#include <sys/un.h>
#include <sys/socket.h>
#include <linux/netlink.h>
+#include <linux/rtnetlink.h>
#include <sys/eventfd.h>
+#include <net/if.h>
+
#ifdef HAVE_SELINUX
#include <selinux/selinux.h>
#endif
#include "sd-daemon.h"
#include "sd-bus.h"
#include "sd-id128.h"
+#include "sd-rtnl.h"
#include "log.h"
#include "util.h"
#include "mkdir.h"
#include "bus-kernel.h"
#include "env-util.h"
#include "def.h"
+#include "rtnl-util.h"
typedef enum LinkJournal {
LINK_NO,
static bool arg_quiet = false;
static bool arg_share_system = false;
static bool arg_register = true;
+static bool arg_keep_unit = false;
+static char **arg_network_interfaces = NULL;
static int help(void) {
" 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"
" --bind-ro=PATH[:PATH] Similar, but creates a read-only bind mount\n"
" --setenv=NAME=VALUE Pass an environment variable to PID 1\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",
program_invocation_short_name);
ARG_BIND_RO,
ARG_SETENV,
ARG_SHARE_SYSTEM,
- ARG_REGISTER
+ ARG_REGISTER,
+ ARG_KEEP_UNIT,
+ ARG_NETWORK_INTERFACE
};
static const struct option options[] = {
- { "help", no_argument, NULL, 'h' },
- { "version", no_argument, NULL, ARG_VERSION },
- { "directory", required_argument, NULL, 'D' },
- { "user", required_argument, NULL, 'u' },
- { "private-network", no_argument, NULL, ARG_PRIVATE_NETWORK },
- { "boot", no_argument, NULL, 'b' },
- { "uuid", required_argument, NULL, ARG_UUID },
- { "read-only", no_argument, NULL, ARG_READ_ONLY },
- { "capability", required_argument, NULL, ARG_CAPABILITY },
- { "drop-capability", required_argument, NULL, ARG_DROP_CAPABILITY },
- { "link-journal", required_argument, NULL, ARG_LINK_JOURNAL },
- { "bind", required_argument, NULL, ARG_BIND },
- { "bind-ro", required_argument, NULL, ARG_BIND_RO },
- { "machine", required_argument, NULL, 'M' },
- { "slice", required_argument, NULL, 'S' },
- { "setenv", required_argument, NULL, ARG_SETENV },
- { "selinux-context", required_argument, NULL, 'Z' },
- { "selinux-apifs-context", required_argument, NULL, 'L' },
- { "quiet", no_argument, NULL, 'q' },
- { "share-system", no_argument, NULL, ARG_SHARE_SYSTEM },
- { "register", required_argument, NULL, ARG_REGISTER },
+ { "help", no_argument, NULL, 'h' },
+ { "version", no_argument, NULL, ARG_VERSION },
+ { "directory", required_argument, NULL, 'D' },
+ { "user", required_argument, NULL, 'u' },
+ { "private-network", no_argument, NULL, ARG_PRIVATE_NETWORK },
+ { "boot", no_argument, NULL, 'b' },
+ { "uuid", required_argument, NULL, ARG_UUID },
+ { "read-only", no_argument, NULL, ARG_READ_ONLY },
+ { "capability", required_argument, NULL, ARG_CAPABILITY },
+ { "drop-capability", required_argument, NULL, ARG_DROP_CAPABILITY },
+ { "link-journal", required_argument, NULL, ARG_LINK_JOURNAL },
+ { "bind", required_argument, NULL, ARG_BIND },
+ { "bind-ro", required_argument, NULL, ARG_BIND_RO },
+ { "machine", required_argument, NULL, 'M' },
+ { "slice", required_argument, NULL, 'S' },
+ { "setenv", required_argument, NULL, ARG_SETENV },
+ { "selinux-context", required_argument, NULL, 'Z' },
+ { "selinux-apifs-context", required_argument, NULL, 'L' },
+ { "quiet", no_argument, NULL, 'q' },
+ { "share-system", no_argument, NULL, ARG_SHARE_SYSTEM },
+ { "register", required_argument, NULL, ARG_REGISTER },
+ { "keep-unit", no_argument, NULL, ARG_KEEP_UNIT },
+ { "network-interface", required_argument, NULL, ARG_NETWORK_INTERFACE },
{}
};
break;
+ case ARG_NETWORK_INTERFACE:
+ if (strv_push(&arg_network_interfaces, optarg) < 0)
+ return log_oom();
+
+ /* fall through */
+
case ARG_PRIVATE_NETWORK:
arg_private_network = true;
break;
size_t length;
FOREACH_WORD_SEPARATOR(word, length, optarg, ",", state) {
+ _cleanup_free_ char *t;
cap_value_t cap;
- char *t;
t = strndup(word, length);
if (!t)
return log_oom();
- if (cap_from_name(t, &cap) < 0) {
- log_error("Failed to parse capability %s.", t);
- free(t);
- return -EINVAL;
+ if (streq(t, "all")) {
+ if (c == ARG_CAPABILITY)
+ arg_retain = (uint64_t) -1;
+ else
+ arg_retain = 0;
+ } else {
+ if (cap_from_name(t, &cap) < 0) {
+ log_error("Failed to parse capability %s.", t);
+ return -EINVAL;
+ }
+
+ if (c == ARG_CAPABILITY)
+ arg_retain |= 1ULL << (uint64_t) cap;
+ else
+ arg_retain &= ~(1ULL << (uint64_t) cap);
}
-
- free(t);
-
- if (c == ARG_CAPABILITY)
- arg_retain |= 1ULL << (uint64_t) cap;
- else
- arg_retain &= ~(1ULL << (uint64_t) cap);
}
break;
arg_register = r;
break;
+ case ARG_KEEP_UNIT:
+ arg_keep_unit = true;
+ break;
+
case '?':
return -EINVAL;
return -EINVAL;
}
+ if (arg_keep_unit && cg_pid_get_owner_uid(0, NULL) >= 0) {
+ log_error("--keep-unit may not be used when invoked from a user session.");
+ return -EINVAL;
+ }
+
return 1;
}
return r;
}
- r = sd_bus_call_method(
- bus,
- "org.freedesktop.machine1",
- "/org/freedesktop/machine1",
- "org.freedesktop.machine1.Manager",
- "CreateMachine",
- &error,
- NULL,
- "sayssusa(sv)",
- arg_machine,
- SD_BUS_MESSAGE_APPEND_ID128(arg_uuid),
- "nspawn",
- "container",
- (uint32_t) pid,
- strempty(arg_directory),
- !isempty(arg_slice), "Slice", "s", arg_slice);
+ if (arg_keep_unit) {
+ r = sd_bus_call_method(
+ bus,
+ "org.freedesktop.machine1",
+ "/org/freedesktop/machine1",
+ "org.freedesktop.machine1.Manager",
+ "RegisterMachine",
+ &error,
+ NULL,
+ "sayssus",
+ arg_machine,
+ SD_BUS_MESSAGE_APPEND_ID128(arg_uuid),
+ "nspawn",
+ "container",
+ (uint32_t) pid,
+ strempty(arg_directory));
+ } else {
+ r = sd_bus_call_method(
+ bus,
+ "org.freedesktop.machine1",
+ "/org/freedesktop/machine1",
+ "org.freedesktop.machine1.Manager",
+ "CreateMachine",
+ &error,
+ NULL,
+ "sayssusa(sv)",
+ arg_machine,
+ SD_BUS_MESSAGE_APPEND_ID128(arg_uuid),
+ "nspawn",
+ "container",
+ (uint32_t) pid,
+ strempty(arg_directory),
+ !isempty(arg_slice), "Slice", "s", arg_slice);
+ }
+
if (r < 0) {
log_error("Failed to register machine: %s", bus_error_message(&error, r));
return r;
return 0;
}
-static bool audit_enabled(void) {
- int fd;
+static int reset_audit_loginuid(void) {
+ _cleanup_free_ char *p = NULL;
+ int r;
+
+ if (arg_share_system)
+ return 0;
+
+ r = read_one_line_file("/proc/self/loginuid", &p);
+ if (r == -EEXIST)
+ return 0;
+ if (r < 0) {
+ log_error("Failed to read /proc/self/loginuid: %s", strerror(-r));
+ return r;
+ }
+
+ /* Already reset? */
+ if (streq(p, "4294967295"))
+ return 0;
+
+ r = write_string_file("/proc/self/loginuid", "4294967295");
+ if (r < 0) {
+ log_error("Failed to reset audit login UID. This probably means that your kernel is too\n"
+ "old and you have audit enabled. Note that the auditing subsystem is known to\n"
+ "be incompatible with containers on old kernels. Please make sure to upgrade\n"
+ "your kernel or to off auditing with 'audit=0' on the kernel command line before\n"
+ "using systemd-nspawn. Sleeping for 5s... (%s)\n", strerror(-r));
+
+ sleep(5);
+ }
+
+ return 0;
+}
+
+static int move_network_interfaces(pid_t pid) {
+ _cleanup_sd_rtnl_unref_ sd_rtnl *rtnl = NULL;
+ char **i;
+ int r;
+
+ if (!arg_private_network)
+ return 0;
+
+ if (strv_isempty(arg_network_interfaces))
+ return 0;
+
+ r = sd_rtnl_open(NETLINK_ROUTE, &rtnl);
+ if (r < 0) {
+ log_error("Failed to connect to netlink: %s", strerror(-r));
+ return r;
+ }
+
+ STRV_FOREACH(i, arg_network_interfaces) {
+ _cleanup_sd_rtnl_message_unref_ sd_rtnl_message *m = NULL;
+ unsigned ifi;
+
+ ifi = if_nametoindex(*i);
+ if (ifi == 0) {
+ log_error("Failed to resolve interface %s: %m", *i);
+ return -errno;
+ }
+
+ r = sd_rtnl_message_link_new(RTM_NEWLINK, ifi, &m);
+ if (r < 0) {
+ log_error("Failed to allocate netlink message: %s", strerror(-r));
+ return r;
+ }
+
+ r = sd_rtnl_message_append_u32(m, IFLA_NET_NS_PID, pid);
+ if (r < 0) {
+ log_error("Failed to append namespace PID to netlink message: %s", strerror(-r));
+ return r;
+ }
- fd = socket(AF_NETLINK, SOCK_RAW, NETLINK_AUDIT);
- if (fd >= 0) {
- close_nointr_nofail(fd);
- return true;
+ r = sd_rtnl_call(rtnl, m, 0, NULL);
+ if (r < 0) {
+ log_error("Failed to move interface to namespace: %s", strerror(-r));
+ return r;
+ }
}
- return false;
+
+ return 0;
}
int main(int argc, char *argv[]) {
goto finish;
}
- if (arg_boot && audit_enabled()) {
- log_warning("The kernel auditing subsystem is known to be incompatible with containers.\n"
- "Please make sure to turn off auditing with 'audit=0' on the kernel command\n"
- "line before using systemd-nspawn. Sleeping for 5s...\n");
- sleep(5);
- }
-
if (path_equal(arg_directory, "/")) {
log_error("Spawning container on root directory not supported.");
goto finish;
goto child_fail;
}
+ if (reset_audit_loginuid() < 0)
+ goto child_fail;
+
if (prctl(PR_SET_PDEATHSIG, SIGKILL) < 0) {
log_error("PR_SET_PDEATHSIG failed: %m");
goto child_fail;
if (r < 0)
goto finish;
+ r = move_network_interfaces(pid);
+ if (r < 0)
+ goto finish;
+
eventfd_write(sync_fd, 1);
close_nointr_nofail(sync_fd);
sync_fd = -1;
free(arg_directory);
free(arg_machine);
free(arg_setenv);
+ free(arg_network_interfaces);
return r;
}