ocfs2/dlm: add DEREF_DONE message

This series of patches is to fix the dis-order issue of setting/clearing
refmap bit described below.

Node 1                               Node 2(master)
dlmlock
dlm_do_master_request
                                dlm_master_request_handler
                                -> dlm_lockres_set_refmap_bit
dlmlock succeed
dlmunlock succeed

dlm_purge_lockres
                                dlm_deref_handler
                                -> find lock resource is in
                                   DLM_LOCK_RES_SETREF_INPROG state,
                                   so dispatch a deref work
dlm_purge_lockres succeed.

call dlmlock again
dlm_do_master_request
                                dlm_master_request_handler
                                -> dlm_lockres_set_refmap_bit

                                deref work trigger, call
                                dlm_lockres_clear_refmap_bit
                                to clear Node 1 from refmap

                                dlm_purge_lockres succeed

dlm_send_remote_lock_request
                                return DLM_IVLOCKID because
                                the lockres is not exist
BUG if the lockres is $RECOVERY

This series of patches add a new message to keep the order of set and
clear.  Other nodes can purge the lock resource only after the refmap bit
on master is cleared.

This patch is to add DEREF_DONE message and corresponding handler.  Node
can purge the lock resource after receiving this message.  As a new
message is added, so increase the minor number of dlm protocol version.

Signed-off-by: xuejiufei <xuejiufei@huawei.com>
Cc: Mark Fasheh <mfasheh@suse.de>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Reviewed-by: Joseph Qi <joseph.qi@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

1 files changed, 116 insertions, 0 deletions

diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c
index 9477d6e..8913e7d 100644
--- a/fs/ocfs2/dlm/dlmmaster.c
+++ b/fs/ocfs2/dlm/dlmmaster.c
@@ -2375,6 +2375,122 @@ done:
 	return ret;
 }
 
+int dlm_deref_lockres_done_handler(struct o2net_msg *msg, u32 len, void *data,
+			      void **ret_data)
+{
+	struct dlm_ctxt *dlm = data;
+	struct dlm_deref_lockres_done *deref
+			= (struct dlm_deref_lockres_done *)msg->buf;
+	struct dlm_lock_resource *res = NULL;
+	char *name;
+	unsigned int namelen;
+	int ret = -EINVAL;
+	u8 node;
+	unsigned int hash;
+
+	if (!dlm_grab(dlm))
+		return 0;
+
+	name = deref->name;
+	namelen = deref->namelen;
+	node = deref->node_idx;
+
+	if (namelen > DLM_LOCKID_NAME_MAX) {
+		mlog(ML_ERROR, "Invalid name length!");
+		goto done;
+	}
+	if (deref->node_idx >= O2NM_MAX_NODES) {
+		mlog(ML_ERROR, "Invalid node number: %u\n", node);
+		goto done;
+	}
+
+	hash = dlm_lockid_hash(name, namelen);
+
+	spin_lock(&dlm->spinlock);
+	res = __dlm_lookup_lockres_full(dlm, name, namelen, hash);
+	if (!res) {
+		spin_unlock(&dlm->spinlock);
+		mlog(ML_ERROR, "%s:%.*s: bad lockres name\n",
+		     dlm->name, namelen, name);
+		goto done;
+	}
+
+	spin_lock(&res->spinlock);
+	BUG_ON(!(res->state & DLM_LOCK_RES_DROPPING_REF));
+	if (!list_empty(&res->purge)) {
+		mlog(0, "%s: Removing res %.*s from purgelist\n",
+			dlm->name, res->lockname.len, res->lockname.name);
+		list_del_init(&res->purge);
+		dlm_lockres_put(res);
+		dlm->purge_count--;
+	}
+
+	if (!__dlm_lockres_unused(res)) {
+		mlog(ML_ERROR, "%s: res %.*s in use after deref\n",
+			dlm->name, res->lockname.len, res->lockname.name);
+		__dlm_print_one_lock_resource(res);
+		BUG();
+	}
+
+	__dlm_unhash_lockres(dlm, res);
+
+	spin_lock(&dlm->track_lock);
+	if (!list_empty(&res->tracking))
+		list_del_init(&res->tracking);
+	else {
+		mlog(ML_ERROR, "%s: Resource %.*s not on the Tracking list\n",
+		     dlm->name, res->lockname.len, res->lockname.name);
+		__dlm_print_one_lock_resource(res);
+	}
+	spin_unlock(&dlm->track_lock);
+
+	/* lockres is not in the hash now. drop the flag and wake up
+	 * any processes waiting in dlm_get_lock_resource.
+	 */
+	res->state &= ~DLM_LOCK_RES_DROPPING_REF;
+	spin_unlock(&res->spinlock);
+	wake_up(&res->wq);
+
+	dlm_lockres_put(res);
+
+	spin_unlock(&dlm->spinlock);
+
+done:
+	dlm_put(dlm);
+	return ret;
+}
+
+static void dlm_drop_lockres_ref_done(struct dlm_ctxt *dlm,
+		struct dlm_lock_resource *res, u8 node)
+{
+	struct dlm_deref_lockres_done deref;
+	int ret = 0, r;
+	const char *lockname;
+	unsigned int namelen;
+
+	lockname = res->lockname.name;
+	namelen = res->lockname.len;
+	BUG_ON(namelen > O2NM_MAX_NAME_LEN);
+
+	memset(&deref, 0, sizeof(deref));
+	deref.node_idx = dlm->node_num;
+	deref.namelen = namelen;
+	memcpy(deref.name, lockname, namelen);
+
+	ret = o2net_send_message(DLM_DEREF_LOCKRES_DONE, dlm->key,
+				 &deref, sizeof(deref), node, &r);
+	if (ret < 0) {
+		mlog(ML_ERROR, "%s: res %.*s, error %d send DEREF DONE "
+				" to node %u\n", dlm->name, namelen,
+				lockname, ret, node);
+	} else if (r < 0) {
+		/* ignore the error */
+		mlog(ML_ERROR, "%s: res %.*s, DEREF to node %u got %d\n",
+		     dlm->name, namelen, lockname, node, r);
+		dlm_print_one_lock_resource(res);
+	}
+}
+
 static void dlm_deref_lockres_worker(struct dlm_work_item *item, void *data)
 {
 	struct dlm_ctxt *dlm;