initial commit

This patch bootstrap the new nft-sync software. Basically, this
software aims to support two different setups:

1) Rule-set repository server. The software serves the nft rule-set to
   clients that request the ruleset.

   Basically from the system that acts as repository, you have to run:

   # nft-sync -c ../contrib/nft-sync.conf.server

   Then, from the client:

   # nft-sync -c ../contrib/nft-sync.conf.client --fetch

   Which displays the nft rule-set in the standard output, so you
   can inspect the nft rule-set.

   Alternatively, the client can also retrieve and apply the nft
   rule-set using the pull command instead:

   # nft-sync -c ../contrib/nft-sync.conf.client --pull

   [ Note that this command above does not work in this bootstrap yet ]

2) Rule-set synchronization: In case of primary-backup and multiprimary
   firewall configurations, the software makes sure that the firewall
   cluster is deploying the same filtering policy. In this case, you have
   to launch the process:

   # nft-sync -c ../contrib/nft-sync.conf --sync

   [ Note that this command above does not work in this bootstrap yet ]

This bootstrap provides the basic infrastructure as a proof-of-concept.
Many of the necessary features are still lacking:

* Implement --sync and --pull commands.

* Interaction with nft through libnftnl, which allows the software to
  retrieve the local nft rule-set, as well as to parse it and apply it.

* SSL support, specifically the repository mode needs it to make sure
  nobody can steal your filtering policy from the network.

* IPv6 support.

* Allow to serve different rule-sets in the repository mode.

And many others that will be added progressively.

Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>

1 files changed, 295 insertions, 0 deletions

diff --git a/src/tcp.c b/src/tcp.c
new file mode 100644
index 0000000..ced350a
--- /dev/null
+++ b/src/tcp.c
@@ -0,0 +1,295 @@
+/*
+ * (C) 2014 by Pablo Neira Ayuso <pablo@netfilter.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU Affero General Public License as published by
+ * the Free Software Foundation; either version 3 of the License, or
+ * (at your option) any later version.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <arpa/inet.h>
+#include <unistd.h>
+#include <string.h>
+#include <sys/ioctl.h>
+#include <net/if.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <limits.h>
+
+#include <tcp.h>
+#include <timer.h>
+
+#include "logging.h"
+
+/*
+ * TCP server side
+ */
+
+struct tcp_server {
+	int fd;
+	union {
+		struct sockaddr_in ipv4;
+		struct sockaddr_in6 ipv6;
+	} addr;
+};
+
+#define TCP_SERVER_LISTEN 20
+
+struct tcp_server *tcp_server_create(struct tcp_conf *conf)
+{
+	int ret, on = 1;
+	struct tcp_server *c;
+	socklen_t socklen = sizeof(int);
+
+	c = calloc(1, sizeof(struct tcp_server));
+	if (c == NULL)
+		return NULL;
+
+	switch (conf->ipproto) {
+	case AF_INET:
+		c->addr.ipv4.sin_family = AF_INET;
+		c->addr.ipv4.sin_port = htons(conf->port);
+		c->addr.ipv4.sin_addr = conf->server.ipv4.inet_addr;
+		socklen = sizeof(struct sockaddr_in);
+		break;
+
+	case AF_INET6:
+		c->addr.ipv6.sin6_family = AF_INET6;
+		c->addr.ipv6.sin6_port = htons(conf->port);
+		c->addr.ipv6.sin6_addr = conf->server.ipv6.inet_addr6;
+		c->addr.ipv6.sin6_scope_id = conf->server.ipv6.scope_id;
+		socklen = sizeof(struct sockaddr_in6);
+		break;
+	}
+
+	c->fd = socket(conf->ipproto, SOCK_STREAM, 0);
+	if (c->fd < 0)
+		goto err1;
+
+	ret = setsockopt(c->fd, SOL_SOCKET, SO_REUSEADDR, &on, sizeof(int));
+	if (ret < 0)
+		goto err2;
+
+	ret = setsockopt(c->fd, SOL_SOCKET, SO_KEEPALIVE, &on, sizeof(int));
+	if (ret < 0)
+		goto err2;
+
+	ret = bind(c->fd, (struct sockaddr *) &c->addr, socklen);
+	if (ret < 0)
+		goto err2;
+
+	ret = listen(c->fd, TCP_SERVER_LISTEN);
+	if (ret < 0)
+		goto err2;
+
+	ret = fcntl(c->fd, F_SETFL, O_NONBLOCK);
+	if (ret < 0)
+		goto err2;
+
+	return c;
+err2:
+	close(c->fd);
+err1:
+	free(c);
+	return NULL;
+}
+
+void tcp_server_destroy(struct tcp_server *c)
+{
+	close(c->fd);
+	free(c);
+}
+
+int tcp_server_get_fd(struct tcp_server *c)
+{
+	return c->fd;
+}
+
+int tcp_server_accept(struct tcp_server *c, struct sockaddr_in *addr)
+{
+	int err, fd;
+	socklen_t socklen = sizeof(struct sockaddr_in);
+
+	err = accept(c->fd, (struct sockaddr *)addr, &socklen);
+	if (err < 0 && errno != EAGAIN)
+		return -1;
+
+	fd = err;
+
+	err = fcntl(fd, F_SETFL, O_NONBLOCK);
+	if (err < 0) {
+		close(fd);
+		return -1;
+	}
+
+	return fd;
+}
+
+/*
+ * TCP client side
+ */
+
+enum tcp_client_state {
+	TCP_DISCONNECTED	= 0,
+	TCP_CONNECTING,
+	TCP_CONNECTED
+};
+
+struct tcp_client {
+	int			fd;
+	enum tcp_client_state	state;
+	union {
+		struct sockaddr_in ipv4;
+		struct sockaddr_in6 ipv6;
+	} addr;
+	socklen_t		socklen;
+	struct nft_timer	timer;
+	void			*data;
+};
+
+#define TCP_CONNECT_TIMEOUT	1
+
+static int tcp_client_init(struct tcp_client *c, struct tcp_conf *conf)
+{
+	int ret = 0;
+
+	c->fd = socket(conf->ipproto, SOCK_STREAM, 0);
+	if (c->fd < 0)
+		return -1;
+
+	switch (conf->ipproto) {
+	case AF_INET:
+		c->addr.ipv4.sin_family = AF_INET;
+		c->addr.ipv4.sin_port = htons(conf->port);
+		c->addr.ipv4.sin_addr = conf->client.inet_addr;
+		c->socklen = sizeof(struct sockaddr_in);
+		break;
+	case AF_INET6:
+		c->addr.ipv6.sin6_family = AF_INET6;
+		c->addr.ipv6.sin6_port = htons(conf->port);
+		c->addr.ipv6.sin6_addr = conf->client.inet_addr6;
+		c->socklen = sizeof(struct sockaddr_in6);
+		break;
+	default:
+		ret = -1;
+		break;
+	}
+
+	if (ret < 0)
+		goto err1;
+
+	ret = fcntl(c->fd, F_SETFL, O_NONBLOCK);
+	if (ret < 0)
+		goto err1;
+
+	ret = connect(c->fd, (struct sockaddr *)&c->addr, c->socklen);
+	if (ret < 0) {
+		switch (errno) {
+		case EINPROGRESS:
+			c->state = TCP_CONNECTING;
+			break;
+		default: /* ECONNREFUSED */
+			c->state = TCP_DISCONNECTED;
+			goto err1;
+		}
+	} else {
+		/* very unlikely at this stage. */
+		c->state = TCP_CONNECTED;
+	}
+	return 0;
+err1:
+	close(c->fd);
+	return ret;
+}
+
+int tcp_client_get_fd(struct tcp_client *c)
+{
+	return c->fd;
+}
+
+struct tcp_client *tcp_client_create(struct tcp_conf *conf)
+{
+	struct tcp_client *c;
+
+	c = calloc(1, sizeof(struct tcp_client));
+	if (c == NULL)
+		return NULL;
+
+	if (tcp_client_init(c, conf) < 0) {
+		free(c);
+		return NULL;
+	}
+
+	return c;
+}
+
+void tcp_client_destroy(struct tcp_client *c)
+{
+	close(c->fd);
+	free(c);
+}
+
+ssize_t tcp_client_send(struct tcp_client *c, const void *data, int size)
+{
+	ssize_t ret = 0;
+
+	switch (c->state) {
+	case TCP_DISCONNECTED:
+		ret = -1;
+		break;
+	case TCP_CONNECTING:
+		ret = connect(c->fd, (struct sockaddr *)&c->addr, c->socklen);
+		if (ret < 0)
+			return ret;
+
+		c->state = TCP_CONNECTED;
+		/* fall through ... */
+	case TCP_CONNECTED:
+		ret = send(c->fd, data, size, 0);
+		if (ret <= 0) {
+			/* errno == EPIPE || errno == ECONNRESET */
+			c->state = TCP_DISCONNECTED;
+			return ret;
+		}
+		break;
+	}
+	return ret;
+}
+
+ssize_t tcp_client_recv(struct tcp_client *c, void *data, int size)
+{
+	ssize_t ret = 0;
+
+	switch (c->state) {
+	case TCP_DISCONNECTED:
+		ret = -1;
+		break;
+	case TCP_CONNECTING:
+		ret = connect(c->fd, (struct sockaddr *)&c->addr, c->socklen);
+		if (ret < 0)
+			return ret;
+
+		c->state = TCP_CONNECTED;
+		/* fall through ... */
+	case TCP_CONNECTED:
+		ret = recv(c->fd, data, size, 0);
+		if (ret <= 0) {
+			/* errno == ENOTCONN */
+			c->state = TCP_DISCONNECTED;
+			return ret;
+		}
+	}
+	return ret;
+}
+
+void tcp_client_set_data(struct tcp_client *c, void *data)
+{
+	c->data = data;
+}
+
+void *tcp_client_get_data(struct tcp_client *c)
+{
+	return c->data;
+}