kdbus: Upgrade driver to newer upstream version

This commit upgrades kdbus ipc driver from v4 patchset, as posted on lkml for review by Greg Kroah-Hartman on Mar 09 2015 to commit 0c05fbdc82f from new upstream kdbus repository (git://github.com/systemd/kdbus). Summary of major changes: * message importer rewritten - considerably reduces internal message processing overhead, * name registration reworked to follow DBus Specification precisely, * attached metadata now follow /proc access checks * reduced in-kernel stack buffer to 256 bytes for small messages Change-Id: I6d849173b4289e1b684ed1a9b48e6e0b361e5d53
author: Karol Lewandowski <k.lewandowsk@samsung.com> 2016-10-06 14:20:24 +0200
committer: Seung-Woo Kim <sw0312.kim@samsung.com> 2016-12-14 13:53:39 +0900
commit: d8934e9b05b82e73a3c78421f372197801c1785e (patch)
tree: 24981389ae0ca646a7997f6e32ea7a39aea147b7 /ipc
parent: ea4654706ad61bc7afc47e6321c5cbee8c175688 (diff)
31 files changed, 2826 insertions, 2497 deletions
diff --git a/ipc/kdbus/bus.c b/ipc/kdbus/bus.c
index cfe997fb4915..a67f825bdeaf 100644
--- a/ipc/kdbus/bus.c
+++ b/ipc/kdbus/bus.c
@@ -66,23 +66,16 @@ static struct kdbus_bus *kdbus_bus_new(struct kdbus_domain *domain,
 				       const char *name,
 				       struct kdbus_bloom_parameter *bloom,
 				       const u64 *pattach_owner,
-				       const u64 *pattach_recv,
 				       u64 flags, kuid_t uid, kgid_t gid)
 {
 	struct kdbus_bus *b;
 	u64 attach_owner;
-	u64 attach_recv;
 	int ret;
 
 	if (bloom->size < 8 || bloom->size > KDBUS_BUS_BLOOM_MAX_SIZE ||
 	    !KDBUS_IS_ALIGNED8(bloom->size) || bloom->n_hash < 1)
 		return ERR_PTR(-EINVAL);
 
-	ret = kdbus_sanitize_attach_flags(pattach_recv ? *pattach_recv : 0,
-					  &attach_recv);
-	if (ret < 0)
-		return ERR_PTR(ret);
-
 	ret = kdbus_sanitize_attach_flags(pattach_owner ? *pattach_owner : 0,
 					  &attach_owner);
 	if (ret < 0)
@@ -111,7 +104,6 @@ static struct kdbus_bus *kdbus_bus_new(struct kdbus_domain *domain,
 
 	b->id = atomic64_inc_return(&domain->last_id);
 	b->bus_flags = flags;
-	b->attach_flags_req = attach_recv;
 	b->attach_flags_owner = attach_owner;
 	generate_random_uuid(b->id128);
 	b->bloom = *bloom;
@@ -240,9 +232,9 @@ struct kdbus_conn *kdbus_bus_find_conn_by_id(struct kdbus_bus *bus, u64 id)
  * kdbus_bus_broadcast() - send a message to all subscribed connections
  * @bus:	The bus the connections are connected to
  * @conn_src:	The source connection, may be %NULL for kernel notifications
- * @kmsg:	The message to send.
+ * @staging:	Staging object containing the message to send
  *
- * Send @kmsg to all connections that are currently active on the bus.
+ * Send message to all connections that are currently active on the bus.
  * Connections must still have matches installed in order to let the message
  * pass.
  *
@@ -250,7 +242,7 @@ struct kdbus_conn *kdbus_bus_find_conn_by_id(struct kdbus_bus *bus, u64 id)
  */
 void kdbus_bus_broadcast(struct kdbus_bus *bus,
 			 struct kdbus_conn *conn_src,
-			 struct kdbus_kmsg *kmsg)
+			 struct kdbus_staging *staging)
 {
 	struct kdbus_conn *conn_dst;
 	unsigned int i;
@@ -267,7 +259,7 @@ void kdbus_bus_broadcast(struct kdbus_bus *bus,
 	 * can re-construct order via sequence numbers), but we should at least
 	 * try to avoid re-ordering for monitors.
 	 */
-	kdbus_bus_eavesdrop(bus, conn_src, kmsg);
+	kdbus_bus_eavesdrop(bus, conn_src, staging);
 
 	down_read(&bus->conn_rwlock);
 	hash_for_each(bus->conn_hash, i, conn_dst, hentry) {
@@ -278,13 +270,11 @@ void kdbus_bus_broadcast(struct kdbus_bus *bus,
 		 * Check if there is a match for the kmsg object in
 		 * the destination connection match db
 		 */
-		if (!kdbus_match_db_match_kmsg(conn_dst->match_db, conn_src,
-					       kmsg))
+		if (!kdbus_match_db_match_msg(conn_dst->match_db, conn_src,
+					      staging))
 			continue;
 
 		if (conn_src) {
-			u64 attach_flags;
-
 			/*
 			 * Anyone can send broadcasts, as they have no
 			 * destination. But a receiver needs TALK access to
@@ -292,20 +282,6 @@ void kdbus_bus_broadcast(struct kdbus_bus *bus,
 			 */
 			if (!kdbus_conn_policy_talk(conn_dst, NULL, conn_src))
 				continue;
-
-			attach_flags = kdbus_meta_calc_attach_flags(conn_src,
-								    conn_dst);
-
-			/*
-			 * Keep sending messages even if we cannot acquire the
-			 * requested metadata. It's up to the receiver to drop
-			 * messages that lack expected metadata.
-			 */
-			if (!conn_src->faked_meta)
-				kdbus_meta_proc_collect(kmsg->proc_meta,
-							attach_flags);
-			kdbus_meta_conn_collect(kmsg->conn_meta, kmsg, conn_src,
-						attach_flags);
 		} else {
 			/*
 			 * Check if there is a policy db that prevents the
@@ -313,11 +289,12 @@ void kdbus_bus_broadcast(struct kdbus_bus *bus,
 			 * notification
 			 */
 			if (!kdbus_conn_policy_see_notification(conn_dst, NULL,
-								kmsg))
+								staging->msg))
 				continue;
 		}
 
-		ret = kdbus_conn_entry_insert(conn_src, conn_dst, kmsg, NULL);
+		ret = kdbus_conn_entry_insert(conn_src, conn_dst, staging,
+					      NULL, NULL);
 		if (ret < 0)
 			kdbus_conn_lost_message(conn_dst);
 	}
@@ -328,16 +305,16 @@ void kdbus_bus_broadcast(struct kdbus_bus *bus,
  * kdbus_bus_eavesdrop() - send a message to all subscribed monitors
  * @bus:	The bus the monitors are connected to
  * @conn_src:	The source connection, may be %NULL for kernel notifications
- * @kmsg:	The message to send.
+ * @staging:	Staging object containing the message to send
  *
- * Send @kmsg to all monitors that are currently active on the bus. Monitors
+ * Send message to all monitors that are currently active on the bus. Monitors
  * must still have matches installed in order to let the message pass.
  *
  * The caller must hold the name-registry lock of @bus.
  */
 void kdbus_bus_eavesdrop(struct kdbus_bus *bus,
 			 struct kdbus_conn *conn_src,
-			 struct kdbus_kmsg *kmsg)
+			 struct kdbus_staging *staging)
 {
 	struct kdbus_conn *conn_dst;
 	int ret;
@@ -351,25 +328,8 @@ void kdbus_bus_eavesdrop(struct kdbus_bus *bus,
 
 	down_read(&bus->conn_rwlock);
 	list_for_each_entry(conn_dst, &bus->monitors_list, monitor_entry) {
-		/*
-		 * Collect metadata requested by the destination connection.
-		 * Ignore errors, as receivers need to check metadata
-		 * availability, anyway. So it's still better to send messages
-		 * that lack data, than to skip it entirely.
-		 */
-		if (conn_src) {
-			u64 attach_flags;
-
-			attach_flags = kdbus_meta_calc_attach_flags(conn_src,
-								    conn_dst);
-			if (!conn_src->faked_meta)
-				kdbus_meta_proc_collect(kmsg->proc_meta,
-							attach_flags);
-			kdbus_meta_conn_collect(kmsg->conn_meta, kmsg, conn_src,
-						attach_flags);
-		}
-
-		ret = kdbus_conn_entry_insert(conn_src, conn_dst, kmsg, NULL);
+		ret = kdbus_conn_entry_insert(conn_src, conn_dst, staging,
+					      NULL, NULL);
 		if (ret < 0)
 			kdbus_conn_lost_message(conn_dst);
 	}
@@ -381,7 +341,7 @@ void kdbus_bus_eavesdrop(struct kdbus_bus *bus,
  * @domain:		domain to operate on
  * @argp:		command payload
  *
- * Return: Newly created bus on success, ERR_PTR on failure.
+ * Return: NULL or newly created bus on success, ERR_PTR on failure.
  */
 struct kdbus_bus *kdbus_cmd_bus_make(struct kdbus_domain *domain,
 				     void __user *argp)
@@ -396,7 +356,6 @@ struct kdbus_bus *kdbus_cmd_bus_make(struct kdbus_domain *domain,
 		{ .type = KDBUS_ITEM_MAKE_NAME, .mandatory = true },
 		{ .type = KDBUS_ITEM_BLOOM_PARAMETER, .mandatory = true },
 		{ .type = KDBUS_ITEM_ATTACH_FLAGS_SEND },
-		{ .type = KDBUS_ITEM_ATTACH_FLAGS_RECV },
 	};
 	struct kdbus_args args = {
 		.allowed_flags = KDBUS_FLAG_NEGOTIATE |
@@ -415,7 +374,6 @@ struct kdbus_bus *kdbus_cmd_bus_make(struct kdbus_domain *domain,
 	bus = kdbus_bus_new(domain,
 			    argv[1].item->str, &argv[2].item->bloom_parameter,
 			    argv[3].item ? argv[3].item->data64 : NULL,
-			    argv[4].item ? argv[4].item->data64 : NULL,
 			    cmd->flags, current_euid(), current_egid());
 	if (IS_ERR(bus)) {
 		ret = PTR_ERR(bus);
@@ -475,20 +433,19 @@ exit:
  * @conn:		connection to operate on
  * @argp:		command payload
  *
- * Return: 0 on success, negative error code on failure.
+ * Return: >=0 on success, negative error code on failure.
  */
 int kdbus_cmd_bus_creator_info(struct kdbus_conn *conn, void __user *argp)
 {
 	struct kdbus_cmd_info *cmd;
 	struct kdbus_bus *bus = conn->ep->bus;
 	struct kdbus_pool_slice *slice = NULL;
+	struct kdbus_item *meta_items = NULL;
 	struct kdbus_item_header item_hdr;
 	struct kdbus_info info = {};
-	size_t meta_size, name_len;
-	struct kvec kvec[5];
-	u64 hdr_size = 0;
-	u64 attach_flags;
-	size_t cnt = 0;
+	size_t meta_size, name_len, cnt = 0;
+	struct kvec kvec[6];
+	u64 attach_flags, size = 0;
 	int ret;
 
 	struct kdbus_arg argv[] = {
@@ -510,8 +467,8 @@ int kdbus_cmd_bus_creator_info(struct kdbus_conn *conn, void __user *argp)
 
 	attach_flags &= bus->attach_flags_owner;
 
-	ret = kdbus_meta_export_prepare(bus->creator_meta, NULL,
-					&attach_flags, &meta_size);
+	ret = kdbus_meta_emit(bus->creator_meta, NULL, NULL, conn,
+			      attach_flags, &meta_items, &meta_size);
 	if (ret < 0)
 		goto exit;
 
@@ -521,26 +478,25 @@ int kdbus_cmd_bus_creator_info(struct kdbus_conn *conn, void __user *argp)
 	item_hdr.type = KDBUS_ITEM_MAKE_NAME;
 	item_hdr.size = KDBUS_ITEM_HEADER_SIZE + name_len;
 
-	kdbus_kvec_set(&kvec[cnt++], &info, sizeof(info), &hdr_size);
-	kdbus_kvec_set(&kvec[cnt++], &item_hdr, sizeof(item_hdr), &hdr_size);
-	kdbus_kvec_set(&kvec[cnt++], bus->node.name, name_len, &hdr_size);
-	cnt += !!kdbus_kvec_pad(&kvec[cnt], &hdr_size);
+	kdbus_kvec_set(&kvec[cnt++], &info, sizeof(info), &size);
+	kdbus_kvec_set(&kvec[cnt++], &item_hdr, sizeof(item_hdr), &size);
+	kdbus_kvec_set(&kvec[cnt++], bus->node.name, name_len, &size);
+	cnt += !!kdbus_kvec_pad(&kvec[cnt], &size);
+	if (meta_size > 0) {
+		kdbus_kvec_set(&kvec[cnt++], meta_items, meta_size, &size);
+		cnt += !!kdbus_kvec_pad(&kvec[cnt], &size);
+	}
+
+	info.size = size;
 
-	slice = kdbus_pool_slice_alloc(conn->pool, hdr_size + meta_size, false);
+	slice = kdbus_pool_slice_alloc(conn->pool, size, false);
 	if (IS_ERR(slice)) {
 		ret = PTR_ERR(slice);
 		slice = NULL;
 		goto exit;
 	}
 
-	ret = kdbus_meta_export(bus->creator_meta, NULL, attach_flags,
-				slice, hdr_size, &meta_size);
-	if (ret < 0)
-		goto exit;
-
-	info.size = hdr_size + meta_size;
-
-	ret = kdbus_pool_slice_copy_kvec(slice, 0, kvec, cnt, hdr_size);
+	ret = kdbus_pool_slice_copy_kvec(slice, 0, kvec, cnt, size);
 	if (ret < 0)
 		goto exit;
 
@@ -553,6 +509,6 @@ int kdbus_cmd_bus_creator_info(struct kdbus_conn *conn, void __user *argp)
 
 exit:
 	kdbus_pool_slice_release(slice);
-
+	kfree(meta_items);
 	return kdbus_args_clear(&args, ret);
 }
diff --git a/ipc/kdbus/bus.h b/ipc/kdbus/bus.h
index 5bea5ef768f1..8c2acaed6258 100644
--- a/ipc/kdbus/bus.h
+++ b/ipc/kdbus/bus.h
@@ -29,7 +29,7 @@
 
 struct kdbus_conn;
 struct kdbus_domain;
-struct kdbus_kmsg;
+struct kdbus_staging;
 struct kdbus_user;
 
 /**
@@ -37,7 +37,6 @@ struct kdbus_user;
  * @node:		kdbus_node
  * @id:			ID of this bus in the domain
  * @bus_flags:		Simple pass-through flags from userspace to userspace
- * @attach_flags_req:	KDBUS_ATTACH_* flags required by connecting peers
  * @attach_flags_owner:	KDBUS_ATTACH_* flags of bus creator that other
  *			connections can see or query
  * @id128:		Unique random 128 bit ID of this bus
@@ -45,6 +44,7 @@ struct kdbus_user;
  * @domain:		Domain of this bus
  * @creator:		Creator of the bus
  * @creator_meta:	Meta information about the bus creator
+ * @last_message_id:	Last used message id
  * @policy_db:		Policy database for this bus
  * @name_registry:	Name registry of this bus
  * @conn_rwlock:	Read/Write lock for all lists of child connections
@@ -60,7 +60,6 @@ struct kdbus_bus {
 	/* static */
 	u64 id;
 	u64 bus_flags;
-	u64 attach_flags_req;
 	u64 attach_flags_owner;
 	u8 id128[16];
 	struct kdbus_bloom_parameter bloom;
@@ -69,6 +68,7 @@ struct kdbus_bus {
 	struct kdbus_meta_proc *creator_meta;
 
 	/* protected by own locks */
+	atomic64_t last_message_id;
 	struct kdbus_policy_db policy_db;
 	struct kdbus_name_registry *name_registry;
 
@@ -89,10 +89,10 @@ struct kdbus_bus *kdbus_bus_unref(struct kdbus_bus *bus);
 struct kdbus_conn *kdbus_bus_find_conn_by_id(struct kdbus_bus *bus, u64 id);
 void kdbus_bus_broadcast(struct kdbus_bus *bus,
 			 struct kdbus_conn *conn_src,
-			 struct kdbus_kmsg *kmsg);
+			 struct kdbus_staging *staging);
 void kdbus_bus_eavesdrop(struct kdbus_bus *bus,
 			 struct kdbus_conn *conn_src,
-			 struct kdbus_kmsg *kmsg);
+			 struct kdbus_staging *staging);
 
 struct kdbus_bus *kdbus_cmd_bus_make(struct kdbus_domain *domain,
 				     void __user *argp);
diff --git a/ipc/kdbus/connection.c b/ipc/kdbus/connection.c
index 8e994ea16e48..f80191e73868 100644
--- a/ipc/kdbus/connection.c
+++ b/ipc/kdbus/connection.c
@@ -53,11 +53,12 @@
 #define KDBUS_CONN_ACTIVE_NEW	(INT_MIN + 1)
 
 /* Disable internal kdbus policy - possibilities of connections to own, see and
- * talk to names are restricted by libdbuspolicy library
+ * talk to names are restricted by libdbuspolicy library and LSM hooks
  */
 #define DISABLE_KDBUS_POLICY
 
-static struct kdbus_conn *kdbus_conn_new(struct kdbus_ep *ep, bool privileged,
+static struct kdbus_conn *kdbus_conn_new(struct kdbus_ep *ep,
+					 struct file *file,
 					 struct kdbus_cmd_hello *hello,
 					 const char *name,
 					 const struct kdbus_creds *creds,
@@ -77,6 +78,8 @@ static struct kdbus_conn *kdbus_conn_new(struct kdbus_ep *ep, bool privileged,
 	bool is_policy_holder;
 	bool is_activator;
 	bool is_monitor;
+	bool privileged;
+	bool owner;
 	struct kvec kvec;
 	int ret;
 
@@ -86,6 +89,9 @@ static struct kdbus_conn *kdbus_conn_new(struct kdbus_ep *ep, bool privileged,
 		struct kdbus_bloom_parameter bloom;
 	} bloom_item;
 
+	privileged = kdbus_ep_is_privileged(ep, file);
+	owner = kdbus_ep_is_owner(ep, file);
+
 	is_monitor = hello->flags & KDBUS_HELLO_MONITOR;
 	is_activator = hello->flags & KDBUS_HELLO_ACTIVATOR;
 	is_policy_holder = hello->flags & KDBUS_HELLO_POLICY_HOLDER;
@@ -102,9 +108,9 @@ static struct kdbus_conn *kdbus_conn_new(struct kdbus_ep *ep, bool privileged,
 		return ERR_PTR(-EINVAL);
 	if (is_monitor && ep->user)
 		return ERR_PTR(-EOPNOTSUPP);
-	if (!privileged && (is_activator || is_policy_holder || is_monitor))
+	if (!owner && (is_activator || is_policy_holder || is_monitor))
 		return ERR_PTR(-EPERM);
-	if ((creds || pids || seclabel) && !privileged)
+	if (!owner && (creds || pids || seclabel))
 		return ERR_PTR(-EPERM);
 
 	ret = kdbus_sanitize_attach_flags(hello->attach_flags_send,
@@ -117,10 +123,6 @@ static struct kdbus_conn *kdbus_conn_new(struct kdbus_ep *ep, bool privileged,
 	if (ret < 0)
 		return ERR_PTR(ret);
 
-	/* The attach flags must always satisfy the bus requirements. */
-	if (bus->attach_flags_req & ~attach_flags_send)
-		return ERR_PTR(-ECONNREFUSED);
-
 	conn = kzalloc(sizeof(*conn), GFP_KERNEL);
 	if (!conn)
 		return ERR_PTR(-ENOMEM);
@@ -132,16 +134,17 @@ static struct kdbus_conn *kdbus_conn_new(struct kdbus_ep *ep, bool privileged,
 #endif
 	mutex_init(&conn->lock);
 	INIT_LIST_HEAD(&conn->names_list);
-	INIT_LIST_HEAD(&conn->names_queue_list);
 	INIT_LIST_HEAD(&conn->reply_list);
-	atomic_set(&conn->name_count, 0);
 	atomic_set(&conn->request_count, 0);
 	atomic_set(&conn->lost_count, 0);
 	INIT_DELAYED_WORK(&conn->work, kdbus_reply_list_scan_work);
-	conn->cred = get_current_cred();
+	conn->cred = get_cred(file->f_cred);
+	conn->pid = get_pid(task_pid(current));
+	get_fs_root(current->fs, &conn->root_path);
 	init_waitqueue_head(&conn->wait);
 	kdbus_queue_init(&conn->queue);
 	conn->privileged = privileged;
+	conn->owner = owner;
 	conn->ep = kdbus_ep_ref(ep);
 	conn->id = atomic64_inc_return(&bus->domain->last_id);
 	conn->flags = hello->flags;
@@ -178,22 +181,28 @@ static struct kdbus_conn *kdbus_conn_new(struct kdbus_ep *ep, bool privileged,
 	BUILD_BUG_ON(sizeof(bus->id128) != sizeof(hello->id128));
 	memcpy(hello->id128, bus->id128, sizeof(hello->id128));
 
-	conn->meta = kdbus_meta_proc_new();
-	if (IS_ERR(conn->meta)) {
-		ret = PTR_ERR(conn->meta);
-		conn->meta = NULL;
-		goto exit_unref;
-	}
-
 	/* privileged processes can impersonate somebody else */
 	if (creds || pids || seclabel) {
-		ret = kdbus_meta_proc_fake(conn->meta, creds, pids, seclabel);
-		if (ret < 0)
+		conn->meta_fake = kdbus_meta_fake_new();
+		if (IS_ERR(conn->meta_fake)) {
+			ret = PTR_ERR(conn->meta_fake);
+			conn->meta_fake = NULL;
 			goto exit_unref;
+		}
 
-		conn->faked_meta = true;
+		ret = kdbus_meta_fake_collect(conn->meta_fake,
+					      creds, pids, seclabel);
+		if (ret < 0)
+			goto exit_unref;
 	} else {
-		ret = kdbus_meta_proc_collect(conn->meta,
+		conn->meta_proc = kdbus_meta_proc_new();
+		if (IS_ERR(conn->meta_proc)) {
+			ret = PTR_ERR(conn->meta_proc);
+			conn->meta_proc = NULL;
+			goto exit_unref;
+		}
+
+		ret = kdbus_meta_proc_collect(conn->meta_proc,
 					      KDBUS_ATTACH_CREDS |
 					      KDBUS_ATTACH_PIDS |
 					      KDBUS_ATTACH_AUXGROUPS |
@@ -215,11 +224,21 @@ static struct kdbus_conn *kdbus_conn_new(struct kdbus_ep *ep, bool privileged,
 	 * Note that limits are always accounted against the real UID, not
 	 * the effective UID (cred->user always points to the accounting of
 	 * cred->uid, not cred->euid).
+	 * In case the caller is privileged, we allow changing the accounting
+	 * to the faked user.
 	 */
 	if (ep->user) {
 		conn->user = kdbus_user_ref(ep->user);
 	} else {
-		conn->user = kdbus_user_lookup(ep->bus->domain, current_uid());
+		kuid_t uid;
+
+		if (conn->meta_fake && uid_valid(conn->meta_fake->uid) &&
+		    conn->privileged)
+			uid = conn->meta_fake->uid;
+		else
+			uid = conn->cred->uid;
+
+		conn->user = kdbus_user_lookup(ep->bus->domain, uid);
 		if (IS_ERR(conn->user)) {
 			ret = PTR_ERR(conn->user);
 			conn->user = NULL;
@@ -268,7 +287,6 @@ static void __kdbus_conn_free(struct kref *kref)
 	WARN_ON(delayed_work_pending(&conn->work));
 	WARN_ON(!list_empty(&conn->queue.msg_list));
 	WARN_ON(!list_empty(&conn->names_list));
-	WARN_ON(!list_empty(&conn->names_queue_list));
 	WARN_ON(!list_empty(&conn->reply_list));
 
 	if (conn->user) {
@@ -276,10 +294,13 @@ static void __kdbus_conn_free(struct kref *kref)
 		kdbus_user_unref(conn->user);
 	}
 
-	kdbus_meta_proc_unref(conn->meta);
+	kdbus_meta_fake_free(conn->meta_fake);
+	kdbus_meta_proc_unref(conn->meta_proc);
 	kdbus_match_db_free(conn->match_db);
 	kdbus_pool_free(conn->pool);
 	kdbus_ep_unref(conn->ep);
+	path_put(&conn->root_path);
+	put_pid(conn->pid);
 	put_cred(conn->cred);
 	kfree(conn->description);
 	kfree(conn->quota);
@@ -437,7 +458,7 @@ static int kdbus_conn_connect(struct kdbus_conn *conn, const char *name)
 	 * directly, and won't cause any notifications.
 	 */
 	if (!kdbus_conn_is_monitor(conn)) {
-		ret = kdbus_notify_id_change(conn->ep->bus, KDBUS_ITEM_ID_ADD,
+		ret = kdbus_notify_id_change(bus, KDBUS_ITEM_ID_ADD,
 					     conn->id, conn->flags);
 		if (ret < 0)
 			goto exit_disconnect;
@@ -564,17 +585,16 @@ int kdbus_conn_disconnect(struct kdbus_conn *conn, bool ensure_queue_empty)
 	hash_for_each(bus->conn_hash, i, c, hentry) {
 		mutex_lock(&c->lock);
 		list_for_each_entry_safe(r, r_tmp, &c->reply_list, entry) {
-			if (r->reply_src == conn) {
-				if (r->sync) {
-					kdbus_sync_reply_wakeup(r, -EPIPE);
-					kdbus_reply_unlink(r);
-					continue;
-				}
+			if (r->reply_src != conn)
+				continue;
 
+			if (r->sync)
+				kdbus_sync_reply_wakeup(r, -EPIPE);
+			else
 				/* send a 'connection dead' notification */
 				kdbus_notify_reply_dead(bus, c->id, r->cookie);
-				kdbus_reply_unlink(r);
-			}
+
+			kdbus_reply_unlink(r);
 		}
 		mutex_unlock(&c->lock);
 	}
@@ -600,21 +620,22 @@ int kdbus_conn_disconnect(struct kdbus_conn *conn, bool ensure_queue_empty)
  */
 bool kdbus_conn_has_name(struct kdbus_conn *conn, const char *name)
 {
-	struct kdbus_name_entry *e;
+	struct kdbus_name_owner *owner;
 
 	lockdep_assert_held(&conn->ep->bus->name_registry->rwlock);
 
-	list_for_each_entry(e, &conn->names_list, conn_entry)
-		if (strcmp(e->name, name) == 0)
+	list_for_each_entry(owner, &conn->names_list, conn_entry)
+		if (!(owner->flags & KDBUS_NAME_IN_QUEUE) &&
+		    !strcmp(name, owner->name->name))
 			return true;
 
 	return false;
 }
 
 struct kdbus_quota {
-	uint32_t memory;
-	uint16_t msgs;
-	uint8_t fds;
+	u32 memory;
+	u16 msgs;
+	u8 fds;
 };
 
 /**
@@ -652,7 +673,7 @@ int kdbus_conn_quota_inc(struct kdbus_conn *c, struct kdbus_user *u,
 	 * allocation schemes. Furthermore, resource utilization should be
 	 * maximized, so only minimal resources stay reserved. However, we need
 	 * to adapt to a dynamic number of users, as we cannot know how many
-	 * users will talk to a connection. Therefore, the current allocations
+	 * users will talk to a connection. Therefore, the current allocation
 	 * works like this:
 	 * We limit the number of bytes in a destination's pool per sending
 	 * user. The space available for a user is 33% of the unused pool space
@@ -694,7 +715,7 @@ int kdbus_conn_quota_inc(struct kdbus_conn *c, struct kdbus_user *u,
 
 	/*
 	 * Pool owner slices are un-accounted slices; they can claim more
-	 * than 50% of the queue. However, the slice we're dealing with here
+	 * than 50% of the queue. However, the slices we're dealing with here
 	 * belong to the incoming queue, hence they are 'accounted' slices
 	 * to which the 50%-limit applies.
 	 */
@@ -758,7 +779,7 @@ void kdbus_conn_quota_dec(struct kdbus_conn *c, struct kdbus_user *u,
  *
  * kdbus is reliable. That means, we try hard to never lose messages. However,
  * memory is limited, so we cannot rely on transmissions to never fail.
- * Therefore, we use quota-limits to let callers know if there unicast message
+ * Therefore, we use quota-limits to let callers know if their unicast message
  * cannot be transmitted to a peer. This works fine for unicasts, but for
  * broadcasts we cannot make the caller handle the transmission failure.
  * Instead, we must let the destination know that it couldn't receive a
@@ -776,12 +797,10 @@ void kdbus_conn_lost_message(struct kdbus_conn *c)
 
 /* Callers should take the conn_dst lock */
 static struct kdbus_queue_entry *
-kdbus_conn_entry_make(struct kdbus_conn *conn_dst,
-		      const struct kdbus_kmsg *kmsg,
-		      struct kdbus_user *user)
+kdbus_conn_entry_make(struct kdbus_conn *conn_src,
+		      struct kdbus_conn *conn_dst,
+		      struct kdbus_staging *staging)
 {
-	struct kdbus_queue_entry *entry;
-
 	/* The remote connection was disconnected */
 	if (!kdbus_conn_active(conn_dst))
 		return ERR_PTR(-ECONNRESET);
@@ -795,14 +814,10 @@ kdbus_conn_entry_make(struct kdbus_conn *conn_dst,
 	 */
 	if (!kdbus_conn_is_monitor(conn_dst) &&
 	    !(conn_dst->flags & KDBUS_HELLO_ACCEPT_FD) &&
-	    kmsg->res && kmsg->res->fds_count > 0)
+	    staging->gaps && staging->gaps->n_fds > 0)
 		return ERR_PTR(-ECOMM);
 
-	entry = kdbus_queue_entry_new(conn_dst, kmsg, user);
-	if (IS_ERR(entry))
-		return entry;
-
-	return entry;
+	return kdbus_queue_entry_new(conn_src, conn_dst, staging);
 }
 
 /*
@@ -811,12 +826,11 @@ kdbus_conn_entry_make(struct kdbus_conn *conn_dst,
  * The connection's queue will never get to see it.
  */
 static int kdbus_conn_entry_sync_attach(struct kdbus_conn *conn_dst,
-					const struct kdbus_kmsg *kmsg,
+					struct kdbus_staging *staging,
 					struct kdbus_reply *reply_wake)
 {
 	struct kdbus_queue_entry *entry;
-	int remote_ret;
-	int ret = 0;
+	int remote_ret, ret = 0;
 
 	mutex_lock(&reply_wake->reply_dst->lock);
 
@@ -825,8 +839,8 @@ static int kdbus_conn_entry_sync_attach(struct kdbus_conn *conn_dst,
 	 * entry and attach it to the reply object
 	 */
 	if (reply_wake->waiting) {
-		entry = kdbus_conn_entry_make(conn_dst, kmsg,
-					      reply_wake->reply_src->user);
+		entry = kdbus_conn_entry_make(reply_wake->reply_src, conn_dst,
+					      staging);
 		if (IS_ERR(entry))
 			ret = PTR_ERR(entry);
 		else
@@ -867,23 +881,24 @@ static int kdbus_conn_entry_sync_attach(struct kdbus_conn *conn_dst,
  * kdbus_conn_entry_insert() - enqueue a message into the receiver's pool
  * @conn_src:		The sending connection
  * @conn_dst:		The connection to queue into
- * @kmsg:		The kmsg to queue
+ * @staging:		Message to send
  * @reply:		The reply tracker to attach to the queue entry
+ * @name:		Destination name this msg is sent to, or NULL
  *
  * Return: 0 on success. negative error otherwise.
  */
 int kdbus_conn_entry_insert(struct kdbus_conn *conn_src,
 			    struct kdbus_conn *conn_dst,
-			    const struct kdbus_kmsg *kmsg,
-			    struct kdbus_reply *reply)
+			    struct kdbus_staging *staging,
+			    struct kdbus_reply *reply,
+			    const struct kdbus_name_entry *name)
 {
 	struct kdbus_queue_entry *entry;
 	int ret;
 
 	kdbus_conn_lock2(conn_src, conn_dst);
 
-	entry = kdbus_conn_entry_make(conn_dst, kmsg,
-				      conn_src ? conn_src->user : NULL);
+	entry = kdbus_conn_entry_make(conn_src, conn_dst, staging);
 	if (IS_ERR(entry)) {
 		ret = PTR_ERR(entry);
 		goto exit_unlock;
@@ -895,6 +910,14 @@ int kdbus_conn_entry_insert(struct kdbus_conn *conn_src,
 			schedule_delayed_work(&conn_src->work, 0);
 	}
 
+	/*
+	 * Record the sequence number of the registered name; it will
+	 * be remembered by the queue, in case messages addressed to a
+	 * name need to be moved from or to an activator.
+	 */
+	if (name)
+		entry->dst_name_id = name->name_id;
+
 	kdbus_queue_entry_enqueue(entry, reply);
 	wake_up_interruptible(&conn_dst->wait);
 
@@ -1027,22 +1050,19 @@ static int kdbus_conn_wait_reply(struct kdbus_conn *conn_src,
 }
 
 static int kdbus_pin_dst(struct kdbus_bus *bus,
-			 struct kdbus_kmsg *kmsg,
+			 struct kdbus_staging *staging,
 			 struct kdbus_name_entry **out_name,
 			 struct kdbus_conn **out_dst)
 {
-	struct kdbus_msg_resources *res = kmsg->res;
+	const struct kdbus_msg *msg = staging->msg;
+	struct kdbus_name_owner *owner = NULL;
 	struct kdbus_name_entry *name = NULL;
 	struct kdbus_conn *dst = NULL;
-	struct kdbus_msg *msg = &kmsg->msg;
 	int ret;
 
-	if (WARN_ON(!res))
-		return -EINVAL;
-
 	lockdep_assert_held(&bus->name_registry->rwlock);
 
-	if (!res->dst_name) {
+	if (!staging->dst_name) {
 		dst = kdbus_bus_find_conn_by_id(bus, msg->dst_id);
 		if (!dst)
 			return -ENXIO;
@@ -1053,8 +1073,10 @@ static int kdbus_pin_dst(struct kdbus_bus *bus,
 		}
 	} else {
 		name = kdbus_name_lookup_unlocked(bus->name_registry,
-						  res->dst_name);
-		if (!name)
+						  staging->dst_name);
+		if (name)
+			owner = kdbus_name_get_owner(name);
+		if (!owner)
 			return -ESRCH;
 
 		/*
@@ -1066,26 +1088,14 @@ static int kdbus_pin_dst(struct kdbus_bus *bus,
 		 * owns the given name.
 		 */
 		if (msg->dst_id != KDBUS_DST_ID_NAME &&
-		    msg->dst_id != name->conn->id)
+		    msg->dst_id != owner->conn->id)
 			return -EREMCHG;
 
-		if (!name->conn && name->activator)
-			dst = kdbus_conn_ref(name->activator);
-		else
-			dst = kdbus_conn_ref(name->conn);
-
 		if ((msg->flags & KDBUS_MSG_NO_AUTO_START) &&
-		    kdbus_conn_is_activator(dst)) {
-			ret = -EADDRNOTAVAIL;
-			goto error;
-		}
+		    kdbus_conn_is_activator(owner->conn))
+			return -EADDRNOTAVAIL;
 
-		/*
-		 * Record the sequence number of the registered name; it will
-		 * be passed on to the queue, in case messages addressed to a
-		 * name need to be moved from or to an activator.
-		 */
-		kmsg->dst_name_id = name->name_id;
+		dst = kdbus_conn_ref(owner->conn);
 	}
 
 	*out_name = name;
@@ -1097,18 +1107,19 @@ error:
 	return ret;
 }
 
-static int kdbus_conn_reply(struct kdbus_conn *src, struct kdbus_kmsg *kmsg)
+static int kdbus_conn_reply(struct kdbus_conn *src,
+			    struct kdbus_staging *staging)
 {
+	const struct kdbus_msg *msg = staging->msg;
 	struct kdbus_name_entry *name = NULL;
 	struct kdbus_reply *reply, *wake = NULL;
 	struct kdbus_conn *dst = NULL;
 	struct kdbus_bus *bus = src->ep->bus;
-	u64 attach;
 	int ret;
 
-	if (WARN_ON(kmsg->msg.dst_id == KDBUS_DST_ID_BROADCAST) ||
-	    WARN_ON(kmsg->msg.flags & KDBUS_MSG_EXPECT_REPLY) ||
-	    WARN_ON(kmsg->msg.flags & KDBUS_MSG_SIGNAL))
+	if (WARN_ON(msg->dst_id == KDBUS_DST_ID_BROADCAST) ||
+	    WARN_ON(msg->flags & KDBUS_MSG_EXPECT_REPLY) ||
+	    WARN_ON(msg->flags & KDBUS_MSG_SIGNAL))
 		return -EINVAL;
 
 	/* name-registry must be locked for lookup *and* collecting data */
@@ -1116,12 +1127,12 @@ static int kdbus_conn_reply(struct kdbus_conn *src, struct kdbus_kmsg *kmsg)
 
 	/* find and pin destination */
 
-	ret = kdbus_pin_dst(bus, kmsg, &name, &dst);
+	ret = kdbus_pin_dst(bus, staging, &name, &dst);
 	if (ret < 0)
 		goto exit;
 
 	mutex_lock(&dst->lock);
-	reply = kdbus_reply_find(src, dst, kmsg->msg.cookie_reply);
+	reply = kdbus_reply_find(src, dst, msg->cookie_reply);
 	if (reply) {
 		if (reply->sync)
 			wake = kdbus_reply_ref(reply);
@@ -1129,28 +1140,19 @@ static int kdbus_conn_reply(struct kdbus_conn *src, struct kdbus_kmsg *kmsg)
 	}
 	mutex_unlock(&dst->lock);
 
-	/* attach metadata */
-
-	attach = kdbus_meta_calc_attach_flags(src, dst);
-
-	if (!src->faked_meta) {
-		ret = kdbus_meta_proc_collect(kmsg->proc_meta, attach);
-		if (ret < 0)
-			goto exit;
-	}
-
-	ret = kdbus_meta_conn_collect(kmsg->conn_meta, kmsg, src, attach);
-	if (ret < 0)
+	if (!reply) {
+		ret = -EBADSLT;
 		goto exit;
+	}
 
 	/* send message */
 
-	kdbus_bus_eavesdrop(bus, src, kmsg);
+	kdbus_bus_eavesdrop(bus, src, staging);
 
 	if (wake)
-		ret = kdbus_conn_entry_sync_attach(dst, kmsg, wake);
+		ret = kdbus_conn_entry_sync_attach(dst, staging, wake);
 	else
-		ret = kdbus_conn_entry_insert(src, dst, kmsg, NULL);
+		ret = kdbus_conn_entry_insert(src, dst, staging, NULL, name);
 
 exit:
 	up_read(&bus->name_registry->rwlock);
@@ -1160,25 +1162,25 @@ exit:
 }
 
 static struct kdbus_reply *kdbus_conn_call(struct kdbus_conn *src,
-					   struct kdbus_kmsg *kmsg,
+					   struct kdbus_staging *staging,
 					   ktime_t exp)
 {
+	const struct kdbus_msg *msg = staging->msg;
 	struct kdbus_name_entry *name = NULL;
 	struct kdbus_reply *wait = NULL;
 	struct kdbus_conn *dst = NULL;
 	struct kdbus_bus *bus = src->ep->bus;
-	u64 attach;
 	int ret;
 
-	if (WARN_ON(kmsg->msg.dst_id == KDBUS_DST_ID_BROADCAST) ||
-	    WARN_ON(kmsg->msg.flags & KDBUS_MSG_SIGNAL) ||
-	    WARN_ON(!(kmsg->msg.flags & KDBUS_MSG_EXPECT_REPLY)))
+	if (WARN_ON(msg->dst_id == KDBUS_DST_ID_BROADCAST) ||
+	    WARN_ON(msg->flags & KDBUS_MSG_SIGNAL) ||
+	    WARN_ON(!(msg->flags & KDBUS_MSG_EXPECT_REPLY)))
 		return ERR_PTR(-EINVAL);
 
 	/* resume previous wait-context, if available */
 
 	mutex_lock(&src->lock);
-	wait = kdbus_reply_find(NULL, src, kmsg->msg.cookie);
+	wait = kdbus_reply_find(NULL, src, msg->cookie);
 	if (wait) {
 		if (wait->interrupted) {
 			kdbus_reply_ref(wait);
@@ -1200,44 +1202,27 @@ static struct kdbus_reply *kdbus_conn_call(struct kdbus_conn *src,
 
 	/* find and pin destination */
 
-	ret = kdbus_pin_dst(bus, kmsg, &name, &dst);
+	ret = kdbus_pin_dst(bus, staging, &name, &dst);
 	if (ret < 0)
 		goto exit;
 
-	/* Disable internal kdbus policy - possibilities of connections to own,
-	 * see and talk to well-known names are restricted by libdbuspolicy
 	if (!kdbus_conn_policy_talk(src, current_cred(), dst)) {
 		ret = -EPERM;
 		goto exit;
 	}
-	*/
 
-	wait = kdbus_reply_new(dst, src, &kmsg->msg, name, true);
+	wait = kdbus_reply_new(dst, src, msg, name, true);
 	if (IS_ERR(wait)) {
 		ret = PTR_ERR(wait);
 		wait = NULL;
 		goto exit;
 	}
 
-	/* attach metadata */
-
-	attach = kdbus_meta_calc_attach_flags(src, dst);
-
-	if (!src->faked_meta) {
-		ret = kdbus_meta_proc_collect(kmsg->proc_meta, attach);
-		if (ret < 0)
-			goto exit;
-	}
-
-	ret = kdbus_meta_conn_collect(kmsg->conn_meta, kmsg, src, attach);
-	if (ret < 0)
-		goto exit;
-
 	/* send message */
 
-	kdbus_bus_eavesdrop(bus, src, kmsg);
+	kdbus_bus_eavesdrop(bus, src, staging);
 
-	ret = kdbus_conn_entry_insert(src, dst, kmsg, wait);
+	ret = kdbus_conn_entry_insert(src, dst, staging, wait, name);
 	if (ret < 0)
 		goto exit;
 
@@ -1253,19 +1238,20 @@ exit:
 	return wait;
 }
 
-static int kdbus_conn_unicast(struct kdbus_conn *src, struct kdbus_kmsg *kmsg)
+static int kdbus_conn_unicast(struct kdbus_conn *src,
+			      struct kdbus_staging *staging)
 {
+	const struct kdbus_msg *msg = staging->msg;
 	struct kdbus_name_entry *name = NULL;
 	struct kdbus_reply *wait = NULL;
 	struct kdbus_conn *dst = NULL;
 	struct kdbus_bus *bus = src->ep->bus;
-	bool is_signal = (kmsg->msg.flags & KDBUS_MSG_SIGNAL);
-	u64 attach;
+	bool is_signal = (msg->flags & KDBUS_MSG_SIGNAL);
 	int ret = 0;
 
-	if (WARN_ON(kmsg->msg.dst_id == KDBUS_DST_ID_BROADCAST) ||
-	    WARN_ON(!(kmsg->msg.flags & KDBUS_MSG_EXPECT_REPLY) &&
-		    kmsg->msg.cookie_reply != 0))
+	if (WARN_ON(msg->dst_id == KDBUS_DST_ID_BROADCAST) ||
+	    WARN_ON(!(msg->flags & KDBUS_MSG_EXPECT_REPLY) &&
+		    msg->cookie_reply != 0))
 		return -EINVAL;
 
 	/* name-registry must be locked for lookup *and* collecting data */
@@ -1273,23 +1259,23 @@ static int kdbus_conn_unicast(struct kdbus_conn *src, struct kdbus_kmsg *kmsg)
 
 	/* find and pin destination */
 
-	ret = kdbus_pin_dst(bus, kmsg, &name, &dst);
+	ret = kdbus_pin_dst(bus, staging, &name, &dst);
 	if (ret < 0)
 		goto exit;
 
 	if (is_signal) {
 		/* like broadcasts we eavesdrop even if the msg is dropped */
-		kdbus_bus_eavesdrop(bus, src, kmsg);
+		kdbus_bus_eavesdrop(bus, src, staging);
 
 		/* drop silently if peer is not interested or not privileged */
-		if (!kdbus_match_db_match_kmsg(dst->match_db, src, kmsg) ||
+		if (!kdbus_match_db_match_msg(dst->match_db, src, staging) ||
 		    !kdbus_conn_policy_talk(dst, NULL, src))
 			goto exit;
 	} else if (!kdbus_conn_policy_talk(src, current_cred(), dst)) {
 		ret = -EPERM;
 		goto exit;
-	} else if (kmsg->msg.flags & KDBUS_MSG_EXPECT_REPLY) {
-		wait = kdbus_reply_new(dst, src, &kmsg->msg, name, false);
+	} else if (msg->flags & KDBUS_MSG_EXPECT_REPLY) {
+		wait = kdbus_reply_new(dst, src, msg, name, false);
 		if (IS_ERR(wait)) {
 			ret = PTR_ERR(wait);
 			wait = NULL;
@@ -1297,26 +1283,12 @@ static int kdbus_conn_unicast(struct kdbus_conn *src, struct kdbus_kmsg *kmsg)
 		}
 	}
 
-	/* attach metadata */
-
-	attach = kdbus_meta_calc_attach_flags(src, dst);
-
-	if (!src->faked_meta) {
-		ret = kdbus_meta_proc_collect(kmsg->proc_meta, attach);
-		if (ret < 0 && !is_signal)
-			goto exit;
-	}
-
-	ret = kdbus_meta_conn_collect(kmsg->conn_meta, kmsg, src, attach);
-	if (ret < 0 && !is_signal)
-		goto exit;
-
 	/* send message */
 
 	if (!is_signal)
-		kdbus_bus_eavesdrop(bus, src, kmsg);
+		kdbus_bus_eavesdrop(bus, src, staging);
 
-	ret = kdbus_conn_entry_insert(src, dst, kmsg, wait);
+	ret = kdbus_conn_entry_insert(src, dst, staging, wait, name);
 	if (ret < 0 && !is_signal)
 		goto exit;
 
@@ -1386,7 +1358,7 @@ void kdbus_conn_move_messages(struct kdbus_conn *conn_dst,
 			continue;
 
 		if (!(conn_dst->flags & KDBUS_HELLO_ACCEPT_FD) &&
-		    e->msg_res && e->msg_res->fds_count > 0) {
+		    e->gaps && e->gaps->n_fds > 0) {
 			kdbus_conn_lost_message(conn_dst);
 			kdbus_queue_entry_free(e);
 			continue;
@@ -1412,7 +1384,7 @@ static bool kdbus_conn_policy_query_all(struct kdbus_conn *conn,
 					struct kdbus_conn *whom,
 					unsigned int access)
 {
-	struct kdbus_name_entry *ne;
+	struct kdbus_name_owner *owner;
 	bool pass = false;
 	int res;
 
@@ -1421,10 +1393,14 @@ static bool kdbus_conn_policy_query_all(struct kdbus_conn *conn,
 	down_read(&db->entries_rwlock);
 	mutex_lock(&whom->lock);
 
-	list_for_each_entry(ne, &whom->names_list, conn_entry) {
-		res = kdbus_policy_query_unlocked(db, conn_creds ? : conn->cred,
-						  ne->name,
-						  kdbus_strhash(ne->name));
+	list_for_each_entry(owner, &whom->names_list, conn_entry) {
+		if (owner->flags & KDBUS_NAME_IN_QUEUE)
+			continue;
+
+		res = kdbus_policy_query_unlocked(db,
+					conn_creds ? : conn->cred,
+					owner->name->name,
+					kdbus_strhash(owner->name->name));
 		if (res >= (int)access) {
 			pass = true;
 			break;
@@ -1468,7 +1444,7 @@ bool kdbus_conn_policy_own_name(struct kdbus_conn *conn,
 			return false;
 	}
 
-	if (conn->privileged)
+	if (conn->owner)
 		return true;
 
 	res = kdbus_policy_query(&conn->ep->bus->policy_db, conn_creds,
@@ -1503,7 +1479,7 @@ bool kdbus_conn_policy_talk(struct kdbus_conn *conn,
 					 to, KDBUS_POLICY_TALK))
 		return false;
 
-	if (conn->privileged)
+	if (conn->owner)
 		return true;
 	if (uid_eq(conn_creds->euid, to->cred->uid))
 		return true;
@@ -1588,19 +1564,16 @@ static bool kdbus_conn_policy_see(struct kdbus_conn *conn,
  *					  receive a given kernel notification
  * @conn:		Connection
  * @conn_creds:		Credentials of @conn to use for policy check
- * @kmsg:		The message carrying the notification
+ * @msg:		Notification message
  *
- * This checks whether @conn is allowed to see the kernel notification @kmsg.
+ * This checks whether @conn is allowed to see the kernel notification.
  *
  * Return: true if allowed, false if not.
  */
 bool kdbus_conn_policy_see_notification(struct kdbus_conn *conn,
 					const struct cred *conn_creds,
-					const struct kdbus_kmsg *kmsg)
+					const struct kdbus_msg *msg)
 {
-	if (WARN_ON(kmsg->msg.src_id != KDBUS_SRC_ID_KERNEL))
-		return false;
-
 	/*
 	 * Depending on the notification type, broadcasted kernel notifications
 	 * have to be filtered:
@@ -1609,26 +1582,24 @@ bool kdbus_conn_policy_see_notification(struct kdbus_conn *conn,
 	 *     to a peer if, and only if, that peer can see the name this
 	 *     notification is for.
 	 *
-	 * KDBUS_ITEM_ID_{ADD,REMOVE}: As new peers cannot have names, and all
-	 *     names are dropped before a peer is removed, those notifications
-	 *     cannot be seen on custom endpoints. Thus, we only pass them
-	 *     through on default endpoints.
+	 * KDBUS_ITEM_ID_{ADD,REMOVE}: Notifications for ID changes are
+	 *     broadcast to everyone, to allow tracking peers.
 	 */
 
-	switch (kmsg->notify_type) {
+	switch (msg->items[0].type) {
 	case KDBUS_ITEM_NAME_ADD:
 	case KDBUS_ITEM_NAME_REMOVE:
 	case KDBUS_ITEM_NAME_CHANGE:
 		return kdbus_conn_policy_see_name(conn, conn_creds,
-						  kmsg->notify_name);
+					msg->items[0].name_change.name);
 
 	case KDBUS_ITEM_ID_ADD:
 	case KDBUS_ITEM_ID_REMOVE:
-		return !conn->ep->user;
+		return true;
 
 	default:
 		WARN(1, "Invalid type for notification broadcast: %llu\n",
-		     (unsigned long long)kmsg->notify_type);
+		     (unsigned long long)msg->items[0].type);
 		return false;
 	}
 }
@@ -1636,12 +1607,12 @@ bool kdbus_conn_policy_see_notification(struct kdbus_conn *conn,
 /**
  * kdbus_cmd_hello() - handle KDBUS_CMD_HELLO
  * @ep:			Endpoint to operate on
- * @privileged:		Whether the caller is privileged
+ * @file:		File this connection is opened on
  * @argp:		Command payload
  *
- * Return: Newly created connection on success, ERR_PTR on failure.
+ * Return: NULL or newly created connection on success, ERR_PTR on failure.
  */
-struct kdbus_conn *kdbus_cmd_hello(struct kdbus_ep *ep, bool privileged,
+struct kdbus_conn *kdbus_cmd_hello(struct kdbus_ep *ep, struct file *file,
 				   void __user *argp)
 {
 	struct kdbus_cmd_hello *cmd;
@@ -1676,7 +1647,7 @@ struct kdbus_conn *kdbus_cmd_hello(struct kdbus_ep *ep, bool privileged,
 
 	item_name = argv[1].item ? argv[1].item->str : NULL;
 
-	c = kdbus_conn_new(ep, privileged, cmd, item_name,
+	c = kdbus_conn_new(ep, file, cmd, item_name,
 			   argv[2].item ? &argv[2].item->creds : NULL,
 			   argv[3].item ? &argv[3].item->pids : NULL,
 			   argv[4].item ? argv[4].item->str : NULL,
@@ -1726,7 +1697,7 @@ exit:
  *
  * The caller must not hold any active reference to @conn or this will deadlock.
  *
- * Return: 0 on success, negative error code on failure.
+ * Return: >=0 on success, negative error code on failure.
  */
 int kdbus_cmd_byebye_unlocked(struct kdbus_conn *conn, void __user *argp)
 {
@@ -1758,21 +1729,23 @@ int kdbus_cmd_byebye_unlocked(struct kdbus_conn *conn, void __user *argp)
  * @conn:		connection to operate on
  * @argp:		command payload
  *
- * Return: 0 on success, negative error code on failure.
+ * Return: >=0 on success, negative error code on failure.
  */
 int kdbus_cmd_conn_info(struct kdbus_conn *conn, void __user *argp)
 {
 	struct kdbus_meta_conn *conn_meta = NULL;
 	struct kdbus_pool_slice *slice = NULL;
 	struct kdbus_name_entry *entry = NULL;
+	struct kdbus_name_owner *owner = NULL;
 	struct kdbus_conn *owner_conn = NULL;
+	struct kdbus_item *meta_items = NULL;
 	struct kdbus_info info = {};
 	struct kdbus_cmd_info *cmd;
 	struct kdbus_bus *bus = conn->ep->bus;
-	struct kvec kvec;
-	size_t meta_size;
+	struct kvec kvec[3];
+	size_t meta_size, cnt = 0;
 	const char *name;
-	u64 attach_flags;
+	u64 attach_flags, size = 0;
 	int ret;
 
 	struct kdbus_arg argv[] = {
@@ -1800,15 +1773,17 @@ int kdbus_cmd_conn_info(struct kdbus_conn *conn, void __user *argp)
 
 	if (name) {
 		entry = kdbus_name_lookup_unlocked(bus->name_registry, name);
-		if (!entry || !entry->conn ||
+		if (entry)
+			owner = kdbus_name_get_owner(entry);
+		if (!owner ||
 		    !kdbus_conn_policy_see_name(conn, current_cred(), name) ||
-		    (cmd->id != 0 && entry->conn->id != cmd->id)) {
+		    (cmd->id != 0 && owner->conn->id != cmd->id)) {
 			/* pretend a name doesn't exist if you cannot see it */
 			ret = -ESRCH;
 			goto exit;
 		}
 
-		owner_conn = kdbus_conn_ref(entry->conn);
+		owner_conn = kdbus_conn_ref(owner->conn);
 	} else if (cmd->id > 0) {
 		owner_conn = kdbus_bus_find_conn_by_id(bus, cmd->id);
 		if (!owner_conn || !kdbus_conn_policy_see(conn, current_cred(),
@@ -1822,10 +1797,6 @@ int kdbus_cmd_conn_info(struct kdbus_conn *conn, void __user *argp)
 		goto exit;
 	}
 
-	info.id = owner_conn->id;
-	info.flags = owner_conn->flags;
-	kdbus_kvec_set(&kvec, &info, sizeof(info), &info.size);
-
 	attach_flags &= atomic64_read(&owner_conn->attach_flags_send);
 
 	conn_meta = kdbus_meta_conn_new();
@@ -1835,32 +1806,35 @@ int kdbus_cmd_conn_info(struct kdbus_conn *conn, void __user *argp)
 		goto exit;
 	}
 
-	ret = kdbus_meta_conn_collect(conn_meta, NULL, owner_conn,
-				      attach_flags);
+	ret = kdbus_meta_conn_collect(conn_meta, owner_conn, 0, attach_flags);
 	if (ret < 0)
 		goto exit;
 
-	ret = kdbus_meta_export_prepare(owner_conn->meta, conn_meta,
-					&attach_flags, &meta_size);
+	ret = kdbus_meta_emit(owner_conn->meta_proc, owner_conn->meta_fake,
+			      conn_meta, conn, attach_flags,
+			      &meta_items, &meta_size);
 	if (ret < 0)
 		goto exit;
 
-	slice = kdbus_pool_slice_alloc(conn->pool,
-				       info.size + meta_size, false);
+	info.id = owner_conn->id;
+	info.flags = owner_conn->flags;
+
+	kdbus_kvec_set(&kvec[cnt++], &info, sizeof(info), &size);
+	if (meta_size > 0) {
+		kdbus_kvec_set(&kvec[cnt++], meta_items, meta_size, &size);
+		cnt += !!kdbus_kvec_pad(&kvec[cnt], &size);
+	}
+
+	info.size = size;
+
+	slice = kdbus_pool_slice_alloc(conn->pool, size, false);
 	if (IS_ERR(slice)) {
 		ret = PTR_ERR(slice);
 		slice = NULL;
 		goto exit;
 	}
 
-	ret = kdbus_meta_export(owner_conn->meta, conn_meta, attach_flags,
-				slice, sizeof(info), &meta_size);
-	if (ret < 0)
-		goto exit;
-
-	info.size += meta_size;
-
-	ret = kdbus_pool_slice_copy_kvec(slice, 0, &kvec, 1, sizeof(info));
+	ret = kdbus_pool_slice_copy_kvec(slice, 0, kvec, cnt, size);
 	if (ret < 0)
 		goto exit;
 
@@ -1878,6 +1852,7 @@ int kdbus_cmd_conn_info(struct kdbus_conn *conn, void __user *argp)
 exit:
 	up_read(&bus->name_registry->rwlock);
 	kdbus_pool_slice_release(slice);
+	kfree(meta_items);
 	kdbus_meta_conn_unref(conn_meta);
 	kdbus_conn_unref(owner_conn);
 	return kdbus_args_clear(&args, ret);
@@ -1888,11 +1863,10 @@ exit:
  * @conn:		connection to operate on
  * @argp:		command payload
  *
- * Return: 0 on success, negative error code on failure.
+ * Return: >=0 on success, negative error code on failure.
  */
 int kdbus_cmd_update(struct kdbus_conn *conn, void __user *argp)
 {
-	struct kdbus_bus *bus = conn->ep->bus;
 	struct kdbus_item *item_policy;
 	u64 *item_attach_send = NULL;
 	u64 *item_attach_recv = NULL;
@@ -1933,11 +1907,6 @@ int kdbus_cmd_update(struct kdbus_conn *conn, void __user *argp)
 						  &attach_send);
 		if (ret < 0)
 			goto exit;
-
-		if (bus->attach_flags_req & ~attach_send) {
-			ret = -EINVAL;
-			goto exit;
-		}
 	}
 
 	if (item_attach_recv) {
@@ -1985,15 +1954,17 @@ exit:
  * @f:			file this command was called on
  * @argp:		command payload
  *
- * Return: 0 on success, negative error code on failure.
+ * Return: >=0 on success, negative error code on failure.
  */
 int kdbus_cmd_send(struct kdbus_conn *conn, struct file *f, void __user *argp)
 {
 	struct kdbus_cmd_send *cmd;
-	struct kdbus_kmsg *kmsg = NULL;
+	struct kdbus_staging *staging = NULL;
+	struct kdbus_msg *msg = NULL;
 	struct file *cancel_fd = NULL;
-	int ret;
+	int ret, ret2;
 
+	/* command arguments */
 	struct kdbus_arg argv[] = {
 		{ .type = KDBUS_ITEM_NEGOTIATE },
 		{ .type = KDBUS_ITEM_CANCEL_FD },
@@ -2005,12 +1976,48 @@ int kdbus_cmd_send(struct kdbus_conn *conn, struct file *f, void __user *argp)
 		.argc = ARRAY_SIZE(argv),
 	};
 
+	/* message arguments */
+	struct kdbus_arg msg_argv[] = {
+		{ .type = KDBUS_ITEM_NEGOTIATE },
+		{ .type = KDBUS_ITEM_PAYLOAD_VEC, .multiple = true },
+		{ .type = KDBUS_ITEM_PAYLOAD_MEMFD, .multiple = true },
+		{ .type = KDBUS_ITEM_FDS },
+		{ .type = KDBUS_ITEM_BLOOM_FILTER },
+		{ .type = KDBUS_ITEM_DST_NAME },
+	};
+	struct kdbus_args msg_args = {
+		.allowed_flags = KDBUS_FLAG_NEGOTIATE |
+				 KDBUS_MSG_EXPECT_REPLY |
+				 KDBUS_MSG_NO_AUTO_START |
+				 KDBUS_MSG_SIGNAL,
+		.argv = msg_argv,
+		.argc = ARRAY_SIZE(msg_argv),
+	};
+
 	if (!kdbus_conn_is_ordinary(conn))
 		return -EOPNOTSUPP;
 
+	/* make sure to parse both, @cmd and @msg on negotiation */
+
 	ret = kdbus_args_parse(&args, argp, &cmd);
-	if (ret != 0)
-		return ret;
+	if (ret < 0)
+		goto exit;
+	else if (ret > 0 && !cmd->msg_address) /* negotiation without msg */
+		goto exit;
+
+	ret2 = kdbus_args_parse_msg(&msg_args, KDBUS_PTR(cmd->msg_address),
+				    &msg);
+	if (ret2 < 0) { /* cannot parse message */
+		ret = ret2;
+		goto exit;
+	} else if (ret2 > 0 && !ret) { /* msg-negot implies cmd-negot */
+		ret = -EINVAL;
+		goto exit;
+	} else if (ret > 0) { /* negotiation */
+		goto exit;
+	}
+
+	/* here we parsed both, @cmd and @msg, and neither wants negotiation */
 
 	cmd->reply.return_flags = 0;
 	kdbus_pool_publish_empty(conn->pool, &cmd->reply.offset,
@@ -2029,23 +2036,30 @@ int kdbus_cmd_send(struct kdbus_conn *conn, struct file *f, void __user *argp)
 		}
 	}
 
-	kmsg = kdbus_kmsg_new_from_cmd(conn, cmd);
-	if (IS_ERR(kmsg)) {
-		ret = PTR_ERR(kmsg);
-		kmsg = NULL;
+	/* patch-in the source of this message */
+	if (msg->src_id > 0 && msg->src_id != conn->id) {
+		ret = -EINVAL;
+		goto exit;
+	}
+	msg->src_id = conn->id;
+
+	staging = kdbus_staging_new_user(conn->ep->bus, cmd, msg);
+	if (IS_ERR(staging)) {
+		ret = PTR_ERR(staging);
+		staging = NULL;
 		goto exit;
 	}
 
-	if (kmsg->msg.dst_id == KDBUS_DST_ID_BROADCAST) {
+	if (msg->dst_id == KDBUS_DST_ID_BROADCAST) {
 		down_read(&conn->ep->bus->name_registry->rwlock);
-		kdbus_bus_broadcast(conn->ep->bus, conn, kmsg);
+		kdbus_bus_broadcast(conn->ep->bus, conn, staging);
 		up_read(&conn->ep->bus->name_registry->rwlock);
 	} else if (cmd->flags & KDBUS_SEND_SYNC_REPLY) {
 		struct kdbus_reply *r;
 		ktime_t exp;
 
-		exp = ns_to_ktime(kmsg->msg.timeout_ns);
-		r = kdbus_conn_call(conn, kmsg, exp);
+		exp = ns_to_ktime(msg->timeout_ns);
+		r = kdbus_conn_call(conn, staging, exp);
 		if (IS_ERR(r)) {
 			ret = PTR_ERR(r);
 			goto exit;
@@ -2055,13 +2069,13 @@ int kdbus_cmd_send(struct kdbus_conn *conn, struct file *f, void __user *argp)
 		kdbus_reply_unref(r);
 		if (ret < 0)
 			goto exit;
-	} else if ((kmsg->msg.flags & KDBUS_MSG_EXPECT_REPLY) ||
-		   kmsg->msg.cookie_reply == 0) {
-		ret = kdbus_conn_unicast(conn, kmsg);
+	} else if ((msg->flags & KDBUS_MSG_EXPECT_REPLY) ||
+		   msg->cookie_reply == 0) {
+		ret = kdbus_conn_unicast(conn, staging);
 		if (ret < 0)
 			goto exit;
 	} else {
-		ret = kdbus_conn_reply(conn, kmsg);
+		ret = kdbus_conn_reply(conn, staging);
 		if (ret < 0)
 			goto exit;
 	}
@@ -2072,7 +2086,8 @@ int kdbus_cmd_send(struct kdbus_conn *conn, struct file *f, void __user *argp)
 exit:
 	if (cancel_fd)
 		fput(cancel_fd);
-	kdbus_kmsg_free(kmsg);
+	kdbus_staging_free(staging);
+	ret = kdbus_args_clear(&msg_args, ret);
 	return kdbus_args_clear(&args, ret);
 }
 
@@ -2081,7 +2096,7 @@ exit:
  * @conn:		connection to operate on
  * @argp:		command payload
  *
- * Return: 0 on success, negative error code on failure.
+ * Return: >=0 on success, negative error code on failure.
  */
 int kdbus_cmd_recv(struct kdbus_conn *conn, void __user *argp)
 {
@@ -2204,7 +2219,7 @@ exit:
  * @conn:		connection to operate on
  * @argp:		command payload
  *
- * Return: 0 on success, negative error code on failure.
+ * Return: >=0 on success, negative error code on failure.
  */
 int kdbus_cmd_free(struct kdbus_conn *conn, void __user *argp)
 {
diff --git a/ipc/kdbus/connection.h b/ipc/kdbus/connection.h
index d1ffe909cb31..1ad082014faa 100644
--- a/ipc/kdbus/connection.h
+++ b/ipc/kdbus/connection.h
@@ -30,8 +30,9 @@
 					 KDBUS_HELLO_POLICY_HOLDER | \
 					 KDBUS_HELLO_MONITOR)
 
+struct kdbus_name_entry;
 struct kdbus_quota;
-struct kdbus_kmsg;
+struct kdbus_staging;
 
 /**
  * struct kdbus_conn - connection to a bus
@@ -54,12 +55,13 @@ struct kdbus_kmsg;
  * @work:		Delayed work to handle timeouts
  *			activator for
  * @match_db:		Subscription filter to broadcast messages
- * @meta:		Active connection creator's metadata/credentials,
- *			either from the handle or from HELLO
+ * @meta_proc:		Process metadata of connection creator, or NULL
+ * @meta_fake:		Faked metadata, or NULL
  * @pool:		The user's buffer to receive messages
  * @user:		Owner of the connection
  * @cred:		The credentials of the connection at creation time
- * @name_count:		Number of owned well-known names
+ * @pid:		Pid at creation time
+ * @root_path:		Root path at creation time
  * @request_count:	Number of pending requests issued by this
  *			connection that are waiting for replies from
  *			other peers
@@ -68,11 +70,10 @@ struct kdbus_kmsg;
  * @queue:		The message queue associated with this connection
  * @quota:		Array of per-user quota indexed by user->id
  * @n_quota:		Number of elements in quota array
- * @activator_of:	Well-known name entry this connection acts as an
  * @names_list:		List of well-known names
- * @names_queue_list:	Well-known names this connection waits for
- * @privileged:		Whether this connection is privileged on the bus
- * @faked_meta:		Whether the metadata was faked on HELLO
+ * @name_count:		Number of owned well-known names
+ * @privileged:		Whether this connection is privileged on the domain
+ * @owner:		Owned by the same user as the bus owner
  */
 struct kdbus_conn {
 	struct kref kref;
@@ -93,11 +94,13 @@ struct kdbus_conn {
 	struct list_head reply_list;
 	struct delayed_work work;
 	struct kdbus_match_db *match_db;
-	struct kdbus_meta_proc *meta;
+	struct kdbus_meta_proc *meta_proc;
+	struct kdbus_meta_fake *meta_fake;
 	struct kdbus_pool *pool;
 	struct kdbus_user *user;
 	const struct cred *cred;
-	atomic_t name_count;
+	struct pid *pid;
+	struct path root_path;
 	atomic_t request_count;
 	atomic_t lost_count;
 	wait_queue_head_t wait;
@@ -107,12 +110,11 @@ struct kdbus_conn {
 	unsigned int n_quota;
 
 	/* protected by registry->rwlock */
-	struct kdbus_name_entry *activator_of;
 	struct list_head names_list;
-	struct list_head names_queue_list;
+	unsigned int name_count;
 
 	bool privileged:1;
-	bool faked_meta:1;
+	bool owner:1;
 };
 
 struct kdbus_conn *kdbus_conn_ref(struct kdbus_conn *conn);
@@ -129,8 +131,9 @@ void kdbus_conn_quota_dec(struct kdbus_conn *c, struct kdbus_user *u,
 void kdbus_conn_lost_message(struct kdbus_conn *c);
 int kdbus_conn_entry_insert(struct kdbus_conn *conn_src,
 			    struct kdbus_conn *conn_dst,
-			    const struct kdbus_kmsg *kmsg,
-			    struct kdbus_reply *reply);
+			    struct kdbus_staging *staging,
+			    struct kdbus_reply *reply,
+			    const struct kdbus_name_entry *name);
 void kdbus_conn_move_messages(struct kdbus_conn *conn_dst,
 			      struct kdbus_conn *conn_src,
 			      u64 name_id);
@@ -147,10 +150,10 @@ bool kdbus_conn_policy_see_name_unlocked(struct kdbus_conn *conn,
 					 const char *name);
 bool kdbus_conn_policy_see_notification(struct kdbus_conn *conn,
 					const struct cred *curr_creds,
-					const struct kdbus_kmsg *kmsg);
+					const struct kdbus_msg *msg);
 
 /* command dispatcher */
-struct kdbus_conn *kdbus_cmd_hello(struct kdbus_ep *ep, bool privileged,
+struct kdbus_conn *kdbus_cmd_hello(struct kdbus_ep *ep, struct file *file,
 				   void __user *argp);
 int kdbus_cmd_byebye_unlocked(struct kdbus_conn *conn, void __user *argp);
 int kdbus_cmd_conn_info(struct kdbus_conn *conn, void __user *argp);
diff --git a/ipc/kdbus/endpoint.c b/ipc/kdbus/endpoint.c
index 174d274b113e..44e7a20de9c2 100644
--- a/ipc/kdbus/endpoint.c
+++ b/ipc/kdbus/endpoint.c
@@ -78,7 +78,7 @@ static void kdbus_ep_release(struct kdbus_node *node, bool was_active)
  * @gid:		The gid of the node
  * @is_custom:		Whether this is a custom endpoint
  *
- * This function will create a new enpoint with the given
+ * This function will create a new endpoint with the given
  * name and properties for a given bus.
  *
  * Return: a new kdbus_ep on success, ERR_PTR on failure.
@@ -184,11 +184,39 @@ struct kdbus_ep *kdbus_ep_unref(struct kdbus_ep *ep)
 }
 
 /**
+ * kdbus_ep_is_privileged() - check whether a file is privileged
+ * @ep:		endpoint to operate on
+ * @file:	file to test
+ *
+ * Return: True if @file is privileged in the domain of @ep.
+ */
+bool kdbus_ep_is_privileged(struct kdbus_ep *ep, struct file *file)
+{
+	return !ep->user &&
+		file_ns_capable(file, ep->bus->domain->user_namespace,
+				CAP_IPC_OWNER);
+}
+
+/**
+ * kdbus_ep_is_owner() - check whether a file should be treated as bus owner
+ * @ep:		endpoint to operate on
+ * @file:	file to test
+ *
+ * Return: True if @file should be treated as bus owner on @ep
+ */
+bool kdbus_ep_is_owner(struct kdbus_ep *ep, struct file *file)
+{
+	return !ep->user &&
+		(uid_eq(file->f_cred->euid, ep->bus->node.uid) ||
+		 kdbus_ep_is_privileged(ep, file));
+}
+
+/**
  * kdbus_cmd_ep_make() - handle KDBUS_CMD_ENDPOINT_MAKE
  * @bus:		bus to operate on
  * @argp:		command payload
  *
- * Return: Newly created endpoint on success, ERR_PTR on failure.
+ * Return: NULL or newly created endpoint on success, ERR_PTR on failure.
  */
 struct kdbus_ep *kdbus_cmd_ep_make(struct kdbus_bus *bus, void __user *argp)
 {
@@ -247,7 +275,7 @@ exit:
  * @ep:			endpoint to operate on
  * @argp:		command payload
  *
- * Return: Newly created endpoint on success, ERR_PTR on failure.
+ * Return: >=0 on success, negative error code on failure.
  */
 int kdbus_cmd_ep_update(struct kdbus_ep *ep, void __user *argp)
 {
diff --git a/ipc/kdbus/endpoint.h b/ipc/kdbus/endpoint.h
index d31954bfba2c..e0da59f01a7a 100644
--- a/ipc/kdbus/endpoint.h
+++ b/ipc/kdbus/endpoint.h
@@ -25,7 +25,7 @@ struct kdbus_bus;
 struct kdbus_user;
 
 /**
- * struct kdbus_ep - enpoint to access a bus
+ * struct kdbus_ep - endpoint to access a bus
  * @node:		The kdbus node
  * @lock:		Endpoint data lock
  * @bus:		Bus behind this endpoint
@@ -33,7 +33,7 @@ struct kdbus_user;
  * @policy_db:		Uploaded policy
  * @conn_list:		Connections of this endpoint
  *
- * An enpoint offers access to a bus; the default endpoint node name is "bus".
+ * An endpoint offers access to a bus; the default endpoint node name is "bus".
  * Additional custom endpoints to the same bus can be created and they can
  * carry their own policies/filters.
  */
@@ -61,6 +61,9 @@ struct kdbus_ep *kdbus_ep_new(struct kdbus_bus *bus, const char *name,
 struct kdbus_ep *kdbus_ep_ref(struct kdbus_ep *ep);
 struct kdbus_ep *kdbus_ep_unref(struct kdbus_ep *ep);
 
+bool kdbus_ep_is_privileged(struct kdbus_ep *ep, struct file *file);
+bool kdbus_ep_is_owner(struct kdbus_ep *ep, struct file *file);
+
 struct kdbus_ep *kdbus_cmd_ep_make(struct kdbus_bus *bus, void __user *argp);
 int kdbus_cmd_ep_update(struct kdbus_ep *ep, void __user *argp);
 
diff --git a/ipc/kdbus/fs.c b/ipc/kdbus/fs.c
index d01f33baaa0d..09c480924b9e 100644
--- a/ipc/kdbus/fs.c
+++ b/ipc/kdbus/fs.c
@@ -74,7 +74,7 @@ static int fs_dir_fop_iterate(struct file *file, struct dir_context *ctx)
 	 * closest node to that position and cannot use our node pointer. This
 	 * means iterating the rb-tree to find the closest match and start over
 	 * from there.
-	 * Note that hash values are not neccessarily unique. Therefore, llseek
+	 * Note that hash values are not necessarily unique. Therefore, llseek
 	 * is not guaranteed to seek to the same node that you got when you
 	 * retrieved the position. Seeking to 0, 1, 2 and >=INT_MAX is safe,
 	 * though. We could use the inode-number as position, but this would
@@ -325,9 +325,7 @@ static void fs_super_kill(struct super_block *sb)
 	}
 
 	kill_anon_super(sb);
-
-	if (domain)
-		kdbus_domain_unref(domain);
+	kdbus_domain_unref(domain);
 }
 
 static int fs_super_set(struct super_block *sb, void *data)
diff --git a/ipc/kdbus/handle.c b/ipc/kdbus/handle.c
index f72dbe513b4a..fc60932d69c7 100644
--- a/ipc/kdbus/handle.c
+++ b/ipc/kdbus/handle.c
@@ -18,6 +18,7 @@
 #include <linux/init.h>
 #include <linux/kdev_t.h>
 #include <linux/module.h>
+#include <linux/mutex.h>
 #include <linux/poll.h>
 #include <linux/rwsem.h>
 #include <linux/sched.h>
@@ -71,10 +72,6 @@ static int kdbus_args_verify(struct kdbus_args *args)
 	if (!KDBUS_ITEMS_END(item, args->items, args->items_size))
 		return -EINVAL;
 
-	for (i = 0; i < args->argc; ++i)
-		if (args->argv[i].mandatory && !args->argv[i].item)
-			return -EINVAL;
-
 	return 0;
 }
 
@@ -129,6 +126,7 @@ static int kdbus_args_negotiate(struct kdbus_args *args)
 /**
  * __kdbus_args_parse() - parse payload of kdbus command
  * @args:		object to parse data into
+ * @is_cmd:		whether this is a command or msg payload
  * @argp:		user-space location of command payload to parse
  * @type_size:		overall size of command payload to parse
  * @items_offset:	offset of items array in command payload
@@ -143,22 +141,49 @@ static int kdbus_args_negotiate(struct kdbus_args *args)
  * If this function succeeded, you must call kdbus_args_clear() to release
  * allocated resources before destroying @args.
  *
+ * This can also be used to import kdbus_msg objects. In that case, @is_cmd must
+ * be set to 'false' and the 'return_flags' field will not be touched (as it
+ * doesn't exist on kdbus_msg).
+ *
  * Return: On failure a negative error code is returned. Otherwise, 1 is
  * returned if negotiation was requested, 0 if not.
  */
-int __kdbus_args_parse(struct kdbus_args *args, void __user *argp,
+int __kdbus_args_parse(struct kdbus_args *args, bool is_cmd, void __user *argp,
 		       size_t type_size, size_t items_offset, void **out)
 {
-	int ret;
+	u64 user_size;
+	int ret, i;
+
+	ret = kdbus_copy_from_user(&user_size, argp, sizeof(user_size));
+	if (ret < 0)
+		return ret;
 
-	args->cmd = kdbus_memdup_user(argp, type_size, KDBUS_CMD_MAX_SIZE);
-	if (IS_ERR(args->cmd))
-		return PTR_ERR(args->cmd);
+	if (user_size < type_size)
+		return -EINVAL;
+	if (user_size > KDBUS_CMD_MAX_SIZE)
+		return -EMSGSIZE;
+
+	if (user_size <= sizeof(args->cmd_buf)) {
+		if (copy_from_user(args->cmd_buf, argp, user_size))
+			return -EFAULT;
+		args->cmd = (void*)args->cmd_buf;
+	} else {
+		args->cmd = memdup_user(argp, user_size);
+		if (IS_ERR(args->cmd))
+			return PTR_ERR(args->cmd);
+	}
+
+	if (args->cmd->size != user_size) {
+		ret = -EINVAL;
+		goto error;
+	}
 
-	args->cmd->return_flags = 0;
+	if (is_cmd)
+		args->cmd->return_flags = 0;
 	args->user = argp;
 	args->items = (void *)((u8 *)args->cmd + items_offset);
 	args->items_size = args->cmd->size - items_offset;
+	args->is_cmd = is_cmd;
 
 	if (args->cmd->flags & ~args->allowed_flags) {
 		ret = -EINVAL;
@@ -173,6 +198,15 @@ int __kdbus_args_parse(struct kdbus_args *args, void __user *argp,
 	if (ret < 0)
 		goto error;
 
+	/* mandatory items must be given (but not on negotiation) */
+	if (!(args->cmd->flags & KDBUS_FLAG_NEGOTIATE)) {
+		for (i = 0; i < args->argc; ++i)
+			if (args->argv[i].mandatory && !args->argv[i].item) {
+				ret = -EINVAL;
+				goto error;
+			}
+	}
+
 	*out = args->cmd;
 	return !!(args->cmd->flags & KDBUS_FLAG_NEGOTIATE);
 
@@ -198,10 +232,11 @@ int kdbus_args_clear(struct kdbus_args *args, int ret)
 		return ret;
 
 	if (!IS_ERR_OR_NULL(args->cmd)) {
-		if (put_user(args->cmd->return_flags,
-			     &args->user->return_flags))
+		if (args->is_cmd && put_user(args->cmd->return_flags,
+					     &args->user->return_flags))
 			ret = -EFAULT;
-		kfree(args->cmd);
+		if (args->cmd != (void*)args->cmd_buf)
+			kfree(args->cmd);
 		args->cmd = NULL;
 	}
 
@@ -224,15 +259,14 @@ enum kdbus_handle_type {
 
 /**
  * struct kdbus_handle - handle to the kdbus system
- * @rwlock:		handle lock
+ * @lock:		handle lock
  * @type:		type of this handle (KDBUS_HANDLE_*)
  * @bus_owner:		bus this handle owns
  * @ep_owner:		endpoint this handle owns
  * @conn:		connection this handle owns
- * @privileged:		Flag to mark a handle as privileged
  */
 struct kdbus_handle {
-	struct rw_semaphore rwlock;
+	struct mutex lock;
 
 	enum kdbus_handle_type type;
 	union {
@@ -240,8 +274,6 @@ struct kdbus_handle {
 		struct kdbus_ep *ep_owner;
 		struct kdbus_conn *conn;
 	};
-
-	bool privileged:1;
 };
 
 static int kdbus_handle_open(struct inode *inode, struct file *file)
@@ -260,26 +292,9 @@ static int kdbus_handle_open(struct inode *inode, struct file *file)
 		goto exit;
 	}
 
-	init_rwsem(&handle->rwlock);
+	mutex_init(&handle->lock);
 	handle->type = KDBUS_HANDLE_NONE;
 
-	if (node->type == KDBUS_NODE_ENDPOINT) {
-		struct kdbus_ep *ep = kdbus_ep_from_node(node);
-		struct kdbus_bus *bus = ep->bus;
-
-		/*
-		 * A connection is privileged if it is opened on an endpoint
-		 * without custom policy and either:
-		 *   * the user has CAP_IPC_OWNER in the domain user namespace
-		 * or
-		 *   * the callers euid matches the uid of the bus creator
-		 */
-		if (!ep->user &&
-		    (ns_capable(bus->domain->user_namespace, CAP_IPC_OWNER) ||
-		     uid_eq(file->f_cred->euid, bus->node.uid)))
-			handle->privileged = true;
-	}
-
 	file->private_data = handle;
 	ret = 0;
 
@@ -350,8 +365,8 @@ static long kdbus_handle_ioctl_control(struct file *file, unsigned int cmd,
 			break;
 		}
 
-		handle->type = KDBUS_HANDLE_BUS_OWNER;
 		handle->bus_owner = bus;
+		ret = KDBUS_HANDLE_BUS_OWNER;
 		break;
 	}
 
@@ -371,6 +386,7 @@ static long kdbus_handle_ioctl_ep(struct file *file, unsigned int cmd,
 	struct kdbus_handle *handle = file->private_data;
 	struct kdbus_node *node = file_inode(file)->i_private;
 	struct kdbus_ep *ep, *file_ep = kdbus_ep_from_node(node);
+	struct kdbus_bus *bus = file_ep->bus;
 	struct kdbus_conn *conn;
 	int ret = 0;
 
@@ -378,32 +394,33 @@ static long kdbus_handle_ioctl_ep(struct file *file, unsigned int cmd,
 		return -ESHUTDOWN;
 
 	switch (cmd) {
-	case KDBUS_CMD_ENDPOINT_MAKE:
+	case KDBUS_CMD_ENDPOINT_MAKE: {
 		/* creating custom endpoints is a privileged operation */
-		if (!handle->privileged) {
+		if (!kdbus_ep_is_owner(file_ep, file)) {
 			ret = -EPERM;
 			break;
 		}
 
-		ep = kdbus_cmd_ep_make(file_ep->bus, buf);
+		ep = kdbus_cmd_ep_make(bus, buf);
 		if (IS_ERR_OR_NULL(ep)) {
 			ret = PTR_ERR_OR_ZERO(ep);
 			break;
 		}
 
-		handle->type = KDBUS_HANDLE_EP_OWNER;
 		handle->ep_owner = ep;
+		ret = KDBUS_HANDLE_EP_OWNER;
 		break;
+	}
 
 	case KDBUS_CMD_HELLO:
-		conn = kdbus_cmd_hello(file_ep, handle->privileged, buf);
+		conn = kdbus_cmd_hello(file_ep, file, buf);
 		if (IS_ERR_OR_NULL(conn)) {
 			ret = PTR_ERR_OR_ZERO(conn);
 			break;
 		}
 
-		handle->type = KDBUS_HANDLE_CONNECTED;
 		handle->conn = conn;
+		ret = KDBUS_HANDLE_CONNECTED;
 		break;
 
 	default:
@@ -517,19 +534,41 @@ static long kdbus_handle_ioctl(struct file *file, unsigned int cmd,
 	case KDBUS_CMD_BUS_MAKE:
 	case KDBUS_CMD_ENDPOINT_MAKE:
 	case KDBUS_CMD_HELLO:
-		/* bail out early if already typed */
-		if (handle->type != KDBUS_HANDLE_NONE)
-			break;
-
-		down_write(&handle->rwlock);
+		mutex_lock(&handle->lock);
 		if (handle->type == KDBUS_HANDLE_NONE) {
 			if (node->type == KDBUS_NODE_CONTROL)
 				ret = kdbus_handle_ioctl_control(file, cmd,
 								 argp);
 			else if (node->type == KDBUS_NODE_ENDPOINT)
 				ret = kdbus_handle_ioctl_ep(file, cmd, argp);
+
+			if (ret > 0) {
+				/*
+				 * The data given via open() is not sufficient
+				 * to setup a kdbus handle. Hence, we require
+				 * the user to perform a setup ioctl. This setup
+				 * can only be performed once and defines the
+				 * type of the handle. The different setup
+				 * ioctls are locked against each other so they
+				 * cannot race. Once the handle type is set,
+				 * the type-dependent ioctls are enabled. To
+				 * improve performance, we don't lock those via
+				 * handle->lock. Instead, we issue a
+				 * write-barrier before performing the
+				 * type-change, which pairs with smp_rmb() in
+				 * all handlers that access the type field. This
+				 * guarantees the handle is fully setup, if
+				 * handle->type is set. If handle->type is
+				 * unset, you must not make any assumptions
+				 * without taking handle->lock.
+				 * Note that handle->type is only set once. It
+				 * will never change afterwards.
+				 */
+				smp_wmb();
+				handle->type = ret;
+			}
 		}
-		up_write(&handle->rwlock);
+		mutex_unlock(&handle->lock);
 		break;
 
 	case KDBUS_CMD_ENDPOINT_UPDATE:
@@ -544,14 +583,30 @@ static long kdbus_handle_ioctl(struct file *file, unsigned int cmd,
 	case KDBUS_CMD_MATCH_REMOVE:
 	case KDBUS_CMD_SEND:
 	case KDBUS_CMD_RECV:
-	case KDBUS_CMD_FREE:
-		down_read(&handle->rwlock);
-		if (handle->type == KDBUS_HANDLE_EP_OWNER)
+	case KDBUS_CMD_FREE: {
+		enum kdbus_handle_type type;
+
+		/*
+		 * This read-barrier pairs with smp_wmb() of the handle setup.
+		 * it guarantees the handle is fully written, in case the
+		 * type has been set. It allows us to access the handle without
+		 * taking handle->lock, given the guarantee that the type is
+		 * only ever set once, and stays constant afterwards.
+		 * Furthermore, the handle object itself is not modified in any
+		 * way after the type is set. That is, the type-field is the
+		 * last field that is written on any handle. If it has not been
+		 * set, we must not access the handle here.
+		 */
+		type = handle->type;
+		smp_rmb();
+
+		if (type == KDBUS_HANDLE_EP_OWNER)
 			ret = kdbus_handle_ioctl_ep_owner(file, cmd, argp);
-		else if (handle->type == KDBUS_HANDLE_CONNECTED)
+		else if (type == KDBUS_HANDLE_CONNECTED)
 			ret = kdbus_handle_ioctl_connected(file, cmd, argp);
-		up_read(&handle->rwlock);
+
 		break;
+	}
 	default:
 		ret = -ENOTTY;
 		break;
@@ -564,42 +619,61 @@ static unsigned int kdbus_handle_poll(struct file *file,
 				      struct poll_table_struct *wait)
 {
 	struct kdbus_handle *handle = file->private_data;
+	enum kdbus_handle_type type;
 	unsigned int mask = POLLOUT | POLLWRNORM;
-	int ret;
 
-	/* Only a connected endpoint can read/write data */
-	down_read(&handle->rwlock);
-	if (handle->type != KDBUS_HANDLE_CONNECTED) {
-		up_read(&handle->rwlock);
-		return POLLERR | POLLHUP;
-	}
-	up_read(&handle->rwlock);
+	/*
+	 * This pairs with smp_wmb() during handle setup. It guarantees that
+	 * _iff_ the handle type is set, handle->conn is valid. Furthermore,
+	 * _iff_ the type is set, the handle object is constant and never
+	 * changed again. If it's not set, we must not access the handle but
+	 * bail out. We also must assume no setup has taken place, yet.
+	 */
+	type = handle->type;
+	smp_rmb();
 
-	ret = kdbus_conn_acquire(handle->conn);
-	if (ret < 0)
+	/* Only a connected endpoint can read/write data */
+	if (type != KDBUS_HANDLE_CONNECTED)
 		return POLLERR | POLLHUP;
 
 	poll_wait(file, &handle->conn->wait, wait);
 
+	/*
+	 * Verify the connection hasn't been deactivated _after_ adding the
+	 * wait-queue. This guarantees, that if the connection is deactivated
+	 * after we checked it, the waitqueue is signaled and we're called
+	 * again.
+	 */
+	if (!kdbus_conn_active(handle->conn))
+		return POLLERR | POLLHUP;
+
 	if (!list_empty(&handle->conn->queue.msg_list) ||
 	    atomic_read(&handle->conn->lost_count) > 0)
 		mask |= POLLIN | POLLRDNORM;
 
-	kdbus_conn_release(handle->conn);
-
 	return mask;
 }
 
 static int kdbus_handle_mmap(struct file *file, struct vm_area_struct *vma)
 {
 	struct kdbus_handle *handle = file->private_data;
+	enum kdbus_handle_type type;
 	int ret = -EBADFD;
 
-	if (down_read_trylock(&handle->rwlock)) {
-		if (handle->type == KDBUS_HANDLE_CONNECTED)
-			ret = kdbus_pool_mmap(handle->conn->pool, vma);
-		up_read(&handle->rwlock);
-	}
+	/*
+	 * This pairs with smp_wmb() during handle setup. It guarantees that
+	 * _iff_ the handle type is set, handle->conn is valid. Furthermore,
+	 * _iff_ the type is set, the handle object is constant and never
+	 * changed again. If it's not set, we must not access the handle but
+	 * bail out. We also must assume no setup has taken place, yet.
+	 */
+	type = handle->type;
+	smp_rmb();
+
+	/* Only connected handles have a pool we can map */
+	if (type == KDBUS_HANDLE_CONNECTED)
+		ret = kdbus_pool_mmap(handle->conn->pool, vma);
+
 	return ret;
 }
 
diff --git a/ipc/kdbus/handle.h b/ipc/kdbus/handle.h
index 93a372d554a2..5dde2c10bed4 100644
--- a/ipc/kdbus/handle.h
+++ b/ipc/kdbus/handle.h
@@ -45,13 +45,18 @@ struct kdbus_arg {
  * @argv:		array of items this command supports
  * @user:		set by parser to user-space location of current command
  * @cmd:		set by parser to kernel copy of command payload
+ * @cmd_buf:		inline buf to avoid kmalloc() on small cmds
  * @items:		points to item array in @cmd
  * @items_size:		size of @items in bytes
+ * @is_cmd:		whether this is a command-payload or msg-payload
  *
  * This structure is used to parse ioctl command payloads on each invocation.
  * The ioctl handler has to pre-fill the flags and allowed items before passing
  * the object to kdbus_args_parse(). The parser will copy the command payload
  * into kernel-space and verify the correctness of the data.
+ *
+ * We use a 256 bytes buffer for small command payloads, to be allocated on
+ * stack on syscall entrance.
  */
 struct kdbus_args {
 	u64 allowed_flags;
@@ -60,12 +65,14 @@ struct kdbus_args {
 
 	struct kdbus_cmd __user *user;
 	struct kdbus_cmd *cmd;
+	u8 cmd_buf[256];
 
 	struct kdbus_item *items;
 	size_t items_size;
+	bool is_cmd : 1;
 };
 
-int __kdbus_args_parse(struct kdbus_args *args, void __user *argp,
+int __kdbus_args_parse(struct kdbus_args *args, bool is_cmd, void __user *argp,
 		       size_t type_size, size_t items_offset, void **out);
 int kdbus_args_clear(struct kdbus_args *args, int ret);
 
@@ -77,7 +84,18 @@ int kdbus_args_clear(struct kdbus_args *args, int ret);
 			     offsetof(struct kdbus_cmd, flags));        \
 		BUILD_BUG_ON(offsetof(typeof(**(_v)), return_flags) !=  \
 			     offsetof(struct kdbus_cmd, return_flags)); \
-		__kdbus_args_parse((_args), (_argp), sizeof(**(_v)),    \
+		__kdbus_args_parse((_args), 1, (_argp), sizeof(**(_v)), \
+				   offsetof(typeof(**(_v)), items),     \
+				   (void **)(_v));                      \
+	})
+
+#define kdbus_args_parse_msg(_args, _argp, _v)                          \
+	({                                                              \
+		BUILD_BUG_ON(offsetof(typeof(**(_v)), size) !=          \
+			     offsetof(struct kdbus_cmd, size));         \
+		BUILD_BUG_ON(offsetof(typeof(**(_v)), flags) !=         \
+			     offsetof(struct kdbus_cmd, flags));        \
+		__kdbus_args_parse((_args), 0, (_argp), sizeof(**(_v)), \
 				   offsetof(typeof(**(_v)), items),     \
 				   (void **)(_v));                      \
 	})
diff --git a/ipc/kdbus/item.c b/ipc/kdbus/item.c
index 745ad5495096..ce78dba03426 100644
--- a/ipc/kdbus/item.c
+++ b/ipc/kdbus/item.c
@@ -96,12 +96,6 @@ int kdbus_item_validate(const struct kdbus_item *item)
 		break;
 
 	case KDBUS_ITEM_PAYLOAD_VEC:
-		if (payload_size != sizeof(struct kdbus_vec))
-			return -EINVAL;
-		if (item->vec.size == 0 || item->vec.size > SIZE_MAX)
-			return -EINVAL;
-		break;
-
 	case KDBUS_ITEM_PAYLOAD_OFF:
 		if (payload_size != sizeof(struct kdbus_vec))
 			return -EINVAL;
@@ -153,6 +147,7 @@ int kdbus_item_validate(const struct kdbus_item *item)
 	case KDBUS_ITEM_ATTACH_FLAGS_SEND:
 	case KDBUS_ITEM_ATTACH_FLAGS_RECV:
 	case KDBUS_ITEM_ID:
+	case KDBUS_ITEM_DST_ID:
 		if (payload_size != sizeof(u64))
 			return -EINVAL;
 		break;
@@ -266,47 +261,6 @@ int kdbus_items_validate(const struct kdbus_item *items, size_t items_size)
 	return 0;
 }
 
-static struct kdbus_item *kdbus_items_get(const struct kdbus_item *items,
-					  size_t items_size,
-					  unsigned int item_type)
-{
-	const struct kdbus_item *iter, *found = NULL;
-
-	KDBUS_ITEMS_FOREACH(iter, items, items_size) {
-		if (iter->type == item_type) {
-			if (found)
-				return ERR_PTR(-EEXIST);
-			found = iter;
-		}
-	}
-
-	return (struct kdbus_item *)found ? : ERR_PTR(-EBADMSG);
-}
-
-/**
- * kdbus_items_get_str() - get string from a list of items
- * @items:		The items to walk
- * @items_size:		The size of all items
- * @item_type:		The item type to look for
- *
- * This function walks a list of items and searches for items of type
- * @item_type. If it finds exactly one such item, @str_ret will be set to
- * the .str member of the item.
- *
- * Return: the string, if the item was found exactly once, ERR_PTR(-EEXIST)
- * if the item was found more than once, and ERR_PTR(-EBADMSG) if there was
- * no item of the given type.
- */
-const char *kdbus_items_get_str(const struct kdbus_item *items,
-				size_t items_size,
-				unsigned int item_type)
-{
-	const struct kdbus_item *item;
-
-	item = kdbus_items_get(items, items_size, item_type);
-	return IS_ERR(item) ? ERR_CAST(item) : item->str;
-}
-
 /**
  * kdbus_item_set() - Set item content
  * @item:	The item to modify
diff --git a/ipc/kdbus/item.h b/ipc/kdbus/item.h
index eeefd8beac3b..3a7e6ccc253c 100644
--- a/ipc/kdbus/item.h
+++ b/ipc/kdbus/item.h
@@ -21,17 +21,17 @@
 #include "util.h"
 
 /* generic access and iterators over a stream of items */
-#define KDBUS_ITEM_NEXT(_i) (typeof(_i))(((u8 *)_i) + KDBUS_ALIGN8((_i)->size))
-#define KDBUS_ITEMS_SIZE(_h, _is) ((_h)->size - offsetof(typeof(*_h), _is))
+#define KDBUS_ITEM_NEXT(_i) (typeof(_i))((u8 *)(_i) + KDBUS_ALIGN8((_i)->size))
+#define KDBUS_ITEMS_SIZE(_h, _is) ((_h)->size - offsetof(typeof(*(_h)), _is))
 #define KDBUS_ITEM_HEADER_SIZE offsetof(struct kdbus_item, data)
 #define KDBUS_ITEM_SIZE(_s) KDBUS_ALIGN8(KDBUS_ITEM_HEADER_SIZE + (_s))
 #define KDBUS_ITEM_PAYLOAD_SIZE(_i) ((_i)->size - KDBUS_ITEM_HEADER_SIZE)
 
 #define KDBUS_ITEMS_FOREACH(_i, _is, _s)				\
-	for (_i = _is;							\
+	for ((_i) = (_is);						\
 	     ((u8 *)(_i) < (u8 *)(_is) + (_s)) &&			\
 	       ((u8 *)(_i) >= (u8 *)(_is));				\
-	     _i = KDBUS_ITEM_NEXT(_i))
+	     (_i) = KDBUS_ITEM_NEXT(_i))
 
 #define KDBUS_ITEM_VALID(_i, _is, _s)					\
 	((_i)->size >= KDBUS_ITEM_HEADER_SIZE &&			\
@@ -40,7 +40,7 @@
 	 (u8 *)(_i) >= (u8 *)(_is))
 
 #define KDBUS_ITEMS_END(_i, _is, _s)					\
-	((u8 *)_i == ((u8 *)(_is) + KDBUS_ALIGN8(_s)))
+	((u8 *)(_i) == ((u8 *)(_is) + KDBUS_ALIGN8(_s)))
 
 /**
  * struct kdbus_item_header - Describes the fix part of an item
@@ -55,9 +55,6 @@ struct kdbus_item_header {
 int kdbus_item_validate_name(const struct kdbus_item *item);
 int kdbus_item_validate(const struct kdbus_item *item);
 int kdbus_items_validate(const struct kdbus_item *items, size_t items_size);
-const char *kdbus_items_get_str(const struct kdbus_item *items,
-				size_t items_size,
-				unsigned int item_type);
 struct kdbus_item *kdbus_item_set(struct kdbus_item *item, u64 type,
 				  const void *data, size_t len);
 
diff --git a/ipc/kdbus/limits.h b/ipc/kdbus/limits.h
index 3ea2f0e42bd9..59e3608ee8c3 100644
--- a/ipc/kdbus/limits.h
+++ b/ipc/kdbus/limits.h
@@ -19,9 +19,6 @@
 /* maximum size of message header and items */
 #define KDBUS_MSG_MAX_SIZE		SZ_8K
 
-/* maximum number of message items */
-#define KDBUS_MSG_MAX_ITEMS		128
-
 /* maximum number of memfd items per message */
 #define KDBUS_MSG_MAX_MEMFD_ITEMS	16
 
@@ -53,7 +50,7 @@
 #define KDBUS_CONN_MAX_NAMES			256
 
 /* maximum number of queued requests waiting for a reply */
-#define KDBUS_CONN_MAX_REQUESTS_PENDING		1024	/* 128 -> 1024 */
+#define KDBUS_CONN_MAX_REQUESTS_PENDING		1024
 
 /* maximum number of connections per user in one domain */
 #define KDBUS_USER_MAX_CONN			1024
diff --git a/ipc/kdbus/main.c b/ipc/kdbus/main.c
index 785f529d98b7..1ad4dc8dafa1 100644
--- a/ipc/kdbus/main.c
+++ b/ipc/kdbus/main.c
@@ -15,7 +15,6 @@
 #include <linux/fs.h>
 #include <linux/init.h>
 #include <linux/module.h>
-#include <linux/moduleparam.h>
 
 #include "util.h"
 #include "fs.h"
@@ -79,17 +78,6 @@
 /* kdbus mount-point /sys/fs/kdbus */
 static struct kobject *kdbus_dir;
 
-/* global module option to apply a mask to exported metadata */
-unsigned long long kdbus_meta_attach_mask = KDBUS_ATTACH_TIMESTAMP |
-					    KDBUS_ATTACH_CREDS |
-					    KDBUS_ATTACH_PIDS |
-					    KDBUS_ATTACH_AUXGROUPS |
-					    KDBUS_ATTACH_NAMES |
-					    KDBUS_ATTACH_SECLABEL |
-					    KDBUS_ATTACH_CONN_DESCRIPTION;
-MODULE_PARM_DESC(attach_flags_mask, "Attach-flags mask for exported metadata");
-module_param_named(attach_flags_mask, kdbus_meta_attach_mask, ullong, 0644);
-
 static int __init kdbus_init(void)
 {
 	int ret;
@@ -116,6 +104,7 @@ static void __exit kdbus_exit(void)
 {
 	kdbus_fs_exit();
 	kobject_put(kdbus_dir);
+	ida_destroy(&kdbus_node_ida);
 }
 
 module_init(kdbus_init);
diff --git a/ipc/kdbus/match.c b/ipc/kdbus/match.c
index 30cec1ca819f..4ee6a1f2e1bc 100644
--- a/ipc/kdbus/match.c
+++ b/ipc/kdbus/match.c
@@ -66,7 +66,7 @@ struct kdbus_bloom_mask {
 
 /**
  * struct kdbus_match_rule - a rule appended to a match entry
- * @type:		An item type to match agains
+ * @type:		An item type to match against
  * @bloom_mask:		Bloom mask to match a message's filter against, used
  *			with KDBUS_ITEM_BLOOM_MASK
  * @name:		Name to match against, used with KDBUS_ITEM_NAME,
@@ -78,6 +78,7 @@ struct kdbus_bloom_mask {
  *			KDBUS_ITEM_NAME_{ADD,REMOVE,CHANGE},
  *			KDBUS_ITEM_ID_REMOVE
  * @src_id:		ID to match against, used with KDBUS_ITEM_ID
+ * @dst_id:		Message destination ID, used with KDBUS_ITEM_DST_ID
  * @rules_entry:	Entry in the entry's rules list
  */
 struct kdbus_match_rule {
@@ -90,6 +91,7 @@ struct kdbus_match_rule {
 			u64 new_id;
 		};
 		u64 src_id;
+		u64 dst_id;
 	};
 	struct list_head rules_entry;
 };
@@ -112,6 +114,7 @@ static void kdbus_match_rule_free(struct kdbus_match_rule *rule)
 		break;
 
 	case KDBUS_ITEM_ID:
+	case KDBUS_ITEM_DST_ID:
 	case KDBUS_ITEM_ID_ADD:
 	case KDBUS_ITEM_ID_REMOVE:
 		break;
@@ -204,96 +207,74 @@ static bool kdbus_match_bloom(const struct kdbus_bloom_filter *filter,
 	return true;
 }
 
-static bool kdbus_match_rules(const struct kdbus_match_entry *entry,
-			      struct kdbus_conn *conn_src,
-			      struct kdbus_kmsg *kmsg)
+static bool kdbus_match_rule_conn(const struct kdbus_match_rule *r,
+				  struct kdbus_conn *c,
+				  const struct kdbus_staging *s)
 {
-	struct kdbus_match_rule *r;
-
-	if (conn_src)
-		lockdep_assert_held(&conn_src->ep->bus->name_registry->rwlock);
-
-	/*
-	 * Walk all the rules and bail out immediately
-	 * if any of them is unsatisfied.
-	 */
-
-	list_for_each_entry(r, &entry->rules_list, rules_entry) {
-		if (conn_src) {
-			/* messages from userspace */
-
-			switch (r->type) {
-			case KDBUS_ITEM_BLOOM_MASK:
-				if (!kdbus_match_bloom(kmsg->bloom_filter,
-						       &r->bloom_mask,
-						       conn_src))
-					return false;
-				break;
-
-			case KDBUS_ITEM_ID:
-				if (r->src_id != conn_src->id &&
-				    r->src_id != KDBUS_MATCH_ID_ANY)
-					return false;
-
-				break;
-
-			case KDBUS_ITEM_NAME:
-				if (!kdbus_conn_has_name(conn_src, r->name))
-					return false;
-
-				break;
-
-			default:
-				return false;
-			}
-		} else {
-			/* kernel notifications */
+	lockdep_assert_held(&c->ep->bus->name_registry->rwlock);
 
-			if (kmsg->notify_type != r->type)
-				return false;
-
-			switch (r->type) {
-			case KDBUS_ITEM_ID_ADD:
-				if (r->new_id != KDBUS_MATCH_ID_ANY &&
-				    r->new_id != kmsg->notify_new_id)
-					return false;
-
-				break;
+	switch (r->type) {
+	case KDBUS_ITEM_BLOOM_MASK:
+		return kdbus_match_bloom(s->bloom_filter, &r->bloom_mask, c);
+	case KDBUS_ITEM_ID:
+		return r->src_id == c->id || r->src_id == KDBUS_MATCH_ID_ANY;
+	case KDBUS_ITEM_DST_ID:
+		return r->dst_id == s->msg->dst_id ||
+		       r->dst_id == KDBUS_MATCH_ID_ANY;
+	case KDBUS_ITEM_NAME:
+		return kdbus_conn_has_name(c, r->name);
+	default:
+		return false;
+	}
+}
 
-			case KDBUS_ITEM_ID_REMOVE:
-				if (r->old_id != KDBUS_MATCH_ID_ANY &&
-				    r->old_id != kmsg->notify_old_id)
-					return false;
+static bool kdbus_match_rule_kernel(const struct kdbus_match_rule *r,
+				    const struct kdbus_staging *s)
+{
+	struct kdbus_item *n = s->notify;
 
-				break;
+	if (WARN_ON(!n) || n->type != r->type)
+		return false;
 
-			case KDBUS_ITEM_NAME_ADD:
-			case KDBUS_ITEM_NAME_CHANGE:
-			case KDBUS_ITEM_NAME_REMOVE:
-				if ((r->old_id != KDBUS_MATCH_ID_ANY &&
-				     r->old_id != kmsg->notify_old_id) ||
-				    (r->new_id != KDBUS_MATCH_ID_ANY &&
-				     r->new_id != kmsg->notify_new_id) ||
-				    (r->name && kmsg->notify_name &&
-				     strcmp(r->name, kmsg->notify_name) != 0))
-					return false;
+	switch (r->type) {
+	case KDBUS_ITEM_ID_ADD:
+		return r->new_id == KDBUS_MATCH_ID_ANY ||
+		       r->new_id == n->id_change.id;
+	case KDBUS_ITEM_ID_REMOVE:
+		return r->old_id == KDBUS_MATCH_ID_ANY ||
+		       r->old_id == n->id_change.id;
+	case KDBUS_ITEM_NAME_ADD:
+	case KDBUS_ITEM_NAME_CHANGE:
+	case KDBUS_ITEM_NAME_REMOVE:
+		return (r->old_id == KDBUS_MATCH_ID_ANY ||
+		        r->old_id == n->name_change.old_id.id) &&
+		       (r->new_id == KDBUS_MATCH_ID_ANY ||
+		        r->new_id == n->name_change.new_id.id) &&
+		       (!r->name || !strcmp(r->name, n->name_change.name));
+	default:
+		return false;
+	}
+}
 
-				break;
+static bool kdbus_match_rules(const struct kdbus_match_entry *entry,
+			      struct kdbus_conn *c,
+			      const struct kdbus_staging *s)
+{
+	struct kdbus_match_rule *r;
 
-			default:
-				return false;
-			}
-		}
-	}
+	list_for_each_entry(r, &entry->rules_list, rules_entry)
+		if ((c && !kdbus_match_rule_conn(r, c, s)) ||
+		    (!c && !kdbus_match_rule_kernel(r, s)))
+			return false;
 
 	return true;
 }
 
 /**
- * kdbus_match_db_match_kmsg() - match a kmsg object agains the database entries
+ * kdbus_match_db_match_msg() - match a msg object agains the database entries
  * @mdb:		The match database
  * @conn_src:		The connection object originating the message
- * @kmsg:		The kmsg to perform the match on
+ * @staging:		Staging object containing the message to match against
  *
  * This function will walk through all the database entries previously uploaded
  * with kdbus_match_db_add(). As soon as any of them has an all-satisfied rule
@@ -304,16 +285,16 @@ static bool kdbus_match_rules(const struct kdbus_match_entry *entry,
  *
  * Return: true if there was a matching database entry, false otherwise.
  */
-bool kdbus_match_db_match_kmsg(struct kdbus_match_db *mdb,
-			       struct kdbus_conn *conn_src,
-			       struct kdbus_kmsg *kmsg)
+bool kdbus_match_db_match_msg(struct kdbus_match_db *mdb,
+			      struct kdbus_conn *conn_src,
+			      const struct kdbus_staging *staging)
 {
 	struct kdbus_match_entry *entry;
 	bool matched = false;
 
 	down_read(&mdb->mdb_rwlock);
 	list_for_each_entry(entry, &mdb->entries_list, list_entry) {
-		matched = kdbus_match_rules(entry, conn_src, kmsg);
+		matched = kdbus_match_rules(entry, conn_src, staging);
 		if (matched)
 			break;
 	}
@@ -353,6 +334,7 @@ static int kdbus_match_db_remove_unlocked(struct kdbus_match_db *mdb,
  * KDBUS_ITEM_BLOOM_MASK:	A bloom mask
  * KDBUS_ITEM_NAME:		A connection's source name
  * KDBUS_ITEM_ID:		A connection ID
+ * KDBUS_ITEM_DST_ID:		A connection ID
  * KDBUS_ITEM_NAME_ADD:
  * KDBUS_ITEM_NAME_REMOVE:
  * KDBUS_ITEM_NAME_CHANGE:	Well-known name changes, carry
@@ -364,11 +346,11 @@ static int kdbus_match_db_remove_unlocked(struct kdbus_match_db *mdb,
  * For kdbus_notify_{id,name}_change structs, only the ID and name fields
  * are looked at when adding an entry. The flags are unused.
  *
- * Also note that KDBUS_ITEM_BLOOM_MASK, KDBUS_ITEM_NAME and KDBUS_ITEM_ID
- * are used to match messages from userspace, while the others apply to
- * kernel-generated notifications.
+ * Also note that KDBUS_ITEM_BLOOM_MASK, KDBUS_ITEM_NAME, KDBUS_ITEM_ID,
+ * and KDBUS_ITEM_DST_ID are used to match messages from userspace, while the
+ * others apply to kernel-generated notifications.
  *
- * Return: 0 on success, negative error code on failure.
+ * Return: >=0 on success, negative error code on failure.
  */
 int kdbus_cmd_match_add(struct kdbus_conn *conn, void __user *argp)
 {
@@ -383,6 +365,7 @@ int kdbus_cmd_match_add(struct kdbus_conn *conn, void __user *argp)
 		{ .type = KDBUS_ITEM_BLOOM_MASK, .multiple = true },
 		{ .type = KDBUS_ITEM_NAME, .multiple = true },
 		{ .type = KDBUS_ITEM_ID, .multiple = true },
+		{ .type = KDBUS_ITEM_DST_ID, .multiple = true },
 		{ .type = KDBUS_ITEM_NAME_ADD, .multiple = true },
 		{ .type = KDBUS_ITEM_NAME_REMOVE, .multiple = true },
 		{ .type = KDBUS_ITEM_NAME_CHANGE, .multiple = true },
@@ -465,6 +448,10 @@ int kdbus_cmd_match_add(struct kdbus_conn *conn, void __user *argp)
 			rule->src_id = item->id;
 			break;
 
+		case KDBUS_ITEM_DST_ID:
+			rule->dst_id = item->id;
+			break;
+
 		case KDBUS_ITEM_NAME_ADD:
 		case KDBUS_ITEM_NAME_REMOVE:
 		case KDBUS_ITEM_NAME_CHANGE:
@@ -528,7 +515,7 @@ exit:
  * @conn:		connection to operate on
  * @argp:		command payload
  *
- * Return: 0 on success, negative error code on failure.
+ * Return: >=0 on success, negative error code on failure.
  */
 int kdbus_cmd_match_remove(struct kdbus_conn *conn, void __user *argp)
 {
diff --git a/ipc/kdbus/match.h b/ipc/kdbus/match.h
index ea4292938deb..ceb492f8e51c 100644
--- a/ipc/kdbus/match.h
+++ b/ipc/kdbus/match.h
@@ -16,8 +16,8 @@
 #define __KDBUS_MATCH_H
 
 struct kdbus_conn;
-struct kdbus_kmsg;
 struct kdbus_match_db;
+struct kdbus_staging;
 
 struct kdbus_match_db *kdbus_match_db_new(void);
 void kdbus_match_db_free(struct kdbus_match_db *db);
@@ -25,9 +25,9 @@ int kdbus_match_db_add(struct kdbus_conn *conn,
 		       struct kdbus_cmd_match *cmd);
 int kdbus_match_db_remove(struct kdbus_conn *conn,
 			  struct kdbus_cmd_match *cmd);
-bool kdbus_match_db_match_kmsg(struct kdbus_match_db *db,
-			       struct kdbus_conn *conn_src,
-			       struct kdbus_kmsg *kmsg);
+bool kdbus_match_db_match_msg(struct kdbus_match_db *db,
+			      struct kdbus_conn *conn_src,
+			      const struct kdbus_staging *staging);
 
 int kdbus_cmd_match_add(struct kdbus_conn *conn, void __user *argp);
 int kdbus_cmd_match_remove(struct kdbus_conn *conn, void __user *argp);
diff --git a/ipc/kdbus/message.c b/ipc/kdbus/message.c
index 4b1487a88c49..ae565cd343f8 100644
--- a/ipc/kdbus/message.c
+++ b/ipc/kdbus/message.c
@@ -37,567 +37,1004 @@
 #include "names.h"
 #include "policy.h"
 
-#define KDBUS_KMSG_HEADER_SIZE offsetof(struct kdbus_kmsg, msg)
+static const char * const zeros = "\0\0\0\0\0\0\0";
 
-static struct kdbus_msg_resources *kdbus_msg_resources_new(void)
+static struct kdbus_gaps *kdbus_gaps_new(size_t n_memfds, size_t n_fds)
 {
-	struct kdbus_msg_resources *r;
+	size_t size_offsets, size_memfds, size_fds, size;
+	struct kdbus_gaps *gaps;
+
+	size_offsets = n_memfds * sizeof(*gaps->memfd_offsets);
+	size_memfds = n_memfds * sizeof(*gaps->memfd_files);
+	size_fds = n_fds * sizeof(*gaps->fd_files);
+	size = sizeof(*gaps) + size_offsets + size_memfds + size_fds;
 
-	r = kzalloc(sizeof(*r), GFP_KERNEL);
-	if (!r)
+	gaps = kzalloc(size, GFP_KERNEL);
+	if (!gaps)
 		return ERR_PTR(-ENOMEM);
 
-	kref_init(&r->kref);
+	kref_init(&gaps->kref);
+	gaps->n_memfds = 0; /* we reserve n_memfds, but don't enforce them */
+	gaps->memfd_offsets = (void *)(gaps + 1);
+	gaps->memfd_files = (void *)((u8 *)gaps->memfd_offsets + size_offsets);
+	gaps->n_fds = 0; /* we reserve n_fds, but don't enforce them */
+	gaps->fd_files = (void *)((u8 *)gaps->memfd_files + size_memfds);
 
-	return r;
+	return gaps;
 }
 
-static void __kdbus_msg_resources_free(struct kref *kref)
+static void kdbus_gaps_free(struct kref *kref)
 {
-	struct kdbus_msg_resources *r =
-		container_of(kref, struct kdbus_msg_resources, kref);
+	struct kdbus_gaps *gaps = container_of(kref, struct kdbus_gaps, kref);
 	size_t i;
 
-	for (i = 0; i < r->data_count; ++i) {
-		switch (r->data[i].type) {
-		case KDBUS_MSG_DATA_VEC:
-			/* nothing to do */
-			break;
-		case KDBUS_MSG_DATA_MEMFD:
-			if (r->data[i].memfd.file)
-				fput(r->data[i].memfd.file);
-			break;
-		}
-	}
-
-	for (i = 0; i < r->fds_count; i++)
-		if (r->fds[i])
-			fput(r->fds[i]);
+	for (i = 0; i < gaps->n_fds; ++i)
+		if (gaps->fd_files[i])
+			fput(gaps->fd_files[i]);
+	for (i = 0; i < gaps->n_memfds; ++i)
+		if (gaps->memfd_files[i])
+			fput(gaps->memfd_files[i]);
 
-	kfree(r->dst_name);
-	kfree(r->data);
-	kfree(r->fds);
-	kfree(r);
+	kfree(gaps);
 }
 
 /**
- * kdbus_msg_resources_ref() - Acquire reference to msg resources
- * @r:		resources to acquire ref to
+ * kdbus_gaps_ref() - gain reference
+ * @gaps:	gaps object
  *
- * Return: The acquired resource
+ * Return: @gaps is returned
  */
-struct kdbus_msg_resources *
-kdbus_msg_resources_ref(struct kdbus_msg_resources *r)
+struct kdbus_gaps *kdbus_gaps_ref(struct kdbus_gaps *gaps)
 {
-	if (r)
-		kref_get(&r->kref);
-	return r;
+	if (gaps)
+		kref_get(&gaps->kref);
+	return gaps;
 }
 
 /**
- * kdbus_msg_resources_unref() - Drop reference to msg resources
- * @r:		resources to drop reference of
+ * kdbus_gaps_unref() - drop reference
+ * @gaps:	gaps object
  *
  * Return: NULL
  */
-struct kdbus_msg_resources *
-kdbus_msg_resources_unref(struct kdbus_msg_resources *r)
+struct kdbus_gaps *kdbus_gaps_unref(struct kdbus_gaps *gaps)
 {
-	if (r)
-		kref_put(&r->kref, __kdbus_msg_resources_free);
+	if (gaps)
+		kref_put(&gaps->kref, kdbus_gaps_free);
 	return NULL;
 }
 
 /**
- * kdbus_kmsg_free() - free allocated message
- * @kmsg:		Message
- */
-void kdbus_kmsg_free(struct kdbus_kmsg *kmsg)
-{
-	if (!kmsg)
-		return;
-
-	kdbus_msg_resources_unref(kmsg->res);
-	kdbus_meta_conn_unref(kmsg->conn_meta);
-	kdbus_meta_proc_unref(kmsg->proc_meta);
-	kfree(kmsg->iov);
-	kfree(kmsg);
-}
-
-/**
- * kdbus_kmsg_new() - allocate message
- * @bus:		Bus this message is allocated on
- * @extra_size:		Additional size to reserve for data
+ * kdbus_gaps_install() - install file-descriptors
+ * @gaps:		gaps object, or NULL
+ * @slice:		pool slice that contains the message
+ * @out_incomplete	output variable to note incomplete fds
+ *
+ * This function installs all file-descriptors of @gaps into the current
+ * process and copies the file-descriptor numbers into the target pool slice.
  *
- * Return: new kdbus_kmsg on success, ERR_PTR on failure.
+ * If the file-descriptors were only partially installed, then @out_incomplete
+ * will be set to true. Otherwise, it's set to false.
+ *
+ * Return: 0 on success, negative error code on failure
  */
-struct kdbus_kmsg *kdbus_kmsg_new(struct kdbus_bus *bus, size_t extra_size)
+int kdbus_gaps_install(struct kdbus_gaps *gaps, struct kdbus_pool_slice *slice,
+		       bool *out_incomplete)
 {
-	struct kdbus_kmsg *m;
-	size_t size;
-	int ret;
+	bool incomplete_fds = false;
+	struct kvec kvec;
+	size_t i, n_fds;
+	int ret, *fds;
+
+	if (!gaps) {
+		/* nothing to do */
+		*out_incomplete = incomplete_fds;
+		return 0;
+	}
 
-	size = sizeof(struct kdbus_kmsg) + KDBUS_ITEM_SIZE(extra_size);
-	m = kzalloc(size, GFP_KERNEL);
-	if (!m)
-		return ERR_PTR(-ENOMEM);
+	n_fds = gaps->n_fds + gaps->n_memfds;
+	if (n_fds < 1) {
+		/* nothing to do */
+		*out_incomplete = incomplete_fds;
+		return 0;
+	}
 
-	m->seq = atomic64_inc_return(&bus->domain->last_id);
-	m->msg.size = size - KDBUS_KMSG_HEADER_SIZE;
-	m->msg.items[0].size = KDBUS_ITEM_SIZE(extra_size);
+	fds = kmalloc_array(n_fds, sizeof(*fds), GFP_TEMPORARY);
+	n_fds = 0;
+	if (!fds)
+		return -ENOMEM;
 
-	m->proc_meta = kdbus_meta_proc_new();
-	if (IS_ERR(m->proc_meta)) {
-		ret = PTR_ERR(m->proc_meta);
-		m->proc_meta = NULL;
-		goto exit;
+	/* 1) allocate fds and copy them over */
+
+	if (gaps->n_fds > 0) {
+		for (i = 0; i < gaps->n_fds; ++i) {
+			int fd;
+
+			fd = get_unused_fd_flags(O_CLOEXEC);
+			if (fd < 0)
+				incomplete_fds = true;
+
+			WARN_ON(!gaps->fd_files[i]);
+
+			fds[n_fds++] = fd < 0 ? -1 : fd;
+		}
+
+		/*
+		 * The file-descriptor array can only be present once per
+		 * message. Hence, prepare all fds and then copy them over with
+		 * a single kvec.
+		 */
+
+		WARN_ON(!gaps->fd_offset);
+
+		kvec.iov_base = fds;
+		kvec.iov_len = gaps->n_fds * sizeof(*fds);
+		ret = kdbus_pool_slice_copy_kvec(slice, gaps->fd_offset,
+						 &kvec, 1, kvec.iov_len);
+		if (ret < 0)
+			goto exit;
 	}
 
-	m->conn_meta = kdbus_meta_conn_new();
-	if (IS_ERR(m->conn_meta)) {
-		ret = PTR_ERR(m->conn_meta);
-		m->conn_meta = NULL;
-		goto exit;
+	for (i = 0; i < gaps->n_memfds; ++i) {
+		int memfd;
+
+		memfd = get_unused_fd_flags(O_CLOEXEC);
+		if (memfd < 0) {
+			incomplete_fds = true;
+			/* memfds are initialized to -1, skip copying it */
+			continue;
+		}
+
+		fds[n_fds++] = memfd;
+
+		/*
+		 * memfds have to be copied individually as they each are put
+		 * into a separate item. This should not be an issue, though,
+		 * as usually there is no need to send more than one memfd per
+		 * message.
+		 */
+
+		WARN_ON(!gaps->memfd_offsets[i]);
+		WARN_ON(!gaps->memfd_files[i]);
+
+		kvec.iov_base = &memfd;
+		kvec.iov_len = sizeof(memfd);
+		ret = kdbus_pool_slice_copy_kvec(slice, gaps->memfd_offsets[i],
+						 &kvec, 1, kvec.iov_len);
+		if (ret < 0)
+			goto exit;
 	}
 
-	return m;
+	/* 2) install fds now that everything was successful */
+
+	for (i = 0; i < gaps->n_fds; ++i)
+		if (fds[i] >= 0)
+			fd_install(fds[i], get_file(gaps->fd_files[i]));
+	for (i = 0; i < gaps->n_memfds; ++i)
+		if (fds[gaps->n_fds + i] >= 0)
+			fd_install(fds[gaps->n_fds + i],
+				   get_file(gaps->memfd_files[i]));
+
+	ret = 0;
 
 exit:
-	kdbus_kmsg_free(m);
-	return ERR_PTR(ret);
+	if (ret < 0)
+		for (i = 0; i < n_fds; ++i)
+			put_unused_fd(fds[i]);
+	kfree(fds);
+	*out_incomplete = incomplete_fds;
+	return ret;
 }
 
-static int kdbus_handle_check_file(struct file *file)
+static struct file *kdbus_get_fd(int fd)
 {
-	/*
-	 * Don't allow file descriptors in the transport that themselves allow
-	 * file descriptor queueing. This will eventually be allowed once both
-	 * unix domain sockets and kdbus share a generic garbage collector.
-	 */
+	struct file *f, *ret;
+	struct inode *inode;
+	struct socket *sock;
 
-	if (file->f_op == &kdbus_handle_ops)
-		return -EOPNOTSUPP;
+	if (fd < 0)
+		return ERR_PTR(-EBADF);
 
-	return 0;
+	f = fget_raw(fd);
+	if (!f)
+		return ERR_PTR(-EBADF);
+
+	inode = file_inode(f);
+	sock = S_ISSOCK(inode->i_mode) ? SOCKET_I(inode) : NULL;
+
+	if (f->f_mode & FMODE_PATH)
+		ret = f; /* O_PATH is always allowed */
+	else if (f->f_op == &kdbus_handle_ops)
+		ret = ERR_PTR(-EOPNOTSUPP); /* disallow kdbus-fd over kdbus */
+	else if (sock && sock->sk && sock->ops && sock->ops->family == PF_UNIX)
+		ret = ERR_PTR(-EOPNOTSUPP); /* disallow UDS over kdbus */
+	else
+		ret = f; /* all other are allowed */
+
+	if (f != ret)
+		fput(f);
+
+	return ret;
 }
 
-static const char * const zeros = "\0\0\0\0\0\0\0";
+static struct file *kdbus_get_memfd(const struct kdbus_memfd *memfd)
+{
+	const int m = F_SEAL_SHRINK | F_SEAL_GROW | F_SEAL_WRITE | F_SEAL_SEAL;
+	struct file *f, *ret;
+	int s;
+
+	if (memfd->fd < 0)
+		return ERR_PTR(-EBADF);
+
+	f = fget(memfd->fd);
+	if (!f)
+		return ERR_PTR(-EBADF);
+
+	s = shmem_get_seals(f);
+	if (s < 0)
+		ret = ERR_PTR(-EMEDIUMTYPE);
+	else if ((s & m) != m)
+		ret = ERR_PTR(-ETXTBSY);
+	else if (memfd->start + memfd->size > (u64)i_size_read(file_inode(f)))
+		ret = ERR_PTR(-EFAULT);
+	else
+		ret = f;
+
+	if (f != ret)
+		fput(f);
+
+	return ret;
+}
 
-/*
- * kdbus_msg_scan_items() - validate incoming data and prepare parsing
- * @kmsg:		Message
- * @bus:		Bus the message is sent over
- *
- * Return: 0 on success, negative errno on failure.
- *
- * Files references in MEMFD or FDS items are pinned.
- *
- * On errors, the caller should drop any taken reference with
- * kdbus_kmsg_free()
- */
-static int kdbus_msg_scan_items(struct kdbus_kmsg *kmsg,
-				struct kdbus_bus *bus)
+static int kdbus_msg_examine(struct kdbus_msg *msg, struct kdbus_bus *bus,
+			     struct kdbus_cmd_send *cmd, size_t *out_n_memfds,
+			     size_t *out_n_fds, size_t *out_n_parts)
 {
-	struct kdbus_msg_resources *res = kmsg->res;
-	const struct kdbus_msg *msg = &kmsg->msg;
-	const struct kdbus_item *item;
-	size_t n, n_vecs, n_memfds;
-	bool has_bloom = false;
-	bool has_name = false;
-	bool has_fds = false;
-	bool is_broadcast;
-	bool is_signal;
-	u64 vec_size;
-
-	is_broadcast = (msg->dst_id == KDBUS_DST_ID_BROADCAST);
-	is_signal = !!(msg->flags & KDBUS_MSG_SIGNAL);
-
-	/* count data payloads */
-	n_vecs = 0;
-	n_memfds = 0;
-	KDBUS_ITEMS_FOREACH(item, msg->items, KDBUS_ITEMS_SIZE(msg, items)) {
-		switch (item->type) {
-		case KDBUS_ITEM_PAYLOAD_VEC:
-			++n_vecs;
-			break;
-		case KDBUS_ITEM_PAYLOAD_MEMFD:
-			++n_memfds;
-			if (item->memfd.size % 8)
-				++n_vecs;
-			break;
-		default:
-			break;
-		}
-	}
+	struct kdbus_item *item, *fds = NULL, *bloom = NULL, *dstname = NULL;
+	u64 n_parts, n_memfds, n_fds, vec_size;
 
-	n = n_vecs + n_memfds;
-	if (n > 0) {
-		res->data = kcalloc(n, sizeof(*res->data), GFP_KERNEL);
-		if (!res->data)
-			return -ENOMEM;
+	/*
+	 * Step 1:
+	 * Validate the message and command parameters.
+	 */
+
+	/* KDBUS_PAYLOAD_KERNEL is reserved to kernel messages */
+	if (msg->payload_type == KDBUS_PAYLOAD_KERNEL)
+		return -EINVAL;
+
+	if (msg->dst_id == KDBUS_DST_ID_BROADCAST) {
+		/* broadcasts must be marked as signals */
+		if (!(msg->flags & KDBUS_MSG_SIGNAL))
+			return -EBADMSG;
+		/* broadcasts cannot have timeouts */
+		if (msg->timeout_ns > 0)
+			return -ENOTUNIQ;
 	}
 
-	if (n_vecs > 0) {
-		kmsg->iov = kcalloc(n_vecs, sizeof(*kmsg->iov), GFP_KERNEL);
-		if (!kmsg->iov)
-			return -ENOMEM;
+	if (msg->flags & KDBUS_MSG_EXPECT_REPLY) {
+		/* if you expect a reply, you must specify a timeout */
+		if (msg->timeout_ns == 0)
+			return -EINVAL;
+		/* signals cannot have replies */
+		if (msg->flags & KDBUS_MSG_SIGNAL)
+			return -ENOTUNIQ;
+	} else {
+		/* must expect reply if sent as synchronous call */
+		if (cmd->flags & KDBUS_SEND_SYNC_REPLY)
+			return -EINVAL;
+		/* cannot mark replies as signal */
+		if (msg->cookie_reply && (msg->flags & KDBUS_MSG_SIGNAL))
+			return -EINVAL;
 	}
 
-	/* import data payloads */
-	n = 0;
-	vec_size = 0;
-	KDBUS_ITEMS_FOREACH(item, msg->items, KDBUS_ITEMS_SIZE(msg, items)) {
-		size_t payload_size = KDBUS_ITEM_PAYLOAD_SIZE(item);
-		struct iovec *iov = kmsg->iov + kmsg->iov_count;
+	/*
+	 * Step 2:
+	 * Validate all passed items. While at it, select some statistics that
+	 * are required to allocate state objects later on.
+	 *
+	 * Generic item validation has already been done via
+	 * kdbus_item_validate(). Furthermore, the number of items is naturally
+	 * limited by the maximum message size. Hence, only non-generic item
+	 * checks are performed here (mainly integer overflow tests).
+	 */
 
-		if (++n > KDBUS_MSG_MAX_ITEMS)
-			return -E2BIG;
+	n_parts = 0;
+	n_memfds = 0;
+	n_fds = 0;
+	vec_size = 0;
 
+	KDBUS_ITEMS_FOREACH(item, msg->items, KDBUS_ITEMS_SIZE(msg, items)) {
 		switch (item->type) {
 		case KDBUS_ITEM_PAYLOAD_VEC: {
-			struct kdbus_msg_data *d = res->data + res->data_count;
 			void __force __user *ptr = KDBUS_PTR(item->vec.address);
-			size_t size = item->vec.size;
+			u64 size = item->vec.size;
 
 			if (vec_size + size < vec_size)
 				return -EMSGSIZE;
 			if (vec_size + size > KDBUS_MSG_MAX_PAYLOAD_VEC_SIZE)
 				return -EMSGSIZE;
+			if (ptr && unlikely(!access_ok(VERIFY_READ, ptr, size)))
+				return -EFAULT;
 
-			d->type = KDBUS_MSG_DATA_VEC;
-			d->size = size;
-
-			if (ptr) {
-				if (unlikely(!access_ok(VERIFY_READ, ptr,
-							size)))
-					return -EFAULT;
-
-				d->vec.off = kmsg->pool_size;
-				iov->iov_base = ptr;
-				iov->iov_len = size;
-			} else {
-				d->vec.off = ~0ULL;
-				iov->iov_base = (char __user *)zeros;
-				iov->iov_len = size % 8;
-			}
-
-			if (kmsg->pool_size + iov->iov_len < kmsg->pool_size)
-				return -EMSGSIZE;
-
-			kmsg->pool_size += iov->iov_len;
-			++kmsg->iov_count;
-			++res->vec_count;
-			++res->data_count;
-			vec_size += size;
-
+			if (ptr || size % 8) /* data or padding */
+				++n_parts;
 			break;
 		}
-
 		case KDBUS_ITEM_PAYLOAD_MEMFD: {
-			struct kdbus_msg_data *d = res->data + res->data_count;
 			u64 start = item->memfd.start;
 			u64 size = item->memfd.size;
-			size_t pad = size % 8;
-			int seals, mask;
-			struct file *f;
 
-			if (kmsg->pool_size + size % 8 < kmsg->pool_size)
-				return -EMSGSIZE;
 			if (start + size < start)
 				return -EMSGSIZE;
-
-			if (item->memfd.fd < 0)
-				return -EBADF;
-
-			if (res->memfd_count >= KDBUS_MSG_MAX_MEMFD_ITEMS)
+			if (n_memfds >= KDBUS_MSG_MAX_MEMFD_ITEMS)
 				return -E2BIG;
 
-			f = fget(item->memfd.fd);
-			if (!f)
-				return -EBADF;
-
-			if (pad) {
-				iov->iov_base = (char __user *)zeros;
-				iov->iov_len = pad;
+			++n_memfds;
+			if (size % 8) /* vec-padding required */
+				++n_parts;
+			break;
+		}
+		case KDBUS_ITEM_FDS: {
+			if (fds)
+				return -EEXIST;
 
-				kmsg->pool_size += pad;
-				++kmsg->iov_count;
-			}
+			fds = item;
+			n_fds = KDBUS_ITEM_PAYLOAD_SIZE(item) / sizeof(int);
+			if (n_fds > KDBUS_CONN_MAX_FDS_PER_USER)
+				return -EMFILE;
 
-			++res->data_count;
-			++res->memfd_count;
+			break;
+		}
+		case KDBUS_ITEM_BLOOM_FILTER: {
+			u64 bloom_size;
 
-			d->type = KDBUS_MSG_DATA_MEMFD;
-			d->size = size;
-			d->memfd.start = start;
-			d->memfd.file = f;
+			if (bloom)
+				return -EEXIST;
 
-			/*
-			 * We only accept a sealed memfd file whose content
-			 * cannot be altered by the sender or anybody else
-			 * while it is shared or in-flight. Other files need
-			 * to be passed with KDBUS_MSG_FDS.
-			 */
-			seals = shmem_get_seals(f);
-			if (seals < 0)
-				return -EMEDIUMTYPE;
+			bloom = item;
+			bloom_size = KDBUS_ITEM_PAYLOAD_SIZE(item) -
+				     offsetof(struct kdbus_bloom_filter, data);
+			if (!KDBUS_IS_ALIGNED8(bloom_size))
+				return -EFAULT;
+			if (bloom_size != bus->bloom.size)
+				return -EDOM;
 
-			mask = F_SEAL_SHRINK | F_SEAL_GROW |
-				F_SEAL_WRITE | F_SEAL_SEAL;
-			if ((seals & mask) != mask)
-				return -ETXTBSY;
+			break;
+		}
+		case KDBUS_ITEM_DST_NAME: {
+			if (dstname)
+				return -EEXIST;
 
-			if (start + size > (u64)i_size_read(file_inode(f)))
-				return -EBADF;
+			dstname = item;
+			if (!kdbus_name_is_valid(item->str, false))
+				return -EINVAL;
+			if (msg->dst_id == KDBUS_DST_ID_BROADCAST)
+				return -EBADMSG;
 
 			break;
 		}
+		default:
+			return -EINVAL;
+		}
+	}
 
-		case KDBUS_ITEM_FDS: {
-			unsigned int i;
-			unsigned int fds_count = payload_size / sizeof(int);
+	/*
+	 * Step 3:
+	 * Validate that required items were actually passed, and that no item
+	 * contradicts the message flags.
+	 */
 
-			/* do not allow multiple fd arrays */
-			if (has_fds)
-				return -EEXIST;
-			has_fds = true;
+	/* bloom filters must be attached _iff_ it's a signal */
+	if (!(msg->flags & KDBUS_MSG_SIGNAL) != !bloom)
+		return -EBADMSG;
+	/* destination name is required if no ID is given */
+	if (msg->dst_id == KDBUS_DST_ID_NAME && !dstname)
+		return -EDESTADDRREQ;
+	/* cannot send file-descriptors attached to broadcasts */
+	if (msg->dst_id == KDBUS_DST_ID_BROADCAST && fds)
+		return -ENOTUNIQ;
 
-			/* Do not allow to broadcast file descriptors */
-			if (is_broadcast)
-				return -ENOTUNIQ;
+	*out_n_memfds = n_memfds;
+	*out_n_fds = n_fds;
+	*out_n_parts = n_parts;
 
-			if (fds_count > KDBUS_CONN_MAX_FDS_PER_USER)
-				return -EMFILE;
+	return 0;
+}
+
+static bool kdbus_staging_merge_vecs(struct kdbus_staging *staging,
+				     struct kdbus_item **prev_item,
+				     struct iovec **prev_vec,
+				     const struct kdbus_item *merge)
+{
+	void __user *ptr = (void __user *)KDBUS_PTR(merge->vec.address);
+	u64 padding = merge->vec.size % 8;
+	struct kdbus_item *prev = *prev_item;
+	struct iovec *vec = *prev_vec;
+
+	/* XXX: merging is disabled so far */
+	if (0 && prev && prev->type == KDBUS_ITEM_PAYLOAD_OFF &&
+	    !merge->vec.address == !prev->vec.address) {
+		/*
+		 * If we merge two VECs, we can always drop the second
+		 * PAYLOAD_VEC item. Hence, include its size in the previous
+		 * one.
+		 */
+		prev->vec.size += merge->vec.size;
+
+		if (ptr) {
+			/*
+			 * If we merge two data VECs, we need two iovecs to copy
+			 * the data. But the items can be easily merged by
+			 * summing their lengths.
+			 */
+			vec = &staging->parts[staging->n_parts++];
+			vec->iov_len = merge->vec.size;
+			vec->iov_base = ptr;
+			staging->n_payload += vec->iov_len;
+		} else if (padding) {
+			/*
+			 * If we merge two 0-vecs with the second 0-vec
+			 * requiring padding, we need to insert an iovec to copy
+			 * the 0-padding. We try merging it with the previous
+			 * 0-padding iovec. This might end up with an
+			 * iov_len==0, in which case we simply drop the iovec.
+			 */
+			if (vec) {
+				staging->n_payload -= vec->iov_len;
+				vec->iov_len = prev->vec.size % 8;
+				if (!vec->iov_len) {
+					--staging->n_parts;
+					vec = NULL;
+				} else {
+					staging->n_payload += vec->iov_len;
+				}
+			} else {
+				vec = &staging->parts[staging->n_parts++];
+				vec->iov_len = padding;
+				vec->iov_base = (char __user *)zeros;
+				staging->n_payload += vec->iov_len;
+			}
+		} else {
+			/*
+			 * If we merge two 0-vecs with the second 0-vec having
+			 * no padding, we know the padding of the first stays
+			 * the same. Hence, @vec needs no adjustment.
+			 */
+		}
 
-			res->fds = kcalloc(fds_count, sizeof(struct file *),
-					   GFP_KERNEL);
-			if (!res->fds)
-				return -ENOMEM;
+		/* successfully merged with previous item */
+		merge = prev;
+	} else {
+		/*
+		 * If we cannot merge the payload item with the previous one,
+		 * we simply insert a new iovec for the data/padding.
+		 */
+		if (ptr) {
+			vec = &staging->parts[staging->n_parts++];
+			vec->iov_len = merge->vec.size;
+			vec->iov_base = ptr;
+			staging->n_payload += vec->iov_len;
+		} else if (padding) {
+			vec = &staging->parts[staging->n_parts++];
+			vec->iov_len = padding;
+			vec->iov_base = (char __user *)zeros;
+			staging->n_payload += vec->iov_len;
+		} else {
+			vec = NULL;
+		}
+	}
 
-			for (i = 0; i < fds_count; i++) {
-				int fd = item->fds[i];
-				int ret;
+	*prev_item = (struct kdbus_item *)merge;
+	*prev_vec = vec;
 
-				/*
-				 * Verify the fd and increment the usage count.
-				 * Use fget_raw() to allow passing O_PATH fds.
-				 */
-				if (fd < 0)
-					return -EBADF;
+	return merge == prev;
+}
 
-				res->fds[i] = fget_raw(fd);
-				if (!res->fds[i])
-					return -EBADF;
+static int kdbus_staging_import(struct kdbus_staging *staging)
+{
+	struct kdbus_item *it, *item, *last, *prev_payload;
+	struct kdbus_gaps *gaps = staging->gaps;
+	struct kdbus_msg *msg = staging->msg;
+	struct iovec *part, *prev_part;
+	bool drop_item;
 
-				res->fds_count++;
+	drop_item = false;
+	last = NULL;
+	prev_payload = NULL;
+	prev_part = NULL;
 
-				ret = kdbus_handle_check_file(res->fds[i]);
-				if (ret < 0)
-					return ret;
+	/*
+	 * We modify msg->items along the way; make sure to use @item as offset
+	 * to the next item (instead of the iterator @it).
+	 */
+	for (it = item = msg->items;
+	     it >= msg->items &&
+	             (u8 *)it < (u8 *)msg + msg->size &&
+	             (u8 *)it + it->size <= (u8 *)msg + msg->size; ) {
+		/*
+		 * If we dropped items along the way, move current item to
+		 * front. We must not access @it afterwards, but use @item
+		 * instead!
+		 */
+		if (it != item)
+			memmove(item, it, it->size);
+		it = (void *)((u8 *)it + KDBUS_ALIGN8(item->size));
+
+		switch (item->type) {
+		case KDBUS_ITEM_PAYLOAD_VEC: {
+			size_t offset = staging->n_payload;
+
+			if (kdbus_staging_merge_vecs(staging, &prev_payload,
+						     &prev_part, item)) {
+				drop_item = true;
+			} else if (item->vec.address) {
+				/* real offset is patched later on */
+				item->type = KDBUS_ITEM_PAYLOAD_OFF;
+				item->vec.offset = offset;
+			} else {
+				item->type = KDBUS_ITEM_PAYLOAD_OFF;
+				item->vec.offset = ~0ULL;
 			}
 
 			break;
 		}
+		case KDBUS_ITEM_PAYLOAD_MEMFD: {
+			struct file *f;
 
-		case KDBUS_ITEM_BLOOM_FILTER: {
-			u64 bloom_size;
+			f = kdbus_get_memfd(&item->memfd);
+			if (IS_ERR(f))
+				return PTR_ERR(f);
+
+			gaps->memfd_files[gaps->n_memfds] = f;
+			gaps->memfd_offsets[gaps->n_memfds] =
+					(u8 *)&item->memfd.fd - (u8 *)msg;
+			++gaps->n_memfds;
+
+			/* memfds cannot be merged */
+			prev_payload = item;
+			prev_part = NULL;
+
+			/* insert padding to make following VECs aligned */
+			if (item->memfd.size % 8) {
+				part = &staging->parts[staging->n_parts++];
+				part->iov_len = item->memfd.size % 8;
+				part->iov_base = (char __user *)zeros;
+				staging->n_payload += part->iov_len;
+			}
 
-			/* do not allow multiple bloom filters */
-			if (has_bloom)
-				return -EEXIST;
-			has_bloom = true;
+			break;
+		}
+		case KDBUS_ITEM_FDS: {
+			size_t i, n_fds;
 
-			bloom_size = payload_size -
-				     offsetof(struct kdbus_bloom_filter, data);
+			n_fds = KDBUS_ITEM_PAYLOAD_SIZE(item) / sizeof(int);
+			for (i = 0; i < n_fds; ++i) {
+				struct file *f;
 
-			/*
-			* Allow only bloom filter sizes of a multiple of 64bit.
-			*/
-			if (!KDBUS_IS_ALIGNED8(bloom_size))
-				return -EFAULT;
+				f = kdbus_get_fd(item->fds[i]);
+				if (IS_ERR(f))
+					return PTR_ERR(f);
 
-			/* do not allow mismatching bloom filter sizes */
-			if (bloom_size != bus->bloom.size)
-				return -EDOM;
+				gaps->fd_files[gaps->n_fds++] = f;
+			}
+
+			gaps->fd_offset = (u8 *)item->fds - (u8 *)msg;
 
-			kmsg->bloom_filter = &item->bloom_filter;
 			break;
 		}
-
+		case KDBUS_ITEM_BLOOM_FILTER:
+			staging->bloom_filter = &item->bloom_filter;
+			break;
 		case KDBUS_ITEM_DST_NAME:
-			/* do not allow multiple names */
-			if (has_name)
-				return -EEXIST;
-			has_name = true;
-
-			if (!kdbus_name_is_valid(item->str, false))
-				return -EINVAL;
-
-			res->dst_name = kstrdup(item->str, GFP_KERNEL);
-			if (!res->dst_name)
-				return -ENOMEM;
+			staging->dst_name = item->str;
 			break;
+		}
 
-		default:
-			return -EINVAL;
+		/* drop item if we merged it with a previous one */
+		if (drop_item) {
+			drop_item = false;
+		} else {
+			last = item;
+			item = KDBUS_ITEM_NEXT(item);
 		}
 	}
 
-	/* name is needed if no ID is given */
-	if (msg->dst_id == KDBUS_DST_ID_NAME && !has_name)
-		return -EDESTADDRREQ;
+	/* adjust message size regarding dropped items */
+	msg->size = offsetof(struct kdbus_msg, items);
+	if (last)
+		msg->size += ((u8 *)last - (u8 *)msg->items) + last->size;
 
-	if (is_broadcast) {
-		/* Broadcasts can't take names */
-		if (has_name)
-			return -EBADMSG;
+	return 0;
+}
 
-		/* All broadcasts have to be signals */
-		if (!is_signal)
-			return -EBADMSG;
+static void kdbus_staging_reserve(struct kdbus_staging *staging)
+{
+	struct iovec *part;
 
-		/* Timeouts are not allowed for broadcasts */
-		if (msg->timeout_ns > 0)
-			return -ENOTUNIQ;
+	part = &staging->parts[staging->n_parts++];
+	part->iov_base = (void __user *)zeros;
+	part->iov_len = 0;
+}
+
+static struct kdbus_staging *kdbus_staging_new(struct kdbus_bus *bus,
+					       size_t n_parts,
+					       size_t msg_extra_size)
+{
+	const size_t reserved_parts = 5; /* see below for explanation */
+	struct kdbus_staging *staging;
+	int ret;
+
+	n_parts += reserved_parts;
+
+	staging = kzalloc(sizeof(*staging) + n_parts * sizeof(*staging->parts) +
+			  msg_extra_size, GFP_TEMPORARY);
+	if (!staging)
+		return ERR_PTR(-ENOMEM);
+
+	staging->msg_seqnum = atomic64_inc_return(&bus->last_message_id);
+	staging->n_parts = 0; /* we reserve n_parts, but don't enforce them */
+	staging->parts = (void *)(staging + 1);
+
+	if (msg_extra_size) /* if requested, allocate message, too */
+		staging->msg = (void *)((u8 *)staging->parts +
+				        n_parts * sizeof(*staging->parts));
+
+	staging->meta_proc = kdbus_meta_proc_new();
+	if (IS_ERR(staging->meta_proc)) {
+		ret = PTR_ERR(staging->meta_proc);
+		staging->meta_proc = NULL;
+		goto error;
+	}
+
+	staging->meta_conn = kdbus_meta_conn_new();
+	if (IS_ERR(staging->meta_conn)) {
+		ret = PTR_ERR(staging->meta_conn);
+		staging->meta_conn = NULL;
+		goto error;
 	}
 
 	/*
-	 * Signal messages require a bloom filter, and bloom filters are
-	 * only valid with signals.
+	 * Prepare iovecs to copy the message into the target pool. We use the
+	 * following iovecs:
+	 *   * iovec to copy "kdbus_msg.size"
+	 *   * iovec to copy "struct kdbus_msg" (minus size) plus items
+	 *   * iovec for possible padding after the items
+	 *   * iovec for metadata items
+	 *   * iovec for possible padding after the items
+	 *
+	 * Make sure to update @reserved_parts if you add more parts here.
 	 */
-	if (is_signal ^ has_bloom)
-		return -EBADMSG;
 
-	return 0;
+	kdbus_staging_reserve(staging); /* msg.size */
+	kdbus_staging_reserve(staging); /* msg (minus msg.size) plus items */
+	kdbus_staging_reserve(staging); /* msg padding */
+	kdbus_staging_reserve(staging); /* meta */
+	kdbus_staging_reserve(staging); /* meta padding */
+
+	return staging;
+
+error:
+	kdbus_staging_free(staging);
+	return ERR_PTR(ret);
 }
 
-/**
- * kdbus_kmsg_new_from_cmd() - create kernel message from send payload
- * @conn:		Connection
- * @cmd_send:		Payload of KDBUS_CMD_SEND
- *
- * Return: a new kdbus_kmsg on success, ERR_PTR on failure.
- */
-struct kdbus_kmsg *kdbus_kmsg_new_from_cmd(struct kdbus_conn *conn,
-					   struct kdbus_cmd_send *cmd_send)
+struct kdbus_staging *kdbus_staging_new_kernel(struct kdbus_bus *bus,
+					       u64 dst, u64 cookie_timeout,
+					       size_t it_size, size_t it_type)
+{
+	struct kdbus_staging *staging;
+	size_t size;
+
+	size = offsetof(struct kdbus_msg, items) +
+	       KDBUS_ITEM_HEADER_SIZE + it_size;
+
+	staging = kdbus_staging_new(bus, 0, KDBUS_ALIGN8(size));
+	if (IS_ERR(staging))
+		return ERR_CAST(staging);
+
+	staging->msg->size = size;
+	staging->msg->flags = (dst == KDBUS_DST_ID_BROADCAST) ?
+							KDBUS_MSG_SIGNAL : 0;
+	staging->msg->dst_id = dst;
+	staging->msg->src_id = KDBUS_SRC_ID_KERNEL;
+	staging->msg->payload_type = KDBUS_PAYLOAD_KERNEL;
+	staging->msg->cookie_reply = cookie_timeout;
+	staging->notify = staging->msg->items;
+	staging->notify->size = KDBUS_ITEM_HEADER_SIZE + it_size;
+	staging->notify->type = it_type;
+
+	return staging;
+}
+
+struct kdbus_staging *kdbus_staging_new_user(struct kdbus_bus *bus,
+					     struct kdbus_cmd_send *cmd,
+					     struct kdbus_msg *msg)
 {
-	struct kdbus_kmsg *m;
-	u64 size;
+	const size_t reserved_parts = 1; /* see below for explanation */
+	size_t n_memfds, n_fds, n_parts;
+	struct kdbus_staging *staging;
 	int ret;
 
-	ret = kdbus_copy_from_user(&size, KDBUS_PTR(cmd_send->msg_address),
-				   sizeof(size));
+	/*
+	 * Examine user-supplied message and figure out how many resources we
+	 * need to allocate in our staging area. This requires us to iterate
+	 * the message twice, but saves us from re-allocating our resources
+	 * all the time.
+	 */
+
+	ret = kdbus_msg_examine(msg, bus, cmd, &n_memfds, &n_fds, &n_parts);
 	if (ret < 0)
 		return ERR_PTR(ret);
 
-	if (size < sizeof(struct kdbus_msg) || size > KDBUS_MSG_MAX_SIZE)
-		return ERR_PTR(-EINVAL);
+	n_parts += reserved_parts;
 
-	m = kmalloc(size + KDBUS_KMSG_HEADER_SIZE, GFP_KERNEL);
-	if (!m)
-		return ERR_PTR(-ENOMEM);
+	/*
+	 * Allocate staging area with the number of required resources. Make
+	 * sure that we have enough iovecs for all required parts pre-allocated
+	 * so this will hopefully be the only memory allocation for this
+	 * message transaction.
+	 */
 
-	memset(m, 0, KDBUS_KMSG_HEADER_SIZE);
-	m->seq = atomic64_inc_return(&conn->ep->bus->domain->last_id);
+	staging = kdbus_staging_new(bus, n_parts, 0);
+	if (IS_ERR(staging))
+		return ERR_CAST(staging);
 
-	m->proc_meta = kdbus_meta_proc_new();
-	if (IS_ERR(m->proc_meta)) {
-		ret = PTR_ERR(m->proc_meta);
-		m->proc_meta = NULL;
-		goto exit_free;
-	}
+	staging->msg = msg;
 
-	m->conn_meta = kdbus_meta_conn_new();
-	if (IS_ERR(m->conn_meta)) {
-		ret = PTR_ERR(m->conn_meta);
-		m->conn_meta = NULL;
-		goto exit_free;
-	}
+	/*
+	 * If the message contains memfds or fd items, we need to remember some
+	 * state so we can fill in the requested information at RECV time.
+	 * File-descriptors cannot be passed at SEND time. Hence, allocate a
+	 * gaps-object to remember that state. That gaps object is linked to
+	 * from the staging area, but will also be linked to from the message
+	 * queue of each peer. Hence, each receiver owns a reference to it, and
+	 * it will later be used to fill the 'gaps' in message that couldn't be
+	 * filled at SEND time.
+	 * Note that the 'gaps' object is read-only once the staging-allocator
+	 * returns. There might be connections receiving a queued message while
+	 * the sender still broadcasts the message to other receivers.
+	 */
 
-	if (copy_from_user(&m->msg, KDBUS_PTR(cmd_send->msg_address), size)) {
-		ret = -EFAULT;
-		goto exit_free;
+	if (n_memfds > 0 || n_fds > 0) {
+		staging->gaps = kdbus_gaps_new(n_memfds, n_fds);
+		if (IS_ERR(staging->gaps)) {
+			ret = PTR_ERR(staging->gaps);
+			staging->gaps = NULL;
+			kdbus_staging_free(staging);
+			return ERR_PTR(ret);
+		}
 	}
 
-	if (m->msg.size != size) {
-		ret = -EINVAL;
-		goto exit_free;
-	}
+	/*
+	 * kdbus_staging_new() already reserves parts for message setup. For
+	 * user-supplied messages, we add the following iovecs:
+	 *   ... variable number of iovecs for payload ...
+	 *   * final iovec for possible padding of payload
+	 *
+	 * Make sure to update @reserved_parts if you add more parts here.
+	 */
+
+	ret = kdbus_staging_import(staging); /* payload */
+	kdbus_staging_reserve(staging); /* payload padding */
+
+	if (ret < 0)
+		goto error;
+
+	return staging;
 
-	if (m->msg.flags & ~(KDBUS_MSG_EXPECT_REPLY |
-			     KDBUS_MSG_NO_AUTO_START |
-			     KDBUS_MSG_SIGNAL)) {
-		ret = -EINVAL;
-		goto exit_free;
+error:
+	kdbus_staging_free(staging);
+	return ERR_PTR(ret);
+}
+
+struct kdbus_staging *kdbus_staging_free(struct kdbus_staging *staging)
+{
+	if (!staging)
+		return NULL;
+
+	kdbus_meta_conn_unref(staging->meta_conn);
+	kdbus_meta_proc_unref(staging->meta_proc);
+	kdbus_gaps_unref(staging->gaps);
+	kfree(staging);
+
+	return NULL;
+}
+
+static int kdbus_staging_collect_metadata(struct kdbus_staging *staging,
+					  struct kdbus_conn *src,
+					  struct kdbus_conn *dst,
+					  u64 *out_attach)
+{
+	u64 attach;
+	int ret;
+
+	if (src)
+		attach = kdbus_meta_msg_mask(src, dst);
+	else
+		attach = KDBUS_ATTACH_TIMESTAMP; /* metadata for kernel msgs */
+
+	if (src && !src->meta_fake) {
+		ret = kdbus_meta_proc_collect(staging->meta_proc, attach);
+		if (ret < 0)
+			return ret;
 	}
 
-	ret = kdbus_items_validate(m->msg.items,
-				   KDBUS_ITEMS_SIZE(&m->msg, items));
+	ret = kdbus_meta_conn_collect(staging->meta_conn, src,
+				      staging->msg_seqnum, attach);
 	if (ret < 0)
-		goto exit_free;
+		return ret;
 
-	m->res = kdbus_msg_resources_new();
-	if (IS_ERR(m->res)) {
-		ret = PTR_ERR(m->res);
-		m->res = NULL;
-		goto exit_free;
+	*out_attach = attach;
+	return 0;
+}
+
+/**
+ * kdbus_staging_emit() - emit linearized message in target pool
+ * @staging:		staging object to create message from
+ * @src:		sender of the message (or NULL)
+ * @dst:		target connection to allocate message for
+ *
+ * This allocates a pool-slice for @dst and copies the message provided by
+ * @staging into it. The new slice is then returned to the caller for further
+ * processing. It's not linked into any queue, yet.
+ *
+ * Return: Newly allocated slice or ERR_PTR on failure.
+ */
+struct kdbus_pool_slice *kdbus_staging_emit(struct kdbus_staging *staging,
+					    struct kdbus_conn *src,
+					    struct kdbus_conn *dst)
+{
+	struct kdbus_item *item, *meta_items = NULL;
+	struct kdbus_pool_slice *slice = NULL;
+	size_t off, size, meta_size;
+	struct iovec *v;
+	u64 attach, msg_size;
+	int ret;
+
+	/*
+	 * Step 1:
+	 * Collect metadata from @src depending on the attach-flags allowed for
+	 * @dst. Translate it into the namespaces pinned by @dst.
+	 */
+
+	ret = kdbus_staging_collect_metadata(staging, src, dst, &attach);
+	if (ret < 0)
+		goto error;
+
+	ret = kdbus_meta_emit(staging->meta_proc, NULL, staging->meta_conn,
+			      dst, attach, &meta_items, &meta_size);
+	if (ret < 0)
+		goto error;
+
+	/*
+	 * Step 2:
+	 * Setup iovecs for the message. See kdbus_staging_new() for allocation
+	 * of those iovecs. All reserved iovecs have been initialized with
+	 * iov_len=0 + iov_base=zeros. Furthermore, the iovecs to copy the
+	 * actual message payload have already been initialized and need not be
+	 * touched.
+	 */
+
+	v = staging->parts;
+	msg_size = staging->msg->size;
+
+	/* msg.size */
+	v->iov_len = sizeof(msg_size);
+	v->iov_base = (void __user *)&msg_size;
+	++v;
+
+	/* msg (after msg.size) plus items */
+	v->iov_len = staging->msg->size - sizeof(staging->msg->size);
+	v->iov_base = (void __user *)((u8 *)staging->msg +
+				      sizeof(staging->msg->size));
+	++v;
+
+	/* padding after msg */
+	v->iov_len = KDBUS_ALIGN8(staging->msg->size) - staging->msg->size;
+	v->iov_base = (void __user *)zeros;
+	++v;
+
+	if (meta_size > 0) {
+		/* metadata items */
+		v->iov_len = meta_size;
+		v->iov_base = (void __user *)meta_items;
+		++v;
+
+		/* padding after metadata */
+		v->iov_len = KDBUS_ALIGN8(meta_size) - meta_size;
+		v->iov_base = (void __user *)zeros;
+		++v;
+
+		msg_size = KDBUS_ALIGN8(msg_size) + meta_size;
+	} else {
+		/* metadata items */
+		v->iov_len = 0;
+		v->iov_base = (void __user *)zeros;
+		++v;
+
+		/* padding after metadata */
+		v->iov_len = 0;
+		v->iov_base = (void __user *)zeros;
+		++v;
 	}
 
-	/* do not accept kernel-generated messages */
-	if (m->msg.payload_type == KDBUS_PAYLOAD_KERNEL) {
-		ret = -EINVAL;
-		goto exit_free;
+	/* ... payload iovecs are already filled in ... */
+
+	/* compute overall size and fill in padding after payload */
+	size = KDBUS_ALIGN8(msg_size);
+
+	if (staging->n_payload > 0) {
+		size += staging->n_payload;
+
+		v = &staging->parts[staging->n_parts - 1];
+		v->iov_len = KDBUS_ALIGN8(size) - size;
+		v->iov_base = (void __user *)zeros;
+
+		size = KDBUS_ALIGN8(size);
 	}
 
-	if (m->msg.flags & KDBUS_MSG_EXPECT_REPLY) {
-		/* requests for replies need timeout and cookie */
-		if (m->msg.timeout_ns == 0 || m->msg.cookie == 0) {
-			ret = -EINVAL;
-			goto exit_free;
-		}
+	/*
+	 * Step 3:
+	 * The PAYLOAD_OFF items in the message contain a relative 'offset'
+	 * field that tells the receiver where to find the actual payload. This
+	 * offset is relative to the start of the message, and as such depends
+	 * on the size of the metadata items we inserted. This size is variable
+	 * and changes for each peer we send the message to. Hence, we remember
+	 * the last relative offset that was used to calculate the 'offset'
+	 * fields. For each message, we re-calculate it and patch all items, in
+	 * case it changed.
+	 */
 
-		/* replies may not be expected for broadcasts */
-		if (m->msg.dst_id == KDBUS_DST_ID_BROADCAST) {
-			ret = -ENOTUNIQ;
-			goto exit_free;
-		}
+	off = KDBUS_ALIGN8(msg_size);
 
-		/* replies may not be expected for signals */
-		if (m->msg.flags & KDBUS_MSG_SIGNAL) {
-			ret = -EINVAL;
-			goto exit_free;
-		}
-	} else {
-		/*
-		 * KDBUS_SEND_SYNC_REPLY is only valid together with
-		 * KDBUS_MSG_EXPECT_REPLY
-		 */
-		if (cmd_send->flags & KDBUS_SEND_SYNC_REPLY) {
-			ret = -EINVAL;
-			goto exit_free;
-		}
+	if (off != staging->i_payload) {
+		KDBUS_ITEMS_FOREACH(item, staging->msg->items,
+				    KDBUS_ITEMS_SIZE(staging->msg, items)) {
+			if (item->type != KDBUS_ITEM_PAYLOAD_OFF)
+				continue;
 
-		/* replies cannot be signals */
-		if (m->msg.cookie_reply && (m->msg.flags & KDBUS_MSG_SIGNAL)) {
-			ret = -EINVAL;
-			goto exit_free;
+			item->vec.offset -= staging->i_payload;
+			item->vec.offset += off;
 		}
+
+		staging->i_payload = off;
 	}
 
-	ret = kdbus_msg_scan_items(m, conn->ep->bus);
+	/*
+	 * Step 4:
+	 * Allocate pool slice and copy over all data. Make sure to properly
+	 * account on user quota.
+	 */
+
+	ret = kdbus_conn_quota_inc(dst, src ? src->user : NULL, size,
+				   staging->gaps ? staging->gaps->n_fds : 0);
 	if (ret < 0)
-		goto exit_free;
+		goto error;
 
-	/* patch-in the source of this message */
-	if (m->msg.src_id > 0 && m->msg.src_id != conn->id) {
-		ret = -EINVAL;
-		goto exit_free;
+	slice = kdbus_pool_slice_alloc(dst->pool, size, true);
+	if (IS_ERR(slice)) {
+		ret = PTR_ERR(slice);
+		slice = NULL;
+		goto error;
 	}
-	m->msg.src_id = conn->id;
 
-	return m;
+	WARN_ON(kdbus_pool_slice_size(slice) != size);
 
-exit_free:
-	kdbus_kmsg_free(m);
-	return ERR_PTR(ret);
+	ret = kdbus_pool_slice_copy_iovec(slice, 0, staging->parts,
+					  staging->n_parts, size);
+	if (ret < 0)
+		goto error;
+
+	/* all done, return slice to caller */
+	goto exit;
+
+error:
+	if (slice)
+		kdbus_conn_quota_dec(dst, src ? src->user : NULL, size,
+				     staging->gaps ? staging->gaps->n_fds : 0);
+	kdbus_pool_slice_release(slice);
+	slice = ERR_PTR(ret);
+exit:
+	kfree(meta_items);
+	return slice;
 }
diff --git a/ipc/kdbus/message.h b/ipc/kdbus/message.h
index af4775850235..298f9c99dfcf 100644
--- a/ipc/kdbus/message.h
+++ b/ipc/kdbus/message.h
@@ -14,120 +14,107 @@
 #ifndef __KDBUS_MESSAGE_H
 #define __KDBUS_MESSAGE_H
 
-#include "util.h"
-#include "metadata.h"
+#include <linux/fs.h>
+#include <linux/kref.h>
+#include <uapi/linux/kdbus.h>
 
-/**
- * enum kdbus_msg_data_type - Type of kdbus_msg_data payloads
- * @KDBUS_MSG_DATA_VEC:		Data vector provided by user-space
- * @KDBUS_MSG_DATA_MEMFD:	Memfd payload
- */
-enum kdbus_msg_data_type {
-	KDBUS_MSG_DATA_VEC,
-	KDBUS_MSG_DATA_MEMFD,
-};
-
-/**
- * struct kdbus_msg_data - Data payload as stored by messages
- * @type:	Type of payload (KDBUS_MSG_DATA_*)
- * @size:	Size of the described payload
- * @off:	The offset, relative to the vec slice
- * @start:	Offset inside the memfd
- * @file:	Backing file referenced by the memfd
- */
-struct kdbus_msg_data {
-	unsigned int type;
-	u64 size;
-
-	union {
-		struct {
-			u64 off;
-		} vec;
-		struct {
-			u64 start;
-			struct file *file;
-		} memfd;
-	};
-};
+struct kdbus_bus;
+struct kdbus_conn;
+struct kdbus_meta_conn;
+struct kdbus_meta_proc;
+struct kdbus_pool_slice;
 
 /**
- * struct kdbus_kmsg_resources - resources of a message
+ * struct kdbus_gaps - gaps in message to be filled later
  * @kref:		Reference counter
- * @dst_name:		Short-cut to msg for faster lookup
- * @fds:		Array of file descriptors to pass
- * @fds_count:		Number of file descriptors to pass
- * @data:		Array of data payloads
- * @vec_count:		Number of VEC entries
- * @memfd_count:	Number of MEMFD entries in @data
- * @data_count:		Sum of @vec_count + @memfd_count
+ * @n_memfd_offs:	Number of memfds
+ * @memfd_offs:		Offsets of kdbus_memfd items in target slice
+ * @n_fds:		Number of fds
+ * @fds:		Array of sent fds
+ * @fds_offset:		Offset of fd-array in target slice
+ *
+ * The 'gaps' object is used to track data that is needed to fill gaps in a
+ * message at RECV time. Usually, we try to compile the whole message at SEND
+ * time. This has the advantage, that we don't have to cache any information and
+ * can keep the memory consumption small. Furthermore, all copy operations can
+ * be combined into a single function call, which speeds up transactions
+ * considerably.
+ * However, things like file-descriptors can only be fully installed at RECV
+ * time. The gaps object tracks this data and pins it until a message is
+ * received. The gaps object is shared between all receivers of the same
+ * message.
  */
-struct kdbus_msg_resources {
+struct kdbus_gaps {
 	struct kref kref;
-	const char *dst_name;
 
-	struct file **fds;
-	unsigned int fds_count;
+	/* state tracking for KDBUS_ITEM_PAYLOAD_MEMFD entries */
+	size_t n_memfds;
+	u64 *memfd_offsets;
+	struct file **memfd_files;
 
-	struct kdbus_msg_data *data;
-	size_t vec_count;
-	size_t memfd_count;
-	size_t data_count;
+	/* state tracking for KDBUS_ITEM_FDS */
+	size_t n_fds;
+	struct file **fd_files;
+	u64 fd_offset;
 };
 
-struct kdbus_msg_resources *
-kdbus_msg_resources_ref(struct kdbus_msg_resources *r);
-struct kdbus_msg_resources *
-kdbus_msg_resources_unref(struct kdbus_msg_resources *r);
+struct kdbus_gaps *kdbus_gaps_ref(struct kdbus_gaps *gaps);
+struct kdbus_gaps *kdbus_gaps_unref(struct kdbus_gaps *gaps);
+int kdbus_gaps_install(struct kdbus_gaps *gaps, struct kdbus_pool_slice *slice,
+		       bool *out_incomplete);
 
 /**
- * struct kdbus_kmsg - internal message handling data
- * @seq:		Domain-global message sequence number
- * @notify_type:	Short-cut for faster lookup
- * @notify_old_id:	Short-cut for faster lookup
- * @notify_new_id:	Short-cut for faster lookup
- * @notify_name:	Short-cut for faster lookup
- * @dst_name_id:	Short-cut to msg for faster lookup
- * @bloom_filter:	Bloom filter to match message properties
- * @bloom_generation:	Generation of bloom element set
- * @notify_entry:	List of kernel-generated notifications
- * @iov:		Array of iovec, describing the payload to copy
- * @iov_count:		Number of array members in @iov
- * @pool_size:		Overall size of inlined data referenced by @iov
- * @proc_meta:		Appended SCM-like metadata of the sending process
- * @conn_meta:		Appended SCM-like metadata of the sending connection
- * @res:		Message resources
- * @msg:		Message from or to userspace
+ * struct kdbus_staging - staging area to import messages
+ * @msg:		User-supplied message
+ * @gaps:		Gaps-object created during import (or NULL if empty)
+ * @msg_seqnum:		Message sequence number
+ * @notify_entry:	Entry into list of kernel-generated notifications
+ * @i_payload:		Current relative index of start of payload
+ * @n_payload:		Total number of bytes needed for payload
+ * @n_parts:		Number of parts
+ * @parts:		Array of iovecs that make up the whole message
+ * @meta_proc:		Process metadata of the sender (or NULL if empty)
+ * @meta_conn:		Connection metadata of the sender (or NULL if empty)
+ * @bloom_filter:	Pointer to the bloom-item in @msg, or NULL
+ * @dst_name:		Pointer to the dst-name-item in @msg, or NULL
+ * @notify:		Pointer to the notification item in @msg, or NULL
+ *
+ * The kdbus_staging object is a temporary staging area to import user-supplied
+ * messages into the kernel. It is only used during SEND and dropped once the
+ * message is queued. Any data that cannot be collected during SEND, is
+ * collected in a kdbus_gaps object and attached to the message queue.
  */
-struct kdbus_kmsg {
-	u64 seq;
-	u64 notify_type;
-	u64 notify_old_id;
-	u64 notify_new_id;
-	const char *notify_name;
-
-	u64 dst_name_id;
-	const struct kdbus_bloom_filter *bloom_filter;
-	u64 bloom_generation;
+struct kdbus_staging {
+	struct kdbus_msg *msg;
+	struct kdbus_gaps *gaps;
+	u64 msg_seqnum;
 	struct list_head notify_entry;
 
-	struct iovec *iov;
-	size_t iov_count;
-	u64 pool_size;
+	/* crafted iovecs to copy the message */
+	size_t i_payload;
+	size_t n_payload;
+	size_t n_parts;
+	struct iovec *parts;
 
-	struct kdbus_meta_proc *proc_meta;
-	struct kdbus_meta_conn *conn_meta;
-	struct kdbus_msg_resources *res;
+	/* metadata state */
+	struct kdbus_meta_proc *meta_proc;
+	struct kdbus_meta_conn *meta_conn;
 
-	/* variable size, must be the last member */
-	struct kdbus_msg msg;
+	/* cached pointers into @msg */
+	const struct kdbus_bloom_filter *bloom_filter;
+	const char *dst_name;
+	struct kdbus_item *notify;
 };
 
-struct kdbus_bus;
-struct kdbus_conn;
-
-struct kdbus_kmsg *kdbus_kmsg_new(struct kdbus_bus *bus, size_t extra_size);
-struct kdbus_kmsg *kdbus_kmsg_new_from_cmd(struct kdbus_conn *conn,
-					   struct kdbus_cmd_send *cmd_send);
-void kdbus_kmsg_free(struct kdbus_kmsg *kmsg);
+struct kdbus_staging *kdbus_staging_new_kernel(struct kdbus_bus *bus,
+					       u64 dst, u64 cookie_timeout,
+					       size_t it_size, size_t it_type);
+struct kdbus_staging *kdbus_staging_new_user(struct kdbus_bus *bus,
+					     struct kdbus_cmd_send *cmd,
+					     struct kdbus_msg *msg);
+struct kdbus_staging *kdbus_staging_free(struct kdbus_staging *staging);
+struct kdbus_pool_slice *kdbus_staging_emit(struct kdbus_staging *staging,
+					    struct kdbus_conn *src,
+					    struct kdbus_conn *dst);
 
 #endif
diff --git a/ipc/kdbus/metadata.c b/ipc/kdbus/metadata.c
index 501bebda7d84..71ca475a80d5 100644
--- a/ipc/kdbus/metadata.c
+++ b/ipc/kdbus/metadata.c
@@ -29,7 +29,6 @@
 #include <linux/uidgid.h>
 #include <linux/uio.h>
 #include <linux/user_namespace.h>
-#include <linux/version.h>
 
 #include "bus.h"
 #include "connection.h"
@@ -45,27 +44,16 @@
  * @lock:		Object lock
  * @collected:		Bitmask of collected items
  * @valid:		Bitmask of collected and valid items
- * @uid:		UID of process
- * @euid:		EUID of process
- * @suid:		SUID of process
- * @fsuid:		FSUID of process
- * @gid:		GID of process
- * @egid:		EGID of process
- * @sgid:		SGID of process
- * @fsgid:		FSGID of process
+ * @cred:		Credentials
  * @pid:		PID of process
  * @tgid:		TGID of process
  * @ppid:		PPID of process
- * @auxgrps:		Auxiliary groups
- * @n_auxgrps:		Number of items in @auxgrps
  * @tid_comm:		TID comm line
  * @pid_comm:		PID comm line
  * @exe_path:		Executable path
  * @root_path:		Root-FS path
  * @cmdline:		Command-line
  * @cgroup:		Full cgroup path
- * @caps:		Capabilities
- * @caps_namespace:	User-namespace of @caps
  * @seclabel:		Seclabel
  * @audit_loginuid:	Audit login-UID
  * @audit_sessionid:	Audit session-ID
@@ -77,18 +65,15 @@ struct kdbus_meta_proc {
 	u64 valid;
 
 	/* KDBUS_ITEM_CREDS */
-	kuid_t uid, euid, suid, fsuid;
-	kgid_t gid, egid, sgid, fsgid;
+	/* KDBUS_ITEM_AUXGROUPS */
+	/* KDBUS_ITEM_CAPS */
+	const struct cred *cred;
 
 	/* KDBUS_ITEM_PIDS */
 	struct pid *pid;
 	struct pid *tgid;
 	struct pid *ppid;
 
-	/* KDBUS_ITEM_AUXGROUPS */
-	kgid_t *auxgrps;
-	size_t n_auxgrps;
-
 	/* KDBUS_ITEM_TID_COMM */
 	char tid_comm[TASK_COMM_LEN];
 	/* KDBUS_ITEM_PID_COMM */
@@ -104,16 +89,6 @@ struct kdbus_meta_proc {
 	/* KDBUS_ITEM_CGROUP */
 	char *cgroup;
 
-	/* KDBUS_ITEM_CAPS */
-	struct caps {
-		/* binary compatible to kdbus_caps */
-		u32 last_cap;
-		struct {
-			u32 caps[_KERNEL_CAPABILITY_U32S];
-		} set[4];
-	} caps;
-	struct user_namespace *caps_namespace;
-
 	/* KDBUS_ITEM_SECLABEL */
 	char *seclabel;
 
@@ -150,6 +125,14 @@ struct kdbus_meta_conn {
 	char *conn_description;
 };
 
+/* fixed size equivalent of "kdbus_caps" */
+struct kdbus_meta_caps {
+	u32 last_cap;
+	struct {
+		u32 caps[_KERNEL_CAPABILITY_U32S];
+	} set[4];
+};
+
 /**
  * kdbus_meta_proc_new() - Create process metadata object
  *
@@ -176,13 +159,13 @@ static void kdbus_meta_proc_free(struct kref *kref)
 
 	path_put(&mp->exe_path);
 	path_put(&mp->root_path);
-	put_user_ns(mp->caps_namespace);
+	if (mp->cred)
+		put_cred(mp->cred);
 	put_pid(mp->ppid);
 	put_pid(mp->tgid);
 	put_pid(mp->pid);
 
 	kfree(mp->seclabel);
-	kfree(mp->auxgrps);
 	kfree(mp->cmdline);
 	kfree(mp->cgroup);
 	kfree(mp);
@@ -214,21 +197,6 @@ struct kdbus_meta_proc *kdbus_meta_proc_unref(struct kdbus_meta_proc *mp)
 	return NULL;
 }
 
-static void kdbus_meta_proc_collect_creds(struct kdbus_meta_proc *mp)
-{
-	mp->uid		= current_uid();
-	mp->euid	= current_euid();
-	mp->suid	= current_suid();
-	mp->fsuid	= current_fsuid();
-
-	mp->gid		= current_gid();
-	mp->egid	= current_egid();
-	mp->sgid	= current_sgid();
-	mp->fsgid	= current_fsgid();
-
-	mp->valid |= KDBUS_ATTACH_CREDS;
-}
-
 static void kdbus_meta_proc_collect_pids(struct kdbus_meta_proc *mp)
 {
 	struct task_struct *parent;
@@ -244,32 +212,6 @@ static void kdbus_meta_proc_collect_pids(struct kdbus_meta_proc *mp)
 	mp->valid |= KDBUS_ATTACH_PIDS;
 }
 
-static int kdbus_meta_proc_collect_auxgroups(struct kdbus_meta_proc *mp)
-{
-	struct group_info *info;
-	size_t i;
-
-	info = get_current_groups();
-
-	if (info->ngroups > 0) {
-		mp->auxgrps = kmalloc_array(info->ngroups, sizeof(kgid_t),
-					    GFP_KERNEL);
-		if (!mp->auxgrps) {
-			put_group_info(info);
-			return -ENOMEM;
-		}
-
-		for (i = 0; i < info->ngroups; i++)
-			mp->auxgrps[i] = GROUP_AT(info, i);
-	}
-
-	mp->n_auxgrps = info->ngroups;
-	put_group_info(info);
-	mp->valid |= KDBUS_ATTACH_AUXGROUPS;
-
-	return 0;
-}
-
 static void kdbus_meta_proc_collect_tid_comm(struct kdbus_meta_proc *mp)
 {
 	get_task_comm(mp->tid_comm, current);
@@ -284,42 +226,29 @@ static void kdbus_meta_proc_collect_pid_comm(struct kdbus_meta_proc *mp)
 
 static void kdbus_meta_proc_collect_exe(struct kdbus_meta_proc *mp)
 {
-	struct mm_struct *mm;
-
-	mm = get_task_mm(current);
-	if (!mm)
-		return;
+	struct file *exe_file;
 
-	down_read(&mm->mmap_sem);
-	if (mm->exe_file) {
-		mp->exe_path = mm->exe_file->f_path;
+	rcu_read_lock();
+	exe_file = rcu_dereference(current->mm->exe_file);
+	if (exe_file) {
+		mp->exe_path = exe_file->f_path;
 		path_get(&mp->exe_path);
 		get_fs_root(current->fs, &mp->root_path);
 		mp->valid |= KDBUS_ATTACH_EXE;
 	}
-	up_read(&mm->mmap_sem);
-
-	mmput(mm);
+	rcu_read_unlock();
 }
 
 static int kdbus_meta_proc_collect_cmdline(struct kdbus_meta_proc *mp)
 {
-	struct mm_struct *mm;
+	struct mm_struct *mm = current->mm;
 	char *cmdline;
 
-	mm = get_task_mm(current);
-	if (!mm)
-		return 0;
-
-	if (mm->arg_start >= mm->arg_end) {
-		mmput(mm);
+	if (!mm->arg_end)
 		return 0;
-	}
 
 	cmdline = strndup_user((const char __user *)mm->arg_start,
 			       mm->arg_end - mm->arg_start);
-	mmput(mm);
-
 	if (IS_ERR(cmdline))
 		return PTR_ERR(cmdline);
 
@@ -355,30 +284,6 @@ static int kdbus_meta_proc_collect_cgroup(struct kdbus_meta_proc *mp)
 	return 0;
 }
 
-static void kdbus_meta_proc_collect_caps(struct kdbus_meta_proc *mp)
-{
-	const struct cred *c = current_cred();
-	int i;
-
-	/* ABI: "last_cap" equals /proc/sys/kernel/cap_last_cap */
-	mp->caps.last_cap = CAP_LAST_CAP;
-	mp->caps_namespace = get_user_ns(current_user_ns());
-
-	CAP_FOR_EACH_U32(i) {
-		mp->caps.set[0].caps[i] = c->cap_inheritable.cap[i];
-		mp->caps.set[1].caps[i] = c->cap_permitted.cap[i];
-		mp->caps.set[2].caps[i] = c->cap_effective.cap[i];
-		mp->caps.set[3].caps[i] = c->cap_bset.cap[i];
-	}
-
-	/* clear unused bits */
-	for (i = 0; i < 4; i++)
-		mp->caps.set[i].caps[CAP_TO_INDEX(CAP_LAST_CAP)] &=
-						CAP_LAST_U32_VALID_MASK;
-
-	mp->valid |= KDBUS_ATTACH_CAPS;
-}
-
 static int kdbus_meta_proc_collect_seclabel(struct kdbus_meta_proc *mp)
 {
 #ifdef CONFIG_SECURITY
@@ -445,10 +350,17 @@ int kdbus_meta_proc_collect(struct kdbus_meta_proc *mp, u64 what)
 
 	mutex_lock(&mp->lock);
 
-	if ((what & KDBUS_ATTACH_CREDS) &&
-	    !(mp->collected & KDBUS_ATTACH_CREDS)) {
-		kdbus_meta_proc_collect_creds(mp);
-		mp->collected |= KDBUS_ATTACH_CREDS;
+	/* creds, auxgrps and caps share "struct cred" as context */
+	{
+		const u64 m_cred = KDBUS_ATTACH_CREDS |
+				   KDBUS_ATTACH_AUXGROUPS |
+				   KDBUS_ATTACH_CAPS;
+
+		if ((what & m_cred) && !(mp->collected & m_cred)) {
+			mp->cred = get_current_cred();
+			mp->valid |= m_cred;
+			mp->collected |= m_cred;
+		}
 	}
 
 	if ((what & KDBUS_ATTACH_PIDS) &&
@@ -457,14 +369,6 @@ int kdbus_meta_proc_collect(struct kdbus_meta_proc *mp, u64 what)
 		mp->collected |= KDBUS_ATTACH_PIDS;
 	}
 
-	if ((what & KDBUS_ATTACH_AUXGROUPS) &&
-	    !(mp->collected & KDBUS_ATTACH_AUXGROUPS)) {
-		ret = kdbus_meta_proc_collect_auxgroups(mp);
-		if (ret < 0)
-			goto exit_unlock;
-		mp->collected |= KDBUS_ATTACH_AUXGROUPS;
-	}
-
 	if ((what & KDBUS_ATTACH_TID_COMM) &&
 	    !(mp->collected & KDBUS_ATTACH_TID_COMM)) {
 		kdbus_meta_proc_collect_tid_comm(mp);
@@ -499,12 +403,6 @@ int kdbus_meta_proc_collect(struct kdbus_meta_proc *mp, u64 what)
 		mp->collected |= KDBUS_ATTACH_CGROUP;
 	}
 
-	if ((what & KDBUS_ATTACH_CAPS) &&
-	    !(mp->collected & KDBUS_ATTACH_CAPS)) {
-		kdbus_meta_proc_collect_caps(mp);
-		mp->collected |= KDBUS_ATTACH_CAPS;
-	}
-
 	if ((what & KDBUS_ATTACH_SECLABEL) &&
 	    !(mp->collected & KDBUS_ATTACH_SECLABEL)) {
 		ret = kdbus_meta_proc_collect_seclabel(mp);
@@ -527,101 +425,116 @@ exit_unlock:
 }
 
 /**
- * kdbus_meta_proc_fake() - Fill process metadata from faked credentials
- * @mp:		Metadata
+ * kdbus_meta_fake_new() - Create fake metadata object
+ *
+ * Return: Pointer to new object on success, ERR_PTR on failure.
+ */
+struct kdbus_meta_fake *kdbus_meta_fake_new(void)
+{
+	struct kdbus_meta_fake *mf;
+
+	mf = kzalloc(sizeof(*mf), GFP_KERNEL);
+	if (!mf)
+		return ERR_PTR(-ENOMEM);
+
+	return mf;
+}
+
+/**
+ * kdbus_meta_fake_free() - Free fake metadata object
+ * @mf:		Fake metadata object
+ *
+ * Return: NULL
+ */
+struct kdbus_meta_fake *kdbus_meta_fake_free(struct kdbus_meta_fake *mf)
+{
+	if (mf) {
+		put_pid(mf->ppid);
+		put_pid(mf->tgid);
+		put_pid(mf->pid);
+		kfree(mf->seclabel);
+		kfree(mf);
+	}
+
+	return NULL;
+}
+
+/**
+ * kdbus_meta_fake_collect() - Fill fake metadata from faked credentials
+ * @mf:		Fake metadata object
  * @creds:	Creds to set, may be %NULL
  * @pids:	PIDs to set, may be %NULL
  * @seclabel:	Seclabel to set, may be %NULL
  *
  * This function takes information stored in @creds, @pids and @seclabel and
- * resolves them to kernel-representations, if possible. A call to this function
- * is considered an alternative to calling kdbus_meta_add_current(), which
- * derives the same information from the 'current' task.
- *
- * This call uses the current task's namespaces to resolve the given
- * information.
+ * resolves them to kernel-representations, if possible. This call uses the
+ * current task's namespaces to resolve the given information.
  *
- * Return: 0 on success, negative error number otherwise.
+ * Return: 0 on success, negative error code on failure.
  */
-int kdbus_meta_proc_fake(struct kdbus_meta_proc *mp,
-			 const struct kdbus_creds *creds,
-			 const struct kdbus_pids *pids,
-			 const char *seclabel)
+int kdbus_meta_fake_collect(struct kdbus_meta_fake *mf,
+			    const struct kdbus_creds *creds,
+			    const struct kdbus_pids *pids,
+			    const char *seclabel)
 {
-	int ret;
+	if (mf->valid)
+		return -EALREADY;
 
-	if (!mp)
-		return 0;
-
-	mutex_lock(&mp->lock);
-
-	if (creds && !(mp->collected & KDBUS_ATTACH_CREDS)) {
+	if (creds) {
 		struct user_namespace *ns = current_user_ns();
 
-		mp->uid		= make_kuid(ns, creds->uid);
-		mp->euid	= make_kuid(ns, creds->euid);
-		mp->suid	= make_kuid(ns, creds->suid);
-		mp->fsuid	= make_kuid(ns, creds->fsuid);
-
-		mp->gid		= make_kgid(ns, creds->gid);
-		mp->egid	= make_kgid(ns, creds->egid);
-		mp->sgid	= make_kgid(ns, creds->sgid);
-		mp->fsgid	= make_kgid(ns, creds->fsgid);
-
-		if ((creds->uid   != (uid_t)-1 && !uid_valid(mp->uid))   ||
-		    (creds->euid  != (uid_t)-1 && !uid_valid(mp->euid))  ||
-		    (creds->suid  != (uid_t)-1 && !uid_valid(mp->suid))  ||
-		    (creds->fsuid != (uid_t)-1 && !uid_valid(mp->fsuid)) ||
-		    (creds->gid   != (gid_t)-1 && !gid_valid(mp->gid))   ||
-		    (creds->egid  != (gid_t)-1 && !gid_valid(mp->egid))  ||
-		    (creds->sgid  != (gid_t)-1 && !gid_valid(mp->sgid))  ||
-		    (creds->fsgid != (gid_t)-1 && !gid_valid(mp->fsgid))) {
-			ret = -EINVAL;
-			goto exit_unlock;
-		}
-
-		mp->valid |= KDBUS_ATTACH_CREDS;
-		mp->collected |= KDBUS_ATTACH_CREDS;
+		mf->uid		= make_kuid(ns, creds->uid);
+		mf->euid	= make_kuid(ns, creds->euid);
+		mf->suid	= make_kuid(ns, creds->suid);
+		mf->fsuid	= make_kuid(ns, creds->fsuid);
+
+		mf->gid		= make_kgid(ns, creds->gid);
+		mf->egid	= make_kgid(ns, creds->egid);
+		mf->sgid	= make_kgid(ns, creds->sgid);
+		mf->fsgid	= make_kgid(ns, creds->fsgid);
+
+		if ((creds->uid   != (uid_t)-1 && !uid_valid(mf->uid))   ||
+		    (creds->euid  != (uid_t)-1 && !uid_valid(mf->euid))  ||
+		    (creds->suid  != (uid_t)-1 && !uid_valid(mf->suid))  ||
+		    (creds->fsuid != (uid_t)-1 && !uid_valid(mf->fsuid)) ||
+		    (creds->gid   != (gid_t)-1 && !gid_valid(mf->gid))   ||
+		    (creds->egid  != (gid_t)-1 && !gid_valid(mf->egid))  ||
+		    (creds->sgid  != (gid_t)-1 && !gid_valid(mf->sgid))  ||
+		    (creds->fsgid != (gid_t)-1 && !gid_valid(mf->fsgid)))
+			return -EINVAL;
+
+		mf->valid |= KDBUS_ATTACH_CREDS;
 	}
 
-	if (pids && !(mp->collected & KDBUS_ATTACH_PIDS)) {
-		mp->pid = get_pid(find_vpid(pids->tid));
-		mp->tgid = get_pid(find_vpid(pids->pid));
-		mp->ppid = get_pid(find_vpid(pids->ppid));
-
-		if ((pids->tid != 0 && !mp->pid) ||
-		    (pids->pid != 0 && !mp->tgid) ||
-		    (pids->ppid != 0 && !mp->ppid)) {
-			put_pid(mp->pid);
-			put_pid(mp->tgid);
-			put_pid(mp->ppid);
-			mp->pid = NULL;
-			mp->tgid = NULL;
-			mp->ppid = NULL;
-			ret = -EINVAL;
-			goto exit_unlock;
+	if (pids) {
+		mf->pid = get_pid(find_vpid(pids->tid));
+		mf->tgid = get_pid(find_vpid(pids->pid));
+		mf->ppid = get_pid(find_vpid(pids->ppid));
+
+		if ((pids->tid != 0 && !mf->pid) ||
+		    (pids->pid != 0 && !mf->tgid) ||
+		    (pids->ppid != 0 && !mf->ppid)) {
+			put_pid(mf->pid);
+			put_pid(mf->tgid);
+			put_pid(mf->ppid);
+			mf->pid = NULL;
+			mf->tgid = NULL;
+			mf->ppid = NULL;
+			return -EINVAL;
 		}
 
-		mp->valid |= KDBUS_ATTACH_PIDS;
-		mp->collected |= KDBUS_ATTACH_PIDS;
+		mf->valid |= KDBUS_ATTACH_PIDS;
 	}
 
-	if (seclabel && !(mp->collected & KDBUS_ATTACH_SECLABEL)) {
-		mp->seclabel = kstrdup(seclabel, GFP_KERNEL);
-		if (!mp->seclabel) {
-			ret = -ENOMEM;
-			goto exit_unlock;
-		}
+	if (seclabel) {
+		mf->seclabel = kstrdup(seclabel, GFP_KERNEL);
+		if (!mf->seclabel)
+			return -ENOMEM;
 
-		mp->valid |= KDBUS_ATTACH_SECLABEL;
-		mp->collected |= KDBUS_ATTACH_SECLABEL;
+		mf->valid |= KDBUS_ATTACH_SECLABEL;
 	}
 
-	ret = 0;
-
-exit_unlock:
-	mutex_unlock(&mp->lock);
-	return ret;
+	return 0;
 }
 
 /**
@@ -676,13 +589,13 @@ struct kdbus_meta_conn *kdbus_meta_conn_unref(struct kdbus_meta_conn *mc)
 }
 
 static void kdbus_meta_conn_collect_timestamp(struct kdbus_meta_conn *mc,
-					      struct kdbus_kmsg *kmsg)
+					      u64 msg_seqnum)
 {
 	mc->ts.monotonic_ns = ktime_get_ns();
 	mc->ts.realtime_ns = ktime_get_real_ns();
 
-	if (kmsg)
-		mc->ts.seqnum = kmsg->seq;
+	if (msg_seqnum)
+		mc->ts.seqnum = msg_seqnum;
 
 	mc->valid |= KDBUS_ATTACH_TIMESTAMP;
 }
@@ -690,38 +603,46 @@ static void kdbus_meta_conn_collect_timestamp(struct kdbus_meta_conn *mc,
 static int kdbus_meta_conn_collect_names(struct kdbus_meta_conn *mc,
 					 struct kdbus_conn *conn)
 {
-	const struct kdbus_name_entry *e;
+	const struct kdbus_name_owner *owner;
 	struct kdbus_item *item;
 	size_t slen, size;
 
 	lockdep_assert_held(&conn->ep->bus->name_registry->rwlock);
 
 	size = 0;
-	list_for_each_entry(e, &conn->names_list, conn_entry)
-		size += KDBUS_ITEM_SIZE(sizeof(struct kdbus_name) +
-					strlen(e->name) + 1);
+	/* open-code length calculation to avoid final padding */
+	list_for_each_entry(owner, &conn->names_list, conn_entry)
+		if (!(owner->flags & KDBUS_NAME_IN_QUEUE))
+			size = KDBUS_ALIGN8(size) + KDBUS_ITEM_HEADER_SIZE +
+				sizeof(struct kdbus_name) +
+				strlen(owner->name->name) + 1;
 
 	if (!size)
 		return 0;
 
-	item = kmalloc(size, GFP_KERNEL);
+	/* make sure we include zeroed padding for convenience helpers */
+	item = kmalloc(KDBUS_ALIGN8(size), GFP_KERNEL);
 	if (!item)
 		return -ENOMEM;
 
 	mc->owned_names_items = item;
 	mc->owned_names_size = size;
 
-	list_for_each_entry(e, &conn->names_list, conn_entry) {
-		slen = strlen(e->name) + 1;
+	list_for_each_entry(owner, &conn->names_list, conn_entry) {
+		if (owner->flags & KDBUS_NAME_IN_QUEUE)
+			continue;
+
+		slen = strlen(owner->name->name) + 1;
 		kdbus_item_set(item, KDBUS_ITEM_OWNED_NAME, NULL,
 			       sizeof(struct kdbus_name) + slen);
-		item->name.flags = e->flags;
-		memcpy(item->name.name, e->name, slen);
+		item->name.flags = owner->flags;
+		memcpy(item->name.name, owner->name->name, slen);
 		item = KDBUS_ITEM_NEXT(item);
 	}
 
 	/* sanity check: the buffer should be completely written now */
-	WARN_ON((u8 *)item != (u8 *)mc->owned_names_items + size);
+	WARN_ON((u8 *)item !=
+			(u8 *)mc->owned_names_items + KDBUS_ALIGN8(size));
 
 	mc->valid |= KDBUS_ATTACH_NAMES;
 	return 0;
@@ -744,11 +665,12 @@ static int kdbus_meta_conn_collect_description(struct kdbus_meta_conn *mc,
 /**
  * kdbus_meta_conn_collect() - Collect connection metadata
  * @mc:		Message metadata object
- * @kmsg:	Kmsg to collect data from
  * @conn:	Connection to collect data from
+ * @msg_seqnum:	Sequence number of the message to send
  * @what:	Attach flags to collect
  *
- * This collects connection metadata from @kmsg and @conn and saves it in @mc.
+ * This collects connection metadata from @msg_seqnum and @conn and saves it
+ * in @mc.
  *
  * If KDBUS_ATTACH_NAMES is set in @what and @conn is non-NULL, the caller must
  * hold the name-registry read-lock of conn->ep->bus->registry.
@@ -756,9 +678,8 @@ static int kdbus_meta_conn_collect_description(struct kdbus_meta_conn *mc,
  * Return: 0 on success, negative error code on failure.
  */
 int kdbus_meta_conn_collect(struct kdbus_meta_conn *mc,
-			    struct kdbus_kmsg *kmsg,
 			    struct kdbus_conn *conn,
-			    u64 what)
+			    u64 msg_seqnum, u64 what)
 {
 	int ret;
 
@@ -769,9 +690,9 @@ int kdbus_meta_conn_collect(struct kdbus_meta_conn *mc,
 
 	mutex_lock(&mc->lock);
 
-	if (kmsg && (what & KDBUS_ATTACH_TIMESTAMP) &&
+	if (msg_seqnum && (what & KDBUS_ATTACH_TIMESTAMP) &&
 	    !(mc->collected & KDBUS_ATTACH_TIMESTAMP)) {
-		kdbus_meta_conn_collect_timestamp(mc, kmsg);
+		kdbus_meta_conn_collect_timestamp(mc, msg_seqnum);
 		mc->collected |= KDBUS_ATTACH_TIMESTAMP;
 	}
 
@@ -798,258 +719,392 @@ exit_unlock:
 	return ret;
 }
 
-/*
- * kdbus_meta_export_prepare() - Prepare metadata for export
- * @mp:		Process metadata, or NULL
- * @mc:		Connection metadata, or NULL
- * @mask:	Pointer to mask of KDBUS_ATTACH_* flags to export
- * @sz:		Pointer to return the size needed by the metadata
- *
- * Does a conservative calculation of how much space metadata information
- * will take up during export. It is 'conservative' because for string
- * translations in namespaces, it will use the kernel namespaces, which is
- * the longest possible version.
- *
- * The actual size consumed by kdbus_meta_export() may hence vary from the
- * one reported here, but it is guaranteed never to be greater.
- *
- * Return: 0 on success, negative error number otherwise.
- */
-int kdbus_meta_export_prepare(struct kdbus_meta_proc *mp,
-			      struct kdbus_meta_conn *mc,
-			      u64 *mask, size_t *sz)
+static void kdbus_meta_export_caps(struct kdbus_meta_caps *out,
+				   const struct kdbus_meta_proc *mp,
+				   struct user_namespace *user_ns)
 {
-	char *exe_pathname = NULL;
-	void *exe_page = NULL;
-	size_t size = 0;
-	u64 valid = 0;
-	int ret = 0;
+	struct user_namespace *iter;
+	const struct cred *cred = mp->cred;
+	bool parent = false, owner = false;
+	int i;
 
-	if (mp) {
-		mutex_lock(&mp->lock);
-		valid |= mp->valid;
-		mutex_unlock(&mp->lock);
+	/*
+	 * This translates the effective capabilities of 'cred' into the given
+	 * user-namespace. If the given user-namespace is a child-namespace of
+	 * the user-namespace of 'cred', the mask can be copied verbatim. If
+	 * not, the mask is cleared.
+	 * There's one exception: If 'cred' is the owner of any user-namespace
+	 * in the path between the given user-namespace and the user-namespace
+	 * of 'cred', then it has all effective capabilities set. This means,
+	 * the user who created a user-namespace always has all effective
+	 * capabilities in any child namespaces. Note that this is based on the
+	 * uid of the namespace creator, not the task hierarchy.
+	 */
+	for (iter = user_ns; iter; iter = iter->parent) {
+		if (iter == cred->user_ns) {
+			parent = true;
+			break;
+		}
+
+		if (iter == &init_user_ns)
+			break;
+
+		if ((iter->parent == cred->user_ns) &&
+		    uid_eq(iter->owner, cred->euid)) {
+			owner = true;
+			break;
+		}
 	}
 
-	if (mc) {
-		mutex_lock(&mc->lock);
-		valid |= mc->valid;
-		mutex_unlock(&mc->lock);
+	out->last_cap = CAP_LAST_CAP;
+
+	CAP_FOR_EACH_U32(i) {
+		if (parent) {
+			out->set[0].caps[i] = cred->cap_inheritable.cap[i];
+			out->set[1].caps[i] = cred->cap_permitted.cap[i];
+			out->set[2].caps[i] = cred->cap_effective.cap[i];
+			out->set[3].caps[i] = cred->cap_bset.cap[i];
+		} else if (owner) {
+			out->set[0].caps[i] = 0U;
+			out->set[1].caps[i] = ~0U;
+			out->set[2].caps[i] = ~0U;
+			out->set[3].caps[i] = ~0U;
+		} else {
+			out->set[0].caps[i] = 0U;
+			out->set[1].caps[i] = 0U;
+			out->set[2].caps[i] = 0U;
+			out->set[3].caps[i] = 0U;
+		}
 	}
 
-	*mask &= valid;
-	*mask &= kdbus_meta_attach_mask;
+	/* clear unused bits */
+	for (i = 0; i < 4; i++)
+		out->set[i].caps[CAP_TO_INDEX(CAP_LAST_CAP)] &=
+					CAP_LAST_U32_VALID_MASK;
+}
 
-	if (!*mask)
-		goto exit;
+/* This is equivalent to from_kuid_munged(), but maps INVALID_UID to itself */
+static uid_t kdbus_from_kuid_keep(struct user_namespace *ns, kuid_t uid)
+{
+	return uid_valid(uid) ? from_kuid_munged(ns, uid) : ((uid_t)-1);
+}
+
+/* This is equivalent to from_kgid_munged(), but maps INVALID_GID to itself */
+static gid_t kdbus_from_kgid_keep(struct user_namespace *ns, kgid_t gid)
+{
+	return gid_valid(gid) ? from_kgid_munged(ns, gid) : ((gid_t)-1);
+}
+
+struct kdbus_meta_staging {
+	const struct kdbus_meta_proc *mp;
+	const struct kdbus_meta_fake *mf;
+	const struct kdbus_meta_conn *mc;
+	const struct kdbus_conn *conn;
+	u64 mask;
+
+	void *exe;
+	const char *exe_path;
+};
+
+static size_t kdbus_meta_measure(struct kdbus_meta_staging *staging)
+{
+	const struct kdbus_meta_proc *mp = staging->mp;
+	const struct kdbus_meta_fake *mf = staging->mf;
+	const struct kdbus_meta_conn *mc = staging->mc;
+	const u64 mask = staging->mask;
+	size_t size = 0;
 
 	/* process metadata */
 
-	if (mp && (*mask & KDBUS_ATTACH_CREDS))
+	if (mf && (mask & KDBUS_ATTACH_CREDS))
+		size += KDBUS_ITEM_SIZE(sizeof(struct kdbus_creds));
+	else if (mp && (mask & KDBUS_ATTACH_CREDS))
 		size += KDBUS_ITEM_SIZE(sizeof(struct kdbus_creds));
 
-	if (mp && (*mask & KDBUS_ATTACH_PIDS))
+	if (mf && (mask & KDBUS_ATTACH_PIDS))
+		size += KDBUS_ITEM_SIZE(sizeof(struct kdbus_pids));
+	else if (mp && (mask & KDBUS_ATTACH_PIDS))
 		size += KDBUS_ITEM_SIZE(sizeof(struct kdbus_pids));
 
-	if (mp && (*mask & KDBUS_ATTACH_AUXGROUPS))
-		size += KDBUS_ITEM_SIZE(mp->n_auxgrps * sizeof(u64));
+	if (mp && (mask & KDBUS_ATTACH_AUXGROUPS))
+		size += KDBUS_ITEM_SIZE(mp->cred->group_info->ngroups *
+					sizeof(u64));
 
-	if (mp && (*mask & KDBUS_ATTACH_TID_COMM))
+	if (mp && (mask & KDBUS_ATTACH_TID_COMM))
 		size += KDBUS_ITEM_SIZE(strlen(mp->tid_comm) + 1);
 
-	if (mp && (*mask & KDBUS_ATTACH_PID_COMM))
+	if (mp && (mask & KDBUS_ATTACH_PID_COMM))
 		size += KDBUS_ITEM_SIZE(strlen(mp->pid_comm) + 1);
 
-	if (mp && (*mask & KDBUS_ATTACH_EXE)) {
-		exe_page = (void *)__get_free_page(GFP_TEMPORARY);
-		if (!exe_page) {
-			ret = -ENOMEM;
-			goto exit;
-		}
-
-		exe_pathname = d_path(&mp->exe_path, exe_page, PAGE_SIZE);
-		if (IS_ERR(exe_pathname)) {
-			ret = PTR_ERR(exe_pathname);
-			goto exit;
-		}
+	if (staging->exe_path && (mask & KDBUS_ATTACH_EXE))
+		size += KDBUS_ITEM_SIZE(strlen(staging->exe_path) + 1);
 
-		size += KDBUS_ITEM_SIZE(strlen(exe_pathname) + 1);
-		free_page((unsigned long)exe_page);
-	}
-
-	if (mp && (*mask & KDBUS_ATTACH_CMDLINE))
+	if (mp && (mask & KDBUS_ATTACH_CMDLINE))
 		size += KDBUS_ITEM_SIZE(strlen(mp->cmdline) + 1);
 
-	if (mp && (*mask & KDBUS_ATTACH_CGROUP))
+	if (mp && (mask & KDBUS_ATTACH_CGROUP))
 		size += KDBUS_ITEM_SIZE(strlen(mp->cgroup) + 1);
 
-	if (mp && (*mask & KDBUS_ATTACH_CAPS))
-		size += KDBUS_ITEM_SIZE(sizeof(mp->caps));
+	if (mp && (mask & KDBUS_ATTACH_CAPS))
+		size += KDBUS_ITEM_SIZE(sizeof(struct kdbus_meta_caps));
 
-	if (mp && (*mask & KDBUS_ATTACH_SECLABEL))
+	if (mf && (mask & KDBUS_ATTACH_SECLABEL))
+		size += KDBUS_ITEM_SIZE(strlen(mf->seclabel) + 1);
+	else if (mp && (mask & KDBUS_ATTACH_SECLABEL))
 		size += KDBUS_ITEM_SIZE(strlen(mp->seclabel) + 1);
 
-	if (mp && (*mask & KDBUS_ATTACH_AUDIT))
+	if (mp && (mask & KDBUS_ATTACH_AUDIT))
 		size += KDBUS_ITEM_SIZE(sizeof(struct kdbus_audit));
 
 	/* connection metadata */
 
-	if (mc && (*mask & KDBUS_ATTACH_NAMES))
-		size += mc->owned_names_size;
+	if (mc && (mask & KDBUS_ATTACH_NAMES))
+		size += KDBUS_ALIGN8(mc->owned_names_size);
 
-	if (mc && (*mask & KDBUS_ATTACH_CONN_DESCRIPTION))
+	if (mc && (mask & KDBUS_ATTACH_CONN_DESCRIPTION))
 		size += KDBUS_ITEM_SIZE(strlen(mc->conn_description) + 1);
 
-	if (mc && (*mask & KDBUS_ATTACH_TIMESTAMP))
+	if (mc && (mask & KDBUS_ATTACH_TIMESTAMP))
 		size += KDBUS_ITEM_SIZE(sizeof(struct kdbus_timestamp));
 
-exit:
-	*sz = size;
-
-	return ret;
+	return size;
 }
 
-static int kdbus_meta_push_kvec(struct kvec *kvec,
-				struct kdbus_item_header *hdr,
-				u64 type, void *payload,
-				size_t payload_size, u64 *size)
+static struct kdbus_item *kdbus_write_head(struct kdbus_item **iter,
+					   u64 type, u64 size)
 {
-	hdr->type = type;
-	hdr->size = KDBUS_ITEM_HEADER_SIZE + payload_size;
-	kdbus_kvec_set(kvec++, hdr, sizeof(*hdr), size);
-	kdbus_kvec_set(kvec++, payload, payload_size, size);
-	return 2 + !!kdbus_kvec_pad(kvec++, size);
-}
+	struct kdbus_item *item = *iter;
+	size_t padding;
 
-/* This is equivalent to from_kuid_munged(), but maps INVALID_UID to itself */
-static uid_t kdbus_from_kuid_keep(kuid_t uid)
-{
-	return uid_valid(uid) ?
-		from_kuid_munged(current_user_ns(), uid) : ((uid_t)-1);
+	item->type = type;
+	item->size = KDBUS_ITEM_HEADER_SIZE + size;
+
+	/* clear padding */
+	padding = KDBUS_ALIGN8(item->size) - item->size;
+	if (padding)
+		memset(item->data + size, 0, padding);
+
+	*iter = KDBUS_ITEM_NEXT(item);
+	return item;
 }
 
-/* This is equivalent to from_kgid_munged(), but maps INVALID_GID to itself */
-static gid_t kdbus_from_kgid_keep(kgid_t gid)
+static struct kdbus_item *kdbus_write_full(struct kdbus_item **iter,
+					   u64 type, u64 size, const void *data)
 {
-	return gid_valid(gid) ?
-		from_kgid_munged(current_user_ns(), gid) : ((gid_t)-1);
+	struct kdbus_item *item;
+
+	item = kdbus_write_head(iter, type, size);
+	memcpy(item->data, data, size);
+	return item;
 }
 
-/**
- * kdbus_meta_export() - export information from metadata into a slice
- * @mp:		Process metadata, or NULL
- * @mc:		Connection metadata, or NULL
- * @mask:	Mask of KDBUS_ATTACH_* flags to export
- * @slice:	The slice to export to
- * @offset:	The offset inside @slice to write to
- * @real_size:	The real size the metadata consumed
- *
- * This function exports information from metadata into @slice at offset
- * @offset inside that slice. Only information that is requested in @mask
- * and that has been collected before is exported.
- *
- * In order to make sure not to write out of bounds, @mask must be the same
- * value that was previously returned from kdbus_meta_export_prepare(). The
- * function will, however, not necessarily write as many bytes as returned by
- * kdbus_meta_export_prepare(); depending on the namespaces in question, it
- * might use up less than that.
- *
- * All information will be translated using the current namespaces.
- *
- * Return: 0 on success, negative error number otherwise.
- */
-int kdbus_meta_export(struct kdbus_meta_proc *mp,
-		      struct kdbus_meta_conn *mc,
-		      u64 mask,
-		      struct kdbus_pool_slice *slice,
-		      off_t offset,
-		      size_t *real_size)
+static size_t kdbus_meta_write(struct kdbus_meta_staging *staging, void *mem,
+			       size_t size)
 {
-	struct user_namespace *user_ns = current_user_ns();
-	struct kdbus_item_header item_hdr[13], *hdr;
-	char *exe_pathname = NULL;
-	struct kdbus_creds creds;
-	struct kdbus_pids pids;
-	void *exe_page = NULL;
-	struct kvec kvec[40];
-	u64 *auxgrps = NULL;
-	size_t cnt = 0;
-	u64 size = 0;
-	int ret = 0;
-
-	hdr = &item_hdr[0];
+	struct user_namespace *user_ns = staging->conn->cred->user_ns;
+	struct pid_namespace *pid_ns = ns_of_pid(staging->conn->pid);
+	struct kdbus_item *item = NULL, *items = mem;
+	u8 *end, *owned_names_end = NULL;
 
-	/*
-	 * TODO: We currently have no sane way of translating a set of caps
-	 * between different user namespaces. Until that changes, we have
-	 * to drop such items.
-	 */
-	if (mp && mp->caps_namespace != user_ns)
-		mask &= ~KDBUS_ATTACH_CAPS;
+	/* process metadata */
 
-	if (mask == 0) {
-		*real_size = 0;
-		return 0;
+	if (staging->mf && (staging->mask & KDBUS_ATTACH_CREDS)) {
+		const struct kdbus_meta_fake *mf = staging->mf;
+
+		item = kdbus_write_head(&items, KDBUS_ITEM_CREDS,
+					sizeof(struct kdbus_creds));
+		item->creds = (struct kdbus_creds){
+			.uid	= kdbus_from_kuid_keep(user_ns, mf->uid),
+			.euid	= kdbus_from_kuid_keep(user_ns, mf->euid),
+			.suid	= kdbus_from_kuid_keep(user_ns, mf->suid),
+			.fsuid	= kdbus_from_kuid_keep(user_ns, mf->fsuid),
+			.gid	= kdbus_from_kgid_keep(user_ns, mf->gid),
+			.egid	= kdbus_from_kgid_keep(user_ns, mf->egid),
+			.sgid	= kdbus_from_kgid_keep(user_ns, mf->sgid),
+			.fsgid	= kdbus_from_kgid_keep(user_ns, mf->fsgid),
+		};
+	} else if (staging->mp && (staging->mask & KDBUS_ATTACH_CREDS)) {
+		const struct cred *c = staging->mp->cred;
+
+		item = kdbus_write_head(&items, KDBUS_ITEM_CREDS,
+					sizeof(struct kdbus_creds));
+		item->creds = (struct kdbus_creds){
+			.uid	= kdbus_from_kuid_keep(user_ns, c->uid),
+			.euid	= kdbus_from_kuid_keep(user_ns, c->euid),
+			.suid	= kdbus_from_kuid_keep(user_ns, c->suid),
+			.fsuid	= kdbus_from_kuid_keep(user_ns, c->fsuid),
+			.gid	= kdbus_from_kgid_keep(user_ns, c->gid),
+			.egid	= kdbus_from_kgid_keep(user_ns, c->egid),
+			.sgid	= kdbus_from_kgid_keep(user_ns, c->sgid),
+			.fsgid	= kdbus_from_kgid_keep(user_ns, c->fsgid),
+		};
 	}
 
-	/* process metadata */
+	if (staging->mf && (staging->mask & KDBUS_ATTACH_PIDS)) {
+		item = kdbus_write_head(&items, KDBUS_ITEM_PIDS,
+					sizeof(struct kdbus_pids));
+		item->pids = (struct kdbus_pids){
+			.pid = pid_nr_ns(staging->mf->tgid, pid_ns),
+			.tid = pid_nr_ns(staging->mf->pid, pid_ns),
+			.ppid = pid_nr_ns(staging->mf->ppid, pid_ns),
+		};
+	} else if (staging->mp && (staging->mask & KDBUS_ATTACH_PIDS)) {
+		item = kdbus_write_head(&items, KDBUS_ITEM_PIDS,
+					sizeof(struct kdbus_pids));
+		item->pids = (struct kdbus_pids){
+			.pid = pid_nr_ns(staging->mp->tgid, pid_ns),
+			.tid = pid_nr_ns(staging->mp->pid, pid_ns),
+			.ppid = pid_nr_ns(staging->mp->ppid, pid_ns),
+		};
+	}
+
+	if (staging->mp && (staging->mask & KDBUS_ATTACH_AUXGROUPS)) {
+		const struct group_info *info = staging->mp->cred->group_info;
+		size_t i;
 
-	if (mp && (mask & KDBUS_ATTACH_CREDS)) {
-		creds.uid	= kdbus_from_kuid_keep(mp->uid);
-		creds.euid	= kdbus_from_kuid_keep(mp->euid);
-		creds.suid	= kdbus_from_kuid_keep(mp->suid);
-		creds.fsuid	= kdbus_from_kuid_keep(mp->fsuid);
-		creds.gid	= kdbus_from_kgid_keep(mp->gid);
-		creds.egid	= kdbus_from_kgid_keep(mp->egid);
-		creds.sgid	= kdbus_from_kgid_keep(mp->sgid);
-		creds.fsgid	= kdbus_from_kgid_keep(mp->fsgid);
-
-		cnt += kdbus_meta_push_kvec(kvec + cnt, hdr++, KDBUS_ITEM_CREDS,
-					    &creds, sizeof(creds), &size);
+		item = kdbus_write_head(&items, KDBUS_ITEM_AUXGROUPS,
+					info->ngroups * sizeof(u64));
+		for (i = 0; i < info->ngroups; ++i)
+			item->data64[i] = from_kgid_munged(user_ns,
+							   GROUP_AT(info, i));
 	}
 
-	if (mp && (mask & KDBUS_ATTACH_PIDS)) {
-		pids.pid = pid_vnr(mp->tgid);
-		pids.tid = pid_vnr(mp->pid);
-		pids.ppid = pid_vnr(mp->ppid);
+	if (staging->mp && (staging->mask & KDBUS_ATTACH_TID_COMM))
+		item = kdbus_write_full(&items, KDBUS_ITEM_TID_COMM,
+					strlen(staging->mp->tid_comm) + 1,
+					staging->mp->tid_comm);
+
+	if (staging->mp && (staging->mask & KDBUS_ATTACH_PID_COMM))
+		item = kdbus_write_full(&items, KDBUS_ITEM_PID_COMM,
+					strlen(staging->mp->pid_comm) + 1,
+					staging->mp->pid_comm);
+
+	if (staging->exe_path && (staging->mask & KDBUS_ATTACH_EXE))
+		item = kdbus_write_full(&items, KDBUS_ITEM_EXE,
+					strlen(staging->exe_path) + 1,
+					staging->exe_path);
+
+	if (staging->mp && (staging->mask & KDBUS_ATTACH_CMDLINE))
+		item = kdbus_write_full(&items, KDBUS_ITEM_CMDLINE,
+					strlen(staging->mp->cmdline) + 1,
+					staging->mp->cmdline);
+
+	if (staging->mp && (staging->mask & KDBUS_ATTACH_CGROUP))
+		item = kdbus_write_full(&items, KDBUS_ITEM_CGROUP,
+					strlen(staging->mp->cgroup) + 1,
+					staging->mp->cgroup);
+
+	if (staging->mp && (staging->mask & KDBUS_ATTACH_CAPS)) {
+		item = kdbus_write_head(&items, KDBUS_ITEM_CAPS,
+					sizeof(struct kdbus_meta_caps));
+		kdbus_meta_export_caps((void*)&item->caps, staging->mp,
+				       user_ns);
+	}
 
-		cnt += kdbus_meta_push_kvec(kvec + cnt, hdr++, KDBUS_ITEM_PIDS,
-					    &pids, sizeof(pids), &size);
+	if (staging->mf && (staging->mask & KDBUS_ATTACH_SECLABEL))
+		item = kdbus_write_full(&items, KDBUS_ITEM_SECLABEL,
+					strlen(staging->mf->seclabel) + 1,
+					staging->mf->seclabel);
+	else if (staging->mp && (staging->mask & KDBUS_ATTACH_SECLABEL))
+		item = kdbus_write_full(&items, KDBUS_ITEM_SECLABEL,
+					strlen(staging->mp->seclabel) + 1,
+					staging->mp->seclabel);
+
+	if (staging->mp && (staging->mask & KDBUS_ATTACH_AUDIT)) {
+		item = kdbus_write_head(&items, KDBUS_ITEM_AUDIT,
+					sizeof(struct kdbus_audit));
+		item->audit = (struct kdbus_audit){
+			.loginuid = from_kuid(user_ns,
+					      staging->mp->audit_loginuid),
+			.sessionid = staging->mp->audit_sessionid,
+		};
 	}
 
-	if (mp && (mask & KDBUS_ATTACH_AUXGROUPS)) {
-		size_t payload_size = mp->n_auxgrps * sizeof(u64);
-		int i;
+	/* connection metadata */
 
-		auxgrps = kmalloc(payload_size, GFP_KERNEL);
-		if (!auxgrps) {
-			ret = -ENOMEM;
-			goto exit;
-		}
+	if (staging->mc && (staging->mask & KDBUS_ATTACH_NAMES)) {
+		memcpy(items, staging->mc->owned_names_items,
+		       KDBUS_ALIGN8(staging->mc->owned_names_size));
+		owned_names_end = (u8 *)items + staging->mc->owned_names_size;
+		items = (void *)KDBUS_ALIGN8((unsigned long)owned_names_end);
+	}
+
+	if (staging->mc && (staging->mask & KDBUS_ATTACH_CONN_DESCRIPTION))
+		item = kdbus_write_full(&items, KDBUS_ITEM_CONN_DESCRIPTION,
+				strlen(staging->mc->conn_description) + 1,
+				staging->mc->conn_description);
 
-		for (i = 0; i < mp->n_auxgrps; i++)
-			auxgrps[i] = from_kgid_munged(user_ns, mp->auxgrps[i]);
+	if (staging->mc && (staging->mask & KDBUS_ATTACH_TIMESTAMP))
+		item = kdbus_write_full(&items, KDBUS_ITEM_TIMESTAMP,
+					sizeof(staging->mc->ts),
+					&staging->mc->ts);
 
-		cnt += kdbus_meta_push_kvec(kvec + cnt, hdr++,
-					    KDBUS_ITEM_AUXGROUPS,
-					    auxgrps, payload_size, &size);
+	/*
+	 * Return real size (minus trailing padding). In case of 'owned_names'
+	 * we cannot deduce it from item->size, so treat it special.
+	 */
+
+	if (items == (void *)KDBUS_ALIGN8((unsigned long)owned_names_end))
+		end = owned_names_end;
+	else if (item)
+		end = (u8 *)item + item->size;
+	else
+		end = mem;
+
+	WARN_ON((u8 *)items - (u8 *)mem != size);
+	WARN_ON((void *)KDBUS_ALIGN8((unsigned long)end) != (void *)items);
+
+	return end - (u8 *)mem;
+}
+
+int kdbus_meta_emit(struct kdbus_meta_proc *mp,
+		    struct kdbus_meta_fake *mf,
+		    struct kdbus_meta_conn *mc,
+		    struct kdbus_conn *conn,
+		    u64 mask,
+		    struct kdbus_item **out_items,
+		    size_t *out_size)
+{
+	struct kdbus_meta_staging staging = {};
+	struct kdbus_item *items = NULL;
+	size_t size = 0;
+	int ret;
+
+	if (WARN_ON(mf && mp))
+		mp = NULL;
+
+	staging.mp = mp;
+	staging.mf = mf;
+	staging.mc = mc;
+	staging.conn = conn;
+
+	/* get mask of valid items */
+	if (mf)
+		staging.mask |= mf->valid;
+	if (mp) {
+		mutex_lock(&mp->lock);
+		staging.mask |= mp->valid;
+		mutex_unlock(&mp->lock);
+	}
+	if (mc) {
+		mutex_lock(&mc->lock);
+		staging.mask |= mc->valid;
+		mutex_unlock(&mc->lock);
 	}
 
-	if (mp && (mask & KDBUS_ATTACH_TID_COMM))
-		cnt += kdbus_meta_push_kvec(kvec + cnt, hdr++,
-					    KDBUS_ITEM_TID_COMM, mp->tid_comm,
-					    strlen(mp->tid_comm) + 1, &size);
+	staging.mask &= mask;
 
-	if (mp && (mask & KDBUS_ATTACH_PID_COMM))
-		cnt += kdbus_meta_push_kvec(kvec + cnt, hdr++,
-					    KDBUS_ITEM_PID_COMM, mp->pid_comm,
-					    strlen(mp->pid_comm) + 1, &size);
+	if (!staging.mask) { /* bail out if nothing to do */
+		ret = 0;
+		goto exit;
+	}
 
-	if (mp && (mask & KDBUS_ATTACH_EXE)) {
+	/* EXE is special as it needs a temporary page to assemble */
+	if (mp && (staging.mask & KDBUS_ATTACH_EXE)) {
 		struct path p;
 
 		/*
-		 * TODO: We need access to __d_path() so we can write the path
+		 * XXX: We need access to __d_path() so we can write the path
 		 * relative to conn->root_path. Once upstream, we need
 		 * EXPORT_SYMBOL(__d_path) or an equivalent of d_path() that
 		 * takes the root path directly. Until then, we drop this item
@@ -1057,103 +1112,236 @@ int kdbus_meta_export(struct kdbus_meta_proc *mp,
 		 */
 
 		get_fs_root(current->fs, &p);
-		if (path_equal(&p, &mp->root_path)) {
-			exe_page = (void *)__get_free_page(GFP_TEMPORARY);
-			if (!exe_page) {
+		if (path_equal(&p, &conn->root_path)) {
+			staging.exe = (void *)__get_free_page(GFP_TEMPORARY);
+			if (!staging.exe) {
 				path_put(&p);
 				ret = -ENOMEM;
 				goto exit;
 			}
 
-			exe_pathname = d_path(&mp->exe_path, exe_page,
-					      PAGE_SIZE);
-			if (IS_ERR(exe_pathname)) {
+			staging.exe_path = d_path(&mp->exe_path, staging.exe,
+						  PAGE_SIZE);
+			if (IS_ERR(staging.exe_path)) {
 				path_put(&p);
-				ret = PTR_ERR(exe_pathname);
+				ret = PTR_ERR(staging.exe_path);
 				goto exit;
 			}
-
-			cnt += kdbus_meta_push_kvec(kvec + cnt, hdr++,
-						    KDBUS_ITEM_EXE,
-						    exe_pathname,
-						    strlen(exe_pathname) + 1,
-						    &size);
 		}
 		path_put(&p);
 	}
 
-	if (mp && (mask & KDBUS_ATTACH_CMDLINE))
-		cnt += kdbus_meta_push_kvec(kvec + cnt, hdr++,
-					    KDBUS_ITEM_CMDLINE, mp->cmdline,
-					    strlen(mp->cmdline) + 1, &size);
+	size = kdbus_meta_measure(&staging);
+	if (!size) { /* bail out if nothing to do */
+		ret = 0;
+		goto exit;
+	}
 
-	if (mp && (mask & KDBUS_ATTACH_CGROUP))
-		cnt += kdbus_meta_push_kvec(kvec + cnt, hdr++,
-					    KDBUS_ITEM_CGROUP, mp->cgroup,
-					    strlen(mp->cgroup) + 1, &size);
+	items = kmalloc(size, GFP_KERNEL);
+	if (!items) {
+		ret = -ENOMEM;
+		goto exit;
+	}
 
-	if (mp && (mask & KDBUS_ATTACH_CAPS))
-		cnt += kdbus_meta_push_kvec(kvec + cnt, hdr++,
-					    KDBUS_ITEM_CAPS, &mp->caps,
-					    sizeof(mp->caps), &size);
-
-	if (mp && (mask & KDBUS_ATTACH_SECLABEL))
-		cnt += kdbus_meta_push_kvec(kvec + cnt, hdr++,
-					    KDBUS_ITEM_SECLABEL, mp->seclabel,
-					    strlen(mp->seclabel) + 1, &size);
-
-	if (mp && (mask & KDBUS_ATTACH_AUDIT)) {
-		struct kdbus_audit a = {
-			.loginuid = from_kuid(user_ns, mp->audit_loginuid),
-			.sessionid = mp->audit_sessionid,
-		};
+	size = kdbus_meta_write(&staging, items, size);
+	if (!size) {
+		kfree(items);
+		items = NULL;
+	}
+
+	ret = 0;
 
-		cnt += kdbus_meta_push_kvec(kvec + cnt, hdr++, KDBUS_ITEM_AUDIT,
-					    &a, sizeof(a), &size);
+exit:
+	if (staging.exe)
+		free_page((unsigned long)staging.exe);
+	if (ret >= 0) {
+		*out_items = items;
+		*out_size = size;
 	}
+	return ret;
+}
 
-	/* connection metadata */
+enum {
+	KDBUS_META_PROC_NONE,
+	KDBUS_META_PROC_NORMAL,
+};
 
-	if (mc && (mask & KDBUS_ATTACH_NAMES))
-		kdbus_kvec_set(&kvec[cnt++], mc->owned_names_items,
-			       mc->owned_names_size, &size);
+/**
+ * kdbus_proc_permission() - check /proc permissions on target pid
+ * @pid_ns:		namespace we operate in
+ * @cred:		credentials of requestor
+ * @target:		target process
+ *
+ * This checks whether a process with credentials @cred can access information
+ * of @target in the namespace @pid_ns. This tries to follow /proc permissions,
+ * but is slightly more restrictive.
+ *
+ * Return: The /proc access level (KDBUS_META_PROC_*) is returned.
+ */
+static unsigned int kdbus_proc_permission(const struct pid_namespace *pid_ns,
+					  const struct cred *cred,
+					  struct pid *target)
+{
+	if (pid_ns->hide_pid < 1)
+		return KDBUS_META_PROC_NORMAL;
 
-	if (mc && (mask & KDBUS_ATTACH_CONN_DESCRIPTION))
-		cnt += kdbus_meta_push_kvec(kvec + cnt, hdr++,
-					    KDBUS_ITEM_CONN_DESCRIPTION,
-					    mc->conn_description,
-					    strlen(mc->conn_description) + 1,
-					    &size);
+	/* XXX: we need groups_search() exported for aux-groups */
+	if (gid_eq(cred->egid, pid_ns->pid_gid))
+		return KDBUS_META_PROC_NORMAL;
 
-	if (mc && (mask & KDBUS_ATTACH_TIMESTAMP))
-		cnt += kdbus_meta_push_kvec(kvec + cnt, hdr++,
-					    KDBUS_ITEM_TIMESTAMP, &mc->ts,
-					    sizeof(mc->ts), &size);
+	/*
+	 * XXX: If ptrace_may_access(PTRACE_MODE_READ) is granted, you can
+	 * overwrite hide_pid. However, ptrace_may_access() only supports
+	 * checking 'current', hence, we cannot use this here. But we
+	 * simply decide to not support this override, so no need to worry.
+	 */
 
-	ret = kdbus_pool_slice_copy_kvec(slice, offset, kvec, cnt, size);
-	*real_size = size;
+	return KDBUS_META_PROC_NONE;
+}
 
-exit:
-	kfree(auxgrps);
+/**
+ * kdbus_meta_proc_mask() - calculate which metadata would be visible to
+ *			    a connection via /proc
+ * @prv_pid:		pid of metadata provider
+ * @req_pid:		pid of metadata requestor
+ * @req_cred:		credentials of metadata reqeuestor
+ * @wanted:		metadata that is requested
+ *
+ * This checks which metadata items of @prv_pid can be read via /proc by the
+ * requestor @req_pid.
+ *
+ * Return: Set of metadata flags the requestor can see (limited by @wanted).
+ */
+static u64 kdbus_meta_proc_mask(struct pid *prv_pid,
+				struct pid *req_pid,
+				const struct cred *req_cred,
+				u64 wanted)
+{
+	struct pid_namespace *prv_ns, *req_ns;
+	unsigned int proc;
 
-	if (exe_page)
-		free_page((unsigned long)exe_page);
+	prv_ns = ns_of_pid(prv_pid);
+	req_ns = ns_of_pid(req_pid);
 
-	return ret;
+	/*
+	 * If the sender is not visible in the receiver namespace, then the
+	 * receiver cannot access the sender via its own procfs. Hence, we do
+	 * not attach any additional metadata.
+	 */
+	if (!pid_nr_ns(prv_pid, req_ns))
+		return 0;
+
+	/*
+	 * If the pid-namespace of the receiver has hide_pid set, it cannot see
+	 * any process but its own. We shortcut this /proc permission check if
+	 * provider and requestor are the same. If not, we perform rather
+	 * expensive /proc permission checks.
+	 */
+	if (prv_pid == req_pid)
+		proc = KDBUS_META_PROC_NORMAL;
+	else
+		proc = kdbus_proc_permission(req_ns, req_cred, prv_pid);
+
+	/* you need /proc access to read standard process attributes */
+	if (proc < KDBUS_META_PROC_NORMAL)
+		wanted &= ~(KDBUS_ATTACH_TID_COMM |
+			    KDBUS_ATTACH_PID_COMM |
+			    KDBUS_ATTACH_SECLABEL |
+			    KDBUS_ATTACH_CMDLINE |
+			    KDBUS_ATTACH_CGROUP |
+			    KDBUS_ATTACH_AUDIT |
+			    KDBUS_ATTACH_CAPS |
+			    KDBUS_ATTACH_EXE);
+
+	/* clear all non-/proc flags */
+	return wanted & (KDBUS_ATTACH_TID_COMM |
+			 KDBUS_ATTACH_PID_COMM |
+			 KDBUS_ATTACH_SECLABEL |
+			 KDBUS_ATTACH_CMDLINE |
+			 KDBUS_ATTACH_CGROUP |
+			 KDBUS_ATTACH_AUDIT |
+			 KDBUS_ATTACH_CAPS |
+			 KDBUS_ATTACH_EXE);
 }
 
 /**
- * kdbus_meta_calc_attach_flags() - calculate attach flags for a sender
- *				    and a receiver
- * @sender:		Sending connection
- * @receiver:		Receiving connection
+ * kdbus_meta_get_mask() - calculate attach flags mask for metadata request
+ * @prv_pid:		pid of metadata provider
+ * @prv_mask:		mask of metadata the provide grants unchecked
+ * @req_pid:		pid of metadata requestor
+ * @req_cred:		credentials of metadata requestor
+ * @req_mask:		mask of metadata that is requested
  *
- * Return: the attach flags both the sender and the receiver have opted-in
- * for.
+ * This calculates the metadata items that the requestor @req_pid can access
+ * from the metadata provider @prv_pid. This permission check consists of
+ * several different parts:
+ *  - Providers can grant metadata items unchecked. Regardless of their type,
+ *    they're always granted to the requestor. This mask is passed as @prv_mask.
+ *  - Basic items (credentials and connection metadata) are granted implicitly
+ *    to everyone. They're publicly available to any bus-user that can see the
+ *    provider.
+ *  - Process credentials that are not granted implicitly follow the same
+ *    permission checks as /proc. This means, we always assume a requestor
+ *    process has access to their *own* /proc mount, if they have access to
+ *    kdbusfs.
+ *
+ * Return: Mask of metadata that is granted.
+ */
+static u64 kdbus_meta_get_mask(struct pid *prv_pid, u64 prv_mask,
+			       struct pid *req_pid,
+			       const struct cred *req_cred, u64 req_mask)
+{
+	u64 missing, impl_mask, proc_mask = 0;
+
+	/*
+	 * Connection metadata and basic unix process credentials are
+	 * transmitted implicitly, and cannot be suppressed. Both are required
+	 * to perform user-space policies on the receiver-side. Furthermore,
+	 * connection metadata is public state, anyway, and unix credentials
+	 * are needed for UDS-compatibility. We extend them slightly by
+	 * auxiliary groups and additional uids/gids/pids.
+	 */
+	impl_mask = /* connection metadata */
+		    KDBUS_ATTACH_CONN_DESCRIPTION |
+		    KDBUS_ATTACH_TIMESTAMP |
+		    KDBUS_ATTACH_NAMES |
+		    /* credentials and pids */
+		    KDBUS_ATTACH_AUXGROUPS |
+		    KDBUS_ATTACH_CREDS |
+		    KDBUS_ATTACH_PIDS;
+
+	/*
+	 * Calculate the set of metadata that is not granted implicitly nor by
+	 * the sender, but still requested by the receiver. If any are left,
+	 * perform rather expensive /proc access checks for them.
+	 */
+	missing = req_mask & ~((prv_mask | impl_mask) & req_mask);
+	if (missing)
+		proc_mask = kdbus_meta_proc_mask(prv_pid, req_pid, req_cred,
+						 missing);
+
+	return (prv_mask | impl_mask | proc_mask) & req_mask;
+}
+
+/**
+ */
+u64 kdbus_meta_info_mask(const struct kdbus_conn *conn, u64 mask)
+{
+	return kdbus_meta_get_mask(conn->pid,
+				   atomic64_read(&conn->attach_flags_send),
+				   task_pid(current),
+				   current_cred(),
+				   mask);
+}
+
+/**
  */
-u64 kdbus_meta_calc_attach_flags(const struct kdbus_conn *sender,
-				 const struct kdbus_conn *receiver)
+u64 kdbus_meta_msg_mask(const struct kdbus_conn *snd,
+			const struct kdbus_conn *rcv)
 {
-	return atomic64_read(&sender->attach_flags_send) &
-	       atomic64_read(&receiver->attach_flags_recv);
+	return kdbus_meta_get_mask(task_pid(current),
+				   atomic64_read(&snd->attach_flags_send),
+				   rcv->pid,
+				   rcv->cred,
+				   atomic64_read(&rcv->attach_flags_recv));
 }
diff --git a/ipc/kdbus/metadata.h b/ipc/kdbus/metadata.h
index 0c977cd9e28d..dba7cc7fdbcb 100644
--- a/ipc/kdbus/metadata.h
+++ b/ipc/kdbus/metadata.h
@@ -16,43 +16,71 @@
 #define __KDBUS_METADATA_H
 
 #include <linux/kernel.h>
-#include <uapi/linux/kdbus.h>
 
 struct kdbus_conn;
-struct kdbus_kmsg;
 struct kdbus_pool_slice;
 
 struct kdbus_meta_proc;
 struct kdbus_meta_conn;
 
-extern unsigned long long kdbus_meta_attach_mask;
+/**
+ * struct kdbus_meta_fake - Fake metadata
+ * @valid:		Bitmask of collected and valid items
+ * @uid:		UID of process
+ * @euid:		EUID of process
+ * @suid:		SUID of process
+ * @fsuid:		FSUID of process
+ * @gid:		GID of process
+ * @egid:		EGID of process
+ * @sgid:		SGID of process
+ * @fsgid:		FSGID of process
+ * @pid:		PID of process
+ * @tgid:		TGID of process
+ * @ppid:		PPID of process
+ * @seclabel:		Seclabel
+ */
+struct kdbus_meta_fake {
+	u64 valid;
+
+	/* KDBUS_ITEM_CREDS */
+	kuid_t uid, euid, suid, fsuid;
+	kgid_t gid, egid, sgid, fsgid;
+
+	/* KDBUS_ITEM_PIDS */
+	struct pid *pid, *tgid, *ppid;
+
+	/* KDBUS_ITEM_SECLABEL */
+	char *seclabel;
+};
 
 struct kdbus_meta_proc *kdbus_meta_proc_new(void);
 struct kdbus_meta_proc *kdbus_meta_proc_ref(struct kdbus_meta_proc *mp);
 struct kdbus_meta_proc *kdbus_meta_proc_unref(struct kdbus_meta_proc *mp);
 int kdbus_meta_proc_collect(struct kdbus_meta_proc *mp, u64 what);
-int kdbus_meta_proc_fake(struct kdbus_meta_proc *mp,
-			 const struct kdbus_creds *creds,
-			 const struct kdbus_pids *pids,
-			 const char *seclabel);
+
+struct kdbus_meta_fake *kdbus_meta_fake_new(void);
+struct kdbus_meta_fake *kdbus_meta_fake_free(struct kdbus_meta_fake *mf);
+int kdbus_meta_fake_collect(struct kdbus_meta_fake *mf,
+			    const struct kdbus_creds *creds,
+			    const struct kdbus_pids *pids,
+			    const char *seclabel);
 
 struct kdbus_meta_conn *kdbus_meta_conn_new(void);
 struct kdbus_meta_conn *kdbus_meta_conn_ref(struct kdbus_meta_conn *mc);
 struct kdbus_meta_conn *kdbus_meta_conn_unref(struct kdbus_meta_conn *mc);
 int kdbus_meta_conn_collect(struct kdbus_meta_conn *mc,
-			    struct kdbus_kmsg *kmsg,
 			    struct kdbus_conn *conn,
-			    u64 what);
-
-int kdbus_meta_export_prepare(struct kdbus_meta_proc *mp,
-			      struct kdbus_meta_conn *mc,
-			      u64 *mask, size_t *sz);
-int kdbus_meta_export(struct kdbus_meta_proc *mp,
-		      struct kdbus_meta_conn *mc,
-		      u64 mask,
-		      struct kdbus_pool_slice *slice,
-		      off_t offset, size_t *real_size);
-u64 kdbus_meta_calc_attach_flags(const struct kdbus_conn *sender,
-				 const struct kdbus_conn *receiver);
+			    u64 msg_seqnum, u64 what);
+
+int kdbus_meta_emit(struct kdbus_meta_proc *mp,
+		    struct kdbus_meta_fake *mf,
+		    struct kdbus_meta_conn *mc,
+		    struct kdbus_conn *conn,
+		    u64 mask,
+		    struct kdbus_item **out_items,
+		    size_t *out_size);
+u64 kdbus_meta_info_mask(const struct kdbus_conn *conn, u64 mask);
+u64 kdbus_meta_msg_mask(const struct kdbus_conn *snd,
+			const struct kdbus_conn *rcv);
 
 #endif
diff --git a/ipc/kdbus/names.c b/ipc/kdbus/names.c
index df99e4df815b..bf44ca3f12b6 100644
--- a/ipc/kdbus/names.c
+++ b/ipc/kdbus/names.c
@@ -34,167 +34,128 @@
 #include "notify.h"
 #include "policy.h"
 
-struct kdbus_name_pending {
-	u64 flags;
-	struct kdbus_conn *conn;
-	struct kdbus_name_entry *name;
-	struct list_head conn_entry;
-	struct list_head name_entry;
-};
+#define KDBUS_NAME_SAVED_MASK (KDBUS_NAME_ALLOW_REPLACEMENT |	\
+			       KDBUS_NAME_QUEUE)
 
-static int kdbus_name_pending_new(struct kdbus_name_entry *e,
-				  struct kdbus_conn *conn, u64 flags)
+static bool kdbus_name_owner_is_used(struct kdbus_name_owner *owner)
 {
-	struct kdbus_name_pending *p;
-
-	kdbus_conn_assert_active(conn);
-
-	p = kmalloc(sizeof(*p), GFP_KERNEL);
-	if (!p)
-		return -ENOMEM;
-
-	p->flags = flags;
-	p->conn = conn;
-	p->name = e;
-	list_add_tail(&p->conn_entry, &conn->names_queue_list);
-	list_add_tail(&p->name_entry, &e->queue);
-
-	return 0;
+	return !list_empty(&owner->name_entry) ||
+	       owner == owner->name->activator;
 }
 
-static void kdbus_name_pending_free(struct kdbus_name_pending *p)
+static struct kdbus_name_owner *
+kdbus_name_owner_new(struct kdbus_conn *conn, struct kdbus_name_entry *name,
+		     u64 flags)
 {
-	if (!p)
-		return;
+	struct kdbus_name_owner *owner;
 
-	list_del(&p->name_entry);
-	list_del(&p->conn_entry);
-	kfree(p);
-}
-
-static struct kdbus_name_entry *
-kdbus_name_entry_new(struct kdbus_name_registry *r, u32 hash, const char *name)
-{
-	struct kdbus_name_entry *e;
-	size_t namelen;
+	kdbus_conn_assert_active(conn);
 
-	namelen = strlen(name);
+	if (conn->name_count >= KDBUS_CONN_MAX_NAMES)
+		return ERR_PTR(-E2BIG);
 
-	e = kmalloc(sizeof(*e) + namelen + 1, GFP_KERNEL);
-	if (!e)
+	owner = kmalloc(sizeof(*owner), GFP_KERNEL);
+	if (!owner)
 		return ERR_PTR(-ENOMEM);
 
-	e->name_id = ++r->name_seq_last;
-	e->flags = 0;
-	e->conn = NULL;
-	e->activator = NULL;
-	INIT_LIST_HEAD(&e->queue);
-	INIT_LIST_HEAD(&e->conn_entry);
-	hash_add(r->entries_hash, &e->hentry, hash);
-	memcpy(e->name, name, namelen + 1);
+	owner->flags = flags & KDBUS_NAME_SAVED_MASK;
+	owner->conn = conn;
+	owner->name = name;
+	list_add_tail(&owner->conn_entry, &conn->names_list);
+	INIT_LIST_HEAD(&owner->name_entry);
 
-	return e;
+	++conn->name_count;
+	return owner;
 }
 
-static void kdbus_name_entry_free(struct kdbus_name_entry *e)
+static void kdbus_name_owner_free(struct kdbus_name_owner *owner)
 {
-	if (!e)
+	if (!owner)
 		return;
 
-	WARN_ON(!list_empty(&e->conn_entry));
-	WARN_ON(!list_empty(&e->queue));
-	WARN_ON(e->activator);
-	WARN_ON(e->conn);
-
-	hash_del(&e->hentry);
-	kfree(e);
+	WARN_ON(kdbus_name_owner_is_used(owner));
+	--owner->conn->name_count;
+	list_del(&owner->conn_entry);
+	kfree(owner);
 }
 
-static void kdbus_name_entry_set_owner(struct kdbus_name_entry *e,
-				       struct kdbus_conn *conn, u64 flags)
+static struct kdbus_name_owner *
+kdbus_name_owner_find(struct kdbus_name_entry *name, struct kdbus_conn *conn)
 {
-	WARN_ON(e->conn);
+	struct kdbus_name_owner *owner;
 
-	e->conn = kdbus_conn_ref(conn);
-	e->flags = flags;
-	atomic_inc(&conn->name_count);
-	list_add_tail(&e->conn_entry, &e->conn->names_list);
+	/*
+	 * Use conn->names_list over name->queue to make sure boundaries of
+	 * this linear search are controlled by the connection itself.
+	 * Furthermore, this will find normal owners as well as activators
+	 * without any additional code.
+	 */
+	list_for_each_entry(owner, &conn->names_list, conn_entry)
+		if (owner->name == name)
+			return owner;
+
+	return NULL;
 }
 
-static void kdbus_name_entry_remove_owner(struct kdbus_name_entry *e)
+static bool kdbus_name_entry_is_used(struct kdbus_name_entry *name)
 {
-	WARN_ON(!e->conn);
-
-	list_del_init(&e->conn_entry);
-	atomic_dec(&e->conn->name_count);
-	e->flags = 0;
-	e->conn = kdbus_conn_unref(e->conn);
+	return !list_empty(&name->queue) || name->activator;
 }
 
-static void kdbus_name_entry_replace_owner(struct kdbus_name_entry *e,
-					   struct kdbus_conn *conn, u64 flags)
+static struct kdbus_name_owner *
+kdbus_name_entry_first(struct kdbus_name_entry *name)
 {
-	if (WARN_ON(!e->conn) || WARN_ON(conn == e->conn))
-		return;
-
-	kdbus_notify_name_change(conn->ep->bus, KDBUS_ITEM_NAME_CHANGE,
-				 e->conn->id, conn->id,
-				 e->flags, flags, e->name);
-	kdbus_name_entry_remove_owner(e);
-	kdbus_name_entry_set_owner(e, conn, flags);
+	return list_first_entry_or_null(&name->queue, struct kdbus_name_owner,
+					name_entry);
 }
 
-/**
- * kdbus_name_is_valid() - check if a name is valid
- * @p:			The name to check
- * @allow_wildcard:	Whether or not to allow a wildcard name
- *
- * A name is valid if all of the following criterias are met:
- *
- *  - The name has two or more elements separated by a period ('.') character.
- *  - All elements must contain at least one character.
- *  - Each element must only contain the ASCII characters "[A-Z][a-z][0-9]_-"
- *    and must not begin with a digit.
- *  - The name must not exceed KDBUS_NAME_MAX_LEN.
- *  - If @allow_wildcard is true, the name may end on '.*'
- */
-bool kdbus_name_is_valid(const char *p, bool allow_wildcard)
+static struct kdbus_name_entry *
+kdbus_name_entry_new(struct kdbus_name_registry *r, u32 hash,
+		     const char *name_str)
 {
-	bool dot, found_dot = false;
-	const char *q;
+	struct kdbus_name_entry *name;
+	size_t namelen;
 
-	for (dot = true, q = p; *q; q++) {
-		if (*q == '.') {
-			if (dot)
-				return false;
+	lockdep_assert_held(&r->rwlock);
 
-			found_dot = true;
-			dot = true;
-		} else {
-			bool good;
+	namelen = strlen(name_str);
 
-			good = isalpha(*q) || (!dot && isdigit(*q)) ||
-				*q == '_' || *q == '-' ||
-				(allow_wildcard && dot &&
-					*q == '*' && *(q + 1) == '\0');
+	name = kmalloc(sizeof(*name) + namelen + 1, GFP_KERNEL);
+	if (!name)
+		return ERR_PTR(-ENOMEM);
 
-			if (!good)
-				return false;
+	name->name_id = ++r->name_seq_last;
+	name->activator = NULL;
+	INIT_LIST_HEAD(&name->queue);
+	hash_add(r->entries_hash, &name->hentry, hash);
+	memcpy(name->name, name_str, namelen + 1);
 
-			dot = false;
-		}
-	}
+	return name;
+}
 
-	if (q - p > KDBUS_NAME_MAX_LEN)
-		return false;
+static void kdbus_name_entry_free(struct kdbus_name_entry *name)
+{
+	if (!name)
+		return;
 
-	if (dot)
-		return false;
+	WARN_ON(kdbus_name_entry_is_used(name));
+	hash_del(&name->hentry);
+	kfree(name);
+}
 
-	if (!found_dot)
-		return false;
+static struct kdbus_name_entry *
+kdbus_name_entry_find(struct kdbus_name_registry *r, u32 hash,
+		      const char *name_str)
+{
+	struct kdbus_name_entry *name;
 
-	return true;
+	lockdep_assert_held(&r->rwlock);
+
+	hash_for_each_possible(r->entries_hash, name, hentry, hash)
+		if (!strcmp(name->name, name_str))
+			return name;
+
+	return NULL;
 }
 
 /**
@@ -218,32 +179,19 @@ struct kdbus_name_registry *kdbus_name_registry_new(void)
 }
 
 /**
- * kdbus_name_registry_free() - drop a name reg's reference
- * @reg:		The name registry, may be %NULL
+ * kdbus_name_registry_free() - free name registry
+ * @r:		name registry to free, or NULL
  *
- * Cleanup the name registry's internal structures.
+ * Free a name registry and cleanup all internal objects. This is a no-op if
+ * you pass NULL as registry.
  */
-void kdbus_name_registry_free(struct kdbus_name_registry *reg)
+void kdbus_name_registry_free(struct kdbus_name_registry *r)
 {
-	if (!reg)
+	if (!r)
 		return;
 
-	WARN_ON(!hash_empty(reg->entries_hash));
-	kfree(reg);
-}
-
-static struct kdbus_name_entry *
-kdbus_name_find(struct kdbus_name_registry *reg, u32 hash, const char *name)
-{
-	struct kdbus_name_entry *e;
-
-	lockdep_assert_held(&reg->rwlock);
-
-	hash_for_each_possible(reg->entries_hash, e, hentry, hash)
-		if (strcmp(e->name, name) == 0)
-			return e;
-
-	return NULL;
+	WARN_ON(!hash_empty(r->entries_hash));
+	kfree(r);
 }
 
 /**
@@ -260,183 +208,286 @@ kdbus_name_find(struct kdbus_name_registry *reg, u32 hash, const char *name)
 struct kdbus_name_entry *
 kdbus_name_lookup_unlocked(struct kdbus_name_registry *reg, const char *name)
 {
-	return kdbus_name_find(reg, kdbus_strhash(name), name);
+	return kdbus_name_entry_find(reg, kdbus_strhash(name), name);
 }
 
-/**
- * kdbus_name_acquire() - acquire a name
- * @reg:		The name registry
- * @conn:		The connection to pin this entry to
- * @name:		The name to acquire
- * @flags:		Acquisition flags (KDBUS_NAME_*)
- * @return_flags:	Pointer to return flags for the acquired name
- *			(KDBUS_NAME_*), may be %NULL
- *
- * Callers must ensure that @conn is either a privileged bus user or has
- * sufficient privileges in the policy-db to own the well-known name @name.
- *
- * Return: 0 success, negative error number on failure.
- */
-int kdbus_name_acquire(struct kdbus_name_registry *reg,
-		       struct kdbus_conn *conn, const char *name,
-		       u64 flags, u64 *return_flags)
+static int kdbus_name_become_activator(struct kdbus_name_owner *owner,
+				       u64 *return_flags)
 {
-	struct kdbus_name_entry *e;
-	u64 rflags = 0;
-	int ret = 0;
-	u32 hash;
+	if (kdbus_name_owner_is_used(owner))
+		return -EALREADY;
+	if (owner->name->activator)
+		return -EEXIST;
 
-	kdbus_conn_assert_active(conn);
+	owner->name->activator = owner;
+	owner->flags |= KDBUS_NAME_ACTIVATOR;
 
-	down_write(&reg->rwlock);
-
-	if (!kdbus_conn_policy_own_name(conn, current_cred(), name)) {
-		ret = -EPERM;
-		goto exit_unlock;
+	if (kdbus_name_entry_first(owner->name)) {
+		owner->flags |= KDBUS_NAME_IN_QUEUE;
+	} else {
+		owner->flags |= KDBUS_NAME_PRIMARY;
+		kdbus_notify_name_change(owner->conn->ep->bus,
+					 KDBUS_ITEM_NAME_ADD,
+					 0, owner->conn->id,
+					 0, owner->flags,
+					 owner->name->name);
 	}
 
-	hash = kdbus_strhash(name);
-	e = kdbus_name_find(reg, hash, name);
-	if (!e) {
-		/* claim new name */
+	if (return_flags)
+		*return_flags = owner->flags | KDBUS_NAME_ACQUIRED;
 
-		if (conn->activator_of) {
-			ret = -EINVAL;
-			goto exit_unlock;
-		}
+	return 0;
+}
 
-		e = kdbus_name_entry_new(reg, hash, name);
-		if (IS_ERR(e)) {
-			ret = PTR_ERR(e);
-			goto exit_unlock;
-		}
+static int kdbus_name_update(struct kdbus_name_owner *owner, u64 flags,
+			     u64 *return_flags)
+{
+	struct kdbus_name_owner *primary, *activator;
+	struct kdbus_name_entry *name;
+	struct kdbus_bus *bus;
+	u64 nflags = 0;
+	int ret = 0;
 
-		if (kdbus_conn_is_activator(conn)) {
-			e->activator = kdbus_conn_ref(conn);
-			conn->activator_of = e;
+	name = owner->name;
+	bus = owner->conn->ep->bus;
+	primary = kdbus_name_entry_first(name);
+	activator = name->activator;
+
+	/* cannot be activator and acquire a name */
+	if (owner == activator)
+		return -EUCLEAN;
+
+	/* update saved flags */
+	owner->flags = flags & KDBUS_NAME_SAVED_MASK;
+
+	if (!primary) {
+		/*
+		 * No primary owner (but maybe an activator). Take over the
+		 * name.
+		 */
+
+		list_add(&owner->name_entry, &name->queue);
+		owner->flags |= KDBUS_NAME_PRIMARY;
+		nflags |= KDBUS_NAME_ACQUIRED;
+
+		/* move messages to new owner on activation */
+		if (activator) {
+			kdbus_conn_move_messages(owner->conn, activator->conn,
+						 name->name_id);
+			kdbus_notify_name_change(bus, KDBUS_ITEM_NAME_CHANGE,
+					activator->conn->id, owner->conn->id,
+					activator->flags, owner->flags,
+					name->name);
+			activator->flags &= ~KDBUS_NAME_PRIMARY;
+			activator->flags |= KDBUS_NAME_IN_QUEUE;
+		} else {
+			kdbus_notify_name_change(bus, KDBUS_ITEM_NAME_ADD,
+						 0, owner->conn->id,
+						 0, owner->flags,
+						 name->name);
 		}
 
-		kdbus_name_entry_set_owner(e, conn, flags);
-		kdbus_notify_name_change(e->conn->ep->bus, KDBUS_ITEM_NAME_ADD,
-					 0, e->conn->id, 0, e->flags, e->name);
-	} else if (e->conn == conn || e == conn->activator_of) {
-		/* connection already owns that name */
-		ret = -EALREADY;
-	} else if (kdbus_conn_is_activator(conn)) {
-		/* activator claims existing name */
-
-		if (conn->activator_of) {
-			ret = -EINVAL; /* multiple names not allowed */
-		} else if (e->activator) {
-			ret = -EEXIST; /* only one activator per name */
+	} else if (owner == primary) {
+		/*
+		 * Already the primary owner of the name, flags were already
+		 * updated. Nothing to do.
+		 */
+
+		owner->flags |= KDBUS_NAME_PRIMARY;
+
+	} else if ((primary->flags & KDBUS_NAME_ALLOW_REPLACEMENT) &&
+		   (flags & KDBUS_NAME_REPLACE_EXISTING)) {
+		/*
+		 * We're not the primary owner but can replace it. Move us
+		 * ahead of the primary owner and acquire the name (possibly
+		 * skipping queued owners ahead of us).
+		 */
+
+		list_del_init(&owner->name_entry);
+		list_add(&owner->name_entry, &name->queue);
+		owner->flags |= KDBUS_NAME_PRIMARY;
+		nflags |= KDBUS_NAME_ACQUIRED;
+
+		kdbus_notify_name_change(bus, KDBUS_ITEM_NAME_CHANGE,
+					 primary->conn->id, owner->conn->id,
+					 primary->flags, owner->flags,
+					 name->name);
+
+		/* requeue old primary, or drop if queueing not wanted */
+		if (primary->flags & KDBUS_NAME_QUEUE) {
+			primary->flags &= ~KDBUS_NAME_PRIMARY;
+			primary->flags |= KDBUS_NAME_IN_QUEUE;
 		} else {
-			e->activator = kdbus_conn_ref(conn);
-			conn->activator_of = e;
-		}
-	} else if (e->flags & KDBUS_NAME_ACTIVATOR) {
-		/* claim name of an activator */
-
-		kdbus_conn_move_messages(conn, e->activator, 0);
-		kdbus_name_entry_replace_owner(e, conn, flags);
-	} else if ((flags & KDBUS_NAME_REPLACE_EXISTING) &&
-		   (e->flags & KDBUS_NAME_ALLOW_REPLACEMENT)) {
-		/* claim name of a previous owner */
-
-		if (e->flags & KDBUS_NAME_QUEUE) {
-			/* move owner back to queue if they asked for it */
-			ret = kdbus_name_pending_new(e, e->conn, e->flags);
-			if (ret < 0)
-				goto exit_unlock;
+			list_del_init(&primary->name_entry);
+			kdbus_name_owner_free(primary);
 		}
 
-		kdbus_name_entry_replace_owner(e, conn, flags);
 	} else if (flags & KDBUS_NAME_QUEUE) {
-		/* add to waiting-queue of the name */
+		/*
+		 * Name is already occupied and we cannot take it over, but
+		 * queuing is allowed. Put us silently on the queue, if not
+		 * already there.
+		 */
+
+		owner->flags |= KDBUS_NAME_IN_QUEUE;
+		if (!kdbus_name_owner_is_used(owner)) {
+			list_add_tail(&owner->name_entry, &name->queue);
+			nflags |= KDBUS_NAME_ACQUIRED;
+		}
+	} else if (kdbus_name_owner_is_used(owner)) {
+		/*
+		 * Already queued on name, but re-queueing was not requested.
+		 * Make sure to unlink it from the name, the caller is
+		 * responsible for releasing it.
+		 */
+
+		list_del_init(&owner->name_entry);
+	} else {
+		/*
+		 * Name is already claimed and queueing is not requested.
+		 * Return error to the caller.
+		 */
+
+		ret = -EEXIST;
+	}
 
-		struct kdbus_name_pending *p;
-		bool in_queue = false;
+	if (return_flags)
+		*return_flags = owner->flags | nflags;
 
-		list_for_each_entry(p, &e->queue, name_entry) {
-			if (p->conn == conn) {
-				/* connection is already queued */
-				rflags |= KDBUS_NAME_IN_QUEUE;
-				in_queue = true;
-				break;
-			}
+	return ret;
+}
+
+int kdbus_name_acquire(struct kdbus_name_registry *reg,
+		       struct kdbus_conn *conn, const char *name_str,
+		       u64 flags, u64 *return_flags)
+{
+	struct kdbus_name_entry *name = NULL;
+	struct kdbus_name_owner *owner = NULL;
+	u32 hash;
+	int ret;
+
+	kdbus_conn_assert_active(conn);
+
+	down_write(&reg->rwlock);
+
+	/*
+	 * Verify the connection has access to the name. Do this before testing
+	 * for double-acquisitions and other errors to make sure we do not leak
+	 * information about this name through possible custom endpoints.
+	 */
+	if (!kdbus_conn_policy_own_name(conn, current_cred(), name_str)) {
+		ret = -EPERM;
+		goto exit;
+	}
+
+	/*
+	 * Lookup the name entry. If it already exists, search for an owner
+	 * entry as we might already own that name. If either does not exist,
+	 * we will allocate a fresh one.
+	 */
+	hash = kdbus_strhash(name_str);
+	name = kdbus_name_entry_find(reg, hash, name_str);
+	if (name) {
+		owner = kdbus_name_owner_find(name, conn);
+	} else {
+		name = kdbus_name_entry_new(reg, hash, name_str);
+		if (IS_ERR(name)) {
+			ret = PTR_ERR(name);
+			name = NULL;
+			goto exit;
 		}
+	}
 
-		if (!in_queue) {
-			ret = kdbus_name_pending_new(e, conn, flags);
-			if (ret >= 0)
-				/* tell the caller that we queued it */
-				rflags |= KDBUS_NAME_IN_QUEUE;
+	/* create name owner object if not already queued */
+	if (!owner) {
+		owner = kdbus_name_owner_new(conn, name, flags);
+		if (IS_ERR(owner)) {
+			ret = PTR_ERR(owner);
+			owner = NULL;
+			goto exit;
 		}
-	} else {
-		/* the name is busy, return a failure */
-		ret = -EEXIST;
 	}
 
-	if (ret == 0 && return_flags)
-		*return_flags = rflags;
+	if (flags & KDBUS_NAME_ACTIVATOR)
+		ret = kdbus_name_become_activator(owner, return_flags);
+	else
+		ret = kdbus_name_update(owner, flags, return_flags);
+	if (ret < 0)
+		goto exit;
 
-exit_unlock:
+exit:
+	if (owner && !kdbus_name_owner_is_used(owner))
+		kdbus_name_owner_free(owner);
+	if (name && !kdbus_name_entry_is_used(name))
+		kdbus_name_entry_free(name);
 	up_write(&reg->rwlock);
 	kdbus_notify_flush(conn->ep->bus);
 	return ret;
 }
 
-static void kdbus_name_release_unlocked(struct kdbus_name_registry *reg,
-					struct kdbus_name_entry *e)
+static void kdbus_name_release_unlocked(struct kdbus_name_owner *owner)
 {
-	struct kdbus_name_pending *p;
-
-	lockdep_assert_held(&reg->rwlock);
-
-	p = list_first_entry_or_null(&e->queue, struct kdbus_name_pending,
-				     name_entry);
-
-	if (p) {
-		/* give it to first active waiter in the queue */
-		kdbus_name_entry_replace_owner(e, p->conn, p->flags);
-		kdbus_name_pending_free(p);
-	} else if (e->activator && e->activator != e->conn) {
-		/* hand it back to an active activator connection */
-		kdbus_conn_move_messages(e->activator, e->conn, e->name_id);
-		kdbus_name_entry_replace_owner(e, e->activator,
-					       KDBUS_NAME_ACTIVATOR);
-	} else {
-		/* release the name */
-		kdbus_notify_name_change(e->conn->ep->bus,
-					 KDBUS_ITEM_NAME_REMOVE,
-					 e->conn->id, 0, e->flags, 0, e->name);
-		kdbus_name_entry_remove_owner(e);
-		kdbus_name_entry_free(e);
+	struct kdbus_name_owner *primary, *next;
+	struct kdbus_name_entry *name;
+
+	name = owner->name;
+	primary = kdbus_name_entry_first(name);
+
+	list_del_init(&owner->name_entry);
+	if (owner == name->activator)
+		name->activator = NULL;
+
+	if (!primary || owner == primary) {
+		next = kdbus_name_entry_first(name);
+		if (!next)
+			next = name->activator;
+
+		if (next) {
+			/* hand to next in queue */
+			next->flags &= ~KDBUS_NAME_IN_QUEUE;
+			next->flags |= KDBUS_NAME_PRIMARY;
+			if (next == name->activator)
+				kdbus_conn_move_messages(next->conn,
+							 owner->conn,
+							 name->name_id);
+
+			kdbus_notify_name_change(owner->conn->ep->bus,
+					KDBUS_ITEM_NAME_CHANGE,
+					owner->conn->id, next->conn->id,
+					owner->flags, next->flags,
+					name->name);
+		} else {
+			kdbus_notify_name_change(owner->conn->ep->bus,
+						 KDBUS_ITEM_NAME_REMOVE,
+						 owner->conn->id, 0,
+						 owner->flags, 0,
+						 name->name);
+		}
 	}
+
+	kdbus_name_owner_free(owner);
+	if (!kdbus_name_entry_is_used(name))
+		kdbus_name_entry_free(name);
 }
 
 static int kdbus_name_release(struct kdbus_name_registry *reg,
 			      struct kdbus_conn *conn,
-			      const char *name)
+			      const char *name_str)
 {
-	struct kdbus_name_pending *p;
-	struct kdbus_name_entry *e;
+	struct kdbus_name_owner *owner;
+	struct kdbus_name_entry *name;
 	int ret = 0;
 
 	down_write(&reg->rwlock);
-	e = kdbus_name_find(reg, kdbus_strhash(name), name);
-	if (!e) {
-		ret = -ESRCH;
-	} else if (e->conn == conn) {
-		kdbus_name_release_unlocked(reg, e);
+	name = kdbus_name_entry_find(reg, kdbus_strhash(name_str), name_str);
+	if (name) {
+		owner = kdbus_name_owner_find(name, conn);
+		if (owner)
+			kdbus_name_release_unlocked(owner);
+		else
+			ret = -EADDRINUSE;
 	} else {
-		ret = -EADDRINUSE;
-		list_for_each_entry(p, &e->queue, name_entry) {
-			if (p->conn == conn) {
-				kdbus_name_pending_free(p);
-				ret = 0;
-				break;
-			}
-		}
+		ret = -ESRCH;
 	}
 	up_write(&reg->rwlock);
 
@@ -452,38 +503,79 @@ static int kdbus_name_release(struct kdbus_name_registry *reg,
 void kdbus_name_release_all(struct kdbus_name_registry *reg,
 			    struct kdbus_conn *conn)
 {
-	struct kdbus_name_pending *p;
-	struct kdbus_conn *activator = NULL;
-	struct kdbus_name_entry *e;
+	struct kdbus_name_owner *owner;
 
 	down_write(&reg->rwlock);
 
-	if (kdbus_conn_is_activator(conn)) {
-		activator = conn->activator_of->activator;
-		conn->activator_of->activator = NULL;
-	}
-
-	while ((p = list_first_entry_or_null(&conn->names_queue_list,
-					     struct kdbus_name_pending,
-					     conn_entry)))
-		kdbus_name_pending_free(p);
-	while ((e = list_first_entry_or_null(&conn->names_list,
-					     struct kdbus_name_entry,
-					     conn_entry)))
-		kdbus_name_release_unlocked(reg, e);
+	while ((owner = list_first_entry_or_null(&conn->names_list,
+						 struct kdbus_name_owner,
+						 conn_entry)))
+		kdbus_name_release_unlocked(owner);
 
 	up_write(&reg->rwlock);
 
-	kdbus_conn_unref(activator);
 	kdbus_notify_flush(conn->ep->bus);
 }
 
 /**
+ * kdbus_name_is_valid() - check if a name is valid
+ * @p:			The name to check
+ * @allow_wildcard:	Whether or not to allow a wildcard name
+ *
+ * A name is valid if all of the following criterias are met:
+ *
+ *  - The name has two or more elements separated by a period ('.') character.
+ *  - All elements must contain at least one character.
+ *  - Each element must only contain the ASCII characters "[A-Z][a-z][0-9]_-"
+ *    and must not begin with a digit.
+ *  - The name must not exceed KDBUS_NAME_MAX_LEN.
+ *  - If @allow_wildcard is true, the name may end on '.*'
+ */
+bool kdbus_name_is_valid(const char *p, bool allow_wildcard)
+{
+	bool dot, found_dot = false;
+	const char *q;
+
+	for (dot = true, q = p; *q; q++) {
+		if (*q == '.') {
+			if (dot)
+				return false;
+
+			found_dot = true;
+			dot = true;
+		} else {
+			bool good;
+
+			good = isalpha(*q) || (!dot && isdigit(*q)) ||
+				*q == '_' || *q == '-' ||
+				(allow_wildcard && dot &&
+					*q == '*' && *(q + 1) == '\0');
+
+			if (!good)
+				return false;
+
+			dot = false;
+		}
+	}
+
+	if (q - p > KDBUS_NAME_MAX_LEN)
+		return false;
+
+	if (dot)
+		return false;
+
+	if (!found_dot)
+		return false;
+
+	return true;
+}
+
+/**
  * kdbus_cmd_name_acquire() - handle KDBUS_CMD_NAME_ACQUIRE
  * @conn:		connection to operate on
  * @argp:		command payload
  *
- * Return: 0 on success, negative error code on failure.
+ * Return: >=0 on success, negative error code on failure.
  */
 int kdbus_cmd_name_acquire(struct kdbus_conn *conn, void __user *argp)
 {
@@ -517,22 +609,9 @@ int kdbus_cmd_name_acquire(struct kdbus_conn *conn, void __user *argp)
 		goto exit;
 	}
 
-	/*
-	 * Do atomic_inc_return here to reserve our slot, then decrement
-	 * it before returning.
-	 */
-	if (atomic_inc_return(&conn->name_count) > KDBUS_CONN_MAX_NAMES) {
-		ret = -E2BIG;
-		goto exit_dec;
-	}
-
 	ret = kdbus_name_acquire(conn->ep->bus->name_registry, conn, item_name,
 				 cmd->flags, &cmd->return_flags);
-	if (ret < 0)
-		goto exit_dec;
 
-exit_dec:
-	atomic_dec(&conn->name_count);
 exit:
 	return kdbus_args_clear(&args, ret);
 }
@@ -542,7 +621,7 @@ exit:
  * @conn:		connection to operate on
  * @argp:		command payload
  *
- * Return: 0 on success, negative error code on failure.
+ * Return: >=0 on success, negative error code on failure.
  */
 int kdbus_cmd_name_release(struct kdbus_conn *conn, void __user *argp)
 {
@@ -575,7 +654,7 @@ static int kdbus_list_write(struct kdbus_conn *conn,
 			    struct kdbus_conn *c,
 			    struct kdbus_pool_slice *slice,
 			    size_t *pos,
-			    struct kdbus_name_entry *e,
+			    struct kdbus_name_owner *o,
 			    bool write)
 {
 	struct kvec kvec[4];
@@ -596,22 +675,22 @@ static int kdbus_list_write(struct kdbus_conn *conn,
 		u64 flags;
 	} h = {};
 
-	if (e && !kdbus_conn_policy_see_name_unlocked(conn, current_cred(),
-						      e->name))
+	if (o && !kdbus_conn_policy_see_name_unlocked(conn, current_cred(),
+						      o->name->name))
 		return 0;
 
 	kdbus_kvec_set(&kvec[cnt++], &info, sizeof(info), &info.size);
 
 	/* append name */
-	if (e) {
-		size_t slen = strlen(e->name) + 1;
+	if (o) {
+		size_t slen = strlen(o->name->name) + 1;
 
 		h.size = offsetof(struct kdbus_item, name.name) + slen;
 		h.type = KDBUS_ITEM_OWNED_NAME;
-		h.flags = e->flags;
+		h.flags = o->flags;
 
 		kdbus_kvec_set(&kvec[cnt++], &h, sizeof(h), &info.size);
-		kdbus_kvec_set(&kvec[cnt++], e->name, slen, &info.size);
+		kdbus_kvec_set(&kvec[cnt++], o->name->name, slen, &info.size);
 		cnt += !!kdbus_kvec_pad(&kvec[cnt], &info.size);
 	}
 
@@ -641,63 +720,52 @@ static int kdbus_list_all(struct kdbus_conn *conn, u64 flags,
 		if (kdbus_conn_is_monitor(c))
 			continue;
 
-		/* skip activators */
-		if (!(flags & KDBUS_LIST_ACTIVATORS) &&
-		    kdbus_conn_is_activator(c))
-			continue;
-
 		/* all names the connection owns */
-		if (flags & (KDBUS_LIST_NAMES | KDBUS_LIST_ACTIVATORS)) {
-			struct kdbus_name_entry *e;
+		if (flags & (KDBUS_LIST_NAMES |
+			     KDBUS_LIST_ACTIVATORS |
+			     KDBUS_LIST_QUEUED)) {
+			struct kdbus_name_owner *o;
 
-			list_for_each_entry(e, &c->names_list, conn_entry) {
-				struct kdbus_conn *a = e->activator;
+			list_for_each_entry(o, &c->names_list, conn_entry) {
+				if (o->flags & KDBUS_NAME_ACTIVATOR) {
+					if (!(flags & KDBUS_LIST_ACTIVATORS))
+						continue;
 
-				if ((flags & KDBUS_LIST_ACTIVATORS) &&
-				    a && a != c) {
-					ret = kdbus_list_write(conn, a, slice,
-							       &p, e, write);
+					ret = kdbus_list_write(conn, c, slice,
+							       &p, o, write);
 					if (ret < 0) {
 						mutex_unlock(&c->lock);
 						return ret;
 					}
 
 					added = true;
-				}
+				} else if (o->flags & KDBUS_NAME_IN_QUEUE) {
+					if (!(flags & KDBUS_LIST_QUEUED))
+						continue;
 
-				if (flags & KDBUS_LIST_NAMES ||
-				    kdbus_conn_is_activator(c)) {
 					ret = kdbus_list_write(conn, c, slice,
-							       &p, e, write);
+							       &p, o, write);
 					if (ret < 0) {
 						mutex_unlock(&c->lock);
 						return ret;
 					}
 
 					added = true;
-				}
-			}
-		}
+				} else if (flags & KDBUS_LIST_NAMES) {
+					ret = kdbus_list_write(conn, c, slice,
+							       &p, o, write);
+					if (ret < 0) {
+						mutex_unlock(&c->lock);
+						return ret;
+					}
 
-		/* queue of names the connection is currently waiting for */
-		if (flags & KDBUS_LIST_QUEUED) {
-			struct kdbus_name_pending *q;
-
-			list_for_each_entry(q, &c->names_queue_list,
-					    conn_entry) {
-				ret = kdbus_list_write(conn, c, slice, &p,
-						       q->name, write);
-				if (ret < 0) {
-					mutex_unlock(&c->lock);
-					return ret;
+					added = true;
 				}
-
-				added = true;
 			}
 		}
 
 		/* nothing added so far, just add the unique ID */
-		if (!added && flags & KDBUS_LIST_UNIQUE) {
+		if (!added && (flags & KDBUS_LIST_UNIQUE)) {
 			ret = kdbus_list_write(conn, c, slice, &p, NULL, write);
 			if (ret < 0)
 				return ret;
@@ -713,7 +781,7 @@ static int kdbus_list_all(struct kdbus_conn *conn, u64 flags,
  * @conn:		connection to operate on
  * @argp:		command payload
  *
- * Return: 0 on success, negative error code on failure.
+ * Return: >=0 on success, negative error code on failure.
  */
 int kdbus_cmd_list(struct kdbus_conn *conn, void __user *argp)
 {
diff --git a/ipc/kdbus/names.h b/ipc/kdbus/names.h
index 3dd2589293e0..edac59ddd8ee 100644
--- a/ipc/kdbus/names.h
+++ b/ipc/kdbus/names.h
@@ -18,6 +18,10 @@
 #include <linux/hashtable.h>
 #include <linux/rwsem.h>
 
+struct kdbus_name_entry;
+struct kdbus_name_owner;
+struct kdbus_name_registry;
+
 /**
  * struct kdbus_name_registry - names registered for a bus
  * @entries_hash:	Map of entries
@@ -32,27 +36,37 @@ struct kdbus_name_registry {
 
 /**
  * struct kdbus_name_entry - well-know name entry
- * @name_id:		Sequence number of name entry to be able to uniquely
+ * @name_id:		sequence number of name entry to be able to uniquely
  *			identify a name over its registration lifetime
- * @flags:		KDBUS_NAME_* flags
- * @conn:		Connection owning the name
- * @activator:		Connection of the activator queuing incoming messages
- * @queue:		List of queued connections
- * @conn_entry:		Entry in connection
- * @hentry:		Entry in registry map
- * @name:		The well-known name
+ * @activator:		activator of this name, or NULL
+ * @queue:		list of queued owners
+ * @hentry:		entry in registry map
+ * @name:		well-known name
  */
 struct kdbus_name_entry {
 	u64 name_id;
-	u64 flags;
-	struct kdbus_conn *conn;
-	struct kdbus_conn *activator;
+	struct kdbus_name_owner *activator;
 	struct list_head queue;
-	struct list_head conn_entry;
 	struct hlist_node hentry;
 	char name[];
 };
 
+/**
+ * struct kdbus_name_owner - owner of a well-known name
+ * @flags:		KDBUS_NAME_* flags of this owner
+ * @conn:		connection owning the name
+ * @name:		name that is owned
+ * @conn_entry:		link into @conn
+ * @name_entry:		link into @name
+ */
+struct kdbus_name_owner {
+	u64 flags;
+	struct kdbus_conn *conn;
+	struct kdbus_name_entry *name;
+	struct list_head conn_entry;
+	struct list_head name_entry;
+};
+
 bool kdbus_name_is_valid(const char *p, bool allow_wildcard);
 
 struct kdbus_name_registry *kdbus_name_registry_new(void);
@@ -71,4 +85,21 @@ int kdbus_cmd_name_acquire(struct kdbus_conn *conn, void __user *argp);
 int kdbus_cmd_name_release(struct kdbus_conn *conn, void __user *argp);
 int kdbus_cmd_list(struct kdbus_conn *conn, void __user *argp);
 
+/**
+ * kdbus_name_get_owner() - get current owner of a name
+ * @name:	name to get current owner of
+ *
+ * This returns a pointer to the current owner of a name (or its activator if
+ * there is no owner). The caller must make sure @name is valid and does not
+ * vanish.
+ *
+ * Return: Pointer to current owner or NULL if there is none.
+ */
+static inline struct kdbus_name_owner *
+kdbus_name_get_owner(struct kdbus_name_entry *name)
+{
+	return list_first_entry_or_null(&name->queue, struct kdbus_name_owner,
+					name_entry) ? : name->activator;
+}
+
 #endif
diff --git a/ipc/kdbus/node.c b/ipc/kdbus/node.c
index 520df00e676a..89f58bc85433 100644
--- a/ipc/kdbus/node.c
+++ b/ipc/kdbus/node.c
@@ -120,7 +120,7 @@
  * new active references can be acquired.
  * Once all active references are dropped, the node is considered 'drained'. Now
  * kdbus_node_deactivate() is called on each child of the node before we
- * continue deactvating our node. That is, once all children are entirely
+ * continue deactivating our node. That is, once all children are entirely
  * deactivated, we call ->release_cb() of our node. ->release_cb() can release
  * any resources on that node which are bound to the "active" state of a node.
  * When done, we unlink the node from its parent rb-tree, mark it as
@@ -178,7 +178,7 @@
  *                   accessed by other callers to properly initialize
  *                   filesystem nodes.
  *
- *     * node->id: This is an unsigned 32bit integer allocated by an IDR. It is
+ *     * node->id: This is an unsigned 32bit integer allocated by an IDA. It is
  *                 always kept as small as possible during allocation and is
  *                 globally unique across all nodes allocated by this module. 0
  *                 is reserved as "not assigned" and is the default.
@@ -233,8 +233,7 @@
 #define KDBUS_NODE_NEW			(KDBUS_NODE_BIAS - 4)
 
 /* global unique ID mapping for kdbus nodes */
-static DEFINE_IDR(kdbus_node_idr);
-static DECLARE_RWSEM(kdbus_node_idr_lock);
+DEFINE_IDA(kdbus_node_ida);
 
 /**
  * kdbus_node_name_hash() - hash a name
@@ -337,15 +336,11 @@ int kdbus_node_link(struct kdbus_node *node, struct kdbus_node *parent,
 		node->hash = kdbus_node_name_hash(name);
 	}
 
-	down_write(&kdbus_node_idr_lock);
-	ret = idr_alloc(&kdbus_node_idr, node, 1, 0, GFP_KERNEL);
-	if (ret >= 0)
-		node->id = ret;
-	up_write(&kdbus_node_idr_lock);
-
+	ret = ida_simple_get(&kdbus_node_ida, 1, 0, GFP_KERNEL);
 	if (ret < 0)
 		return ret;
 
+	node->id = ret;
 	ret = 0;
 
 	if (parent) {
@@ -440,16 +435,8 @@ struct kdbus_node *kdbus_node_unref(struct kdbus_node *node)
 
 		if (node->free_cb)
 			node->free_cb(node);
-
-		down_write(&kdbus_node_idr_lock);
 		if (safe.id > 0)
-			idr_remove(&kdbus_node_idr, safe.id);
-		/* drop caches after last node to not leak memory on unload */
-		if (idr_is_empty(&kdbus_node_idr)) {
-			idr_destroy(&kdbus_node_idr);
-			idr_init(&kdbus_node_idr);
-		}
-		up_write(&kdbus_node_idr_lock);
+			ida_simple_remove(&kdbus_node_ida, safe.id);
 
 		kfree(safe.name);
 
@@ -650,7 +637,7 @@ void kdbus_node_deactivate(struct kdbus_node *node)
 			kdbus_fs_flush(pos);
 
 			/*
-			 * If the node was activated and somone subtracted BIAS
+			 * If the node was activated and someone subtracted BIAS
 			 * from it to deactivate it, we, and only us, are
 			 * responsible to release the extra ref-count that was
 			 * taken once in kdbus_node_activate().
diff --git a/ipc/kdbus/node.h b/ipc/kdbus/node.h
index be125ce4fd58..970e02b08e9f 100644
--- a/ipc/kdbus/node.h
+++ b/ipc/kdbus/node.h
@@ -58,6 +58,8 @@ struct kdbus_node {
 
 #define kdbus_node_from_rb(_node) rb_entry((_node), struct kdbus_node, rb)
 
+extern struct ida kdbus_node_ida;
+
 void kdbus_node_init(struct kdbus_node *node, unsigned int type);
 
 int kdbus_node_link(struct kdbus_node *node, struct kdbus_node *parent,
diff --git a/ipc/kdbus/notify.c b/ipc/kdbus/notify.c
index e4a454222f09..375758c4896a 100644
--- a/ipc/kdbus/notify.c
+++ b/ipc/kdbus/notify.c
@@ -28,40 +28,24 @@
 #include "message.h"
 #include "notify.h"
 
-static inline void kdbus_notify_add_tail(struct kdbus_kmsg *kmsg,
+static inline void kdbus_notify_add_tail(struct kdbus_staging *staging,
 					 struct kdbus_bus *bus)
 {
 	spin_lock(&bus->notify_lock);
-	list_add_tail(&kmsg->notify_entry, &bus->notify_list);
+	list_add_tail(&staging->notify_entry, &bus->notify_list);
 	spin_unlock(&bus->notify_lock);
 }
 
 static int kdbus_notify_reply(struct kdbus_bus *bus, u64 id,
 			      u64 cookie, u64 msg_type)
 {
-	struct kdbus_kmsg *kmsg = NULL;
+	struct kdbus_staging *s;
 
-	WARN_ON(id == 0);
-
-	kmsg = kdbus_kmsg_new(bus, 0);
-	if (IS_ERR(kmsg))
-		return PTR_ERR(kmsg);
-
-	/*
-	 * a kernel-generated notification can only contain one
-	 * struct kdbus_item, so make a shortcut here for
-	 * faster lookup in the match db.
-	 */
-	kmsg->notify_type = msg_type;
-	kmsg->msg.flags = KDBUS_MSG_SIGNAL;
-	kmsg->msg.dst_id = id;
-	kmsg->msg.src_id = KDBUS_SRC_ID_KERNEL;
-	kmsg->msg.payload_type = KDBUS_PAYLOAD_KERNEL;
-	kmsg->msg.cookie_reply = cookie;
-	kmsg->msg.items[0].type = msg_type;
-
-	kdbus_notify_add_tail(kmsg, bus);
+	s = kdbus_staging_new_kernel(bus, id, cookie, 0, msg_type);
+	if (IS_ERR(s))
+		return PTR_ERR(s);
 
+	kdbus_notify_add_tail(s, bus);
 	return 0;
 }
 
@@ -115,32 +99,24 @@ int kdbus_notify_name_change(struct kdbus_bus *bus, u64 type,
 			     u64 old_flags, u64 new_flags,
 			     const char *name)
 {
-	struct kdbus_kmsg *kmsg = NULL;
 	size_t name_len, extra_size;
+	struct kdbus_staging *s;
 
 	name_len = strlen(name) + 1;
 	extra_size = sizeof(struct kdbus_notify_name_change) + name_len;
-	kmsg = kdbus_kmsg_new(bus, extra_size);
-	if (IS_ERR(kmsg))
-		return PTR_ERR(kmsg);
-
-	kmsg->msg.flags = KDBUS_MSG_SIGNAL;
-	kmsg->msg.dst_id = KDBUS_DST_ID_BROADCAST;
-	kmsg->msg.src_id = KDBUS_SRC_ID_KERNEL;
-	kmsg->msg.payload_type = KDBUS_PAYLOAD_KERNEL;
-	kmsg->notify_type = type;
-	kmsg->notify_old_id = old_id;
-	kmsg->notify_new_id = new_id;
-	kmsg->msg.items[0].type = type;
-	kmsg->msg.items[0].name_change.old_id.id = old_id;
-	kmsg->msg.items[0].name_change.old_id.flags = old_flags;
-	kmsg->msg.items[0].name_change.new_id.id = new_id;
-	kmsg->msg.items[0].name_change.new_id.flags = new_flags;
-	memcpy(kmsg->msg.items[0].name_change.name, name, name_len);
-	kmsg->notify_name = kmsg->msg.items[0].name_change.name;
-
-	kdbus_notify_add_tail(kmsg, bus);
 
+	s = kdbus_staging_new_kernel(bus, KDBUS_DST_ID_BROADCAST, 0,
+				     extra_size, type);
+	if (IS_ERR(s))
+		return PTR_ERR(s);
+
+	s->notify->name_change.old_id.id = old_id;
+	s->notify->name_change.old_id.flags = old_flags;
+	s->notify->name_change.new_id.id = new_id;
+	s->notify->name_change.new_id.flags = new_flags;
+	memcpy(s->notify->name_change.name, name, name_len);
+
+	kdbus_notify_add_tail(s, bus);
 	return 0;
 }
 
@@ -156,37 +132,19 @@ int kdbus_notify_name_change(struct kdbus_bus *bus, u64 type,
  */
 int kdbus_notify_id_change(struct kdbus_bus *bus, u64 type, u64 id, u64 flags)
 {
-	struct kdbus_kmsg *kmsg = NULL;
-
-	kmsg = kdbus_kmsg_new(bus, sizeof(struct kdbus_notify_id_change));
-	if (IS_ERR(kmsg))
-		return PTR_ERR(kmsg);
+	struct kdbus_staging *s;
+	size_t extra_size;
 
-	kmsg->msg.flags = KDBUS_MSG_SIGNAL;
-	kmsg->msg.dst_id = KDBUS_DST_ID_BROADCAST;
-	kmsg->msg.src_id = KDBUS_SRC_ID_KERNEL;
-	kmsg->msg.payload_type = KDBUS_PAYLOAD_KERNEL;
-	kmsg->notify_type = type;
+	extra_size = sizeof(struct kdbus_notify_id_change);
+	s = kdbus_staging_new_kernel(bus, KDBUS_DST_ID_BROADCAST, 0,
+				     extra_size, type);
+	if (IS_ERR(s))
+		return PTR_ERR(s);
 
-	switch (type) {
-	case KDBUS_ITEM_ID_ADD:
-		kmsg->notify_new_id = id;
-		break;
-
-	case KDBUS_ITEM_ID_REMOVE:
-		kmsg->notify_old_id = id;
-		break;
-
-	default:
-		BUG();
-	}
-
-	kmsg->msg.items[0].type = type;
-	kmsg->msg.items[0].id_change.id = id;
-	kmsg->msg.items[0].id_change.flags = flags;
-
-	kdbus_notify_add_tail(kmsg, bus);
+	s->notify->id_change.id = id;
+	s->notify->id_change.flags = flags;
 
+	kdbus_notify_add_tail(s, bus);
 	return 0;
 }
 
@@ -199,7 +157,7 @@ int kdbus_notify_id_change(struct kdbus_bus *bus, u64 type, u64 id, u64 flags)
 void kdbus_notify_flush(struct kdbus_bus *bus)
 {
 	LIST_HEAD(notify_list);
-	struct kdbus_kmsg *kmsg, *tmp;
+	struct kdbus_staging *s, *tmp;
 
 	mutex_lock(&bus->notify_flush_lock);
 	down_read(&bus->name_registry->rwlock);
@@ -208,25 +166,23 @@ void kdbus_notify_flush(struct kdbus_bus *bus)
 	list_splice_init(&bus->notify_list, &notify_list);
 	spin_unlock(&bus->notify_lock);
 
-	list_for_each_entry_safe(kmsg, tmp, &notify_list, notify_entry) {
-		kdbus_meta_conn_collect(kmsg->conn_meta, kmsg, NULL,
-					KDBUS_ATTACH_TIMESTAMP);
-
-		if (kmsg->msg.dst_id != KDBUS_DST_ID_BROADCAST) {
+	list_for_each_entry_safe(s, tmp, &notify_list, notify_entry) {
+		if (s->msg->dst_id != KDBUS_DST_ID_BROADCAST) {
 			struct kdbus_conn *conn;
 
-			conn = kdbus_bus_find_conn_by_id(bus, kmsg->msg.dst_id);
+			conn = kdbus_bus_find_conn_by_id(bus, s->msg->dst_id);
 			if (conn) {
-				kdbus_bus_eavesdrop(bus, NULL, kmsg);
-				kdbus_conn_entry_insert(NULL, conn, kmsg, NULL);
+				kdbus_bus_eavesdrop(bus, NULL, s);
+				kdbus_conn_entry_insert(NULL, conn, s, NULL,
+							NULL);
 				kdbus_conn_unref(conn);
 			}
 		} else {
-			kdbus_bus_broadcast(bus, NULL, kmsg);
+			kdbus_bus_broadcast(bus, NULL, s);
 		}
 
-		list_del(&kmsg->notify_entry);
-		kdbus_kmsg_free(kmsg);
+		list_del(&s->notify_entry);
+		kdbus_staging_free(s);
 	}
 
 	up_read(&bus->name_registry->rwlock);
@@ -239,10 +195,10 @@ void kdbus_notify_flush(struct kdbus_bus *bus)
  */
 void kdbus_notify_free(struct kdbus_bus *bus)
 {
-	struct kdbus_kmsg *kmsg, *tmp;
+	struct kdbus_staging *s, *tmp;
 
-	list_for_each_entry_safe(kmsg, tmp, &bus->notify_list, notify_entry) {
-		list_del(&kmsg->notify_entry);
-		kdbus_kmsg_free(kmsg);
+	list_for_each_entry_safe(s, tmp, &bus->notify_list, notify_entry) {
+		list_del(&s->notify_entry);
+		kdbus_staging_free(s);
 	}
 }
diff --git a/ipc/kdbus/policy.c b/ipc/kdbus/policy.c
index dd7fffaafa84..f2618e15e78d 100644
--- a/ipc/kdbus/policy.c
+++ b/ipc/kdbus/policy.c
@@ -344,7 +344,7 @@ err:
  * In order to allow atomic replacement of rules, the function first removes
  * all entries that have been created for the given owner previously.
  *
- * Callers to this function must make sur that the owner is a custom
+ * Callers to this function must make sure that the owner is a custom
  * endpoint, or if the endpoint is a default endpoint, then it must be
  * either a policy holder or an activator.
  *
diff --git a/ipc/kdbus/pool.c b/ipc/kdbus/pool.c
index 45dcdea505f4..63ccd55713c7 100644
--- a/ipc/kdbus/pool.c
+++ b/ipc/kdbus/pool.c
@@ -44,7 +44,7 @@
  * The receiver's buffer, managed as a pool of allocated and free
  * slices containing the queued messages.
  *
- * Messages sent with KDBUS_CMD_SEND are copied direcly by the
+ * Messages sent with KDBUS_CMD_SEND are copied directly by the
  * sending process into the receiver's pool.
  *
  * Messages received with KDBUS_CMD_RECV just return the offset
diff --git a/ipc/kdbus/queue.c b/ipc/kdbus/queue.c
index a449464a3975..f9c44d7bae6d 100644
--- a/ipc/kdbus/queue.c
+++ b/ipc/kdbus/queue.c
@@ -171,242 +171,43 @@ static void kdbus_queue_entry_unlink(struct kdbus_queue_entry *entry)
 
 /**
  * kdbus_queue_entry_new() - allocate a queue entry
- * @conn_dst:	destination connection
- * @kmsg:	kmsg object the queue entry should track
- * @user:	user to account message on (or NULL for kernel messages)
+ * @src:	source connection, or NULL
+ * @dst:	destination connection
+ * @s:		staging object carrying the message
  *
- * Allocates a queue entry based on a given kmsg and allocate space for
+ * Allocates a queue entry based on a given msg and allocate space for
  * the message payload and the requested metadata in the connection's pool.
  * The entry is not actually added to the queue's lists at this point.
  *
  * Return: the allocated entry on success, or an ERR_PTR on failures.
  */
-struct kdbus_queue_entry *kdbus_queue_entry_new(struct kdbus_conn *conn_dst,
-						const struct kdbus_kmsg *kmsg,
-						struct kdbus_user *user)
+struct kdbus_queue_entry *kdbus_queue_entry_new(struct kdbus_conn *src,
+						struct kdbus_conn *dst,
+						struct kdbus_staging *s)
 {
-	struct kdbus_msg_resources *res = kmsg->res;
-	const struct kdbus_msg *msg = &kmsg->msg;
 	struct kdbus_queue_entry *entry;
-	size_t memfd_cnt = 0;
-	struct kvec kvec[2];
-	size_t meta_size;
-	size_t msg_size;
-	u64 payload_off;
-	u64 size = 0;
-	int ret = 0;
+	int ret;
 
 	entry = kzalloc(sizeof(*entry), GFP_KERNEL);
 	if (!entry)
 		return ERR_PTR(-ENOMEM);
 
 	INIT_LIST_HEAD(&entry->entry);
-	entry->priority = msg->priority;
-	entry->dst_name_id = kmsg->dst_name_id;
-	entry->msg_res = kdbus_msg_resources_ref(res);
-	entry->proc_meta = kdbus_meta_proc_ref(kmsg->proc_meta);
-	entry->conn_meta = kdbus_meta_conn_ref(kmsg->conn_meta);
-	entry->conn = kdbus_conn_ref(conn_dst);
-
-	if (kmsg->msg.src_id == KDBUS_SRC_ID_KERNEL)
-		msg_size = msg->size;
-	else
-		msg_size = offsetof(struct kdbus_msg, items);
-
-	/* sum up the size of the needed slice */
-	size = msg_size;
-
-	if (res) {
-		size += res->vec_count *
-			KDBUS_ITEM_SIZE(sizeof(struct kdbus_vec));
-
-		if (res->memfd_count) {
-			entry->memfd_offset =
-				kcalloc(res->memfd_count, sizeof(size_t),
-					GFP_KERNEL);
-			if (!entry->memfd_offset) {
-				ret = -ENOMEM;
-				goto exit_free_entry;
-			}
-
-			size += res->memfd_count *
-				KDBUS_ITEM_SIZE(sizeof(struct kdbus_memfd));
-		}
-
-		if (res->fds_count)
-			size += KDBUS_ITEM_SIZE(sizeof(int) * res->fds_count);
-
-		if (res->dst_name)
-			size += KDBUS_ITEM_SIZE(strlen(res->dst_name) + 1);
-	}
-
-	/*
-	 * Remember the offset of the metadata part, so we can override
-	 * this part later during kdbus_queue_entry_install().
-	 */
-	entry->meta_offset = size;
-
-	if (entry->proc_meta || entry->conn_meta) {
-		entry->attach_flags =
-			atomic64_read(&conn_dst->attach_flags_recv);
-
-		ret = kdbus_meta_export_prepare(entry->proc_meta,
-						entry->conn_meta,
-						&entry->attach_flags,
-						&meta_size);
-		if (ret < 0)
-			goto exit_free_entry;
+	entry->priority = s->msg->priority;
+	entry->conn = kdbus_conn_ref(dst);
+	entry->gaps = kdbus_gaps_ref(s->gaps);
 
-		size += meta_size;
-	}
-
-	payload_off = size;
-	size += kmsg->pool_size;
-	size = KDBUS_ALIGN8(size);
-
-	ret = kdbus_conn_quota_inc(conn_dst, user, size,
-				   res ? res->fds_count : 0);
-	if (ret < 0)
-		goto exit_free_entry;
-
-	entry->slice = kdbus_pool_slice_alloc(conn_dst->pool, size, true);
+	entry->slice = kdbus_staging_emit(s, src, dst);
 	if (IS_ERR(entry->slice)) {
 		ret = PTR_ERR(entry->slice);
 		entry->slice = NULL;
-		kdbus_conn_quota_dec(conn_dst, user, size,
-				     res ? res->fds_count : 0);
-		goto exit_free_entry;
-	}
-
-	/* we accounted for exactly 'size' bytes, make sure it didn't grow */
-	WARN_ON(kdbus_pool_slice_size(entry->slice) != size);
-	entry->user = kdbus_user_ref(user);
-
-	/* copy message header */
-	kvec[0].iov_base = (char *)msg;
-	kvec[0].iov_len = msg_size;
-
-	ret = kdbus_pool_slice_copy_kvec(entry->slice, 0, kvec, 1, msg_size);
-	if (ret < 0)
-		goto exit_free_entry;
-
-	/* 'size' will now track the write position */
-	size = msg_size;
-
-	/* create message payload items */
-	if (res) {
-		size_t dst_name_len = 0;
-		unsigned int i;
-		size_t sz = 0;
-
-		if (res->dst_name) {
-			dst_name_len = strlen(res->dst_name) + 1;
-			sz += KDBUS_ITEM_SIZE(dst_name_len);
-		}
-
-		for (i = 0; i < res->data_count; ++i) {
-			struct kdbus_vec v;
-			struct kdbus_memfd m;
-
-			switch (res->data[i].type) {
-			case KDBUS_MSG_DATA_VEC:
-				sz += KDBUS_ITEM_SIZE(sizeof(v));
-				break;
-
-			case KDBUS_MSG_DATA_MEMFD:
-				sz += KDBUS_ITEM_SIZE(sizeof(m));
-				break;
-			}
-		}
-
-		if (sz) {
-			struct kdbus_item *items, *item;
-
-			items = kmalloc(sz, GFP_KERNEL);
-			if (!items) {
-				ret = -ENOMEM;
-				goto exit_free_entry;
-			}
-
-			item = items;
-
-			if (res->dst_name)
-				item = kdbus_item_set(item, KDBUS_ITEM_DST_NAME,
-						      res->dst_name,
-						      dst_name_len);
-
-			for (i = 0; i < res->data_count; ++i) {
-				struct kdbus_msg_data *d = res->data + i;
-				struct kdbus_memfd m = {};
-				struct kdbus_vec v = {};
-
-				switch (d->type) {
-				case KDBUS_MSG_DATA_VEC:
-					v.size = d->size;
-					v.offset = d->vec.off;
-					if (v.offset != ~0ULL)
-						v.offset += payload_off;
-
-					item = kdbus_item_set(item,
-							KDBUS_ITEM_PAYLOAD_OFF,
-							&v, sizeof(v));
-					break;
-
-				case KDBUS_MSG_DATA_MEMFD:
-					/*
-					 * Remember the location of memfds, so
-					 * we can override the content from
-					 * kdbus_queue_entry_install().
-					 */
-					entry->memfd_offset[memfd_cnt++] =
-						msg_size +
-						(char *)item - (char *)items +
-						offsetof(struct kdbus_item,
-							 memfd);
-
-					item = kdbus_item_set(item,
-						       KDBUS_ITEM_PAYLOAD_MEMFD,
-						       &m, sizeof(m));
-					break;
-				}
-			}
-
-			kvec[0].iov_base = items;
-			kvec[0].iov_len = sz;
-
-			ret = kdbus_pool_slice_copy_kvec(entry->slice, size,
-							 kvec, 1, sz);
-			kfree(items);
-
-			if (ret < 0)
-				goto exit_free_entry;
-
-			size += sz;
-		}
-
-		/*
-		 * Remember the location of the FD part, so we can override the
-		 * content in kdbus_queue_entry_install().
-		 */
-		if (res->fds_count) {
-			entry->fds_offset = size;
-			size += KDBUS_ITEM_SIZE(sizeof(int) * res->fds_count);
-		}
-	}
-
-	/* finally, copy over the actual message payload */
-	if (kmsg->iov_count) {
-		ret = kdbus_pool_slice_copy_iovec(entry->slice, payload_off,
-						  kmsg->iov,
-						  kmsg->iov_count,
-						  kmsg->pool_size);
-		if (ret < 0)
-			goto exit_free_entry;
+		goto error;
 	}
 
+	entry->user = src ? kdbus_user_ref(src->user) : NULL;
 	return entry;
 
-exit_free_entry:
+error:
 	kdbus_queue_entry_free(entry);
 	return ERR_PTR(ret);
 }
@@ -431,17 +232,13 @@ void kdbus_queue_entry_free(struct kdbus_queue_entry *entry)
 	if (entry->slice) {
 		kdbus_conn_quota_dec(entry->conn, entry->user,
 				     kdbus_pool_slice_size(entry->slice),
-				     entry->msg_res ?
-						entry->msg_res->fds_count : 0);
+				     entry->gaps ? entry->gaps->n_fds : 0);
 		kdbus_pool_slice_release(entry->slice);
-		kdbus_user_unref(entry->user);
 	}
 
-	kdbus_msg_resources_unref(entry->msg_res);
-	kdbus_meta_conn_unref(entry->conn_meta);
-	kdbus_meta_proc_unref(entry->proc_meta);
+	kdbus_user_unref(entry->user);
+	kdbus_gaps_unref(entry->gaps);
 	kdbus_conn_unref(entry->conn);
-	kfree(entry->memfd_offset);
 	kfree(entry);
 }
 
@@ -452,134 +249,22 @@ void kdbus_queue_entry_free(struct kdbus_queue_entry *entry)
  * @return_flags:	Pointer to store the return flags for userspace
  * @install_fds:	Whether or not to install associated file descriptors
  *
- * This function will create a slice to transport the message header, the
- * metadata items and other items for information stored in @entry, and
- * store it as entry->slice.
- *
- * If @install_fds is %true, file descriptors will as well be installed.
- * This function must always be called from the task context of the receiver.
- *
  * Return: 0 on success.
  */
 int kdbus_queue_entry_install(struct kdbus_queue_entry *entry,
 			      u64 *return_flags, bool install_fds)
 {
-	u64 msg_size = entry->meta_offset;
-	struct kdbus_conn *conn_dst = entry->conn;
-	struct kdbus_msg_resources *res;
 	bool incomplete_fds = false;
-	struct kvec kvec[2];
-	size_t memfds = 0;
-	int i, ret;
-
-	lockdep_assert_held(&conn_dst->lock);
-
-	if (entry->proc_meta || entry->conn_meta) {
-		size_t meta_size;
-
-		ret = kdbus_meta_export(entry->proc_meta,
-					entry->conn_meta,
-					entry->attach_flags,
-					entry->slice,
-					entry->meta_offset,
-					&meta_size);
-		if (ret < 0)
-			return ret;
-
-		msg_size += meta_size;
-	}
+	int ret;
 
-	/* Update message size at offset 0 */
-	kvec[0].iov_base = &msg_size;
-	kvec[0].iov_len = sizeof(msg_size);
+	lockdep_assert_held(&entry->conn->lock);
 
-	ret = kdbus_pool_slice_copy_kvec(entry->slice, 0, kvec, 1,
-					 sizeof(msg_size));
+	ret = kdbus_gaps_install(entry->gaps, entry->slice, &incomplete_fds);
 	if (ret < 0)
 		return ret;
 
-	res = entry->msg_res;
-
-	if (!res)
-		return 0;
-
-	if (res->fds_count) {
-		struct kdbus_item_header hdr;
-		size_t off;
-		int *fds;
-
-		fds = kmalloc_array(res->fds_count, sizeof(int), GFP_KERNEL);
-		if (!fds)
-			return -ENOMEM;
-
-		for (i = 0; i < res->fds_count; i++) {
-			if (install_fds) {
-				fds[i] = get_unused_fd_flags(O_CLOEXEC);
-				if (fds[i] >= 0)
-					fd_install(fds[i],
-						   get_file(res->fds[i]));
-				else
-					incomplete_fds = true;
-			} else {
-				fds[i] = -1;
-			}
-		}
-
-		off = entry->fds_offset;
-
-		hdr.type = KDBUS_ITEM_FDS;
-		hdr.size = KDBUS_ITEM_HEADER_SIZE +
-			   sizeof(int) * res->fds_count;
-
-		kvec[0].iov_base = &hdr;
-		kvec[0].iov_len = sizeof(hdr);
-
-		kvec[1].iov_base = fds;
-		kvec[1].iov_len = sizeof(int) * res->fds_count;
-
-		ret = kdbus_pool_slice_copy_kvec(entry->slice, off,
-						 kvec, 2, hdr.size);
-		kfree(fds);
-
-		if (ret < 0)
-			return ret;
-	}
-
-	for (i = 0; i < res->data_count; ++i) {
-		struct kdbus_msg_data *d = res->data + i;
-		struct kdbus_memfd m;
-
-		if (d->type != KDBUS_MSG_DATA_MEMFD)
-			continue;
-
-		m.start = d->memfd.start;
-		m.size = d->size;
-		m.fd = -1;
-
-		if (install_fds) {
-			m.fd = get_unused_fd_flags(O_CLOEXEC);
-			if (m.fd < 0) {
-				m.fd = -1;
-				incomplete_fds = true;
-			} else {
-				fd_install(m.fd,
-					   get_file(d->memfd.file));
-			}
-		}
-
-		kvec[0].iov_base = &m;
-		kvec[0].iov_len = sizeof(m);
-
-		ret = kdbus_pool_slice_copy_kvec(entry->slice,
-						 entry->memfd_offset[memfds++],
-						 kvec, 1, sizeof(m));
-		if (ret < 0)
-			return ret;
-	}
-
 	if (incomplete_fds)
 		*return_flags |= KDBUS_RECV_RETURN_INCOMPLETE_FDS;
-
 	return 0;
 }
 
@@ -637,13 +322,13 @@ int kdbus_queue_entry_move(struct kdbus_queue_entry *e,
 	lockdep_assert_held(&src->lock);
 	lockdep_assert_held(&dst->lock);
 
-	if (WARN_ON(IS_ERR(e->user)) || WARN_ON(list_empty(&e->entry)))
+	if (WARN_ON(list_empty(&e->entry)))
 		return -EINVAL;
 	if (src == dst)
 		return 0;
 
 	size = kdbus_pool_slice_size(e->slice);
-	fds = e->msg_res ? e->msg_res->fds_count : 0;
+	fds = e->gaps ? e->gaps->n_fds : 0;
 
 	ret = kdbus_conn_quota_inc(dst, e->user, size, fds);
 	if (ret < 0)
diff --git a/ipc/kdbus/queue.h b/ipc/kdbus/queue.h
index 7f2db96fe308..bf686d182ce1 100644
--- a/ipc/kdbus/queue.h
+++ b/ipc/kdbus/queue.h
@@ -15,6 +15,13 @@
 #ifndef __KDBUS_QUEUE_H
 #define __KDBUS_QUEUE_H
 
+#include <linux/list.h>
+#include <linux/rbtree.h>
+
+struct kdbus_conn;
+struct kdbus_pool_slice;
+struct kdbus_reply;
+struct kdbus_staging;
 struct kdbus_user;
 
 /**
@@ -35,52 +42,37 @@ struct kdbus_queue {
  * @entry:		Entry in the connection's list
  * @prio_node:		Entry in the priority queue tree
  * @prio_entry:		Queue tree node entry in the list of one priority
- * @slice:		Slice in the receiver's pool for the message
- * @attach_flags:	Attach flags used during slice allocation
- * @meta_offset:	Offset of first metadata item in slice
- * @fds_offset:		Offset of FD item in slice
- * @memfd_offset:	Array of slice-offsets for all memfd items
  * @priority:		Message priority
  * @dst_name_id:	The sequence number of the name this message is
  *			addressed to, 0 for messages sent to an ID
- * @msg_res:		Message resources
- * @proc_meta:		Process metadata, captured at message arrival
- * @conn_meta:		Connection metadata, captured at message arrival
- * @reply:		The reply block if a reply to this message is expected
+ * @conn:		Connection this entry is queued on
+ * @gaps:		Gaps object to fill message gaps at RECV time
  * @user:		User used for accounting
+ * @slice:		Slice in the receiver's pool for the message
+ * @reply:		The reply block if a reply to this message is expected
  */
 struct kdbus_queue_entry {
 	struct list_head entry;
 	struct rb_node prio_node;
 	struct list_head prio_entry;
 
-	struct kdbus_pool_slice *slice;
-
-	u64 attach_flags;
-	size_t meta_offset;
-	size_t fds_offset;
-	size_t *memfd_offset;
-
 	s64 priority;
 	u64 dst_name_id;
 
-	struct kdbus_msg_resources *msg_res;
-	struct kdbus_meta_proc *proc_meta;
-	struct kdbus_meta_conn *conn_meta;
-	struct kdbus_reply *reply;
 	struct kdbus_conn *conn;
+	struct kdbus_gaps *gaps;
 	struct kdbus_user *user;
+	struct kdbus_pool_slice *slice;
+	struct kdbus_reply *reply;
 };
 
-struct kdbus_kmsg;
-
 void kdbus_queue_init(struct kdbus_queue *queue);
 struct kdbus_queue_entry *kdbus_queue_peek(struct kdbus_queue *queue,
 					   s64 priority, bool use_priority);
 
-struct kdbus_queue_entry *kdbus_queue_entry_new(struct kdbus_conn *conn_dst,
-						const struct kdbus_kmsg *kmsg,
-						struct kdbus_user *user);
+struct kdbus_queue_entry *kdbus_queue_entry_new(struct kdbus_conn *src,
+						struct kdbus_conn *dst,
+						struct kdbus_staging *s);
 void kdbus_queue_entry_free(struct kdbus_queue_entry *entry);
 int kdbus_queue_entry_install(struct kdbus_queue_entry *entry,
 			      u64 *return_flags, bool install_fds);
diff --git a/ipc/kdbus/reply.c b/ipc/kdbus/reply.c
index 008dca801627..e6791d86ec92 100644
--- a/ipc/kdbus/reply.c
+++ b/ipc/kdbus/reply.c
@@ -37,7 +37,7 @@ struct kdbus_reply *kdbus_reply_new(struct kdbus_conn *reply_src,
 				    bool sync)
 {
 	struct kdbus_reply *r;
-	int ret = 0;
+	int ret;
 
 	if (atomic_inc_return(&reply_dst->request_count) >
 	    KDBUS_CONN_MAX_REQUESTS_PENDING) {
@@ -64,13 +64,11 @@ struct kdbus_reply *kdbus_reply_new(struct kdbus_conn *reply_src,
 		r->waiting = true;
 	}
 
-exit_dec_request_count:
-	if (ret < 0) {
-		atomic_dec(&reply_dst->request_count);
-		return ERR_PTR(ret);
-	}
-
 	return r;
+
+exit_dec_request_count:
+	atomic_dec(&reply_dst->request_count);
+	return ERR_PTR(ret);
 }
 
 static void __kdbus_reply_free(struct kref *kref)
@@ -140,8 +138,7 @@ void kdbus_reply_unlink(struct kdbus_reply *r)
  * @reply:	The reply object
  * @err:	Error code to set on the remote side
  *
- * Remove the synchronous reply object from its connection reply_list, and
- * wake up remote peer (method origin) with the appropriate synchronous reply
+ * Wake up remote peer (method origin) with the appropriate synchronous reply
  * code.
  */
 void kdbus_sync_reply_wakeup(struct kdbus_reply *reply, int err)
@@ -172,17 +169,15 @@ struct kdbus_reply *kdbus_reply_find(struct kdbus_conn *replying,
 				     struct kdbus_conn *reply_dst,
 				     u64 cookie)
 {
-	struct kdbus_reply *r, *reply = NULL;
+	struct kdbus_reply *r;
 
 	list_for_each_entry(r, &reply_dst->reply_list, entry) {
 		if (r->cookie == cookie &&
-		    (!replying || r->reply_src == replying)) {
-			reply = r;
-			break;
-		}
+		    (!replying || r->reply_src == replying))
+			return r;
 	}
 
-	return reply;
+	return NULL;
 }
 
 /**
diff --git a/ipc/kdbus/util.c b/ipc/kdbus/util.c
index eaa806a27997..72b188330896 100644
--- a/ipc/kdbus/util.c
+++ b/ipc/kdbus/util.c
@@ -50,51 +50,6 @@ int kdbus_copy_from_user(void *dest, void __user *user_ptr, size_t size)
 }
 
 /**
- * kdbus_memdup_user() - copy dynamically sized object from user-space
- * @user_ptr:	user-provided source buffer
- * @sz_min:	minimum object size
- * @sz_max:	maximum object size
- *
- * This copies a dynamically sized object from user-space into kernel-space. We
- * require the object to have a 64bit size field at offset 0. We read it out
- * first, allocate a suitably sized buffer and then copy all data.
- *
- * The @sz_min and @sz_max parameters define possible min and max object sizes
- * so user-space cannot trigger un-bound kernel-space allocations.
- *
- * The same alignment-restrictions as described in kdbus_copy_from_user() apply.
- *
- * Return: pointer to dynamically allocated copy, or ERR_PTR() on failure.
- */
-void *kdbus_memdup_user(void __user *user_ptr, size_t sz_min, size_t sz_max)
-{
-	void *ptr;
-	u64 size;
-	int ret;
-
-	ret = kdbus_copy_from_user(&size, user_ptr, sizeof(size));
-	if (ret < 0)
-		return ERR_PTR(ret);
-
-	if (size < sz_min)
-		return ERR_PTR(-EINVAL);
-
-	if (size > sz_max)
-		return ERR_PTR(-EMSGSIZE);
-
-	ptr = memdup_user(user_ptr, size);
-	if (IS_ERR(ptr))
-		return ptr;
-
-	if (*(u64 *)ptr != size) {
-		kfree(ptr);
-		return ERR_PTR(-EINVAL);
-	}
-
-	return ptr;
-}
-
-/**
  * kdbus_verify_uid_prefix() - verify UID prefix of a user-supplied name
  * @name:	user-supplied name to verify
  * @user_ns:	user-namespace to act in
diff --git a/ipc/kdbus/util.h b/ipc/kdbus/util.h
index 740b19880985..529716669fe7 100644
--- a/ipc/kdbus/util.h
+++ b/ipc/kdbus/util.h
@@ -40,7 +40,7 @@
 ({									\
 	u64 __user *_sz =						\
 		(void __user *)((u8 __user *)(_b) + offsetof(_t, _m));	\
-	copy_to_user(_sz, _s, sizeof(((_t *)0)->_m));			\
+	copy_to_user(_sz, _s, FIELD_SIZEOF(_t, _m));			\
 })
 
 /**
@@ -64,7 +64,6 @@ int kdbus_verify_uid_prefix(const char *name, struct user_namespace *user_ns,
 int kdbus_sanitize_attach_flags(u64 flags, u64 *attach_flags);
 
 int kdbus_copy_from_user(void *dest, void __user *user_ptr, size_t size);
-void *kdbus_memdup_user(void __user *user_ptr, size_t sz_min, size_t sz_max);
 
 struct kvec;
author	Karol Lewandowski <k.lewandowsk@samsung.com>	2016-10-06 14:20:24 +0200
committer	Seung-Woo Kim <sw0312.kim@samsung.com>	2016-12-14 13:53:39 +0900
commit	d8934e9b05b82e73a3c78421f372197801c1785e (patch)
tree	24981389ae0ca646a7997f6e32ea7a39aea147b7 /ipc
parent	ea4654706ad61bc7afc47e6321c5cbee8c175688 (diff)