monitor: an interface of acrn-dm

A monitor component will be added to acrn-dm, which crteats socket,
bind and listening at /run/acrn/vmname. Acrnctl & acrnd could conn
-ect to the socket for communication, using defined message, in
include/monitor_msg.h
For each defined message, a message handler callback could be
registered via monitor_add_msg_handler(). On received of a defined
message, a certain call back will be called. Each callback can only
see the message sender's socket-fd.
When acrn-dm want report something, not triggered by incoming message
it can send broadcast message, use monitor_broadcast().

Acked-by: Eddie Dong <eddie.dong@intel.com>
Reviewed-by: Geoffroy Van Cutsem <geoffroy.vancutsem@intel.com>
Reviewed-by: Zhao Yakui <yakui.zhao@intel.com>
Reviewed-by: Yin, Fengwei <fengwei.yin@intel.com>
Signed-off-by: Tao, Yuhong <yuhong.tao@intel.com>

Makefile

@@ -81,6 +81,7 @@ SRCS += hw/acpi/acpi.c
 
 # core
 #SRCS += core/bootrom.c
+SRCS += core/monitor.c
 SRCS += core/sw_load_common.c
 SRCS += core/sw_load_bzimage.c
 SRCS += core/sw_load_vsbl.c

core/main.c

@@ -63,6 +63,7 @@
 #include "rtc.h"
 #include "version.h"
 #include "sw_load.h"
+#include "monitor.h"
 
 #define GUEST_NIO_PORT		0x488	/* guest upcalls via i/o port */
 
@@ -801,6 +802,7 @@ main(int argc, char *argv[])
 
 		vrtc_init(ctx, rtc_localtime);
 		sci_init(ctx);
+		monitor_init(ctx);
 
 		/*
 		 * Exit if a device emulation finds an error in its
@@ -863,6 +865,7 @@ main(int argc, char *argv[])
 		 */
 		mevent_dispatch();
 
+		monitor_close();
 		vm_pause(ctx);
 		fbsdrun_deletecpu(ctx, BSP);
 		vm_unsetup_memory(ctx);

core/monitor.c

@@ -0,0 +1,472 @@
+/*
+ * Project Acrn
+ * Acrn-dm-monitor
+ *
+ * Copyright (C) 2018 Intel Corporation. All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ *   * Redistributions of source code must retain the above copyright
+ *     notice, this list of conditions and the following disclaimer.
+ *   * Redistributions in binary form must reproduce the above copyright
+ *     notice, this list of conditions and the following disclaimer in
+ *     the documentation and/or other materials provided with the
+ *     distribution.
+ *   * Neither the name of Intel Corporation nor the names of its
+ *     contributors may be used to endorse or promote products derived
+ *     from this software without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ *
+ * Author: TaoYuhong <yuhong.tao@intel.com>
+ */
+
+#include <stdio.h>
+#include <unistd.h>
+#include <sys/socket.h>
+#include <sys/un.h>
+#include <pthread.h>
+#include <stdlib.h>
+#include <sys/queue.h>
+#include <errno.h>
+#include <time.h>
+#include "dm.h"
+#include "vmmapi.h"
+#include "mevent.h"
+#include "monitor.h"
+
+/* Data structure and functions for processing received messages */
+struct monitor_msg_handle {
+	struct vmm_msg msg;
+	void (*callback) (struct vmm_msg * msg, struct msg_sender * sender,
+			  void *priv);
+	void *priv;
+	LIST_ENTRY(monitor_msg_handle) list;
+};
+
+static LIST_HEAD(mmh_list_struct, monitor_msg_handle) mmh_head;
+static pthread_mutex_t mmh_mutex = PTHREAD_MUTEX_INITIALIZER;
+static int can_register_handler = 0;	/* Do not allow anyone add his handler, 
+					   untill we have added some researved ones */
+static int monitor_add_handler(struct monitor_msg_handle *handle)
+{
+	struct monitor_msg_handle *hp;
+
+	pthread_mutex_lock(&mmh_mutex);
+
+	LIST_FOREACH(hp, &mmh_head, list)
+	    if (hp->msg.msgid == handle->msg.msgid) {
+		fprintf(stderr, "%s %d\r\n", __FUNCTION__, __LINE__);
+		pthread_mutex_unlock(&mmh_mutex);
+		return -1;
+	}
+	LIST_INSERT_HEAD(&mmh_head, handle, list);
+	pthread_mutex_unlock(&mmh_mutex);
+
+	return 0;
+}
+
+int monitor_register_handler(struct vmm_msg *msg,
+			    void (*callback) (struct vmm_msg * msg,
+					      struct msg_sender * client,
+					      void *priv), void *priv)
+{
+	struct monitor_msg_handle *handle;
+	int ret;
+
+	if (!can_register_handler)
+		return -1;
+
+	handle = calloc(1, sizeof(struct monitor_msg_handle));
+	if (!handle) {
+		fprintf(stderr, "%s %d\r\n", __FUNCTION__, __LINE__);
+		return -1;
+	}
+
+	handle->msg.msgid = msg->msgid;
+	handle->callback = callback;
+	handle->priv = priv;
+
+	ret = monitor_add_handler(handle);
+	if (ret)
+		free(handle);
+
+	return ret;
+}
+
+/* messages handled by monitor */
+static int write_msg_to(int fd, void *data, unsigned long timeout_usec)
+{
+	struct vmm_msg *msg = data;
+	fd_set wfd;
+	struct timeval timeout;
+	int ret = 0;
+
+	if (msg->len < sizeof(struct vmm_msg)) {
+		fprintf(stderr, "%s %d\r\n", __FUNCTION__, __LINE__);
+		return -1;
+	}
+
+	if (msg->msgid > MSGID_MAX) {
+		fprintf(stderr, "%s %d\r\n", __FUNCTION__, __LINE__);
+		return -1;
+	}
+
+	if (msg->magic != VMM_MSG_MAGIC) {
+		fprintf(stderr, "%s %d\r\n", __FUNCTION__, __LINE__);
+		msg->magic = VMM_MSG_MAGIC;
+	}
+
+	msg->timestamp = time(NULL);
+
+	FD_ZERO(&wfd);
+	FD_SET(fd, &wfd);
+	timeout.tv_sec = 0;
+	timeout.tv_usec = timeout_usec;
+	select(fd + 1, NULL, &wfd, NULL, &timeout);
+
+	if (FD_ISSET(fd, &wfd))
+		ret = write(fd, msg, msg->len);
+
+	return ret;
+}
+
+/* MSG_HANDSHAKE, handshake message handler*/
+#define TIMEOUT_USEC	100000
+static VMM_MSG_STR(handshake_badname, "Error: bad name!");
+static VMM_MSG_STR(handshake_ok, "acrn-dm read you request");
+
+static void handshake_acrn_dm(struct vmm_msg *msg, struct msg_sender *sender,
+			      void *priv)
+{
+	struct vmm_msg_handshake *hsk = (void *)msg;
+	int ret;
+
+	ret = strnlen(hsk->name, CLIENT_NAME_LEN);
+	if (ret >= CLIENT_NAME_LEN) {
+		write_msg_to(sender->fd, &handshake_badname, TIMEOUT_USEC);
+		return;
+	}
+
+	strncpy(sender->name, hsk->name, CLIENT_NAME_LEN);
+	sender->broadcast = hsk->broadcast;
+
+	write_msg_to(sender->fd, &handshake_ok, TIMEOUT_USEC);
+}
+
+static struct monitor_msg_handle handle_handshake = {
+	.msg = {.msgid = MSG_HANDSHAKE},
+	.callback = handshake_acrn_dm,
+};
+
+/* vm manager can comunicate with dm-monitor, use unix socket,
+ * the monitor is the server, and there may have many clients,
+ * a client send a message, trigger right msg handler. And msg handler
+ * should only reply to message sender.
+ */
+
+static struct sockaddr_un monitor_addr;	/* one monitor */
+static int monitor_fd;
+
+struct vmm_client {
+	/* msg_sender will be seen/modify by msg handler */
+	struct msg_sender sender;
+
+	/* the rest should be invisible for msg_handler */
+	struct sockaddr_un addr;
+	int fd;
+	socklen_t addr_len;
+	void *buf;
+	int len;		/* buf len */
+	struct mevent *mev;
+	LIST_ENTRY(vmm_client) list;
+};
+
+static LIST_HEAD(client_list_struct, vmm_client) client_head;
+static int num_client = 0;
+static pthread_mutex_t client_mutex = PTHREAD_MUTEX_INITIALIZER;
+
+static void vmm_client_free_res(struct vmm_client *client)
+{
+	mevent_delete(client->mev);
+	close(client->fd);
+	client->fd = -1;
+	free(client->buf);
+	client->buf = NULL;
+	free(client);
+}
+
+static void vmm_client_free(struct vmm_client *client)
+{
+	pthread_mutex_lock(&client_mutex);
+	LIST_REMOVE(client, list);
+	num_client--;
+	pthread_mutex_unlock(&client_mutex);
+
+	vmm_client_free_res(client);
+}
+
+static VMM_MSG_STR(unsupported_msgid, "Error: unsupported msgid!");
+
+static int monitor_parse_buf(struct vmm_client *client)
+{
+	struct vmm_msg *msg;
+	struct monitor_msg_handle *handle;
+	size_t p = 0;
+	int handled = 0;
+
+	if (client->len < sizeof(struct vmm_msg))
+		return -1;
+	do {
+		msg = client->buf + p;
+
+		/* do we out-of-bounary? */
+		if (p + msg->len > client->len) {
+			fprintf(stderr, "%s %d\r\n", __FUNCTION__, __LINE__);
+			break;
+		}
+
+		LIST_FOREACH(handle, &mmh_head, list) {
+			if (msg->magic != VMM_MSG_MAGIC)
+				return -1;
+			if (handle->msg.msgid != msg->msgid)
+				continue;
+			client->sender.fd = client->fd;
+			handle->callback(msg, &client->sender, handle->priv);
+			handled = 1;
+			break;
+		}
+		p += msg->len;
+	} while (p < client->len);
+
+	if (!handled)
+		write_msg_to(client->fd, &unsupported_msgid, TIMEOUT_USEC);
+
+	return 0;
+}
+
+static void mevent_read_func(int fd, enum ev_type type, void *param)
+{
+	struct vmm_client *client = param;
+
+	client->len = read(fd, client->buf, VMM_MSG_MAX_LEN);
+	if (client->len <= 0) {
+		fprintf(stderr, "Disconnect(%d)!\r\n", client->fd);
+		vmm_client_free(client);
+		return;
+	}
+
+	if (client->len == VMM_MSG_MAX_LEN) {
+		fprintf(stderr, "TODO: buf overflow!\r\n");
+		return;
+	}
+
+	monitor_parse_buf(client);
+}
+
+static struct vmm_client *vmm_client_new(void)
+{
+	struct vmm_client *client;
+
+	client = calloc(1, sizeof(struct vmm_client));
+	if (!client) {
+		fprintf(stderr, "%s %d\r\n", __FUNCTION__, __LINE__);
+		goto alloc_client;
+	}
+	memset(client, 0, sizeof(struct vmm_client));
+
+	client->buf = calloc(1, VMM_MSG_MAX_LEN);
+	if (!client->buf) {
+		fprintf(stderr, "%s %d\r\n", __FUNCTION__, __LINE__);
+		goto alloc_buf;
+	}
+
+	client->addr_len = sizeof(client->addr);
+	client->fd =
+	    accept(monitor_fd, (struct sockaddr *)&client->addr, &client->addr_len);
+	if (client->fd < 0) {
+		fprintf(stderr, "%s %d\r\n", __FUNCTION__, __LINE__);
+		goto accept_con;
+	}
+
+	client->mev =
+	    mevent_add(client->fd, EVF_READ, mevent_read_func, client);
+	if (!client->mev) {
+		fprintf(stderr, "%s %d\r\n", __FUNCTION__, __LINE__);
+		goto add_mev;
+	}
+
+	pthread_mutex_lock(&client_mutex);
+	LIST_INSERT_HEAD(&client_head, client, list);
+	num_client++;
+	pthread_mutex_unlock(&client_mutex);
+
+	return client;
+
+ add_mev:
+	close(client->fd);
+	client->fd = -1;
+ accept_con:
+	free(client->buf);
+	client->buf = NULL;
+ alloc_buf:
+	free(client);
+ alloc_client:
+	return NULL;
+}
+
+int monitor_broadcast(struct vmm_msg *msg)
+{
+	struct vmm_client *client;
+	fd_set wfd;
+	int max_fd = 0;
+	struct timeval timeout;
+	int ret = 0;
+
+	if (msg->len < sizeof(struct vmm_msg)) {
+		fprintf(stderr, "%s %d\r\n", __FUNCTION__, __LINE__);
+		return -1;
+	}
+
+	if (msg->msgid > MSGID_MAX) {
+		fprintf(stderr, "%s %d\r\n", __FUNCTION__, __LINE__);
+		return -1;
+	}
+
+	if (msg->magic != VMM_MSG_MAGIC) {
+		fprintf(stderr, "%s %d\r\n", __FUNCTION__, __LINE__);
+		msg->magic = VMM_MSG_MAGIC;
+	}
+
+	msg->timestamp = time(NULL);
+
+	pthread_mutex_lock(&client_mutex);
+
+	FD_ZERO(&wfd);
+	LIST_FOREACH(client, &client_head, list) {
+		if (!client->sender.broadcast)
+			continue;
+		FD_SET(client->fd, &wfd);
+		if (client->fd > max_fd)
+			max_fd = client->fd;
+	}
+	timeout.tv_sec = 0;
+	timeout.tv_usec = 10000;
+	select(max_fd + 1, NULL, &wfd, NULL, &timeout);
+
+	LIST_FOREACH(client, &client_head, list) {
+		if (!client->sender.broadcast)
+			continue;
+		if (FD_ISSET(client->fd, &wfd)) {
+			ret = write(client->fd, msg->payload,
+				    msg->len - sizeof(struct vmm_msg));
+			if (ret < 0)
+				continue;
+		}
+	}
+
+	pthread_mutex_unlock(&client_mutex);
+	return 0;
+}
+
+/* monitor thread */
+static int monitor_running = 1;
+static pthread_t monitor_thread;
+
+static void *monitor_server_func(void *arg)
+{
+	struct vmm_client *client;
+	while (monitor_running) {
+		client = vmm_client_new();
+		if (!client) {
+			usleep(10000);
+			continue;
+		}
+		fprintf(stderr, "Connected:%d\r\n", client->fd);
+	}
+
+	fprintf(stderr, "%s quit!\r\n", __FUNCTION__);
+	return NULL;
+}
+
+int monitor_init(struct vmctx *ctx)
+{
+	int ret;
+	char path[128] = { };
+
+	ret = system("mkdir -p /run/acrn/");
+	if (ret) {
+		fprintf(stderr, "%s %d\r\n", __FUNCTION__, __LINE__);
+		goto socket_err;
+	}
+	memset(&monitor_addr, 0, sizeof(monitor_addr));
+	snprintf(path, sizeof(path), "/run/acrn/%s-monitor.socket", vmname);
+	unlink(path);
+	monitor_fd = socket(AF_UNIX, SOCK_STREAM, 0);
+	if (monitor_fd < 0) {
+		fprintf(stderr, "%s %d\r\n", __FUNCTION__, __LINE__);
+		goto socket_err;
+	}
+
+	monitor_addr.sun_family = AF_UNIX;
+	strncpy(monitor_addr.sun_path, path, sizeof(monitor_addr.sun_path));
+	ret = bind(monitor_fd, (struct sockaddr *)&monitor_addr, sizeof(monitor_addr));
+	if (ret < 0) {
+		fprintf(stderr, "%s %d\r\n", __FUNCTION__, __LINE__);
+		goto bind_err;
+	}
+
+	listen(monitor_fd, 1);
+	ret = pthread_create(&monitor_thread, NULL, monitor_server_func, NULL);
+	if (ret) {
+		fprintf(stderr, "%s %d\r\n", __FUNCTION__, __LINE__);
+		goto thread_err;
+	}
+
+	/* Messages handled by monitor */
+	monitor_add_handler(&handle_handshake);
+
+	__sync_fetch_and_add(&can_register_handler, 1);
+	return 0;
+
+ thread_err:
+	monitor_thread = 0;
+	unlink(path);
+ bind_err:
+	close(monitor_fd);
+ socket_err:
+	return -1;
+}
+
+void monitor_close(void)
+{
+	struct vmm_client *client;
+	if (!monitor_thread)
+		return;
+	shutdown(monitor_fd, SHUT_RDWR);
+	close(monitor_fd);
+	monitor_running = 0;
+	pthread_join(monitor_thread, NULL);
+	unlink(monitor_addr.sun_path);
+
+	/* client buf-mem and fd may be still in use by msg-handler */
+	/* which is handled by mevent */
+	pthread_mutex_lock(&client_mutex);
+	LIST_FOREACH(client, &client_head, list) {
+		vmm_client_free_res(client);
+	}
+	pthread_mutex_unlock(&client_mutex);
+}

include/monitor.h

@@ -0,0 +1,82 @@
+/*
+ * Project Acrn
+ * Acrn-dm-monitor
+ *
+ * Copyright (C) 2018 Intel Corporation. All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ *   * Redistributions of source code must retain the above copyright
+ *     notice, this list of conditions and the following disclaimer.
+ *   * Redistributions in binary form must reproduce the above copyright
+ *     notice, this list of conditions and the following disclaimer in
+ *     the documentation and/or other materials provided with the
+ *     distribution.
+ *   * Neither the name of Intel Corporation nor the names of its
+ *     contributors may be used to endorse or promote products derived
+ *     from this software without specific prior written permission.
+ *
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ *
+ * Author: TaoYuhong <yuhong.tao@intel.com>
+ */
+
+/* acrn-dm monitor APIS */
+
+#ifndef MONITOR_H
+#define MONITOR_H
+
+#include "monitor_msg.h"
+
+int monitor_init(struct vmctx *ctx);
+void monitor_close(void);
+
+/**
+ * monitor broadcast()
+ * Developer can use monitor_broadcast() inside acrn-dm, send vmm_msg to all client.
+ * @arguements:
+ * @msg: any valid vmm_msg data structure, which you want send
+ */
+
+int monitor_broadcast(struct vmm_msg *msg);
+
+/* msg_sender will be seen/modify by msg handler */
+struct msg_sender {
+	int fd;			/* msg handler can replay to this fd */
+	char name[CLIENT_NAME_LEN];	/* client have a chance to name itsself */
+	int broadcast;
+};
+
+/**
+ * monitor_register_handler()
+ * Developer can add vmm_msg handler inside acrn-dm, that means, when a client send a
+ * vmm_msg to acrn-dm, the correspoding callback will be called
+ * @arguements:
+ * @msg: msg->msgid must be set, for which handler will be add.
+ * @callback: when a received message match msg->msgid, callback will be envoked.
+ *   And these data are pass in to help developer: (a)msg, the received message, from
+ *   socket. (b)sender, tell you who send this message, anything wite to sender->fd
+ *   will be able to read out by client socket. Developer should only write valid
+ *   vmm_msg. (c)priv, that is what you pass to monitor_add_msg_handler();
+ * @priv, callback will see this value.
+*/
+
+int monitor_register_handler(struct vmm_msg *msg,
+			     void (*callback) (struct vmm_msg * msg,
+					       struct msg_sender * sender,
+					       void *priv), void *priv);
+#endif

include/monitor_msg.h

@@ -0,0 +1,107 @@
+/*
+ * Project Acrn
+ * Acrn-dm-monitor
+ *
+ * Copyright (C) 2018 Intel Corporation. All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ *
+ *   * Redistributions of source code must retain the above copyright
+ *     notice, this list of conditions and the following disclaimer.
+ *   * Redistributions in binary form must reproduce the above copyright
+ *     notice, this list of conditions and the following disclaimer in
+ *     the documentation and/or other materials provided with the
+ *     distribution.
+ *   * Neither the name of Intel Corporation nor the names of its
+ *     contributors may be used to endorse or promote products derived
+ *     from this software without specific prior written permission.
+ *
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ *
+ * Author: TaoYuhong <yuhong.tao@intel.com>
+ */
+
+/* this file will be shared, by anyone who want talk to acrn-dm */
+#ifndef MONITOR_MSG_H
+#define MONITOR_MSG_H
+
+enum msgid {
+	REQ_STARTALL = 0,	/* SOS lifecycle service -> VM Mngr */
+	REQ_PAUSEALL,		/* SOS lifecycle service -> VM Mngr */
+	REQ_START,		/* VM Mngr -> ACRN-DM(vm) */
+	REQ_STOP,		/* VM Mngr -> ACRN-DM(vm) */
+	REQ_PAUSE,
+	REQ_RESUME,		/* VM Mngr -> ACRN-DM(vm) */
+	REQ_RESET,
+	REQ_QUERY,
+	NTF_ALLSTOPPED,		/* VM Mngr -> SOS lifecycle service */
+	NTF_ALLPAUSED,		/* VM Mngr -> SOS lifecycle service */
+	NTF_STARTED,		/* ACRN-DM -> VM Mngr */
+	NTF_STOPPED,		/* ACRN-DM -> VM Mngr */
+	NTF_PAUSED,		/* ACRN-DM -> VM Mngr */
+	NTF_RESUMED,		/* ACRN-DM -> VM Mngr */
+
+	MSG_STR,
+	MSG_HANDSHAKE,		/* handshake */
+
+	MSGID_MAX
+};
+
+#define VMM_MSG_MAGIC	0x67736d206d6d76	/* that is char[8] "vmm msg", on X86 */
+#define VMM_MSG_MAX_LEN		4096
+
+struct vmm_msg {
+	unsigned long long magic;	/* Make sure you get a vmm_msg */
+	unsigned int msgid;
+	unsigned long timestamp;
+	size_t len;		/* vmm_msg + payload size */
+	char payload[0];
+};
+
+/* practical messages, and helpers
+ *  each message defined in msgid should have its own data structure,
+ *  and shared by acrn-dm and vm-mngr. So that such that message data 
+ *  structure can be recognized by both sides of the communication
+ */
+
+/* For test, generate a message who carry a string
+ * eg., VMM_MSG_STR(hello_msg, "Hello\n") will create hello_msg,
+ * then you can write(sock_fd, hello_msg, sizeof(hello_msg))
+ */
+#define VMM_MSG_STR(var, str)	\
+struct vmm_msg_##var {		\
+struct vmm_msg vmsg;		\
+        char raw[sizeof(str)];		\
+} var = {				\
+        .vmsg = {			\
+                .magic = VMM_MSG_MAGIC,	\
+                .msgid = MSG_STR,	\
+                .len = sizeof(struct vmm_msg_##var),	\
+        },						\
+        .raw = str,				\
+}
+
+#define CLIENT_NAME_LEN	16
+struct vmm_msg_handshake {
+	struct vmm_msg vmsg;
+	char name[CLIENT_NAME_LEN];	/* name should be a string, end with '\0' */
+	/* can be "acrnd", "vm-mngr" or "acrnctl" */
+	int broadcast;		/* if set, allow acrn-dm send broadcast */
+	/*   message to such client */
+};
+
+#endif