| /*-*- Mode: C; c-basic-offset: 8; indent-tabs-mode: nil -*-*/ |
| |
| /*** |
| This file is part of systemd. |
| |
| Copyright 2010 Lennart Poettering |
| |
| systemd is free software; you can redistribute it and/or modify it |
| under the terms of the GNU Lesser General Public License as published by |
| the Free Software Foundation; either version 2.1 of the License, or |
| (at your option) any later version. |
| |
| systemd is distributed in the hope that it will be useful, but |
| WITHOUT ANY WARRANTY; without even the implied warranty of |
| MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| Lesser General Public License for more details. |
| |
| You should have received a copy of the GNU Lesser General Public License |
| along with systemd; If not, see <http://www.gnu.org/licenses/>. |
| ***/ |
| |
| #include <errno.h> |
| #include <string.h> |
| |
| #include "path-util.h" |
| #include "bus-label.h" |
| #include "util.h" |
| #include "unit-name.h" |
| #include "def.h" |
| #include "strv.h" |
| |
| #define VALID_CHARS \ |
| DIGITS LETTERS \ |
| ":-_.\\" |
| |
| bool unit_name_is_valid(const char *n, UnitNameFlags flags) { |
| const char *e, *i, *at; |
| |
| assert((flags & ~(UNIT_NAME_PLAIN|UNIT_NAME_INSTANCE|UNIT_NAME_TEMPLATE)) == 0); |
| |
| if (_unlikely_(flags == 0)) |
| return false; |
| |
| if (isempty(n)) |
| return false; |
| |
| if (strlen(n) >= UNIT_NAME_MAX) |
| return false; |
| |
| e = strrchr(n, '.'); |
| if (!e || e == n) |
| return false; |
| |
| if (unit_type_from_string(e + 1) < 0) |
| return false; |
| |
| for (i = n, at = NULL; i < e; i++) { |
| |
| if (*i == '@' && !at) |
| at = i; |
| |
| if (!strchr("@" VALID_CHARS, *i)) |
| return false; |
| } |
| |
| if (at == n) |
| return false; |
| |
| if (flags & UNIT_NAME_PLAIN) |
| if (!at) |
| return true; |
| |
| if (flags & UNIT_NAME_INSTANCE) |
| if (at && e > at + 1) |
| return true; |
| |
| if (flags & UNIT_NAME_TEMPLATE) |
| if (at && e == at + 1) |
| return true; |
| |
| return false; |
| } |
| |
| bool unit_prefix_is_valid(const char *p) { |
| |
| /* We don't allow additional @ in the prefix string */ |
| |
| if (isempty(p)) |
| return false; |
| |
| return in_charset(p, VALID_CHARS); |
| } |
| |
| bool unit_instance_is_valid(const char *i) { |
| |
| /* The max length depends on the length of the string, so we |
| * don't really check this here. */ |
| |
| if (isempty(i)) |
| return false; |
| |
| /* We allow additional @ in the instance string, we do not |
| * allow them in the prefix! */ |
| |
| return in_charset(i, "@" VALID_CHARS); |
| } |
| |
| bool unit_suffix_is_valid(const char *s) { |
| if (isempty(s)) |
| return false; |
| |
| if (s[0] != '.') |
| return false; |
| |
| if (unit_type_from_string(s + 1) < 0) |
| return false; |
| |
| return true; |
| } |
| |
| int unit_name_to_prefix(const char *n, char **ret) { |
| const char *p; |
| char *s; |
| |
| assert(n); |
| assert(ret); |
| |
| if (!unit_name_is_valid(n, UNIT_NAME_ANY)) |
| return -EINVAL; |
| |
| p = strchr(n, '@'); |
| if (!p) |
| p = strrchr(n, '.'); |
| |
| assert_se(p); |
| |
| s = strndup(n, p - n); |
| if (!s) |
| return -ENOMEM; |
| |
| *ret = s; |
| return 0; |
| } |
| |
| int unit_name_to_instance(const char *n, char **instance) { |
| const char *p, *d; |
| char *i; |
| |
| assert(n); |
| assert(instance); |
| |
| if (!unit_name_is_valid(n, UNIT_NAME_ANY)) |
| return -EINVAL; |
| |
| /* Everything past the first @ and before the last . is the instance */ |
| p = strchr(n, '@'); |
| if (!p) { |
| *instance = NULL; |
| return 0; |
| } |
| |
| p++; |
| |
| d = strrchr(p, '.'); |
| if (!d) |
| return -EINVAL; |
| |
| i = strndup(p, d-p); |
| if (!i) |
| return -ENOMEM; |
| |
| *instance = i; |
| return 1; |
| } |
| |
| int unit_name_to_prefix_and_instance(const char *n, char **ret) { |
| const char *d; |
| char *s; |
| |
| assert(n); |
| assert(ret); |
| |
| if (!unit_name_is_valid(n, UNIT_NAME_ANY)) |
| return -EINVAL; |
| |
| d = strrchr(n, '.'); |
| if (!d) |
| return -EINVAL; |
| |
| s = strndup(n, d - n); |
| if (!s) |
| return -ENOMEM; |
| |
| *ret = s; |
| return 0; |
| } |
| |
| UnitType unit_name_to_type(const char *n) { |
| const char *e; |
| |
| assert(n); |
| |
| if (!unit_name_is_valid(n, UNIT_NAME_ANY)) |
| return _UNIT_TYPE_INVALID; |
| |
| assert_se(e = strrchr(n, '.')); |
| |
| return unit_type_from_string(e + 1); |
| } |
| |
| int unit_name_change_suffix(const char *n, const char *suffix, char **ret) { |
| char *e, *s; |
| size_t a, b; |
| |
| assert(n); |
| assert(suffix); |
| assert(ret); |
| |
| if (!unit_name_is_valid(n, UNIT_NAME_ANY)) |
| return -EINVAL; |
| |
| if (!unit_suffix_is_valid(suffix)) |
| return -EINVAL; |
| |
| assert_se(e = strrchr(n, '.')); |
| |
| a = e - n; |
| b = strlen(suffix); |
| |
| s = new(char, a + b + 1); |
| if (!s) |
| return -ENOMEM; |
| |
| strcpy(mempcpy(s, n, a), suffix); |
| *ret = s; |
| |
| return 0; |
| } |
| |
| int unit_name_build(const char *prefix, const char *instance, const char *suffix, char **ret) { |
| char *s; |
| |
| assert(prefix); |
| assert(suffix); |
| assert(ret); |
| |
| if (!unit_prefix_is_valid(prefix)) |
| return -EINVAL; |
| |
| if (instance && !unit_instance_is_valid(instance)) |
| return -EINVAL; |
| |
| if (!unit_suffix_is_valid(suffix)) |
| return -EINVAL; |
| |
| if (!instance) |
| s = strappend(prefix, suffix); |
| else |
| s = strjoin(prefix, "@", instance, suffix, NULL); |
| if (!s) |
| return -ENOMEM; |
| |
| *ret = s; |
| return 0; |
| } |
| |
| static char *do_escape_char(char c, char *t) { |
| assert(t); |
| |
| *(t++) = '\\'; |
| *(t++) = 'x'; |
| *(t++) = hexchar(c >> 4); |
| *(t++) = hexchar(c); |
| |
| return t; |
| } |
| |
| static char *do_escape(const char *f, char *t) { |
| assert(f); |
| assert(t); |
| |
| /* do not create units with a leading '.', like for "/.dotdir" mount points */ |
| if (*f == '.') { |
| t = do_escape_char(*f, t); |
| f++; |
| } |
| |
| for (; *f; f++) { |
| if (*f == '/') |
| *(t++) = '-'; |
| else if (*f == '-' || *f == '\\' || !strchr(VALID_CHARS, *f)) |
| t = do_escape_char(*f, t); |
| else |
| *(t++) = *f; |
| } |
| |
| return t; |
| } |
| |
| char *unit_name_escape(const char *f) { |
| char *r, *t; |
| |
| assert(f); |
| |
| r = new(char, strlen(f)*4+1); |
| if (!r) |
| return NULL; |
| |
| t = do_escape(f, r); |
| *t = 0; |
| |
| return r; |
| } |
| |
| int unit_name_unescape(const char *f, char **ret) { |
| _cleanup_free_ char *r = NULL; |
| char *t; |
| |
| assert(f); |
| |
| r = strdup(f); |
| if (!r) |
| return -ENOMEM; |
| |
| for (t = r; *f; f++) { |
| if (*f == '-') |
| *(t++) = '/'; |
| else if (*f == '\\') { |
| int a, b; |
| |
| if (f[1] != 'x') |
| return -EINVAL; |
| |
| a = unhexchar(f[2]); |
| if (a < 0) |
| return -EINVAL; |
| |
| b = unhexchar(f[3]); |
| if (b < 0) |
| return -EINVAL; |
| |
| *(t++) = (char) (((uint8_t) a << 4U) | (uint8_t) b); |
| f += 3; |
| } else |
| *(t++) = *f; |
| } |
| |
| *t = 0; |
| |
| *ret = r; |
| r = NULL; |
| |
| return 0; |
| } |
| |
| int unit_name_path_escape(const char *f, char **ret) { |
| char *p, *s; |
| |
| assert(f); |
| assert(ret); |
| |
| p = strdupa(f); |
| if (!p) |
| return -ENOMEM; |
| |
| path_kill_slashes(p); |
| |
| if (STR_IN_SET(p, "/", "")) |
| s = strdup("-"); |
| else { |
| char *e; |
| |
| if (!path_is_safe(p)) |
| return -EINVAL; |
| |
| /* Truncate trailing slashes */ |
| e = endswith(p, "/"); |
| if (e) |
| *e = 0; |
| |
| /* Truncate leading slashes */ |
| if (p[0] == '/') |
| p++; |
| |
| s = unit_name_escape(p); |
| } |
| if (!s) |
| return -ENOMEM; |
| |
| *ret = s; |
| return 0; |
| } |
| |
| int unit_name_path_unescape(const char *f, char **ret) { |
| char *s; |
| int r; |
| |
| assert(f); |
| |
| if (isempty(f)) |
| return -EINVAL; |
| |
| if (streq(f, "-")) { |
| s = strdup("/"); |
| if (!s) |
| return -ENOMEM; |
| } else { |
| char *w; |
| |
| r = unit_name_unescape(f, &w); |
| if (r < 0) |
| return r; |
| |
| /* Don't accept trailing or leading slashes */ |
| if (startswith(w, "/") || endswith(w, "/")) { |
| free(w); |
| return -EINVAL; |
| } |
| |
| /* Prefix a slash again */ |
| s = strappend("/", w); |
| free(w); |
| if (!s) |
| return -ENOMEM; |
| |
| if (!path_is_safe(s)) { |
| free(s); |
| return -EINVAL; |
| } |
| } |
| |
| if (ret) |
| *ret = s; |
| else |
| free(s); |
| |
| return 0; |
| } |
| |
| int unit_name_replace_instance(const char *f, const char *i, char **ret) { |
| const char *p, *e; |
| char *s; |
| size_t a, b; |
| |
| assert(f); |
| assert(i); |
| assert(ret); |
| |
| if (!unit_name_is_valid(f, UNIT_NAME_INSTANCE|UNIT_NAME_TEMPLATE)) |
| return -EINVAL; |
| if (!unit_instance_is_valid(i)) |
| return -EINVAL; |
| |
| assert_se(p = strchr(f, '@')); |
| assert_se(e = strrchr(f, '.')); |
| |
| a = p - f; |
| b = strlen(i); |
| |
| s = new(char, a + 1 + b + strlen(e) + 1); |
| if (!s) |
| return -ENOMEM; |
| |
| strcpy(mempcpy(mempcpy(s, f, a + 1), i, b), e); |
| |
| *ret = s; |
| return 0; |
| } |
| |
| int unit_name_template(const char *f, char **ret) { |
| const char *p, *e; |
| char *s; |
| size_t a; |
| |
| assert(f); |
| assert(ret); |
| |
| if (!unit_name_is_valid(f, UNIT_NAME_INSTANCE|UNIT_NAME_TEMPLATE)) |
| return -EINVAL; |
| |
| assert_se(p = strchr(f, '@')); |
| assert_se(e = strrchr(f, '.')); |
| |
| a = p - f; |
| |
| s = new(char, a + 1 + strlen(e) + 1); |
| if (!s) |
| return -ENOMEM; |
| |
| strcpy(mempcpy(s, f, a + 1), e); |
| |
| *ret = s; |
| return 0; |
| } |
| |
| int unit_name_from_path(const char *path, const char *suffix, char **ret) { |
| _cleanup_free_ char *p = NULL; |
| char *s = NULL; |
| int r; |
| |
| assert(path); |
| assert(suffix); |
| assert(ret); |
| |
| if (!unit_suffix_is_valid(suffix)) |
| return -EINVAL; |
| |
| r = unit_name_path_escape(path, &p); |
| if (r < 0) |
| return r; |
| |
| s = strappend(p, suffix); |
| if (!s) |
| return -ENOMEM; |
| |
| *ret = s; |
| return 0; |
| } |
| |
| int unit_name_from_path_instance(const char *prefix, const char *path, const char *suffix, char **ret) { |
| _cleanup_free_ char *p = NULL; |
| char *s; |
| int r; |
| |
| assert(prefix); |
| assert(path); |
| assert(suffix); |
| assert(ret); |
| |
| if (!unit_prefix_is_valid(prefix)) |
| return -EINVAL; |
| |
| if (!unit_suffix_is_valid(suffix)) |
| return -EINVAL; |
| |
| r = unit_name_path_escape(path, &p); |
| if (r < 0) |
| return r; |
| |
| s = strjoin(prefix, "@", p, suffix, NULL); |
| if (!s) |
| return -ENOMEM; |
| |
| *ret = s; |
| return 0; |
| } |
| |
| int unit_name_to_path(const char *name, char **ret) { |
| _cleanup_free_ char *prefix = NULL; |
| int r; |
| |
| assert(name); |
| |
| r = unit_name_to_prefix(name, &prefix); |
| if (r < 0) |
| return r; |
| |
| return unit_name_path_unescape(prefix, ret); |
| } |
| |
| char *unit_dbus_path_from_name(const char *name) { |
| _cleanup_free_ char *e = NULL; |
| |
| assert(name); |
| |
| e = bus_label_escape(name); |
| if (!e) |
| return NULL; |
| |
| return strappend("/org/freedesktop/systemd1/unit/", e); |
| } |
| |
| int unit_name_from_dbus_path(const char *path, char **name) { |
| const char *e; |
| char *n; |
| |
| e = startswith(path, "/org/freedesktop/systemd1/unit/"); |
| if (!e) |
| return -EINVAL; |
| |
| n = bus_label_unescape(e); |
| if (!n) |
| return -ENOMEM; |
| |
| *name = n; |
| return 0; |
| } |
| |
| const char* unit_dbus_interface_from_type(UnitType t) { |
| |
| static const char *const table[_UNIT_TYPE_MAX] = { |
| [UNIT_SERVICE] = "org.freedesktop.systemd1.Service", |
| [UNIT_SOCKET] = "org.freedesktop.systemd1.Socket", |
| [UNIT_BUSNAME] = "org.freedesktop.systemd1.BusName", |
| [UNIT_TARGET] = "org.freedesktop.systemd1.Target", |
| [UNIT_SNAPSHOT] = "org.freedesktop.systemd1.Snapshot", |
| [UNIT_DEVICE] = "org.freedesktop.systemd1.Device", |
| [UNIT_MOUNT] = "org.freedesktop.systemd1.Mount", |
| [UNIT_AUTOMOUNT] = "org.freedesktop.systemd1.Automount", |
| [UNIT_SWAP] = "org.freedesktop.systemd1.Swap", |
| [UNIT_TIMER] = "org.freedesktop.systemd1.Timer", |
| [UNIT_PATH] = "org.freedesktop.systemd1.Path", |
| [UNIT_SLICE] = "org.freedesktop.systemd1.Slice", |
| [UNIT_SCOPE] = "org.freedesktop.systemd1.Scope", |
| }; |
| |
| if (t < 0) |
| return NULL; |
| if (t >= _UNIT_TYPE_MAX) |
| return NULL; |
| |
| return table[t]; |
| } |
| |
| const char *unit_dbus_interface_from_name(const char *name) { |
| UnitType t; |
| |
| t = unit_name_to_type(name); |
| if (t < 0) |
| return NULL; |
| |
| return unit_dbus_interface_from_type(t); |
| } |
| |
| static char *do_escape_mangle(const char *f, UnitNameMangle allow_globs, char *t) { |
| const char *valid_chars; |
| |
| assert(f); |
| assert(IN_SET(allow_globs, UNIT_NAME_GLOB, UNIT_NAME_NOGLOB)); |
| assert(t); |
| |
| /* We'll only escape the obvious characters here, to play |
| * safe. */ |
| |
| valid_chars = allow_globs == UNIT_NAME_GLOB ? "@" VALID_CHARS "[]!-*?" : "@" VALID_CHARS; |
| |
| for (; *f; f++) { |
| if (*f == '/') |
| *(t++) = '-'; |
| else if (!strchr(valid_chars, *f)) |
| t = do_escape_char(*f, t); |
| else |
| *(t++) = *f; |
| } |
| |
| return t; |
| } |
| |
| /** |
| * Convert a string to a unit name. /dev/blah is converted to dev-blah.device, |
| * /blah/blah is converted to blah-blah.mount, anything else is left alone, |
| * except that @suffix is appended if a valid unit suffix is not present. |
| * |
| * If @allow_globs, globs characters are preserved. Otherwise they are escaped. |
| */ |
| int unit_name_mangle_with_suffix(const char *name, UnitNameMangle allow_globs, const char *suffix, char **ret) { |
| char *s, *t; |
| int r; |
| |
| assert(name); |
| assert(suffix); |
| assert(ret); |
| |
| if (isempty(name)) /* We cannot mangle empty unit names to become valid, sorry. */ |
| return -EINVAL; |
| |
| if (!unit_suffix_is_valid(suffix)) |
| return -EINVAL; |
| |
| if (unit_name_is_valid(name, UNIT_NAME_ANY)) { |
| /* No mangling necessary... */ |
| s = strdup(name); |
| if (!s) |
| return -ENOMEM; |
| |
| *ret = s; |
| return 0; |
| } |
| |
| if (is_device_path(name)) { |
| r = unit_name_from_path(name, ".device", ret); |
| if (r >= 0) |
| return 1; |
| if (r != -EINVAL) |
| return r; |
| } |
| |
| if (path_is_absolute(name)) { |
| r = unit_name_from_path(name, ".mount", ret); |
| if (r >= 0) |
| return 1; |
| if (r != -EINVAL) |
| return r; |
| } |
| |
| s = new(char, strlen(name) * 4 + strlen(suffix) + 1); |
| if (!s) |
| return -ENOMEM; |
| |
| t = do_escape_mangle(name, allow_globs, s); |
| *t = 0; |
| |
| if (unit_name_to_type(s) < 0) |
| strcpy(t, suffix); |
| |
| *ret = s; |
| return 1; |
| } |
| |
| int slice_build_parent_slice(const char *slice, char **ret) { |
| char *s, *dash; |
| int r; |
| |
| assert(slice); |
| assert(ret); |
| |
| if (!slice_name_is_valid(slice)) |
| return -EINVAL; |
| |
| if (streq(slice, "-.slice")) { |
| *ret = NULL; |
| return 0; |
| } |
| |
| s = strdup(slice); |
| if (!s) |
| return -ENOMEM; |
| |
| dash = strrchr(s, '-'); |
| if (dash) |
| strcpy(dash, ".slice"); |
| else { |
| r = free_and_strdup(&s, "-.slice"); |
| if (r < 0) { |
| free(s); |
| return r; |
| } |
| } |
| |
| *ret = s; |
| return 1; |
| } |
| |
| int slice_build_subslice(const char *slice, const char*name, char **ret) { |
| char *subslice; |
| |
| assert(slice); |
| assert(name); |
| assert(ret); |
| |
| if (!slice_name_is_valid(slice)) |
| return -EINVAL; |
| |
| if (!unit_prefix_is_valid(name)) |
| return -EINVAL; |
| |
| if (streq(slice, "-.slice")) |
| subslice = strappend(name, ".slice"); |
| else { |
| char *e; |
| |
| assert_se(e = endswith(slice, ".slice")); |
| |
| subslice = new(char, (e - slice) + 1 + strlen(name) + 6 + 1); |
| if (!subslice) |
| return -ENOMEM; |
| |
| stpcpy(stpcpy(stpcpy(mempcpy(subslice, slice, e - slice), "-"), name), ".slice"); |
| } |
| |
| *ret = subslice; |
| return 0; |
| } |
| |
| bool slice_name_is_valid(const char *name) { |
| const char *p, *e; |
| bool dash = false; |
| |
| if (!unit_name_is_valid(name, UNIT_NAME_PLAIN)) |
| return false; |
| |
| if (streq(name, "-.slice")) |
| return true; |
| |
| e = endswith(name, ".slice"); |
| if (!e) |
| return false; |
| |
| for (p = name; p < e; p++) { |
| |
| if (*p == '-') { |
| |
| /* Don't allow initial dash */ |
| if (p == name) |
| return false; |
| |
| /* Don't allow multiple dashes */ |
| if (dash) |
| return false; |
| |
| dash = true; |
| } else |
| dash = false; |
| } |
| |
| /* Don't allow trailing hash */ |
| if (dash) |
| return false; |
| |
| return true; |
| } |
| |
| static const char* const unit_type_table[_UNIT_TYPE_MAX] = { |
| [UNIT_SERVICE] = "service", |
| [UNIT_SOCKET] = "socket", |
| [UNIT_BUSNAME] = "busname", |
| [UNIT_TARGET] = "target", |
| [UNIT_SNAPSHOT] = "snapshot", |
| [UNIT_DEVICE] = "device", |
| [UNIT_MOUNT] = "mount", |
| [UNIT_AUTOMOUNT] = "automount", |
| [UNIT_SWAP] = "swap", |
| [UNIT_TIMER] = "timer", |
| [UNIT_PATH] = "path", |
| [UNIT_SLICE] = "slice", |
| [UNIT_SCOPE] = "scope", |
| }; |
| |
| DEFINE_STRING_TABLE_LOOKUP(unit_type, UnitType); |
| |
| static const char* const unit_load_state_table[_UNIT_LOAD_STATE_MAX] = { |
| [UNIT_STUB] = "stub", |
| [UNIT_LOADED] = "loaded", |
| [UNIT_NOT_FOUND] = "not-found", |
| [UNIT_ERROR] = "error", |
| [UNIT_MERGED] = "merged", |
| [UNIT_MASKED] = "masked" |
| }; |
| |
| DEFINE_STRING_TABLE_LOOKUP(unit_load_state, UnitLoadState); |
| |
| static const char* const unit_active_state_table[_UNIT_ACTIVE_STATE_MAX] = { |
| [UNIT_ACTIVE] = "active", |
| [UNIT_RELOADING] = "reloading", |
| [UNIT_INACTIVE] = "inactive", |
| [UNIT_FAILED] = "failed", |
| [UNIT_ACTIVATING] = "activating", |
| [UNIT_DEACTIVATING] = "deactivating" |
| }; |
| |
| DEFINE_STRING_TABLE_LOOKUP(unit_active_state, UnitActiveState); |
| |
| static const char* const automount_state_table[_AUTOMOUNT_STATE_MAX] = { |
| [AUTOMOUNT_DEAD] = "dead", |
| [AUTOMOUNT_WAITING] = "waiting", |
| [AUTOMOUNT_RUNNING] = "running", |
| [AUTOMOUNT_FAILED] = "failed" |
| }; |
| |
| DEFINE_STRING_TABLE_LOOKUP(automount_state, AutomountState); |
| |
| static const char* const busname_state_table[_BUSNAME_STATE_MAX] = { |
| [BUSNAME_DEAD] = "dead", |
| [BUSNAME_MAKING] = "making", |
| [BUSNAME_REGISTERED] = "registered", |
| [BUSNAME_LISTENING] = "listening", |
| [BUSNAME_RUNNING] = "running", |
| [BUSNAME_SIGTERM] = "sigterm", |
| [BUSNAME_SIGKILL] = "sigkill", |
| [BUSNAME_FAILED] = "failed", |
| }; |
| |
| DEFINE_STRING_TABLE_LOOKUP(busname_state, BusNameState); |
| |
| static const char* const device_state_table[_DEVICE_STATE_MAX] = { |
| [DEVICE_DEAD] = "dead", |
| [DEVICE_TENTATIVE] = "tentative", |
| [DEVICE_PLUGGED] = "plugged", |
| }; |
| |
| DEFINE_STRING_TABLE_LOOKUP(device_state, DeviceState); |
| |
| static const char* const mount_state_table[_MOUNT_STATE_MAX] = { |
| [MOUNT_DEAD] = "dead", |
| [MOUNT_MOUNTING] = "mounting", |
| [MOUNT_MOUNTING_DONE] = "mounting-done", |
| [MOUNT_MOUNTED] = "mounted", |
| [MOUNT_REMOUNTING] = "remounting", |
| [MOUNT_UNMOUNTING] = "unmounting", |
| [MOUNT_MOUNTING_SIGTERM] = "mounting-sigterm", |
| [MOUNT_MOUNTING_SIGKILL] = "mounting-sigkill", |
| [MOUNT_REMOUNTING_SIGTERM] = "remounting-sigterm", |
| [MOUNT_REMOUNTING_SIGKILL] = "remounting-sigkill", |
| [MOUNT_UNMOUNTING_SIGTERM] = "unmounting-sigterm", |
| [MOUNT_UNMOUNTING_SIGKILL] = "unmounting-sigkill", |
| [MOUNT_FAILED] = "failed" |
| }; |
| |
| DEFINE_STRING_TABLE_LOOKUP(mount_state, MountState); |
| |
| static const char* const path_state_table[_PATH_STATE_MAX] = { |
| [PATH_DEAD] = "dead", |
| [PATH_WAITING] = "waiting", |
| [PATH_RUNNING] = "running", |
| [PATH_FAILED] = "failed" |
| }; |
| |
| DEFINE_STRING_TABLE_LOOKUP(path_state, PathState); |
| |
| static const char* const scope_state_table[_SCOPE_STATE_MAX] = { |
| [SCOPE_DEAD] = "dead", |
| [SCOPE_RUNNING] = "running", |
| [SCOPE_ABANDONED] = "abandoned", |
| [SCOPE_STOP_SIGTERM] = "stop-sigterm", |
| [SCOPE_STOP_SIGKILL] = "stop-sigkill", |
| [SCOPE_FAILED] = "failed", |
| }; |
| |
| DEFINE_STRING_TABLE_LOOKUP(scope_state, ScopeState); |
| |
| static const char* const service_state_table[_SERVICE_STATE_MAX] = { |
| [SERVICE_DEAD] = "dead", |
| [SERVICE_START_PRE] = "start-pre", |
| [SERVICE_START] = "start", |
| [SERVICE_START_POST] = "start-post", |
| [SERVICE_RUNNING] = "running", |
| [SERVICE_EXITED] = "exited", |
| [SERVICE_RELOAD] = "reload", |
| [SERVICE_STOP] = "stop", |
| [SERVICE_STOP_SIGABRT] = "stop-sigabrt", |
| [SERVICE_STOP_SIGTERM] = "stop-sigterm", |
| [SERVICE_STOP_SIGKILL] = "stop-sigkill", |
| [SERVICE_STOP_POST] = "stop-post", |
| [SERVICE_FINAL_SIGTERM] = "final-sigterm", |
| [SERVICE_FINAL_SIGKILL] = "final-sigkill", |
| [SERVICE_FAILED] = "failed", |
| [SERVICE_AUTO_RESTART] = "auto-restart", |
| }; |
| |
| DEFINE_STRING_TABLE_LOOKUP(service_state, ServiceState); |
| |
| static const char* const slice_state_table[_SLICE_STATE_MAX] = { |
| [SLICE_DEAD] = "dead", |
| [SLICE_ACTIVE] = "active" |
| }; |
| |
| DEFINE_STRING_TABLE_LOOKUP(slice_state, SliceState); |
| |
| static const char* const snapshot_state_table[_SNAPSHOT_STATE_MAX] = { |
| [SNAPSHOT_DEAD] = "dead", |
| [SNAPSHOT_ACTIVE] = "active" |
| }; |
| |
| DEFINE_STRING_TABLE_LOOKUP(snapshot_state, SnapshotState); |
| |
| static const char* const socket_state_table[_SOCKET_STATE_MAX] = { |
| [SOCKET_DEAD] = "dead", |
| [SOCKET_START_PRE] = "start-pre", |
| [SOCKET_START_CHOWN] = "start-chown", |
| [SOCKET_START_POST] = "start-post", |
| [SOCKET_LISTENING] = "listening", |
| [SOCKET_RUNNING] = "running", |
| [SOCKET_STOP_PRE] = "stop-pre", |
| [SOCKET_STOP_PRE_SIGTERM] = "stop-pre-sigterm", |
| [SOCKET_STOP_PRE_SIGKILL] = "stop-pre-sigkill", |
| [SOCKET_STOP_POST] = "stop-post", |
| [SOCKET_FINAL_SIGTERM] = "final-sigterm", |
| [SOCKET_FINAL_SIGKILL] = "final-sigkill", |
| [SOCKET_FAILED] = "failed" |
| }; |
| |
| DEFINE_STRING_TABLE_LOOKUP(socket_state, SocketState); |
| |
| static const char* const swap_state_table[_SWAP_STATE_MAX] = { |
| [SWAP_DEAD] = "dead", |
| [SWAP_ACTIVATING] = "activating", |
| [SWAP_ACTIVATING_DONE] = "activating-done", |
| [SWAP_ACTIVE] = "active", |
| [SWAP_DEACTIVATING] = "deactivating", |
| [SWAP_ACTIVATING_SIGTERM] = "activating-sigterm", |
| [SWAP_ACTIVATING_SIGKILL] = "activating-sigkill", |
| [SWAP_DEACTIVATING_SIGTERM] = "deactivating-sigterm", |
| [SWAP_DEACTIVATING_SIGKILL] = "deactivating-sigkill", |
| [SWAP_FAILED] = "failed" |
| }; |
| |
| DEFINE_STRING_TABLE_LOOKUP(swap_state, SwapState); |
| |
| static const char* const target_state_table[_TARGET_STATE_MAX] = { |
| [TARGET_DEAD] = "dead", |
| [TARGET_ACTIVE] = "active" |
| }; |
| |
| DEFINE_STRING_TABLE_LOOKUP(target_state, TargetState); |
| |
| static const char* const timer_state_table[_TIMER_STATE_MAX] = { |
| [TIMER_DEAD] = "dead", |
| [TIMER_WAITING] = "waiting", |
| [TIMER_RUNNING] = "running", |
| [TIMER_ELAPSED] = "elapsed", |
| [TIMER_FAILED] = "failed" |
| }; |
| |
| DEFINE_STRING_TABLE_LOOKUP(timer_state, TimerState); |
| |
| static const char* const unit_dependency_table[_UNIT_DEPENDENCY_MAX] = { |
| [UNIT_REQUIRES] = "Requires", |
| [UNIT_REQUIRES_OVERRIDABLE] = "RequiresOverridable", |
| [UNIT_REQUISITE] = "Requisite", |
| [UNIT_REQUISITE_OVERRIDABLE] = "RequisiteOverridable", |
| [UNIT_WANTS] = "Wants", |
| [UNIT_BINDS_TO] = "BindsTo", |
| [UNIT_PART_OF] = "PartOf", |
| [UNIT_REQUIRED_BY] = "RequiredBy", |
| [UNIT_REQUIRED_BY_OVERRIDABLE] = "RequiredByOverridable", |
| [UNIT_REQUISITE_OF] = "RequisiteOf", |
| [UNIT_REQUISITE_OF_OVERRIDABLE] = "RequisiteOfOverridable", |
| [UNIT_WANTED_BY] = "WantedBy", |
| [UNIT_BOUND_BY] = "BoundBy", |
| [UNIT_CONSISTS_OF] = "ConsistsOf", |
| [UNIT_CONFLICTS] = "Conflicts", |
| [UNIT_CONFLICTED_BY] = "ConflictedBy", |
| [UNIT_BEFORE] = "Before", |
| [UNIT_AFTER] = "After", |
| [UNIT_ON_FAILURE] = "OnFailure", |
| [UNIT_TRIGGERS] = "Triggers", |
| [UNIT_TRIGGERED_BY] = "TriggeredBy", |
| [UNIT_PROPAGATES_RELOAD_TO] = "PropagatesReloadTo", |
| [UNIT_RELOAD_PROPAGATED_FROM] = "ReloadPropagatedFrom", |
| [UNIT_JOINS_NAMESPACE_OF] = "JoinsNamespaceOf", |
| [UNIT_REFERENCES] = "References", |
| [UNIT_REFERENCED_BY] = "ReferencedBy", |
| }; |
| |
| DEFINE_STRING_TABLE_LOOKUP(unit_dependency, UnitDependency); |