#include <sys/mount.h>
#include <sys/types.h>
#include <sys/stat.h>
+#include <sys/prctl.h>
+#include <grp.h>
/* we only need some defines missing in sys/mount.h, no libmount linkage */
#include <libmount.h>
#include "nls.h"
#include "c.h"
+#include "caputils.h"
#include "closestream.h"
#include "namespace.h"
#include "exec_shell.h"
#include "xalloc.h"
#include "pathnames.h"
#include "all-io.h"
+#include "signames.h"
+#include "strutils.h"
/* synchronize parent and child by pipe */
#define PIPE_SYNC_BYTE 0x06
static int npersists; /* number of persistent namespaces */
-
enum {
SETGROUPS_NONE = -1,
SETGROUPS_DENY = 0,
SETGROUPS_ALLOW = 1,
};
+enum {
+ MAP_USER_NONE,
+ MAP_USER_ROOT,
+ MAP_USER_CURRENT,
+};
+
static const char *setgroups_strings[] =
{
[SETGROUPS_DENY] = "deny",
fputs(_(" -p, --pid[=<file>] unshare pid namespace\n"), out);
fputs(_(" -U, --user[=<file>] unshare user namespace\n"), out);
fputs(_(" -C, --cgroup[=<file>] unshare cgroup namespace\n"), out);
+ fputs(USAGE_SEPARATOR, out);
fputs(_(" -f, --fork fork before launching <program>\n"), out);
- fputs(_(" --mount-proc[=<dir>] mount proc filesystem first (implies --mount)\n"), out);
fputs(_(" -r, --map-root-user map current user to root (implies --user)\n"), out);
- fputs(_(" --propagation slave|shared|private|unchanged\n"
+ fputs(_(" -c, --map-current-user map current user to itself (implies --user)\n"), out);
+ fputs(USAGE_SEPARATOR, out);
+ fputs(_(" --kill-child[=<signame>] when dying, kill the forked child (implies --fork)\n"
+ " defaults to SIGKILL\n"), out);
+ fputs(_(" --mount-proc[=<dir>] mount proc filesystem first (implies --mount)\n"), out);
+ fputs(_(" --propagation slave|shared|private|unchanged\n"
" modify mount propagation in mount namespace\n"), out);
- fputs(_(" -s, --setgroups allow|deny control the setgroups syscall in user namespaces\n"), out);
+ fputs(_(" --setgroups allow|deny control the setgroups syscall in user namespaces\n"), out);
+ fputs(_(" --keep-caps retain capabilities granted in user namespaces\n"), out);
+ fputs(USAGE_SEPARATOR, out);
+ fputs(_(" -R, --root=<dir> run the command with root directory set to <dir>\n"), out);
+ fputs(_(" -w, --wd=<dir> change working directory to <dir>\n"), out);
+ fputs(_(" -S, --setuid <uid> set uid in entered namespace\n"), out);
+ fputs(_(" -G, --setgid <gid> set gid in entered namespace\n"), out);
fputs(USAGE_SEPARATOR, out);
printf(USAGE_HELP_OPTIONS(27));
enum {
OPT_MOUNTPROC = CHAR_MAX + 1,
OPT_PROPAGATION,
- OPT_SETGROUPS
+ OPT_SETGROUPS,
+ OPT_KILLCHILD,
+ OPT_KEEPCAPS,
};
static const struct option longopts[] = {
{ "help", no_argument, NULL, 'h' },
{ "cgroup", optional_argument, NULL, 'C' },
{ "fork", no_argument, NULL, 'f' },
+ { "kill-child", optional_argument, NULL, OPT_KILLCHILD },
{ "mount-proc", optional_argument, NULL, OPT_MOUNTPROC },
{ "map-root-user", no_argument, NULL, 'r' },
+ { "map-current-user", no_argument, NULL, 'c' },
{ "propagation", required_argument, NULL, OPT_PROPAGATION },
{ "setgroups", required_argument, NULL, OPT_SETGROUPS },
+ { "keep-caps", no_argument, NULL, OPT_KEEPCAPS },
+ { "setuid", required_argument, NULL, 'S' },
+ { "setgid", required_argument, NULL, 'G' },
+ { "root", required_argument, NULL, 'R' },
+ { "wd", required_argument, NULL, 'w' },
{ NULL, 0, NULL, 0 }
};
int setgrpcmd = SETGROUPS_NONE;
int unshare_flags = 0;
- int c, forkit = 0, maproot = 0;
+ int c, forkit = 0, mapuser = MAP_USER_NONE;
+ int kill_child_signo = 0; /* 0 means --kill-child was not used */
const char *procmnt = NULL;
+ const char *newroot = NULL;
+ const char *newdir = NULL;
pid_t pid = 0;
int fds[2];
int status;
unsigned long propagation = UNSHARE_PROPAGATION_DEFAULT;
- uid_t real_euid = geteuid();
- gid_t real_egid = getegid();
+ int force_uid = 0, force_gid = 0;
+ uid_t uid = 0, real_euid = geteuid();
+ gid_t gid = 0, real_egid = getegid();
+ int keepcaps = 0;
setlocale(LC_ALL, "");
bindtextdomain(PACKAGE, LOCALEDIR);
textdomain(PACKAGE);
- atexit(close_stdout);
+ close_stdout_atexit();
- while ((c = getopt_long(argc, argv, "+fhVmuinpCUr", longopts, NULL)) != -1) {
+ while ((c = getopt_long(argc, argv, "+fhVmuinpCUrR:w:S:G:c", longopts, NULL)) != -1) {
switch (c) {
case 'f':
forkit = 1;
break;
- case 'h':
- usage();
- case 'V':
- printf(UTIL_LINUX_VERSION);
- return EXIT_SUCCESS;
case 'm':
unshare_flags |= CLONE_NEWNS;
if (optarg)
procmnt = optarg ? optarg : "/proc";
break;
case 'r':
+ if (mapuser == MAP_USER_CURRENT)
+ errx(EXIT_FAILURE, _("options --map-root-user and "
+ "--map-current-user are mutually exclusive"));
+
+ unshare_flags |= CLONE_NEWUSER;
+ mapuser = MAP_USER_ROOT;
+ break;
+ case 'c':
+ if (mapuser == MAP_USER_ROOT)
+ errx(EXIT_FAILURE, _("options --map-root-user and "
+ "--map-current-user are mutually exclusive"));
+
unshare_flags |= CLONE_NEWUSER;
- maproot = 1;
+ mapuser = MAP_USER_CURRENT;
break;
case OPT_SETGROUPS:
setgrpcmd = setgroups_str2id(optarg);
case OPT_PROPAGATION:
propagation = parse_propagation(optarg);
break;
+ case OPT_KILLCHILD:
+ forkit = 1;
+ if (optarg) {
+ if ((kill_child_signo = signame_to_signum(optarg)) < 0)
+ errx(EXIT_FAILURE, _("unknown signal: %s"),
+ optarg);
+ } else {
+ kill_child_signo = SIGKILL;
+ }
+ break;
+ case OPT_KEEPCAPS:
+ keepcaps = 1;
+ cap_last_cap(); /* Force last cap to be cached before we fork. */
+ break;
+ case 'S':
+ uid = strtoul_or_err(optarg, _("failed to parse uid"));
+ force_uid = 1;
+ break;
+ case 'G':
+ gid = strtoul_or_err(optarg, _("failed to parse gid"));
+ force_gid = 1;
+ break;
+ case 'R':
+ newroot = optarg;
+ break;
+ case 'w':
+ newdir = optarg;
+ break;
+
+ case 'h':
+ usage();
+ case 'V':
+ print_version(EXIT_SUCCESS);
default:
errtryhelp(EXIT_FAILURE);
}
}
}
+ if (kill_child_signo != 0 && prctl(PR_SET_PDEATHSIG, kill_child_signo) < 0)
+ err(EXIT_FAILURE, "prctl failed");
- if (maproot) {
+ /* Since Linux 3.19 unprivileged writing of /proc/self/gid_map
+ * has been disabled unless /proc/self/setgroups is written
+ * first to permanently disable the ability to call setgroups
+ * in that user namespace. */
+ switch (mapuser) {
+ case MAP_USER_ROOT:
if (setgrpcmd == SETGROUPS_ALLOW)
errx(EXIT_FAILURE, _("options --setgroups=allow and "
"--map-root-user are mutually exclusive"));
- /* since Linux 3.19 unprivileged writing of /proc/self/gid_map
- * has s been disabled unless /proc/self/setgroups is written
- * first to permanently disable the ability to call setgroups
- * in that user namespace. */
setgroups_control(SETGROUPS_DENY);
map_id(_PATH_PROC_UIDMAP, 0, real_euid);
map_id(_PATH_PROC_GIDMAP, 0, real_egid);
+ break;
+ case MAP_USER_CURRENT:
+ if (setgrpcmd == SETGROUPS_ALLOW)
+ errx(EXIT_FAILURE, _("options --setgroups=allow and "
+ "--map-current-user are mutually exclusive"));
- } else if (setgrpcmd != SETGROUPS_NONE)
- setgroups_control(setgrpcmd);
+ setgroups_control(SETGROUPS_DENY);
+ map_id(_PATH_PROC_UIDMAP, real_euid, real_euid);
+ map_id(_PATH_PROC_GIDMAP, real_egid, real_egid);
+ break;
+ case MAP_USER_NONE:
+ if (setgrpcmd != SETGROUPS_NONE)
+ setgroups_control(setgrpcmd);
+ }
if ((unshare_flags & CLONE_NEWNS) && propagation)
set_propagation(propagation);
- if (procmnt &&
- (mount("none", procmnt, NULL, MS_PRIVATE|MS_REC, NULL) != 0 ||
- mount("proc", procmnt, "proc", MS_NOSUID|MS_NOEXEC|MS_NODEV, NULL) != 0))
+ if (newroot) {
+ if (chroot(newroot) != 0)
+ err(EXIT_FAILURE,
+ _("cannot change root directory to '%s'"), newroot);
+ newdir = newdir ?: "/";
+ }
+ if (newdir && chdir(newdir))
+ err(EXIT_FAILURE, _("cannot chdir to '%s'"), newdir);
+
+ if (procmnt) {
+ if (!newroot && mount("none", procmnt, NULL, MS_PRIVATE|MS_REC, NULL) != 0)
+ err(EXIT_FAILURE, _("umount %s failed"), procmnt);
+ if (mount("proc", procmnt, "proc", MS_NOSUID|MS_NOEXEC|MS_NODEV, NULL) != 0)
err(EXIT_FAILURE, _("mount %s failed"), procmnt);
+ }
+
+ if (force_gid) {
+ if (setgroups(0, NULL) != 0) /* drop supplementary groups */
+ err(EXIT_FAILURE, _("setgroups failed"));
+ if (setgid(gid) < 0) /* change GID */
+ err(EXIT_FAILURE, _("setgid failed"));
+ }
+ if (force_uid && setuid(uid) < 0) /* change UID */
+ err(EXIT_FAILURE, _("setuid failed"));
+
+ /* We use capabilities system calls to propagate the permitted
+ * capabilities into the ambient set because we have already
+ * forked so are in async-signal-safe context. */
+ if (keepcaps && (unshare_flags & CLONE_NEWUSER)) {
+ struct __user_cap_header_struct header = {
+ .version = _LINUX_CAPABILITY_VERSION_3,
+ .pid = 0,
+ };
+
+ struct __user_cap_data_struct payload[_LINUX_CAPABILITY_U32S_3] = { 0 };
+ int cap;
+ uint64_t effective;
+
+ if (capget(&header, payload) < 0)
+ err(EXIT_FAILURE, _("capget failed"));
+
+ /* In order the make capabilities ambient, we first need to ensure
+ * that they are all inheritable. */
+ payload[0].inheritable = payload[0].permitted;
+ payload[1].inheritable = payload[1].permitted;
+
+ if (capset(&header, payload) < 0)
+ err(EXIT_FAILURE, _("capset failed"));
+
+ effective = ((uint64_t)payload[1].effective << 32) | (uint64_t)payload[0].effective;
+
+ for (cap = 0; cap < 64; cap++) {
+ /* This is the same check as cap_valid(), but using
+ * the runtime value for the last valid cap. */
+ if (cap > cap_last_cap())
+ continue;
+
+ if ((effective & (1 << cap))
+ && prctl(PR_CAP_AMBIENT, PR_CAP_AMBIENT_RAISE, cap, 0, 0) < 0)
+ err(EXIT_FAILURE, _("prctl(PR_CAP_AMBIENT) failed"));
+ }
+ }
if (optind < argc) {
execvp(argv[optind], argv + optind);
- err(EXIT_FAILURE, _("failed to execute %s"), argv[optind]);
+ errexec(argv[optind]);
}
exec_shell();
}
projects / thirdparty / util-linux.git / blobdiff