From 0515adceaa8f69a1d85ae3c8c550c37dd49b0c47 Mon Sep 17 00:00:00 2001 From: Stefano Brivio Date: Mon, 7 Feb 2022 21:11:37 +0100 Subject: passt, pasta: Namespace-based sandboxing, defer seccomp policy application To reach (at least) a conceptually equivalent security level as implemented by --enable-sandbox in slirp4netns, we need to create a new mount namespace and pivot_root() into a new (empty) mountpoint, so that passt and pasta can't access any filesystem resource after initialisation. While at it, also detach IPC, PID (only for passt, to prevent vulnerabilities based on the knowledge of a target PID), and UTS namespaces. With this approach, if we apply the seccomp filters right after the configuration step, the number of allowed syscalls grows further. To prevent this, defer the application of seccomp policies after the initialisation phase, before the main loop, that's where we expect bad things to happen, potentially. This way, we get back to 22 allowed syscalls for passt and 34 for pasta, on x86_64. While at it, move #syscalls notes to specific code paths wherever it conceptually makes sense. We have to open all the file handles we'll ever need before sandboxing: - the packet capture file can only be opened once, drop instance numbers from the default path and use the (pre-sandbox) PID instead - /proc/net/tcp{,v6} and /proc/net/udp{,v6}, for automatic detection of bound ports in pasta mode, are now opened only once, before sandboxing, and their handles are stored in the execution context - the UNIX domain socket for passt is also bound only once, before sandboxing: to reject clients after the first one, instead of closing the listening socket, keep it open, accept and immediately discard new connection if we already have a valid one Clarify the (unchanged) behaviour for --netns-only in the man page. To actually make passt and pasta processes run in a separate PID namespace, we need to unshare(CLONE_NEWPID) before forking to background (if configured to do so). Introduce a small daemon() implementation, __daemon(), that additionally saves the PID file before forking. While running in foreground, the process itself can't move to a new PID namespace (a process can't change the notion of its own PID): mention that in the man page. For some reason, fork() in a detached PID namespace causes SIGTERM and SIGQUIT to be ignored, even if the handler is still reported as SIG_DFL: add a signal handler that just exits. We can now drop most of the pasta_child_handler() implementation, that took care of terminating all processes running in the same namespace, if pasta started a shell: the shell itself is now the init process in that namespace, and all children will terminate once the init process exits. Issuing 'echo $$' in a detached PID namespace won't return the actual namespace PID as seen from the init namespace: adapt demo and test setup scripts to reflect that. Signed-off-by: Stefano Brivio --- util.c | 129 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++-------- 1 file changed, 113 insertions(+), 16 deletions(-) (limited to 'util.c') diff --git a/util.c b/util.c index 94d49a6..e9fca3b 100644 --- a/util.c +++ b/util.c @@ -16,6 +16,7 @@ #include #include #include +#include #include #include #include @@ -23,6 +24,7 @@ #include #include #include +#include #include #include #include @@ -32,6 +34,8 @@ #include #include +#include + #include "util.h" #include "passt.h" @@ -431,31 +435,51 @@ char *line_read(char *buf, size_t len, int fd) /** * procfs_scan_listen() - Set bits for listening TCP or UDP sockets from procfs - * @name: Corresponding name of file under /proc/net/ + * @proto: IPPROTO_TCP or IPPROTO_UDP + * @ip_version: IP version, V4 or V6 + * @ns: Use saved file descriptors for namespace if set * @map: Bitmap where numbers of ports in listening state will be set * @exclude: Bitmap of ports to exclude from setting (and clear) + * + * #syscalls:pasta lseek ppc64le:_llseek ppc64:_llseek */ -void procfs_scan_listen(char *name, uint8_t *map, uint8_t *exclude) +void procfs_scan_listen(struct ctx *c, uint8_t proto, int ip_version, int ns, + uint8_t *map, uint8_t *exclude) { - char line[BUFSIZ], path[PATH_MAX]; + char line[BUFSIZ], *path; unsigned long port; unsigned int state; - int fd; + int *fd; - snprintf(path, PATH_MAX, "/proc/net/%s", name); - if ((fd = open(path, O_RDONLY)) < 0) + if (proto == IPPROTO_TCP) { + fd = &c->proc_net_tcp[ip_version][ns]; + if (ip_version == V4) + path = "/proc/net/tcp"; + else + path = "/proc/net/tcp6"; + } else { + fd = &c->proc_net_udp[ip_version][ns]; + if (ip_version == V4) + path = "/proc/net/udp"; + else + path = "/proc/net/udp6"; + } + + if (*fd != -1) + lseek(*fd, 0, SEEK_SET); + else if ((*fd = open(path, O_RDONLY)) < 0) return; *line = 0; - line_read(line, sizeof(line), fd); - while (line_read(line, sizeof(line), fd)) { + line_read(line, sizeof(line), *fd); + while (line_read(line, sizeof(line), *fd)) { /* NOLINTNEXTLINE(cert-err34-c): != 2 if conversion fails */ if (sscanf(line, "%*u: %*x:%lx %*x:%*x %x", &port, &state) != 2) continue; /* See enum in kernel's include/net/tcp_states.h */ - if ((strstr(name, "tcp") && state != 0x0a) || - (strstr(name, "udp") && state != 0x07)) + if ((proto == IPPROTO_TCP && state != 0x0a) || + (proto == IPPROTO_UDP && state != 0x07)) continue; if (bitmap_isset(exclude, port)) @@ -463,25 +487,98 @@ void procfs_scan_listen(char *name, uint8_t *map, uint8_t *exclude) else bitmap_set(map, port); } +} - close(fd); +/** + * drop_caps() - Drop capabilities we might have except for CAP_NET_BIND_SERVICE + */ +void drop_caps(void) +{ + int i; + + for (i = 0; i < 64; i++) { + if (i == CAP_NET_BIND_SERVICE) + continue; + + prctl(PR_CAPBSET_DROP, i, 0, 0, 0); + } } /** - * ns_enter() - Enter configured network and user namespaces + * ns_enter() - Enter configured user (unless already joined) and network ns * @c: Execution context * - * Return: 0 on success, -1 on failure + * Return: 0, won't return on failure * * #syscalls:pasta setns */ int ns_enter(struct ctx *c) { - if (!c->netns_only && setns(c->pasta_userns_fd, CLONE_NEWUSER)) - return -errno; + if (!c->netns_only && + c->pasta_userns_fd != -1 && + setns(c->pasta_userns_fd, CLONE_NEWUSER)) + exit(EXIT_FAILURE); if (setns(c->pasta_netns_fd, CLONE_NEWNET)) - return -errno; + exit(EXIT_FAILURE); + + return 0; +} + +/** + * pid_file() - Write PID to file, if requested to do so, and close it + * @fd: Open PID file descriptor, closed on exit, -1 to skip writing it + * @pid: PID value to write + */ +void write_pidfile(int fd, pid_t pid) { + char pid_buf[12]; + int n; + + if (fd == -1) + return; + + n = snprintf(pid_buf, sizeof(pid_buf), "%i\n", pid); + + if (write(fd, pid_buf, n) < 0) { + perror("PID file write"); + exit(EXIT_FAILURE); + } + + close(fd); +} + +/** + * __daemon() - daemon()-like function writing PID file before parent exits + * @pidfile_fd: Open PID file descriptor + * @devnull_fd: Open file descriptor for /dev/null + * + * Return: child PID on success, won't return on failure + */ +int __daemon(int pidfile_fd, int devnull_fd) +{ + pid_t pid = fork(); + + if (pid == -1) { + perror("fork"); + exit(EXIT_FAILURE); + } + + if (pid) { + write_pidfile(pidfile_fd, pid); + exit(EXIT_SUCCESS); + } + + errno = 0; + + setsid(); + + dup2(devnull_fd, STDIN_FILENO); + dup2(devnull_fd, STDOUT_FILENO); + dup2(devnull_fd, STDERR_FILENO); + close(devnull_fd); + + if (errno) + exit(EXIT_FAILURE); return 0; } -- cgit v1.2.3