Merge remote-tracking branch 'origin/master'

[unleashed/lotheac.git] / usr / src / uts / common / io / rsm / rsmka_pathmanager.c

/*
 * CDDL HEADER START
 *
 * The contents of this file are subject to the terms of the
 * Common Development and Distribution License, Version 1.0 only
 * (the "License").  You may not use this file except in compliance
 * with the License.
 *
 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
 * or http://www.opensolaris.org/os/licensing.
 * See the License for the specific language governing permissions
 * and limitations under the License.
 *
 * When distributing Covered Code, include this CDDL HEADER in each
 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
 * If applicable, add the following below this CDDL HEADER, with the
 * fields enclosed by brackets "[]" replaced with your own identifying
 * information: Portions Copyright [yyyy] [name of copyright owner]
 *
 * CDDL HEADER END
 */
/*
 * Copyright 2004 Sun Microsystems, Inc.  All rights reserved.
 * Use is subject to license terms.
 * Copyright 2012 Milan Jurik. All rights reserved.
 */

/*
 * This module provides for the management of interconnect adapters
 * inter-node connections (aka paths), and IPC.  Adapter descriptors are
 * maintained on a linked list; one list per adapter devname.  Each
 * adapter descriptor heads a linked list of path descriptors.  There is
 * also a linked list of ipc_info descriptors; one for each node.  Each
 * ipc_info descriptor heads a circular list of ipc tokens (the tokens are
 * embedded within a path descriptor). The tokens are used in round robin
 * fashion.
 *
 *
 * The exported interface consists of the following functions:
 *      - rsmka_add_adapter
 *      - rsmka_remove_adapter
 *
 *      [add_path and remove_path only called for current adapters]
 *      - rsmka_add_path
 *      - rsmka_remove_path     [a path down request is implicit]
 *
 *      - rsmka_path_up           [called at clock ipl for Sun Cluster]
 *      - rsmka_path_down         [called at clock ipl for Sun Cluster]
 *      - rsmka_disconnect_node   [called at clock ipl for Sun Cluster;
 *                              treat like path-down for all node paths;
 *                              can be before node_alive; always before
 *                              node_died.]
 *
 *      [node_alive and node_died are always paired]
 *      - rsmka_node_alive   called after the first cluster path is up
 *                           for this node
 *      - rsmka_node_died
 *
 *      [set the local node id]
 *      - rsmka_set_my_nodeid    called to set the variable my_nodeid to the
 *                           local node id
 *
 * Processing for these functions is setup as a state machine supported
 * by the data structures described above.
 *
 * For Sun Cluster these are called from the Path-Manager/Kernel-Agent
 * Interface (rsmka_pm_interface.cc).
 *
 * The functions rsm_path_up, rsm_path_down, and rsm_disconnect_node are
 * called at clock interrupt level from the Path-Manager/Kernel-Agent
 * Interface which precludes sleeping; so these functions may (optionally)
 * defer processing to an independent thread running at normal ipl.
 *
 *
 * lock definitions:
 *
 *      (mutex) work_queue.work_mutex
 *                      protects linked list of work tokens and used
 *                      with cv_wait/cv_signal thread synchronization.
 *                      No other locks acquired when held.
 *
 *      (mutex) adapter_listhead_base.listlock
 *                      protects linked list of adapter listheads
 *                      Always acquired before listhead->mutex
 *
 *
 *      (mutex) ipc_info_lock
 *                      protects ipc_info list and sendq token lists
 *                      Always acquired before listhead->mutex
 *
 *      (mutex) listhead->mutex
 *                      protects adapter listhead, linked list of
 *                      adapters, and linked list of paths.
 *
 *      (mutex) path->mutex
 *                      protects the path descriptor.
 *                      work_queue.work_mutex may be acquired when holding
 *                      this lock.
 *
 *      (mutex) adapter->mutex
 *                      protects adapter descriptor contents.  used
 *                      mainly for ref_cnt update.
 */

#include <sys/param.h>
#include <sys/kmem.h>
#include <sys/errno.h>
#include <sys/time.h>
#include <sys/devops.h>
#include <sys/ddi_impldefs.h>
#include <sys/cmn_err.h>
#include <sys/ddi.h>
#include <sys/sunddi.h>
#include <sys/proc.h>
#include <sys/thread.h>
#include <sys/taskq.h>
#include <sys/callb.h>

#include <sys/rsm/rsm.h>
#include <rsm_in.h>
#include <sys/rsm/rsmka_path_int.h>

extern void _cplpl_init();
extern void _cplpl_fini();
extern pri_t maxclsyspri;
extern int   rsm_hash_size;

extern rsm_node_id_t my_nodeid;
extern rsmhash_table_t rsm_import_segs;
extern rsm_intr_hand_ret_t rsm_srv_func(rsm_controller_object_t *,
    rsm_intr_q_op_t, rsm_addr_t, void *, size_t, rsm_intr_hand_arg_t);
extern void rsmseg_unload(rsmseg_t *);
extern void rsm_suspend_complete(rsm_node_id_t src_node, int flag);
extern int rsmipc_send_controlmsg(path_t *path, int msgtype);
extern void rsmka_path_monitor_initialize();
extern void rsmka_path_monitor_terminate();

extern adapter_t loopback_adapter;
/*
 * Lint errors and warnings are displayed; informational messages
 * are suppressed.
 */
/* lint -w2 */


/*
 * macros SQ_TOKEN_TO_PATH and WORK_TOKEN_TO_PATH use a null pointer
 * for computational purposes.  Ignore the lint warning.
 */
/* lint -save -e413 */
/* FUNCTION PROTOTYPES */
static adapter_t *init_adapter(char *, int, rsm_addr_t,
    rsm_controller_handle_t, rsm_ops_t *, srv_handler_arg_t *);
adapter_t *rsmka_lookup_adapter(char *, int);
static ipc_info_t *lookup_ipc_info(rsm_node_id_t);
static ipc_info_t *init_ipc_info(rsm_node_id_t, boolean_t);
static path_t *lookup_path(char *, int, rsm_node_id_t, rsm_addr_t);
static void pathup_to_pathactive(ipc_info_t *, rsm_node_id_t);
static void path_importer_disconnect(path_t *);
boolean_t rsmka_do_path_active(path_t *, int);
static boolean_t do_path_up(path_t *, int);
static void do_path_down(path_t *, int);
static void enqueue_work(work_token_t *);
static boolean_t cancel_work(work_token_t *);
static void link_path(path_t *);
static void destroy_path(path_t *);
static void link_sendq_token(sendq_token_t *, rsm_node_id_t);
static void unlink_sendq_token(sendq_token_t *, rsm_node_id_t);
boolean_t rsmka_check_node_alive(rsm_node_id_t);
static void do_deferred_work(caddr_t);
static int create_ipc_sendq(path_t *);
static void destroy_ipc_info(ipc_info_t *);
void rsmka_pathmanager_cleanup();
void rsmka_release_adapter(adapter_t *);

kt_did_t rsm_thread_id;
int rsmka_terminate_workthread_loop = 0;

static struct adapter_listhead_list adapter_listhead_base;
static work_queue_t work_queue;

/* protect ipc_info descriptor manipulation */
static kmutex_t ipc_info_lock;

static ipc_info_t *ipc_info_head = NULL;

static int category = RSM_PATH_MANAGER | RSM_KERNEL_AGENT;

/* for synchronization with rsmipc_send() in rsm.c */
kmutex_t ipc_info_cvlock;
kcondvar_t ipc_info_cv;



/*
 * RSMKA PATHMANAGER INITIALIZATION AND CLEANUP ROUTINES
 *
 */


/*
 * Called from the rsm module (rsm.c)  _init() routine
 */
void
rsmka_pathmanager_init()
{
        kthread_t *tp;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsmka_pathmanager_init enter\n"));

        /* initialization for locks and condition variables  */
        mutex_init(&work_queue.work_mutex, NULL, MUTEX_DEFAULT, NULL);
        mutex_init(&ipc_info_lock, NULL, MUTEX_DEFAULT, NULL);
        mutex_init(&ipc_info_cvlock, NULL, MUTEX_DEFAULT, NULL);
        mutex_init(&adapter_listhead_base.listlock, NULL,
            MUTEX_DEFAULT, NULL);

        cv_init(&work_queue.work_cv, NULL, CV_DEFAULT, NULL);
        cv_init(&ipc_info_cv, NULL, CV_DEFAULT, NULL);

        tp = thread_create(NULL, 0, do_deferred_work, NULL, 0, &p0,
            TS_RUN, maxclsyspri);
        rsm_thread_id = tp->t_did;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsmka_pathmanager_init done\n"));
}

void
rsmka_pathmanager_cleanup()
{
        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsmka_pathmanager_cleanup enter\n"));

        ASSERT(work_queue.head == NULL);

        /*
         * In processing the remove path callbacks from the path monitor
         * object, all deferred work will have been completed. So
         * awaken the deferred work thread to give it a chance to exit
         * the loop.
         */
        mutex_enter(&work_queue.work_mutex);
        rsmka_terminate_workthread_loop++;
        cv_signal(&work_queue.work_cv);
        mutex_exit(&work_queue.work_mutex);

        /*
         * Wait for the deferred work thread to exit before
         * destroying the locks and cleaning up other data
         * structures.
         */
        if (rsm_thread_id)
                thread_join(rsm_thread_id);

        /*
         * Destroy locks & condition variables
         */
        mutex_destroy(&work_queue.work_mutex);
        cv_destroy(&work_queue.work_cv);

        mutex_enter(&ipc_info_lock);
        while (ipc_info_head)
                destroy_ipc_info(ipc_info_head);
        mutex_exit(&ipc_info_lock);

        mutex_destroy(&ipc_info_lock);

        mutex_destroy(&ipc_info_cvlock);
        cv_destroy(&ipc_info_cv);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsmka_pathmanager_cleanup done\n"));

}

void
rsmka_set_my_nodeid(rsm_node_id_t local_nodeid)
{
        my_nodeid = local_nodeid;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsm: node %d \n", my_nodeid));

}

/*
 * DEFERRED WORK THREAD AND WORK QUEUE SUPPORT ROUTINES
 *
 */

/*
 * This function is the executable code of the thread which handles
 * deferred work.  Work is deferred when a function is called at
 * clock ipl and processing may require blocking.
 *
 *
 * The thread is created by a call to taskq_create in rsmka_pathmanager_init.
 * After creation, a call to taskq_dispatch causes this function to
 * execute.  It loops forever - blocked until work is enqueued from
 * rsmka_do_path_active, do_path_down, or rsmka_disconnect_node.
 * rsmka_pathmanager_cleanup (called from _fini) will
 * set rsmka_terminate_workthread_loop and the task processing will
 * terminate.
 */
static void
do_deferred_work(caddr_t arg /*ARGSUSED*/)
{

        adapter_t                       *adapter;
        path_t                          *path;
        work_token_t                    *work_token;
        int                             work_opcode;
        rsm_send_q_handle_t             sendq_handle;
        int                             error;
        timespec_t                      tv;
        callb_cpr_t                     cprinfo;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "do_deferred_work enter\n"));

        CALLB_CPR_INIT(&cprinfo, &work_queue.work_mutex, callb_generic_cpr,
            "rsm_deferred_work");

        for (;;) {
                mutex_enter(&work_queue.work_mutex);

                if (rsmka_terminate_workthread_loop) {
                        goto exit;
                }

                /* When there is no work to do, block here */
                while (work_queue.head == NULL) {
                        /* Since no work to do, Safe to CPR */
                        CALLB_CPR_SAFE_BEGIN(&cprinfo);
                        cv_wait(&work_queue.work_cv, &work_queue.work_mutex);
                        CALLB_CPR_SAFE_END(&cprinfo, &work_queue.work_mutex);

                        if (rsmka_terminate_workthread_loop) {
                                goto exit;
                        }
                }

                /*
                 * Remove a work token and begin work
                 */
                work_token = work_queue.head;
                work_queue.head = work_token->next;
                if (work_queue.tail == work_token)
                        work_queue.tail = NULL;

                work_opcode = work_token->opcode;
                path = WORK_TOKEN_TO_PATH(work_token, work_opcode -1);
                work_token->next = NULL;
                mutex_exit(&work_queue.work_mutex);


                switch (work_opcode) {
                case RSMKA_IPC_UP:
                        DBG_PRINTF((category, RSM_DEBUG,
                            "do_deferred_work:up,  path = %lx\n", path));
                        error = create_ipc_sendq(path);
                        mutex_enter(&path->mutex);
                        if (path->state != RSMKA_PATH_UP) {
                                /*
                                 * path state has changed, if sendq was created,
                                 * destroy it and return. Don't need to worry
                                 * about sendq ref_cnt since no one starts
                                 * using the sendq till path state becomes
                                 * active
                                 */
                                if (error == RSM_SUCCESS) {
                                        sendq_handle = path->sendq_token.
                                            rsmpi_sendq_handle;
                                        path->sendq_token.rsmpi_sendq_handle =
                                            NULL;
                                        adapter = path->local_adapter;
                                        mutex_exit(&path->mutex);

                                        if (sendq_handle != NULL) {
                                                adapter->rsmpi_ops->
                                                    rsm_sendq_destroy(
                                                    sendq_handle);
                                        }
                                        mutex_enter(&path->mutex);
                                }
                                /* free up work token */
                                work_token->opcode = 0;

                                /*
                                 * decrement reference count for the path
                                 * descriptor and signal for synchronization
                                 * with rsmka_remove_path. PATH_HOLD_NOLOCK was
                                 * done by rsmka_path_up.
                                 */
                                PATH_RELE_NOLOCK(path);
                                mutex_exit(&path->mutex);
                                break;
                        }

                        if (error == RSM_SUCCESS) {
                                DBG_PRINTF((category, RSM_DEBUG,
                                    "do_deferred_work:success on up\n"));
                                /* clear flag since sendq_create succeeded */
                                path->flags &= ~RSMKA_SQCREATE_PENDING;
                                path->state = RSMKA_PATH_ACTIVE;

                                /*
                                 * now that path is active we send the
                                 * RSMIPC_MSG_SQREADY to the remote endpoint
                                 */
                                path->procmsg_cnt = 0;
                                path->sendq_token.msgbuf_avail = 0;

                                /* Calculate local incarnation number */
                                gethrestime(&tv);
                                if (tv.tv_sec == RSM_UNKNOWN_INCN)
                                        tv.tv_sec = 1;
                                path->local_incn = (int64_t)tv.tv_sec;

                                /*
                                 * if send fails here its due to some
                                 * non-transient error because QUEUE_FULL is
                                 * not possible here since we are the first
                                 * message on this sendq. The error will cause
                                 * the path to go down anyways, so ignore
                                 * the return value.
                                 */
                                (void) rsmipc_send_controlmsg(path,
                                    RSMIPC_MSG_SQREADY);
                                /* wait for SQREADY_ACK message */
                                path->flags |= RSMKA_WAIT_FOR_SQACK;
                        } else {
                                /*
                                 * sendq create failed possibly because
                                 * the remote end is not yet ready eg.
                                 * handler not registered, set a flag
                                 * so that when there is an indication
                                 * that the remote end is ready
                                 * rsmka_do_path_active will be retried.
                                 */
                                path->flags |= RSMKA_SQCREATE_PENDING;
                        }

                        /* free up work token */
                        work_token->opcode = 0;

                        /*
                         * decrement reference count for the path
                         * descriptor and signal for synchronization with
                         * rsmka_remove_path. PATH_HOLD_NOLOCK was done
                         * by rsmka_path_up.
                         */
                        PATH_RELE_NOLOCK(path);
                        mutex_exit(&path->mutex);

                        break;
                case RSMKA_IPC_DOWN:
                        DBG_PRINTF((category, RSM_DEBUG,
                            "do_deferred_work:down, path = %lx\n", path));

                        /*
                         * Unlike the processing of path_down in the case
                         * where the RSMKA_NO_SLEEP flag is not set, here,
                         * the state of the path is changed directly to
                         * RSMKA_PATH_DOWN. This is because in this case
                         * where the RSMKA_NO_SLEEP flag is set, any other
                         * calls referring this path will just queue up
                         * and will be processed only after the path
                         * down processing has completed.
                         */
                        mutex_enter(&path->mutex);
                        path->state = RSMKA_PATH_DOWN;
                        /*
                         * clear the WAIT_FOR_SQACK flag since path is down.
                         */
                        path->flags &= ~RSMKA_WAIT_FOR_SQACK;

                        /*
                         * this wakes up any thread waiting to receive credits
                         * in rsmipc_send to tell it that the path is down
                         * thus releasing the sendq.
                         */
                        cv_broadcast(&path->sendq_token.sendq_cv);

                        mutex_exit(&path->mutex);

                        /* drain the messages from the receive msgbuf */
                        taskq_wait(path->recv_taskq);

                        /*
                         * The path_importer_disconnect function has to
                         * be called after releasing the mutex on the path
                         * in order to avoid any recursive mutex enter panics
                         */
                        path_importer_disconnect(path);
                        DBG_PRINTF((category, RSM_DEBUG,
                            "do_deferred_work: success on down\n"));
                        /*
                         * decrement reference count for the path
                         * descriptor and signal for synchronization with
                         * rsmka_remove_path. PATH_HOLD_NOLOCK was done
                         * by rsmka_path_down.
                         */
                        mutex_enter(&path->mutex);

#ifdef DEBUG
                        /*
                         * Some IPC messages left in the recv_buf,
                         * they'll be dropped
                         */
                        if (path->msgbuf_cnt != 0)
                                cmn_err(CE_NOTE,
                                    "path=%lx msgbuf_cnt != 0\n",
                                    (uintptr_t)path);
#endif

                        /*
                         * Don't want to destroy a send queue when a token
                         * has been acquired; so wait 'til the token is
                         * no longer referenced (with a cv_wait).
                         */
                        while (path->sendq_token.ref_cnt != 0)
                                cv_wait(&path->sendq_token.sendq_cv,
                                    &path->mutex);

                        sendq_handle = path->sendq_token.rsmpi_sendq_handle;
                        path->sendq_token.rsmpi_sendq_handle = NULL;

                        /* destroy the send queue and release the handle */
                        if (sendq_handle != NULL) {
                                adapter = path->local_adapter;
                                adapter->rsmpi_ops->rsm_sendq_destroy(
                                    sendq_handle);
                        }

                        work_token->opcode = 0;
                        PATH_RELE_NOLOCK(path);
                        mutex_exit(&path->mutex);
                        break;
                default:
                        DBG_PRINTF((category, RSM_DEBUG,
                            "do_deferred_work: bad work token opcode\n"));
                        break;
                }
        }

exit:
        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "do_deferred_work done\n"));
        /*
         * CALLB_CPR_EXIT does a mutex_exit for
         * the work_queue.work_mutex
         */
        CALLB_CPR_EXIT(&cprinfo);
}

/*
 * Work is inserted at the tail of the list and processed from the
 * head of the list.
 */
static void
enqueue_work(work_token_t *token)
{
        work_token_t    *tail_token;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "enqueue_work enter\n"));

        ASSERT(MUTEX_HELD(&work_queue.work_mutex));

        token->next = NULL;
        if (work_queue.head == NULL) {
                work_queue.head = work_queue.tail = token;
        } else {
                tail_token = work_queue.tail;
                work_queue.tail = tail_token->next = token;
        }

        /* wake up deferred work thread */
        cv_signal(&work_queue.work_cv);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "enqueue_work done\n"));
}


/*
 * If the work_token is found on the work queue, the work is cancelled
 * by removing the token from the work queue.
 *
 * Return true if a work_token was found and cancelled, otherwise return false
 *
 * enqueue_work increments the path refcnt to make sure that the path doesn't
 * go away, callers of cancel_work need to decrement the refcnt of the path to
 * which this work_token belongs if a work_token is found in the work_queue
 * and cancelled ie. when the return value is B_TRUE.
 */
static boolean_t
cancel_work(work_token_t *work_token)
{
        work_token_t    *current_token;
        work_token_t    *prev_token = NULL;
        boolean_t       cancelled = B_FALSE;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "cancel_work enter\n"));

        ASSERT(MUTEX_HELD(&work_queue.work_mutex));


        current_token = work_queue.head;
        while (current_token != NULL) {
                if (current_token == work_token) {
                        if (work_token == work_queue.head)
                                work_queue.head = work_token->next;
                        else
                                prev_token->next = work_token->next;
                        if (work_token == work_queue.tail)
                                work_queue.tail = prev_token;

                        current_token->opcode = 0;
                        current_token->next = NULL;
                        /* found and cancelled work */
                        cancelled = B_TRUE;
                        DBG_PRINTF((category, RSM_DEBUG,
                            "cancelled_work = 0x%p\n", work_token));
                        break;
                }
                prev_token = current_token;
                current_token = current_token->next;
        }

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "cancel_work done\n"));
        return (cancelled);
}

/*
 * EXTERNAL INTERFACES
 *
 * For Galileo Clustering, these routine are called from
 * rsmka_pm_interface.cc
 *
 */

/*
 *
 * If the adapter is supported by rsmpi then initialize an adapter descriptor
 * and link it to the list of adapters.  The adapter attributes are obtained
 * from rsmpi and stored in the descriptor.  Finally, a service handler
 * for incoming ipc on this adapter is registered with rsmpi.
 * A pointer for the adapter descriptor is returned as a cookie to the
 * caller.  The cookie may be use with subsequent calls to save the time of
 * adapter descriptor lookup.
 *
 * The adapter descriptor maintains a reference count which is intialized
 * to 1 and incremented on lookups; when a cookie is used in place of
 * a lookup, an explicit ADAPTER_HOLD is required.
 */

void *
rsmka_add_adapter(char *name, int instance, rsm_addr_t hwaddr)
{
        adapter_t               *adapter;
        rsm_controller_object_t rsmpi_adapter_object;
        rsm_controller_handle_t rsmpi_adapter_handle;
        rsm_ops_t               *rsmpi_ops_vector;
        int                     adapter_is_supported;
        rsm_controller_attr_t   *attr;
        srv_handler_arg_t       *srv_hdlr_argp;
        int result;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "rsmka_add_adapter enter\n"));

        DBG_PRINTF((category, RSM_DEBUG,
            "rsmka_add_adapter: name = %s instance = %d hwaddr = %llx \n",
            name, instance, hwaddr));

        /* verify name length */
        if (strlen(name) >= MAXNAMELEN) {
                DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
                    "rsmka_add_adapter done: name too long\n"));
                return (NULL);
        }


        /* Check if rsmpi supports this adapter type */
        adapter_is_supported = rsm_get_controller(name, instance,
            &rsmpi_adapter_object, RSM_VERSION);

        if (adapter_is_supported != RSM_SUCCESS) {
                DBG_PRINTF((category, RSM_ERR,
                    "rsmka_add_adapter done: adapter not supported\n"));
                return (NULL);
        }

        rsmpi_adapter_handle = rsmpi_adapter_object.handle;
        rsmpi_ops_vector = rsmpi_adapter_object.ops;

        /* Get adapter attributes */
        result = rsm_get_controller_attr(rsmpi_adapter_handle, &attr);
        if (result != RSM_SUCCESS) {
                DBG_PRINTF((category, RSM_ERR,
                    "rsm: get_controller_attr(%d) Failed %x\n",
                    instance, result));
                (void) rsm_release_controller(name, instance,
                    &rsmpi_adapter_object);
                return (NULL);
        }

        DBG_PRINTF((category, RSM_DEBUG,
            "rsmka_add_adapter: register service offset = %d\n", hwaddr));

        /*
         * create a srv_handler_arg_t object, initialize it and register
         * it along with rsm_srv_func. This get passed as the
         * rsm_intr_hand_arg_t when the handler gets invoked.
         */
        srv_hdlr_argp = kmem_zalloc(sizeof (srv_handler_arg_t), KM_SLEEP);

        (void) strcpy(srv_hdlr_argp->adapter_name, name);
        srv_hdlr_argp->adapter_instance = instance;
        srv_hdlr_argp->adapter_hwaddr = hwaddr;

        /* Have rsmpi register the ipc receive handler for this adapter */
        /*
         * Currently, we need to pass in a separate service identifier for
         * each adapter. In order to obtain a unique service identifier
         * value for an adapter, we add the hardware address of the
         * adapter to the base service identifier(RSM_SERVICE which is
         * defined as RSM_INTR_T_KA as per the RSMPI specification).
         * NOTE: This may result in using some of the service identifier
         * values defined for RSM_INTR_T_XPORT(the Sun Cluster Transport).
         */
        result = rsmpi_ops_vector->rsm_register_handler(
            rsmpi_adapter_handle, &rsmpi_adapter_object,
            RSM_SERVICE+(uint_t)hwaddr, rsm_srv_func,
            (rsm_intr_hand_arg_t)srv_hdlr_argp, NULL, 0);

        if (result != RSM_SUCCESS) {
                DBG_PRINTF((category, RSM_ERR,
                    "rsmka_add_adapter done: rsm_register_handler"
                    "failed %d\n",
                    instance));
                return (NULL);
        }

        /* Initialize an adapter descriptor and add it to the adapter list */
        adapter = init_adapter(name, instance, hwaddr,
            rsmpi_adapter_handle, rsmpi_ops_vector, srv_hdlr_argp);

        /* Copy over the attributes from the pointer returned to us */
        adapter->rsm_attr = *attr;

        /*
         * With the addition of the topology obtainment interface, applications
         * now get the local nodeid from the topology data structure.
         *
         * adapter->rsm_attr.attr_node_id = my_nodeid;
         */
        DBG_PRINTF((category, RSM_ERR,
            "rsmka_add_adapter: adapter = %lx\n", adapter));

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "rsmka_add_adapter done\n"));

        /* return adapter pointer as a cookie for later fast access */
        return ((void *)adapter);
}


/*
 * Unlink the adapter descriptor and call rsmka_release_adapter which
 * will decrement the reference count and possibly free the desriptor.
 */
boolean_t
rsmka_remove_adapter(char *name, uint_t instance, void *cookie, int flags)
{
        adapter_t               *adapter;
        adapter_listhead_t      *listhead;
        adapter_t               *prev, *current;
        rsm_controller_object_t rsm_cntl_obj;


        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsmka_remove_adapter enter\n"));

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsmka_remove_adapter: cookie = %lx\n", cookie));

        if (flags & RSMKA_USE_COOKIE) {
                adapter = (adapter_t *)cookie;
        } else {
                adapter = rsmka_lookup_adapter(name, instance);
                /*
                 * rsmka_lookup_adapter increments the ref_cnt; need
                 * to decrement here to get true count
                 */
                ADAPTER_RELE(adapter);
        }
        ASSERT(adapter->next_path == NULL);

        listhead = adapter->listhead;

        mutex_enter(&adapter_listhead_base.listlock);

        mutex_enter(&listhead->mutex);

        /* find the adapter in the list and remove it */
        prev = NULL;
        current = listhead->next_adapter;
        while (current != NULL) {
                if (adapter->instance == current->instance) {
                        break;
                } else {
                        prev = current;
                        current = current->next;
                }
        }
        ASSERT(current != NULL);

        if (prev == NULL)
                listhead->next_adapter = current->next;
        else
                prev->next = current->next;

        listhead->adapter_count--;

        mutex_exit(&listhead->mutex);

        mutex_exit(&adapter_listhead_base.listlock);

        mutex_enter(&current->mutex);

        /*
         * unregister the handler
         */
        current->rsmpi_ops->rsm_unregister_handler(current->rsmpi_handle,
            RSM_SERVICE+current->hwaddr, rsm_srv_func,
            (rsm_intr_hand_arg_t)current->hdlr_argp);

        DBG_PRINTF((category, RSM_DEBUG, "rsmka_remove_adapter: unreg hdlr "
            ":adapter=%lx, hwaddr=%lx\n", current, current->hwaddr));

        rsm_cntl_obj.handle = current->rsmpi_handle;
        rsm_cntl_obj.ops = current->rsmpi_ops;

        (void) rsm_release_controller(current->listhead->adapter_devname,
            current->instance, &rsm_cntl_obj);

        mutex_exit(&current->mutex);

        rsmka_release_adapter(current);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsmka_remove_adapter done\n"));

        return (B_TRUE);
}

/*
 * An adapter descriptor will exist from an earlier add_adapter. This
 * function does:
 *              initialize the path descriptor
 *              initialize the ipc descriptor (it may already exist)
 *              initialize and link a sendq token for this path
 */
void *
rsmka_add_path(char *adapter_name, int adapter_instance,
    rsm_node_id_t remote_node,
    rsm_addr_t remote_hwaddr, int rem_adapt_instance,
    void *cookie, int flags)
{

        path_t                  *path;
        adapter_t               *adapter;
        char                    tq_name[TASKQ_NAMELEN];

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "rsmka_add_path enter\n"));

        /* allocate new path descriptor */
        path = kmem_zalloc(sizeof (path_t), KM_SLEEP);

        if (flags & RSMKA_USE_COOKIE) {
                adapter = (adapter_t *)cookie;
                ADAPTER_HOLD(adapter);
        } else {
                adapter = rsmka_lookup_adapter(adapter_name, adapter_instance);
        }

        DBG_PRINTF((category, RSM_DEBUG,
            "rsmka_add_path: adapter = %lx\n", adapter));

        /*
         * initialize path descriptor
         * don't need to increment adapter reference count because
         * it can't be removed if paths exist for it.
         */
        mutex_init(&path->mutex, NULL, MUTEX_DEFAULT, NULL);

        PATH_HOLD(path);
        path->state = RSMKA_PATH_DOWN;
        path->remote_node = remote_node;
        path->remote_hwaddr = remote_hwaddr;
        path->remote_devinst = rem_adapt_instance;
        path->local_adapter = adapter;

        /* taskq is for sendq on adapter with remote_hwaddr on remote_node */
        (void) snprintf(tq_name, sizeof (tq_name), "%x_%llx",
            remote_node, (unsigned long long) remote_hwaddr);

        path->recv_taskq = taskq_create_instance(tq_name, adapter_instance,
            RSMKA_ONE_THREAD, maxclsyspri, RSMIPC_MAX_MESSAGES,
            RSMIPC_MAX_MESSAGES, TASKQ_PREPOPULATE);

        /* allocate the message buffer array */
        path->msgbuf_queue = (msgbuf_elem_t *)kmem_zalloc(
            RSMIPC_MAX_MESSAGES * sizeof (msgbuf_elem_t), KM_SLEEP);

        /*
         * init cond variables for synch with rsmipc_send()
         * and rsmka_remove_path
         */
        cv_init(&path->sendq_token.sendq_cv, NULL, CV_DEFAULT, NULL);
        cv_init(&path->hold_cv, NULL, CV_DEFAULT, NULL);

        /* link path descriptor on adapter path list */
        link_path(path);

        /* link the path sendq token on the ipc_info token list */
        link_sendq_token(&path->sendq_token, remote_node);

        /* ADAPTER_HOLD done above by rsmka_lookup_adapter */
        ADAPTER_RELE(adapter);

        DBG_PRINTF((category, RSM_DEBUG, "rsmka_add_path: path = %lx\n", path));

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "rsmka_add_path done\n"));
        return ((void *)path);
}

/*
 * Wait for the path descriptor reference count to become zero then
 * directly call path down processing.  Finally, unlink the sendq token and
 * free the path descriptor memory.
 *
 * Note: lookup_path locks the path and increments the path hold count
 */
void
rsmka_remove_path(char *adapter_name, int instance, rsm_node_id_t remote_node,
    rsm_addr_t remote_hwaddr, void *path_cookie, int flags)
{
        path_t          *path;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "rsmka_remove_path enter\n"));

        if (flags & RSMKA_USE_COOKIE) {
                path = (path_t *)path_cookie;
                mutex_enter(&path->mutex);
        } else {
                path = lookup_path(adapter_name, instance,  remote_node,
                    remote_hwaddr);

                /*
                 * remember, lookup_path increments the reference
                 * count - so decrement now so we can get to zero
                 */
                PATH_RELE_NOLOCK(path);
        }

        DBG_PRINTF((category, RSM_DEBUG,
            "rsmka_remove_path: path = %lx\n", path));

        while (path->state == RSMKA_PATH_GOING_DOWN)
                cv_wait(&path->hold_cv, &path->mutex);

        /* attempt to cancel any possibly pending work */
        mutex_enter(&work_queue.work_mutex);
        if (cancel_work(&path->work_token[RSMKA_IPC_UP_INDEX])) {
                PATH_RELE_NOLOCK(path);
        }
        if (cancel_work(&path->work_token[RSMKA_IPC_DOWN_INDEX])) {
                PATH_RELE_NOLOCK(path);
        }
        mutex_exit(&work_queue.work_mutex);

        /*
         * The path descriptor ref cnt was set to 1 initially when
         * the path was added.  So we need to do a decrement here to
         * balance that.
         */
        PATH_RELE_NOLOCK(path);

        switch (path->state) {
        case RSMKA_PATH_UP:
                /* clear the flag */
                path->flags &= ~RSMKA_SQCREATE_PENDING;
                path->state = RSMKA_PATH_DOWN;
                break;
        case RSMKA_PATH_DOWN:
                break;

        case RSMKA_PATH_ACTIVE:
                /*
                 * rsmka_remove_path should not call do_path_down
                 * with the RSMKA_NO_SLEEP flag set since for
                 * this code path, the deferred work would
                 * incorrectly do a PATH_RELE_NOLOCK.
                 */
                do_path_down(path, 0);
                break;
        default:
                mutex_exit(&path->mutex);
                DBG_PRINTF((category, RSM_ERR,
                    "rsm_remove_path: invalid path state %d\n",
                    path->state));
                return;

        }

        /*
         * wait for all references to the path to be released. If a thread
         * was waiting to receive credits do_path_down should wake it up
         * since the path is going down and that will cause the sleeping
         * thread to release its hold on the path.
         */
        while (path->ref_cnt != 0) {
                cv_wait(&path->hold_cv, &path->mutex);
        }

        mutex_exit(&path->mutex);

        /*
         * remove from ipc token list
         * NOTE: use the remote_node value from the path structure
         * since for RSMKA_USE_COOKIE being set, the remote_node
         * value passed into rsmka_remove_path is 0.
         */
        unlink_sendq_token(&path->sendq_token, path->remote_node);

        /* unlink from adapter path list and free path descriptor */
        destroy_path(path);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "rsmka_remove_path done\n"));
}

/*
 *
 * LOCKING:
 * lookup_path locks the path and increments the path hold count. If the remote
 * node is not in the alive state, do_path_up will release the lock and
 * decrement the hold count.  Otherwise rsmka_do_path_active will release the
 * lock prior to waking up the work thread.
 *
 * REF_CNT:
 * The path descriptor ref_cnt is incremented here; it will be decremented
 * when path up processing is completed in do_path_up or by the work thread
 * if the path up is deferred.
 *
 */
boolean_t
rsmka_path_up(char *adapter_name, uint_t adapter_instance,
    rsm_node_id_t remote_node, rsm_addr_t remote_hwaddr,
    void *path_cookie, int flags)
{

        path_t                  *path;
        boolean_t               rval = B_TRUE;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "rsmka_path_up enter\n"));

        if (flags & RSMKA_USE_COOKIE) {
                path = (path_t *)path_cookie;
                mutex_enter(&path->mutex);
                PATH_HOLD_NOLOCK(path);
        } else {
                path = lookup_path(adapter_name, adapter_instance,
                    remote_node, remote_hwaddr);
        }

        while (path->state == RSMKA_PATH_GOING_DOWN)
                cv_wait(&path->hold_cv, &path->mutex);

        DBG_PRINTF((category, RSM_DEBUG, "rsmka_path_up: path = %lx\n", path));
        rval = do_path_up(path, flags);
        mutex_exit(&path->mutex);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "rsmka_path_up done\n"));
        return (rval);
}

/*
 *
 * LOCKING:
 * lookup_path locks the path and increments the path hold count. If the
 * current state is ACTIVE the path lock is release prior to waking up
 * the work thread in do_path_down .  The work thread will decrement the hold
 * count when the work for this is finished.
 *
 *
 * REF_CNT:
 * The path descriptor ref_cnt is incremented here; it will be decremented
 * when path down processing is completed in do_path_down or by the work thread
 * if the path down is deferred.
 *
 */
boolean_t
rsmka_path_down(char *adapter_devname, int instance, rsm_node_id_t remote_node,
    rsm_addr_t remote_hwaddr,  void *path_cookie, int flags)
{
        path_t                  *path;
        boolean_t               rval = B_TRUE;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "rsmka_path_down enter\n"));

        if (flags & RSMKA_USE_COOKIE) {
                path = (path_t *)path_cookie;
                mutex_enter(&path->mutex);
                PATH_HOLD_NOLOCK(path);
        } else {
                path = lookup_path(adapter_devname, instance, remote_node,
                    remote_hwaddr);
        }

        while (path->state == RSMKA_PATH_GOING_DOWN)
                cv_wait(&path->hold_cv, &path->mutex);

        DBG_PRINTF((category, RSM_DEBUG,
            "rsmka_path_down: path = %lx\n", path));

        switch (path->state) {
        case RSMKA_PATH_UP:
                /* clear the flag */
                path->flags &= ~RSMKA_SQCREATE_PENDING;
                path->state = RSMKA_PATH_GOING_DOWN;
                mutex_exit(&path->mutex);

                /*
                 * release path->mutex since enqueued tasks acquire it.
                 * Drain all the enqueued tasks.
                 */
                taskq_wait(path->recv_taskq);

                mutex_enter(&path->mutex);
                path->state = RSMKA_PATH_DOWN;
                PATH_RELE_NOLOCK(path);
                break;
        case RSMKA_PATH_DOWN:
                PATH_RELE_NOLOCK(path);
                break;
        case RSMKA_PATH_ACTIVE:
                do_path_down(path, flags);
                /*
                 * Need to release the path refcnt. Either done in do_path_down
                 * or do_deferred_work for RSMKA_NO_SLEEP being set. Has to be
                 * done here for RSMKA_NO_SLEEP not set.
                 */
                if (!(flags & RSMKA_NO_SLEEP))
                        PATH_RELE_NOLOCK(path);
                break;
        default:
                DBG_PRINTF((category, RSM_ERR,
                    "rsm_path_down: invalid path state %d\n", path->state));
                rval = B_FALSE;
        }

        mutex_exit(&path->mutex);
        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "rsmka_path_down done\n"));
        return (rval);
}


/*
 * Paths cannot become active until node_is_alive is marked true
 * in the ipc_info descriptor for the node
 *
 * In the event this is called before any paths have been added,
 * init_ipc_info if called here.
 *
 */
boolean_t
rsmka_node_alive(rsm_node_id_t remote_node)
{
        ipc_info_t *ipc_info;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "rsmka_node_alive enter\n"));

        DBG_PRINTF((category, RSM_DEBUG,
            "rsmka_node_alive: remote_node = %x\n", remote_node));

        ipc_info = lookup_ipc_info(remote_node);

        if (ipc_info == NULL) {
                ipc_info = init_ipc_info(remote_node, B_TRUE);
                DBG_PRINTF((category, RSM_DEBUG,
                    "rsmka_node_alive: new ipc_info = %lx\n", ipc_info));
        } else {
                ASSERT(ipc_info->node_is_alive == B_FALSE);
                ipc_info->node_is_alive = B_TRUE;
        }

        pathup_to_pathactive(ipc_info, remote_node);

        mutex_exit(&ipc_info_lock);

        /* rsmipc_send() may be waiting for a sendq_token */
        mutex_enter(&ipc_info_cvlock);
        cv_broadcast(&ipc_info_cv);
        mutex_exit(&ipc_info_cvlock);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "rsmka_node_alive done\n"));

        return (B_TRUE);
}



/*
 * Paths cannot become active when node_is_alive is marked false
 * in the ipc_info descriptor for the node
 */
boolean_t
rsmka_node_died(rsm_node_id_t remote_node)
{
        ipc_info_t *ipc_info;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "rsmka_node_died enter\n"));

        DBG_PRINTF((category, RSM_DEBUG,
            "rsmka_node_died: remote_node = %x\n", remote_node));

        ipc_info = lookup_ipc_info(remote_node);
        if (ipc_info == NULL)
                return (B_FALSE);

        ASSERT(ipc_info->node_is_alive == B_TRUE);
        ipc_info->node_is_alive = B_FALSE;

        rsm_suspend_complete(remote_node, RSM_SUSPEND_NODEDEAD);

        mutex_exit(&ipc_info_lock);

        /* rsmipc_send() may be waiting for a sendq_token */
        mutex_enter(&ipc_info_cvlock);
        cv_broadcast(&ipc_info_cv);
        mutex_exit(&ipc_info_cvlock);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "rsmka_node_died done\n"));

        return (B_TRUE);
}

/*
 * Treat like path_down for all paths for the specified remote node.
 * Always invoked before node died.
 *
 * NOTE: This routine is not called from the cluster path interface; the
 * rsmka_path_down is called directly for each path.
 */
void
rsmka_disconnect_node(rsm_node_id_t remote_node, int flags)
{
        ipc_info_t      *ipc_info;
        path_t          *path;
        sendq_token_t   *sendq_token;
        work_token_t    *up_token;
        work_token_t    *down_token;
        boolean_t       do_work = B_FALSE;
        boolean_t       cancelled = B_FALSE;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsmka_disconnect_node enter\n"));

        DBG_PRINTF((category, RSM_DEBUG,
            "rsmka_disconnect_node: node = %d\n", remote_node));

        if (flags & RSMKA_NO_SLEEP) {
                ipc_info = lookup_ipc_info(remote_node);

                sendq_token = ipc_info->token_list;

                while (sendq_token != NULL) {
                        path = SQ_TOKEN_TO_PATH(sendq_token);
                        PATH_HOLD(path);
                        up_token = &path->work_token[RSMKA_IPC_UP_INDEX];
                        down_token = &path->work_token[RSMKA_IPC_DOWN_INDEX];

                        mutex_enter(&work_queue.work_mutex);

                        /* if an up token is enqueued, remove it */
                        cancelled = cancel_work(up_token);

                        /*
                         * If the path is active and down work hasn't
                         * already been setup then down work is needed.
                         * else
                         * if up work wasn't canceled because it was
                         * already being processed then down work is needed
                         */
                        if (path->state == RSMKA_PATH_ACTIVE) {
                                if (down_token->opcode == 0)
                                        do_work = B_TRUE;
                        } else
                                if (up_token->opcode == RSMKA_IPC_UP)
                                        do_work = B_TRUE;

                        if (do_work == B_TRUE) {
                                down_token->opcode = RSMKA_IPC_DOWN;
                                enqueue_work(down_token);
                        }
                        mutex_exit(&work_queue.work_mutex);

                        if (do_work == B_FALSE)
                                PATH_RELE(path);

                        if (cancelled) {
                                PATH_RELE(path);
                        }
                        sendq_token = sendq_token->next;
                }

                /*
                 * Now that all the work is enqueued, wakeup the work
                 * thread.
                 */
                mutex_enter(&work_queue.work_mutex);
                cv_signal(&work_queue.work_cv);
                mutex_exit(&work_queue.work_mutex);

                IPCINFO_RELE_NOLOCK(ipc_info);
                mutex_exit(&ipc_info_lock);

        } else {
                /* get locked ipc_info descriptor */
                ipc_info = lookup_ipc_info(remote_node);

                sendq_token = ipc_info->token_list;
                while (sendq_token != NULL) {
                        path = SQ_TOKEN_TO_PATH(sendq_token);
                        DBG_PRINTF((category, RSM_DEBUG,
                            "rsmka_disconnect_node: path_down"
                            "for path = %x\n",
                            path));
                        (void) rsmka_path_down(0, 0, 0, 0,
                            path, RSMKA_USE_COOKIE);
                        sendq_token = sendq_token->next;
                        if (sendq_token == ipc_info->token_list)
                                break;
                }
                mutex_exit(&ipc_info_lock);
        }

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsmka_disconnect_node done\n"));
}


/*
 * Called from rsm_node_alive - if a path to a remote node is in
 * state RSMKA_PATH_UP, transition the state to RSMKA_PATH_ACTIVE with a
 * call to rsmka_do_path_active.
 *
 * REF_CNT:
 * The path descriptor ref_cnt is incremented here; it will be decremented
 * when path up processing is completed in rsmka_do_path_active or by the work
 * thread if the path up is deferred.
 */
static void
pathup_to_pathactive(ipc_info_t *ipc_info, rsm_node_id_t remote_node)
{
        path_t          *path;
        sendq_token_t   *token;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "pathup_to_pathactive enter\n"));

        remote_node = remote_node;

        ASSERT(MUTEX_HELD(&ipc_info_lock));

        token = ipc_info->token_list;
        while (token != NULL) {
                path = SQ_TOKEN_TO_PATH(token);
                mutex_enter(&path->mutex);
                if (path->state == RSMKA_PATH_UP)  {
                        PATH_HOLD_NOLOCK(path);
                        (void) rsmka_do_path_active(path, 0);
                }
                mutex_exit(&path->mutex);
                token = token->next;
                if (token == ipc_info->token_list)
                        break;
        }

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "pathup_to_pathactive done\n"));
}

/*
 * Called from pathup_to_pathactive and do_path_up. The objective is to
 * create an ipc send queue and transition to state RSMKA_PATH_ACTIVE.
 * For the no sleep case we may need to defer the work using a token.
 *
 */
boolean_t
rsmka_do_path_active(path_t *path, int flags)
{
        work_token_t    *up_token = &path->work_token[RSMKA_IPC_UP_INDEX];
        work_token_t    *down_token = &path->work_token[RSMKA_IPC_DOWN_INDEX];
        boolean_t       do_work = B_FALSE;
        int             error;
        timespec_t      tv;
        adapter_t       *adapter;
        rsm_send_q_handle_t     sqhdl;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsmka_do_path_active enter\n"));

        ASSERT(MUTEX_HELD(&path->mutex));

        if (flags & RSMKA_NO_SLEEP) {
                mutex_enter(&work_queue.work_mutex);

                /* if a down token is enqueued, remove it */
                if (cancel_work(down_token)) {
                        PATH_RELE_NOLOCK(path);
                }

                /*
                 * If the path is not active and up work hasn't
                 * already been setup then up work is needed.
                 * else
                 * if down work wasn't canceled because it was
                 * already being processed then up work is needed
                 */
                if (path->state != RSMKA_PATH_ACTIVE) {
                        if (up_token->opcode == 0)
                                do_work = B_TRUE;
                } else
                        if (down_token->opcode == RSMKA_IPC_DOWN)
                                do_work = B_TRUE;

                if (do_work == B_TRUE) {
                        up_token->opcode = RSMKA_IPC_UP;
                        enqueue_work(up_token);
                }
                else
                        PATH_RELE_NOLOCK(path);

                mutex_exit(&work_queue.work_mutex);
                DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
                    "rsmka_do_path_active done\n"));
                return (B_TRUE);
        } else {
                /*
                 * Drop the path lock before calling create_ipc_sendq, shouldn't
                 * hold locks across calls to RSMPI routines.
                 */
                mutex_exit(&path->mutex);

                error = create_ipc_sendq(path);

                mutex_enter(&path->mutex);
                if (path->state != RSMKA_PATH_UP) {
                        /*
                         * path state has changed, if sendq was created,
                         * destroy it and return
                         */
                        if (error == RSM_SUCCESS) {
                                sqhdl = path->sendq_token.rsmpi_sendq_handle;
                                path->sendq_token.rsmpi_sendq_handle = NULL;
                                adapter = path->local_adapter;
                                mutex_exit(&path->mutex);

                                if (sqhdl != NULL) {
                                        adapter->rsmpi_ops->rsm_sendq_destroy(
                                            sqhdl);
                                }
                                mutex_enter(&path->mutex);
                        }
                        PATH_RELE_NOLOCK(path);

                        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
                            "rsmka_do_path_active done: path=%lx not UP\n",
                            (uintptr_t)path));
                        return (error ? B_FALSE : B_TRUE);
                }

                if (error == RSM_SUCCESS) {
                        /* clear flag since sendq_create succeeded */
                        path->flags &= ~RSMKA_SQCREATE_PENDING;
                        path->state = RSMKA_PATH_ACTIVE;
                        /*
                         * now that path is active we send the
                         * RSMIPC_MSG_SQREADY to the remote endpoint
                         */
                        path->procmsg_cnt = 0;
                        path->sendq_token.msgbuf_avail = 0;

                        /* Calculate local incarnation number */
                        gethrestime(&tv);
                        if (tv.tv_sec == RSM_UNKNOWN_INCN)
                                tv.tv_sec = 1;
                        path->local_incn = (int64_t)tv.tv_sec;

                        /*
                         * if send fails here its due to some non-transient
                         * error because QUEUE_FULL is not possible here since
                         * we are the first message on this sendq. The error
                         * will cause the path to go down anyways so ignore
                         * the return value
                         */
                        (void) rsmipc_send_controlmsg(path, RSMIPC_MSG_SQREADY);
                        /* wait for SQREADY_ACK message */
                        path->flags |= RSMKA_WAIT_FOR_SQACK;

                        DBG_PRINTF((category, RSM_DEBUG,
                            "rsmka_do_path_active success\n"));
                } else {
                        /*
                         * sendq create failed possibly because
                         * the remote end is not yet ready eg.
                         * handler not registered, set a flag
                         * so that when there is an indication
                         * that the remote end is ready rsmka_do_path_active
                         * will be retried.
                         */
                        path->flags |= RSMKA_SQCREATE_PENDING;
                }

                PATH_RELE_NOLOCK(path);

                DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
                    "rsmka_do_path_active done\n"));
                return (error ? B_FALSE : B_TRUE);
        }

}

/*
 * Called from rsm_path_up.
 * If the remote node state is "alive" then call rsmka_do_path_active
 * otherwise just transition path state to RSMKA_PATH_UP.
 */
static boolean_t
do_path_up(path_t *path, int flags)
{
        boolean_t       rval;
        boolean_t       node_alive;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "do_path_up enter\n"));

        ASSERT(MUTEX_HELD(&path->mutex));

        /* path moved to ACTIVE by rsm_sqcreateop_callback - just return */
        if (path->state == RSMKA_PATH_ACTIVE) {
                DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
                    "do_path_up done: already ACTIVE\n"));
                PATH_RELE_NOLOCK(path);
                return (B_TRUE);
        }

        path->state = RSMKA_PATH_UP;

        /* initialize the receive msgbuf counters */
        path->msgbuf_head = 0;
        path->msgbuf_tail = RSMIPC_MAX_MESSAGES - 1;
        path->msgbuf_cnt = 0;
        path->procmsg_cnt = 0;
        /*
         * rsmka_check_node_alive acquires ipc_info_lock, in order to maintain
         * correct lock ordering drop the path lock before calling it.
         */
        mutex_exit(&path->mutex);

        node_alive = rsmka_check_node_alive(path->remote_node);

        mutex_enter(&path->mutex);
        if (node_alive == B_TRUE)
                rval = rsmka_do_path_active(path, flags);
        else {
                PATH_RELE_NOLOCK(path);
                rval = B_TRUE;
        }

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "do_path_up done\n"));
        return (rval);
}



/*
 * Called from rsm_remove_path, rsm_path_down, deferred_work.
 * Destroy the send queue on this path.
 * Disconnect segments being imported from the remote node
 * Disconnect segments being imported by the remote node
 *
 */
static void
do_path_down(path_t *path, int flags)
{
        work_token_t *up_token = &path->work_token[RSMKA_IPC_UP_INDEX];
        work_token_t *down_token = &path->work_token[RSMKA_IPC_DOWN_INDEX];
        boolean_t do_work = B_FALSE;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "do_path_down enter\n"));

        ASSERT(MUTEX_HELD(&path->mutex));

        if (flags & RSMKA_NO_SLEEP) {
                mutex_enter(&work_queue.work_mutex);
                DBG_PRINTF((category, RSM_DEBUG,
                    "do_path_down: after work_mutex\n"));

                /* if an up token is enqueued, remove it */
                if (cancel_work(up_token)) {
                        PATH_RELE_NOLOCK(path);
                }

                /*
                 * If the path is active and down work hasn't
                 * already been setup then down work is needed.
                 * else
                 * if up work wasn't canceled because it was
                 * already being processed then down work is needed
                 */
                if (path->state == RSMKA_PATH_ACTIVE) {
                        if (down_token->opcode == 0)
                                do_work = B_TRUE;
                } else
                        if (up_token->opcode == RSMKA_IPC_UP)
                                do_work = B_TRUE;

                if (do_work == B_TRUE) {
                        down_token->opcode = RSMKA_IPC_DOWN;
                        enqueue_work(down_token);
                } else
                        PATH_RELE_NOLOCK(path);


                mutex_exit(&work_queue.work_mutex);

        } else {

                /*
                 * Change state of the path to RSMKA_PATH_GOING_DOWN and
                 * release the path mutex. Any other thread referring
                 * this path would cv_wait till the state of the path
                 * remains RSMKA_PATH_GOING_DOWN.
                 * On completing the path down processing, change the
                 * state of RSMKA_PATH_DOWN indicating that the path
                 * is indeed down.
                 */
                path->state = RSMKA_PATH_GOING_DOWN;

                /*
                 * clear the WAIT_FOR_SQACK flag since path is going down.
                 */
                path->flags &= ~RSMKA_WAIT_FOR_SQACK;

                /*
                 * this wakes up any thread waiting to receive credits
                 * in rsmipc_send to tell it that the path is going down
                 */
                cv_broadcast(&path->sendq_token.sendq_cv);

                mutex_exit(&path->mutex);

                /*
                 * drain the messages from the receive msgbuf, the
                 * tasks in the taskq_thread acquire the path->mutex
                 * so we drop the path mutex before taskq_wait.
                 */
                taskq_wait(path->recv_taskq);

                /*
                 * Disconnect segments being imported from the remote node
                 * The path_importer_disconnect function needs to be called
                 * only after releasing the mutex on the path. This is to
                 * avoid a recursive mutex enter when doing the
                 * rsmka_get_sendq_token.
                 */
                path_importer_disconnect(path);

                /*
                 * Get the path mutex, change the state of the path to
                 * RSMKA_PATH_DOWN since the path down processing has
                 * completed and cv_signal anyone who was waiting since
                 * the state was RSMKA_PATH_GOING_DOWN.
                 * NOTE: Do not do a mutex_exit here. We entered this
                 * routine with the path lock held by the caller. The
                 * caller eventually releases the path lock by doing a
                 * mutex_exit.
                 */
                mutex_enter(&path->mutex);

#ifdef DEBUG
                /*
                 * Some IPC messages left in the recv_buf,
                 * they'll be dropped
                 */
                if (path->msgbuf_cnt != 0)
                        cmn_err(CE_NOTE, "path=%lx msgbuf_cnt != 0\n",
                            (uintptr_t)path);
#endif
                while (path->sendq_token.ref_cnt != 0)
                        cv_wait(&path->sendq_token.sendq_cv,
                            &path->mutex);

                /* release the rsmpi handle */
                if (path->sendq_token.rsmpi_sendq_handle != NULL)
                        path->local_adapter->rsmpi_ops->rsm_sendq_destroy(
                            path->sendq_token.rsmpi_sendq_handle);

                path->sendq_token.rsmpi_sendq_handle = NULL;

                path->state = RSMKA_PATH_DOWN;

                cv_signal(&path->hold_cv);

        }

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "do_path_down done\n"));

}

/*
 * Search through the list of imported segments for segments using this path
 * and unload the memory mappings for each one.  The application will
 * get an error return when a barrier close is invoked.
 * NOTE: This function has to be called only after releasing the mutex on
 * the path. This is to avoid any recursive mutex panics on the path mutex
 * since the path_importer_disconnect function would end up calling
 * rsmka_get_sendq_token which requires the path mutex.
 */

static void
path_importer_disconnect(path_t *path)
{
        int i;
        adapter_t *adapter = path->local_adapter;
        rsm_node_id_t remote_node = path->remote_node;
        rsmresource_t           *p = NULL;
        rsmseg_t *seg;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "path_importer_disconnect enter\n"));

        rw_enter(&rsm_import_segs.rsmhash_rw, RW_READER);

        if (rsm_import_segs.bucket != NULL) {
                for (i = 0; i < rsm_hash_size; i++) {
                        p = rsm_import_segs.bucket[i];
                        for (; p; p = p->rsmrc_next) {
                                if ((p->rsmrc_node == remote_node) &&
                                    (p->rsmrc_adapter == adapter)) {
                                        seg = (rsmseg_t *)p;
                        /*
                         * In order to make rsmseg_unload and
                         * path_importer_disconnect thread safe, acquire the
                         * segment lock here. rsmseg_unload is responsible for
                         * releasing the lock. rsmseg_unload releases the lock
                         * just before a call to rsmipc_send or in case of an
                         * early exit which occurs if the segment was in the
                         * state RSM_STATE_CONNECTING or RSM_STATE_NEW.
                         */
                                        rsmseglock_acquire(seg);
                                        seg->s_flags |= RSM_FORCE_DISCONNECT;
                                        rsmseg_unload(seg);
                                }
                        }
                }
        }
        rw_exit(&rsm_import_segs.rsmhash_rw);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "path_importer_disconnect done\n"));
}




/*
 *
 * ADAPTER UTILITY FUNCTIONS
 *
 */



/*
 * Allocate new adapter list head structure and add it to the beginning of
 * the list of adapter list heads.  There is one list for each adapter
 * device name (or type).
 */
static adapter_listhead_t *
init_listhead(char *name)
{
        adapter_listhead_t *listhead;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "init_listhead enter\n"));

        /* allocation and initialization */
        listhead = kmem_zalloc(sizeof (adapter_listhead_t), KM_SLEEP);
        mutex_init(&listhead->mutex, NULL, MUTEX_DEFAULT, NULL);
        (void) strcpy(listhead->adapter_devname, name);

        /* link into list of listheads */
        mutex_enter(&adapter_listhead_base.listlock);
        if (adapter_listhead_base.next == NULL) {
                adapter_listhead_base.next = listhead;
                listhead->next_listhead = NULL;
        } else {
                listhead->next_listhead = adapter_listhead_base.next;
                adapter_listhead_base.next = listhead;
        }
        mutex_exit(&adapter_listhead_base.listlock);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "init_listhead done\n"));

        return (listhead);
}


/*
 * Search the list of adapter list heads for a match on name.
 *
 */
static adapter_listhead_t *
lookup_adapter_listhead(char *name)
{
        adapter_listhead_t *listhead;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "lookup_adapter_listhead enter\n"));

        mutex_enter(&adapter_listhead_base.listlock);
        listhead = adapter_listhead_base.next;
        while (listhead != NULL) {
                if (strcmp(name, listhead->adapter_devname) == 0)
                        break;
                listhead = listhead->next_listhead;
        }
        mutex_exit(&adapter_listhead_base.listlock);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "lookup_adapter_listhead done\n"));

        return (listhead);
}


/*
 * Get the adapter list head corresponding to devname and search for
 * an adapter descriptor with a match on the instance number. If
 * successful, increment the descriptor reference count and return
 * the descriptor pointer to the caller.
 *
 */
adapter_t *
rsmka_lookup_adapter(char *devname, int instance)
{
        adapter_listhead_t *listhead;
        adapter_t *current = NULL;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsmka_lookup_adapter enter\n"));

        listhead = lookup_adapter_listhead(devname);
        if (listhead != NULL) {
                mutex_enter(&listhead->mutex);

                current = listhead->next_adapter;
                while (current != NULL) {
                        if (current->instance == instance) {
                                ADAPTER_HOLD(current);
                                break;
                        } else
                                current = current->next;
                }

                mutex_exit(&listhead->mutex);
        }

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsmka_lookup_adapter done\n"));

        return (current);
}

/*
 * Called from rsmka_remove_adapter or rsmseg_free.
 * rsm_bind() and rsm_connect() store the adapter pointer returned
 * from rsmka_getadapter.  The pointer is kept in the segment descriptor.
 * When the segment is freed, this routine is called by rsmseg_free to decrement
 * the adapter descriptor reference count and possibly free the
 * descriptor.
 */
void
rsmka_release_adapter(adapter_t *adapter)
{
        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsmka_release_adapter enter\n"));

        if (adapter == &loopback_adapter) {
                DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
                    "rsmka_release_adapter done\n"));
                return;
        }

        mutex_enter(&adapter->mutex);

        /* decrement reference count */
        ADAPTER_RELE_NOLOCK(adapter);

        /*
         * if the adapter descriptor reference count is equal to the
         * initialization value of one, then the descriptor has been
         * unlinked and can now be freed.
         */
        if (adapter->ref_cnt == 1) {
                mutex_exit(&adapter->mutex);

                mutex_destroy(&adapter->mutex);
                kmem_free(adapter->hdlr_argp, sizeof (srv_handler_arg_t));
                kmem_free(adapter, sizeof (adapter_t));
        }
        else
                mutex_exit(&adapter->mutex);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsmka_release_adapter done\n"));

}



/*
 * Singly linked list. Add to the front.
 */
static void
link_adapter(adapter_t *adapter)
{

        adapter_listhead_t *listhead;
        adapter_t *current;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "link_adapter enter\n"));

        mutex_enter(&adapter_listhead_base.listlock);

        mutex_enter(&adapter->listhead->mutex);

        listhead = adapter->listhead;
        current = listhead->next_adapter;
        listhead->next_adapter = adapter;
        adapter->next = current;
        ADAPTER_HOLD(adapter);

        adapter->listhead->adapter_count++;

        mutex_exit(&adapter->listhead->mutex);

        mutex_exit(&adapter_listhead_base.listlock);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "link_adapter done\n"));
}


/*
 * Return adapter descriptor
 *
 * lookup_adapter_listhead returns with the the list of adapter listheads
 * locked.  After adding the adapter descriptor, the adapter listhead list
 * lock is dropped.
 */
static adapter_t *
init_adapter(char *name, int instance, rsm_addr_t hwaddr,
    rsm_controller_handle_t handle, rsm_ops_t *ops,
    srv_handler_arg_t *hdlr_argp)
{
        adapter_t *adapter;
        adapter_listhead_t *listhead;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "init_adapter enter\n"));

        adapter = kmem_zalloc(sizeof (adapter_t), KM_SLEEP);
        adapter->instance = instance;
        adapter->hwaddr = hwaddr;
        adapter->rsmpi_handle = handle;
        adapter->rsmpi_ops = ops;
        adapter->hdlr_argp = hdlr_argp;
        mutex_init(&adapter->mutex, NULL, MUTEX_DEFAULT, NULL);
        ADAPTER_HOLD(adapter);


        listhead = lookup_adapter_listhead(name);
        if (listhead == NULL)  {
                listhead = init_listhead(name);
        }

        adapter->listhead = listhead;

        link_adapter(adapter);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "init_adapter done\n"));

        return (adapter);
}

/*
 *
 * PATH UTILITY FUNCTIONS
 *
 */


/*
 * Search the per adapter path list for a match on remote node and
 * hwaddr.  The path ref_cnt must be greater than zero or the path
 * is in the process of being removed.
 *
 * Acquire the path lock and increment the path hold count.
 */
static path_t *
lookup_path(char *adapter_devname, int adapter_instance,
    rsm_node_id_t remote_node, rsm_addr_t hwaddr)
{
        path_t          *current;
        adapter_t       *adapter;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "lookup_path enter\n"));

        adapter = rsmka_lookup_adapter(adapter_devname, adapter_instance);
        ASSERT(adapter != NULL);

        mutex_enter(&adapter->listhead->mutex);

        /* start at the list head */
        current = adapter->next_path;

        while (current != NULL) {
                if ((current->remote_node == remote_node) &&
                    (current->remote_hwaddr == hwaddr) &&
                    (current->ref_cnt > 0))
                        break;
                else
                        current = current->next_path;
        }
        if (current != NULL) {
                mutex_enter(&current->mutex);
                PATH_HOLD_NOLOCK(current);
        }

        mutex_exit(&adapter->listhead->mutex);
        ADAPTER_RELE(adapter);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "lookup_path done\n"));

        return (current);
}

/*
 * This interface is similar to lookup_path but takes only the local
 * adapter name, instance and remote adapters hwaddr to identify the
 * path. This is used in the interrupt handler routines where nodeid
 * is not always available.
 */
path_t *
rsm_find_path(char *adapter_devname, int adapter_instance, rsm_addr_t hwaddr)
{
        path_t          *current;
        adapter_t       *adapter;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "rsm_find_path enter\n"));

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsm_find_path:adapter=%s:%d,rem=%llx\n",
            adapter_devname, adapter_instance, hwaddr));

        adapter = rsmka_lookup_adapter(adapter_devname, adapter_instance);

        /*
         * its possible that we are here due to an interrupt but the adapter
         * has been removed after we received the callback.
         */
        if (adapter == NULL)
                return (NULL);

        mutex_enter(&adapter->listhead->mutex);

        /* start at the list head */
        current = adapter->next_path;

        while (current != NULL) {
                if ((current->remote_hwaddr == hwaddr) &&
                    (current->ref_cnt > 0))
                        break;
                else
                        current = current->next_path;
        }
        if (current != NULL) {
                mutex_enter(&current->mutex);
                PATH_HOLD_NOLOCK(current);
        }

        mutex_exit(&adapter->listhead->mutex);

        rsmka_release_adapter(adapter);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "rsm_find_path done\n"));

        return (current);
}


/*
 * Add the path to the head of the (per adapter) list of paths
 */
static void
link_path(path_t *path)
{

        adapter_t *adapter = path->local_adapter;
        path_t *first_path;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "link_path enter\n"));

        mutex_enter(&adapter_listhead_base.listlock);

        mutex_enter(&adapter->listhead->mutex);

        first_path = adapter->next_path;
        adapter->next_path = path;
        path->next_path = first_path;

        adapter->listhead->path_count++;

        mutex_exit(&adapter->listhead->mutex);

        mutex_exit(&adapter_listhead_base.listlock);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "link_path done\n"));
}

/*
 * Search the per-adapter list of paths for the specified path, beginning
 * at the head of the list.  Unlink the path and free the descriptor
 * memory.
 */
static void
destroy_path(path_t *path)
{

        adapter_t *adapter = path->local_adapter;
        path_t *prev, *current;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "destroy_path enter\n"));

        mutex_enter(&adapter_listhead_base.listlock);

        mutex_enter(&path->local_adapter->listhead->mutex);
        ASSERT(path->ref_cnt == 0);

        /* start at the list head */
        prev = NULL;
        current =  adapter->next_path;

        while (current != NULL) {
                if (path->remote_node == current->remote_node &&
                    path->remote_hwaddr == current->remote_hwaddr)
                        break;
                else {
                        prev = current;
                        current = current->next_path;
                }
        }

        if (prev == NULL)
                adapter->next_path = current->next_path;
        else
                prev->next_path = current->next_path;

        path->local_adapter->listhead->path_count--;

        mutex_exit(&path->local_adapter->listhead->mutex);

        mutex_exit(&adapter_listhead_base.listlock);

        taskq_destroy(path->recv_taskq);

        kmem_free(path->msgbuf_queue,
            RSMIPC_MAX_MESSAGES * sizeof (msgbuf_elem_t));

        mutex_destroy(&current->mutex);
        cv_destroy(&current->sendq_token.sendq_cv);
        cv_destroy(&path->hold_cv);
        kmem_free(current, sizeof (path_t));

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "destroy_path done\n"));
}

void
rsmka_enqueue_msgbuf(path_t *path, void *data)
{
        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsmka_enqueue_msgbuf enter\n"));

        ASSERT(MUTEX_HELD(&path->mutex));

        ASSERT(path->msgbuf_cnt < RSMIPC_MAX_MESSAGES);

        /* increment the count and advance the tail */

        path->msgbuf_cnt++;

        if (path->msgbuf_tail == RSMIPC_MAX_MESSAGES - 1) {
                path->msgbuf_tail = 0;
        } else {
                path->msgbuf_tail++;
        }

        path->msgbuf_queue[path->msgbuf_tail].active = B_TRUE;

        bcopy(data, &(path->msgbuf_queue[path->msgbuf_tail].msg),
            sizeof (rsmipc_request_t));

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsmka_enqueue_msgbuf done\n"));

}

/*
 * get the head of the queue using rsmka_gethead_msgbuf and then call
 * rsmka_dequeue_msgbuf to remove it.
 */
void
rsmka_dequeue_msgbuf(path_t *path)
{
        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsmka_dequeue_msgbuf enter\n"));

        ASSERT(MUTEX_HELD(&path->mutex));

        if (path->msgbuf_cnt == 0)
                return;

        path->msgbuf_cnt--;

        path->msgbuf_queue[path->msgbuf_head].active = B_FALSE;

        if (path->msgbuf_head == RSMIPC_MAX_MESSAGES - 1) {
                path->msgbuf_head = 0;
        } else {
                path->msgbuf_head++;
        }

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsmka_dequeue_msgbuf done\n"));

}

msgbuf_elem_t *
rsmka_gethead_msgbuf(path_t *path)
{
        msgbuf_elem_t   *head;

        ASSERT(MUTEX_HELD(&path->mutex));

        if (path->msgbuf_cnt == 0)
                return (NULL);

        head = &path->msgbuf_queue[path->msgbuf_head];

        return (head);

}
/*
 * Called by rsm_connect which needs the hardware address of the
 * remote adapter.  A search is done through the paths for the local
 * adapter for a match on the specified remote node.
 */
rsm_node_id_t
get_remote_nodeid(adapter_t *adapter, rsm_addr_t remote_hwaddr)
{

        rsm_node_id_t remote_node;
        path_t     *current = adapter->next_path;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "get_remote_nodeid enter\n"));

        mutex_enter(&adapter->listhead->mutex);
        while (current != NULL) {
                if (current->remote_hwaddr == remote_hwaddr) {
                        remote_node = current->remote_node;
                        break;
                }
                current = current->next_path;
        }

        if (current == NULL)
                remote_node = (rsm_node_id_t)-1;

        mutex_exit(&adapter->listhead->mutex);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "get_remote_nodeid done\n"));

        return (remote_node);
}

/*
 * Called by rsm_connect which needs the hardware address of the
 * remote adapter.  A search is done through the paths for the local
 * adapter for a match on the specified remote node.
 */
rsm_addr_t
get_remote_hwaddr(adapter_t *adapter, rsm_node_id_t remote_node)
{

        rsm_addr_t remote_hwaddr;
        path_t     *current = adapter->next_path;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "get_remote_hwaddr enter\n"));

        mutex_enter(&adapter->listhead->mutex);
        while (current != NULL) {
                if (current->remote_node == remote_node) {
                        remote_hwaddr = current->remote_hwaddr;
                        break;
                }
                current = current->next_path;
        }
        if (current == NULL)
                remote_hwaddr = -1;
        mutex_exit(&adapter->listhead->mutex);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "get_remote_hwaddr done\n"));

        return (remote_hwaddr);
}
/*
 * IPC UTILITY FUNCTIONS
 */


/*
 * If an entry exists, return with the ipc_info_lock held
 */
static ipc_info_t *
lookup_ipc_info(rsm_node_id_t remote_node)
{
        ipc_info_t  *ipc_info;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "lookup_ipc_info enter\n"));

        mutex_enter(&ipc_info_lock);

        ipc_info = ipc_info_head;
        if (ipc_info == NULL) {
                mutex_exit(&ipc_info_lock);
                DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
                    "lookup_ipc_info done: ipc_info is NULL\n"));
                return (NULL);
        }

        while (ipc_info->remote_node != remote_node) {
                ipc_info = ipc_info->next;
                if (ipc_info == NULL) {
                        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
                            "lookup_ipc_info: ipc_info not found\n"));
                        mutex_exit(&ipc_info_lock);
                        break;
                }
        }

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "lookup_ipc_info done\n"));

        return (ipc_info);
}

/*
 * Create an ipc_info descriptor and return with ipc_info_lock held
 */
static ipc_info_t *
init_ipc_info(rsm_node_id_t remote_node, boolean_t state)
{
        ipc_info_t *ipc_info;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "init_ipc_info enter\n"));

        /*
         * allocate an ipc_info descriptor and add it to a
         * singly linked list
         */

        ipc_info = kmem_zalloc(sizeof (ipc_info_t), KM_SLEEP);
        ipc_info->remote_node = remote_node;
        ipc_info->node_is_alive = state;

        mutex_enter(&ipc_info_lock);
        if (ipc_info_head == NULL) {
                DBG_PRINTF((category, RSM_DEBUG,
                    "init_ipc_info:ipc_info_head = %lx\n", ipc_info));
                ipc_info_head = ipc_info;
                ipc_info->next = NULL;
        } else {
                ipc_info->next = ipc_info_head;
                ipc_info_head = ipc_info;
        }

        ipc_info->remote_node = remote_node;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "init_ipc_info done\n"));

        return (ipc_info);
}

static void
destroy_ipc_info(ipc_info_t *ipc_info)
{
        ipc_info_t *current = ipc_info_head;
        ipc_info_t *prev;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "destroy_ipc_info enter\n"));

        ASSERT(MUTEX_HELD(&ipc_info_lock));

        while (current != ipc_info) {
                prev = current;
                current = current->next;
        }
        ASSERT(current != NULL);

        if (current != ipc_info_head)
                prev->next = current->next;
        else
                ipc_info_head = current->next;

        kmem_free(current, sizeof (ipc_info_t));

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "destroy_ipc_info done\n"));

}

/*
 * Sendq tokens are kept on a circular list.  If tokens A, B, C, & D are
 * on the list headed by ipc_info, then ipc_info points to A, A points to
 * D, D to C, C to B, and B to A.
 */
static void
link_sendq_token(sendq_token_t *token, rsm_node_id_t remote_node)
{
        ipc_info_t *ipc_info;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "link_sendq_token enter\n"));

        ipc_info = lookup_ipc_info(remote_node);
        if (ipc_info == NULL) {
                ipc_info = init_ipc_info(remote_node, B_FALSE);
                DBG_PRINTF((category, RSM_DEBUG,
                    "link_sendq_token: new ipc_info = %lx\n", ipc_info));
        }
        else
                DBG_PRINTF((category, RSM_DEBUG,
                    "link_sendq_token: ipc_info = %lx\n", ipc_info));

        if (ipc_info->token_list == NULL) {
                ipc_info->token_list = token;
                ipc_info->current_token = token;
                DBG_PRINTF((category, RSM_DEBUG,
                    "link_sendq_token: current = %lx\n", token));
                token->next = token;
        } else {
                DBG_PRINTF((category, RSM_DEBUG,
                    "link_sendq_token: token = %lx\n", token));
                token->next = ipc_info->token_list->next;
                ipc_info->token_list->next = token;
                ipc_info->token_list = token;
        }


        mutex_exit(&ipc_info_lock);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "link_sendq_token done\n"));

}

static void
unlink_sendq_token(sendq_token_t *token, rsm_node_id_t remote_node)
{
        sendq_token_t *prev, *start,  *current;
        ipc_info_t *ipc_info;
        path_t *path = SQ_TOKEN_TO_PATH(token);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "unlink_sendq_token enter\n"));

        ASSERT(path->ref_cnt == 0);

        ipc_info = lookup_ipc_info(remote_node);
        if (ipc_info == NULL) {
                DBG_PRINTF((category, RSM_DEBUG,
                    "ipc_info for %d not found\n", remote_node));
                return;
        }

        prev = ipc_info->token_list;
        start = current = ipc_info->token_list->next;

        for (;;) {
                if (current == token) {
                        if (current->next != current) {
                                DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
                                    "found token, removed it\n"));
                                prev->next = token->next;
                                if (ipc_info->token_list == token)
                                        ipc_info->token_list = prev;
                                ipc_info->current_token = token->next;
                        } else {
                                /* list will be empty  */
                                DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
                                    "removed token, list empty\n"));
                                ipc_info->token_list = NULL;
                                ipc_info->current_token = NULL;
                        }
                        break;
                }
                prev = current;
                current = current->next;
                if (current == start) {
                        DBG_PRINTF((category, RSM_DEBUG,
                            "unlink_sendq_token: token not found\n"));
                        break;
                }
        }
        mutex_exit(&ipc_info_lock);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "unlink_sendq_token done\n"));
}


void
rele_sendq_token(sendq_token_t *token)
{
        path_t *path;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "rele_sendq_token enter\n"));

        path = SQ_TOKEN_TO_PATH(token);
        mutex_enter(&path->mutex);
        PATH_RELE_NOLOCK(path);
        SENDQ_TOKEN_RELE(path);
        mutex_exit(&path->mutex);

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "rele_sendq_token done\n"));

}

/*
 * A valid ipc token can only be returned if the remote node is alive.
 * Tokens are on a circular list.  Starting with the current token
 * search for a token with an endpoint in state RSM_PATH_ACTIVE.
 * rsmipc_send which calls rsmka_get_sendq_token expects that if there are
 * multiple paths available between a node-pair then consecutive calls from
 * a particular invocation of rsmipc_send will return a sendq that is
 * different from the one that was used in the previous iteration. When
 * prev_used is NULL it indicates that this is the first interation in a
 * specific rsmipc_send invocation.
 *
 * Updating the current token provides round robin selection and this
 * is done only in the first iteration ie. when prev_used is NULL
 */
sendq_token_t *
rsmka_get_sendq_token(rsm_node_id_t remote_node, sendq_token_t *prev_used)
{
        sendq_token_t *token, *first_token;
        path_t *path;
        ipc_info_t *ipc_info;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
            "rsmka_get_sendq_token enter\n"));

        ipc_info = lookup_ipc_info(remote_node);
        if (ipc_info == NULL) {
                DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
                    "rsmka_get_sendq_token done: ipc_info is NULL\n"));
                return (NULL);
        }

        if (ipc_info->node_is_alive == B_TRUE) {
                token = first_token = ipc_info->current_token;
                if (token == NULL) {
                        mutex_exit(&ipc_info_lock);
                        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
                            "rsmka_get_sendq_token done: token=NULL\n"));
                        return (NULL);
                }

                for (;;) {
                        path = SQ_TOKEN_TO_PATH(token);
                        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
                            "path %lx\n", path));
                        mutex_enter(&path->mutex);
                        if (path->state != RSMKA_PATH_ACTIVE ||
                            path->ref_cnt == 0) {
                                mutex_exit(&path->mutex);
                        } else {
                                if (token != prev_used) {
                                        /* found a new token */
                                        break;
                                }
                                mutex_exit(&path->mutex);
                        }

                        token = token->next;
                        if (token == first_token) {
                                /*
                                 * we didn't find a new token reuse prev_used
                                 * if the corresponding path is still up
                                 */
                                if (prev_used) {
                                        path = SQ_TOKEN_TO_PATH(prev_used);
                                        DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
                                            "path %lx\n", path));
                                        mutex_enter(&path->mutex);
                                        if (path->state != RSMKA_PATH_ACTIVE ||
                                            path->ref_cnt == 0) {
                                                mutex_exit(&path->mutex);
                                        } else {
                                                token = prev_used;
                                                break;
                                        }
                                }
                                mutex_exit(&ipc_info_lock);
                                DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
                                    "rsmka_get_sendq_token: token=NULL\n"));
                                return (NULL);
                        }
                }

                PATH_HOLD_NOLOCK(path);
                SENDQ_TOKEN_HOLD(path);
                if (prev_used == NULL) {
                        /* change current_token only the first time */
                        ipc_info->current_token = token->next;
                }

                mutex_exit(&path->mutex);
                mutex_exit(&ipc_info_lock);

                DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
                    "rsmka_get_sendq_token done\n"));
                return (token);
        } else {
                mutex_exit(&ipc_info_lock);
                DBG_PRINTF((category, RSM_DEBUG_VERBOSE,
                    "rsmka_get_sendq_token done\n"));
                return (NULL);
        }
}



/*
 */
static int
create_ipc_sendq(path_t *path)
{
        int             rval;
        sendq_token_t   *token;
        adapter_t       *adapter;
        int64_t         srvc_offset;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "create_ipc_sendq enter\n"));

        DBG_PRINTF((category, RSM_DEBUG, "create_ipc_sendq: path = %lx\n",
            path));

        adapter = path->local_adapter;
        token = &path->sendq_token;

        srvc_offset = path->remote_hwaddr;

        DBG_PRINTF((category, RSM_DEBUG,
            "create_ipc_sendq: srvc_offset = %lld\n",
            srvc_offset));

        rval = adapter->rsmpi_ops->rsm_sendq_create(adapter->rsmpi_handle,
            path->remote_hwaddr,
            (rsm_intr_service_t)(RSM_SERVICE+srvc_offset),
            (rsm_intr_pri_t)RSM_PRI, (size_t)RSM_QUEUE_SZ,
            RSM_INTR_SEND_Q_NO_FENCE,
            RSM_RESOURCE_SLEEP, NULL, &token->rsmpi_sendq_handle);
        if (rval == RSM_SUCCESS) {
                /* rsmipc_send() may be waiting for a sendq_token */
                mutex_enter(&ipc_info_cvlock);
                cv_broadcast(&ipc_info_cv);
                mutex_exit(&ipc_info_cvlock);
        }

        DBG_PRINTF((category, RSM_DEBUG, "create_ipc_sendq: handle = %lx\n",
            token->rsmpi_sendq_handle));
        DBG_PRINTF((category, RSM_DEBUG, "create_ipc_sendq: rval = %d\n",
            rval));
        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "create_ipc_sendq done\n"));

        return (rval);
}


boolean_t
rsmka_check_node_alive(rsm_node_id_t remote_node)
{
        ipc_info_t *ipc_info;

        DBG_PRINTF((category, RSM_DEBUG, "rsmka_check_node_alive enter\n"));

        ipc_info = lookup_ipc_info(remote_node);
        if (ipc_info == NULL) {
                DBG_PRINTF((category, RSM_DEBUG,
                    "rsmka_check_node_alive done: ipc_info NULL\n"));
                return (B_FALSE);
        }

        if (ipc_info->node_is_alive == B_TRUE) {
                mutex_exit(&ipc_info_lock);
                DBG_PRINTF((category, RSM_DEBUG,
                    "rsmka_check_node_alive done: node is alive\n"));
                return (B_TRUE);
        } else {
                mutex_exit(&ipc_info_lock);
                DBG_PRINTF((category, RSM_DEBUG,
                    "rsmka_check_node_alive done: node is not alive\n"));
                return (B_FALSE);
        }
}




/*
 *  TOPOLOGY IOCTL SUPPORT
 */

static uint32_t
get_topology_size(int mode)
{
        uint32_t        topology_size;
        int             pointer_area_size;
        adapter_listhead_t      *listhead;
        int             total_num_of_adapters;
        int             total_num_of_paths;

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "get_topology_size enter\n"));

        /*
         * Find the total number of adapters and paths by adding up the
         * individual adapter and path counts from all the listheads
         */
        total_num_of_adapters = 0;
        total_num_of_paths = 0;
        listhead = adapter_listhead_base.next;
        while (listhead != NULL) {
                total_num_of_adapters += listhead->adapter_count;
                total_num_of_paths += listhead->path_count;
                listhead = listhead->next_listhead;
        }

#ifdef  _MULTI_DATAMODEL
        if ((mode & DATAMODEL_MASK) == DATAMODEL_ILP32)
                /*
                 * Add extra 4-bytes to make sure connections header
                 * is double-word aligned
                 */
                pointer_area_size =
                    (total_num_of_adapters + total_num_of_adapters%2) *
                    sizeof (caddr32_t);
        else
                pointer_area_size = total_num_of_adapters * sizeof (caddr_t);
#else   /* _MULTI_DATAMODEL */
        mode = mode;
        pointer_area_size = total_num_of_adapters * sizeof (caddr_t);
#endif  /* _MULTI_DATAMODEL */


        topology_size = sizeof (rsmka_topology_hdr_t) +
            pointer_area_size +
            (total_num_of_adapters * sizeof (rsmka_connections_hdr_t)) +
            (total_num_of_paths * sizeof (rsmka_remote_cntlr_t));

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "get_topology_size done\n"));

        return (topology_size);
}



static void
get_topology(caddr_t arg, char *bufp, int mode)
{

        rsmka_topology_t        *tp = (rsmka_topology_t *)bufp;
        adapter_listhead_t      *listhead;
        adapter_t               *adapter;
        path_t                  *path;
        int                     cntlr = 0;
        rsmka_connections_t     *connection;
        rsmka_remote_cntlr_t    *rem_cntlr;
        int                     total_num_of_adapters;

#ifdef  _MULTI_DATAMODEL
        rsmka_topology32_t      *tp32 = (rsmka_topology32_t *)bufp;
#else
        mode = mode;
#endif  /* _MULTI_DATAMODEL */

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "get_topology enter\n"));

        /*
         * Find the total number of adapters by adding up the
         * individual adapter counts from all the listheads
         */
        total_num_of_adapters = 0;
        listhead = adapter_listhead_base.next;
        while (listhead != NULL) {
                total_num_of_adapters += listhead->adapter_count;
                listhead = listhead->next_listhead;
        }

        /* fill topology header and adjust bufp */
        tp->topology_hdr.local_nodeid = my_nodeid;
        tp->topology_hdr.local_cntlr_count = total_num_of_adapters;
        bufp = (char *)&tp->connections[0];

        /* leave room for connection pointer area */
#ifdef  _MULTI_DATAMODEL
        if ((mode & DATAMODEL_MASK) == DATAMODEL_ILP32)
                /* make sure bufp is double-word aligned */
                bufp += (total_num_of_adapters + total_num_of_adapters%2) *
                    sizeof (caddr32_t);
        else
                bufp += total_num_of_adapters * sizeof (caddr_t);
#else   /* _MULTI_DATAMODEL */
        bufp += total_num_of_adapters * sizeof (caddr_t);
#endif  /* _MULTI_DATAMODEL */

        /* fill topology from the adapter and path data */
        listhead = adapter_listhead_base.next;
        while (listhead != NULL) {
                adapter = listhead->next_adapter;
                while (adapter != NULL) {
                        /* fill in user based connection pointer */
#ifdef  _MULTI_DATAMODEL
                        if ((mode & DATAMODEL_MASK) == DATAMODEL_ILP32) {
                                ulong_t delta = (ulong_t)bufp - (ulong_t)tp32;
                                caddr32_t userbase = (caddr32_t)((ulong_t)arg &
                                    0xffffffff);
                                tp32->connections[cntlr++] = userbase + delta;
                        } else {
                                tp->connections[cntlr++] = arg +
                                    (ulong_t)bufp -
                                    (ulong_t)tp;
                        }
#else   /* _MULTI_DATAMODEL */
                                tp->connections[cntlr++] = arg +
                                    (ulong_t)bufp -
                                    (ulong_t)tp;
#endif  /* _MULTI_DATAMODEL */
                        connection = (rsmka_connections_t *)bufp;
                        (void) snprintf(connection->hdr.cntlr_name,
                            MAXNAMELEN, "%s%d",
                            listhead->adapter_devname,
                            adapter->instance);
                        connection->hdr.local_hwaddr = adapter->hwaddr;
                        connection->hdr.remote_cntlr_count = 0;
                        bufp += sizeof (rsmka_connections_hdr_t);
                        rem_cntlr = (rsmka_remote_cntlr_t *)bufp;
                        path = adapter->next_path;
                        while (path != NULL) {
                                connection->hdr.remote_cntlr_count++;
                                rem_cntlr->remote_nodeid = path->remote_node;
                                (void) snprintf(rem_cntlr->remote_cntlrname,
                                    MAXNAMELEN, "%s%d",
                                    listhead->adapter_devname,
                                    path->remote_devinst);
                                rem_cntlr->remote_hwaddr = path->remote_hwaddr;
                                rem_cntlr->connection_state = path->state;
                                ++rem_cntlr;
                                path = path->next_path;
                        }
                        adapter = adapter->next;
                        bufp = (char *)rem_cntlr;
                }
                listhead = listhead->next_listhead;
        }

        DBG_PRINTF((category, RSM_DEBUG_VERBOSE, "get_topology done\n"));

}


/*
 * Called from rsm_ioctl() in rsm.c
 * Make sure there is no possiblity of blocking while holding
 * adapter_listhead_base.lock
 */
int
rsmka_topology_ioctl(caddr_t arg, int cmd, int mode)
{
        uint32_t        topology_size;
        uint32_t        request_size;
        char            *bufp;
        int             error = RSM_SUCCESS;
        size_t          max_toposize;

        DBG_PRINTF((category | RSM_IOCTL, RSM_DEBUG_VERBOSE,
            "rsmka_topology_ioctl enter\n"));

        switch (cmd) {
        case RSM_IOCTL_TOPOLOGY_SIZE:
                mutex_enter(&adapter_listhead_base.listlock);
                topology_size = get_topology_size(mode);
                mutex_exit(&adapter_listhead_base.listlock);
                if (ddi_copyout((caddr_t)&topology_size,
                    (caddr_t)arg, sizeof (uint32_t), mode))
                        error = RSMERR_BAD_ADDR;
                break;
        case RSM_IOCTL_TOPOLOGY_DATA:
                /*
                 * The size of the buffer which the caller has allocated
                 * is passed in.  If the size needed for the topology data
                 * is not sufficient, E2BIG is returned
                 */
                if (ddi_copyin(arg, &request_size, sizeof (uint32_t), mode)) {
                        DBG_PRINTF((category | RSM_IOCTL, RSM_DEBUG_VERBOSE,
                            "rsmka_topology_ioctl done: BAD_ADDR\n"));
                        return (RSMERR_BAD_ADDR);
                }
                /* calculate the max size of the topology structure */
                max_toposize = sizeof (rsmka_topology_hdr_t) +
                    RSM_MAX_CTRL * (sizeof (caddr_t) +
                    sizeof (rsmka_connections_hdr_t)) +
                    RSM_MAX_NODE * sizeof (rsmka_remote_cntlr_t);

                if (request_size > max_toposize) { /* validate request_size */
                        DBG_PRINTF((category | RSM_IOCTL, RSM_DEBUG_VERBOSE,
                            "rsmka_topology_ioctl done: size too large\n"));
                        return (EINVAL);
                }
                bufp = kmem_zalloc(request_size, KM_SLEEP);
                mutex_enter(&adapter_listhead_base.listlock);
                topology_size = get_topology_size(mode);
                if (request_size < topology_size) {
                        kmem_free(bufp, request_size);
                        mutex_exit(&adapter_listhead_base.listlock);
                        DBG_PRINTF((category | RSM_IOCTL, RSM_DEBUG_VERBOSE,
                            "rsmka_topology_ioctl done: E2BIG\n"));
                        return (E2BIG);
                }

                /* get the topology data and copyout to the caller */
                get_topology(arg, bufp, mode);
                mutex_exit(&adapter_listhead_base.listlock);
                if (ddi_copyout((caddr_t)bufp, (caddr_t)arg,
                    topology_size, mode))
                        error = RSMERR_BAD_ADDR;

                kmem_free(bufp, request_size);
                break;
        default:
                DBG_PRINTF((category | RSM_IOCTL, RSM_DEBUG,
                    "rsmka_topology_ioctl: cmd not supported\n"));
                error = DDI_FAILURE;
        }

        DBG_PRINTF((category | RSM_IOCTL, RSM_DEBUG_VERBOSE,
            "rsmka_topology_ioctl done: %d\n", error));
        return (error);
}