X-Git-Url: https://www.chiark.greenend.org.uk/ucgi/~ianmdlvl/git?p=elogind.git;a=blobdiff_plain;f=src%2Fcore%2Fnamespace.c;h=4b8dbdd0d83d975a2e5f135d280d05f056047f98;hp=926ff7133b98e21f881158134817cf88c74b2b89;hb=4968105790c65af58d4ab42bffa2a4bedc0be8ee;hpb=76cd584b8d6e0df5f662f193803ce2bdc530ac86 diff --git a/src/core/namespace.c b/src/core/namespace.c index 926ff7133..4b8dbdd0d 100644 --- a/src/core/namespace.c +++ b/src/core/namespace.c @@ -35,10 +35,15 @@ #include "strv.h" #include "util.h" #include "path-util.h" -#include "namespace.h" #include "missing.h" #include "execute.h" #include "loopback-setup.h" +#include "mkdir.h" +#include "dev-setup.h" +#include "def.h" +#include "label.h" +#include "selinux-util.h" +#include "namespace.h" typedef enum MountMode { /* This is ordered by priority! */ @@ -46,6 +51,8 @@ typedef enum MountMode { READONLY, PRIVATE_TMP, PRIVATE_VAR_TMP, + PRIVATE_DEV, + PRIVATE_BUS_ENDPOINT, READWRITE } MountMode; @@ -64,8 +71,9 @@ static int append_mounts(BindMount **p, char **strv, MountMode mode) { STRV_FOREACH(i, strv) { (*p)->ignore = false; + (*p)->done = false; - if ((mode == INACCESSIBLE || mode == READONLY) && (*i)[0] == '-') { + if ((mode == INACCESSIBLE || mode == READONLY || mode == READWRITE) && (*i)[0] == '-') { (*p)->ignore = true; (*i)++; } @@ -118,8 +126,7 @@ static void drop_duplicates(BindMount *m, unsigned *n) { if (previous && path_equal(f->path, previous->path)) continue; - t->path = f->path; - t->mode = f->mode; + *t = *f; previous = t; @@ -129,6 +136,207 @@ static void drop_duplicates(BindMount *m, unsigned *n) { *n = t - m; } +static int mount_dev(BindMount *m) { + static const char devnodes[] = + "/dev/null\0" + "/dev/zero\0" + "/dev/full\0" + "/dev/random\0" + "/dev/urandom\0" + "/dev/tty\0"; + + char temporary_mount[] = "/tmp/namespace-dev-XXXXXX"; + const char *d, *dev = NULL, *devpts = NULL, *devshm = NULL, *devhugepages = NULL, *devmqueue = NULL, *devlog = NULL, *devptmx = NULL; + _cleanup_umask_ mode_t u; + int r; + + assert(m); + + u = umask(0000); + + if (!mkdtemp(temporary_mount)) + return -errno; + + dev = strappenda(temporary_mount, "/dev"); + (void)mkdir(dev, 0755); + if (mount("tmpfs", dev, "tmpfs", MS_NOSUID|MS_STRICTATIME, "mode=755") < 0) { + r = -errno; + goto fail; + } + + devpts = strappenda(temporary_mount, "/dev/pts"); + (void)mkdir(devpts, 0755); + if (mount("/dev/pts", devpts, NULL, MS_BIND, NULL) < 0) { + r = -errno; + goto fail; + } + + devptmx = strappenda(temporary_mount, "/dev/ptmx"); + symlink("pts/ptmx", devptmx); + + devshm = strappenda(temporary_mount, "/dev/shm"); + (void)mkdir(devshm, 01777); + r = mount("/dev/shm", devshm, NULL, MS_BIND, NULL); + if (r < 0) { + r = -errno; + goto fail; + } + + devmqueue = strappenda(temporary_mount, "/dev/mqueue"); + (void)mkdir(devmqueue, 0755); + mount("/dev/mqueue", devmqueue, NULL, MS_BIND, NULL); + + devhugepages = strappenda(temporary_mount, "/dev/hugepages"); + (void)mkdir(devhugepages, 0755); + mount("/dev/hugepages", devhugepages, NULL, MS_BIND, NULL); + + devlog = strappenda(temporary_mount, "/dev/log"); + symlink("/run/systemd/journal/dev-log", devlog); + + NULSTR_FOREACH(d, devnodes) { + _cleanup_free_ char *dn = NULL; + struct stat st; + + r = stat(d, &st); + if (r < 0) { + + if (errno == ENOENT) + continue; + + r = -errno; + goto fail; + } + + if (!S_ISBLK(st.st_mode) && + !S_ISCHR(st.st_mode)) { + r = -EINVAL; + goto fail; + } + + if (st.st_rdev == 0) + continue; + + dn = strappend(temporary_mount, d); + if (!dn) { + r = -ENOMEM; + goto fail; + } + + mac_selinux_create_file_prepare(d, st.st_mode); + r = mknod(dn, st.st_mode, st.st_rdev); + mac_selinux_create_file_clear(); + + if (r < 0) { + r = -errno; + goto fail; + } + } + + dev_setup(temporary_mount); + + if (mount(dev, "/dev/", NULL, MS_MOVE, NULL) < 0) { + r = -errno; + goto fail; + } + + rmdir(dev); + rmdir(temporary_mount); + + return 0; + +fail: + if (devpts) + umount(devpts); + + if (devshm) + umount(devshm); + + if (devhugepages) + umount(devhugepages); + + if (devmqueue) + umount(devmqueue); + + umount(dev); + rmdir(dev); + rmdir(temporary_mount); + + return r; +} + +static int mount_kdbus(BindMount *m) { + + char temporary_mount[] = "/tmp/kdbus-dev-XXXXXX"; + _cleanup_free_ char *basepath = NULL; + _cleanup_umask_ mode_t u; + char *busnode = NULL, *root; + struct stat st; + int r; + + assert(m); + + u = umask(0000); + + if (!mkdtemp(temporary_mount)) + return log_error_errno(errno, "Failed create temp dir: %m"); + + root = strappenda(temporary_mount, "/kdbus"); + (void)mkdir(root, 0755); + if (mount("tmpfs", root, "tmpfs", MS_NOSUID|MS_STRICTATIME, "mode=777") < 0) { + r = -errno; + goto fail; + } + + /* create a new /dev/null dev node copy so we have some fodder to + * bind-mount the custom endpoint over. */ + if (stat("/dev/null", &st) < 0) { + log_error_errno(errno, "Failed to stat /dev/null: %m"); + r = -errno; + goto fail; + } + + busnode = strappenda(root, "/bus"); + if (mknod(busnode, (st.st_mode & ~07777) | 0600, st.st_rdev) < 0) { + log_error_errno(errno, "mknod() for %s failed: %m", busnode); + r = -errno; + goto fail; + } + + r = mount(m->path, busnode, "bind", MS_BIND, NULL); + if (r < 0) { + log_error_errno(errno, "bind mount of %s failed: %m", m->path); + r = -errno; + goto fail; + } + + basepath = dirname_malloc(m->path); + if (!basepath) { + r = -ENOMEM; + goto fail; + } + + if (mount(root, basepath, NULL, MS_MOVE, NULL) < 0) { + log_error_errno(errno, "bind mount of %s failed: %m", basepath); + r = -errno; + goto fail; + } + + rmdir(temporary_mount); + return 0; + +fail: + if (busnode) { + umount(busnode); + unlink(busnode); + } + + umount(root); + rmdir(root); + rmdir(temporary_mount); + + return r; +} + static int apply_mount( BindMount *m, const char *tmp_dir, @@ -142,13 +350,20 @@ static int apply_mount( switch (m->mode) { case INACCESSIBLE: + + /* First, get rid of everything that is below if there + * is anything... Then, overmount it with an + * inaccessible directory. */ + umount_recursive(m->path, 0); + what = "/run/systemd/inaccessible"; break; case READONLY: case READWRITE: - what = m->path; - break; + /* Nothing to mount here, we just later toggle the + * MS_RDONLY bit for the mount point */ + return 0; case PRIVATE_TMP: what = tmp_dir; @@ -158,6 +373,12 @@ static int apply_mount( what = var_tmp_dir; break; + case PRIVATE_DEV: + return mount_dev(m); + + case PRIVATE_BUS_ENDPOINT: + return mount_kdbus(m); + default: assert_not_reached("Unknown mode"); } @@ -168,7 +389,7 @@ static int apply_mount( if (r >= 0) log_debug("Successfully mounted %s to %s", what, m->path); else if (m->ignore && errno == ENOENT) - r = 0; + return 0; return r; } @@ -178,23 +399,30 @@ static int make_read_only(BindMount *m) { assert(m); - if (m->mode != INACCESSIBLE && m->mode != READONLY) - return 0; + if (IN_SET(m->mode, INACCESSIBLE, READONLY)) + r = bind_remount_recursive(m->path, true); + else if (IN_SET(m->mode, READWRITE, PRIVATE_TMP, PRIVATE_VAR_TMP, PRIVATE_DEV)) + r = bind_remount_recursive(m->path, false); + else + r = 0; - r = mount(NULL, m->path, NULL, MS_BIND|MS_REMOUNT|MS_RDONLY|MS_REC, NULL); - if (r < 0 && !(m->ignore && errno == ENOENT)) - return -errno; + if (m->ignore && r == -ENOENT) + return 0; - return 0; + return r; } int setup_namespace( char** read_write_dirs, char** read_only_dirs, char** inaccessible_dirs, - char* tmp_dir, - char* var_tmp_dir, - unsigned mount_flags) { + const char* tmp_dir, + const char* var_tmp_dir, + const char* bus_endpoint_path, + bool private_dev, + ProtectHome protect_home, + ProtectSystem protect_system, + unsigned long mount_flags) { BindMount *m, *mounts = NULL; unsigned n; @@ -206,13 +434,17 @@ int setup_namespace( if (unshare(CLONE_NEWNS) < 0) return -errno; - n = !!tmp_dir + !!var_tmp_dir + + n = !!tmp_dir + !!var_tmp_dir + !!bus_endpoint_path + strv_length(read_write_dirs) + strv_length(read_only_dirs) + - strv_length(inaccessible_dirs); + strv_length(inaccessible_dirs) + + private_dev + + (protect_home != PROTECT_HOME_NO ? 3 : 0) + + (protect_system != PROTECT_SYSTEM_NO ? 2 : 0) + + (protect_system == PROTECT_SYSTEM_FULL ? 1 : 0); if (n > 0) { - m = mounts = (BindMount *) alloca(n * sizeof(BindMount)); + m = mounts = (BindMount *) alloca0(n * sizeof(BindMount)); r = append_mounts(&m, read_write_dirs, READWRITE); if (r < 0) return r; @@ -237,30 +469,58 @@ int setup_namespace( m++; } + if (private_dev) { + m->path = "/dev"; + m->mode = PRIVATE_DEV; + m++; + } + + if (bus_endpoint_path) { + m->path = bus_endpoint_path; + m->mode = PRIVATE_BUS_ENDPOINT; + m++; + } + + if (protect_home != PROTECT_HOME_NO) { + r = append_mounts(&m, STRV_MAKE("-/home", "-/run/user", "-/root"), protect_home == PROTECT_HOME_READ_ONLY ? READONLY : INACCESSIBLE); + if (r < 0) + return r; + } + + if (protect_system != PROTECT_SYSTEM_NO) { + r = append_mounts(&m, protect_system == PROTECT_SYSTEM_FULL ? STRV_MAKE("/usr", "-/boot", "/etc") : STRV_MAKE("/usr", "-/boot"), READONLY); + if (r < 0) + return r; + } + assert(mounts + n == m); qsort(mounts, n, sizeof(BindMount), mount_path_compare); drop_duplicates(mounts, &n); } - /* Remount / as SLAVE so that nothing now mounted in the namespace - shows up in the parent */ - if (mount(NULL, "/", NULL, MS_SLAVE|MS_REC, NULL) < 0) - return -errno; + if (n > 0) { + /* Remount / as SLAVE so that nothing now mounted in the namespace + shows up in the parent */ + if (mount(NULL, "/", NULL, MS_SLAVE|MS_REC, NULL) < 0) + return -errno; - for (m = mounts; m < mounts + n; ++m) { - r = apply_mount(m, tmp_dir, var_tmp_dir); - if (r < 0) - goto fail; - } + for (m = mounts; m < mounts + n; ++m) { + r = apply_mount(m, tmp_dir, var_tmp_dir); + if (r < 0) + goto fail; + } - for (m = mounts; m < mounts + n; ++m) { - r = make_read_only(m); - if (r < 0) - goto fail; + for (m = mounts; m < mounts + n; ++m) { + r = make_read_only(m); + if (r < 0) + goto fail; + } } - /* Remount / as the desired mode */ + /* Remount / as the desired mode. Not that this will not + * reestablish propagation from our side to the host, since + * what's disconnected is disconnected. */ if (mount(NULL, "/", NULL, mount_flags | MS_REC, NULL) < 0) { r = -errno; goto fail; @@ -269,21 +529,33 @@ int setup_namespace( return 0; fail: - for (m = mounts; m < mounts + n; ++m) - if (m->done) - umount2(m->path, MNT_DETACH); + if (n > 0) { + for (m = mounts; m < mounts + n; ++m) + if (m->done) + umount2(m->path, MNT_DETACH); + } return r; } static int setup_one_tmp_dir(const char *id, const char *prefix, char **path) { _cleanup_free_ char *x = NULL; + char bid[SD_ID128_STRING_MAX]; + sd_id128_t boot_id; + int r; assert(id); assert(prefix); assert(path); - x = strjoin(prefix, "/systemd-", id, "-XXXXXX", NULL); + /* We include the boot id in the directory so that after a + * reboot we can easily identify obsolete directories. */ + + r = sd_id128_get_boot(&boot_id); + if (r < 0) + return r; + + x = strjoin(prefix, "/systemd-private-", sd_id128_to_string(boot_id, bid), "-", id, "-XXXXXX", NULL); if (!x) return -ENOMEM; @@ -421,3 +693,19 @@ fail: return r; } + +static const char *const protect_home_table[_PROTECT_HOME_MAX] = { + [PROTECT_HOME_NO] = "no", + [PROTECT_HOME_YES] = "yes", + [PROTECT_HOME_READ_ONLY] = "read-only", +}; + +DEFINE_STRING_TABLE_LOOKUP(protect_home, ProtectHome); + +static const char *const protect_system_table[_PROTECT_SYSTEM_MAX] = { + [PROTECT_SYSTEM_NO] = "no", + [PROTECT_SYSTEM_YES] = "yes", + [PROTECT_SYSTEM_FULL] = "full", +}; + +DEFINE_STRING_TABLE_LOOKUP(protect_system, ProtectSystem);