pesto, conf: Have pesto connect to passt and check versions

Start implementing pesto in earnest.  Create a control/configuration
socket in passt.  Have pesto connect to it and retrieve a server greeting
Perform some basic version checking.

Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
[sbrivio: Avoid potential recursive calling between conf_accept() and
 conf_close(), reported by clang-tidy]
[sbrivio: In conf(), check we're not exceeding sizeof(c->control_path)
 instead of sizeof(c->socket_path), and, in pesto's main(), print
 argv[optind] instead of argv[1] to indicate an invalid socket path,
 both reported by Jon Maloy]
[sbrivio: In pesto's main(), drop unnecessary newline from error
 message, reported by Laurent]
[sbrivio: Don't use SOCK_NONBLOCK on accept4(), as that only applies
 to the *new* file descriptor, which we don't want -- set O_NONBLOCK
 on the listening file descriptor using fcntl()]
[sbrivio: Switch to protocol version 1, and reflect the true magic
 behind pesto, i.e. basil, into the magic string]
[sbrivio: Fix conflicts in the Makefile caused by the fact that I'm
 not merging a previous series reworking it]
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>

10 files changed, 275 insertions, 4 deletions

diff --git a/Makefile b/Makefile
index 2639472..b1003d8 100644
--- a/Makefile
+++ b/Makefile
@@ -45,7 +45,7 @@ PASST_SRCS = arch.c arp.c bitmap.c checksum.c conf.c dhcp.c dhcpv6.c \
 	vhost_user.c virtio.c vu_common.c
 QRAP_SRCS = qrap.c
 PASST_REPAIR_SRCS = passt-repair.c
-PESTO_SRCS = pesto.c
+PESTO_SRCS = pesto.c serialise.c
 SRCS = $(PASST_SRCS) $(QRAP_SRCS) $(PASST_REPAIR_SRCS) $(PESTO_SRCS)
 
 MANPAGES = passt.1 pasta.1 pesto.1 qrap.1 passt-repair.1
diff --git a/conf.c b/conf.c
index 05e93db..e830368 100644
--- a/conf.c
+++ b/conf.c
@@ -48,6 +48,10 @@
 #include "isolation.h"
 #include "log.h"
 #include "vhost_user.h"
+#include "epoll_ctl.h"
+#include "conf.h"
+#include "pesto.h"
+#include "serialise.h"
 
 #define NETNS_RUN_DIR	"/run/netns"
 
@@ -543,6 +547,7 @@ static void usage(const char *name, FILE *f, int status)
 		"  --runas UID|UID:GID 	Run as given UID, GID, which can be\n"
 		"    numeric, or login and group names\n"
 		"    default: drop to user \"nobody\"\n"
+		"  -c, --conf-path PATH	Configuration socket path\n"
 		"  -h, --help		Display this help message and exit\n"
 		"  --version		Show version and exit\n");
 
@@ -781,6 +786,9 @@ static void conf_print(const struct ctx *c)
 	char buf[INANY_ADDRSTRLEN];
 	int i;
 
+	if (c->fd_control_listen >= 0)
+		info("Configuration socket: %s", c->control_path);
+
 	if (c->ifi4 > 0 || c->ifi6 > 0) {
 		char ifn[IFNAMSIZ];
 
@@ -1074,6 +1082,19 @@ static void conf_open_files(struct ctx *c)
 		if (c->pidfile_fd < 0)
 			die_perror("Couldn't open PID file %s", c->pidfile);
 	}
+
+	c->fd_control = -1;
+	if (*c->control_path) {
+		c->fd_control_listen = sock_unix(c->control_path);
+		if (c->fd_control_listen < 0) {
+			die_perror("Couldn't open control socket %s",
+				   c->control_path);
+		}
+		if (fcntl(c->fd_control_listen, F_SETFL, O_NONBLOCK))
+			die_perror("Couldn't set O_NONBLOCK on control socket");
+	} else {
+		c->fd_control_listen = -1;
+	}
 }
 
 /**
@@ -1110,6 +1131,25 @@ fail:
 }
 
 /**
+ * conf_sock_listen() - Start listening for connections on configuration socket
+ * @c:		Execution context
+ */
+static void conf_sock_listen(const struct ctx *c)
+{
+	union epoll_ref ref = { .type = EPOLL_TYPE_CONF_LISTEN };
+
+	if (c->fd_control_listen < 0)
+		return;
+
+	if (listen(c->fd_control_listen, 0))
+		die_perror("Couldn't listen on configuration socket");
+
+	ref.fd = c->fd_control_listen;
+	if (epoll_add(c->epollfd, EPOLLIN | EPOLLET, ref))
+		die_perror("Couldn't add configuration socket to epoll");
+}
+
+/**
  * conf() - Process command-line arguments and set configuration
  * @c:		Execution context
  * @argc:	Argument count
@@ -1191,9 +1231,10 @@ void conf(struct ctx *c, int argc, char **argv)
 		{"migrate-exit", no_argument,		NULL,		29 },
 		{"migrate-no-linger", no_argument,	NULL,		30 },
 		{"stats", required_argument,		NULL,		31 },
+		{"conf-path",	required_argument,	NULL,		'c' },
 		{ 0 },
 	};
-	const char *optstring = "+dqfel:hs:F:I:p:P:m:a:n:M:g:i:o:D:S:H:461t:u:T:U:";
+	const char *optstring = "+dqfel:hs:c:F:I:p:P:m:a:n:M:g:i:o:D:S:H:461t:u:T:U:";
 	const char *logname = (c->mode == MODE_PASTA) ? "pasta" : "passt";
 	bool opt_t = false, opt_T = false, opt_u = false, opt_U = false;
 	char userns[PATH_MAX] = { 0 }, netns[PATH_MAX] = { 0 };
@@ -1451,6 +1492,13 @@ void conf(struct ctx *c, int argc, char **argv)
 
 			c->fd_tap = -1;
 			break;
+		case 'c':
+			ret = snprintf(c->control_path, sizeof(c->control_path),
+				       "%s", optarg);
+			if (ret <= 0 || ret >= (int)sizeof(c->control_path))
+				die("Invalid configuration path: %s", optarg);
+			c->fd_control_listen = c->fd_control = -1;
+			break;
 		case 'F':
 			errno = 0;
 			fd_tap_opt = strtol(optarg, NULL, 0);
@@ -1873,6 +1921,136 @@ void conf(struct ctx *c, int argc, char **argv)
 			fwd_rule_parse('U', "auto", c->fwd[PIF_SPLICE]);
 	}
 
+	conf_sock_listen(c);
+
 	if (!c->quiet)
 		conf_print(c);
 }
+
+static void conf_accept(struct ctx *c);
+
+/**
+ * conf_close() - Close configuration / control socket and clean up
+ * @c:		Execution context
+ */
+static void conf_close(struct ctx *c)
+{
+	debug("Closing configuration socket");
+	epoll_ctl(c->epollfd, EPOLL_CTL_DEL, c->fd_control, NULL);
+	close(c->fd_control);
+	c->fd_control = -1;
+}
+
+/**
+ * conf_listen_handler() - Handle events on configuration listening socket
+ * @c:		Execution context
+ * @events:	epoll events
+ */
+void conf_listen_handler(struct ctx *c, uint32_t events)
+{
+	if (events != EPOLLIN) {
+		err("Unexpected event 0x%04x on configuration socket", events);
+		return;
+	}
+
+	if (c->fd_control >= 0) {
+		/* Ignore the new connection for now, blocking it until the
+		 * current one finishes.
+		 */
+		return;
+	}
+
+	conf_accept(c);
+}
+
+/**
+ * conf_accept() - Accept a new control connection
+ * @c:		Execution context
+ */
+static void conf_accept(struct ctx *c)
+{
+	struct pesto_hello hello = {
+		.magic = PESTO_SERVER_MAGIC,
+		.version = htonl(PESTO_PROTOCOL_VERSION),
+	};
+	union epoll_ref ref = { .type = EPOLL_TYPE_CONF };
+	struct ucred uc = { 0 };
+	socklen_t len = sizeof(uc);
+	int fd, rc;
+
+retry:
+	fd = accept4(c->fd_control_listen, NULL, NULL, SOCK_CLOEXEC);
+	if (fd < 0) {
+		if (errno != EAGAIN)
+			warn_perror("accept4() on configuration listening socket");
+		return;
+	}
+
+	if (getsockopt(fd, SOL_SOCKET, SO_PEERCRED, &uc, &len) < 0)
+		warn_perror("Can't get configuration client credentials");
+
+	c->fd_control = ref.fd = fd;
+	rc = epoll_add(c->epollfd, EPOLLIN | EPOLLET, ref);
+	if (rc < 0) {
+		warn_perror("epoll_ctl() on configuration socket");
+		goto fail;
+	}
+
+	rc = write_all_buf(fd, &hello, sizeof(hello));
+	if (rc < 0) {
+		warn_perror("Error writing configuration protocol hello");
+		goto fail;
+	}
+
+	info("Accepted configuration client, PID %i", uc.pid);
+	if (!PESTO_PROTOCOL_VERSION) {
+		warn(
+"Warning: Using experimental unsupported configuration protocol");
+	}
+
+	return;
+
+fail:
+	conf_close(c);
+	goto retry;
+}
+
+/**
+ * conf_handler() - Handle events on configuration socket
+ * @c:		Execution context
+ * @events:	epoll events
+ */
+void conf_handler(struct ctx *c, uint32_t events)
+{
+	if (events & EPOLLIN) {
+		char discard[BUFSIZ];
+		ssize_t n;
+
+		do {
+			n = read(c->fd_control, discard, sizeof(discard));
+			if (n > 0)
+				debug("Discarded %zd bytes of config data", n);
+		} while (n > 0);
+		if (n == 0) {
+			debug("Configuration client EOF");
+			goto close;
+		}
+		if (errno != EAGAIN && errno != EWOULDBLOCK) {
+			err_perror("Error reading config data");
+			goto close;
+		}
+	}
+
+	if (events & EPOLLHUP) {
+		debug("Configuration client hangup");
+		goto close;
+	}
+
+	return;
+
+close:
+	conf_close(c);
+
+	/* Check if any other clients are waiting to connect */
+	conf_accept(c);
+}
diff --git a/conf.h b/conf.h
index b45ad74..16f9718 100644
--- a/conf.h
+++ b/conf.h
@@ -8,5 +8,7 @@
 
 enum passt_modes conf_mode(int argc, char *argv[]);
 void conf(struct ctx *c, int argc, char **argv);
+void conf_listen_handler(struct ctx *c, uint32_t events);
+void conf_handler(struct ctx *c, uint32_t events);
 
 #endif /* CONF_H */
diff --git a/epoll_type.h b/epoll_type.h
index a90ffb6..061325a 100644
--- a/epoll_type.h
+++ b/epoll_type.h
@@ -46,6 +46,10 @@ enum epoll_type {
 	EPOLL_TYPE_REPAIR,
 	/* Netlink neighbour subscription socket */
 	EPOLL_TYPE_NL_NEIGH,
+	/* Configuration listening socket */
+	EPOLL_TYPE_CONF_LISTEN,
+	/* Configuration socket */
+	EPOLL_TYPE_CONF,
 
 	EPOLL_NUM_TYPES,
 };
diff --git a/passt.1 b/passt.1
index 6303aeb..908fd4a 100644
--- a/passt.1
+++ b/passt.1
@@ -128,6 +128,11 @@ initial effective UID 0 or CAP_SETUID capability) to work.
 Default is to change to user \fInobody\fR if started as root.
 
 .TP
+.BR \-c ", " \-\-conf-path " " \fIpath " " (EXPERIMENTAL)
+Path for configuration and control socket used by \fBpesto\fR(1) to
+dynamically update passt or pasta's configuration.
+
+.TP
 .BR \-h ", " \-\-help
 Display a help message and exit.
 
diff --git a/passt.c b/passt.c
index f84419c..bc42ea3 100644
--- a/passt.c
+++ b/passt.c
@@ -80,6 +80,8 @@ char *epoll_type_str[] = {
 	[EPOLL_TYPE_REPAIR_LISTEN]	= "TCP_REPAIR helper listening socket",
 	[EPOLL_TYPE_REPAIR]		= "TCP_REPAIR helper socket",
 	[EPOLL_TYPE_NL_NEIGH]		= "netlink neighbour notifier socket",
+	[EPOLL_TYPE_CONF_LISTEN]	= "configuration listening socket",
+	[EPOLL_TYPE_CONF]		= "configuration socket",
 };
 static_assert(ARRAY_SIZE(epoll_type_str) == EPOLL_NUM_TYPES,
 	      "epoll_type_str[] doesn't match enum epoll_type");
@@ -303,6 +305,12 @@ static void passt_worker(void *opaque, int nfds, struct epoll_event *events)
 		case EPOLL_TYPE_NL_NEIGH:
 			nl_neigh_notify_handler(c);
 			break;
+		case EPOLL_TYPE_CONF_LISTEN:
+			conf_listen_handler(c, eventmask);
+			break;
+		case EPOLL_TYPE_CONF:
+			conf_handler(c, eventmask);
+			break;
 		default:
 			/* Can't happen */
 			assert(0);
diff --git a/passt.h b/passt.h
index 62b8dcd..b3f049d 100644
--- a/passt.h
+++ b/passt.h
@@ -158,6 +158,7 @@ struct ip6_ctx {
  * @foreground:		Run in foreground, don't log to stderr by default
  * @nofile:		Maximum number of open files (ulimit -n)
  * @sock_path:		Path for UNIX domain socket
+ * @control_path:	Path for control/configuration UNIX domain socket
  * @repair_path:	TCP_REPAIR helper path, can be "none", empty for default
  * @pcap:		Path for packet capture file
  * @pidfile:		Path to PID file, empty string if not configured
@@ -169,6 +170,8 @@ struct ip6_ctx {
  * @epollfd:		File descriptor for epoll instance
  * @fd_tap_listen:	File descriptor for listening AF_UNIX socket, if any
  * @fd_tap:		AF_UNIX socket, tuntap device, or pre-opened socket
+ * @fd_control_listen:	Listening control/configuration socket, if any
+ * @fd_control:		Control/configuration socket, if any
  * @fd_repair_listen:	File descriptor for listening TCP_REPAIR socket, if any
  * @fd_repair:		Connected AF_UNIX socket for TCP_REPAIR helper
  * @our_tap_mac:	Pasta/passt's MAC on the tap link
@@ -223,6 +226,7 @@ struct ctx {
 	int foreground;
 	int nofile;
 	char sock_path[UNIX_PATH_MAX];
+	char control_path[UNIX_PATH_MAX];
 	char repair_path[UNIX_PATH_MAX];
 	char pcap[PATH_MAX];
 
@@ -240,6 +244,8 @@ struct ctx {
 	int epollfd;
 	int fd_tap_listen;
 	int fd_tap;
+	int fd_control_listen;
+	int fd_control;
 	int fd_repair_listen;
 	int fd_repair;
 	unsigned char our_tap_mac[ETH_ALEN];
diff --git a/pesto.c b/pesto.c
index f0916e8..ab476c5 100644
--- a/pesto.c
+++ b/pesto.c
@@ -33,6 +33,7 @@
 
 #include "common.h"
 #include "seccomp_pesto.h"
+#include "serialise.h"
 #include "pesto.h"
 #include "log.h"
 
@@ -66,6 +67,8 @@ static void usage(const char *name, FILE *f, int status)
  *
  * Return: 0 on success, won't return on failure
  *
+ * #syscalls:pesto socket s390x:socketcall i686:socketcall
+ * #syscalls:pesto connect shutdown close
  * #syscalls:pesto exit_group fstat read write
  */
 int main(int argc, char **argv)
@@ -76,9 +79,12 @@ int main(int argc, char **argv)
 		{"version",	no_argument,		NULL,		1 },
 		{ 0 },
 	};
+	struct sockaddr_un a = { AF_UNIX, "" };
 	const char *optstring = "dh";
+	struct pesto_hello hello;
 	struct sock_fprog prog;
-	int optname;
+	int optname, ret, s;
+	uint32_t s_version;
 
 	prctl(PR_SET_DUMPABLE, 0);
 
@@ -122,5 +128,42 @@ int main(int argc, char **argv)
 
 	debug("debug_flag=%d, path=\"%s\"", debug_flag, argv[optind]);
 
-	die("pesto is not implemented yet");
+	if ((s = socket(AF_UNIX, SOCK_STREAM, 0)) < 0)
+		die_perror("Failed to create AF_UNIX socket");
+
+	ret = snprintf(a.sun_path, sizeof(a.sun_path), "%s", argv[optind]);
+	if (ret <= 0 || ret >= (int)sizeof(a.sun_path))
+		die("Invalid socket path \"%s\"", argv[optind]);
+
+	ret = connect(s, (struct sockaddr *)&a, sizeof(a));
+	if (ret < 0) {
+		die_perror("Failed to connect to %s", a.sun_path);
+	}
+
+	ret = read_all_buf(s, &hello, sizeof(hello));
+	if (ret < 0)
+		die_perror("Couldn't read server greeting");
+
+	if (memcmp(hello.magic, PESTO_SERVER_MAGIC, sizeof(hello.magic)))
+		die("Bad magic number from server");
+
+	s_version = ntohl(hello.version);
+
+	if (s_version > PESTO_PROTOCOL_VERSION) {
+		die("Unknown server protocol version %"PRIu32" > %"PRIu32,
+		    s_version, PESTO_PROTOCOL_VERSION);
+	}
+
+	/* cppcheck-suppress knownConditionTrueFalse */
+	if (!s_version) {
+		if (PESTO_PROTOCOL_VERSION)
+			die("Unsupported experimental server protocol");
+		FPRINTF(stderr,
+"Warning: Using experimental protocol version, client and server must match\n");
+	}
+
+	if (shutdown(s, SHUT_RDWR) < 0 || close(s) < 0)
+		die_perror("Error shutting down control socket");
+
+	exit(0);
 }
diff --git a/pesto.h b/pesto.h
index e9b329f..3c93d3e 100644
--- a/pesto.h
+++ b/pesto.h
@@ -9,4 +9,26 @@
 #ifndef PESTO_H
 #define PESTO_H
 
+#include <assert.h>
+#include <stdint.h>
+
+#define PESTO_SERVER_MAGIC	"basil:s"
+
+/* Version 0 is reserved for unreleased / unsupported experimental versions */
+#define PESTO_PROTOCOL_VERSION	1
+
+/**
+ * struct pesto_hello - Server introduction message
+ * @magic:	PESTO_SERVER_MAGIC
+ * @version:	Version number
+ */
+struct pesto_hello {
+	char magic[8];
+	uint32_t version;
+} __attribute__ ((__packed__));
+
+static_assert(sizeof(PESTO_SERVER_MAGIC)
+	      == sizeof(((struct pesto_hello *)0)->magic),
+	      "PESTO_SERVER_MAGIC has wrong size");
+
 #endif /* PESTO_H */
diff --git a/serialise.c b/serialise.c
index 944e741..346df99 100644
--- a/serialise.c
+++ b/serialise.c
@@ -6,6 +6,9 @@
  * PASTA - Pack A Subtle Tap Abstraction
  *  for network namespace/tap device mode
  *
+ * PESTO - Programmable Extensible Socket Translation Orchestrator
+ *  front-end for passt(1) and pasta(1) forwarding configuration
+ *
  * serialise.c - Serialisation of data structures over bytestreams
  *
  * Copyright Red Hat