/* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*-
* vim:expandtab:shiftwidth=8:tabstop=8:
*
- * Portal-RPC reconnection and replay operations, for use in recovery.
+ * GPL HEADER START
+ *
+ * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 only,
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * General Public License version 2 for more details (a copy is included
+ * in the LICENSE file that accompanied this code).
*
- * Copyright (c) 2003 Cluster File Systems, Inc.
- * Authors: Phil Schwan <phil@clusterfs.com>
- * Mike Shaver <shaver@clusterfs.com>
+ * You should have received a copy of the GNU General Public License
+ * version 2 along with this program; If not, see
+ * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
*
- * This file is part of the Lustre file system, http://www.lustre.org
- * Lustre is a trademark of Cluster File Systems, Inc.
+ * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
+ * CA 95054 USA or visit www.sun.com if you need additional information or
+ * have any questions.
*
- * You may have signed or agreed to another license before downloading
- * this software. If so, you are bound by the terms and conditions
- * of that agreement, and the following does not apply to you. See the
- * LICENSE file included with this distribution for more information.
+ * GPL HEADER END
+ */
+/*
+ * Copyright 2008 Sun Microsystems, Inc. All rights reserved
+ * Use is subject to license terms.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
*
- * If you did not agree to a different license, then this copy of Lustre
- * is open source software; you can redistribute it and/or modify it
- * under the terms of version 2 of the GNU General Public License as
- * published by the Free Software Foundation.
+ * lustre/ptlrpc/pinger.c
*
- * In either case, Lustre is distributed in the hope that it will be
- * useful, but WITHOUT ANY WARRANTY; without even the implied warranty
- * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
- * license text for more details.
+ * Portal-RPC reconnection and replay operations, for use in recovery.
*/
#ifndef __KERNEL__
#include "ptlrpc_internal.h"
struct semaphore pinger_sem;
-static struct list_head pinger_imports = CFS_LIST_HEAD_INIT(pinger_imports);
+static CFS_LIST_HEAD(pinger_imports);
int ptlrpc_ping(struct obd_import *imp)
{
imp->imp_obd->obd_uuid.uuid, obd2cli_tgt(imp->imp_obd));
req->rq_no_resend = req->rq_no_delay = 1;
ptlrpc_request_set_replen(req);
- req->rq_timeout = PING_INTERVAL;
- ptlrpcd_add_req(req);
+ ptlrpcd_add_req(req, PSCOPE_OTHER);
RETURN(0);
}
void ptlrpc_update_next_ping(struct obd_import *imp)
{
#ifdef ENABLE_PINGER
- imp->imp_next_ping = cfs_time_shift(
- (imp->imp_state == LUSTRE_IMP_DISCON ?
- RECONNECT_INTERVAL : PING_INTERVAL));
+ int time = PING_INTERVAL;
+ if (imp->imp_state == LUSTRE_IMP_DISCON) {
+ int dtime = max_t(int, CONNECTION_SWITCH_MIN,
+ AT_OFF ? 0 :
+ at_get(&imp->imp_at.iat_net_latency));
+ time = min(time, dtime);
+ }
+ imp->imp_next_ping = cfs_time_shift(time);
#endif /* ENABLE_PINGER */
}
/* wait at least a timeout before trying recovery again */
imp->imp_next_ping = ptlrpc_next_reconnect(imp);
ptlrpc_initiate_recovery(imp);
- } else if (level != LUSTRE_IMP_FULL ||
+ } else if (level != LUSTRE_IMP_FULL ||
imp->imp_obd->obd_no_recov ||
imp->imp_deactive) {
CDEBUG(D_HA, "not pinging %s (in recovery "
ptlrpc_pinger_process_import(imp, this_ping);
/* obd_timeout might have changed */
if (imp->imp_pingable && imp->imp_next_ping &&
- cfs_time_after(imp->imp_next_ping,
- cfs_time_add(this_ping,
+ cfs_time_after(imp->imp_next_ping,
+ cfs_time_add(this_ping,
cfs_time_seconds(PING_INTERVAL))))
ptlrpc_update_next_ping(imp);
}
obd_update_maxusage();
/* Wait until the next ping time, or until we're stopped. */
- time_to_next_ping = cfs_time_sub(cfs_time_add(this_ping,
- cfs_time_seconds(PING_INTERVAL)),
- cfs_time_current());
+ time_to_next_ping = cfs_time_sub(cfs_time_add(this_ping,
+ cfs_time_seconds(PING_INTERVAL)),
+ cfs_time_current());
/* The ping sent by ptlrpc_send_rpc may get sent out
say .01 second after this.
- ptlrpc_pinger_eending_on_import will then set the
+ ptlrpc_pinger_sending_on_import will then set the
next ping time to next_ping + .01 sec, which means
we will SKIP the next ping at next_ping, and the
ping will get sent 2 timeouts from now! Beware. */
- CDEBUG(D_INFO, "next ping in "CFS_DURATION_T" ("CFS_TIME_T")\n",
- time_to_next_ping,
- cfs_time_add(this_ping, cfs_time_seconds(PING_INTERVAL)));
+ CDEBUG(D_INFO, "next ping in "CFS_DURATION_T" ("CFS_TIME_T")\n",
+ time_to_next_ping,
+ cfs_time_add(this_ping,
+ cfs_time_seconds(PING_INTERVAL)));
if (time_to_next_ping > 0) {
- lwi = LWI_TIMEOUT(max_t(cfs_duration_t, time_to_next_ping, cfs_time_seconds(1)),
+ lwi = LWI_TIMEOUT(max_t(cfs_duration_t,
+ time_to_next_ping,
+ cfs_time_seconds(1)),
NULL, NULL);
l_wait_event(thread->t_ctl_waitq,
thread->t_flags & (SVC_STOPPING|SVC_EVENT),
RETURN(-ENOMEM);
cfs_waitq_init(&pinger_thread->t_ctl_waitq);
cfs_waitq_init(&suspend_timeouts_waitq);
-
+
d.name = "ll_ping";
d.thread = pinger_thread;
obd = pet_exp->exp_obd;
spin_unlock(&pet_lock);
- expire_time = CURRENT_SECONDS - (3 * obd_timeout / 2);
+ expire_time = cfs_time_current_sec() - PING_EVICT_TIMEOUT;
CDEBUG(D_HA, "evicting all exports of obd %s older than %ld\n",
obd->obd_name, expire_time);
obd->obd_name,
obd_uuid2str(&exp->exp_client_uuid),
obd_export_nid2str(exp),
- (long)(CURRENT_SECONDS -
+ (long)(cfs_time_current_sec() -
exp->exp_last_request_time),
- exp, (long)CURRENT_SECONDS,
+ exp, (long)cfs_time_current_sec(),
(long)expire_time,
(long)exp->exp_last_request_time);
CDEBUG(D_HA, "Last request was at %ld\n",
struct ptlrpc_request *req;
struct ptlrpc_request_set *set;
struct list_head *iter;
+ struct obd_import *imp;
struct pinger_data *pd = &pinger_args;
int rc;
list_entry(iter, struct obd_import, imp_pinger_chain);
int generation, level;
- if (cfs_time_aftereq(pd->pd_this_ping,
+ if (cfs_time_aftereq(pd->pd_this_ping,
imp->imp_next_ping - 5 * CFS_TICK)) {
/* Add a ping. */
spin_lock(&imp->imp_lock);
req->rq_no_resend = 1;
ptlrpc_request_set_replen(req);
req->rq_send_state = LUSTRE_IMP_FULL;
- req->rq_phase = RQ_PHASE_RPC;
+ ptlrpc_rqphase_move(req, RQ_PHASE_RPC);
req->rq_import_generation = generation;
ptlrpc_set_add_req(set, req);
} else {
}
do_check_set:
- rc = ptlrpc_check_set(set);
+ rc = ptlrpc_check_set(NULL, set);
/* not finished, and we are not expired, simply return */
if (!rc && cfs_time_before(curtime, cfs_time_add(pd->pd_this_ping,
if (req->rq_phase == RQ_PHASE_COMPLETE)
continue;
- req->rq_phase = RQ_PHASE_COMPLETE;
- atomic_dec(&req->rq_import->imp_inflight);
- set->set_remaining--;
- /* If it was disconnected, don't sweat it. */
- if (list_empty(&req->rq_import->imp_pinger_chain)) {
- ptlrpc_unregister_reply(req);
- continue;
- }
+ CDEBUG(D_RPCTRACE, "Pinger initiate expire request(%p)\n",
+ req);
+
+ /* This will also unregister reply. */
+ ptlrpc_expire_one_request(req, 0);
- CDEBUG(D_RPCTRACE, "pinger initiate expire_one_request\n");
- ptlrpc_expire_one_request(req);
+ /* We're done with this req, let's finally move it to complete
+ * phase and take care of inflights. */
+ ptlrpc_rqphase_move(req, RQ_PHASE_COMPLETE);
+ imp = req->rq_import;
+ spin_lock(&imp->imp_lock);
+ if (!list_empty(&req->rq_list)) {
+ list_del_init(&req->rq_list);
+ atomic_dec(&imp->imp_inflight);
+ }
+ spin_unlock(&imp->imp_lock);
+ set->set_remaining--;
}
mutex_up(&pinger_sem);