#include <sys/un.h>
#include <sys/socket.h>
#include <linux/netlink.h>
-#include <linux/rtnetlink.h>
#include <sys/eventfd.h>
#include <net/if.h>
+#include <linux/veth.h>
+#include <sys/personality.h>
#ifdef HAVE_SELINUX
#include <selinux/selinux.h>
#endif
+#ifdef HAVE_SECCOMP
+#include <seccomp.h>
+#endif
+
#include "sd-daemon.h"
#include "sd-bus.h"
#include "sd-id128.h"
#include "rtnl-util.h"
#include "udev-util.h"
+#ifdef HAVE_SECCOMP
+#include "seccomp-util.h"
+#endif
+
typedef enum LinkJournal {
LINK_NO,
LINK_AUTO,
static bool arg_register = true;
static bool arg_keep_unit = false;
static char **arg_network_interfaces = NULL;
+static bool arg_network_veth = false;
+static char *arg_network_bridge = NULL;
+static unsigned long arg_personality = 0xffffffffLU;
static int help(void) {
"Spawn a minimal namespace container for debugging, testing and building.\n\n"
" -h --help Show this help\n"
" --version Print version string\n"
+ " -q --quiet Do not show status information\n"
" -D --directory=NAME Root directory for the container\n"
" -b --boot Boot up full system (i.e. invoke init)\n"
" -u --user=USER Run the command under specified user or uid\n"
- " --uuid=UUID Set a specific machine UUID for the container\n"
" -M --machine=NAME Set the machine name for the container\n"
+ " --uuid=UUID Set a specific machine UUID for the container\n"
" -S --slice=SLICE Place the container in the specified slice\n"
+ " --private-network Disable network in container\n"
+ " --network-interface=INTERFACE\n"
+ " Assign an existing network interface to the\n"
+ " container\n"
+ " --network-veth Add a virtual ethernet connection between host\n"
+ " and container\n"
+ " --network-bridge=INTERFACE\n"
+ " Add a virtual ethernet connection between host\n"
+ " and container and add it to an existing bridge on\n"
+ " the host\n"
" -Z --selinux-context=SECLABEL\n"
" Set the SELinux security context to be used by\n"
" processes in the container\n"
" -L --selinux-apifs-context=SECLABEL\n"
" Set the SELinux security context to be used by\n"
" API/tmpfs file systems in the container\n"
- " --private-network Disable network in container\n"
- " --network-interface=INTERFACE\n"
- " Assign an existing network interface to the container\n"
- " --share-system Share system namespaces with host\n"
- " --read-only Mount the root directory read-only\n"
" --capability=CAP In addition to the default, retain specified\n"
" capability\n"
" --drop-capability=CAP Drop the specified capability from the default set\n"
" --link-journal=MODE Link up guest journal, one of no, auto, guest, host\n"
" -j Equivalent to --link-journal=host\n"
+ " --read-only Mount the root directory read-only\n"
" --bind=PATH[:PATH] Bind mount a file or directory from the host into\n"
" the container\n"
" --bind-ro=PATH[:PATH] Similar, but creates a read-only bind mount\n"
" --setenv=NAME=VALUE Pass an environment variable to PID 1\n"
+ " --share-system Share system namespaces with host\n"
" --register=BOOLEAN Register container as machine\n"
" --keep-unit Do not register a scope for the machine, reuse\n"
- " the service unit nspawn is running in\n"
- " -q --quiet Do not show status information\n",
+ " the service unit nspawn is running in\n",
program_invocation_short_name);
return 0;
ARG_SHARE_SYSTEM,
ARG_REGISTER,
ARG_KEEP_UNIT,
- ARG_NETWORK_INTERFACE
+ ARG_NETWORK_INTERFACE,
+ ARG_NETWORK_VETH,
+ ARG_NETWORK_BRIDGE,
+ ARG_PERSONALITY,
};
static const struct option options[] = {
{ "register", required_argument, NULL, ARG_REGISTER },
{ "keep-unit", no_argument, NULL, ARG_KEEP_UNIT },
{ "network-interface", required_argument, NULL, ARG_NETWORK_INTERFACE },
+ { "network-veth", no_argument, NULL, ARG_NETWORK_VETH },
+ { "network-bridge", required_argument, NULL, ARG_NETWORK_BRIDGE },
+ { "personality", required_argument, NULL, ARG_PERSONALITY },
{}
};
break;
+ case ARG_NETWORK_BRIDGE:
+ arg_network_bridge = strdup(optarg);
+ if (!arg_network_bridge)
+ return log_oom();
+
+ /* fall through */
+
+ case ARG_NETWORK_VETH:
+ arg_network_veth = true;
+ arg_private_network = true;
+ break;
+
case ARG_NETWORK_INTERFACE:
if (strv_push(&arg_network_interfaces, optarg) < 0)
return log_oom();
arg_keep_unit = true;
break;
+ case ARG_PERSONALITY:
+
+ arg_personality = personality_from_string(optarg);
+ if (arg_personality == 0xffffffffLU) {
+ log_error("Unknown or unsupported personality '%s'.", optarg);
+ return -EINVAL;
+ }
+
+ break;
+
case '?':
return -EINVAL;
static int setup_boot_id(const char *dest) {
_cleanup_free_ char *from = NULL, *to = NULL;
- sd_id128_t rnd;
+ sd_id128_t rnd = {};
char as_uuid[37];
int r;
return 0;
}
-static int move_network_interfaces(pid_t pid) {
+static int setup_veth(pid_t pid, char iface_name[]) {
+ _cleanup_rtnl_message_unref_ sd_rtnl_message *m = NULL;
+ _cleanup_rtnl_unref_ sd_rtnl *rtnl = NULL;
+ int r;
+
+ if (!arg_private_network)
+ return 0;
+
+ if (!arg_network_veth)
+ return 0;
+
+ strncpy(iface_name+3, arg_machine, IFNAMSIZ - 3);
+
+ r = sd_rtnl_open(&rtnl, 0);
+ if (r < 0) {
+ log_error("Failed to connect to netlink: %s", strerror(-r));
+ return r;
+ }
+
+ r = sd_rtnl_message_new_link(rtnl, &m, RTM_NEWLINK, 0);
+ if (r < 0) {
+ log_error("Failed to allocate netlink message: %s", strerror(-r));
+ return r;
+ }
+
+ r = sd_rtnl_message_append_string(m, IFLA_IFNAME, iface_name);
+ if (r < 0) {
+ log_error("Failed to add netlink interface name: %s", strerror(-r));
+ return r;
+ }
+
+ r = sd_rtnl_message_open_container(m, IFLA_LINKINFO);
+ if (r < 0) {
+ log_error("Failed to open netlink container: %s", strerror(-r));
+ return r;
+ }
+
+ r = sd_rtnl_message_append_string(m, IFLA_INFO_KIND, "veth");
+ if (r < 0) {
+ log_error("Failed to append netlink kind: %s", strerror(-r));
+ return r;
+ }
+
+ r = sd_rtnl_message_open_container(m, IFLA_INFO_DATA);
+ if (r < 0) {
+ log_error("Failed to open netlink container: %s", strerror(-r));
+ return r;
+ }
+
+ r = sd_rtnl_message_open_container(m, VETH_INFO_PEER);
+ if (r < 0) {
+ log_error("Failed to open netlink container: %s", strerror(-r));
+ return r;
+ }
+
+ r = sd_rtnl_message_append_string(m, IFLA_IFNAME, "host0");
+ if (r < 0) {
+ log_error("Failed to add netlink interface name: %s", strerror(-r));
+ return r;
+ }
+
+ r = sd_rtnl_message_append_u32(m, IFLA_NET_NS_PID, pid);
+ if (r < 0) {
+ log_error("Failed to add netlink namespace field: %s", strerror(-r));
+ return r;
+ }
+
+ r = sd_rtnl_message_close_container(m);
+ if (r < 0) {
+ log_error("Failed to close netlink container: %s", strerror(-r));
+ return r;
+ }
+
+ r = sd_rtnl_message_close_container(m);
+ if (r < 0) {
+ log_error("Failed to close netlink container: %s", strerror(-r));
+ return r;
+ }
+
+ r = sd_rtnl_message_close_container(m);
+ if (r < 0) {
+ log_error("Failed to close netlink container: %s", strerror(-r));
+ return r;
+ }
+
+ r = sd_rtnl_call(rtnl, m, 0, NULL);
+ if (r < 0) {
+ log_error("Failed to add new veth interfaces: %s", strerror(-r));
+ return r;
+ }
+
+ return 0;
+}
+
+static int setup_bridge(const char veth_name[]) {
+ _cleanup_rtnl_message_unref_ sd_rtnl_message *m = NULL;
_cleanup_rtnl_unref_ sd_rtnl *rtnl = NULL;
+ int r, bridge;
+
+ if (!arg_private_network)
+ return 0;
+
+ if (!arg_network_veth)
+ return 0;
+
+ if (!arg_network_bridge)
+ return 0;
+
+ bridge = (int) if_nametoindex(arg_network_bridge);
+ if (bridge <= 0) {
+ log_error("Failed to resolve interface %s: %m", arg_network_bridge);
+ return -errno;
+ }
+
+ r = sd_rtnl_open(&rtnl, 0);
+ if (r < 0) {
+ log_error("Failed to connect to netlink: %s", strerror(-r));
+ return r;
+ }
+
+ r = sd_rtnl_message_new_link(rtnl, &m, RTM_SETLINK, 0);
+ if (r < 0) {
+ log_error("Failed to allocate netlink message: %s", strerror(-r));
+ return r;
+ }
+
+ r = sd_rtnl_message_append_string(m, IFLA_IFNAME, veth_name);
+ if (r < 0) {
+ log_error("Failed to add netlink interface name field: %s", strerror(-r));
+ return r;
+ }
+
+ r = sd_rtnl_message_append_u32(m, IFLA_MASTER, bridge);
+ if (r < 0) {
+ log_error("Failed to add netlink master field: %s", strerror(-r));
+ return r;
+ }
+
+ r = sd_rtnl_call(rtnl, m, 0, NULL);
+ if (r < 0) {
+ log_error("Failed to add veth interface to bridge: %s", strerror(-r));
+ return r;
+ }
+
+ return 0;
+}
+
+static int move_network_interfaces(pid_t pid) {
_cleanup_udev_unref_ struct udev *udev = NULL;
+ _cleanup_rtnl_unref_ sd_rtnl *rtnl = NULL;
char **i;
int r;
if (strv_isempty(arg_network_interfaces))
return 0;
- r = sd_rtnl_open(0, &rtnl);
+ r = sd_rtnl_open(&rtnl, 0);
if (r < 0) {
log_error("Failed to connect to netlink: %s", strerror(-r));
return r;
return -EBUSY;
}
- r = sd_rtnl_message_new_link(RTM_NEWLINK, ifi, &m);
+ r = sd_rtnl_message_new_link(rtnl, &m, RTM_NEWLINK, ifi);
if (r < 0) {
log_error("Failed to allocate netlink message: %s", strerror(-r));
return r;
return 0;
}
+static int audit_still_doesnt_work_in_containers(void) {
+
+#ifdef HAVE_SECCOMP
+ scmp_filter_ctx seccomp;
+ int r;
+
+ /*
+ Audit is broken in containers, much of the userspace audit
+ hookup will fail if running inside a container. We don't
+ care and just turn off creation of audit sockets.
+
+ This will make socket(AF_NETLINK, *, NETLINK_AUDIT) fail
+ with EAFNOSUPPORT which audit userspace uses as indication
+ that audit is disabled in the kernel.
+ */
+
+ seccomp = seccomp_init(SCMP_ACT_ALLOW);
+ if (!seccomp)
+ return log_oom();
+
+ r = seccomp_add_secondary_archs(seccomp);
+ if (r < 0 && r != -EEXIST) {
+ log_error("Failed to add secondary archs to seccomp filter: %s", strerror(-r));
+ goto finish;
+ }
+
+ r = seccomp_rule_add(
+ seccomp,
+ SCMP_ACT_ERRNO(EAFNOSUPPORT),
+ SCMP_SYS(socket),
+ 2,
+ SCMP_A0(SCMP_CMP_EQ, AF_NETLINK),
+ SCMP_A2(SCMP_CMP_EQ, NETLINK_AUDIT));
+ if (r < 0) {
+ log_error("Failed to add audit seccomp rule: %s", strerror(-r));
+ goto finish;
+ }
+
+ r = seccomp_attr_set(seccomp, SCMP_FLTATR_CTL_NNP, 0);
+ if (r < 0) {
+ log_error("Failed to unset NO_NEW_PRIVS: %s", strerror(-r));
+ goto finish;
+ }
+
+ r = seccomp_load(seccomp);
+ if (r < 0)
+ log_error("Failed to install seccomp audit filter: %s", strerror(-r));
+
+finish:
+ seccomp_release(seccomp);
+ return r;
+#else
+ return 0;
+#endif
+
+}
+
int main(int argc, char *argv[]) {
- pid_t pid = 0;
- int r = EXIT_FAILURE, k;
+
_cleanup_close_ int master = -1, kdbus_fd = -1, sync_fd = -1;
- int n_fd_passed;
- const char *console = NULL;
- sigset_t mask;
_cleanup_close_pipe_ int kmsg_socket_pair[2] = { -1, -1 };
- _cleanup_fdset_free_ FDSet *fds = NULL;
_cleanup_free_ char *kdbus_domain = NULL;
+ _cleanup_fdset_free_ FDSet *fds = NULL;
+ const char *console = NULL;
+ int r = EXIT_FAILURE, k;
+ int n_fd_passed;
+ pid_t pid = 0;
+ sigset_t mask;
+ char veth_name[IFNAMSIZ] = "ve-";
log_parse_environment();
log_open();
goto finish;
}
- if (arg_boot && path_is_os_tree(arg_directory) <= 0) {
- log_error("Directory %s doesn't look like an OS root directory (/etc/os-release is missing). Refusing.", arg_directory);
- goto finish;
+ if (arg_boot) {
+ if (path_is_os_tree(arg_directory) <= 0) {
+ log_error("Directory %s doesn't look like an OS root directory (/etc/os-release is missing). Refusing.", arg_directory);
+ goto finish;
+ }
+ } else {
+ const char *p;
+
+ p = strappenda(arg_directory,
+ argc > optind && path_is_absolute(argv[optind]) ? argv[optind] : "/usr/bin/");
+ if (access(p, F_OK) < 0) {
+ log_error("Directory %s lacks the binary to execute or doesn't look like a binary tree. Refusing.", arg_directory);
+ goto finish;
+
+ }
}
log_close();
goto finish;
}
-
if (access("/dev/kdbus/control", F_OK) >= 0) {
if (arg_share_system) {
dev_setup(arg_directory);
+ if (audit_still_doesnt_work_in_containers() < 0)
+ goto child_fail;
+
if (setup_dev_console(arg_directory, console) < 0)
goto child_fail;
setup_hostname();
+ if (arg_personality != 0xffffffffLU) {
+ if (personality(arg_personality) < 0) {
+ log_error("personality() failed: %m");
+ goto child_fail;
+ }
+ }
+
eventfd_read(sync_fd, &x);
close_nointr_nofail(sync_fd);
sync_fd = -1;
else {
chdir(home ? home : "/root");
execle("/bin/bash", "-bash", NULL, env_use);
+ execle("/bin/sh", "-sh", NULL, env_use);
}
log_error("execv() failed: %m");
if (r < 0)
goto finish;
+ r = setup_veth(pid, veth_name);
+ if (r < 0)
+ goto finish;
+
+ r = setup_bridge(veth_name);
+ if (r < 0)
+ goto finish;
+
eventfd_write(sync_fd, 1);
close_nointr_nofail(sync_fd);
sync_fd = -1;