summaryrefslogtreecommitdiff
path: root/fs/ocfs2/dlm
diff options
context:
space:
mode:
authorJoel Becker <joel.becker@oracle.com>2010-10-15 13:03:09 -0700
committerJoel Becker <joel.becker@oracle.com>2010-10-15 13:03:09 -0700
commitfc3718918f13ad72827d62d36ea0f5fb55090644 (patch)
tree4f9551256e02d08be37bab137f3d94182a67504c /fs/ocfs2/dlm
parent7bdb0d18bfd381cc5491eb95973ec5604b356c7e (diff)
parentd4396eafe402b710a8535137b3bf2abe6c059a15 (diff)
Merge branch 'globalheartbeat-2' of git://oss.oracle.com/git/smushran/linux-2.6 into ocfs2-merge-window
Conflicts: fs/ocfs2/ocfs2.h
Diffstat (limited to 'fs/ocfs2/dlm')
-rw-r--r--fs/ocfs2/dlm/dlmcommon.h30
-rw-r--r--fs/ocfs2/dlm/dlmdebug.c13
-rw-r--r--fs/ocfs2/dlm/dlmdomain.c401
-rw-r--r--fs/ocfs2/dlm/dlmmaster.c40
4 files changed, 480 insertions, 4 deletions
diff --git a/fs/ocfs2/dlm/dlmcommon.h b/fs/ocfs2/dlm/dlmcommon.h
index 4b6ae2c13b47..b36d0bf77a5a 100644
--- a/fs/ocfs2/dlm/dlmcommon.h
+++ b/fs/ocfs2/dlm/dlmcommon.h
@@ -445,7 +445,9 @@ enum {
DLM_LOCK_REQUEST_MSG, /* 515 */
DLM_RECO_DATA_DONE_MSG, /* 516 */
DLM_BEGIN_RECO_MSG, /* 517 */
- DLM_FINALIZE_RECO_MSG /* 518 */
+ DLM_FINALIZE_RECO_MSG, /* 518 */
+ DLM_QUERY_REGION, /* 519 */
+ DLM_QUERY_NODEINFO, /* 520 */
};
struct dlm_reco_node_data
@@ -727,6 +729,31 @@ struct dlm_cancel_join
u8 domain[O2NM_MAX_NAME_LEN];
};
+struct dlm_query_region {
+ u8 qr_node;
+ u8 qr_numregions;
+ u8 qr_namelen;
+ u8 pad1;
+ u8 qr_domain[O2NM_MAX_NAME_LEN];
+ u8 qr_regions[O2HB_MAX_REGION_NAME_LEN * O2NM_MAX_REGIONS];
+};
+
+struct dlm_node_info {
+ u8 ni_nodenum;
+ u8 pad1;
+ u16 ni_ipv4_port;
+ u32 ni_ipv4_address;
+};
+
+struct dlm_query_nodeinfo {
+ u8 qn_nodenum;
+ u8 qn_numnodes;
+ u8 qn_namelen;
+ u8 pad1;
+ u8 qn_domain[O2NM_MAX_NAME_LEN];
+ struct dlm_node_info qn_nodes[O2NM_MAX_NODES];
+};
+
struct dlm_exit_domain
{
u8 node_idx;
@@ -1030,6 +1057,7 @@ int dlm_drop_lockres_ref(struct dlm_ctxt *dlm,
struct dlm_lock_resource *res);
void dlm_clean_master_list(struct dlm_ctxt *dlm,
u8 dead_node);
+void dlm_force_free_mles(struct dlm_ctxt *dlm);
int dlm_lock_basts_flushed(struct dlm_ctxt *dlm, struct dlm_lock *lock);
int __dlm_lockres_has_locks(struct dlm_lock_resource *res);
int __dlm_lockres_unused(struct dlm_lock_resource *res);
diff --git a/fs/ocfs2/dlm/dlmdebug.c b/fs/ocfs2/dlm/dlmdebug.c
index 26ff2e185b1e..272ec8631a51 100644
--- a/fs/ocfs2/dlm/dlmdebug.c
+++ b/fs/ocfs2/dlm/dlmdebug.c
@@ -636,8 +636,14 @@ static void *lockres_seq_start(struct seq_file *m, loff_t *pos)
spin_lock(&dlm->track_lock);
if (oldres)
track_list = &oldres->tracking;
- else
+ else {
track_list = &dlm->tracking_list;
+ if (list_empty(track_list)) {
+ dl = NULL;
+ spin_unlock(&dlm->track_lock);
+ goto bail;
+ }
+ }
list_for_each_entry(res, track_list, tracking) {
if (&res->tracking == &dlm->tracking_list)
@@ -660,6 +666,7 @@ static void *lockres_seq_start(struct seq_file *m, loff_t *pos)
} else
dl = NULL;
+bail:
/* passed to seq_show */
return dl;
}
@@ -775,7 +782,9 @@ static int debug_state_print(struct dlm_ctxt *dlm, struct debug_buffer *db)
/* Domain: xxxxxxxxxx Key: 0xdfbac769 */
out += snprintf(db->buf + out, db->len - out,
- "Domain: %s Key: 0x%08x\n", dlm->name, dlm->key);
+ "Domain: %s Key: 0x%08x Protocol: %d.%d\n",
+ dlm->name, dlm->key, dlm->dlm_locking_proto.pv_major,
+ dlm->dlm_locking_proto.pv_minor);
/* Thread Pid: xxx Node: xxx State: xxxxx */
out += snprintf(db->buf + out, db->len - out,
diff --git a/fs/ocfs2/dlm/dlmdomain.c b/fs/ocfs2/dlm/dlmdomain.c
index 153abb5abef0..58a93b953735 100644
--- a/fs/ocfs2/dlm/dlmdomain.c
+++ b/fs/ocfs2/dlm/dlmdomain.c
@@ -128,10 +128,14 @@ static DECLARE_WAIT_QUEUE_HEAD(dlm_domain_events);
* will have a negotiated version with the same major number and a minor
* number equal or smaller. The dlm_ctxt->dlm_locking_proto field should
* be used to determine what a running domain is actually using.
+ *
+ * New in version 1.1:
+ * - Message DLM_QUERY_REGION added to support global heartbeat
+ * - Message DLM_QUERY_NODEINFO added to allow online node removes
*/
static const struct dlm_protocol_version dlm_protocol = {
.pv_major = 1,
- .pv_minor = 0,
+ .pv_minor = 1,
};
#define DLM_DOMAIN_BACKOFF_MS 200
@@ -142,6 +146,8 @@ static int dlm_assert_joined_handler(struct o2net_msg *msg, u32 len, void *data,
void **ret_data);
static int dlm_cancel_join_handler(struct o2net_msg *msg, u32 len, void *data,
void **ret_data);
+static int dlm_query_region_handler(struct o2net_msg *msg, u32 len,
+ void *data, void **ret_data);
static int dlm_exit_domain_handler(struct o2net_msg *msg, u32 len, void *data,
void **ret_data);
static int dlm_protocol_compare(struct dlm_protocol_version *existing,
@@ -693,6 +699,7 @@ void dlm_unregister_domain(struct dlm_ctxt *dlm)
dlm_mark_domain_leaving(dlm);
dlm_leave_domain(dlm);
+ dlm_force_free_mles(dlm);
dlm_complete_dlm_shutdown(dlm);
}
dlm_put(dlm);
@@ -920,6 +927,370 @@ static int dlm_assert_joined_handler(struct o2net_msg *msg, u32 len, void *data,
return 0;
}
+static int dlm_match_regions(struct dlm_ctxt *dlm,
+ struct dlm_query_region *qr)
+{
+ char *local = NULL, *remote = qr->qr_regions;
+ char *l, *r;
+ int localnr, i, j, foundit;
+ int status = 0;
+
+ if (!o2hb_global_heartbeat_active()) {
+ if (qr->qr_numregions) {
+ mlog(ML_ERROR, "Domain %s: Joining node %d has global "
+ "heartbeat enabled but local node %d does not\n",
+ qr->qr_domain, qr->qr_node, dlm->node_num);
+ status = -EINVAL;
+ }
+ goto bail;
+ }
+
+ if (o2hb_global_heartbeat_active() && !qr->qr_numregions) {
+ mlog(ML_ERROR, "Domain %s: Local node %d has global "
+ "heartbeat enabled but joining node %d does not\n",
+ qr->qr_domain, dlm->node_num, qr->qr_node);
+ status = -EINVAL;
+ goto bail;
+ }
+
+ r = remote;
+ for (i = 0; i < qr->qr_numregions; ++i) {
+ mlog(0, "Region %.*s\n", O2HB_MAX_REGION_NAME_LEN, r);
+ r += O2HB_MAX_REGION_NAME_LEN;
+ }
+
+ local = kmalloc(sizeof(qr->qr_regions), GFP_KERNEL);
+ if (!local) {
+ status = -ENOMEM;
+ goto bail;
+ }
+
+ localnr = o2hb_get_all_regions(local, O2NM_MAX_REGIONS);
+
+ /* compare local regions with remote */
+ l = local;
+ for (i = 0; i < localnr; ++i) {
+ foundit = 0;
+ r = remote;
+ for (j = 0; j <= qr->qr_numregions; ++j) {
+ if (!memcmp(l, r, O2HB_MAX_REGION_NAME_LEN)) {
+ foundit = 1;
+ break;
+ }
+ r += O2HB_MAX_REGION_NAME_LEN;
+ }
+ if (!foundit) {
+ status = -EINVAL;
+ mlog(ML_ERROR, "Domain %s: Region '%.*s' registered "
+ "in local node %d but not in joining node %d\n",
+ qr->qr_domain, O2HB_MAX_REGION_NAME_LEN, l,
+ dlm->node_num, qr->qr_node);
+ goto bail;
+ }
+ l += O2HB_MAX_REGION_NAME_LEN;
+ }
+
+ /* compare remote with local regions */
+ r = remote;
+ for (i = 0; i < qr->qr_numregions; ++i) {
+ foundit = 0;
+ l = local;
+ for (j = 0; j < localnr; ++j) {
+ if (!memcmp(r, l, O2HB_MAX_REGION_NAME_LEN)) {
+ foundit = 1;
+ break;
+ }
+ l += O2HB_MAX_REGION_NAME_LEN;
+ }
+ if (!foundit) {
+ status = -EINVAL;
+ mlog(ML_ERROR, "Domain %s: Region '%.*s' registered "
+ "in joining node %d but not in local node %d\n",
+ qr->qr_domain, O2HB_MAX_REGION_NAME_LEN, r,
+ qr->qr_node, dlm->node_num);
+ goto bail;
+ }
+ r += O2HB_MAX_REGION_NAME_LEN;
+ }
+
+bail:
+ kfree(local);
+
+ return status;
+}
+
+static int dlm_send_regions(struct dlm_ctxt *dlm, unsigned long *node_map)
+{
+ struct dlm_query_region *qr = NULL;
+ int status, ret = 0, i;
+ char *p;
+
+ if (find_next_bit(node_map, O2NM_MAX_NODES, 0) >= O2NM_MAX_NODES)
+ goto bail;
+
+ qr = kzalloc(sizeof(struct dlm_query_region), GFP_KERNEL);
+ if (!qr) {
+ ret = -ENOMEM;
+ mlog_errno(ret);
+ goto bail;
+ }
+
+ qr->qr_node = dlm->node_num;
+ qr->qr_namelen = strlen(dlm->name);
+ memcpy(qr->qr_domain, dlm->name, qr->qr_namelen);
+ /* if local hb, the numregions will be zero */
+ if (o2hb_global_heartbeat_active())
+ qr->qr_numregions = o2hb_get_all_regions(qr->qr_regions,
+ O2NM_MAX_REGIONS);
+
+ p = qr->qr_regions;
+ for (i = 0; i < qr->qr_numregions; ++i, p += O2HB_MAX_REGION_NAME_LEN)
+ mlog(0, "Region %.*s\n", O2HB_MAX_REGION_NAME_LEN, p);
+
+ i = -1;
+ while ((i = find_next_bit(node_map, O2NM_MAX_NODES,
+ i + 1)) < O2NM_MAX_NODES) {
+ if (i == dlm->node_num)
+ continue;
+
+ mlog(0, "Sending regions to node %d\n", i);
+
+ ret = o2net_send_message(DLM_QUERY_REGION, DLM_MOD_KEY, qr,
+ sizeof(struct dlm_query_region),
+ i, &status);
+ if (ret >= 0)
+ ret = status;
+ if (ret) {
+ mlog(ML_ERROR, "Region mismatch %d, node %d\n",
+ ret, i);
+ break;
+ }
+ }
+
+bail:
+ kfree(qr);
+ return ret;
+}
+
+static int dlm_query_region_handler(struct o2net_msg *msg, u32 len,
+ void *data, void **ret_data)
+{
+ struct dlm_query_region *qr;
+ struct dlm_ctxt *dlm = NULL;
+ int status = 0;
+ int locked = 0;
+
+ qr = (struct dlm_query_region *) msg->buf;
+
+ mlog(0, "Node %u queries hb regions on domain %s\n", qr->qr_node,
+ qr->qr_domain);
+
+ status = -EINVAL;
+
+ spin_lock(&dlm_domain_lock);
+ dlm = __dlm_lookup_domain_full(qr->qr_domain, qr->qr_namelen);
+ if (!dlm) {
+ mlog(ML_ERROR, "Node %d queried hb regions on domain %s "
+ "before join domain\n", qr->qr_node, qr->qr_domain);
+ goto bail;
+ }
+
+ spin_lock(&dlm->spinlock);
+ locked = 1;
+ if (dlm->joining_node != qr->qr_node) {
+ mlog(ML_ERROR, "Node %d queried hb regions on domain %s "
+ "but joining node is %d\n", qr->qr_node, qr->qr_domain,
+ dlm->joining_node);
+ goto bail;
+ }
+
+ /* Support for global heartbeat was added in 1.1 */
+ if (dlm->dlm_locking_proto.pv_major == 1 &&
+ dlm->dlm_locking_proto.pv_minor == 0) {
+ mlog(ML_ERROR, "Node %d queried hb regions on domain %s "
+ "but active dlm protocol is %d.%d\n", qr->qr_node,
+ qr->qr_domain, dlm->dlm_locking_proto.pv_major,
+ dlm->dlm_locking_proto.pv_minor);
+ goto bail;
+ }
+
+ status = dlm_match_regions(dlm, qr);
+
+bail:
+ if (locked)
+ spin_unlock(&dlm->spinlock);
+ spin_unlock(&dlm_domain_lock);
+
+ return status;
+}
+
+static int dlm_match_nodes(struct dlm_ctxt *dlm, struct dlm_query_nodeinfo *qn)
+{
+ struct o2nm_node *local;
+ struct dlm_node_info *remote;
+ int i, j;
+ int status = 0;
+
+ for (j = 0; j < qn->qn_numnodes; ++j)
+ mlog(0, "Node %3d, %pI4:%u\n", qn->qn_nodes[j].ni_nodenum,
+ &(qn->qn_nodes[j].ni_ipv4_address),
+ ntohs(qn->qn_nodes[j].ni_ipv4_port));
+
+ for (i = 0; i < O2NM_MAX_NODES && !status; ++i) {
+ local = o2nm_get_node_by_num(i);
+ remote = NULL;
+ for (j = 0; j < qn->qn_numnodes; ++j) {
+ if (qn->qn_nodes[j].ni_nodenum == i) {
+ remote = &(qn->qn_nodes[j]);
+ break;
+ }
+ }
+
+ if (!local && !remote)
+ continue;
+
+ if ((local && !remote) || (!local && remote))
+ status = -EINVAL;
+
+ if (!status &&
+ ((remote->ni_nodenum != local->nd_num) ||
+ (remote->ni_ipv4_port != local->nd_ipv4_port) ||
+ (remote->ni_ipv4_address != local->nd_ipv4_address)))
+ status = -EINVAL;
+
+ if (status) {
+ if (remote && !local)
+ mlog(ML_ERROR, "Domain %s: Node %d (%pI4:%u) "
+ "registered in joining node %d but not in "
+ "local node %d\n", qn->qn_domain,
+ remote->ni_nodenum,
+ &(remote->ni_ipv4_address),
+ ntohs(remote->ni_ipv4_port),
+ qn->qn_nodenum, dlm->node_num);
+ if (local && !remote)
+ mlog(ML_ERROR, "Domain %s: Node %d (%pI4:%u) "
+ "registered in local node %d but not in "
+ "joining node %d\n", qn->qn_domain,
+ local->nd_num, &(local->nd_ipv4_address),
+ ntohs(local->nd_ipv4_port),
+ dlm->node_num, qn->qn_nodenum);
+ BUG_ON((!local && !remote));
+ }
+
+ if (local)
+ o2nm_node_put(local);
+ }
+
+ return status;
+}
+
+static int dlm_send_nodeinfo(struct dlm_ctxt *dlm, unsigned long *node_map)
+{
+ struct dlm_query_nodeinfo *qn = NULL;
+ struct o2nm_node *node;
+ int ret = 0, status, count, i;
+
+ if (find_next_bit(node_map, O2NM_MAX_NODES, 0) >= O2NM_MAX_NODES)
+ goto bail;
+
+ qn = kzalloc(sizeof(struct dlm_query_nodeinfo), GFP_KERNEL);
+ if (!qn) {
+ ret = -ENOMEM;
+ mlog_errno(ret);
+ goto bail;
+ }
+
+ for (i = 0, count = 0; i < O2NM_MAX_NODES; ++i) {
+ node = o2nm_get_node_by_num(i);
+ if (!node)
+ continue;
+ qn->qn_nodes[count].ni_nodenum = node->nd_num;
+ qn->qn_nodes[count].ni_ipv4_port = node->nd_ipv4_port;
+ qn->qn_nodes[count].ni_ipv4_address = node->nd_ipv4_address;
+ mlog(0, "Node %3d, %pI4:%u\n", node->nd_num,
+ &(node->nd_ipv4_address), ntohs(node->nd_ipv4_port));
+ ++count;
+ o2nm_node_put(node);
+ }
+
+ qn->qn_nodenum = dlm->node_num;
+ qn->qn_numnodes = count;
+ qn->qn_namelen = strlen(dlm->name);
+ memcpy(qn->qn_domain, dlm->name, qn->qn_namelen);
+
+ i = -1;
+ while ((i = find_next_bit(node_map, O2NM_MAX_NODES,
+ i + 1)) < O2NM_MAX_NODES) {
+ if (i == dlm->node_num)
+ continue;
+
+ mlog(0, "Sending nodeinfo to node %d\n", i);
+
+ ret = o2net_send_message(DLM_QUERY_NODEINFO, DLM_MOD_KEY,
+ qn, sizeof(struct dlm_query_nodeinfo),
+ i, &status);
+ if (ret >= 0)
+ ret = status;
+ if (ret) {
+ mlog(ML_ERROR, "node mismatch %d, node %d\n", ret, i);
+ break;
+ }
+ }
+
+bail:
+ kfree(qn);
+ return ret;
+}
+
+static int dlm_query_nodeinfo_handler(struct o2net_msg *msg, u32 len,
+ void *data, void **ret_data)
+{
+ struct dlm_query_nodeinfo *qn;
+ struct dlm_ctxt *dlm = NULL;
+ int locked = 0, status = -EINVAL;
+
+ qn = (struct dlm_query_nodeinfo *) msg->buf;
+
+ mlog(0, "Node %u queries nodes on domain %s\n", qn->qn_nodenum,
+ qn->qn_domain);
+
+ spin_lock(&dlm_domain_lock);
+ dlm = __dlm_lookup_domain_full(qn->qn_domain, qn->qn_namelen);
+ if (!dlm) {
+ mlog(ML_ERROR, "Node %d queried nodes on domain %s before "
+ "join domain\n", qn->qn_nodenum, qn->qn_domain);
+ goto bail;
+ }
+
+ spin_lock(&dlm->spinlock);
+ locked = 1;
+ if (dlm->joining_node != qn->qn_nodenum) {
+ mlog(ML_ERROR, "Node %d queried nodes on domain %s but "
+ "joining node is %d\n", qn->qn_nodenum, qn->qn_domain,
+ dlm->joining_node);
+ goto bail;
+ }
+
+ /* Support for node query was added in 1.1 */
+ if (dlm->dlm_locking_proto.pv_major == 1 &&
+ dlm->dlm_locking_proto.pv_minor == 0) {
+ mlog(ML_ERROR, "Node %d queried nodes on domain %s "
+ "but active dlm protocol is %d.%d\n", qn->qn_nodenum,
+ qn->qn_domain, dlm->dlm_locking_proto.pv_major,
+ dlm->dlm_locking_proto.pv_minor);
+ goto bail;
+ }
+
+ status = dlm_match_nodes(dlm, qn);
+
+bail:
+ if (locked)
+ spin_unlock(&dlm->spinlock);
+ spin_unlock(&dlm_domain_lock);
+
+ return status;
+}
+
static int dlm_cancel_join_handler(struct o2net_msg *msg, u32 len, void *data,
void **ret_data)
{
@@ -1240,6 +1611,20 @@ static int dlm_try_to_join_domain(struct dlm_ctxt *dlm)
set_bit(dlm->node_num, dlm->domain_map);
spin_unlock(&dlm->spinlock);
+ /* Support for global heartbeat and node info was added in 1.1 */
+ if (dlm_protocol.pv_major > 1 || dlm_protocol.pv_minor > 0) {
+ status = dlm_send_nodeinfo(dlm, ctxt->yes_resp_map);
+ if (status) {
+ mlog_errno(status);
+ goto bail;
+ }
+ status = dlm_send_regions(dlm, ctxt->yes_resp_map);
+ if (status) {
+ mlog_errno(status);
+ goto bail;
+ }
+ }
+
dlm_send_join_asserts(dlm, ctxt->yes_resp_map);
/* Joined state *must* be set before the joining node
@@ -1806,7 +2191,21 @@ static int dlm_register_net_handlers(void)
sizeof(struct dlm_cancel_join),
dlm_cancel_join_handler,
NULL, NULL, &dlm_join_handlers);
+ if (status)
+ goto bail;
+
+ status = o2net_register_handler(DLM_QUERY_REGION, DLM_MOD_KEY,
+ sizeof(struct dlm_query_region),
+ dlm_query_region_handler,
+ NULL, NULL, &dlm_join_handlers);
+ if (status)
+ goto bail;
+
+ status = o2net_register_handler(DLM_QUERY_NODEINFO, DLM_MOD_KEY,
+ sizeof(struct dlm_query_nodeinfo),
+ dlm_query_nodeinfo_handler,
+ NULL, NULL, &dlm_join_handlers);
bail:
if (status < 0)
dlm_unregister_net_handlers();
diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c
index ffb4c68dafa4..f564b0e5f80d 100644
--- a/fs/ocfs2/dlm/dlmmaster.c
+++ b/fs/ocfs2/dlm/dlmmaster.c
@@ -3433,3 +3433,43 @@ void dlm_lockres_release_ast(struct dlm_ctxt *dlm,
wake_up(&res->wq);
wake_up(&dlm->migration_wq);
}
+
+void dlm_force_free_mles(struct dlm_ctxt *dlm)
+{
+ int i;
+ struct hlist_head *bucket;
+ struct dlm_master_list_entry *mle;
+ struct hlist_node *tmp, *list;
+
+ /*
+ * We notified all other nodes that we are exiting the domain and
+ * marked the dlm state to DLM_CTXT_LEAVING. If any mles are still
+ * around we force free them and wake any processes that are waiting
+ * on the mles
+ */
+ spin_lock(&dlm->spinlock);
+ spin_lock(&dlm->master_lock);
+
+ BUG_ON(dlm->dlm_state != DLM_CTXT_LEAVING);
+ BUG_ON((find_next_bit(dlm->domain_map, O2NM_MAX_NODES, 0) < O2NM_MAX_NODES));
+
+ for (i = 0; i < DLM_HASH_BUCKETS; i++) {
+ bucket = dlm_master_hash(dlm, i);
+ hlist_for_each_safe(list, tmp, bucket) {
+ mle = hlist_entry(list, struct dlm_master_list_entry,
+ master_hash_node);
+ if (mle->type != DLM_MLE_BLOCK) {
+ mlog(ML_ERROR, "bad mle: %p\n", mle);
+ dlm_print_one_mle(mle);
+ }
+ atomic_set(&mle->woken, 1);
+ wake_up(&mle->wq);
+
+ __dlm_unlink_mle(dlm, mle);
+ __dlm_mle_detach_hb_events(dlm, mle);
+ __dlm_put_mle(mle);
+ }
+ }
+ spin_unlock(&dlm->master_lock);
+ spin_unlock(&dlm->spinlock);
+}