. Let's remove structures of all
+	 * devices not re-validated (meaning they've been removed).
+	 */
+	sbp2_remove_unvalidated_devices(hi);
+
+	/*
+	 * Complete any pending commands with busy (so they get retried) and remove them from our queue
+	 */
+	for (i=0; i<SBP2SCSI_MAX_SCSI_IDS; i++) {
+		if (hi->scsi_id[i]) {
+			sbp2scsi_complete_all_commands(hi, hi->scsi_id[i], DID_BUS_BUSY);
+		}
+	}
+
+	/*
+	 * Now, note that the bus reset is complete (finally!)
+	 */
+	hi->bus_reset_in_progress = 0;
+
+	/*
+	 * Deal with the initial scsi bus scan if needed (since we only now know if there are
+	 * any sbp2 devices attached)
+	 */
+	if (!no_bus_scan && !hi->initial_scsi_bus_scan_complete && hi->bus_scan_SCpnt) {
+
+		hi->initial_scsi_bus_scan_complete = 1;			
+		scsi_id = hi->scsi_id[hi->bus_scan_SCpnt->target];
+
+		/* 
+		 * If the sbp2 device exists, then let's now execute the command.
+		 * If not, then just complete it as a selection time-out. 
+		 */
+                if (scsi_id) {
+			if (sbp2_send_command(hi, scsi_id, hi->bus_scan_SCpnt, hi->bus_scan_done)) {
+				SBP2_ERR("sbp2: Error sending SCSI command");
+				sbp2scsi_complete_command(hi, scsi_id, SBP2_SCSI_STATUS_SELECTION_TIMEOUT,
+							  hi->bus_scan_SCpnt, hi->bus_scan_done);
+			}
+		} else {
+			void (*done)(Scsi_Cmnd *) = hi->bus_scan_done;
+			hi->bus_scan_SCpnt->result = DID_NO_CONNECT << 16;
+			done (hi->bus_scan_SCpnt);
+		}
+	}
+
+	sbp2_spin_unlock(&hi->sbp2_command_lock, flags);
+
+	return;
+}
+
+
+/*
+ * This is called from the host's bh when a bus reset is complete. We wake up our detection thread
+ * to deal with the reset
+ */
+static void sbp2_host_reset(struct hpsb_host *host)
+{
+	unsigned long flags;
+	struct sbp2scsi_host_info *hi;
+	int i;
+
+	SBP2_INFO("sbp2: IEEE-1394 bus reset");
+	sbp2_spin_lock(&sbp2_host_info_lock, flags);
+	hi = sbp2_find_host_info(host);
+
+	if (hi != NULL) {
+
+		/*
+		 * Wake up our detection thread, only if it's not already handling a reset
+		 */
+		if (!hi->bus_reset_in_progress) {
+
+			/*
+			 * First thing to do. Invalidate all SBP-2 devices. This is needed so that
+			 * we stop sending down I/O requests to the device, and also so that we can
+			 * figure out which devices have disappeared after a bus reset.
+			 */
+			for (i=0; i<SBP2SCSI_MAX_SCSI_IDS; i++) {
+				if (hi->scsi_id[i]) {
+					hi->scsi_id[i]->validated = 0;
+				}
+			}
+
+			hi->bus_reset_in_progress = 1;
+
+			wake_up(&hi->sbp2_detection_wait);
+		}
+	}
+	sbp2_spin_unlock(&sbp2_host_info_lock, flags);
+	return;
+}
+
+/* XXX: How best to handle these with DMA interface? */
+
+#if 0
+/*
+ * This function deals with physical dma write requests (for adapters that do not support
+ * physical dma in hardware).
+ */
+static int sbp2_handle_physdma_write(struct hpsb_host *host, int nodeid, quadlet_t *data,
+				     u64 addr, unsigned int length)
+{
+
+	/*
+	 * Manually put the data in the right place.
+	 */
+	memcpy(bus_to_virt((u32)addr), data, length);
+	return(RCODE_COMPLETE);
+}
+
+/*
+ * This function deals with physical dma read requests (for adapters that do not support
+ * physical dma in hardware).
+ */
+static int sbp2_handle_physdma_read(struct hpsb_host *host, int nodeid, quadlet_t *data,
+				    u64 addr, unsigned int length)
+{
+
+	/*
+	 * Grab data from memory and send a read response.
+	 */
+	memcpy(data, bus_to_virt((u32)addr), length);
+	return(RCODE_COMPLETE);
+}
+#endif
+
+/**************************************
+ * SBP-2 protocol related section
+ **************************************/
+
+/*
+ * This function is called in order to login to a particular SBP-2 device, after a bus reset
+ */
+static int sbp2_login_device(struct sbp2scsi_host_info *hi, struct scsi_id_instance_data *scsi_id) 
+{
+	quadlet_t data[2];
+	unsigned long flags;
+
+	SBP2_DEBUG("sbp2: sbp2_login_device");
+
+	if (!scsi_id->login_orb) {
+		SBP2_DEBUG("sbp2: sbp2_login_device: login_orb not alloc'd!");
+		return(-EIO);
+	}
+
+	/*
+	 * Set-up login ORB
+	 */
+	scsi_id->login_orb->password_hi = 0;		/* Assume no password */
+	scsi_id->login_orb->password_lo = 0;
+	SBP2_DEBUG("sbp2: sbp2_login_device: password_hi/lo initialized");
+#ifdef SBP2_NEED_LOGIN_DESCRIPTOR_WORKAROUND
+	scsi_id->login_orb->login_response_lo = cpu_to_le32(scsi_id->login_response_dma);
+	scsi_id->login_orb->login_response_hi = cpu_to_le32(ORB_SET_NODE_ID(hi->host->node_id));
+#else
+	scsi_id->login_orb->login_response_lo = scsi_id->login_response_dma;
+	scsi_id->login_orb->login_response_hi = ORB_SET_NODE_ID(hi->host->node_id);
+#endif
+	SBP2_DEBUG("sbp2: sbp2_login_device: login_response_hi/lo initialized");
+	scsi_id->login_orb->lun_misc = ORB_SET_FUNCTION(LOGIN_REQUEST);
+	scsi_id->login_orb->lun_misc |= ORB_SET_RECONNECT(0);	/* One second reconnect time */
+	scsi_id->login_orb->lun_misc |= ORB_SET_EXCLUSIVE(1);	/* Exclusive access to device */
+	scsi_id->login_orb->lun_misc |= ORB_SET_NOTIFY(1);		/* Notify us of login complete */
+	SBP2_DEBUG("sbp2: sbp2_login_device: lun_misc initialized");
+	scsi_id->login_orb->passwd_resp_lengths = ORB_SET_LOGIN_RESP_LENGTH(sizeof(struct sbp2_login_response));
+	SBP2_DEBUG("sbp2: sbp2_login_device: passwd_resp_lengths initialized");
+#ifdef SBP2_NEED_LOGIN_DESCRIPTOR_WORKAROUND
+	scsi_id->login_orb->status_FIFO_lo = cpu_to_le32((u32)SBP2_STATUS_FIFO_ADDRESS_LO);
+	scsi_id->login_orb->status_FIFO_hi = (ORB_SET_NODE_ID(hi->host->node_id) | cpu_to_le16(SBP2_STATUS_FIFO_ADDRESS_HI));
+#else
+	scsi_id->login_orb->status_FIFO_lo = SBP2_STATUS_FIFO_ADDRESS_LO;
+	scsi_id->login_orb->status_FIFO_hi = (ORB_SET_NODE_ID(hi->host->node_id) | SBP2_STATUS_FIFO_ADDRESS_HI);
+#endif
+	SBP2_DEBUG("sbp2: sbp2_login_device: status FIFO initialized");
+
+	/*
+	 * Byte swap ORB if necessary
+	 */
+	sbp2util_cpu_to_be32_buffer(scsi_id->login_orb, sizeof(struct sbp2_login_orb));
+
+	SBP2_DEBUG("sbp2: sbp2_login_device: orb byte-swapped");
+
+	/*
+	 * Initialize login response and status fifo
+	 */
+	memset(scsi_id->login_response, 0, sizeof(struct sbp2_login_response));
+	memset(&scsi_id->status_block, 0, sizeof(struct sbp2_status_block));
+
+	SBP2_DEBUG("sbp2: sbp2_login_device: login_response/status FIFO memset");
+
+	/*
+	 * Ok, let's write to the target's management agent register
+	 */
+	data[0] = ORB_SET_NODE_ID(hi->host->node_id);
+	data[1] = scsi_id->login_orb_dma;
+	sbp2util_cpu_to_be32_buffer(data, 8);
+
+	SBP2_DEBUG("sbp2: sbp2_login_device: prepared to write");
+
+	hpsb_write(hi->host, LOCAL_BUS | scsi_id->node_id, scsi_id->sbp2_management_agent_addr, data, 8);
+
+	/*
+	 * Wait for login status... but, only if the device has not already logged-in (some devices are fast)
+	 */
+
+	SBP2_DEBUG("sbp2: sbp2_login_device: written");
+	save_flags(flags);
+	cli();
+	if (scsi_id->status_block.ORB_offset_lo != scsi_id->login_orb_dma) {
+		interruptible_sleep_on_timeout(&scsi_id->sbp2_login_wait, 10*HZ);		/* 10 second timeout */
+	}
+	restore_flags(flags);
+
+	SBP2_DEBUG("sbp2: sbp2_login_device: initial check");
+
+	/*
+	 * Match status to the login orb. If they do not match, it's probably because the login timed-out
+	 */
+	if (scsi_id->status_block.ORB_offset_lo != scsi_id->login_orb_dma) {
+		SBP2_ERR("sbp2: Error logging into SBP-2 device - login timed-out");
+		return(-EIO);
+	}
+
+	SBP2_DEBUG("sbp2: sbp2_login_device: second check");
+
+	/*
+	 * Check status
+	 */				       
+	if (STATUS_GET_RESP(scsi_id->status_block.ORB_offset_hi_misc) ||
+	    STATUS_GET_DEAD_BIT(scsi_id->status_block.ORB_offset_hi_misc) ||
+	    STATUS_GET_SBP_STATUS(scsi_id->status_block.ORB_offset_hi_misc)) {
+
+		SBP2_ERR("sbp2: Error logging into SBP-2 device - login failed");
+		return(-EIO);
+	}
+
+	/*
+	 * Byte swap the login response, for use when reconnecting or logging out.
+	 */
+	sbp2util_cpu_to_be32_buffer(scsi_id->login_response, sizeof(struct sbp2_login_response));
+
+	/*
+	 * Grab our command block agent address from the login response
+	 */
+	SBP2_DEBUG("sbp2: command_block_agent_hi = %x", (unsigned int)scsi_id->login_response->command_block_agent_hi);
+	SBP2_DEBUG("sbp2: command_block_agent_lo = %x", (unsigned int)scsi_id->login_response->command_block_agent_lo);
+
+	scsi_id->sbp2_command_block_agent_addr = ((u64)scsi_id->login_response->command_block_agent_hi) << 32;
+	scsi_id->sbp2_command_block_agent_addr |= ((u64)scsi_id->login_response->command_block_agent_lo);
+	scsi_id->sbp2_command_block_agent_addr &= 0x0000ffffffffffffULL;
+
+	SBP2_INFO("sbp2: Logged into SBP-2 device");
+
+	return(0);
+
+}
+
+/*
+ * This function is called in order to logout from a particular SBP-2 device, usually called during driver
+ * unload
+ */
+static int sbp2_logout_device(struct sbp2scsi_host_info *hi, struct scsi_id_instance_data *scsi_id) 
+{
+	quadlet_t data[2];
+
+	SBP2_DEBUG("sbp2: sbp2_logout_device");
+
+	/*
+	 * Set-up logout ORB
+	 */
+	scsi_id->logout_orb->reserved1 = 0x0;
+	scsi_id->logout_orb->reserved2 = 0x0;
+	scsi_id->logout_orb->reserved3 = 0x0;
+	scsi_id->logout_orb->reserved4 = 0x0;
+	scsi_id->logout_orb->login_ID_misc = ORB_SET_FUNCTION(LOGOUT_REQUEST);
+	scsi_id->logout_orb->login_ID_misc |= ORB_SET_LOGIN_ID(scsi_id->login_response->length_login_ID);
+	scsi_id->logout_orb->login_ID_misc |= ORB_SET_NOTIFY(1);		/* Notify us when complete */
+	scsi_id->logout_orb->reserved5 = 0x0;
+#ifdef SBP2_NEED_LOGIN_DESCRIPTOR_WORKAROUND
+	scsi_id->logout_orb->status_FIFO_lo = cpu_to_le32((u32)SBP2_STATUS_FIFO_ADDRESS_LO);
+	scsi_id->logout_orb->status_FIFO_hi = (ORB_SET_NODE_ID(hi->host->node_id) | cpu_to_le16(SBP2_STATUS_FIFO_ADDRESS_HI));
+#else
+	scsi_id->logout_orb->status_FIFO_lo = SBP2_STATUS_FIFO_ADDRESS_LO;
+	scsi_id->logout_orb->status_FIFO_hi = (ORB_SET_NODE_ID(hi->host->node_id) | SBP2_STATUS_FIFO_ADDRESS_HI);
+#endif
+
+	/*
+	 * Byte swap ORB if necessary
+	 */
+	sbp2util_cpu_to_be32_buffer(scsi_id->logout_orb, sizeof(struct sbp2_logout_orb));
+
+	/*
+	 * Ok, let's write to the target's management agent register
+	 */
+	data[0] = ORB_SET_NODE_ID(hi->host->node_id);
+	data[1] = scsi_id->logout_orb_dma;
+	sbp2util_cpu_to_be32_buffer(data, 8);
+
+	hpsb_write(hi->host, LOCAL_BUS | scsi_id->node_id, scsi_id->sbp2_management_agent_addr, data, 8);
+
+	/*
+	 * Wait for device to logout... 
+	 */
+	interruptible_sleep_on_timeout(&scsi_id->sbp2_login_wait, HZ);		/* 1 second timeout */
+
+	SBP2_INFO("sbp2: Logged out of SBP-2 device");
+
+	return(0);
+
+}
+
+/*
+ * This function is called in order to reconnect to a particular SBP-2 device, after a bus reset
+ */
+static int sbp2_reconnect_device(struct sbp2scsi_host_info *hi, struct scsi_id_instance_data *scsi_id) 
+{
+	quadlet_t data[2];
+	unsigned long flags;
+
+	SBP2_DEBUG("sbp2: sbp2_reconnect_device");
+
+	/*
+	 * Set-up reconnect ORB
+	 */
+	scsi_id->reconnect_orb->reserved1 = 0x0;
+	scsi_id->reconnect_orb->reserved2 = 0x0;
+	scsi_id->reconnect_orb->reserved3 = 0x0;
+	scsi_id->reconnect_orb->reserved4 = 0x0;
+	scsi_id->reconnect_orb->login_ID_misc = ORB_SET_FUNCTION(RECONNECT_REQUEST);
+	scsi_id->reconnect_orb->login_ID_misc |= ORB_SET_LOGIN_ID(scsi_id->login_response->length_login_ID);
+	scsi_id->reconnect_orb->login_ID_misc |= ORB_SET_NOTIFY(1);		/* Notify us when complete */
+	scsi_id->reconnect_orb->reserved5 = 0x0;
+#ifdef SBP2_NEED_LOGIN_DESCRIPTOR_WORKAROUND
+	scsi_id->reconnect_orb->status_FIFO_lo = cpu_to_le32((u32)SBP2_STATUS_FIFO_ADDRESS_LO);
+	scsi_id->reconnect_orb->status_FIFO_hi = (ORB_SET_NODE_ID(hi->host->node_id) | cpu_to_le16(SBP2_STATUS_FIFO_ADDRESS_HI));
+#else
+	scsi_id->reconnect_orb->status_FIFO_lo = SBP2_STATUS_FIFO_ADDRESS_LO;
+	scsi_id->reconnect_orb->status_FIFO_hi = (ORB_SET_NODE_ID(hi->host->node_id) | SBP2_STATUS_FIFO_ADDRESS_HI);
+#endif
+	
+	/*
+	 * Byte swap ORB if necessary
+	 */
+	sbp2util_cpu_to_be32_buffer(scsi_id->reconnect_orb, sizeof(struct sbp2_reconnect_orb));
+
+	/*
+	 * Initialize status fifo
+	 */
+	memset(&scsi_id->status_block, 0, sizeof(struct sbp2_status_block));
+
+	/*
+	 * Ok, let's write to the target's management agent register
+	 */
+	data[0] = ORB_SET_NODE_ID(hi->host->node_id);
+	data[1] = scsi_id->reconnect_orb_dma;
+	sbp2util_cpu_to_be32_buffer(data, 8);
+
+	hpsb_write(hi->host, LOCAL_BUS | scsi_id->node_id, scsi_id->sbp2_management_agent_addr, data, 8);
+
+	/*
+	 * Wait for reconnect status... but, only if the device has not already reconnected (some devices are fast)
+	 */
+	save_flags(flags);
+	cli();
+	if (scsi_id->status_block.ORB_offset_lo != scsi_id->reconnect_orb_dma) {
+		interruptible_sleep_on_timeout(&scsi_id->sbp2_login_wait, HZ);		/* one second timeout */
+	}
+	restore_flags(flags);
+
+	/*
+	 * Match status to the reconnect orb. If they do not match, it's probably because the reconnect timed-out
+	 */
+	if (scsi_id->status_block.ORB_offset_lo != scsi_id->reconnect_orb_dma) {
+		SBP2_ERR("sbp2: Error reconnecting to SBP-2 device - reconnect timed-out");
+		return(-EIO);
+	}
+
+	/*
+	 * Check status
+	 */
+	if (STATUS_GET_RESP(scsi_id->status_block.ORB_offset_hi_misc) ||
+	    STATUS_GET_DEAD_BIT(scsi_id->status_block.ORB_offset_hi_misc) ||
+	    STATUS_GET_SBP_STATUS(scsi_id->status_block.ORB_offset_hi_misc)) {
+
+		SBP2_ERR("sbp2: Error reconnecting to SBP-2 device - reconnect failed");
+		return(-EIO);
+	}
+
+	SBP2_INFO("sbp2: Reconnected to SBP-2 device");
+
+	return(0);
+
+}
+
+/*
+ * This function is called in order to set the busy timeout (number of retries to attempt) on the sbp2 device. 
+ */
+static int sbp2_set_busy_timeout(struct sbp2scsi_host_info *hi, struct scsi_id_instance_data *scsi_id)
+{      
+	quadlet_t data;
+
+	SBP2_DEBUG("sbp2: sbp2_set_busy_timeout");
+
+	/*
+	 * Ok, let's write to the target's busy timeout register
+	 */
+	data = cpu_to_be32(SBP2_BUSY_TIMEOUT_VALUE);
+
+	if (hpsb_write(hi->host, LOCAL_BUS | scsi_id->node_id, SBP2_BUSY_TIMEOUT_ADDRESS, &data, 4)) {
+		SBP2_ERR("sbp2: sbp2_set_busy_timeout error");
+	}
+
+	return(0);
+}
+
+/*
+ * This function is called to parse sbp2 device's config rom unit directory. Used to determine
+ * things like sbp2 management agent offset, and command set used (SCSI or RBC). 
+ */
+static int sbp2_parse_unit_directory(struct sbp2scsi_host_info *hi, struct scsi_id_instance_data *scsi_id)
+{
+	quadlet_t unit_directory_length, unit_directory_data;
+	u64 unit_directory_addr;
+	u32 i;
+
+	SBP2_DEBUG("sbp2: sbp2_parse_unit_directory");
+
+	if (sbp2util_unit_directory(hi, LOCAL_BUS | scsi_id->node_id, &unit_directory_addr)) {
+		SBP2_DEBUG("sbp2: Error reading unit directory address - bad status");
+		return(-EIO);   
+	}
+
+	/*
+	 * Read the size of the unit directory
+	 */
+	if (sbp2util_read_quadlet(hi, LOCAL_BUS | scsi_id->node_id, unit_directory_addr, 
+				  &unit_directory_length)) {
+		SBP2_DEBUG("sbp2: Error reading unit directory length - bad status");
+		return(-EIO);   
+	}
+
+	unit_directory_length = ((be32_to_cpu(unit_directory_length)) >> 16);
+
+	/*
+	 * Now, sweep through the unit directory looking for the management agent offset
+	 * Give up if we hit any error or somehow miss it...
+	 */
+	for (i=0; i<unit_directory_length; i++) {
+
+		if (sbp2util_read_quadlet(hi, LOCAL_BUS | scsi_id->node_id, unit_directory_addr + (i<<2) + 4, 
+					  &unit_directory_data)) {
+			SBP2_DEBUG("sbp2: Error reading unit directory - bad status");
+			return(-EIO);
+		}
+
+		/* 
+		 * Handle different fields in the unit directory, based on keys
+		 */
+		unit_directory_data = be32_to_cpu(unit_directory_data);
+		switch (unit_directory_data >> 24) {
+			
+			case SBP2_CSR_OFFSET_KEY:
+
+				/*
+				 * Save off the management agent address
+				 */
+				scsi_id->sbp2_management_agent_addr = CONFIG_ROM_INITIAL_MEMORY_SPACE + 
+								      ((unit_directory_data & 0x00ffffff) << 2);
+
+				SBP2_DEBUG("sbp2: sbp2_management_agent_addr = %x", (unsigned int) scsi_id->sbp2_management_agent_addr);
+				break;
+
+			case SBP2_COMMAND_SET_SPEC_ID_KEY:
+
+				/*
+				 * Command spec organization
+				 */
+				scsi_id->sbp2_command_set_spec_id = unit_directory_data & 0xffffff;
+				SBP2_DEBUG("sbp2: sbp2_command_set_spec_id = %x", (unsigned int) scsi_id->sbp2_command_set_spec_id);
+				break;
+
+			case SBP2_COMMAND_SET_KEY:
+
+				/*
+				 * Command set used by sbp2 device
+				 */
+				scsi_id->sbp2_command_set = unit_directory_data & 0xffffff;
+				SBP2_DEBUG("sbp2: sbp2_command_set = %x", (unsigned int) scsi_id->sbp2_command_set);
+				break;
+
+			case SBP2_UNIT_CHARACTERISTICS_KEY:
+
+				/*
+				 * Unit characterisitcs (orb related stuff that I'm not yet paying attention to)
+				 */
+				scsi_id->sbp2_unit_characteristics = unit_directory_data & 0xffffff;
+				SBP2_DEBUG("sbp2: sbp2_unit_characteristics = %x", (unsigned int) scsi_id->sbp2_unit_characteristics);
+				break;
+
+			case SBP2_DEVICE_TYPE_AND_LUN_KEY:
+
+				/*
+				 * Device type and lun (used for detemining type of sbp2 device)
+				 */
+				scsi_id->sbp2_device_type_and_lun = unit_directory_data & 0xffffff;
+				SBP2_DEBUG("sbp2: sbp2_device_type_and_lun = %x", (unsigned int) scsi_id->sbp2_device_type_and_lun);
+				break;
+
+			case SBP2_UNIT_SPEC_ID_KEY:
+
+				/*
+				 * Unit spec id (used for protocol detection)
+				 */
+				scsi_id->sbp2_unit_spec_id = unit_directory_data & 0xffffff;
+				SBP2_DEBUG("sbp2: sbp2_unit_spec_id = %x", (unsigned int) scsi_id->sbp2_unit_spec_id);
+				break;
+
+			case SBP2_UNIT_SW_VERSION_KEY:
+
+				/*
+				 * Unit sw version (used for protocol detection) 
+				 */
+				scsi_id->sbp2_unit_sw_version = unit_directory_data & 0xffffff;
+				SBP2_DEBUG("sbp2: sbp2_unit_sw_version = %x", (unsigned int) scsi_id->sbp2_unit_sw_version);
+				break;
+
+			case SBP2_FIRMWARE_REVISION_KEY:
+
+				/*
+				 * Firmware revision (used to find broken devices). If the vendor id is 0xa0b8 
+				 * (Symbios vendor id), then we have a bridge with 128KB max transfer size limitation.
+				 */
+				scsi_id->sbp2_firmware_revision = unit_directory_data & 0xffff00;
+				if (scsi_id->sbp2_firmware_revision == SBP2_128KB_BROKEN_FIRMWARE) {
+					SBP2_WARN("sbp2: warning: Bridge chipset supports 128KB max transfer size");
+				}
+				break;
+
+			default:
+				break;
+		}
+
+	}
+
+	return(0);
+}
+
+/*
+ * This function is called in order to determine the max speed and packet size we can use in our ORBs. 
+ */
+static int sbp2_max_speed_and_size(struct sbp2scsi_host_info *hi, struct scsi_id_instance_data *scsi_id)
+{
+	quadlet_t node_options, max_rec;
+	u8 speed_code;
+
+	SBP2_DEBUG("sbp2: sbp2_max_speed_and_size");
+
+	/*
+	 * Get speed code from internal host structure. There should be a better way to obtain this.
+	 */
+	speed_code = hi->host->speed_map[(hi->host->node_id & NODE_MASK) * 64 + (scsi_id->node_id & NODE_MASK)];
+
+	/*
+	 * Bump down our speed if there is a module parameter forcing us slower
+	 */
+	if (speed_code > max_speed) {
+		speed_code = max_speed;
+		SBP2_ERR("sbp2: Reducing SBP-2 max speed allowed (%x)", max_speed); 
+	}
+
+	switch (speed_code) {
+		case SPEED_S100:
+			scsi_id->speed_code = SPEED_S100;
+			scsi_id->max_payload_size = MAX_PAYLOAD_S100;
+			SBP2_INFO("sbp2: SBP-2 device max speed S100 and payload 512 bytes"); 
+			break;
+		case SPEED_S200:
+			scsi_id->speed_code = SPEED_S200;
+			scsi_id->max_payload_size = MAX_PAYLOAD_S200;
+			SBP2_INFO("sbp2: SBP-2 device max speed S200 and payload 1KB"); 
+			break;
+		case SPEED_S400:
+			scsi_id->speed_code = SPEED_S400;
+			scsi_id->max_payload_size = MAX_PAYLOAD_S400;
+			SBP2_INFO("sbp2: SBP-2 device max speed S400 and payload 2KB"); 
+			break;
+		default:
+			scsi_id->speed_code = SPEED_S100;
+			scsi_id->max_payload_size = MAX_PAYLOAD_S100;
+			SBP2_ERR("sbp2: Undefined speed: Using SBP-2 device max speed S100 and payload 512 bytes"); 
+			break;
+	}
+
+	/*
+	 * Finally, check the adapter's capabilities to further bump down our max payload size
+	 * if necessary. For instance, TILynx may not support the default max payload at a 
+	 * particular speed.
+	 */
+	if (!hpsb_read(hi->host, hi->host->node_id | LOCAL_BUS, CONFIG_ROM_NODE_OPTIONS, &node_options, 4)) {
+
+		/* 
+		 * Grab max_rec (max payload = 2 ^ (max_rec+1)) from node options. Sbp2 max payload is 
+		 * defined as 2 ^ (max_pay+2)... so, have to subtract one from max rec for comparison...
+		 * confusing, eh?   ;-)
+		 */
+		max_rec = (be32_to_cpu(node_options) & 0x0000f000) >> 12;
+		if (scsi_id->max_payload_size > (max_rec - 1)) {
+			scsi_id->max_payload_size = (max_rec - 1);
+			SBP2_ERR("sbp2: Reducing SBP-2 max payload allowed (%x)", (max_rec - 1)); 
+		}
+
+	}
+
+	return(0);
+}
+
+/*
+ * This function is called in order to perform a SBP-2 agent reset. 
+ */
+static int sbp2_agent_reset(struct sbp2scsi_host_info *hi, struct scsi_id_instance_data *scsi_id, u32 flags) 
+{
+	struct sbp2_request_packet *agent_reset_request_packet;
+
+	SBP2_DEBUG("sbp2: sbp2_agent_reset");
+
+	/*
+	 * Ok, let's write to the target's management agent register
+	 */
+	agent_reset_request_packet = sbp2util_allocate_write_request_packet(hi, LOCAL_BUS | scsi_id->node_id,
+									    scsi_id->sbp2_command_block_agent_addr + SBP2_AGENT_RESET_OFFSET,
+									    0, ntohl(SBP2_AGENT_RESET_DATA));
+
+	if (!agent_reset_request_packet) {
+		SBP2_ERR("sbp2: sbp2util_allocate_write_request_packet failed");
+		return(-EIO);
+	}
+
+	if (!hpsb_send_packet(agent_reset_request_packet->packet)) {
+		SBP2_ERR("sbp2: hpsb_send_packet failed");
+		sbp2util_free_request_packet(agent_reset_request_packet); 
+		return(-EIO);
+	}
+
+	if (!(flags & SBP2_SEND_NO_WAIT)) {
+		down(&agent_reset_request_packet->packet->state_change);
+		down(&agent_reset_request_packet->packet->state_change);
+	}
+
+	/*
+	 * Need to make sure orb pointer is written on next command
+	 */
+	scsi_id->last_orb = NULL;
+
+	return(0);
+
+}
+
+/*
+ * This function is called to create the actual command orb and s/g list out of the 
+ * scsi command itself. 
+ */
+static int sbp2_create_command_orb(struct sbp2scsi_host_info *hi, 
+				   struct scsi_id_instance_data *scsi_id,
+				   struct sbp2_command_info *command,
+				   unchar *scsi_cmd,
+				   unsigned int scsi_use_sg,
+				   unsigned int scsi_request_bufflen,
+				   void *scsi_request_buffer, int dma_dir)
+{
+	struct scatterlist *sgpnt = (struct scatterlist *) scsi_request_buffer;
+	struct sbp2_command_orb *command_orb = &command->command_orb;
+	struct sbp2_unrestricted_page_table *scatter_gather_element =
+		&command->scatter_gather_element[0];
+	u32 sg_count, sg_len;
+	dma_addr_t sg_addr;
+	int i;
+
+	/*
+	 * Set-up our command ORB..
+	 *
+	 * NOTE: We're doing unrestricted page tables (s/g), as this is best performance 
+	 * (at least with the devices I have). This means that data_size becomes the number 
+	 * of s/g elements, and page_size should be zero (for unrestricted).
+	 */
+	command_orb->next_ORB_hi = 0xffffffff;
+	command_orb->next_ORB_lo = 0xffffffff;
+	command_orb->misc = ORB_SET_MAX_PAYLOAD(scsi_id->max_payload_size);
+	command_orb->misc |= ORB_SET_SPEED(scsi_id->speed_code);
+	command_orb->misc |= ORB_SET_NOTIFY(1);		/* Notify us when complete */
+
+	/*
+	 * Set-up our pagetable stuff... unfortunately, this has become messier than I'd like. Need to 
+	 * clean this up a bit.   ;-)
+	 */
+	if (sbp2scsi_direction_table[*scsi_cmd] == ORB_DIRECTION_NO_DATA_TRANSFER) {
+
+		SBP2_DEBUG("sbp2: No data transfer");
+
+		/*
+		 * Handle no data transfer
+		 */
+		command_orb->data_descriptor_hi = 0xffffffff;
+		command_orb->data_descriptor_lo = 0xffffffff;
+		command_orb->misc |= ORB_SET_DIRECTION(1);
+
+	} else if (scsi_use_sg) {
+
+		SBP2_DEBUG("sbp2: Use scatter/gather");
+
+		/*
+		 * Special case if only one element (and less than 64KB in size)
+		 */
+		if ((scsi_use_sg == 1) && (sgpnt[0].length <= SBP2_MAX_SG_ELEMENT_LENGTH)) {
+
+			SBP2_DEBUG("sbp2: Only one s/g element");
+			command->dma_dir = dma_dir;
+			command->dma_size = sgpnt[0].length;
+			command->cmd_dma = pci_map_single (hi->host->pdev, sgpnt[0].address,
+							   command->dma_size,
+							   command->dma_dir);
+			SBP2_DMA_ALLOC("single scatter element");
+
+			command_orb->data_descriptor_hi = ORB_SET_NODE_ID(hi->host->node_id);
+			command_orb->data_descriptor_lo = command->cmd_dma;
+			command_orb->misc |= ORB_SET_DATA_SIZE(command->dma_size);
+			command_orb->misc |= ORB_SET_DIRECTION(sbp2scsi_direction_table[*scsi_cmd]);
+
+		} else {
+			int count = pci_map_sg(hi->host->pdev, sgpnt, scsi_use_sg, dma_dir);
+			SBP2_DMA_ALLOC("scatter list");
+
+			command->dma_size = scsi_use_sg;
+			command->dma_dir = dma_dir;
+			command->sge_buffer = sgpnt;
+
+			/* use page tables (s/g) */
+			command_orb->misc |= ORB_SET_PAGE_TABLE_PRESENT(0x1);
+			command_orb->misc |= ORB_SET_DIRECTION(sbp2scsi_direction_table[*scsi_cmd]);
+			command_orb->data_descriptor_hi = ORB_SET_NODE_ID(hi->host->node_id);
+			command_orb->data_descriptor_lo = command->sge_dma;
+
+			/*
+			 * Loop through and fill out our sbp-2 page tables
+			 * (and split up anything too large)
+			 */
+			for (i = 0, sg_count = 0 ; i < count; i++, sgpnt++) {
+				sg_len = sg_dma_len(sgpnt);
+				sg_addr = sg_dma_address(sgpnt);
+				while (sg_len) {
+					scatter_gather_element[sg_count].segment_base_lo = sg_addr;
+					if (sg_len > SBP2_MAX_SG_ELEMENT_LENGTH) {
+						scatter_gather_element[sg_count].length_segment_base_hi =  
+							PAGE_TABLE_SET_SEGMENT_LENGTH(SBP2_MAX_SG_ELEMENT_LENGTH);
+						sg_addr += SBP2_MAX_SG_ELEMENT_LENGTH;
+						sg_len -= SBP2_MAX_SG_ELEMENT_LENGTH;
+					} else {
+						scatter_gather_element[sg_count].length_segment_base_hi = 
+							PAGE_TABLE_SET_SEGMENT_LENGTH(sg_len);
+						sg_len = 0;
+					}
+					sg_count++;
+				}
+			}
+
+			command_orb->misc |= ORB_SET_DATA_SIZE(sg_count); /* number of page table (s/g) elements */
+
+			/*
+			 * Byte swap page tables if necessary
+			 */
+			sbp2util_cpu_to_be32_buffer(scatter_gather_element, 
+						    (sizeof(struct sbp2_unrestricted_page_table)) * sg_count);
+
+		}
+
+	} else {
+
+		SBP2_DEBUG("sbp2: No scatter/gather");
+
+		command->dma_dir = dma_dir;
+		command->dma_size = scsi_request_bufflen;
+		command->cmd_dma = pci_map_single (hi->host->pdev, scsi_request_buffer,
+						   command->dma_size,
+						   command->dma_dir);
+		SBP2_DMA_ALLOC("single bulk");
+
+		/*
+		 * Handle case where we get a command w/o s/g enabled (but check
+		 * for transfers larger than 64K)
+		 */
+		if (scsi_request_bufflen <= SBP2_MAX_SG_ELEMENT_LENGTH) {
+
+			command_orb->data_descriptor_hi = ORB_SET_NODE_ID(hi->host->node_id);
+			command_orb->data_descriptor_lo = command->cmd_dma;
+			command_orb->misc |= ORB_SET_DATA_SIZE(scsi_request_bufflen);
+			command_orb->misc |= ORB_SET_DIRECTION(sbp2scsi_direction_table[*scsi_cmd]);
+
+			/*
+			 * Sanity, in case our direction table is not up-to-date
+			 */
+			if (!scsi_request_bufflen) {
+				command_orb->data_descriptor_hi = 0xffffffff;
+				command_orb->data_descriptor_lo = 0xffffffff;
+				command_orb->misc |= ORB_SET_DIRECTION(1);
+			}
+
+		} else {
+			/*
+			 * Need to turn this into page tables, since the buffer is too large.
+			 */                     
+			command_orb->data_descriptor_hi = ORB_SET_NODE_ID(hi->host->node_id);
+			command_orb->data_descriptor_lo = command->sge_dma;
+			command_orb->misc |= ORB_SET_PAGE_TABLE_PRESENT(0x1);	/* use page tables (s/g) */
+			command_orb->misc |= ORB_SET_DIRECTION(sbp2scsi_direction_table[*scsi_cmd]);
+
+			/*
+			 * fill out our sbp-2 page tables (and split up the large buffer)
+			 */
+			sg_count = 0;
+			sg_len = scsi_request_bufflen;
+			sg_addr = command->cmd_dma;
+			while (sg_len) {
+				scatter_gather_element[sg_count].segment_base_lo = sg_addr;
+				if (sg_len > SBP2_MAX_SG_ELEMENT_LENGTH) {
+					scatter_gather_element[sg_count].length_segment_base_hi = 
+						PAGE_TABLE_SET_SEGMENT_LENGTH(SBP2_MAX_SG_ELEMENT_LENGTH);
+					sg_addr += SBP2_MAX_SG_ELEMENT_LENGTH;
+					sg_len -= SBP2_MAX_SG_ELEMENT_LENGTH;
+				} else {
+					scatter_gather_element[sg_count].length_segment_base_hi = 
+						PAGE_TABLE_SET_SEGMENT_LENGTH(sg_len);
+					sg_len = 0;
+				}
+				sg_count++;
+			}
+
+			command_orb->misc |= ORB_SET_DATA_SIZE(sg_count); /* number of page table (s/g) elements */
+
+			/*
+			 * Byte swap page tables if necessary
+			 */
+			sbp2util_cpu_to_be32_buffer(scatter_gather_element, 
+						    (sizeof(struct sbp2_unrestricted_page_table)) *
+						     sg_count);
+
+		}
+
+	}
+
+	/*
+	 * Byte swap command ORB if necessary
+	 */
+	sbp2util_cpu_to_be32_buffer(command_orb, sizeof(struct sbp2_command_orb));
+
+	/*
+	 * Put our scsi command in the command ORB
+	 */
+	memset(command_orb->cdb, 0, 12);
+	memcpy(command_orb->cdb, scsi_cmd, COMMAND_SIZE(*scsi_cmd));
+
+	return(0);
+}
+ 
+/*
+ * This function is called in order to begin a regular SBP-2 command. 
+ */
+static int sbp2_link_orb_command(struct sbp2scsi_host_info *hi, struct scsi_id_instance_data *scsi_id,
+				 struct sbp2_command_info *command)
+{
+        struct sbp2_request_packet *command_request_packet;
+	struct sbp2_command_orb *command_orb = &command->command_orb;
+
+	outstanding_orb_incr;
+	SBP2_ORB_DEBUG("sending command orb %p, linked = %x, total orbs = %x",
+			command_orb, command->linked, global_outstanding_command_orbs);
+
+	/*
+	 * Check to see if there are any previous orbs to use
+	 */
+	if (scsi_id->last_orb == NULL) {
+	
+		/*
+		 * Ok, let's write to the target's management agent register
+		 */
+		if (!hi->bus_reset_in_progress) {
+
+			command_request_packet = sbp2util_allocate_write_request_packet(hi, LOCAL_BUS | scsi_id->node_id,
+											scsi_id->sbp2_command_block_agent_addr + SBP2_ORB_POINTER_OFFSET,
+											8, 0);
+		
+			if (!command_request_packet) {
+				SBP2_ERR("sbp2: sbp2util_allocate_write_request_packet failed");
+				return(-EIO);
+			}
+		
+			command_request_packet->packet->data[0] = ORB_SET_NODE_ID(hi->host->node_id);
+			command_request_packet->packet->data[1] = command->command_orb_dma;
+			sbp2util_cpu_to_be32_buffer(command_request_packet->packet->data, 8);
+		
+			SBP2_ORB_DEBUG("write command agent, command orb %p", command_orb);
+
+			if (!hpsb_send_packet(command_request_packet->packet)) {
+				SBP2_ERR("sbp2: hpsb_send_packet failed");
+				sbp2util_free_request_packet(command_request_packet); 
+				return(-EIO);
+			}
+
+			SBP2_ORB_DEBUG("write command agent complete");
+		}
+
+		scsi_id->last_orb = command_orb;
+
+	} else {
+
+		/*
+		 * We have an orb already sent (maybe or maybe not
+		 * processed) that we can append this orb to. So do so,
+		 * and ring the doorbell. Have to be very careful
+		 * modifying these next orb pointers, as they are accessed
+		 * both by the sbp2 device and us.
+		 */
+		scsi_id->last_orb->next_ORB_lo = cpu_to_be32(command->command_orb_dma);
+		scsi_id->last_orb->next_ORB_hi = 0x0;	/* Tells hardware that this pointer is valid */
+		
+		/*
+		 * Only ring the doorbell if we need to (first parts of linked orbs don't need this)
+		 */
+		if (!command->linked && !hi->bus_reset_in_progress) {
+
+			command_request_packet = sbp2util_allocate_write_request_packet(hi,
+				LOCAL_BUS | scsi_id->node_id,
+				scsi_id->sbp2_command_block_agent_addr + SBP2_DOORBELL_OFFSET,
+				0, cpu_to_be32(command->command_orb_dma));
+	
+			if (!command_request_packet) {
+				SBP2_ERR("sbp2: sbp2util_allocate_write_request_packet failed");
+				return(-EIO);
+			}
+
+			SBP2_ORB_DEBUG("ring doorbell, command orb %p", command_orb);
+
+			if (!hpsb_send_packet(command_request_packet->packet)) {
+				SBP2_ERR("sbp2: hpsb_send_packet failed");
+				sbp2util_free_request_packet(command_request_packet);
+				return(-EIO);
+			}
+		}
+
+		scsi_id->last_orb = command_orb;
+
+	}
+       	return(0);
+}
+
+/*
+ * This function is called in order to begin a regular SBP-2 command. 
+ */
+static int sbp2_send_command(struct sbp2scsi_host_info *hi, struct scsi_id_instance_data *scsi_id,
+			     Scsi_Cmnd *SCpnt, void (*done)(Scsi_Cmnd *))
+{
+	unchar *cmd = (unchar *) SCpnt->cmnd;
+	u32 device_type = (scsi_id->sbp2_device_type_and_lun & 0x00ff0000) >> 16;
+	struct sbp2_command_info *command;
+
+	SBP2_DEBUG("sbp2: sbp2_send_command");
+	SBP2_DEBUG("sbp2: SCSI command = %02x", *cmd);
+	SBP2_DEBUG("sbp2: SCSI transfer size = %x", SCpnt->request_bufflen);
+	SBP2_DEBUG("sbp2: SCSI s/g elements = %x", (unsigned int)SCpnt->use_sg);
+
+	/*
+	 * Check for broken devices that can't handle greater than 128K transfers, and deal with them in a 
+	 * hacked ugly way.
+	 */
+	if ((scsi_id->sbp2_firmware_revision == SBP2_128KB_BROKEN_FIRMWARE) && 
+	    (SCpnt->request_bufflen > SBP2_BROKEN_FIRMWARE_MAX_TRANSFER) && 
+	    (device_type == TYPE_DISK) &&
+	    (SCpnt->use_sg) &&
+	    (*cmd == 0x28 || *cmd == 0x2a || *cmd == 0x0a || *cmd == 0x08)) {
+
+		/*
+		 * Darn, a broken device. We'll need to split up the transfer ourselves
+		 */
+		sbp2_send_split_command(hi, scsi_id, SCpnt, done);
+		return(0);
+	}
+
+	/*
+	 * Allocate a command orb and s/g structure
+	 */
+	command = sbp2util_allocate_command_orb(scsi_id, SCpnt, done, hi);
+	if (!command) {
+		return(-EIO);
+	}
+
+	/*
+	 * Now actually fill in the comamnd orb and sbp2 s/g list
+	 */
+	sbp2_create_command_orb(hi, scsi_id, command, cmd, SCpnt->use_sg, 
+				SCpnt->request_bufflen, SCpnt->request_buffer,
+				scsi_to_pci_dma_dir(SCpnt->sc_data_direction));
+	
+	/*
+	 * Update our cdb if necessary (to handle sbp2 RBC command set differences).
+	 * This is where the command set hacks go!   =)
+	 */
+	if ((device_type == TYPE_DISK) ||
+	    (device_type == TYPE_SDAD) ||
+	    (device_type == TYPE_ROM)) {
+		sbp2_check_sbp2_command(command->command_orb.cdb);
+	}
+
+	/*
+	 * Initialize status fifo
+	 */
+	memset(&scsi_id->status_block, 0, sizeof(struct sbp2_status_block));
+
+	/*
+	 * Link up the orb, and ring the doorbell if needed
+	 */
+	sbp2_link_orb_command(hi, scsi_id, command);
+	
+	return(0);
+}
+
+/*
+ * This function is called for broken sbp2 device, where we have to break up large transfers. 
+ */
+static int sbp2_send_split_command(struct sbp2scsi_host_info *hi, struct scsi_id_instance_data *scsi_id,
+				   Scsi_Cmnd *SCpnt, void (*done)(Scsi_Cmnd *))
+{
+	unchar *cmd = (unchar *) SCpnt->cmnd;
+	struct scatterlist *sgpnt = (struct scatterlist *) SCpnt->request_buffer;
+	struct sbp2_command_info *command;
+	unsigned int i, block_count, block_address, block_size;
+	unsigned int current_sg = 0;
+	unsigned int total_transfer = 0;
+	unsigned int total_sg = 0;
+	unchar new_cmd[12];
+
+	memset(new_cmd, 0, 12);
+	memcpy(new_cmd, cmd, COMMAND_SIZE(*cmd));
+
+	/*
+	 * Turns command into 10 byte version
+	 */
+	sbp2_check_sbp2_command(new_cmd);
+        
+	/*
+	 * Pull block size, block address, block count from command sent down
+	 */
+	block_count = (cmd[7] << 8) | cmd[8];
+	block_address = (cmd[2] << 24) | (cmd[3] << 16) | (cmd[4] << 8) | cmd[5]; 
+	block_size = SCpnt->request_bufflen/block_count;
+
+	/*
+	 * Walk the scsi s/g list to determine how much we can transfer in one pop
+	 */
+	for (i=0; i<SCpnt->use_sg; i++) {
+
+		total_transfer+=sgpnt[i].length;
+		total_sg++;
+
+		if (total_transfer > SBP2_BROKEN_FIRMWARE_MAX_TRANSFER) {
+
+			/*
+			 * Back everything up one, so that we're less than 128KB
+			 */
+			total_transfer-=sgpnt[i].length;
+			total_sg--;
+			i--;
+
+			command = sbp2util_allocate_command_orb(scsi_id, SCpnt, done, hi);
+			if (!command) {
+				return(-EIO);
+			}
+
+			/*
+			 * This is not the final piece, so mark it as linked
+			 */
+			command->linked = 1;
+
+			block_count = total_transfer/block_size;
+			new_cmd[2] = (unchar) (block_address >> 24) & 0xff;
+			new_cmd[3] = (unchar) (block_address >> 16) & 0xff;
+			new_cmd[4] = (unchar) (block_address >> 8) & 0xff;
+			new_cmd[5] = (unchar) block_address & 0xff;
+			new_cmd[7] = (unchar) (block_count >> 8) & 0xff;
+			new_cmd[8] = (unchar) block_count & 0xff;
+			block_address+=block_count;
+
+			sbp2_create_command_orb(hi, scsi_id, command, new_cmd, total_sg, 
+						total_transfer, &sgpnt[current_sg],
+						scsi_to_pci_dma_dir(SCpnt->sc_data_direction));
+
+			/*
+			 * Link up the orb, and ring the doorbell if needed
+			 */
+			memset(&scsi_id->status_block, 0, sizeof(struct sbp2_status_block));
+			sbp2_link_orb_command(hi, scsi_id, command);
+
+			current_sg += total_sg;
+			total_sg = 0;
+			total_transfer = 0;
+
+		}
+
+	}
+
+	/*
+	 * Get the last piece...
+	 */
+	command = sbp2util_allocate_command_orb(scsi_id, SCpnt, done, hi);
+	if (!command) {
+		return(-EIO);
+	}
+
+	block_count = total_transfer/block_size;
+	new_cmd[2] = (unchar) (block_address >> 24) & 0xff;
+	new_cmd[3] = (unchar) (block_address >> 16) & 0xff;
+	new_cmd[4] = (unchar) (block_address >> 8) & 0xff;
+	new_cmd[5] = (unchar) block_address & 0xff;
+	new_cmd[7] = (unchar) (block_count >> 8) & 0xff;
+	new_cmd[8] = (unchar) block_count & 0xff;
+
+	sbp2_create_command_orb(hi, scsi_id, command, new_cmd, total_sg, 
+				total_transfer, &sgpnt[current_sg],
+				scsi_to_pci_dma_dir(SCpnt->sc_data_direction));
+
+	/*
+	 * Link up the orb, and ring the doorbell if needed
+	 */
+	memset(&scsi_id->status_block, 0, sizeof(struct sbp2_status_block));
+	sbp2_link_orb_command(hi, scsi_id, command);
+
+
+	return(0);
+}
+
+/*
+ * This function deals with command set differences between Linux scsi command set and sbp2 RBC
+ * command set.
+ */
+static void sbp2_check_sbp2_command(unchar *cmd)
+{
+	unchar new_cmd[16];
+
+	SBP2_DEBUG("sbp2: sbp2_check_sbp2_command");
+
+	switch (*cmd) {
+		
+		case READ_6:
+
+			SBP2_DEBUG("sbp2: Convert READ_6 to READ_10");
+
+			/*
+			 * Need to turn read_6 into read_10
+			 */
+			new_cmd[0] = 0x28;
+			new_cmd[1] = (cmd[1] & 0xe0);
+			new_cmd[2] = 0x0;
+			new_cmd[3] = (cmd[1] & 0x1f);
+			new_cmd[4] = cmd[2];
+			new_cmd[5] = cmd[3];
+			new_cmd[6] = 0x0;
+			new_cmd[7] = 0x0;
+			new_cmd[8] = cmd[4];
+			new_cmd[9] = cmd[5];
+
+			memcpy(cmd, new_cmd, 10);
+
+			break;
+
+		case WRITE_6:
+
+			SBP2_DEBUG("sbp2: Convert WRITE_6 to WRITE_10");
+
+			/*
+			 * Need to turn write_6 into write_10
+			 */
+			new_cmd[0] = 0x2a;
+			new_cmd[1] = (cmd[1] & 0xe0);
+			new_cmd[2] = 0x0;
+			new_cmd[3] = (cmd[1] & 0x1f);
+			new_cmd[4] = cmd[2];
+			new_cmd[5] = cmd[3];
+			new_cmd[6] = 0x0;
+			new_cmd[7] = 0x0;
+			new_cmd[8] = cmd[4];
+			new_cmd[9] = cmd[5];
+
+			memcpy(cmd, new_cmd, 10);
+
+			break;
+
+		case MODE_SENSE:
+
+			SBP2_DEBUG("sbp2: Convert MODE_SENSE_6 to MOSE_SENSE_10");
+
+			/*
+			 * Need to turn mode_sense_6 into mode_sense_10
+			 */
+			new_cmd[0] = 0x5a;
+			new_cmd[1] = cmd[1];
+			new_cmd[2] = cmd[2];
+			new_cmd[3] = 0x0;
+			new_cmd[4] = 0x0;
+			new_cmd[5] = 0x0;
+			new_cmd[6] = 0x0;
+			new_cmd[7] = 0x0;
+			new_cmd[8] = cmd[4];
+			new_cmd[9] = cmd[5];
+
+			memcpy(cmd, new_cmd, 10);
+
+			break;
+
+		case MODE_SELECT:
+
+			/*
+			 * TODO. Probably need to change mode select to 10 byte version
+			 */
+
+		default:
+			break;
+	}
+
+	return;
+}
+
+/*
+ * Translates SBP-2 status into SCSI sense data for check conditions
+ */
+static unsigned int sbp2_status_to_sense_data(unchar *sbp2_status, unchar *sense_data)
+{
+	SBP2_DEBUG("sbp2: sbp2_status_to_sense_data");
+
+	/*
+	 * Ok, it's pretty ugly...   ;-)
+	 */
+	sense_data[0] = 0x70;
+	sense_data[1] = 0x0;
+	sense_data[2] = sbp2_status[9];
+	sense_data[3] = sbp2_status[12];
+	sense_data[4] = sbp2_status[13];
+	sense_data[5] = sbp2_status[14];
+	sense_data[6] = sbp2_status[15];
+	sense_data[7] = 10;
+	sense_data[8] = sbp2_status[16];
+	sense_data[9] = sbp2_status[17];
+	sense_data[10] = sbp2_status[18];
+	sense_data[11] = sbp2_status[19];
+	sense_data[12] = sbp2_status[10];
+	sense_data[13] = sbp2_status[11];
+	sense_data[14] = sbp2_status[20];
+	sense_data[15] = sbp2_status[21];
+
+	return(sbp2_status[8] & 0x3f);	/* return scsi status */
+}
+
+/*
+ * This function is called after a command is completed, in order to do any necessary SBP-2
+ * response data translations for the SCSI stack
+ */
+static void sbp2_check_sbp2_response(struct sbp2scsi_host_info *hi,
+				     struct scsi_id_instance_data *scsi_id, 
+				     Scsi_Cmnd *SCpnt)
+{
+	u8 *scsi_buf = SCpnt->request_buffer;
+	u32 device_type = (scsi_id->sbp2_device_type_and_lun & 0x00ff0000) >> 16;
+        
+	SBP2_DEBUG("sbp2: sbp2_check_sbp2_response");
+
+	switch (SCpnt->cmnd[0]) {
+		
+		case INQUIRY:
+
+			SBP2_DEBUG("sbp2: Check Inquiry data");
+
+			/*
+			 * Check for Simple Direct Access Device and change it to TYPE_DISK
+			 */
+			if ((scsi_buf[0] & 0x1f) == TYPE_SDAD) {
+				SBP2_DEBUG("sbp2: Changing TYPE_SDAD to TYPE_DISK");
+				scsi_buf[0] &= 0xe0;
+			}
+
+			/*
+			 * Fix ansi revision and response data format
+			 */
+			scsi_buf[2] |= 2;
+			scsi_buf[3] = (scsi_buf[3] & 0xf0) | 2;
+
+			break;
+
+		case MODE_SENSE:
+
+			if ((device_type == TYPE_DISK) ||
+			    (device_type == TYPE_SDAD) ||
+			    (device_type == TYPE_ROM)) {
+
+				SBP2_DEBUG("sbp2: Modify mode sense response (10 byte version)");
+	
+				scsi_buf[0] = scsi_buf[1];	/* Mode data length */
+				scsi_buf[1] = scsi_buf[2];	/* Medium type */
+				scsi_buf[2] = scsi_buf[3];	/* Device specific parameter */
+				scsi_buf[3] = scsi_buf[7];	/* Block descriptor length */
+				memcpy(scsi_buf + 4, scsi_buf + 8, scsi_buf[0]);
+
+			}
+
+			break;
+
+		case MODE_SELECT:
+
+			/*
+			 * TODO. Probably need to change mode select to 10 byte version
+			 */
+
+		default:
+			break;
+	}
+	return;
+}
+
+/*
+ * This function deals with status writes from the SBP-2 device
+ */
+static int sbp2_handle_status_write(struct hpsb_host *host, int nodeid, int destid,
+				    quadlet_t *data, u64 addr, unsigned int length)
+{
+	struct sbp2scsi_host_info *hi = NULL;
+	struct scsi_id_instance_data *scsi_id = NULL;
+	int i;
+	unsigned long flags;
+	Scsi_Cmnd *SCpnt = NULL;
+	u32 scsi_status = SBP2_SCSI_STATUS_GOOD;
+	struct sbp2_command_info *command;
+
+	SBP2_DEBUG("sbp2: sbp2_handle_status_write");
+
+	if (!host) {
+		SBP2_ERR("sbp2: host is NULL - this is bad!");
+		return(RCODE_ADDRESS_ERROR);
+	}
+
+	sbp2_spin_lock(&sbp2_host_info_lock, flags);
+	hi = sbp2_find_host_info(host);
+	sbp2_spin_unlock(&sbp2_host_info_lock, flags);
+
+	if (!hi) {
+		SBP2_ERR("sbp2: host info is NULL - this is bad!");
+		return(RCODE_ADDRESS_ERROR);
+	}
+
+	sbp2_spin_lock(&hi->sbp2_command_lock, flags);
+
+	/*
+	 * Find our scsi_id structure
+	 */
+	for (i=0; i<SBP2SCSI_MAX_SCSI_IDS; i++) {
+		if (hi->scsi_id[i]) {
+			if (hi->scsi_id[i]->node_id == (nodeid & NODE_MASK)) {
+				scsi_id = hi->scsi_id[i];
+				SBP2_DEBUG("sbp2: SBP-2 status write from node %x", scsi_id->node_id);
+				break;
+			}
+		}
+	}
+
+	if (!scsi_id) {
+		SBP2_ERR("sbp2: scsi_id is NULL - device is gone?");
+		sbp2_spin_unlock(&hi->sbp2_command_lock, flags);
+		return(RCODE_ADDRESS_ERROR);
+	}
+
+	/*
+	 * Put response into scsi_id status fifo... 
+	 */
+	memcpy(&scsi_id->status_block, data, length);
+
+	/*
+	 * Byte swap first two quadlets (8 bytes) of status for processing
+	 */
+	sbp2util_be32_to_cpu_buffer(&scsi_id->status_block, 8);
+
+	/*
+	 * Handle command ORB status here if necessary. First, need to match status with command.
+	 */
+	command = sbp2util_find_command_for_orb(scsi_id, scsi_id->status_block.ORB_offset_lo);
+	if (command) {
+
+		SBP2_DEBUG("sbp2: Found status for command ORB");
+
+		SBP2_ORB_DEBUG("matched command orb %p", &command->command_orb);
+		outstanding_orb_decr;
+
+		/*
+		 * Matched status with command, now grab scsi command pointers and check status
+		 */
+		SCpnt = command->Current_SCpnt;
+		sbp2util_mark_command_completed(scsi_id, command);
+
+		if (SCpnt && !command->linked) {
+
+			/*
+			 * Handle check conditions
+			 */
+			if (STATUS_GET_SBP_STATUS(scsi_id->status_block.ORB_offset_hi_misc)) {
+
+				SBP2_DEBUG("sbp2: CHECK CONDITION");
+
+				/*
+				 * Translate SBP-2 status to SCSI sense data
+				 */
+				scsi_status = sbp2_status_to_sense_data((unchar *)&scsi_id->status_block, SCpnt->sense_buffer);
+
+				/*
+				 * Initiate a fetch agent reset. 
+				 */
+				sbp2_agent_reset(hi, scsi_id, SBP2_SEND_NO_WAIT);
+
+			}
+
+			SBP2_ORB_DEBUG("completing command orb %p", &command->command_orb);
+
+			/*
+			 * Complete the SCSI command
+			 */
+			SBP2_DEBUG("sbp2: Completing SCSI command");
+			sbp2scsi_complete_command(hi, scsi_id, scsi_status, SCpnt, command->Current_done);
+			SBP2_ORB_DEBUG("command orb completed");
+		}
+
+		/*
+		 * Check here to see if there are no commands in-use. If there are none, we can
+		 * null out last orb so that next time around we write directly to the orb pointer... 
+		 * Quick start saves one 1394 bus transaction.
+		 */
+		if (list_empty(&scsi_id->sbp2_command_orb_inuse)) {
+			scsi_id->last_orb = NULL;
+		}
+
+	}
+
+	sbp2_spin_unlock(&hi->sbp2_command_lock, flags);
+	wake_up(&scsi_id->sbp2_login_wait);
+	return(RCODE_COMPLETE);
+}
+
+
+/**************************************
+ * SCSI interface related section
+ **************************************/
+
+/*
+ * This routine is the main request entry routine for doing I/O. It is 
+ * called from the scsi stack directly.
+ */
+static int sbp2scsi_queuecommand (Scsi_Cmnd *SCpnt, void (*done)(Scsi_Cmnd *)) 
+{
+	struct sbp2scsi_host_info *hi = NULL;
+	struct scsi_id_instance_data *scsi_id = NULL;
+	unsigned long flags;
+
+	SBP2_DEBUG("sbp2: sbp2scsi_queuecommand");
+
+	/*
+	 * Pull our host info and scsi id instance data from the scsi command
+	 */
+	hi = (struct sbp2scsi_host_info *) SCpnt->host->hostdata[0];
+
+	if (!hi) {
+		SBP2_ERR("sbp2: sbp2scsi_host_info is NULL - this is bad!");
+		SCpnt->result = DID_NO_CONNECT << 16;
+		done (SCpnt);
+		return(0);
+	}
+
+	scsi_id = hi->scsi_id[SCpnt->target];
+
+	/*
+	 * Save off the command if this is the initial bus scan... so that we can
+	 * complete it after we find all our sbp2 devices on the 1394 bus
+	 */
+	if (!no_bus_scan && !hi->initial_scsi_bus_scan_complete) {
+		hi->bus_scan_SCpnt = SCpnt;
+		hi->bus_scan_done = done;
+		return(0);
+	}
+
+	/*
+	 * If scsi_id is null, it means there is no device in this slot, so we should return 
+	 * selection timeout.
+	 */
+	if (!scsi_id) {
+		SCpnt->result = DID_NO_CONNECT << 16;
+		done (SCpnt);
+		return(0);
+	}
+
+	/*
+	 * Until we handle multiple luns, just return selection time-out to any IO directed at non-zero LUNs
+	 */
+	if (SCpnt->lun) {
+		SCpnt->result = DID_NO_CONNECT << 16;
+		done (SCpnt);
+		return(0);
+	}
+
+	/*
+	 * Check for request sense command, and handle it here (autorequest sense)
+	 */
+	if (SCpnt->cmnd[0] == REQUEST_SENSE) {
+		SBP2_DEBUG("sbp2: REQUEST_SENSE");
+		memcpy(SCpnt->request_buffer, SCpnt->sense_buffer, SCpnt->request_bufflen);
+		memset(SCpnt->sense_buffer, 0, sizeof(SCpnt->sense_buffer));
+		sbp2scsi_complete_command(hi, scsi_id, SBP2_SCSI_STATUS_GOOD, SCpnt, done);
+		return(0);
+	}
+
+	/*
+	 * Check to see if there is a command in progress and just return busy (to be queued later)
+	 */
+	if (hi->bus_reset_in_progress) {
+		SBP2_ERR("sbp2: Bus reset in progress - rejecting command");
+		SCpnt->result = DID_BUS_BUSY << 16;
+		done (SCpnt);
+		return(0);
+	}
+
+	/*
+	 * Try and send our SCSI command
+	 */
+	sbp2_spin_lock(&hi->sbp2_command_lock, flags);
+	if (sbp2_send_command(hi, scsi_id, SCpnt, done)) {
+		SBP2_ERR("sbp2: Error sending SCSI command");
+		sbp2scsi_complete_command(hi, scsi_id, SBP2_SCSI_STATUS_SELECTION_TIMEOUT, SCpnt, done);
+	}
+	sbp2_spin_unlock(&hi->sbp2_command_lock, flags);
+
+	return(0);
+}
+
+/*
+ * This function is called in order to complete all outstanding SBP-2 commands (in case of resets, etc.). 
+ */
+static void sbp2scsi_complete_all_commands(struct sbp2scsi_host_info *hi, struct scsi_id_instance_data *scsi_id, 
+					   u32 status)
+{
+	struct list_head *lh;
+	struct sbp2_command_info *command;
+
+	SBP2_DEBUG("sbp2: sbp2_complete_all_commands");
+
+	while (!list_empty(&scsi_id->sbp2_command_orb_inuse)) {
+		SBP2_DEBUG("sbp2: Found pending command to complete");
+		lh = scsi_id->sbp2_command_orb_inuse.next;
+		command = list_entry(lh, struct sbp2_command_info, list);
+		sbp2util_mark_command_completed(scsi_id, command);
+		if (command->Current_SCpnt && !command->linked) {
+			void (*done)(Scsi_Cmnd *) = command->Current_done;
+			command->Current_SCpnt->result = status << 16;
+			done (command->Current_SCpnt);
+		}
+	}
+
+	return;
+}
+
+/*
+ * This function is called in order to complete a regular SBP-2 command. 
+ */
+static void sbp2scsi_complete_command(struct sbp2scsi_host_info *hi, struct scsi_id_instance_data *scsi_id, u32 scsi_status,
+				      Scsi_Cmnd *SCpnt, void (*done)(Scsi_Cmnd *))
+{
+	SBP2_DEBUG("sbp2: sbp2scsi_complete_command");
+
+	/*
+	 * Sanity
+	 */
+	if (!SCpnt) {
+		SBP2_ERR("sbp2: SCpnt is NULL");
+		return;
+	}
+
+	/*
+	 * If a bus reset is in progress and there was an error, don't complete the command,
+	 * just let it get retried at the end of the bus reset.
+	 */
+	if ((hi->bus_reset_in_progress) && (scsi_status != SBP2_SCSI_STATUS_GOOD)) {
+		SBP2_ERR("sbp2: Bus reset in progress - retry command later");
+		return;
+	}
+        
+	/*
+	 * Switch on scsi status
+	 */
+	switch (scsi_status) {
+		case SBP2_SCSI_STATUS_GOOD:
+			SCpnt->result = DID_OK;
+			break;
+
+		case SBP2_SCSI_STATUS_BUSY:
+			SBP2_ERR("sbp2: SBP2_SCSI_STATUS_BUSY");
+			SCpnt->result = DID_BUS_BUSY << 16;
+			break;
+
+		case SBP2_SCSI_STATUS_CHECK_CONDITION:
+			SBP2_DEBUG("sbp2: SBP2_SCSI_STATUS_CHECK_CONDITION");
+			SCpnt->result = CHECK_CONDITION << 1;
+
+			/*
+			 * Debug stuff
+			 */
+			print_sense("bh", SCpnt);
+
+			break;
+
+		case SBP2_SCSI_STATUS_SELECTION_TIMEOUT:
+			SBP2_ERR("sbp2: SBP2_SCSI_STATUS_SELECTION_TIMEOUT");
+			SCpnt->result = DID_NO_CONNECT << 16;
+			break;
+
+		case SBP2_SCSI_STATUS_CONDITION_MET:
+		case SBP2_SCSI_STATUS_RESERVATION_CONFLICT:
+		case SBP2_SCSI_STATUS_COMMAND_TERMINATED:
+			SBP2_ERR("sbp2: Bad SCSI status = %x", scsi_status);
+			SCpnt->result = DID_ERROR << 16;
+			break;
+
+		default:
+			SBP2_ERR("sbp2: Unsupported SCSI status = %x", scsi_status);
+			SCpnt->result = DID_ERROR << 16;
+	}
+
+	/*
+	 * Take care of any sbp2 response data mucking here (RBC stuff, etc.)
+	 */
+	if (SCpnt->result == DID_OK) {
+		sbp2_check_sbp2_response(hi, scsi_id, SCpnt);
+	}
+
+	/*
+	 * One more quick hack (not enabled by default). Some sbp2 devices do not support 
+	 * mode sense. Turn-on this hack to allow the device to pass the sd driver's 
+	 * write-protect test (so that you can mount the device rw).
+	 */
+	if (mode_sense_hack && SCpnt->result != DID_OK && SCpnt->cmnd[0] == MODE_SENSE) {
+		SBP2_INFO("sbp2: Returning success to mode sense command");
+		SCpnt->result = DID_OK;
+		SCpnt->sense_buffer[0] = 0;
+		memset (SCpnt->request_buffer, 0, 8);
+	}
+
+	/*
+	 * If a bus reset is in progress and there was an error, complete the command
+	 * as busy so that it will get retried.
+	 */
+	if ((hi->bus_reset_in_progress) && (scsi_status != SBP2_SCSI_STATUS_GOOD)) {
+		SBP2_ERR("sbp2: Completing command with busy (bus reset)");
+		SCpnt->result = DID_BUS_BUSY << 16;
+	}
+
+	/*
+	 * If a unit attention occurs, return busy status so it gets retried... it could have happened because
+	 * of a 1394 bus reset or hot-plug...
+	 */
+	if ((scsi_status == SBP2_SCSI_STATUS_CHECK_CONDITION) && (SCpnt->sense_buffer[2] == UNIT_ATTENTION)) {
+		SBP2_INFO("sbp2: UNIT ATTENTION - return busy");
+		SCpnt->result = DID_BUS_BUSY << 16;
+	}
+
+	/*
+	 * Tell scsi stack that we're done with this command
+	 */
+	done (SCpnt);
+
+	return;
+}
+
+/*
+ * Called by scsi stack when something has really gone wrong.
+ * Usually called when a command has timed-out for some reason. 
+ */
+static int sbp2scsi_abort (Scsi_Cmnd *SCpnt) 
+{
+	struct sbp2scsi_host_info *hi = (struct sbp2scsi_host_info *) SCpnt->host->hostdata[0];
+	struct scsi_id_instance_data *scsi_id = hi->scsi_id[SCpnt->target];
+	struct sbp2_command_info *command;
+	unsigned long flags;
+
+	SBP2_ERR("sbp2: aborting sbp2 command");
+
+	if (scsi_id) {
+
+		/*
+		 * Right now, just return any matching command structures to the free pool (there may
+		 * be more than one because of broken up/linked commands).
+		 */
+		sbp2_spin_lock(&hi->sbp2_command_lock, flags);
+		do {
+			command = sbp2util_find_command_for_SCpnt(scsi_id, SCpnt);
+			if (command) {
+				SBP2_DEBUG("sbp2: Found command to abort");
+				sbp2util_mark_command_completed(scsi_id, command);
+				if (command->Current_SCpnt && !command->linked) {
+					void (*done)(Scsi_Cmnd *) = command->Current_done;
+					command->Current_SCpnt->result = DID_ABORT << 16;
+					done (command->Current_SCpnt);
+				}
+			}
+		} while (command);
+
+		/*
+		 * Initiate a fetch agent reset. 
+		 */
+		sbp2_agent_reset(hi, scsi_id, SBP2_SEND_NO_WAIT);
+		sbp2scsi_complete_all_commands(hi, scsi_id, DID_BUS_BUSY);		
+		sbp2_spin_unlock(&hi->sbp2_command_lock, flags);
+	}
+
+	return(SCSI_ABORT_SUCCESS);
+}
+
+/*
+ * Called by scsi stack when something has really gone wrong.
+ */
+static int sbp2scsi_reset (Scsi_Cmnd *SCpnt, unsigned int reset_flags) 
+{
+	struct sbp2scsi_host_info *hi = (struct sbp2scsi_host_info *) SCpnt->host->hostdata[0];
+
+	SBP2_ERR("sbp2: reset requested");
+
+	if (hi) {
+		SBP2_ERR("sbp2: generating IEEE-1394 bus reset");
+		hpsb_reset_bus(hi->host, LONG_RESET);
+	}
+
+	return(SCSI_RESET_SUCCESS);
+}
+
+/*
+ * Called by scsi stack to get bios parameters (used by fdisk, and at boot).
+ */
+static int sbp2scsi_biosparam (Scsi_Disk *disk, kdev_t dev, int geom[]) 
+{
+	int heads, sectors, cylinders;
+
+	SBP2_DEBUG("sbp2: request for bios parameters");
+
+	heads = 64;
+	sectors = 32;
+	cylinders = disk->capacity / (heads * sectors);
+
+	if (cylinders > 1024) {
+		heads = 255;
+		sectors = 63;
+		cylinders = disk->capacity / (heads * sectors);
+	}
+
+	geom[0] = heads;
+	geom[1] = sectors;
+	geom[2] = cylinders;
+
+	return(0);
+}
+
+/*
+ * This routine is called at setup (init) and does nothing. Not used here.   =)
+ */
+void sbp2scsi_setup( char *str, int *ints) 
+{
+	SBP2_DEBUG("sbp2: sbp2scsi_setup");
+	return;
+}
+
+/*
+ * This is our detection routine, and is where we init everything.
+ */
+static int sbp2scsi_detect (Scsi_Host_Template *tpnt) 
+{
+	SBP2_DEBUG("sbp2: sbp2scsi_detect");
+
+	global_scsi_tpnt = tpnt;
+
+	global_scsi_tpnt->proc_name = "sbp2";
+
+	/*
+	 * Module load option for force one command at a time
+	 */
+	if (serialize_io) {
+		SBP2_ERR("sbp2: Driver forced to serialize I/O (serialize_io = 1)");
+		global_scsi_tpnt->can_queue = 1;
+		global_scsi_tpnt->cmd_per_lun = 1;
+	}
+
+	/*
+	 * Module load option to limit max size of requests from the scsi drivers
+	 */
+	if (no_large_packets) {
+		SBP2_ERR("sbp2: Driver forced to limit max transfer size (no_large_packets = 1)");
+		global_scsi_tpnt->sg_tablesize = 0x1f;
+		global_scsi_tpnt->use_clustering = DISABLE_CLUSTERING;
+	}
+
+	if (no_bus_scan) {
+		SBP2_ERR("sbp2: Initial scsi bus scan deferred (no_bus_scan = 1)");
+	}
+
+	if (mode_sense_hack) {
+		SBP2_ERR("sbp2: Mode sense emulation enabled (mode_sense_hack = 1)");
+	}
+
+	sbp2_init();
+
+	if (!sbp2_host_count) {
+		SBP2_ERR("sbp2: Please load the lower level IEEE-1394 driver (e.g. ohci1394) before sbp2...");
+		if (sbp2_hl_handle) {
+			hpsb_unregister_highlevel(sbp2_hl_handle);
+			sbp2_hl_handle = NULL;
+		}
+	}
+
+	/*
+	 * Since we are returning this count, it means that sbp2 must be loaded "after" the 
+	 * host adapter module... 	 
+	 */
+	return(sbp2_host_count);
+}
+
+/*
+ * This function is called from sbp2_add_host, and is where we register our scsi host
+ */
+static void sbp2scsi_register_scsi_host(struct sbp2scsi_host_info *hi)
+{
+	struct Scsi_Host *shpnt = NULL;
+
+	SBP2_DEBUG("sbp2: sbp2scsi_register_scsi_host");
+	SBP2_DEBUG("sbp2: sbp2scsi_host_info = %p", hi);
+
+	/*
+	 * Let's register with the scsi stack
+	 */
+	if (global_scsi_tpnt) {
+
+		shpnt = scsi_register (global_scsi_tpnt, sizeof(void *));
+
+		/*
+		 * If successful, save off a context (to be used when SCSI commands are received)
+		 */
+		if (shpnt) {
+			shpnt->hostdata[0] = (unsigned long)hi;
+		}
+	}
+
+	return;
+}
+
+/*
+ * Called when our module is released
+ */
+static int sbp2scsi_release(struct Scsi_Host *host)
+{
+	SBP2_DEBUG("sbp2: sbp2scsi_release");
+	sbp2_cleanup();
+	return(0);
+}
+
+/*
+ * Called for contents of procfs
+ */
+static const char *sbp2scsi_info (struct Scsi_Host *host)
+{
+	return "IEEE-1394 SBP-2 protocol driver";
+}
+
+/*
+ * Module related section
+ */
+
+MODULE_AUTHOR("James Goodwin <jamesg@filanet.com>");
+MODULE_DESCRIPTION("IEEE-1394 SBP-2 protocol driver");
+MODULE_SUPPORTED_DEVICE("sbp2");
+
+/*
+ * SCSI host template
+ */
+static Scsi_Host_Template driver_template = SBP2SCSI;
+
+#include "../scsi/scsi_module.c"

FUNET's LINUX-ADM group, linux-adm@nic.funet.fi
TCL-scripts by Sam Shen (who was at: slshen@lbl.gov)