staging: add Lustre file system client support

Lustre is the most deployed distributed file system
in the HPC (High Performance Computing) world. The patch
adds its client side support.

The code is not very clean and needs to live in drivers/staging
for some time for continuing cleanup work. See
drivers/staging/lustre/TODO for details.

The code is based on Lustre master commit faefbfc04

commit faefbfc0460bc00f2ee4c1c1c86aa1e39b9eea49
Author: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Date:   Tue Apr 30 23:05:21 2013 +0400

    LU-3244 utils: tunefs.lustre should preserve virgin label

Plus a few under-review patches on Whamcloud gerrit:
3.8 kernel support:
http://review.whamcloud.com/#change,5973
http://review.whamcloud.com/#change,5974
http://review.whamcloud.com/#change,5768
http://review.whamcloud.com/#change,5781
http://review.whamcloud.com/#change,5763
http://review.whamcloud.com/#change,5613
http://review.whamcloud.com/#change,5655

3.9 kernel support:
http://review.whamcloud.com/#change,5898
http://review.whamcloud.com/#change,5899

Kconfig/Kbuild:
http://review.whamcloud.com/#change,4646
http://review.whamcloud.com/#change,4644

libcfs cleanup:
http://review.whamcloud.com/#change,2831
http://review.whamcloud.com/#change,4775
http://review.whamcloud.com/#change,4776
http://review.whamcloud.com/#change,4777
http://review.whamcloud.com/#change,4778
http://review.whamcloud.com/#change,4779
http://review.whamcloud.com/#change,4780

All starting/trailing whitespaces are removed, to match kernel
coding style. Also ran scripts/cleanfile on all lustre source files.

[maked the Kconfig depend on BROKEN as the recent procfs changes causes
this to fail - gregkh]

Signed-off-by: Peng Tao <tao.peng@emc.com>
Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

16 files changed, 3616 insertions, 0 deletions

diff --git a/drivers/staging/lustre/include/linux/lnet/api-support.h b/drivers/staging/lustre/include/linux/lnet/api-support.h
new file mode 100644
index 000000000000..a8d91dbe6060
--- /dev/null
+++ b/drivers/staging/lustre/include/linux/lnet/api-support.h
@@ -0,0 +1,44 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * 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 version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
+ *
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ */
+
+#ifndef __LNET_API_SUPPORT_H__
+#define __LNET_API_SUPPORT_H__
+
+#include <linux/lnet/linux/api-support.h>
+
+#include <linux/libcfs/libcfs.h>
+#include <linux/lnet/types.h>
+#include <linux/lnet/lnet.h>
+
+#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/api.h b/drivers/staging/lustre/include/linux/lnet/api.h
new file mode 100644
index 000000000000..e8642e33860d
--- /dev/null
+++ b/drivers/staging/lustre/include/linux/lnet/api.h
@@ -0,0 +1,220 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * 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 version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
+ *
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ */
+
+#ifndef __LNET_API_H__
+#define __LNET_API_H__
+
+/** \defgroup lnet LNet
+ *
+ * The Lustre Networking subsystem.
+ *
+ * LNet is an asynchronous message-passing API, which provides an unreliable
+ * connectionless service that can't guarantee any order. It supports OFA IB,
+ * TCP/IP, and Cray Portals, and routes between heterogeneous networks.
+ *
+ * LNet can run both in OS kernel space and in userspace as a library.
+ * @{
+ */
+
+#include <linux/lnet/types.h>
+
+/** \defgroup lnet_init_fini Initialization and cleanup
+ * The LNet must be properly initialized before any LNet calls can be made.
+ * @{ */
+int LNetInit(void);
+void LNetFini(void);
+
+int LNetNIInit(lnet_pid_t requested_pid);
+int LNetNIFini(void);
+/** @} lnet_init_fini */
+
+/** \defgroup lnet_addr LNet addressing and basic types
+ *
+ * Addressing scheme and basic data types of LNet.
+ *
+ * The LNet API is memory-oriented, so LNet must be able to address not only
+ * end-points but also memory region within a process address space.
+ * An ::lnet_nid_t addresses an end-point. An ::lnet_pid_t identifies a process
+ * in a node. A portal represents an opening in the address space of a
+ * process. Match bits is criteria to identify a region of memory inside a
+ * portal, and offset specifies an offset within the memory region.
+ *
+ * LNet creates a table of portals for each process during initialization.
+ * This table has MAX_PORTALS entries and its size can't be dynamically
+ * changed. A portal stays empty until the owning process starts to add
+ * memory regions to it. A portal is sometimes called an index because
+ * it's an entry in the portals table of a process.
+ *
+ * \see LNetMEAttach
+ * @{ */
+int LNetGetId(unsigned int index, lnet_process_id_t *id);
+int LNetDist(lnet_nid_t nid, lnet_nid_t *srcnid, __u32 *order);
+void LNetSnprintHandle(char *str, int str_len, lnet_handle_any_t handle);
+
+/** @} lnet_addr */
+
+
+/** \defgroup lnet_me Match entries
+ *
+ * A match entry (abbreviated as ME) describes a set of criteria to accept
+ * incoming requests.
+ *
+ * A portal is essentially a match list plus a set of attributes. A match
+ * list is a chain of MEs. Each ME includes a pointer to a memory descriptor
+ * and a set of match criteria. The match criteria can be used to reject
+ * incoming requests based on process ID or the match bits provided in the
+ * request. MEs can be dynamically inserted into a match list by LNetMEAttach()
+ * and LNetMEInsert(), and removed from its list by LNetMEUnlink().
+ * @{ */
+int LNetMEAttach(unsigned int      portal,
+		 lnet_process_id_t match_id_in,
+		 __u64	     match_bits_in,
+		 __u64	     ignore_bits_in,
+		 lnet_unlink_t     unlink_in,
+		 lnet_ins_pos_t    pos_in,
+		 lnet_handle_me_t *handle_out);
+
+int LNetMEInsert(lnet_handle_me_t  current_in,
+		 lnet_process_id_t match_id_in,
+		 __u64	     match_bits_in,
+		 __u64	     ignore_bits_in,
+		 lnet_unlink_t     unlink_in,
+		 lnet_ins_pos_t    position_in,
+		 lnet_handle_me_t *handle_out);
+
+int LNetMEUnlink(lnet_handle_me_t current_in);
+/** @} lnet_me */
+
+/** \defgroup lnet_md Memory descriptors
+ *
+ * A memory descriptor contains information about a region of a user's
+ * memory (either in kernel or user space) and optionally points to an
+ * event queue where information about the operations performed on the
+ * memory descriptor are recorded. Memory descriptor is abbreviated as
+ * MD and can be used interchangeably with the memory region it describes.
+ *
+ * The LNet API provides two operations to create MDs: LNetMDAttach()
+ * and LNetMDBind(); one operation to unlink and release the resources
+ * associated with a MD: LNetMDUnlink().
+ * @{ */
+int LNetMDAttach(lnet_handle_me_t  current_in,
+		 lnet_md_t	 md_in,
+		 lnet_unlink_t     unlink_in,
+		 lnet_handle_md_t *handle_out);
+
+int LNetMDBind(lnet_md_t	 md_in,
+	       lnet_unlink_t     unlink_in,
+	       lnet_handle_md_t *handle_out);
+
+int LNetMDUnlink(lnet_handle_md_t md_in);
+/** @} lnet_md */
+
+/** \defgroup lnet_eq Events and event queues
+ *
+ * Event queues (abbreviated as EQ) are used to log operations performed on
+ * local MDs. In particular, they signal the completion of a data transmission
+ * into or out of a MD. They can also be used to hold acknowledgments for
+ * completed PUT operations and indicate when a MD has been unlinked. Multiple
+ * MDs can share a single EQ. An EQ may have an optional event handler
+ * associated with it. If an event handler exists, it will be run for each
+ * event that is deposited into the EQ.
+ *
+ * In addition to the lnet_handle_eq_t, the LNet API defines two types
+ * associated with events: The ::lnet_event_kind_t defines the kinds of events
+ * that can be stored in an EQ. The lnet_event_t defines a structure that
+ * holds the information about with an event.
+ *
+ * There are five functions for dealing with EQs: LNetEQAlloc() is used to
+ * create an EQ and allocate the resources needed, while LNetEQFree()
+ * releases these resources and free the EQ. LNetEQGet() retrieves the next
+ * event from an EQ, and LNetEQWait() can be used to block a process until
+ * an EQ has at least one event. LNetEQPoll() can be used to test or wait
+ * on multiple EQs.
+ * @{ */
+int LNetEQAlloc(unsigned int       count_in,
+		lnet_eq_handler_t  handler,
+		lnet_handle_eq_t  *handle_out);
+
+int LNetEQFree(lnet_handle_eq_t eventq_in);
+
+int LNetEQGet(lnet_handle_eq_t  eventq_in,
+	      lnet_event_t     *event_out);
+
+
+int LNetEQWait(lnet_handle_eq_t  eventq_in,
+	       lnet_event_t     *event_out);
+
+int LNetEQPoll(lnet_handle_eq_t *eventqs_in,
+	       int	       neq_in,
+	       int	       timeout_ms,
+	       lnet_event_t     *event_out,
+	       int	      *which_eq_out);
+/** @} lnet_eq */
+
+/** \defgroup lnet_data Data movement operations
+ *
+ * The LNet API provides two data movement operations: LNetPut()
+ * and LNetGet().
+ * @{ */
+int LNetPut(lnet_nid_t	self,
+	    lnet_handle_md_t  md_in,
+	    lnet_ack_req_t    ack_req_in,
+	    lnet_process_id_t target_in,
+	    unsigned int      portal_in,
+	    __u64	     match_bits_in,
+	    unsigned int      offset_in,
+	    __u64	     hdr_data_in);
+
+int LNetGet(lnet_nid_t	self,
+	    lnet_handle_md_t  md_in,
+	    lnet_process_id_t target_in,
+	    unsigned int      portal_in,
+	    __u64	     match_bits_in,
+	    unsigned int      offset_in);
+/** @} lnet_data */
+
+
+/** \defgroup lnet_misc Miscellaneous operations.
+ * Miscellaneous operations.
+ * @{ */
+
+int LNetSetLazyPortal(int portal);
+int LNetClearLazyPortal(int portal);
+int LNetCtl(unsigned int cmd, void *arg);
+int LNetSetAsync(lnet_process_id_t id, int nasync);
+
+/** @} lnet_misc */
+
+/** @} lnet */
+#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/lib-lnet.h b/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
new file mode 100644
index 000000000000..59bff0bea816
--- /dev/null
+++ b/drivers/staging/lustre/include/linux/lnet/lib-lnet.h
@@ -0,0 +1,874 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * 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 version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
+ *
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ *
+ * Copyright (c) 2012, Intel Corporation.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ *
+ * lnet/include/lnet/lib-lnet.h
+ *
+ * Top level include for library side routines
+ */
+
+#ifndef __LNET_LIB_LNET_H__
+#define __LNET_LIB_LNET_H__
+
+#include <linux/lnet/linux/lib-lnet.h>
+
+#include <linux/libcfs/libcfs.h>
+#include <linux/lnet/types.h>
+#include <linux/lnet/lnet.h>
+#include <linux/lnet/lib-types.h>
+
+extern lnet_t  the_lnet;			/* THE network */
+
+#if  defined(LNET_USE_LIB_FREELIST)
+/* 1 CPT, simplify implementation... */
+# define LNET_CPT_MAX_BITS      0
+
+#else /* KERNEL and no freelist */
+
+# if (BITS_PER_LONG == 32)
+/* 2 CPTs, allowing more CPTs might make us under memory pressure */
+#  define LNET_CPT_MAX_BITS     1
+
+# else /* 64-bit system */
+/*
+ * 256 CPTs for thousands of CPUs, allowing more CPTs might make us
+ * under risk of consuming all lh_cookie.
+ */
+#  define LNET_CPT_MAX_BITS     8
+# endif /* BITS_PER_LONG == 32 */
+#endif
+
+/* max allowed CPT number */
+#define LNET_CPT_MAX	    (1 << LNET_CPT_MAX_BITS)
+
+#define LNET_CPT_NUMBER	 (the_lnet.ln_cpt_number)
+#define LNET_CPT_BITS	   (the_lnet.ln_cpt_bits)
+#define LNET_CPT_MASK	   ((1ULL << LNET_CPT_BITS) - 1)
+
+/** exclusive lock */
+#define LNET_LOCK_EX	    CFS_PERCPT_LOCK_EX
+
+static inline int lnet_is_wire_handle_none (lnet_handle_wire_t *wh)
+{
+	return (wh->wh_interface_cookie == LNET_WIRE_HANDLE_COOKIE_NONE &&
+		wh->wh_object_cookie == LNET_WIRE_HANDLE_COOKIE_NONE);
+}
+
+static inline int lnet_md_exhausted (lnet_libmd_t *md)
+{
+	return (md->md_threshold == 0 ||
+		((md->md_options & LNET_MD_MAX_SIZE) != 0 &&
+		 md->md_offset + md->md_max_size > md->md_length));
+}
+
+static inline int lnet_md_unlinkable (lnet_libmd_t *md)
+{
+	/* Should unlink md when its refcount is 0 and either:
+	 *  - md has been flagged for deletion (by auto unlink or LNetM[DE]Unlink,
+	 *    in the latter case md may not be exhausted).
+	 *  - auto unlink is on and md is exhausted.
+	 */
+	if (md->md_refcount != 0)
+		return 0;
+
+	if ((md->md_flags & LNET_MD_FLAG_ZOMBIE) != 0)
+		return 1;
+
+	return ((md->md_flags & LNET_MD_FLAG_AUTO_UNLINK) != 0 &&
+		lnet_md_exhausted(md));
+}
+
+#define lnet_cpt_table()	(the_lnet.ln_cpt_table)
+#define lnet_cpt_current()	cfs_cpt_current(the_lnet.ln_cpt_table, 1)
+
+static inline int
+lnet_cpt_of_cookie(__u64 cookie)
+{
+	unsigned int cpt = (cookie >> LNET_COOKIE_TYPE_BITS) & LNET_CPT_MASK;
+
+	/* LNET_CPT_NUMBER doesn't have to be power2, which means we can
+	 * get illegal cpt from it's invalid cookie */
+	return cpt < LNET_CPT_NUMBER ? cpt : cpt % LNET_CPT_NUMBER;
+}
+
+static inline void
+lnet_res_lock(int cpt)
+{
+	cfs_percpt_lock(the_lnet.ln_res_lock, cpt);
+}
+
+static inline void
+lnet_res_unlock(int cpt)
+{
+	cfs_percpt_unlock(the_lnet.ln_res_lock, cpt);
+}
+
+static inline int
+lnet_res_lock_current(void)
+{
+	int cpt = lnet_cpt_current();
+
+	lnet_res_lock(cpt);
+	return cpt;
+}
+
+static inline void
+lnet_net_lock(int cpt)
+{
+	cfs_percpt_lock(the_lnet.ln_net_lock, cpt);
+}
+
+static inline void
+lnet_net_unlock(int cpt)
+{
+	cfs_percpt_unlock(the_lnet.ln_net_lock, cpt);
+}
+
+static inline int
+lnet_net_lock_current(void)
+{
+	int cpt = lnet_cpt_current();
+
+	lnet_net_lock(cpt);
+	return cpt;
+}
+
+#define LNET_LOCK()		lnet_net_lock(LNET_LOCK_EX)
+#define LNET_UNLOCK()		lnet_net_unlock(LNET_LOCK_EX)
+
+
+#define lnet_ptl_lock(ptl)	spin_lock(&(ptl)->ptl_lock)
+#define lnet_ptl_unlock(ptl)	spin_unlock(&(ptl)->ptl_lock)
+#define lnet_eq_wait_lock()	spin_lock(&the_lnet.ln_eq_wait_lock)
+#define lnet_eq_wait_unlock()	spin_unlock(&the_lnet.ln_eq_wait_lock)
+#define lnet_ni_lock(ni)	spin_lock(&(ni)->ni_lock)
+#define lnet_ni_unlock(ni)	spin_unlock(&(ni)->ni_lock)
+#define LNET_MUTEX_LOCK(m)	mutex_lock(m)
+#define LNET_MUTEX_UNLOCK(m)	mutex_unlock(m)
+
+
+#define MAX_PORTALS     64
+
+/* these are only used by code with LNET_USE_LIB_FREELIST, but we still
+ * exported them to !LNET_USE_LIB_FREELIST for easy implemetation */
+#define LNET_FL_MAX_MES		2048
+#define LNET_FL_MAX_MDS		2048
+#define LNET_FL_MAX_EQS		512
+#define LNET_FL_MAX_MSGS	2048    /* Outstanding messages */
+
+#ifdef LNET_USE_LIB_FREELIST
+
+int lnet_freelist_init(lnet_freelist_t *fl, int n, int size);
+void lnet_freelist_fini(lnet_freelist_t *fl);
+
+static inline void *
+lnet_freelist_alloc (lnet_freelist_t *fl)
+{
+	/* ALWAYS called with liblock held */
+	lnet_freeobj_t *o;
+
+	if (list_empty (&fl->fl_list))
+		return (NULL);
+
+	o = list_entry (fl->fl_list.next, lnet_freeobj_t, fo_list);
+	list_del (&o->fo_list);
+	return ((void *)&o->fo_contents);
+}
+
+static inline void
+lnet_freelist_free (lnet_freelist_t *fl, void *obj)
+{
+	/* ALWAYS called with liblock held */
+	lnet_freeobj_t *o = list_entry (obj, lnet_freeobj_t, fo_contents);
+
+	list_add (&o->fo_list, &fl->fl_list);
+}
+
+
+static inline lnet_eq_t *
+lnet_eq_alloc (void)
+{
+	/* NEVER called with resource lock held */
+	struct lnet_res_container *rec = &the_lnet.ln_eq_container;
+	lnet_eq_t		  *eq;
+
+	LASSERT(LNET_CPT_NUMBER == 1);
+
+	lnet_res_lock(0);
+	eq = (lnet_eq_t *)lnet_freelist_alloc(&rec->rec_freelist);
+	lnet_res_unlock(0);
+
+	return eq;
+}
+
+static inline void
+lnet_eq_free_locked(lnet_eq_t *eq)
+{
+	/* ALWAYS called with resource lock held */
+	struct lnet_res_container *rec = &the_lnet.ln_eq_container;
+
+	LASSERT(LNET_CPT_NUMBER == 1);
+	lnet_freelist_free(&rec->rec_freelist, eq);
+}
+
+static inline void
+lnet_eq_free(lnet_eq_t *eq)
+{
+	lnet_res_lock(0);
+	lnet_eq_free_locked(eq);
+	lnet_res_unlock(0);
+}
+
+static inline lnet_libmd_t *
+lnet_md_alloc (lnet_md_t *umd)
+{
+	/* NEVER called with resource lock held */
+	struct lnet_res_container *rec = the_lnet.ln_md_containers[0];
+	lnet_libmd_t		  *md;
+
+	LASSERT(LNET_CPT_NUMBER == 1);
+
+	lnet_res_lock(0);
+	md = (lnet_libmd_t *)lnet_freelist_alloc(&rec->rec_freelist);
+	lnet_res_unlock(0);
+
+	if (md != NULL)
+		INIT_LIST_HEAD(&md->md_list);
+
+	return md;
+}
+
+static inline void
+lnet_md_free_locked(lnet_libmd_t *md)
+{
+	/* ALWAYS called with resource lock held */
+	struct lnet_res_container *rec = the_lnet.ln_md_containers[0];
+
+	LASSERT(LNET_CPT_NUMBER == 1);
+	lnet_freelist_free(&rec->rec_freelist, md);
+}
+
+static inline void
+lnet_md_free(lnet_libmd_t *md)
+{
+	lnet_res_lock(0);
+	lnet_md_free_locked(md);
+	lnet_res_unlock(0);
+}
+
+static inline lnet_me_t *
+lnet_me_alloc(void)
+{
+	/* NEVER called with resource lock held */
+	struct lnet_res_container *rec = the_lnet.ln_me_containers[0];
+	lnet_me_t		  *me;
+
+	LASSERT(LNET_CPT_NUMBER == 1);
+
+	lnet_res_lock(0);
+	me = (lnet_me_t *)lnet_freelist_alloc(&rec->rec_freelist);
+	lnet_res_unlock(0);
+
+	return me;
+}
+
+static inline void
+lnet_me_free_locked(lnet_me_t *me)
+{
+	/* ALWAYS called with resource lock held */
+	struct lnet_res_container *rec = the_lnet.ln_me_containers[0];
+
+	LASSERT(LNET_CPT_NUMBER == 1);
+	lnet_freelist_free(&rec->rec_freelist, me);
+}
+
+static inline void
+lnet_me_free(lnet_me_t *me)
+{
+	lnet_res_lock(0);
+	lnet_me_free_locked(me);
+	lnet_res_unlock(0);
+}
+
+static inline lnet_msg_t *
+lnet_msg_alloc (void)
+{
+	/* NEVER called with network lock held */
+	struct lnet_msg_container *msc = the_lnet.ln_msg_containers[0];
+	lnet_msg_t		  *msg;
+
+	LASSERT(LNET_CPT_NUMBER == 1);
+
+	lnet_net_lock(0);
+	msg = (lnet_msg_t *)lnet_freelist_alloc(&msc->msc_freelist);
+	lnet_net_unlock(0);
+
+	if (msg != NULL) {
+		/* NULL pointers, clear flags etc */
+		memset(msg, 0, sizeof(*msg));
+	}
+	return msg;
+}
+
+static inline void
+lnet_msg_free_locked(lnet_msg_t *msg)
+{
+	/* ALWAYS called with network lock held */
+	struct lnet_msg_container *msc = the_lnet.ln_msg_containers[0];
+
+	LASSERT(LNET_CPT_NUMBER == 1);
+	LASSERT(!msg->msg_onactivelist);
+	lnet_freelist_free(&msc->msc_freelist, msg);
+}
+
+static inline void
+lnet_msg_free (lnet_msg_t *msg)
+{
+	lnet_net_lock(0);
+	lnet_msg_free_locked(msg);
+	lnet_net_unlock(0);
+}
+
+#else /* !LNET_USE_LIB_FREELIST */
+
+static inline lnet_eq_t *
+lnet_eq_alloc (void)
+{
+	/* NEVER called with liblock held */
+	lnet_eq_t *eq;
+
+	LIBCFS_ALLOC(eq, sizeof(*eq));
+	return (eq);
+}
+
+static inline void
+lnet_eq_free(lnet_eq_t *eq)
+{
+	/* ALWAYS called with resource lock held */
+	LIBCFS_FREE(eq, sizeof(*eq));
+}
+
+static inline lnet_libmd_t *
+lnet_md_alloc (lnet_md_t *umd)
+{
+	/* NEVER called with liblock held */
+	lnet_libmd_t *md;
+	unsigned int  size;
+	unsigned int  niov;
+
+	if ((umd->options & LNET_MD_KIOV) != 0) {
+		niov = umd->length;
+		size = offsetof(lnet_libmd_t, md_iov.kiov[niov]);
+	} else {
+		niov = ((umd->options & LNET_MD_IOVEC) != 0) ?
+		       umd->length : 1;
+		size = offsetof(lnet_libmd_t, md_iov.iov[niov]);
+	}
+
+	LIBCFS_ALLOC(md, size);
+
+	if (md != NULL) {
+		/* Set here in case of early free */
+		md->md_options = umd->options;
+		md->md_niov = niov;
+		INIT_LIST_HEAD(&md->md_list);
+	}
+
+	return (md);
+}
+
+static inline void
+lnet_md_free(lnet_libmd_t *md)
+{
+	/* ALWAYS called with resource lock held */
+	unsigned int  size;
+
+	if ((md->md_options & LNET_MD_KIOV) != 0)
+		size = offsetof(lnet_libmd_t, md_iov.kiov[md->md_niov]);
+	else
+		size = offsetof(lnet_libmd_t, md_iov.iov[md->md_niov]);
+
+	LIBCFS_FREE(md, size);
+}
+
+static inline lnet_me_t *
+lnet_me_alloc (void)
+{
+	/* NEVER called with liblock held */
+	lnet_me_t *me;
+
+	LIBCFS_ALLOC(me, sizeof(*me));
+	return (me);
+}
+
+static inline void
+lnet_me_free(lnet_me_t *me)
+{
+	/* ALWAYS called with resource lock held */
+	LIBCFS_FREE(me, sizeof(*me));
+}
+
+static inline lnet_msg_t *
+lnet_msg_alloc(void)
+{
+	/* NEVER called with liblock held */
+	lnet_msg_t *msg;
+
+	LIBCFS_ALLOC(msg, sizeof(*msg));
+
+	/* no need to zero, LIBCFS_ALLOC does for us */
+	return (msg);
+}
+
+static inline void
+lnet_msg_free(lnet_msg_t *msg)
+{
+	/* ALWAYS called with network lock held */
+	LASSERT(!msg->msg_onactivelist);
+	LIBCFS_FREE(msg, sizeof(*msg));
+}
+
+#define lnet_eq_free_locked(eq)		lnet_eq_free(eq)
+#define lnet_md_free_locked(md)		lnet_md_free(md)
+#define lnet_me_free_locked(me)		lnet_me_free(me)
+#define lnet_msg_free_locked(msg)	lnet_msg_free(msg)
+
+#endif /* LNET_USE_LIB_FREELIST */
+
+lnet_libhandle_t *lnet_res_lh_lookup(struct lnet_res_container *rec,
+				     __u64 cookie);
+void lnet_res_lh_initialize(struct lnet_res_container *rec,
+			    lnet_libhandle_t *lh);
+static inline void
+lnet_res_lh_invalidate(lnet_libhandle_t *lh)
+{
+	/* ALWAYS called with resource lock held */
+	/* NB: cookie is still useful, don't reset it */
+	list_del(&lh->lh_hash_chain);
+}
+
+static inline void
+lnet_eq2handle (lnet_handle_eq_t *handle, lnet_eq_t *eq)
+{
+	if (eq == NULL) {
+		LNetInvalidateHandle(handle);
+		return;
+	}
+
+	handle->cookie = eq->eq_lh.lh_cookie;
+}
+
+static inline lnet_eq_t *
+lnet_handle2eq(lnet_handle_eq_t *handle)
+{
+	/* ALWAYS called with resource lock held */
+	lnet_libhandle_t *lh;
+
+	lh = lnet_res_lh_lookup(&the_lnet.ln_eq_container, handle->cookie);
+	if (lh == NULL)
+		return NULL;
+
+	return lh_entry(lh, lnet_eq_t, eq_lh);
+}
+
+static inline void
+lnet_md2handle (lnet_handle_md_t *handle, lnet_libmd_t *md)
+{
+	handle->cookie = md->md_lh.lh_cookie;
+}
+
+static inline lnet_libmd_t *
+lnet_handle2md(lnet_handle_md_t *handle)
+{
+	/* ALWAYS called with resource lock held */
+	lnet_libhandle_t *lh;
+	int		 cpt;
+
+	cpt = lnet_cpt_of_cookie(handle->cookie);
+	lh = lnet_res_lh_lookup(the_lnet.ln_md_containers[cpt],
+				handle->cookie);
+	if (lh == NULL)
+		return NULL;
+
+	return lh_entry(lh, lnet_libmd_t, md_lh);
+}
+
+static inline lnet_libmd_t *
+lnet_wire_handle2md(lnet_handle_wire_t *wh)
+{
+	/* ALWAYS called with resource lock held */
+	lnet_libhandle_t *lh;
+	int		 cpt;
+
+	if (wh->wh_interface_cookie != the_lnet.ln_interface_cookie)
+		return NULL;
+
+	cpt = lnet_cpt_of_cookie(wh->wh_object_cookie);
+	lh = lnet_res_lh_lookup(the_lnet.ln_md_containers[cpt],
+				wh->wh_object_cookie);
+	if (lh == NULL)
+		return NULL;
+
+	return lh_entry(lh, lnet_libmd_t, md_lh);
+}
+
+static inline void
+lnet_me2handle (lnet_handle_me_t *handle, lnet_me_t *me)
+{
+	handle->cookie = me->me_lh.lh_cookie;
+}
+
+static inline lnet_me_t *
+lnet_handle2me(lnet_handle_me_t *handle)
+{
+	/* ALWAYS called with resource lock held */
+	lnet_libhandle_t *lh;
+	int		 cpt;
+
+	cpt = lnet_cpt_of_cookie(handle->cookie);
+	lh = lnet_res_lh_lookup(the_lnet.ln_me_containers[cpt],
+				handle->cookie);
+	if (lh == NULL)
+		return NULL;
+
+	return lh_entry(lh, lnet_me_t, me_lh);
+}
+
+static inline void
+lnet_peer_addref_locked(lnet_peer_t *lp)
+{
+	LASSERT (lp->lp_refcount > 0);
+	lp->lp_refcount++;
+}
+
+extern void lnet_destroy_peer_locked(lnet_peer_t *lp);
+
+static inline void
+lnet_peer_decref_locked(lnet_peer_t *lp)
+{
+	LASSERT (lp->lp_refcount > 0);
+	lp->lp_refcount--;
+	if (lp->lp_refcount == 0)
+		lnet_destroy_peer_locked(lp);
+}
+
+static inline int
+lnet_isrouter(lnet_peer_t *lp)
+{
+	return lp->lp_rtr_refcount != 0;
+}
+
+static inline void
+lnet_ni_addref_locked(lnet_ni_t *ni, int cpt)
+{
+	LASSERT(cpt >= 0 && cpt < LNET_CPT_NUMBER);
+	LASSERT(*ni->ni_refs[cpt] >= 0);
+
+	(*ni->ni_refs[cpt])++;
+}
+
+static inline void
+lnet_ni_addref(lnet_ni_t *ni)
+{
+	lnet_net_lock(0);
+	lnet_ni_addref_locked(ni, 0);
+	lnet_net_unlock(0);
+}
+
+static inline void
+lnet_ni_decref_locked(lnet_ni_t *ni, int cpt)
+{
+	LASSERT(cpt >= 0 && cpt < LNET_CPT_NUMBER);
+	LASSERT(*ni->ni_refs[cpt] > 0);
+
+	(*ni->ni_refs[cpt])--;
+}
+
+static inline void
+lnet_ni_decref(lnet_ni_t *ni)
+{
+	lnet_net_lock(0);
+	lnet_ni_decref_locked(ni, 0);
+	lnet_net_unlock(0);
+}
+
+void lnet_ni_free(lnet_ni_t *ni);
+
+static inline int
+lnet_nid2peerhash(lnet_nid_t nid)
+{
+	return cfs_hash_long(nid, LNET_PEER_HASH_BITS);
+}
+
+static inline struct list_head *
+lnet_net2rnethash(__u32 net)
+{
+	return &the_lnet.ln_remote_nets_hash[(LNET_NETNUM(net) +
+		LNET_NETTYP(net)) &
+		((1U << the_lnet.ln_remote_nets_hbits) - 1)];
+}
+
+extern lnd_t the_lolnd;
+
+
+extern int lnet_cpt_of_nid_locked(lnet_nid_t nid);
+extern int lnet_cpt_of_nid(lnet_nid_t nid);
+extern lnet_ni_t *lnet_nid2ni_locked(lnet_nid_t nid, int cpt);
+extern lnet_ni_t *lnet_net2ni_locked(__u32 net, int cpt);
+extern lnet_ni_t *lnet_net2ni(__u32 net);
+
+int lnet_notify(lnet_ni_t *ni, lnet_nid_t peer, int alive, cfs_time_t when);
+void lnet_notify_locked(lnet_peer_t *lp, int notifylnd, int alive, cfs_time_t when);
+int lnet_add_route(__u32 net, unsigned int hops, lnet_nid_t gateway_nid);
+int lnet_check_routes(void);
+int lnet_del_route(__u32 net, lnet_nid_t gw_nid);
+void lnet_destroy_routes(void);
+int lnet_get_route(int idx, __u32 *net, __u32 *hops,
+		   lnet_nid_t *gateway, __u32 *alive);
+void lnet_proc_init(void);
+void lnet_proc_fini(void);
+int  lnet_rtrpools_alloc(int im_a_router);
+void lnet_rtrpools_free(void);
+lnet_remotenet_t *lnet_find_net_locked (__u32 net);
+
+int lnet_islocalnid(lnet_nid_t nid);
+int lnet_islocalnet(__u32 net);
+
+void lnet_msg_attach_md(lnet_msg_t *msg, lnet_libmd_t *md,
+			unsigned int offset, unsigned int mlen);
+void lnet_msg_detach_md(lnet_msg_t *msg, int status);
+void lnet_build_unlink_event(lnet_libmd_t *md, lnet_event_t *ev);
+void lnet_build_msg_event(lnet_msg_t *msg, lnet_event_kind_t ev_type);
+void lnet_msg_commit(lnet_msg_t *msg, int cpt);
+void lnet_msg_decommit(lnet_msg_t *msg, int cpt, int status);
+
+void lnet_eq_enqueue_event(lnet_eq_t *eq, lnet_event_t *ev);
+void lnet_prep_send(lnet_msg_t *msg, int type, lnet_process_id_t target,
+		    unsigned int offset, unsigned int len);
+int lnet_send(lnet_nid_t nid, lnet_msg_t *msg, lnet_nid_t rtr_nid);
+void lnet_return_tx_credits_locked(lnet_msg_t *msg);
+void lnet_return_rx_credits_locked(lnet_msg_t *msg);
+
+/* portals functions */
+/* portals attributes */
+static inline int
+lnet_ptl_is_lazy(lnet_portal_t *ptl)
+{
+	return !!(ptl->ptl_options & LNET_PTL_LAZY);
+}
+
+static inline int
+lnet_ptl_is_unique(lnet_portal_t *ptl)
+{
+	return !!(ptl->ptl_options & LNET_PTL_MATCH_UNIQUE);
+}
+
+static inline int
+lnet_ptl_is_wildcard(lnet_portal_t *ptl)
+{
+	return !!(ptl->ptl_options & LNET_PTL_MATCH_WILDCARD);
+}
+
+static inline void
+lnet_ptl_setopt(lnet_portal_t *ptl, int opt)
+{
+	ptl->ptl_options |= opt;
+}
+
+static inline void
+lnet_ptl_unsetopt(lnet_portal_t *ptl, int opt)
+{
+	ptl->ptl_options &= ~opt;
+}
+
+/* match-table functions */
+struct list_head *lnet_mt_match_head(struct lnet_match_table *mtable,
+			       lnet_process_id_t id, __u64 mbits);
+struct lnet_match_table *lnet_mt_of_attach(unsigned int index,
+					   lnet_process_id_t id, __u64 mbits,
+					   __u64 ignore_bits,
+					   lnet_ins_pos_t pos);
+int lnet_mt_match_md(struct lnet_match_table *mtable,
+		     struct lnet_match_info *info, struct lnet_msg *msg);
+
+/* portals match/attach functions */
+void lnet_ptl_attach_md(lnet_me_t *me, lnet_libmd_t *md,
+			struct list_head *matches, struct list_head *drops);
+void lnet_ptl_detach_md(lnet_me_t *me, lnet_libmd_t *md);
+int lnet_ptl_match_md(struct lnet_match_info *info, struct lnet_msg *msg);
+
+/* initialized and finalize portals */
+int lnet_portals_create(void);
+void lnet_portals_destroy(void);
+
+/* message functions */
+int lnet_parse (lnet_ni_t *ni, lnet_hdr_t *hdr,
+		lnet_nid_t fromnid, void *private, int rdma_req);
+void lnet_recv(lnet_ni_t *ni, void *private, lnet_msg_t *msg, int delayed,
+	       unsigned int offset, unsigned int mlen, unsigned int rlen);
+lnet_msg_t *lnet_create_reply_msg (lnet_ni_t *ni, lnet_msg_t *get_msg);
+void lnet_set_reply_msg_len(lnet_ni_t *ni, lnet_msg_t *msg, unsigned int len);
+void lnet_finalize(lnet_ni_t *ni, lnet_msg_t *msg, int rc);
+void lnet_drop_delayed_msg_list(struct list_head *head, char *reason);
+void lnet_recv_delayed_msg_list(struct list_head *head);
+
+int lnet_msg_container_setup(struct lnet_msg_container *container, int cpt);
+void lnet_msg_container_cleanup(struct lnet_msg_container *container);
+void lnet_msg_containers_destroy(void);
+int lnet_msg_containers_create(void);
+
+char *lnet_msgtyp2str (int type);
+void lnet_print_hdr (lnet_hdr_t * hdr);
+int lnet_fail_nid(lnet_nid_t nid, unsigned int threshold);
+
+void lnet_counters_get(lnet_counters_t *counters);
+void lnet_counters_reset(void);
+
+unsigned int lnet_iov_nob (unsigned int niov, struct iovec *iov);
+int lnet_extract_iov (int dst_niov, struct iovec *dst,
+		      int src_niov, struct iovec *src,
+		      unsigned int offset, unsigned int len);
+
+unsigned int lnet_kiov_nob (unsigned int niov, lnet_kiov_t *iov);
+int lnet_extract_kiov (int dst_niov, lnet_kiov_t *dst,
+		      int src_niov, lnet_kiov_t *src,
+		      unsigned int offset, unsigned int len);
+
+void lnet_copy_iov2iov (unsigned int ndiov, struct iovec *diov,
+			unsigned int doffset,
+			unsigned int nsiov, struct iovec *siov,
+			unsigned int soffset, unsigned int nob);
+void lnet_copy_kiov2iov (unsigned int niov, struct iovec *iov,
+			 unsigned int iovoffset,
+			 unsigned int nkiov, lnet_kiov_t *kiov,
+			 unsigned int kiovoffset, unsigned int nob);
+void lnet_copy_iov2kiov (unsigned int nkiov, lnet_kiov_t *kiov,
+			 unsigned int kiovoffset,
+			 unsigned int niov, struct iovec *iov,
+			 unsigned int iovoffset, unsigned int nob);
+void lnet_copy_kiov2kiov (unsigned int ndkiov, lnet_kiov_t *dkiov,
+			  unsigned int doffset,
+			  unsigned int nskiov, lnet_kiov_t *skiov,
+			  unsigned int soffset, unsigned int nob);
+
+static inline void
+lnet_copy_iov2flat(int dlen, void *dest, unsigned int doffset,
+		   unsigned int nsiov, struct iovec *siov, unsigned int soffset,
+		   unsigned int nob)
+{
+	struct iovec diov = {/*.iov_base = */ dest, /*.iov_len = */ dlen};
+
+	lnet_copy_iov2iov(1, &diov, doffset,
+			  nsiov, siov, soffset, nob);
+}
+
+static inline void
+lnet_copy_kiov2flat(int dlen, void *dest, unsigned int doffset,
+		    unsigned int nsiov, lnet_kiov_t *skiov, unsigned int soffset,
+		    unsigned int nob)
+{
+	struct iovec diov = {/* .iov_base = */ dest, /* .iov_len = */ dlen};
+
+	lnet_copy_kiov2iov(1, &diov, doffset,
+			   nsiov, skiov, soffset, nob);
+}
+
+static inline void
+lnet_copy_flat2iov(unsigned int ndiov, struct iovec *diov, unsigned int doffset,
+		   int slen, void *src, unsigned int soffset, unsigned int nob)
+{
+	struct iovec siov = {/*.iov_base = */ src, /*.iov_len = */slen};
+	lnet_copy_iov2iov(ndiov, diov, doffset,
+			  1, &siov, soffset, nob);
+}
+
+static inline void
+lnet_copy_flat2kiov(unsigned int ndiov, lnet_kiov_t *dkiov, unsigned int doffset,
+		    int slen, void *src, unsigned int soffset, unsigned int nob)
+{
+	struct iovec siov = {/* .iov_base = */ src, /* .iov_len = */ slen};
+	lnet_copy_iov2kiov(ndiov, dkiov, doffset,
+			   1, &siov, soffset, nob);
+}
+
+void lnet_me_unlink(lnet_me_t *me);
+
+void lnet_md_unlink(lnet_libmd_t *md);
+void lnet_md_deconstruct(lnet_libmd_t *lmd, lnet_md_t *umd);
+
+void lnet_register_lnd(lnd_t *lnd);
+void lnet_unregister_lnd(lnd_t *lnd);
+int lnet_set_ip_niaddr (lnet_ni_t *ni);
+
+int lnet_connect(socket_t **sockp, lnet_nid_t peer_nid,
+		 __u32 local_ip, __u32 peer_ip, int peer_port);
+void lnet_connect_console_error(int rc, lnet_nid_t peer_nid,
+				__u32 peer_ip, int port);
+int lnet_count_acceptor_nis(void);
+int lnet_acceptor_timeout(void);
+int lnet_acceptor_port(void);
+
+int lnet_count_acceptor_nis(void);
+int lnet_acceptor_port(void);
+
+int lnet_acceptor_start(void);
+void lnet_acceptor_stop(void);
+
+void lnet_get_tunables(void);
+int lnet_peers_start_down(void);
+int lnet_peer_buffer_credits(lnet_ni_t *ni);
+
+int lnet_router_checker_start(void);
+void lnet_router_checker_stop(void);
+void lnet_swap_pinginfo(lnet_ping_info_t *info);
+
+int lnet_ping_target_init(void);
+void lnet_ping_target_fini(void);
+int lnet_ping(lnet_process_id_t id, int timeout_ms,
+	      lnet_process_id_t *ids, int n_ids);
+
+int lnet_parse_ip2nets (char **networksp, char *ip2nets);
+int lnet_parse_routes (char *route_str, int *im_a_router);
+int lnet_parse_networks (struct list_head *nilist, char *networks);
+
+int lnet_nid2peer_locked(lnet_peer_t **lpp, lnet_nid_t nid, int cpt);
+lnet_peer_t *lnet_find_peer_locked(struct lnet_peer_table *ptable,
+				   lnet_nid_t nid);
+void lnet_peer_tables_cleanup(void);
+void lnet_peer_tables_destroy(void);
+int lnet_peer_tables_create(void);
+void lnet_debug_peer(lnet_nid_t nid);
+
+
+#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/lib-types.h b/drivers/staging/lustre/include/linux/lnet/lib-types.h
new file mode 100644
index 000000000000..86428d4b993e
--- /dev/null
+++ b/drivers/staging/lustre/include/linux/lnet/lib-types.h
@@ -0,0 +1,765 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * 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 version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
+ *
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ *
+ * Copyright (c) 2012, Intel Corporation.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ *
+ * lnet/include/lnet/lib-types.h
+ *
+ * Types used by the library side routines that do not need to be
+ * exposed to the user application
+ */
+
+#ifndef __LNET_LIB_TYPES_H__
+#define __LNET_LIB_TYPES_H__
+
+#include <linux/lnet/linux/lib-types.h>
+
+#include <linux/libcfs/libcfs.h>
+#include <linux/list.h>
+#include <linux/lnet/types.h>
+
+#define WIRE_ATTR       __attribute__((packed))
+
+/* Packed version of lnet_process_id_t to transfer via network */
+typedef struct {
+	lnet_nid_t nid;
+	lnet_pid_t pid;   /* node id / process id */
+} WIRE_ATTR lnet_process_id_packed_t;
+
+/* The wire handle's interface cookie only matches one network interface in
+ * one epoch (i.e. new cookie when the interface restarts or the node
+ * reboots).  The object cookie only matches one object on that interface
+ * during that object's lifetime (i.e. no cookie re-use). */
+typedef struct {
+	__u64 wh_interface_cookie;
+	__u64 wh_object_cookie;
+} WIRE_ATTR lnet_handle_wire_t;
+
+typedef enum {
+	LNET_MSG_ACK = 0,
+	LNET_MSG_PUT,
+	LNET_MSG_GET,
+	LNET_MSG_REPLY,
+	LNET_MSG_HELLO,
+} lnet_msg_type_t;
+
+/* The variant fields of the portals message header are aligned on an 8
+ * byte boundary in the message header.  Note that all types used in these
+ * wire structs MUST be fixed size and the smaller types are placed at the
+ * end. */
+typedef struct lnet_ack {
+	lnet_handle_wire_t  dst_wmd;
+	__u64	       match_bits;
+	__u32	       mlength;
+} WIRE_ATTR lnet_ack_t;
+
+typedef struct lnet_put {
+	lnet_handle_wire_t  ack_wmd;
+	__u64	       match_bits;
+	__u64	       hdr_data;
+	__u32	       ptl_index;
+	__u32	       offset;
+} WIRE_ATTR lnet_put_t;
+
+typedef struct lnet_get {
+	lnet_handle_wire_t  return_wmd;
+	__u64	       match_bits;
+	__u32	       ptl_index;
+	__u32	       src_offset;
+	__u32	       sink_length;
+} WIRE_ATTR lnet_get_t;
+
+typedef struct lnet_reply {
+	lnet_handle_wire_t  dst_wmd;
+} WIRE_ATTR lnet_reply_t;
+
+typedef struct lnet_hello {
+	__u64	      incarnation;
+	__u32	      type;
+} WIRE_ATTR lnet_hello_t;
+
+typedef struct {
+	lnet_nid_t	  dest_nid;
+	lnet_nid_t	  src_nid;
+	lnet_pid_t	  dest_pid;
+	lnet_pid_t	  src_pid;
+	__u32	       type;	       /* lnet_msg_type_t */
+	__u32	       payload_length;     /* payload data to follow */
+	/*<------__u64 aligned------->*/
+	union {
+		lnet_ack_t   ack;
+		lnet_put_t   put;
+		lnet_get_t   get;
+		lnet_reply_t reply;
+		lnet_hello_t hello;
+	} msg;
+} WIRE_ATTR lnet_hdr_t;
+
+/* A HELLO message contains a magic number and protocol version
+ * code in the header's dest_nid, the peer's NID in the src_nid, and
+ * LNET_MSG_HELLO in the type field.  All other common fields are zero
+ * (including payload_size; i.e. no payload).
+ * This is for use by byte-stream LNDs (e.g. TCP/IP) to check the peer is
+ * running the same protocol and to find out its NID. These LNDs should
+ * exchange HELLO messages when a connection is first established.  Individual
+ * LNDs can put whatever else they fancy in lnet_hdr_t::msg.
+ */
+typedef struct {
+	__u32   magic;			  /* LNET_PROTO_TCP_MAGIC */
+	__u16   version_major;		  /* increment on incompatible change */
+	__u16   version_minor;		  /* increment on compatible change */
+} WIRE_ATTR lnet_magicversion_t;
+
+/* PROTO MAGIC for LNDs */
+#define LNET_PROTO_IB_MAGIC		 0x0be91b91
+#define LNET_PROTO_RA_MAGIC		 0x0be91b92
+#define LNET_PROTO_QSW_MAGIC		0x0be91b93
+#define LNET_PROTO_GNI_MAGIC		0xb00fbabe /* ask Kim */
+#define LNET_PROTO_TCP_MAGIC		0xeebc0ded
+#define LNET_PROTO_PTL_MAGIC		0x50746C4E /* 'PtlN' unique magic */
+#define LNET_PROTO_MX_MAGIC		 0x4d583130 /* 'MX10'! */
+#define LNET_PROTO_ACCEPTOR_MAGIC	   0xacce7100
+#define LNET_PROTO_PING_MAGIC	       0x70696E67 /* 'ping' */
+
+/* Placeholder for a future "unified" protocol across all LNDs */
+/* Current LNDs that receive a request with this magic will respond with a
+ * "stub" reply using their current protocol */
+#define LNET_PROTO_MAGIC		    0x45726963 /* ! */
+
+
+#define LNET_PROTO_TCP_VERSION_MAJOR	1
+#define LNET_PROTO_TCP_VERSION_MINOR	0
+
+/* Acceptor connection request */
+typedef struct {
+	__u32       acr_magic;		  /* PTL_ACCEPTOR_PROTO_MAGIC */
+	__u32       acr_version;		/* protocol version */
+	__u64       acr_nid;		    /* target NID */
+} WIRE_ATTR lnet_acceptor_connreq_t;
+
+#define LNET_PROTO_ACCEPTOR_VERSION       1
+
+/* forward refs */
+struct lnet_libmd;
+
+typedef struct lnet_msg {
+	struct list_head	    msg_activelist;
+	struct list_head	    msg_list;	   /* Q for credits/MD */
+
+	lnet_process_id_t     msg_target;
+	/* where is it from, it's only for building event */
+	lnet_nid_t		msg_from;
+	__u32			msg_type;
+
+	/* commited for sending */
+	unsigned int		msg_tx_committed:1;
+	/* CPT # this message committed for sending */
+	unsigned int		msg_tx_cpt:15;
+	/* commited for receiving */
+	unsigned int		msg_rx_committed:1;
+	/* CPT # this message committed for receiving */
+	unsigned int		msg_rx_cpt:15;
+	/* queued for tx credit */
+	unsigned int		msg_tx_delayed:1;
+	/* queued for RX buffer */
+	unsigned int		msg_rx_delayed:1;
+	/* ready for pending on RX delay list */
+	unsigned int		msg_rx_ready_delay:1;
+
+	unsigned int	  msg_vmflush:1;      /* VM trying to free memory */
+	unsigned int	  msg_target_is_router:1; /* sending to a router */
+	unsigned int	  msg_routing:1;      /* being forwarded */
+	unsigned int	  msg_ack:1;	  /* ack on finalize (PUT) */
+	unsigned int	  msg_sending:1;      /* outgoing message */
+	unsigned int	  msg_receiving:1;    /* being received */
+	unsigned int	  msg_txcredit:1;     /* taken an NI send credit */
+	unsigned int	  msg_peertxcredit:1; /* taken a peer send credit */
+	unsigned int	  msg_rtrcredit:1;    /* taken a globel router credit */
+	unsigned int	  msg_peerrtrcredit:1; /* taken a peer router credit */
+	unsigned int	  msg_onactivelist:1; /* on the activelist */
+
+	struct lnet_peer     *msg_txpeer;	 /* peer I'm sending to */
+	struct lnet_peer     *msg_rxpeer;	 /* peer I received from */
+
+	void		 *msg_private;
+	struct lnet_libmd    *msg_md;
+
+	unsigned int	  msg_len;
+	unsigned int	  msg_wanted;
+	unsigned int	  msg_offset;
+	unsigned int	  msg_niov;
+	struct iovec	 *msg_iov;
+	lnet_kiov_t	  *msg_kiov;
+
+	lnet_event_t	  msg_ev;
+	lnet_hdr_t	    msg_hdr;
+} lnet_msg_t;
+
+
+typedef struct lnet_libhandle {
+	struct list_head	    lh_hash_chain;
+	__u64		 lh_cookie;
+} lnet_libhandle_t;
+
+#define lh_entry(ptr, type, member) \
+	((type *)((char *)(ptr)-(char *)(&((type *)0)->member)))
+
+typedef struct lnet_eq {
+	struct list_head		eq_list;
+	lnet_libhandle_t	eq_lh;
+	lnet_seq_t		eq_enq_seq;
+	lnet_seq_t		eq_deq_seq;
+	unsigned int		eq_size;
+	lnet_eq_handler_t	eq_callback;
+	lnet_event_t		*eq_events;
+	int			**eq_refs;	/* percpt refcount for EQ */
+} lnet_eq_t;
+
+typedef struct lnet_me {
+	struct list_head	     me_list;
+	lnet_libhandle_t       me_lh;
+	lnet_process_id_t      me_match_id;
+	unsigned int	   me_portal;
+	unsigned int	   me_pos;		/* hash offset in mt_hash */
+	__u64		  me_match_bits;
+	__u64		  me_ignore_bits;
+	lnet_unlink_t	  me_unlink;
+	struct lnet_libmd     *me_md;
+} lnet_me_t;
+
+typedef struct lnet_libmd {
+	struct list_head	    md_list;
+	lnet_libhandle_t      md_lh;
+	lnet_me_t	    *md_me;
+	char		 *md_start;
+	unsigned int	  md_offset;
+	unsigned int	  md_length;
+	unsigned int	  md_max_size;
+	int		   md_threshold;
+	int		   md_refcount;
+	unsigned int	  md_options;
+	unsigned int	  md_flags;
+	void		 *md_user_ptr;
+	lnet_eq_t	    *md_eq;
+	unsigned int	  md_niov;		/* # frags */
+	union {
+		struct iovec  iov[LNET_MAX_IOV];
+		lnet_kiov_t   kiov[LNET_MAX_IOV];
+	} md_iov;
+} lnet_libmd_t;
+
+#define LNET_MD_FLAG_ZOMBIE	   (1 << 0)
+#define LNET_MD_FLAG_AUTO_UNLINK      (1 << 1)
+
+#ifdef LNET_USE_LIB_FREELIST
+typedef struct
+{
+	void		  *fl_objs;	  /* single contiguous array of objects */
+	int		    fl_nobjs;	 /* the number of them */
+	int		    fl_objsize;       /* the size (including overhead) of each of them */
+	struct list_head	     fl_list;	  /* where they are enqueued */
+} lnet_freelist_t;
+
+typedef struct
+{
+	struct list_head	     fo_list;	     /* enqueue on fl_list */
+	void		  *fo_contents;	 /* aligned contents */
+} lnet_freeobj_t;
+#endif
+
+typedef struct {
+	/* info about peers we are trying to fail */
+	struct list_head	     tp_list;	     /* ln_test_peers */
+	lnet_nid_t	     tp_nid;	      /* matching nid */
+	unsigned int	   tp_threshold;	/* # failures to simulate */
+} lnet_test_peer_t;
+
+#define LNET_COOKIE_TYPE_MD    1
+#define LNET_COOKIE_TYPE_ME    2
+#define LNET_COOKIE_TYPE_EQ    3
+#define LNET_COOKIE_TYPE_BITS  2
+#define LNET_COOKIE_MASK	((1ULL << LNET_COOKIE_TYPE_BITS) - 1ULL)
+
+struct lnet_ni;				  /* forward ref */
+
+typedef struct lnet_lnd
+{
+	/* fields managed by portals */
+	struct list_head	    lnd_list;	     /* stash in the LND table */
+	int		   lnd_refcount;	 /* # active instances */
+
+	/* fields initialised by the LND */
+	unsigned int	  lnd_type;
+
+	int  (*lnd_startup) (struct lnet_ni *ni);
+	void (*lnd_shutdown) (struct lnet_ni *ni);
+	int  (*lnd_ctl)(struct lnet_ni *ni, unsigned int cmd, void *arg);
+
+	/* In data movement APIs below, payload buffers are described as a set
+	 * of 'niov' fragments which are...
+	 * EITHER
+	 *    in virtual memory (struct iovec *iov != NULL)
+	 * OR
+	 *    in pages (kernel only: plt_kiov_t *kiov != NULL).
+	 * The LND may NOT overwrite these fragment descriptors.
+	 * An 'offset' and may specify a byte offset within the set of
+	 * fragments to start from
+	 */
+
+	/* Start sending a preformatted message.  'private' is NULL for PUT and
+	 * GET messages; otherwise this is a response to an incoming message
+	 * and 'private' is the 'private' passed to lnet_parse().  Return
+	 * non-zero for immediate failure, otherwise complete later with
+	 * lnet_finalize() */
+	int (*lnd_send)(struct lnet_ni *ni, void *private, lnet_msg_t *msg);
+
+	/* Start receiving 'mlen' bytes of payload data, skipping the following
+	 * 'rlen' - 'mlen' bytes. 'private' is the 'private' passed to
+	 * lnet_parse().  Return non-zero for immedaite failure, otherwise
+	 * complete later with lnet_finalize().  This also gives back a receive
+	 * credit if the LND does flow control. */
+	int (*lnd_recv)(struct lnet_ni *ni, void *private, lnet_msg_t *msg,
+			int delayed, unsigned int niov,
+			struct iovec *iov, lnet_kiov_t *kiov,
+			unsigned int offset, unsigned int mlen, unsigned int rlen);
+
+	/* lnet_parse() has had to delay processing of this message
+	 * (e.g. waiting for a forwarding buffer or send credits).  Give the
+	 * LND a chance to free urgently needed resources.  If called, return 0
+	 * for success and do NOT give back a receive credit; that has to wait
+	 * until lnd_recv() gets called.  On failure return < 0 and
+	 * release resources; lnd_recv() will not be called. */
+	int (*lnd_eager_recv)(struct lnet_ni *ni, void *private, lnet_msg_t *msg,
+			      void **new_privatep);
+
+	/* notification of peer health */
+	void (*lnd_notify)(struct lnet_ni *ni, lnet_nid_t peer, int alive);
+
+	/* query of peer aliveness */
+	void (*lnd_query)(struct lnet_ni *ni, lnet_nid_t peer, cfs_time_t *when);
+
+	/* accept a new connection */
+	int (*lnd_accept)(struct lnet_ni *ni, socket_t *sock);
+
+} lnd_t;
+
+#define LNET_NI_STATUS_UP      0x15aac0de
+#define LNET_NI_STATUS_DOWN    0xdeadface
+#define LNET_NI_STATUS_INVALID 0x00000000
+typedef struct {
+	lnet_nid_t ns_nid;
+	__u32      ns_status;
+	__u32      ns_unused;
+} WIRE_ATTR lnet_ni_status_t;
+
+struct lnet_tx_queue {
+	int			tq_credits;	/* # tx credits free */
+	int			tq_credits_min;	/* lowest it's been */
+	int			tq_credits_max;	/* total # tx credits */
+	struct list_head		tq_delayed;	/* delayed TXs */
+};
+
+#define LNET_MAX_INTERFACES   16
+
+typedef struct lnet_ni {
+	spinlock_t		ni_lock;
+	struct list_head		ni_list;	/* chain on ln_nis */
+	struct list_head		ni_cptlist;	/* chain on ln_nis_cpt */
+	int			ni_maxtxcredits; /* # tx credits  */
+	/* # per-peer send credits */
+	int			ni_peertxcredits;
+	/* # per-peer router buffer credits */
+	int			ni_peerrtrcredits;
+	/* seconds to consider peer dead */
+	int			ni_peertimeout;
+	int			ni_ncpts;	/* number of CPTs */
+	__u32			*ni_cpts;	/* bond NI on some CPTs */
+	lnet_nid_t		ni_nid;		/* interface's NID */
+	void			*ni_data;	/* instance-specific data */
+	lnd_t			*ni_lnd;	/* procedural interface */
+	struct lnet_tx_queue	**ni_tx_queues;	/* percpt TX queues */
+	int			**ni_refs;	/* percpt reference count */
+	long			ni_last_alive;	/* when I was last alive */
+	lnet_ni_status_t	*ni_status;	/* my health status */
+	/* equivalent interfaces to use */
+	char			*ni_interfaces[LNET_MAX_INTERFACES];
+} lnet_ni_t;
+
+#define LNET_PROTO_PING_MATCHBITS	0x8000000000000000LL
+
+/* NB: value of these features equal to LNET_PROTO_PING_VERSION_x
+ * of old LNet, so there shouldn't be any compatibility issue */
+#define LNET_PING_FEAT_INVAL		(0)		/* no feature */
+#define LNET_PING_FEAT_BASE		(1 << 0)	/* just a ping */
+#define LNET_PING_FEAT_NI_STATUS	(1 << 1)	/* return NI status */
+
+#define LNET_PING_FEAT_MASK		(LNET_PING_FEAT_BASE | \
+					 LNET_PING_FEAT_NI_STATUS)
+
+typedef struct {
+	__u32			pi_magic;
+	__u32			pi_features;
+	lnet_pid_t		pi_pid;
+	__u32			pi_nnis;
+	lnet_ni_status_t	pi_ni[0];
+} WIRE_ATTR lnet_ping_info_t;
+
+/* router checker data, per router */
+#define LNET_MAX_RTR_NIS   16
+#define LNET_PINGINFO_SIZE offsetof(lnet_ping_info_t, pi_ni[LNET_MAX_RTR_NIS])
+typedef struct {
+	/* chain on the_lnet.ln_zombie_rcd or ln_deathrow_rcd */
+	struct list_head		rcd_list;
+	lnet_handle_md_t	rcd_mdh;	/* ping buffer MD */
+	struct lnet_peer	*rcd_gateway;	/* reference to gateway */
+	lnet_ping_info_t	*rcd_pinginfo;	/* ping buffer */
+} lnet_rc_data_t;
+
+typedef struct lnet_peer {
+	struct list_head	lp_hashlist;	  /* chain on peer hash */
+	struct list_head	lp_txq;	       /* messages blocking for tx credits */
+	struct list_head	lp_rtrq;	      /* messages blocking for router credits */
+	struct list_head	lp_rtr_list;	  /* chain on router list */
+	int	       lp_txcredits;	 /* # tx credits available */
+	int	       lp_mintxcredits;      /* low water mark */
+	int	       lp_rtrcredits;	/* # router credits */
+	int	       lp_minrtrcredits;     /* low water mark */
+	unsigned int      lp_alive:1;	   /* alive/dead? */
+	unsigned int      lp_notify:1;	  /* notification outstanding? */
+	unsigned int      lp_notifylnd:1;       /* outstanding notification for LND? */
+	unsigned int      lp_notifying:1;       /* some thread is handling notification */
+	unsigned int      lp_ping_notsent;      /* SEND event outstanding from ping */
+	int	       lp_alive_count;       /* # times router went dead<->alive */
+	long	      lp_txqnob;	    /* bytes queued for sending */
+	cfs_time_t	lp_timestamp;	 /* time of last aliveness news */
+	cfs_time_t	lp_ping_timestamp;    /* time of last ping attempt */
+	cfs_time_t	lp_ping_deadline;     /* != 0 if ping reply expected */
+	cfs_time_t	lp_last_alive;	/* when I was last alive */
+	cfs_time_t	lp_last_query;	/* when lp_ni was queried last time */
+	lnet_ni_t	*lp_ni;		/* interface peer is on */
+	lnet_nid_t	lp_nid;	       /* peer's NID */
+	int	       lp_refcount;	  /* # refs */
+	int			lp_cpt;		/* CPT this peer attached on */
+	/* # refs from lnet_route_t::lr_gateway */
+	int			lp_rtr_refcount;
+	/* returned RC ping features */
+	unsigned int		lp_ping_feats;
+	struct list_head		lp_routes;	/* routers on this peer */
+	lnet_rc_data_t		*lp_rcd;	/* router checker state */
+} lnet_peer_t;
+
+
+/* peer hash size */
+#define LNET_PEER_HASH_BITS     9
+#define LNET_PEER_HASH_SIZE     (1 << LNET_PEER_HASH_BITS)
+
+/* peer hash table */
+struct lnet_peer_table {
+	int			pt_version;	/* /proc validity stamp */
+	int			pt_number;	/* # peers extant */
+	struct list_head		pt_deathrow;	/* zombie peers */
+	struct list_head		*pt_hash;	/* NID->peer hash */
+};
+
+/* peer aliveness is enabled only on routers for peers in a network where the
+ * lnet_ni_t::ni_peertimeout has been set to a positive value */
+#define lnet_peer_aliveness_enabled(lp) (the_lnet.ln_routing != 0 && \
+					 (lp)->lp_ni->ni_peertimeout > 0)
+
+typedef struct {
+	struct list_head		lr_list;	/* chain on net */
+	struct list_head		lr_gwlist;	/* chain on gateway */
+	lnet_peer_t		*lr_gateway;	/* router node */
+	__u32			lr_net;		/* remote network number */
+	int			lr_seq;		/* sequence for round-robin */
+	unsigned int		lr_downis;	/* number of down NIs */
+	unsigned int		lr_hops;	/* how far I am */
+} lnet_route_t;
+
+#define LNET_REMOTE_NETS_HASH_DEFAULT	(1U << 7)
+#define LNET_REMOTE_NETS_HASH_MAX	(1U << 16)
+#define LNET_REMOTE_NETS_HASH_SIZE	(1 << the_lnet.ln_remote_nets_hbits)
+
+typedef struct {
+	struct list_head	      lrn_list;       /* chain on ln_remote_nets_hash */
+	struct list_head	      lrn_routes;     /* routes to me */
+	__u32		   lrn_net;	/* my net number */
+} lnet_remotenet_t;
+
+typedef struct {
+	struct list_head rbp_bufs;	     /* my free buffer pool */
+	struct list_head rbp_msgs;	     /* messages blocking for a buffer */
+	int	rbp_npages;	   /* # pages in each buffer */
+	int	rbp_nbuffers;	 /* # buffers */
+	int	rbp_credits;	  /* # free buffers / blocked messages */
+	int	rbp_mincredits;       /* low water mark */
+} lnet_rtrbufpool_t;
+
+typedef struct {
+	struct list_head	     rb_list;	     /* chain on rbp_bufs */
+	lnet_rtrbufpool_t     *rb_pool;	     /* owning pool */
+	lnet_kiov_t	    rb_kiov[0];	  /* the buffer space */
+} lnet_rtrbuf_t;
+
+typedef struct {
+	__u32	msgs_alloc;
+	__u32	msgs_max;
+	__u32	errors;
+	__u32	send_count;
+	__u32	recv_count;
+	__u32	route_count;
+	__u32	drop_count;
+	__u64	send_length;
+	__u64	recv_length;
+	__u64	route_length;
+	__u64	drop_length;
+} WIRE_ATTR lnet_counters_t;
+
+#define LNET_PEER_HASHSIZE   503		/* prime! */
+
+#define LNET_NRBPOOLS	 3		 /* # different router buffer pools */
+
+enum {
+	/* Didn't match anything */
+	LNET_MATCHMD_NONE	= (1 << 0),
+	/* Matched OK */
+	LNET_MATCHMD_OK		= (1 << 1),
+	/* Must be discarded */
+	LNET_MATCHMD_DROP	= (1 << 2),
+	/* match and buffer is exhausted */
+	LNET_MATCHMD_EXHAUSTED  = (1 << 3),
+	/* match or drop */
+	LNET_MATCHMD_FINISH     = (LNET_MATCHMD_OK | LNET_MATCHMD_DROP),
+};
+
+/* Options for lnet_portal_t::ptl_options */
+#define LNET_PTL_LAZY	       (1 << 0)
+#define LNET_PTL_MATCH_UNIQUE       (1 << 1)    /* unique match, for RDMA */
+#define LNET_PTL_MATCH_WILDCARD     (1 << 2)    /* wildcard match, request portal */
+
+/* parameter for matching operations (GET, PUT) */
+struct lnet_match_info {
+	__u64			mi_mbits;
+	lnet_process_id_t	mi_id;
+	unsigned int		mi_opc;
+	unsigned int		mi_portal;
+	unsigned int		mi_rlength;
+	unsigned int		mi_roffset;
+};
+
+/* ME hash of RDMA portal */
+#define LNET_MT_HASH_BITS		8
+#define LNET_MT_HASH_SIZE		(1 << LNET_MT_HASH_BITS)
+#define LNET_MT_HASH_MASK		(LNET_MT_HASH_SIZE - 1)
+/* we allocate (LNET_MT_HASH_SIZE + 1) entries for lnet_match_table::mt_hash,
+ * the last entry is reserved for MEs with ignore-bits */
+#define LNET_MT_HASH_IGNORE		LNET_MT_HASH_SIZE
+/* __u64 has 2^6 bits, so need 2^(LNET_MT_HASH_BITS - LNET_MT_BITS_U64) which
+ * is 4 __u64s as bit-map, and add an extra __u64 (only use one bit) for the
+ * ME-list with ignore-bits, which is mtable::mt_hash[LNET_MT_HASH_IGNORE] */
+#define LNET_MT_BITS_U64		6	/* 2^6 bits */
+#define LNET_MT_EXHAUSTED_BITS		(LNET_MT_HASH_BITS - LNET_MT_BITS_U64)
+#define LNET_MT_EXHAUSTED_BMAP		((1 << LNET_MT_EXHAUSTED_BITS) + 1)
+
+/* portal match table */
+struct lnet_match_table {
+	/* reserved for upcoming patches, CPU partition ID */
+	unsigned int		mt_cpt;
+	unsigned int		mt_portal;      /* portal index */
+	/* match table is set as "enabled" if there's non-exhausted MD
+	 * attached on mt_mhash, it's only valide for wildcard portal */
+	unsigned int		mt_enabled;
+	/* bitmap to flag whether MEs on mt_hash are exhausted or not */
+	__u64			mt_exhausted[LNET_MT_EXHAUSTED_BMAP];
+	struct list_head		*mt_mhash;      /* matching hash */
+};
+
+/* these are only useful for wildcard portal */
+/* Turn off message rotor for wildcard portals */
+#define	LNET_PTL_ROTOR_OFF	0
+/* round-robin dispatch all PUT messages for wildcard portals */
+#define	LNET_PTL_ROTOR_ON	1
+/* round-robin dispatch routed PUT message for wildcard portals */
+#define	LNET_PTL_ROTOR_RR_RT	2
+/* dispatch routed PUT message by hashing source NID for wildcard portals */
+#define	LNET_PTL_ROTOR_HASH_RT	3
+
+typedef struct lnet_portal {
+	spinlock_t		ptl_lock;
+	unsigned int		ptl_index;	/* portal ID, reserved */
+	/* flags on this portal: lazy, unique... */
+	unsigned int		ptl_options;
+	/* list of messags which are stealing buffer */
+	struct list_head		ptl_msg_stealing;
+	/* messages blocking for MD */
+	struct list_head		ptl_msg_delayed;
+	/* Match table for each CPT */
+	struct lnet_match_table	**ptl_mtables;
+	/* spread rotor of incoming "PUT" */
+	int			ptl_rotor;
+	/* # active entries for this portal */
+	int		     ptl_mt_nmaps;
+	/* array of active entries' cpu-partition-id */
+	int		     ptl_mt_maps[0];
+} lnet_portal_t;
+
+#define LNET_LH_HASH_BITS	12
+#define LNET_LH_HASH_SIZE	(1ULL << LNET_LH_HASH_BITS)
+#define LNET_LH_HASH_MASK	(LNET_LH_HASH_SIZE - 1)
+
+/* resource container (ME, MD, EQ) */
+struct lnet_res_container {
+	unsigned int		rec_type;	/* container type */
+	__u64			rec_lh_cookie;	/* cookie generator */
+	struct list_head		rec_active;	/* active resource list */
+	struct list_head		*rec_lh_hash;	/* handle hash */
+#ifdef LNET_USE_LIB_FREELIST
+	lnet_freelist_t		rec_freelist;	/* freelist for resources */
+#endif
+};
+
+/* message container */
+struct lnet_msg_container {
+	int			msc_init;	/* initialized or not */
+	/* max # threads finalizing */
+	int			msc_nfinalizers;
+	/* msgs waiting to complete finalizing */
+	struct list_head		msc_finalizing;
+	struct list_head		msc_active;	/* active message list */
+	/* threads doing finalization */
+	void			**msc_finalizers;
+#ifdef LNET_USE_LIB_FREELIST
+	lnet_freelist_t		msc_freelist;	/* freelist for messages */
+#endif
+};
+
+/* Router Checker states */
+#define LNET_RC_STATE_SHUTDOWN		0	/* not started */
+#define LNET_RC_STATE_RUNNING		1	/* started up OK */
+#define LNET_RC_STATE_STOPPING		2	/* telling thread to stop */
+
+typedef struct
+{
+	/* CPU partition table of LNet */
+	struct cfs_cpt_table		*ln_cpt_table;
+	/* number of CPTs in ln_cpt_table */
+	unsigned int			ln_cpt_number;
+	unsigned int			ln_cpt_bits;
+
+	/* protect LNet resources (ME/MD/EQ) */
+	struct cfs_percpt_lock		*ln_res_lock;
+	/* # portals */
+	int				ln_nportals;
+	/* the vector of portals */
+	lnet_portal_t			**ln_portals;
+	/* percpt ME containers */
+	struct lnet_res_container	**ln_me_containers;
+	/* percpt MD container */
+	struct lnet_res_container	**ln_md_containers;
+
+	/* Event Queue container */
+	struct lnet_res_container	ln_eq_container;
+	wait_queue_head_t			ln_eq_waitq;
+	spinlock_t			ln_eq_wait_lock;
+	unsigned int			ln_remote_nets_hbits;
+
+	/* protect NI, peer table, credits, routers, rtrbuf... */
+	struct cfs_percpt_lock		*ln_net_lock;
+	/* percpt message containers for active/finalizing/freed message */
+	struct lnet_msg_container	**ln_msg_containers;
+	lnet_counters_t			**ln_counters;
+	struct lnet_peer_table		**ln_peer_tables;
+	/* failure simulation */
+	struct list_head			ln_test_peers;
+
+	struct list_head			ln_nis;		/* LND instances */
+	/* NIs bond on specific CPT(s) */
+	struct list_head			ln_nis_cpt;
+	/* dying LND instances */
+	struct list_head			ln_nis_zombie;
+	lnet_ni_t			*ln_loni;	/* the loopback NI */
+	/* NI to wait for events in */
+	lnet_ni_t			*ln_eq_waitni;
+
+	/* remote networks with routes to them */
+	struct list_head			*ln_remote_nets_hash;
+	/* validity stamp */
+	__u64				ln_remote_nets_version;
+	/* list of all known routers */
+	struct list_head			ln_routers;
+	/* validity stamp */
+	__u64				ln_routers_version;
+	/* percpt router buffer pools */
+	lnet_rtrbufpool_t		**ln_rtrpools;
+
+	lnet_handle_md_t		ln_ping_target_md;
+	lnet_handle_eq_t		ln_ping_target_eq;
+	lnet_ping_info_t		*ln_ping_info;
+
+	/* router checker startup/shutdown state */
+	int				ln_rc_state;
+	/* router checker's event queue */
+	lnet_handle_eq_t		ln_rc_eqh;
+	/* rcd still pending on net */
+	struct list_head			ln_rcd_deathrow;
+	/* rcd ready for free */
+	struct list_head			ln_rcd_zombie;
+	/* serialise startup/shutdown */
+	struct semaphore		ln_rc_signal;
+
+	struct mutex			ln_api_mutex;
+	struct mutex			ln_lnd_mutex;
+	int				ln_init;	/* LNetInit() called? */
+	/* Have I called LNetNIInit myself? */
+	int				ln_niinit_self;
+	/* LNetNIInit/LNetNIFini counter */
+	int				ln_refcount;
+	/* shutdown in progress */
+	int				ln_shutdown;
+
+	int				ln_routing;	/* am I a router? */
+	lnet_pid_t			ln_pid;		/* requested pid */
+	/* uniquely identifies this ni in this epoch */
+	__u64				ln_interface_cookie;
+	/* registered LNDs */
+	struct list_head			ln_lnds;
+
+	/* space for network names */
+	char				*ln_network_tokens;
+	int				ln_network_tokens_nob;
+	/* test protocol compatibility flags */
+	int				ln_testprotocompat;
+
+} lnet_t;
+
+#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/linux/api-support.h b/drivers/staging/lustre/include/linux/lnet/linux/api-support.h
new file mode 100644
index 000000000000..ca78a0a4e908
--- /dev/null
+++ b/drivers/staging/lustre/include/linux/lnet/linux/api-support.h
@@ -0,0 +1,43 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * 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 version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
+ *
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ */
+
+#ifndef __LINUX_API_SUPPORT_H__
+#define __LINUX_API_SUPPORT_H__
+
+#ifndef __LNET_API_SUPPORT_H__
+#error Do not #include this file directly. #include <lnet /api-support.h> instead
+#endif
+
+
+#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/linux/lib-lnet.h b/drivers/staging/lustre/include/linux/lnet/linux/lib-lnet.h
new file mode 100644
index 000000000000..d2c0a70f1f7e
--- /dev/null
+++ b/drivers/staging/lustre/include/linux/lnet/linux/lib-lnet.h
@@ -0,0 +1,72 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * 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 version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
+ *
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ */
+
+#ifndef __LNET_LINUX_LIB_LNET_H__
+#define __LNET_LINUX_LIB_LNET_H__
+
+#ifndef __LNET_LIB_LNET_H__
+#error Do not #include this file directly. #include <linux/lnet/lib-lnet.h> instead
+#endif
+
+# include <asm/page.h>
+# include <linux/string.h>
+# include <asm/io.h>
+# include <linux/libcfs/libcfs.h>
+
+static inline __u64
+lnet_page2phys (struct page *p)
+{
+	/* compiler optimizer will elide unused branches */
+
+	switch (sizeof(typeof(page_to_phys(p)))) {
+	case 4:
+		/* page_to_phys returns a 32 bit physical address.  This must
+		 * be a 32 bit machine with <= 4G memory and we must ensure we
+		 * don't sign extend when converting to 64 bits. */
+		return (unsigned long)page_to_phys(p);
+
+	case 8:
+		/* page_to_phys returns a 64 bit physical address :) */
+		return page_to_phys(p);
+
+	default:
+		LBUG();
+		return 0;
+	}
+}
+
+
+#define LNET_ROUTER
+
+#endif /* __LNET_LINUX_LIB_LNET_H__ */
diff --git a/drivers/staging/lustre/include/linux/lnet/linux/lib-types.h b/drivers/staging/lustre/include/linux/lnet/linux/lib-types.h
new file mode 100644
index 000000000000..669e8c038534
--- /dev/null
+++ b/drivers/staging/lustre/include/linux/lnet/linux/lib-types.h
@@ -0,0 +1,45 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * 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 version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
+ *
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ */
+
+#ifndef __LNET_LINUX_LIB_TYPES_H__
+#define __LNET_LINUX_LIB_TYPES_H__
+
+#ifndef __LNET_LIB_TYPES_H__
+#error Do not #include this file directly. #include <linux/lnet/lib-types.h> instead
+#endif
+
+# include <linux/uio.h>
+# include <linux/types.h>
+
+#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/linux/lnet.h b/drivers/staging/lustre/include/linux/lnet/linux/lnet.h
new file mode 100644
index 000000000000..1e888f1efc45
--- /dev/null
+++ b/drivers/staging/lustre/include/linux/lnet/linux/lnet.h
@@ -0,0 +1,56 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * 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 version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
+ *
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ *
+ * Copyright (c) 2011, Intel Corporation.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ */
+
+#ifndef __LNET_LINUX_LNET_H__
+#define __LNET_LINUX_LNET_H__
+
+#ifndef __LNET_H__
+#error Do not #include this file directly. #include <linux/lnet/lnet.h> instead
+#endif
+
+/*
+ * lnet.h
+ *
+ * User application interface file
+ */
+
+#include <linux/uio.h>
+#include <linux/types.h>
+
+#define cfs_tcp_sendpage(sk, page, offset, size, flags) \
+	tcp_sendpage(sk, page, offset, size, flags)
+
+#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/lnet-sysctl.h b/drivers/staging/lustre/include/linux/lnet/lnet-sysctl.h
new file mode 100644
index 000000000000..1bde44ebb911
--- /dev/null
+++ b/drivers/staging/lustre/include/linux/lnet/lnet-sysctl.h
@@ -0,0 +1,51 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * 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 version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
+ *
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2008, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ */
+
+#ifndef __LNET_SYSCTL_H__
+#define __LNET_SYSCTL_H__
+
+#if defined(CONFIG_SYSCTL)
+
+
+#define CTL_KRANAL      201
+#define CTL_O2IBLND     205
+#define CTL_PTLLND      206
+#define CTL_QSWNAL      207
+#define CTL_SOCKLND     208
+#define CTL_GNILND      210
+
+
+#endif
+
+#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/lnet.h b/drivers/staging/lustre/include/linux/lnet/lnet.h
new file mode 100644
index 000000000000..c532b15d7643
--- /dev/null
+++ b/drivers/staging/lustre/include/linux/lnet/lnet.h
@@ -0,0 +1,51 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * 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 version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
+ *
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ */
+
+#ifndef __LNET_H__
+#define __LNET_H__
+
+/*
+ * lnet.h
+ *
+ * User application interface file
+ */
+#include <linux/lnet/linux/lnet.h>
+
+#include <linux/lnet/types.h>
+#include <linux/lnet/api.h>
+
+#define LNET_NIDSTR_COUNT  1024    /* # of nidstrings */
+#define LNET_NIDSTR_SIZE   32      /* size of each one (see below for usage) */
+
+#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/lnetctl.h b/drivers/staging/lustre/include/linux/lnet/lnetctl.h
new file mode 100644
index 000000000000..b22daa234255
--- /dev/null
+++ b/drivers/staging/lustre/include/linux/lnet/lnetctl.h
@@ -0,0 +1,80 @@
+/*
+ *   This file is part of Portals, http://www.sf.net/projects/lustre/
+ *
+ *   Portals is free software; you can redistribute it and/or
+ *   modify it under the terms of version 2 of the GNU General Public
+ *   License as published by the Free Software Foundation.
+ *
+ *   Portals 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 Portals; if not, write to the Free Software
+ *   Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
+ *
+ * header for libptlctl.a
+ */
+#ifndef _PTLCTL_H_
+#define _PTLCTL_H_
+
+#include <linux/libcfs/libcfs.h>
+#include <linux/lnet/types.h>
+
+#define LNET_DEV_ID 0
+#define LNET_DEV_PATH "/dev/lnet"
+#define LNET_DEV_MAJOR 10
+#define LNET_DEV_MINOR 240
+#define OBD_DEV_ID 1
+#define OBD_DEV_NAME "obd"
+#define OBD_DEV_PATH "/dev/" OBD_DEV_NAME
+#define OBD_DEV_MAJOR 10
+#define OBD_DEV_MINOR 241
+#define SMFS_DEV_ID  2
+#define SMFS_DEV_PATH "/dev/snapdev"
+#define SMFS_DEV_MAJOR 10
+#define SMFS_DEV_MINOR 242
+
+int ptl_initialize(int argc, char **argv);
+int jt_ptl_network(int argc, char **argv);
+int jt_ptl_list_nids(int argc, char **argv);
+int jt_ptl_which_nid(int argc, char **argv);
+int jt_ptl_print_interfaces(int argc, char **argv);
+int jt_ptl_add_interface(int argc, char **argv);
+int jt_ptl_del_interface(int argc, char **argv);
+int jt_ptl_print_peers (int argc, char **argv);
+int jt_ptl_add_peer (int argc, char **argv);
+int jt_ptl_del_peer (int argc, char **argv);
+int jt_ptl_print_connections (int argc, char **argv);
+int jt_ptl_disconnect(int argc, char **argv);
+int jt_ptl_push_connection(int argc, char **argv);
+int jt_ptl_print_active_txs(int argc, char **argv);
+int jt_ptl_ping(int argc, char **argv);
+int jt_ptl_mynid(int argc, char **argv);
+int jt_ptl_add_uuid(int argc, char **argv);
+int jt_ptl_add_uuid_old(int argc, char **argv); /* backwards compatibility  */
+int jt_ptl_close_uuid(int argc, char **argv);
+int jt_ptl_del_uuid(int argc, char **argv);
+int jt_ptl_add_route (int argc, char **argv);
+int jt_ptl_del_route (int argc, char **argv);
+int jt_ptl_notify_router (int argc, char **argv);
+int jt_ptl_print_routes (int argc, char **argv);
+int jt_ptl_fail_nid (int argc, char **argv);
+int jt_ptl_lwt(int argc, char **argv);
+int jt_ptl_testprotocompat(int argc, char **argv);
+int jt_ptl_memhog(int argc, char **argv);
+
+int dbg_initialize(int argc, char **argv);
+int jt_dbg_filter(int argc, char **argv);
+int jt_dbg_show(int argc, char **argv);
+int jt_dbg_list(int argc, char **argv);
+int jt_dbg_debug_kernel(int argc, char **argv);
+int jt_dbg_debug_daemon(int argc, char **argv);
+int jt_dbg_debug_file(int argc, char **argv);
+int jt_dbg_clear_debug_buf(int argc, char **argv);
+int jt_dbg_mark_debug_buf(int argc, char **argv);
+int jt_dbg_modules(int argc, char **argv);
+int jt_dbg_panic(int argc, char **argv);
+
+#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/lnetst.h b/drivers/staging/lustre/include/linux/lnet/lnetst.h
new file mode 100644
index 000000000000..d90f94e94601
--- /dev/null
+++ b/drivers/staging/lustre/include/linux/lnet/lnetst.h
@@ -0,0 +1,491 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * 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 version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
+ *
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ *
+ * Copyright (c) 2011, Intel Corporation.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ *
+ * lnet/include/lnet/lnetst.h
+ *
+ * Author: Liang Zhen <liangzhen@clusterfs.com>
+ */
+
+#ifndef __LNET_ST_H__
+#define __LNET_ST_H__
+
+#include <linux/libcfs/libcfs.h>
+#include <linux/lnet/lnet.h>
+#include <linux/lnet/lib-types.h>
+
+#define LST_FEAT_NONE		(0)
+#define LST_FEAT_BULK_LEN	(1 << 0)	/* enable variable page size */
+
+#define LST_FEATS_EMPTY		(LST_FEAT_NONE)
+#define LST_FEATS_MASK		(LST_FEAT_NONE | LST_FEAT_BULK_LEN)
+
+#define LST_NAME_SIZE	   32	      /* max name buffer length */
+
+#define LSTIO_DEBUG	     0xC00	   /* debug */
+#define LSTIO_SESSION_NEW       0xC01	   /* create session */
+#define LSTIO_SESSION_END       0xC02	   /* end session */
+#define LSTIO_SESSION_INFO      0xC03	   /* query session */
+#define LSTIO_GROUP_ADD	 0xC10	   /* add group */
+#define LSTIO_GROUP_LIST	0xC11	   /* list all groups in session */
+#define LSTIO_GROUP_INFO	0xC12	   /* query defailt infomation of specified group */
+#define LSTIO_GROUP_DEL	 0xC13	   /* delete group */
+#define LSTIO_NODES_ADD	 0xC14	   /* add nodes to specified group */
+#define LSTIO_GROUP_UPDATE      0xC15	   /* update group */
+#define LSTIO_BATCH_ADD	 0xC20	   /* add batch */
+#define LSTIO_BATCH_START       0xC21	   /* start batch */
+#define LSTIO_BATCH_STOP	0xC22	   /* stop batch */
+#define LSTIO_BATCH_DEL	 0xC23	   /* delete batch */
+#define LSTIO_BATCH_LIST	0xC24	   /* show all batches in the session */
+#define LSTIO_BATCH_INFO	0xC25	   /* show defail of specified batch */
+#define LSTIO_TEST_ADD	  0xC26	   /* add test (to batch) */
+#define LSTIO_BATCH_QUERY       0xC27	   /* query batch status */
+#define LSTIO_STAT_QUERY	0xC30	   /* get stats */
+
+typedef struct {
+	lnet_nid_t	      ses_nid;		/* nid of console node */
+	__u64		   ses_stamp;	      /* time stamp */
+} lst_sid_t;					    /*** session id */
+
+extern lst_sid_t LST_INVALID_SID;
+
+typedef struct {
+	__u64		   bat_id;		 /* unique id in session */
+} lst_bid_t;					    /*** batch id (group of tests) */
+
+/* Status of test node */
+#define LST_NODE_ACTIVE	 0x1		     /* node in this session */
+#define LST_NODE_BUSY	   0x2		     /* node is taken by other session */
+#define LST_NODE_DOWN	   0x4		     /* node is down */
+#define LST_NODE_UNKNOWN	0x8		     /* node not in session */
+
+typedef struct {
+	lnet_process_id_t       nde_id;		 /* id of node */
+	int		     nde_state;	      /* state of node */
+} lstcon_node_ent_t;				    /*** node entry, for list_group command */
+
+typedef struct {
+	int		     nle_nnode;	      /* # of nodes */
+	int		     nle_nactive;	    /* # of active nodes */
+	int		     nle_nbusy;	      /* # of busy nodes */
+	int		     nle_ndown;	      /* # of down nodes */
+	int		     nle_nunknown;	   /* # of unknown nodes */
+} lstcon_ndlist_ent_t;				  /*** node_list entry, for list_batch command */
+
+typedef struct {
+	int		     tse_type;	       /* test type */
+	int		     tse_loop;	       /* loop count */
+	int		     tse_concur;	     /* concurrency of test */
+} lstcon_test_ent_t;				    /*** test summary entry, for list_batch command */
+
+typedef struct {
+	int		     bae_state;	      /* batch status */
+	int		     bae_timeout;	    /* batch timeout */
+	int		     bae_ntest;	      /* # of tests in the batch */
+} lstcon_batch_ent_t;				   /*** batch summary entry, for list_batch command */
+
+typedef struct {
+	lstcon_ndlist_ent_t     tbe_cli_nle;	    /* client (group) node_list entry */
+	lstcon_ndlist_ent_t     tbe_srv_nle;	    /* server (group) node_list entry */
+	union {
+		lstcon_test_ent_t  tbe_test;	    /* test entry */
+		lstcon_batch_ent_t tbe_batch;	   /* batch entry */
+	} u;
+} lstcon_test_batch_ent_t;			      /*** test/batch verbose information entry,
+							 *** for list_batch command */
+
+typedef struct {
+	struct list_head	      rpe_link;	       /* link chain */
+	lnet_process_id_t       rpe_peer;	       /* peer's id */
+	struct timeval	  rpe_stamp;	      /* time stamp of RPC */
+	int		     rpe_state;	      /* peer's state */
+	int		     rpe_rpc_errno;	  /* RPC errno */
+
+	lst_sid_t	       rpe_sid;		/* peer's session id */
+	int		     rpe_fwk_errno;	  /* framework errno */
+	int		     rpe_priv[4];	    /* private data */
+	char		    rpe_payload[0];	 /* private reply payload */
+} lstcon_rpc_ent_t;
+
+typedef struct {
+	int		     trs_rpc_stat[4];	/* RPCs stat (0: total, 1: failed, 2: finished, 4: reserved */
+	int		     trs_rpc_errno;	  /* RPC errno */
+	int		     trs_fwk_stat[8];	/* framework stat */
+	int		     trs_fwk_errno;	  /* errno of the first remote error */
+	void		   *trs_fwk_private;	/* private framework stat */
+} lstcon_trans_stat_t;
+
+static inline int
+lstcon_rpc_stat_total(lstcon_trans_stat_t *stat, int inc)
+{
+	return inc ? ++stat->trs_rpc_stat[0] : stat->trs_rpc_stat[0];
+}
+
+static inline int
+lstcon_rpc_stat_success(lstcon_trans_stat_t *stat, int inc)
+{
+	return inc ? ++stat->trs_rpc_stat[1] : stat->trs_rpc_stat[1];
+}
+
+static inline int
+lstcon_rpc_stat_failure(lstcon_trans_stat_t *stat, int inc)
+{
+	return inc ? ++stat->trs_rpc_stat[2] : stat->trs_rpc_stat[2];
+}
+
+static inline int
+lstcon_sesop_stat_success(lstcon_trans_stat_t *stat, int inc)
+{
+	return inc ? ++stat->trs_fwk_stat[0] : stat->trs_fwk_stat[0];
+}
+
+static inline int
+lstcon_sesop_stat_failure(lstcon_trans_stat_t *stat, int inc)
+{
+	return inc ? ++stat->trs_fwk_stat[1] : stat->trs_fwk_stat[1];
+}
+
+static inline int
+lstcon_sesqry_stat_active(lstcon_trans_stat_t *stat, int inc)
+{
+	return inc ? ++stat->trs_fwk_stat[0] : stat->trs_fwk_stat[0];
+}
+
+static inline int
+lstcon_sesqry_stat_busy(lstcon_trans_stat_t *stat, int inc)
+{
+	return inc ? ++stat->trs_fwk_stat[1] : stat->trs_fwk_stat[1];
+}
+
+static inline int
+lstcon_sesqry_stat_unknown(lstcon_trans_stat_t *stat, int inc)
+{
+	return inc ? ++stat->trs_fwk_stat[2] : stat->trs_fwk_stat[2];
+}
+
+static inline int
+lstcon_tsbop_stat_success(lstcon_trans_stat_t *stat, int inc)
+{
+	return inc ? ++stat->trs_fwk_stat[0] : stat->trs_fwk_stat[0];
+}
+
+static inline int
+lstcon_tsbop_stat_failure(lstcon_trans_stat_t *stat, int inc)
+{
+	return inc ? ++stat->trs_fwk_stat[1] : stat->trs_fwk_stat[1];
+}
+
+static inline int
+lstcon_tsbqry_stat_idle(lstcon_trans_stat_t *stat, int inc)
+{
+	return inc ? ++stat->trs_fwk_stat[0] : stat->trs_fwk_stat[0];
+}
+
+static inline int
+lstcon_tsbqry_stat_run(lstcon_trans_stat_t *stat, int inc)
+{
+	return inc ? ++stat->trs_fwk_stat[1] : stat->trs_fwk_stat[1];
+}
+
+static inline int
+lstcon_tsbqry_stat_failure(lstcon_trans_stat_t *stat, int inc)
+{
+	return inc ? ++stat->trs_fwk_stat[2] : stat->trs_fwk_stat[2];
+}
+
+static inline int
+lstcon_statqry_stat_success(lstcon_trans_stat_t *stat, int inc)
+{
+	return inc ? ++stat->trs_fwk_stat[0] : stat->trs_fwk_stat[0];
+}
+
+static inline int
+lstcon_statqry_stat_failure(lstcon_trans_stat_t *stat, int inc)
+{
+	return inc ? ++stat->trs_fwk_stat[1] : stat->trs_fwk_stat[1];
+}
+
+/* create a session */
+typedef struct {
+	int		     lstio_ses_key;	  /* IN: local key */
+	int		     lstio_ses_timeout;      /* IN: session timeout */
+	int		     lstio_ses_force;	/* IN: force create ? */
+	/** IN: session features */
+	unsigned		lstio_ses_feats;
+	lst_sid_t	      *lstio_ses_idp;	  /* OUT: session id */
+	int		     lstio_ses_nmlen;	/* IN: name length */
+	char		   *lstio_ses_namep;	/* IN: session name */
+} lstio_session_new_args_t;
+
+/* query current session */
+typedef struct {
+	lst_sid_t	      *lstio_ses_idp;	  /* OUT: session id */
+	int		    *lstio_ses_keyp;	 /* OUT: local key */
+	/** OUT: session features */
+	unsigned	       *lstio_ses_featp;
+	lstcon_ndlist_ent_t    *lstio_ses_ndinfo;       /* OUT: */
+	int		     lstio_ses_nmlen;	/* IN: name length */
+	char		   *lstio_ses_namep;	/* OUT: session name */
+} lstio_session_info_args_t;
+
+/* delete a session */
+typedef struct {
+	int		     lstio_ses_key;	  /* IN: session key */
+} lstio_session_end_args_t;
+
+#define LST_OPC_SESSION	 1
+#define LST_OPC_GROUP	   2
+#define LST_OPC_NODES	   3
+#define LST_OPC_BATCHCLI	4
+#define LST_OPC_BATCHSRV	5
+
+typedef struct {
+	int		     lstio_dbg_key;	  /* IN: session key */
+	int		     lstio_dbg_type;	 /* IN: debug sessin|batch|group|nodes list */
+	int		     lstio_dbg_flags;	/* IN: reserved debug flags */
+	int		     lstio_dbg_timeout;      /* IN: timeout of debug */
+
+	int		     lstio_dbg_nmlen;	/* IN: len of name */
+	char		   *lstio_dbg_namep;	/* IN: name of group|batch */
+	int		     lstio_dbg_count;	/* IN: # of test nodes to debug */
+	lnet_process_id_t      *lstio_dbg_idsp;	 /* IN: id of test nodes */
+	struct list_head	     *lstio_dbg_resultp;      /* OUT: list head of result buffer */
+} lstio_debug_args_t;
+
+typedef struct {
+	int		     lstio_grp_key;	  /* IN: session key */
+	int		     lstio_grp_nmlen;	/* IN: name length */
+	char		   *lstio_grp_namep;	/* IN: group name */
+} lstio_group_add_args_t;
+
+typedef struct {
+	int		     lstio_grp_key;	  /* IN: session key */
+	int		     lstio_grp_nmlen;	/* IN: name length */
+	char		   *lstio_grp_namep;	/* IN: group name */
+} lstio_group_del_args_t;
+
+#define LST_GROUP_CLEAN	 1		       /* remove inactive nodes in the group */
+#define LST_GROUP_REFRESH       2		       /* refresh inactive nodes in the group */
+#define LST_GROUP_RMND	  3		       /* delete nodes from the group */
+
+typedef struct {
+	int		     lstio_grp_key;	  /* IN: session key */
+	int		     lstio_grp_opc;	  /* IN: OPC */
+	int		     lstio_grp_args;	 /* IN: arguments */
+	int		     lstio_grp_nmlen;	/* IN: name length */
+	char		   *lstio_grp_namep;	/* IN: group name */
+	int		     lstio_grp_count;	/* IN: # of nodes id */
+	lnet_process_id_t      *lstio_grp_idsp;	 /* IN: array of nodes */
+	struct list_head	     *lstio_grp_resultp;      /* OUT: list head of result buffer */
+} lstio_group_update_args_t;
+
+typedef struct {
+	int		     lstio_grp_key;	  /* IN: session key */
+	int		     lstio_grp_nmlen;	/* IN: name length */
+	char		   *lstio_grp_namep;	/* IN: group name */
+	int		     lstio_grp_count;	/* IN: # of nodes */
+	/** OUT: session features */
+	unsigned	       *lstio_grp_featp;
+	lnet_process_id_t      *lstio_grp_idsp;	 /* IN: nodes */
+	struct list_head	     *lstio_grp_resultp;      /* OUT: list head of result buffer */
+} lstio_group_nodes_args_t;
+
+typedef struct {
+	int		     lstio_grp_key;	  /* IN: session key */
+	int		     lstio_grp_idx;	  /* IN: group idx */
+	int		     lstio_grp_nmlen;	/* IN: name len */
+	char		   *lstio_grp_namep;	/* OUT: name */
+} lstio_group_list_args_t;
+
+typedef struct {
+	int		     lstio_grp_key;	  /* IN: session key */
+	int		     lstio_grp_nmlen;	/* IN: name len */
+	char		   *lstio_grp_namep;	/* IN: name */
+	lstcon_ndlist_ent_t    *lstio_grp_entp;	 /* OUT: description of group */
+
+	int		    *lstio_grp_idxp;	 /* IN/OUT: node index */
+	int		    *lstio_grp_ndentp;       /* IN/OUT: # of nodent */
+	lstcon_node_ent_t      *lstio_grp_dentsp;       /* OUT: nodent array */
+} lstio_group_info_args_t;
+
+#define LST_DEFAULT_BATCH       "batch"		 /* default batch name */
+
+typedef struct {
+	int		     lstio_bat_key;	  /* IN: session key */
+	int		     lstio_bat_nmlen;	/* IN: name length */
+	char		   *lstio_bat_namep;	/* IN: batch name */
+} lstio_batch_add_args_t;
+
+typedef struct {
+	int		     lstio_bat_key;	  /* IN: session key */
+	int		     lstio_bat_nmlen;	/* IN: name length */
+	char		   *lstio_bat_namep;	/* IN: batch name */
+} lstio_batch_del_args_t;
+
+typedef struct {
+	int		     lstio_bat_key;	  /* IN: session key */
+	int		     lstio_bat_timeout;      /* IN: timeout for the batch */
+	int		     lstio_bat_nmlen;	/* IN: name length */
+	char		   *lstio_bat_namep;	/* IN: batch name */
+	struct list_head	     *lstio_bat_resultp;      /* OUT: list head of result buffer */
+} lstio_batch_run_args_t;
+
+typedef struct {
+	int		     lstio_bat_key;	  /* IN: session key */
+	int		     lstio_bat_force;	/* IN: abort unfinished test RPC */
+	int		     lstio_bat_nmlen;	/* IN: name length */
+	char		   *lstio_bat_namep;	/* IN: batch name */
+	struct list_head	     *lstio_bat_resultp;      /* OUT: list head of result buffer */
+} lstio_batch_stop_args_t;
+
+typedef struct {
+	int		     lstio_bat_key;	  /* IN: session key */
+	int		     lstio_bat_testidx;      /* IN: test index */
+	int		     lstio_bat_client;       /* IN: is test client? */
+	int		     lstio_bat_timeout;      /* IN: timeout for waiting */
+	int		     lstio_bat_nmlen;	/* IN: name length */
+	char		   *lstio_bat_namep;	/* IN: batch name */
+	struct list_head	     *lstio_bat_resultp;      /* OUT: list head of result buffer */
+} lstio_batch_query_args_t;
+
+typedef struct {
+	int		     lstio_bat_key;	  /* IN: session key */
+	int		     lstio_bat_idx;	  /* IN: index */
+	int		     lstio_bat_nmlen;	/* IN: name length */
+	char		   *lstio_bat_namep;	/* IN: batch name */
+} lstio_batch_list_args_t;
+
+typedef struct {
+	int		     lstio_bat_key;	  /* IN: session key */
+	int		     lstio_bat_nmlen;	/* IN: name length */
+	char		   *lstio_bat_namep;	/* IN: name */
+	int		     lstio_bat_server;       /* IN: query server or not */
+	int		     lstio_bat_testidx;      /* IN: test index */
+	lstcon_test_batch_ent_t *lstio_bat_entp;	/* OUT: batch ent */
+
+	int		    *lstio_bat_idxp;	 /* IN/OUT: index of node */
+	int		    *lstio_bat_ndentp;       /* IN/OUT: # of nodent */
+	lstcon_node_ent_t      *lstio_bat_dentsp;       /* array of nodent */
+} lstio_batch_info_args_t;
+
+/* add stat in session */
+typedef struct {
+	int		     lstio_sta_key;	  /* IN: session key */
+	int		     lstio_sta_timeout;      /* IN: timeout for stat requst */
+	int		     lstio_sta_nmlen;	/* IN: group name length */
+	char		   *lstio_sta_namep;	/* IN: group name */
+	int		     lstio_sta_count;	/* IN: # of pid */
+	lnet_process_id_t      *lstio_sta_idsp;	 /* IN: pid */
+	struct list_head	     *lstio_sta_resultp;      /* OUT: list head of result buffer */
+} lstio_stat_args_t;
+
+typedef enum {
+	LST_TEST_BULK   = 1,
+	LST_TEST_PING   = 2
+} lst_test_type_t;
+
+/* create a test in a batch */
+#define LST_MAX_CONCUR	  1024		    /* Max concurrency of test */
+
+typedef struct {
+	int		     lstio_tes_key;	  /* IN: session key */
+	int		     lstio_tes_bat_nmlen;    /* IN: batch name len */
+	char		   *lstio_tes_bat_name;     /* IN: batch name */
+	int		     lstio_tes_type;	 /* IN: test type */
+	int		     lstio_tes_oneside;      /* IN: one sided test */
+	int		     lstio_tes_loop;	 /* IN: loop count */
+	int		     lstio_tes_concur;       /* IN: concurrency */
+
+	int		     lstio_tes_dist;	 /* IN: node distribution in destination groups */
+	int		     lstio_tes_span;	 /* IN: node span in destination groups */
+	int		     lstio_tes_sgrp_nmlen;   /* IN: source group name length */
+	char		   *lstio_tes_sgrp_name;    /* IN: group name */
+	int		     lstio_tes_dgrp_nmlen;   /* IN: destination group name length */
+	char		   *lstio_tes_dgrp_name;    /* IN: group name */
+
+	int		     lstio_tes_param_len;    /* IN: param buffer len */
+	void		   *lstio_tes_param;	/* IN: parameter for specified test:
+							       lstio_bulk_param_t,
+							       lstio_ping_param_t,
+							       ... more */
+	int		    *lstio_tes_retp;	 /* OUT: private returned value */
+	struct list_head	     *lstio_tes_resultp;      /* OUT: list head of result buffer */
+} lstio_test_args_t;
+
+typedef enum {
+	LST_BRW_READ    = 1,
+	LST_BRW_WRITE   = 2
+} lst_brw_type_t;
+
+typedef enum {
+	LST_BRW_CHECK_NONE   = 1,
+	LST_BRW_CHECK_SIMPLE = 2,
+	LST_BRW_CHECK_FULL   = 3
+} lst_brw_flags_t;
+
+typedef struct {
+	int		     blk_opc;		/* bulk operation code */
+	int		     blk_size;	       /* size (bytes) */
+	int		     blk_time;	       /* time of running the test*/
+	int		     blk_flags;	      /* reserved flags */
+} lst_test_bulk_param_t;
+
+typedef struct {
+	int		     png_size;	       /* size of ping message */
+	int		     png_time;	       /* time */
+	int		     png_loop;	       /* loop */
+	int		     png_flags;	      /* reserved flags */
+} lst_test_ping_param_t;
+
+/* more tests */
+typedef struct {
+	__u32 errors;
+	__u32 rpcs_sent;
+	__u32 rpcs_rcvd;
+	__u32 rpcs_dropped;
+	__u32 rpcs_expired;
+	__u64 bulk_get;
+	__u64 bulk_put;
+} WIRE_ATTR srpc_counters_t;
+
+typedef struct {
+	/** milliseconds since current session started */
+	__u32 running_ms;
+	__u32 active_batches;
+	__u32 zombie_sessions;
+	__u32 brw_errors;
+	__u32 ping_errors;
+} WIRE_ATTR sfw_counters_t;
+
+#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/ptllnd.h b/drivers/staging/lustre/include/linux/lnet/ptllnd.h
new file mode 100644
index 000000000000..fc1ce8ed1f8b
--- /dev/null
+++ b/drivers/staging/lustre/include/linux/lnet/ptllnd.h
@@ -0,0 +1,94 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * 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 version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
+ *
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ *
+ * lnet/include/lnet/ptllnd.h
+ *
+ * Author: PJ Kirner <pjkirner@clusterfs.com>
+ */
+
+/*
+ * The PTLLND was designed to support Portals with
+ * Lustre and non-lustre UNLINK semantics.
+ * However for now the two targets are Cray Portals
+ * on the XT3 and Lustre Portals (for testing) both
+ * have Lustre UNLINK semantics, so this is defined
+ * by default.
+ */
+#define LUSTRE_PORTALS_UNLINK_SEMANTICS
+
+
+#ifdef _USING_LUSTRE_PORTALS_
+
+/* NIDs are 64-bits on Lustre Portals */
+#define FMT_NID LPU64
+#define FMT_PID "%d"
+
+/* When using Lustre Portals Lustre completion semantics are imlicit*/
+#define PTL_MD_LUSTRE_COMPLETION_SEMANTICS      0
+
+#else /* _USING_CRAY_PORTALS_ */
+
+/* NIDs are integers on Cray Portals */
+#define FMT_NID "%u"
+#define FMT_PID "%d"
+
+/* When using Cray Portals this is defined in the Cray Portals Header*/
+/*#define PTL_MD_LUSTRE_COMPLETION_SEMANTICS */
+
+/* Can compare handles directly on Cray Portals */
+#define PtlHandleIsEqual(a,b) ((a) == (b))
+
+/* Diffrent error types on Cray Portals*/
+#define ptl_err_t ptl_ni_fail_t
+
+/*
+ * The Cray Portals has no maximum number of IOVs.  The
+ * maximum is limited only by memory and size of the
+ * int parameters (2^31-1).
+ * Lustre only really require that the underyling
+ * implemenation to support at least LNET_MAX_IOV,
+ * so for Cray portals we can safely just use that
+ * value here.
+ *
+ */
+#define PTL_MD_MAX_IOV	  LNET_MAX_IOV
+
+#endif
+
+#define FMT_PTLID "ptlid:"FMT_PID"-"FMT_NID
+
+/* Align incoming small request messages to an 8 byte boundary if this is
+ * supported to avoid alignment issues on some architectures */
+#ifndef PTL_MD_LOCAL_ALIGN8
+# define PTL_MD_LOCAL_ALIGN8 0
+#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/ptllnd_wire.h b/drivers/staging/lustre/include/linux/lnet/ptllnd_wire.h
new file mode 100644
index 000000000000..7d12b3a23a96
--- /dev/null
+++ b/drivers/staging/lustre/include/linux/lnet/ptllnd_wire.h
@@ -0,0 +1,124 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * 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 version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
+ *
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ *
+ * lnet/include/lnet/ptllnd_wire.h
+ *
+ * Author: PJ Kirner <pjkirner@clusterfs.com>
+ */
+
+/* Minimum buffer size that any peer will post to receive ptllnd messages */
+#define PTLLND_MIN_BUFFER_SIZE  256
+
+/************************************************************************
+ * Tunable defaults that {u,k}lnds/ptllnd should have in common.
+ */
+
+#define PTLLND_PORTAL	   9	  /* The same portal PTLPRC used when talking to cray portals */
+#define PTLLND_PID	      9	  /* The Portals PID */
+#define PTLLND_PEERCREDITS      8	  /* concurrent sends to 1 peer */
+
+/* Default buffer size for kernel ptllnds (guaranteed eager) */
+#define PTLLND_MAX_KLND_MSG_SIZE 512
+
+/* Default buffer size for catamount ptllnds (not guaranteed eager) - large
+ * enough to avoid RDMA for anything sent while control is not in liblustre */
+#define PTLLND_MAX_ULND_MSG_SIZE 512
+
+
+/************************************************************************
+ * Portals LND Wire message format.
+ * These are sent in sender's byte order (i.e. receiver flips).
+ */
+
+#define PTL_RESERVED_MATCHBITS  0x100	/* below this value is reserved
+					 * above is for bulk data transfer */
+#define LNET_MSG_MATCHBITS       0      /* the value for the message channel */
+
+typedef struct
+{
+	lnet_hdr_t	kptlim_hdr;	     /* portals header */
+	char	      kptlim_payload[0];      /* piggy-backed payload */
+} WIRE_ATTR kptl_immediate_msg_t;
+
+typedef struct
+{
+	lnet_hdr_t	kptlrm_hdr;	     /* portals header */
+	__u64	     kptlrm_matchbits;       /* matchbits */
+} WIRE_ATTR kptl_rdma_msg_t;
+
+typedef struct
+{
+	__u64	     kptlhm_matchbits;       /* matchbits */
+	__u32	     kptlhm_max_msg_size;    /* max message size */
+} WIRE_ATTR kptl_hello_msg_t;
+
+typedef struct
+{
+	/* First 2 fields fixed FOR ALL TIME */
+	__u32	   ptlm_magic;     /* I'm a Portals LND message */
+	__u16	   ptlm_version;   /* this is my version number */
+	__u8	    ptlm_type;      /* the message type */
+	__u8	    ptlm_credits;   /* returned credits */
+	__u32	   ptlm_nob;       /* # bytes in whole message */
+	__u32	   ptlm_cksum;     /* checksum (0 == no checksum) */
+	__u64	   ptlm_srcnid;    /* sender's NID */
+	__u64	   ptlm_srcstamp;  /* sender's incarnation */
+	__u64	   ptlm_dstnid;    /* destination's NID */
+	__u64	   ptlm_dststamp;  /* destination's incarnation */
+	__u32	   ptlm_srcpid;    /* sender's PID */
+	__u32	   ptlm_dstpid;    /* destination's PID */
+
+	 union {
+		kptl_immediate_msg_t    immediate;
+		kptl_rdma_msg_t	 rdma;
+		kptl_hello_msg_t	hello;
+	} WIRE_ATTR ptlm_u;
+
+} kptl_msg_t;
+
+/* kptl_msg_t::ptlm_credits is only a __u8 */
+#define PTLLND_MSG_MAX_CREDITS ((typeof(((kptl_msg_t*) 0)->ptlm_credits)) -1)
+
+#define PTLLND_MSG_MAGIC		LNET_PROTO_PTL_MAGIC
+#define PTLLND_MSG_VERSION	      0x04
+
+#define PTLLND_RDMA_OK		  0x00
+#define PTLLND_RDMA_FAIL		0x01
+
+#define PTLLND_MSG_TYPE_INVALID	 0x00
+#define PTLLND_MSG_TYPE_PUT	     0x01
+#define PTLLND_MSG_TYPE_GET	     0x02
+#define PTLLND_MSG_TYPE_IMMEDIATE       0x03    /* No bulk data xfer*/
+#define PTLLND_MSG_TYPE_NOOP	    0x04
+#define PTLLND_MSG_TYPE_HELLO	   0x05
+#define PTLLND_MSG_TYPE_NAK	     0x06
diff --git a/drivers/staging/lustre/include/linux/lnet/socklnd.h b/drivers/staging/lustre/include/linux/lnet/socklnd.h
new file mode 100644
index 000000000000..bacc74933a39
--- /dev/null
+++ b/drivers/staging/lustre/include/linux/lnet/socklnd.h
@@ -0,0 +1,103 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * 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 version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
+ *
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ *
+ * lnet/include/lnet/socklnd.h
+ *
+ * #defines shared between socknal implementation and utilities
+ */
+#ifndef __LNET_LNET_SOCKLND_H__
+#define __LNET_LNET_SOCKLND_H__
+
+#include <linux/lnet/types.h>
+#include <linux/lnet/lib-types.h>
+
+#define SOCKLND_CONN_NONE     (-1)
+#define SOCKLND_CONN_ANY	0
+#define SOCKLND_CONN_CONTROL    1
+#define SOCKLND_CONN_BULK_IN    2
+#define SOCKLND_CONN_BULK_OUT   3
+#define SOCKLND_CONN_NTYPES     4
+
+#define SOCKLND_CONN_ACK	SOCKLND_CONN_BULK_IN
+
+typedef struct {
+	__u32		   kshm_magic;     /* magic number of socklnd message */
+	__u32		   kshm_version;   /* version of socklnd message */
+	lnet_nid_t	      kshm_src_nid;   /* sender's nid */
+	lnet_nid_t	      kshm_dst_nid;   /* destination nid */
+	lnet_pid_t	      kshm_src_pid;   /* sender's pid */
+	lnet_pid_t	      kshm_dst_pid;   /* destination pid */
+	__u64		   kshm_src_incarnation; /* sender's incarnation */
+	__u64		   kshm_dst_incarnation; /* destination's incarnation */
+	__u32		   kshm_ctype;     /* connection type */
+	__u32		   kshm_nips;      /* # IP addrs */
+	__u32		   kshm_ips[0];    /* IP addrs */
+} WIRE_ATTR ksock_hello_msg_t;
+
+typedef struct {
+	lnet_hdr_t	      ksnm_hdr;       /* lnet hdr */
+
+	/*
+	 * ksnm_payload is removed because of winnt compiler's limitation:
+	 * zero-sized array can only be placed at the tail of [nested]
+	 * structure definitions. lnet payload will be stored just after
+	 * the body of structure ksock_lnet_msg_t
+	 */
+} WIRE_ATTR ksock_lnet_msg_t;
+
+typedef struct {
+	__u32		   ksm_type;       /* type of socklnd message */
+	__u32		   ksm_csum;       /* checksum if != 0 */
+	__u64		   ksm_zc_cookies[2]; /* Zero-Copy request/ACK cookie */
+	union {
+		ksock_lnet_msg_t lnetmsg;       /* lnet message, it's empty if it's NOOP */
+	} WIRE_ATTR ksm_u;
+} WIRE_ATTR ksock_msg_t;
+
+static inline void
+socklnd_init_msg(ksock_msg_t *msg, int type)
+{
+	msg->ksm_csum	   = 0;
+	msg->ksm_type	   = type;
+	msg->ksm_zc_cookies[0]  = msg->ksm_zc_cookies[1]  = 0;
+}
+
+#define KSOCK_MSG_NOOP	  0xc0	    /* ksm_u empty */
+#define KSOCK_MSG_LNET	  0xc1	    /* lnet msg */
+
+/* We need to know this number to parse hello msg from ksocklnd in
+ * other LND (usocklnd, for example) */
+#define KSOCK_PROTO_V2	  2
+#define KSOCK_PROTO_V3	  3
+
+#endif
diff --git a/drivers/staging/lustre/include/linux/lnet/types.h b/drivers/staging/lustre/include/linux/lnet/types.h
new file mode 100644
index 000000000000..4f63b7acb9d7
--- /dev/null
+++ b/drivers/staging/lustre/include/linux/lnet/types.h
@@ -0,0 +1,503 @@
+/*
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * 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 version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
+ *
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
+ *
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
+ *
+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ *
+ * Copyright (c) 2012, Intel Corporation.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ */
+
+#ifndef __LNET_TYPES_H__
+#define __LNET_TYPES_H__
+
+/** \addtogroup lnet
+ * @{ */
+
+#include <linux/libcfs/libcfs.h>
+
+/** \addtogroup lnet_addr
+ * @{ */
+
+/** Portal reserved for LNet's own use.
+ * \see lustre/include/lustre/lustre_idl.h for Lustre portal assignments.
+ */
+#define LNET_RESERVED_PORTAL      0
+
+/**
+ * Address of an end-point in an LNet network.
+ *
+ * A node can have multiple end-points and hence multiple addresses.
+ * An LNet network can be a simple network (e.g. tcp0) or a network of
+ * LNet networks connected by LNet routers. Therefore an end-point address
+ * has two parts: network ID, and address within a network.
+ *
+ * \see LNET_NIDNET, LNET_NIDADDR, and LNET_MKNID.
+ */
+typedef __u64 lnet_nid_t;
+/**
+ * ID of a process in a node. Shortened as PID to distinguish from
+ * lnet_process_id_t, the global process ID.
+ */
+typedef __u32 lnet_pid_t;
+
+/** wildcard NID that matches any end-point address */
+#define LNET_NID_ANY      ((lnet_nid_t) -1)
+/** wildcard PID that matches any lnet_pid_t */
+#define LNET_PID_ANY      ((lnet_pid_t) -1)
+
+#define LNET_PID_RESERVED 0xf0000000 /* reserved bits in PID */
+#define LNET_PID_USERFLAG 0x80000000 /* set in userspace peers */
+
+#define LNET_TIME_FOREVER    (-1)
+
+/**
+ * Objects maintained by the LNet are accessed through handles. Handle types
+ * have names of the form lnet_handle_xx_t, where xx is one of the two letter
+ * object type codes ('eq' for event queue, 'md' for memory descriptor, and
+ * 'me' for match entry).
+ * Each type of object is given a unique handle type to enhance type checking.
+ * The type lnet_handle_any_t can be used when a generic handle is needed.
+ * Every handle value can be converted into a value of type lnet_handle_any_t
+ * without loss of information.
+ */
+typedef struct {
+	__u64	 cookie;
+} lnet_handle_any_t;
+
+typedef lnet_handle_any_t lnet_handle_eq_t;
+typedef lnet_handle_any_t lnet_handle_md_t;
+typedef lnet_handle_any_t lnet_handle_me_t;
+
+#define LNET_WIRE_HANDLE_COOKIE_NONE   (-1)
+
+/**
+ * Invalidate handle \a h.
+ */
+static inline void LNetInvalidateHandle(lnet_handle_any_t *h)
+{
+	h->cookie = LNET_WIRE_HANDLE_COOKIE_NONE;
+}
+
+/**
+ * Compare handles \a h1 and \a h2.
+ *
+ * \return 1 if handles are equal, 0 if otherwise.
+ */
+static inline int LNetHandleIsEqual (lnet_handle_any_t h1, lnet_handle_any_t h2)
+{
+	return (h1.cookie == h2.cookie);
+}
+
+/**
+ * Check whether handle \a h is invalid.
+ *
+ * \return 1 if handle is invalid, 0 if valid.
+ */
+static inline int LNetHandleIsInvalid(lnet_handle_any_t h)
+{
+	return (LNET_WIRE_HANDLE_COOKIE_NONE == h.cookie);
+}
+
+/**
+ * Global process ID.
+ */
+typedef struct {
+	/** node id */
+	lnet_nid_t nid;
+	/** process id */
+	lnet_pid_t pid;
+} lnet_process_id_t;
+/** @} lnet_addr */
+
+/** \addtogroup lnet_me
+ * @{ */
+
+/**
+ * Specifies whether the match entry or memory descriptor should be unlinked
+ * automatically (LNET_UNLINK) or not (LNET_RETAIN).
+ */
+typedef enum {
+	LNET_RETAIN = 0,
+	LNET_UNLINK
+} lnet_unlink_t;
+
+/**
+ * Values of the type lnet_ins_pos_t are used to control where a new match
+ * entry is inserted. The value LNET_INS_BEFORE is used to insert the new
+ * entry before the current entry or before the head of the list. The value
+ * LNET_INS_AFTER is used to insert the new entry after the current entry
+ * or after the last item in the list.
+ */
+typedef enum {
+	/** insert ME before current position or head of the list */
+	LNET_INS_BEFORE,
+	/** insert ME after current position or tail of the list */
+	LNET_INS_AFTER,
+	/** attach ME at tail of local CPU partition ME list */
+	LNET_INS_LOCAL
+} lnet_ins_pos_t;
+
+/** @} lnet_me */
+
+/** \addtogroup lnet_md
+ * @{ */
+
+/**
+ * Defines the visible parts of a memory descriptor. Values of this type
+ * are used to initialize memory descriptors.
+ */
+typedef struct {
+	/**
+	 * Specify the memory region associated with the memory descriptor.
+	 * If the options field has:
+	 * - LNET_MD_KIOV bit set: The start field points to the starting
+	 * address of an array of lnet_kiov_t and the length field specifies
+	 * the number of entries in the array. The length can't be bigger
+	 * than LNET_MAX_IOV. The lnet_kiov_t is used to describe page-based
+	 * fragments that are not necessarily mapped in virtal memory.
+	 * - LNET_MD_IOVEC bit set: The start field points to the starting
+	 * address of an array of struct iovec and the length field specifies
+	 * the number of entries in the array. The length can't be bigger
+	 * than LNET_MAX_IOV. The struct iovec is used to describe fragments
+	 * that have virtual addresses.
+	 * - Otherwise: The memory region is contiguous. The start field
+	 * specifies the starting address for the memory region and the
+	 * length field specifies its length.
+	 *
+	 * When the memory region is fragmented, all fragments but the first
+	 * one must start on page boundary, and all but the last must end on
+	 * page boundary.
+	 */
+	void	    *start;
+	unsigned int     length;
+	/**
+	 * Specifies the maximum number of operations that can be performed
+	 * on the memory descriptor. An operation is any action that could
+	 * possibly generate an event. In the usual case, the threshold value
+	 * is decremented for each operation on the MD. When the threshold
+	 * drops to zero, the MD becomes inactive and does not respond to
+	 * operations. A threshold value of LNET_MD_THRESH_INF indicates that
+	 * there is no bound on the number of operations that may be applied
+	 * to a MD.
+	 */
+	int	      threshold;
+	/**
+	 * Specifies the largest incoming request that the memory descriptor
+	 * should respond to. When the unused portion of a MD (length -
+	 * local offset) falls below this value, the MD becomes inactive and
+	 * does not respond to further operations. This value is only used
+	 * if the LNET_MD_MAX_SIZE option is set.
+	 */
+	int	      max_size;
+	/**
+	 * Specifies the behavior of the memory descriptor. A bitwise OR
+	 * of the following values can be used:
+	 * - LNET_MD_OP_PUT: The LNet PUT operation is allowed on this MD.
+	 * - LNET_MD_OP_GET: The LNet GET operation is allowed on this MD.
+	 * - LNET_MD_MANAGE_REMOTE: The offset used in accessing the memory
+	 *   region is provided by the incoming request. By default, the
+	 *   offset is maintained locally. When maintained locally, the
+	 *   offset is incremented by the length of the request so that
+	 *   the next operation (PUT or GET) will access the next part of
+	 *   the memory region. Note that only one offset variable exists
+	 *   per memory descriptor. If both PUT and GET operations are
+	 *   performed on a memory descriptor, the offset is updated each time.
+	 * - LNET_MD_TRUNCATE: The length provided in the incoming request can
+	 *   be reduced to match the memory available in the region (determined
+	 *   by subtracting the offset from the length of the memory region).
+	 *   By default, if the length in the incoming operation is greater
+	 *   than the amount of memory available, the operation is rejected.
+	 * - LNET_MD_ACK_DISABLE: An acknowledgment should not be sent for
+	 *   incoming PUT operations, even if requested. By default,
+	 *   acknowledgments are sent for PUT operations that request an
+	 *   acknowledgment. Acknowledgments are never sent for GET operations.
+	 *   The data sent in the REPLY serves as an implicit acknowledgment.
+	 * - LNET_MD_KIOV: The start and length fields specify an array of
+	 *   lnet_kiov_t.
+	 * - LNET_MD_IOVEC: The start and length fields specify an array of
+	 *   struct iovec.
+	 * - LNET_MD_MAX_SIZE: The max_size field is valid.
+	 *
+	 * Note:
+	 * - LNET_MD_KIOV or LNET_MD_IOVEC allows for a scatter/gather
+	 *   capability for memory descriptors. They can't be both set.
+	 * - When LNET_MD_MAX_SIZE is set, the total length of the memory
+	 *   region (i.e. sum of all fragment lengths) must not be less than
+	 *   \a max_size.
+	 */
+	unsigned int     options;
+	/**
+	 * A user-specified value that is associated with the memory
+	 * descriptor. The value does not need to be a pointer, but must fit
+	 * in the space used by a pointer. This value is recorded in events
+	 * associated with operations on this MD.
+	 */
+	void	    *user_ptr;
+	/**
+	 * A handle for the event queue used to log the operations performed on
+	 * the memory region. If this argument is a NULL handle (i.e. nullified
+	 * by LNetInvalidateHandle()), operations performed on this memory
+	 * descriptor are not logged.
+	 */
+	lnet_handle_eq_t eq_handle;
+} lnet_md_t;
+
+/* Max Transfer Unit (minimum supported everywhere).
+ * CAVEAT EMPTOR, with multinet (i.e. routers forwarding between networks)
+ * these limits are system wide and not interface-local. */
+#define LNET_MTU_BITS	20
+#define LNET_MTU	(1 << LNET_MTU_BITS)
+
+/** limit on the number of fragments in discontiguous MDs */
+#define LNET_MAX_IOV    256
+
+/* Max payload size */
+# define LNET_MAX_PAYLOAD	CONFIG_LNET_MAX_PAYLOAD
+# if (LNET_MAX_PAYLOAD < LNET_MTU)
+#  error "LNET_MAX_PAYLOAD too small - error in configure --with-max-payload-mb"
+# else
+#  if (LNET_MAX_PAYLOAD > (PAGE_SIZE * LNET_MAX_IOV))
+/*  PAGE_SIZE is a constant: check with cpp! */
+#   error "LNET_MAX_PAYLOAD too large - error in configure --with-max-payload-mb"
+#  endif
+# endif
+
+/**
+ * Options for the MD structure. See lnet_md_t::options.
+ */
+#define LNET_MD_OP_PUT	       (1 << 0)
+/** See lnet_md_t::options. */
+#define LNET_MD_OP_GET	       (1 << 1)
+/** See lnet_md_t::options. */
+#define LNET_MD_MANAGE_REMOTE	(1 << 2)
+/* unused			    (1 << 3) */
+/** See lnet_md_t::options. */
+#define LNET_MD_TRUNCATE	     (1 << 4)
+/** See lnet_md_t::options. */
+#define LNET_MD_ACK_DISABLE	  (1 << 5)
+/** See lnet_md_t::options. */
+#define LNET_MD_IOVEC		(1 << 6)
+/** See lnet_md_t::options. */
+#define LNET_MD_MAX_SIZE	     (1 << 7)
+/** See lnet_md_t::options. */
+#define LNET_MD_KIOV		 (1 << 8)
+
+/* For compatibility with Cray Portals */
+#define LNET_MD_PHYS			 0
+
+/** Infinite threshold on MD operations. See lnet_md_t::threshold */
+#define LNET_MD_THRESH_INF       (-1)
+
+/* NB lustre portals uses struct iovec internally! */
+typedef struct iovec lnet_md_iovec_t;
+
+/**
+ * A page-based fragment of a MD.
+ */
+typedef struct {
+	/** Pointer to the page where the fragment resides */
+	struct page      *kiov_page;
+	/** Length in bytes of the fragment */
+	unsigned int     kiov_len;
+	/**
+	 * Starting offset of the fragment within the page. Note that the
+	 * end of the fragment must not pass the end of the page; i.e.,
+	 * kiov_len + kiov_offset <= PAGE_CACHE_SIZE.
+	 */
+	unsigned int     kiov_offset;
+} lnet_kiov_t;
+/** @} lnet_md */
+
+/** \addtogroup lnet_eq
+ * @{ */
+
+/**
+ * Six types of events can be logged in an event queue.
+ */
+typedef enum {
+	/** An incoming GET operation has completed on the MD. */
+	LNET_EVENT_GET		= 1,
+	/**
+	 * An incoming PUT operation has completed on the MD. The
+	 * underlying layers will not alter the memory (on behalf of this
+	 * operation) once this event has been logged.
+	 */
+	LNET_EVENT_PUT,
+	/**
+	 * A REPLY operation has completed. This event is logged after the
+	 * data (if any) from the REPLY has been written into the MD.
+	 */
+	LNET_EVENT_REPLY,
+	/** An acknowledgment has been received. */
+	LNET_EVENT_ACK,
+	/**
+	 * An outgoing send (PUT or GET) operation has completed. This event
+	 * is logged after the entire buffer has been sent and it is safe for
+	 * the caller to reuse the buffer.
+	 *
+	 * Note:
+	 * - The LNET_EVENT_SEND doesn't guarantee message delivery. It can
+	 *   happen even when the message has not yet been put out on wire.
+	 * - It's unsafe to assume that in an outgoing GET operation
+	 *   the LNET_EVENT_SEND event would happen before the
+	 *   LNET_EVENT_REPLY event. The same holds for LNET_EVENT_SEND and
+	 *   LNET_EVENT_ACK events in an outgoing PUT operation.
+	 */
+	LNET_EVENT_SEND,
+	/**
+	 * A MD has been unlinked. Note that LNetMDUnlink() does not
+	 * necessarily trigger an LNET_EVENT_UNLINK event.
+	 * \see LNetMDUnlink
+	 */
+	LNET_EVENT_UNLINK,
+} lnet_event_kind_t;
+
+#define LNET_SEQ_BASETYPE       long
+typedef unsigned LNET_SEQ_BASETYPE lnet_seq_t;
+#define LNET_SEQ_GT(a,b)	(((signed LNET_SEQ_BASETYPE)((a) - (b))) > 0)
+
+/* XXX
+ * cygwin need the pragma line, not clear if it's needed in other places.
+ * checking!!!
+ */
+#ifdef __CYGWIN__
+#pragma pack(push, 4)
+#endif
+
+/**
+ * Information about an event on a MD.
+ */
+typedef struct {
+	/** The identifier (nid, pid) of the target. */
+	lnet_process_id_t   target;
+	/** The identifier (nid, pid) of the initiator. */
+	lnet_process_id_t   initiator;
+	/**
+	 * The NID of the immediate sender. If the request has been forwarded
+	 * by routers, this is the NID of the last hop; otherwise it's the
+	 * same as the initiator.
+	 */
+	lnet_nid_t	  sender;
+	/** Indicates the type of the event. */
+	lnet_event_kind_t   type;
+	/** The portal table index specified in the request */
+	unsigned int	pt_index;
+	/** A copy of the match bits specified in the request. */
+	__u64	       match_bits;
+	/** The length (in bytes) specified in the request. */
+	unsigned int	rlength;
+	/**
+	 * The length (in bytes) of the data that was manipulated by the
+	 * operation. For truncated operations, the manipulated length will be
+	 * the number of bytes specified by the MD (possibly with an offset,
+	 * see lnet_md_t). For all other operations, the manipulated length
+	 * will be the length of the requested operation, i.e. rlength.
+	 */
+	unsigned int	mlength;
+	/**
+	 * The handle to the MD associated with the event. The handle may be
+	 * invalid if the MD has been unlinked.
+	 */
+	lnet_handle_md_t    md_handle;
+	/**
+	 * A snapshot of the state of the MD immediately after the event has
+	 * been processed. In particular, the threshold field in md will
+	 * reflect the value of the threshold after the operation occurred.
+	 */
+	lnet_md_t	   md;
+	/**
+	 * 64 bits of out-of-band user data. Only valid for LNET_EVENT_PUT.
+	 * \see LNetPut
+	 */
+	__u64	       hdr_data;
+	/**
+	 * Indicates the completion status of the operation. It's 0 for
+	 * successful operations, otherwise it's an error code.
+	 */
+	int		 status;
+	/**
+	 * Indicates whether the MD has been unlinked. Note that:
+	 * - An event with unlinked set is the last event on the MD.
+	 * - This field is also set for an explicit LNET_EVENT_UNLINK event.
+	 * \see LNetMDUnlink
+	 */
+	int		 unlinked;
+	/**
+	 * The displacement (in bytes) into the memory region that the
+	 * operation used. The offset can be determined by the operation for
+	 * a remote managed MD or by the local MD.
+	 * \see lnet_md_t::options
+	 */
+	unsigned int	offset;
+	/**
+	 * The sequence number for this event. Sequence numbers are unique
+	 * to each event.
+	 */
+	volatile lnet_seq_t sequence;
+} lnet_event_t;
+#ifdef __CYGWIN__
+#pragma pop
+#endif
+
+/**
+ * Event queue handler function type.
+ *
+ * The EQ handler runs for each event that is deposited into the EQ. The
+ * handler is supplied with a pointer to the event that triggered the
+ * handler invocation.
+ *
+ * The handler must not block, must be reentrant, and must not call any LNet
+ * API functions. It should return as quickly as possible.
+ */
+typedef void (*lnet_eq_handler_t)(lnet_event_t *event);
+#define LNET_EQ_HANDLER_NONE NULL
+/** @} lnet_eq */
+
+/** \addtogroup lnet_data
+ * @{ */
+
+/**
+ * Specify whether an acknowledgment should be sent by target when the PUT
+ * operation completes (i.e., when the data has been written to a MD of the
+ * target process).
+ *
+ * \see lnet_md_t::options for the discussion on LNET_MD_ACK_DISABLE by which
+ * acknowledgments can be disabled for a MD.
+ */
+typedef enum {
+	/** Request an acknowledgment */
+	LNET_ACK_REQ,
+	/** Request that no acknowledgment should be generated. */
+	LNET_NOACK_REQ
+} lnet_ack_req_t;
+/** @} lnet_data */
+
+/** @} lnet */
+#endif