1 files changed, 67 insertions, 48 deletions

diff --git a/fs/dlm/recoverd.c b/fs/dlm/recoverd.c
index fd677c8c3d3..6859b4bf971 100644
--- a/fs/dlm/recoverd.c
+++ b/fs/dlm/recoverd.c
@@ -2,7 +2,7 @@
 *******************************************************************************
 **
 **  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
-**  Copyright (C) 2004-2007 Red Hat, Inc.  All rights reserved.
+**  Copyright (C) 2004-2011 Red Hat, Inc.  All rights reserved.
 **
 **  This copyrighted material is made available to anyone wishing to use,
 **  modify, copy, or redistribute it subject to the terms and conditions
@@ -41,6 +41,7 @@ static int enable_locking(struct dlm_ls *ls, uint64_t seq)
 		set_bit(LSFL_RUNNING, &ls->ls_flags);
 		/* unblocks processes waiting to enter the dlm */
 		up_write(&ls->ls_in_recovery);
+		clear_bit(LSFL_RECOVER_LOCK, &ls->ls_flags);
 		error = 0;
 	}
 	spin_unlock(&ls->ls_recover_lock);
@@ -54,24 +55,13 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 	unsigned long start;
 	int error, neg = 0;
 
-	log_debug(ls, "recover %llx", (unsigned long long)rv->seq);
+	log_rinfo(ls, "dlm_recover %llu", (unsigned long long)rv->seq);
 
 	mutex_lock(&ls->ls_recoverd_active);
 
-	/*
-	 * Suspending and resuming dlm_astd ensures that no lkb's from this ls
-	 * will be processed by dlm_astd during recovery.
-	 */
-
-	dlm_astd_suspend();
-	dlm_astd_resume();
-
-	/*
-	 * Free non-master tossed rsb's.  Master rsb's are kept on toss
-	 * list and put on root list to be included in resdir recovery.
-	 */
+	dlm_callback_suspend(ls);
 
-	dlm_clear_toss_list(ls);
+	dlm_clear_toss(ls);
 
 	/*
 	 * This list of root rsb's will be the basis of most of the recovery
@@ -82,14 +72,28 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 
 	/*
 	 * Add or remove nodes from the lockspace's ls_nodes list.
-	 * Also waits for all nodes to complete dlm_recover_members.
 	 */
 
 	error = dlm_recover_members(ls, rv, &neg);
 	if (error) {
-		log_debug(ls, "recover_members failed %d", error);
+		log_rinfo(ls, "dlm_recover_members error %d", error);
+		goto fail;
+	}
+
+	dlm_recover_dir_nodeid(ls);
+
+	ls->ls_recover_dir_sent_res = 0;
+	ls->ls_recover_dir_sent_msg = 0;
+	ls->ls_recover_locks_in = 0;
+
+	dlm_set_recover_status(ls, DLM_RS_NODES);
+
+	error = dlm_recover_members_wait(ls);
+	if (error) {
+		log_rinfo(ls, "dlm_recover_members_wait error %d", error);
 		goto fail;
 	}
+
 	start = jiffies;
 
 	/*
@@ -99,20 +103,21 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 
 	error = dlm_recover_directory(ls);
 	if (error) {
-		log_debug(ls, "recover_directory failed %d", error);
+		log_rinfo(ls, "dlm_recover_directory error %d", error);
 		goto fail;
 	}
 
-	/*
-	 * Wait for all nodes to complete directory rebuild.
-	 */
+	dlm_set_recover_status(ls, DLM_RS_DIR);
 
 	error = dlm_recover_directory_wait(ls);
 	if (error) {
-		log_debug(ls, "recover_directory_wait failed %d", error);
+		log_rinfo(ls, "dlm_recover_directory_wait error %d", error);
 		goto fail;
 	}
 
+	log_rinfo(ls, "dlm_recover_directory %u out %u messages",
+		  ls->ls_recover_dir_sent_res, ls->ls_recover_dir_sent_msg);
+
 	/*
 	 * We may have outstanding operations that are waiting for a reply from
 	 * a failed node.  Mark these to be resent after recovery.  Unlock and
@@ -130,7 +135,7 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 		 * Clear lkb's for departed nodes.
 		 */
 
-		dlm_purge_locks(ls);
+		dlm_recover_purge(ls);
 
 		/*
 		 * Get new master nodeid's for rsb's that were mastered on
@@ -139,7 +144,7 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 
 		error = dlm_recover_masters(ls);
 		if (error) {
-			log_debug(ls, "recover_masters failed %d", error);
+			log_rinfo(ls, "dlm_recover_masters error %d", error);
 			goto fail;
 		}
 
@@ -149,16 +154,21 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 
 		error = dlm_recover_locks(ls);
 		if (error) {
-			log_debug(ls, "recover_locks failed %d", error);
+			log_rinfo(ls, "dlm_recover_locks error %d", error);
 			goto fail;
 		}
 
+		dlm_set_recover_status(ls, DLM_RS_LOCKS);
+
 		error = dlm_recover_locks_wait(ls);
 		if (error) {
-			log_debug(ls, "recover_locks_wait failed %d", error);
+			log_rinfo(ls, "dlm_recover_locks_wait error %d", error);
 			goto fail;
 		}
 
+		log_rinfo(ls, "dlm_recover_locks %u in",
+			  ls->ls_recover_locks_in);
+
 		/*
 		 * Finalize state in master rsb's now that all locks can be
 		 * checked.  This includes conversion resolution and lvb
@@ -176,7 +186,7 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 
 		error = dlm_recover_locks_wait(ls);
 		if (error) {
-			log_debug(ls, "recover_locks_wait failed %d", error);
+			log_rinfo(ls, "dlm_recover_locks_wait error %d", error);
 			goto fail;
 		}
 	}
@@ -192,9 +202,10 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 	dlm_purge_requestqueue(ls);
 
 	dlm_set_recover_status(ls, DLM_RS_DONE);
+
 	error = dlm_recover_done_wait(ls);
 	if (error) {
-		log_debug(ls, "recover_done_wait failed %d", error);
+		log_rinfo(ls, "dlm_recover_done_wait error %d", error);
 		goto fail;
 	}
 
@@ -202,38 +213,39 @@ static int ls_recover(struct dlm_ls *ls, struct dlm_recover *rv)
 
 	dlm_adjust_timeouts(ls);
 
+	dlm_callback_resume(ls);
+
 	error = enable_locking(ls, rv->seq);
 	if (error) {
-		log_debug(ls, "enable_locking failed %d", error);
+		log_rinfo(ls, "enable_locking error %d", error);
 		goto fail;
 	}
 
 	error = dlm_process_requestqueue(ls);
 	if (error) {
-		log_debug(ls, "process_requestqueue failed %d", error);
+		log_rinfo(ls, "dlm_process_requestqueue error %d", error);
 		goto fail;
 	}
 
 	error = dlm_recover_waiters_post(ls);
 	if (error) {
-		log_debug(ls, "recover_waiters_post failed %d", error);
+		log_rinfo(ls, "dlm_recover_waiters_post error %d", error);
 		goto fail;
 	}
 
-	dlm_grant_after_purge(ls);
+	dlm_recover_grant(ls);
 
-	dlm_astd_wake();
-
-	log_debug(ls, "recover %llx done: %u ms",
-		  (unsigned long long)rv->seq,
+	log_rinfo(ls, "dlm_recover %llu generation %u done: %u ms",
+		  (unsigned long long)rv->seq, ls->ls_generation,
 		  jiffies_to_msecs(jiffies - start));
 	mutex_unlock(&ls->ls_recoverd_active);
 
+	dlm_lsop_recover_done(ls);
 	return 0;
 
  fail:
 	dlm_release_root_list(ls);
-	log_debug(ls, "recover %llx error %d",
+	log_rinfo(ls, "dlm_recover %llu error %d",
 		  (unsigned long long)rv->seq, error);
 	mutex_unlock(&ls->ls_recoverd_active);
 	return error;
@@ -251,13 +263,12 @@ static void do_ls_recovery(struct dlm_ls *ls)
 	rv = ls->ls_recover_args;
 	ls->ls_recover_args = NULL;
 	if (rv && ls->ls_recover_seq == rv->seq)
-		clear_bit(LSFL_RECOVERY_STOP, &ls->ls_flags);
+		clear_bit(LSFL_RECOVER_STOP, &ls->ls_flags);
 	spin_unlock(&ls->ls_recover_lock);
 
 	if (rv) {
 		ls_recover(ls, rv);
-		kfree(rv->nodeids);
-		kfree(rv->new);
+		kfree(rv->nodes);
 		kfree(rv);
 	}
 }
@@ -272,26 +283,34 @@ static int dlm_recoverd(void *arg)
 		return -1;
 	}
 
+	down_write(&ls->ls_in_recovery);
+	set_bit(LSFL_RECOVER_LOCK, &ls->ls_flags);
+	wake_up(&ls->ls_recover_lock_wait);
+
 	while (!kthread_should_stop()) {
 		set_current_state(TASK_INTERRUPTIBLE);
-		if (!test_bit(LSFL_WORK, &ls->ls_flags))
+		if (!test_bit(LSFL_RECOVER_WORK, &ls->ls_flags) &&
+		    !test_bit(LSFL_RECOVER_DOWN, &ls->ls_flags))
 			schedule();
 		set_current_state(TASK_RUNNING);
 
-		if (test_and_clear_bit(LSFL_WORK, &ls->ls_flags))
+		if (test_and_clear_bit(LSFL_RECOVER_DOWN, &ls->ls_flags)) {
+			down_write(&ls->ls_in_recovery);
+			set_bit(LSFL_RECOVER_LOCK, &ls->ls_flags);
+			wake_up(&ls->ls_recover_lock_wait);
+		}
+
+		if (test_and_clear_bit(LSFL_RECOVER_WORK, &ls->ls_flags))
 			do_ls_recovery(ls);
 	}
 
+	if (test_bit(LSFL_RECOVER_LOCK, &ls->ls_flags))
+		up_write(&ls->ls_in_recovery);
+
 	dlm_put_lockspace(ls);
 	return 0;
 }
 
-void dlm_recoverd_kick(struct dlm_ls *ls)
-{
-	set_bit(LSFL_WORK, &ls->ls_flags);
-	wake_up_process(ls->ls_recoverd_task);
-}
-
 int dlm_recoverd_start(struct dlm_ls *ls)
 {
 	struct task_struct *p;