[mirror_qemu.git] / net / filter-rewriter.c

/*
 * Copyright (c) 2016 HUAWEI TECHNOLOGIES CO., LTD.
 * Copyright (c) 2016 FUJITSU LIMITED
 * Copyright (c) 2016 Intel Corporation
 *
 * Author: Zhang Chen <zhangchen.fnst@cn.fujitsu.com>
 *
 * This work is licensed under the terms of the GNU GPL, version 2 or
 * later.  See the COPYING file in the top-level directory.
 */

#include "qemu/osdep.h"
#include "trace.h"
#include "net/colo.h"
#include "net/filter.h"
#include "net/net.h"
#include "qemu-common.h"
#include "qapi/error.h"
#include "qapi/qmp/qerror.h"
#include "qapi-visit.h"
#include "qom/object.h"
#include "qemu/main-loop.h"
#include "qemu/iov.h"
#include "net/checksum.h"

#define FILTER_COLO_REWRITER(obj) \
    OBJECT_CHECK(RewriterState, (obj), TYPE_FILTER_REWRITER)

#define TYPE_FILTER_REWRITER "filter-rewriter"

typedef struct RewriterState {
    NetFilterState parent_obj;
    NetQueue *incoming_queue;
    /* hashtable to save connection */
    GHashTable *connection_track_table;
} RewriterState;

static void filter_rewriter_flush(NetFilterState *nf)
{
    RewriterState *s = FILTER_COLO_REWRITER(nf);

    if (!qemu_net_queue_flush(s->incoming_queue)) {
        /* Unable to empty the queue, purge remaining packets */
        qemu_net_queue_purge(s->incoming_queue, nf->netdev);
    }
}

/*
 * Return 1 on success, if return 0 means the pkt
 * is not TCP packet
 */
static int is_tcp_packet(Packet *pkt)
{
    if (!parse_packet_early(pkt) &&
        pkt->ip->ip_p == IPPROTO_TCP) {
        return 1;
    } else {
        return 0;
    }
}

/* handle tcp packet from primary guest */
static int handle_primary_tcp_pkt(NetFilterState *nf,
                                  Connection *conn,
                                  Packet *pkt)
{
    struct tcphdr *tcp_pkt;

    tcp_pkt = (struct tcphdr *)pkt->transport_header;
    if (trace_event_get_state(TRACE_COLO_FILTER_REWRITER_DEBUG)) {
        trace_colo_filter_rewriter_pkt_info(__func__,
                    inet_ntoa(pkt->ip->ip_src), inet_ntoa(pkt->ip->ip_dst),
                    ntohl(tcp_pkt->th_seq), ntohl(tcp_pkt->th_ack),
                    tcp_pkt->th_flags);
        trace_colo_filter_rewriter_conn_offset(conn->offset);
    }

    if (((tcp_pkt->th_flags & (TH_ACK | TH_SYN)) == TH_SYN)) {
        /*
         * we use this flag update offset func
         * run once in independent tcp connection
         */
        conn->syn_flag = 1;
    }

    if (((tcp_pkt->th_flags & (TH_ACK | TH_SYN)) == TH_ACK)) {
        if (conn->syn_flag) {
            /*
             * offset = secondary_seq - primary seq
             * ack packet sent by guest from primary node,
             * so we use th_ack - 1 get primary_seq
             */
            conn->offset -= (ntohl(tcp_pkt->th_ack) - 1);
            conn->syn_flag = 0;
        }
        /* handle packets to the secondary from the primary */
        tcp_pkt->th_ack = htonl(ntohl(tcp_pkt->th_ack) + conn->offset);

        net_checksum_calculate((uint8_t *)pkt->data, pkt->size);
    }

    return 0;
}

/* handle tcp packet from secondary guest */
static int handle_secondary_tcp_pkt(NetFilterState *nf,
                                    Connection *conn,
                                    Packet *pkt)
{
    struct tcphdr *tcp_pkt;

    tcp_pkt = (struct tcphdr *)pkt->transport_header;

    if (trace_event_get_state(TRACE_COLO_FILTER_REWRITER_DEBUG)) {
        trace_colo_filter_rewriter_pkt_info(__func__,
                    inet_ntoa(pkt->ip->ip_src), inet_ntoa(pkt->ip->ip_dst),
                    ntohl(tcp_pkt->th_seq), ntohl(tcp_pkt->th_ack),
                    tcp_pkt->th_flags);
        trace_colo_filter_rewriter_conn_offset(conn->offset);
    }

    if (((tcp_pkt->th_flags & (TH_ACK | TH_SYN)) == (TH_ACK | TH_SYN))) {
        /*
         * save offset = secondary_seq and then
         * in handle_primary_tcp_pkt make offset
         * = secondary_seq - primary_seq
         */
        conn->offset = ntohl(tcp_pkt->th_seq);
    }

    if ((tcp_pkt->th_flags & (TH_ACK | TH_SYN)) == TH_ACK) {
        /* handle packets to the primary from the secondary*/
        tcp_pkt->th_seq = htonl(ntohl(tcp_pkt->th_seq) - conn->offset);

        net_checksum_calculate((uint8_t *)pkt->data, pkt->size);
    }

    return 0;
}

static ssize_t colo_rewriter_receive_iov(NetFilterState *nf,
                                         NetClientState *sender,
                                         unsigned flags,
                                         const struct iovec *iov,
                                         int iovcnt,
                                         NetPacketSent *sent_cb)
{
    RewriterState *s = FILTER_COLO_REWRITER(nf);
    Connection *conn;
    ConnectionKey key;
    Packet *pkt;
    ssize_t size = iov_size(iov, iovcnt);
    char *buf = g_malloc0(size);

    iov_to_buf(iov, iovcnt, 0, buf, size);
    pkt = packet_new(buf, size);
    g_free(buf);

    /*
     * if we get tcp packet
     * we will rewrite it to make secondary guest's
     * connection established successfully
     */
    if (pkt && is_tcp_packet(pkt)) {

        fill_connection_key(pkt, &key);

        if (sender == nf->netdev) {
            /*
             * We need make tcp TX and RX packet
             * into one connection.
             */
            reverse_connection_key(&key);
        }
        conn = connection_get(s->connection_track_table,
                              &key,
                              NULL);

        if (sender == nf->netdev) {
            /* NET_FILTER_DIRECTION_TX */
            if (!handle_primary_tcp_pkt(nf, conn, pkt)) {
                qemu_net_queue_send(s->incoming_queue, sender, 0,
                (const uint8_t *)pkt->data, pkt->size, NULL);
                packet_destroy(pkt, NULL);
                pkt = NULL;
                /*
                 * We block the packet here,after rewrite pkt
                 * and will send it
                 */
                return 1;
            }
        } else {
            /* NET_FILTER_DIRECTION_RX */
            if (!handle_secondary_tcp_pkt(nf, conn, pkt)) {
                qemu_net_queue_send(s->incoming_queue, sender, 0,
                (const uint8_t *)pkt->data, pkt->size, NULL);
                packet_destroy(pkt, NULL);
                pkt = NULL;
                /*
                 * We block the packet here,after rewrite pkt
                 * and will send it
                 */
                return 1;
            }
        }
    }

    packet_destroy(pkt, NULL);
    pkt = NULL;
    return 0;
}

static void colo_rewriter_cleanup(NetFilterState *nf)
{
    RewriterState *s = FILTER_COLO_REWRITER(nf);

    /* flush packets */
    if (s->incoming_queue) {
        filter_rewriter_flush(nf);
        g_free(s->incoming_queue);
    }
}

static void colo_rewriter_setup(NetFilterState *nf, Error **errp)
{
    RewriterState *s = FILTER_COLO_REWRITER(nf);

    s->connection_track_table = g_hash_table_new_full(connection_key_hash,
                                                      connection_key_equal,
                                                      g_free,
                                                      connection_destroy);
    s->incoming_queue = qemu_new_net_queue(qemu_netfilter_pass_to_next, nf);
}

static void colo_rewriter_class_init(ObjectClass *oc, void *data)
{
    NetFilterClass *nfc = NETFILTER_CLASS(oc);

    nfc->setup = colo_rewriter_setup;
    nfc->cleanup = colo_rewriter_cleanup;
    nfc->receive_iov = colo_rewriter_receive_iov;
}

static const TypeInfo colo_rewriter_info = {
    .name = TYPE_FILTER_REWRITER,
    .parent = TYPE_NETFILTER,
    .class_init = colo_rewriter_class_init,
    .instance_size = sizeof(RewriterState),
};

static void register_types(void)
{
    type_register_static(&colo_rewriter_info);
}

type_init(register_types);
Commit	Line	Data
e6eee8ab ZC	1	/*
	2	* Copyright (c) 2016 HUAWEI TECHNOLOGIES CO., LTD.
	3	* Copyright (c) 2016 FUJITSU LIMITED
	4	* Copyright (c) 2016 Intel Corporation
	5	*
	6	* Author: Zhang Chen <zhangchen.fnst@cn.fujitsu.com>
	7	*
	8	* This work is licensed under the terms of the GNU GPL, version 2 or
	9	* later. See the COPYING file in the top-level directory.
	10	*/
	11
	12	#include "qemu/osdep.h"
30656b09	13	#include "trace.h"
e6eee8ab ZC	14	#include "net/colo.h"
	15	#include "net/filter.h"
	16	#include "net/net.h"
	17	#include "qemu-common.h"
	18	#include "qapi/error.h"
	19	#include "qapi/qmp/qerror.h"
	20	#include "qapi-visit.h"
	21	#include "qom/object.h"
	22	#include "qemu/main-loop.h"
	23	#include "qemu/iov.h"
	24	#include "net/checksum.h"
	25
	26	#define FILTER_COLO_REWRITER(obj) \
	27	OBJECT_CHECK(RewriterState, (obj), TYPE_FILTER_REWRITER)
	28
	29	#define TYPE_FILTER_REWRITER "filter-rewriter"
	30
	31	typedef struct RewriterState {
	32	NetFilterState parent_obj;
	33	NetQueue *incoming_queue;
	34	/* hashtable to save connection */
	35	GHashTable *connection_track_table;
	36	} RewriterState;
	37
	38	static void filter_rewriter_flush(NetFilterState *nf)
	39	{
	40	RewriterState *s = FILTER_COLO_REWRITER(nf);
	41
	42	if (!qemu_net_queue_flush(s->incoming_queue)) {
	43	/* Unable to empty the queue, purge remaining packets */
	44	qemu_net_queue_purge(s->incoming_queue, nf->netdev);
	45	}
	46	}
	47
afe46124 ZC	48	/*
	49	* Return 1 on success, if return 0 means the pkt
	50	* is not TCP packet
	51	*/
	52	static int is_tcp_packet(Packet *pkt)
	53	{
	54	if (!parse_packet_early(pkt) &&
	55	pkt->ip->ip_p == IPPROTO_TCP) {
	56	return 1;
	57	} else {
	58	return 0;
	59	}
	60	}
	61
30656b09 ZC	62	/* handle tcp packet from primary guest */
	63	static int handle_primary_tcp_pkt(NetFilterState *nf,
	64	Connection *conn,
	65	Packet *pkt)
	66	{
	67	struct tcphdr *tcp_pkt;
	68
	69	tcp_pkt = (struct tcphdr *)pkt->transport_header;
	70	if (trace_event_get_state(TRACE_COLO_FILTER_REWRITER_DEBUG)) {
2061c14c ZC	71	trace_colo_filter_rewriter_pkt_info(__func__,
2061c14c ZC	72	inet_ntoa(pkt->ip->ip_src), inet_ntoa(pkt->ip->ip_dst),
30656b09 ZC	73	ntohl(tcp_pkt->th_seq), ntohl(tcp_pkt->th_ack),
	74	tcp_pkt->th_flags);
	75	trace_colo_filter_rewriter_conn_offset(conn->offset);
30656b09 ZC	76	}
	77
	78	if (((tcp_pkt->th_flags & (TH_ACK \| TH_SYN)) == TH_SYN)) {
	79	/*
	80	* we use this flag update offset func
	81	* run once in independent tcp connection
	82	*/
	83	conn->syn_flag = 1;
	84	}
	85
	86	if (((tcp_pkt->th_flags & (TH_ACK \| TH_SYN)) == TH_ACK)) {
	87	if (conn->syn_flag) {
	88	/*
	89	* offset = secondary_seq - primary seq
	90	* ack packet sent by guest from primary node,
	91	* so we use th_ack - 1 get primary_seq
	92	*/
	93	conn->offset -= (ntohl(tcp_pkt->th_ack) - 1);
	94	conn->syn_flag = 0;
	95	}
	96	/* handle packets to the secondary from the primary */
	97	tcp_pkt->th_ack = htonl(ntohl(tcp_pkt->th_ack) + conn->offset);
	98
	99	net_checksum_calculate((uint8_t *)pkt->data, pkt->size);
	100	}
	101
	102	return 0;
	103	}
	104
	105	/* handle tcp packet from secondary guest */
	106	static int handle_secondary_tcp_pkt(NetFilterState *nf,
	107	Connection *conn,
	108	Packet *pkt)
	109	{
	110	struct tcphdr *tcp_pkt;
	111
	112	tcp_pkt = (struct tcphdr *)pkt->transport_header;
	113
	114	if (trace_event_get_state(TRACE_COLO_FILTER_REWRITER_DEBUG)) {
2061c14c ZC	115	trace_colo_filter_rewriter_pkt_info(__func__,
2061c14c ZC	116	inet_ntoa(pkt->ip->ip_src), inet_ntoa(pkt->ip->ip_dst),
30656b09 ZC	117	ntohl(tcp_pkt->th_seq), ntohl(tcp_pkt->th_ack),
	118	tcp_pkt->th_flags);
	119	trace_colo_filter_rewriter_conn_offset(conn->offset);
30656b09 ZC	120	}
	121
	122	if (((tcp_pkt->th_flags & (TH_ACK \| TH_SYN)) == (TH_ACK \| TH_SYN))) {
	123	/*
	124	* save offset = secondary_seq and then
	125	* in handle_primary_tcp_pkt make offset
	126	* = secondary_seq - primary_seq
	127	*/
	128	conn->offset = ntohl(tcp_pkt->th_seq);
	129	}
	130
	131	if ((tcp_pkt->th_flags & (TH_ACK \| TH_SYN)) == TH_ACK) {
	132	/* handle packets to the primary from the secondary*/
	133	tcp_pkt->th_seq = htonl(ntohl(tcp_pkt->th_seq) - conn->offset);
	134
	135	net_checksum_calculate((uint8_t *)pkt->data, pkt->size);
	136	}
	137
	138	return 0;
	139	}
	140
e6eee8ab ZC	141	static ssize_t colo_rewriter_receive_iov(NetFilterState *nf,
	142	NetClientState *sender,
	143	unsigned flags,
	144	const struct iovec *iov,
	145	int iovcnt,
	146	NetPacketSent *sent_cb)
	147	{
afe46124 ZC	148	RewriterState *s = FILTER_COLO_REWRITER(nf);
	149	Connection *conn;
	150	ConnectionKey key;
	151	Packet *pkt;
	152	ssize_t size = iov_size(iov, iovcnt);
	153	char *buf = g_malloc0(size);
	154
	155	iov_to_buf(iov, iovcnt, 0, buf, size);
	156	pkt = packet_new(buf, size);
2061c14c	157	g_free(buf);
afe46124	158
e6eee8ab ZC	159	/*
	160	* if we get tcp packet
	161	* we will rewrite it to make secondary guest's
	162	* connection established successfully
	163	*/
afe46124 ZC	164	if (pkt && is_tcp_packet(pkt)) {
	165
	166	fill_connection_key(pkt, &key);
	167
	168	if (sender == nf->netdev) {
	169	/*
	170	* We need make tcp TX and RX packet
	171	* into one connection.
	172	*/
	173	reverse_connection_key(&key);
	174	}
	175	conn = connection_get(s->connection_track_table,
	176	&key,
	177	NULL);
	178
	179	if (sender == nf->netdev) {
	180	/* NET_FILTER_DIRECTION_TX */
30656b09 ZC	181	if (!handle_primary_tcp_pkt(nf, conn, pkt)) {
	182	qemu_net_queue_send(s->incoming_queue, sender, 0,
	183	(const uint8_t *)pkt->data, pkt->size, NULL);
	184	packet_destroy(pkt, NULL);
	185	pkt = NULL;
	186	/*
	187	* We block the packet here,after rewrite pkt
	188	* and will send it
	189	*/
	190	return 1;
	191	}
afe46124 ZC	192	} else {
afe46124 ZC	193	/* NET_FILTER_DIRECTION_RX */
30656b09 ZC	194	if (!handle_secondary_tcp_pkt(nf, conn, pkt)) {
	195	qemu_net_queue_send(s->incoming_queue, sender, 0,
	196	(const uint8_t *)pkt->data, pkt->size, NULL);
	197	packet_destroy(pkt, NULL);
	198	pkt = NULL;
	199	/*
	200	* We block the packet here,after rewrite pkt
	201	* and will send it
	202	*/
	203	return 1;
	204	}
afe46124 ZC	205	}
	206	}
	207
	208	packet_destroy(pkt, NULL);
	209	pkt = NULL;
e6eee8ab ZC	210	return 0;
	211	}
	212
	213	static void colo_rewriter_cleanup(NetFilterState *nf)
	214	{
	215	RewriterState *s = FILTER_COLO_REWRITER(nf);
	216
	217	/* flush packets */
	218	if (s->incoming_queue) {
	219	filter_rewriter_flush(nf);
	220	g_free(s->incoming_queue);
	221	}
	222	}
	223
	224	static void colo_rewriter_setup(NetFilterState nf, Error *errp)
	225	{
	226	RewriterState *s = FILTER_COLO_REWRITER(nf);
	227
	228	s->connection_track_table = g_hash_table_new_full(connection_key_hash,
	229	connection_key_equal,
	230	g_free,
	231	connection_destroy);
	232	s->incoming_queue = qemu_new_net_queue(qemu_netfilter_pass_to_next, nf);
	233	}
	234
	235	static void colo_rewriter_class_init(ObjectClass oc, void data)
	236	{
	237	NetFilterClass *nfc = NETFILTER_CLASS(oc);
	238
	239	nfc->setup = colo_rewriter_setup;
	240	nfc->cleanup = colo_rewriter_cleanup;
	241	nfc->receive_iov = colo_rewriter_receive_iov;
	242	}
	243
	244	static const TypeInfo colo_rewriter_info = {
	245	.name = TYPE_FILTER_REWRITER,
	246	.parent = TYPE_NETFILTER,
	247	.class_init = colo_rewriter_class_init,
	248	.instance_size = sizeof(RewriterState),
	249	};
	250
	251	static void register_types(void)
	252	{
	253	type_register_static(&colo_rewriter_info);
	254	}
	255
	256	type_init(register_types);