/*
* Copyright (C) 2008 The Android Open Source Project
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include <errno.h>
#include <fcntl.h>
#include <net/if.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/socket.h>
#include <sys/mount.h>
#include <sys/resource.h>
#include <sys/time.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/wait.h>
#include <unistd.h>
#include <linux/loop.h>
#include <ext4_crypt_init_extensions.h>
#include <selinux/selinux.h>
#include <selinux/label.h>
#include <fs_mgr.h>
#include <base/stringprintf.h>
#include <cutils/partition_utils.h>
#include <cutils/android_reboot.h>
#include <private/android_filesystem_config.h>
#include "init.h"
#include "keywords.h"
#include "property_service.h"
#include "devices.h"
#include "init_parser.h"
#include "util.h"
#include "log.h"
#define chmod DO_NOT_USE_CHMOD_USE_FCHMODAT_SYMLINK_NOFOLLOW
int add_environment(const char *name, const char *value);
// System call provided by bionic but not in any header file.
extern "C" int init_module(void *, unsigned long, const char *);
static int insmod(const char *filename, char *options)
{
char filename_val[PROP_VALUE_MAX];
if (expand_props(filename_val, filename, sizeof(filename_val)) == -1) {
ERROR("insmod: cannot expand '%s'\n", filename);
return -EINVAL;
}
std::string module;
if (!read_file(filename_val, &module)) {
return -1;
}
// TODO: use finit_module for >= 3.8 kernels.
return init_module(&module[0], module.size(), options);
}
static int __ifupdown(const char *interface, int up)
{
struct ifreq ifr;
int s, ret;
strlcpy(ifr.ifr_name, interface, IFNAMSIZ);
s = socket(AF_INET, SOCK_DGRAM, 0);
if (s < 0)
return -1;
ret = ioctl(s, SIOCGIFFLAGS, &ifr);
if (ret < 0) {
goto done;
}
if (up)
ifr.ifr_flags |= IFF_UP;
else
ifr.ifr_flags &= ~IFF_UP;
ret = ioctl(s, SIOCSIFFLAGS, &ifr);
done:
close(s);
return ret;
}
static void service_start_if_not_disabled(struct service *svc)
{
if (!(svc->flags & SVC_DISABLED)) {
service_start(svc, NULL);
} else {
svc->flags |= SVC_DISABLED_START;
}
}
int do_class_start(int nargs, char **args)
{
/* Starting a class does not start services
* which are explicitly disabled. They must
* be started individually.
*/
service_for_each_class(args[1], service_start_if_not_disabled);
return 0;
}
int do_class_stop(int nargs, char **args)
{
service_for_each_class(args[1], service_stop);
return 0;
}
int do_class_reset(int nargs, char **args)
{
service_for_each_class(args[1], service_reset);
return 0;
}
int do_domainname(int nargs, char **args)
{
return write_file("/proc/sys/kernel/domainname", args[1]);
}
int do_enable(int nargs, char **args)
{
struct service *svc;
svc = service_find_by_name(args[1]);
if (svc) {
svc->flags &= ~(SVC_DISABLED | SVC_RC_DISABLED);
if (svc->flags & SVC_DISABLED_START) {
service_start(svc, NULL);
}
} else {
return -1;
}
return 0;
}
int do_exec(int nargs, char** args) {
service* svc = make_exec_oneshot_service(nargs, args);
if (svc == NULL) {
return -1;
}
service_start(svc, NULL);
return 0;
}
int do_export(int nargs, char **args)
{
return add_environment(args[1], args[2]);
}
int do_hostname(int nargs, char **args)
{
return write_file("/proc/sys/kernel/hostname", args[1]);
}
int do_ifup(int nargs, char **args)
{
return __ifupdown(args[1], 1);
}
static int do_insmod_inner(int nargs, char **args, int opt_len)
{
char options[opt_len + 1];
int i;
options[0] = '\0';
if (nargs > 2) {
strcpy(options, args[2]);
for (i = 3; i < nargs; ++i) {
strcat(options, " ");
strcat(options, args[i]);
}
}
return insmod(args[1], options);
}
int do_insmod(int nargs, char **args)
{
int i;
int size = 0;
if (nargs > 2) {
for (i = 2; i < nargs; ++i)
size += strlen(args[i]) + 1;
}
return do_insmod_inner(nargs, args, size);
}
int do_mkdir(int nargs, char **args)
{
mode_t mode = 0755;
int ret;
/* mkdir <path> [mode] [owner] [group] */
if (nargs >= 3) {
mode = strtoul(args[2], 0, 8);
}
ret = make_dir(args[1], mode);
/* chmod in case the directory already exists */
if (ret == -1 && errno == EEXIST) {
ret = fchmodat(AT_FDCWD, args[1], mode, AT_SYMLINK_NOFOLLOW);
}
if (ret == -1) {
return -errno;
}
if (nargs >= 4) {
uid_t uid = decode_uid(args[3]);
gid_t gid = -1;
if (nargs == 5) {
gid = decode_uid(args[4]);
}
if (lchown(args[1], uid, gid) == -1) {
return -errno;
}
/* chown may have cleared S_ISUID and S_ISGID, chmod again */
if (mode & (S_ISUID | S_ISGID)) {
ret = fchmodat(AT_FDCWD, args[1], mode, AT_SYMLINK_NOFOLLOW);
if (ret == -1) {
return -errno;
}
}
}
return e4crypt_set_directory_policy(args[1]);
}
static struct {
const char *name;
unsigned flag;
} mount_flags[] = {
{ "noatime", MS_NOATIME },
{ "noexec", MS_NOEXEC },
{ "nosuid", MS_NOSUID },
{ "nodev", MS_NODEV },
{ "nodiratime", MS_NODIRATIME },
{ "ro", MS_RDONLY },
{ "rw", 0 },
{ "remount", MS_REMOUNT },
{ "bind", MS_BIND },
{ "rec", MS_REC },
{ "unbindable", MS_UNBINDABLE },
{ "private", MS_PRIVATE },
{ "slave", MS_SLAVE },
{ "shared", MS_SHARED },
{ "defaults", 0 },
{ 0, 0 },
};
#define DATA_MNT_POINT "/data"
/* mount <type> <device> <path> <flags ...> <options> */
int do_mount(int nargs, char **args)
{
char tmp[64];
char *source, *target, *system;
char *options = NULL;
unsigned flags = 0;
int n, i;
int wait = 0;
for (n = 4; n < nargs; n++) {
for (i = 0; mount_flags[i].name; i++) {
if (!strcmp(args[n], mount_flags[i].name)) {
flags |= mount_flags[i].flag;
break;
}
}
if (!mount_flags[i].name) {
if (!strcmp(args[n], "wait"))
wait = 1;
/* if our last argument isn't a flag, wolf it up as an option string */
else if (n + 1 == nargs)
options = args[n];
}
}
system = args[1];
source = args[2];
target = args[3];
if (!strncmp(source, "mtd@", 4)) {
n = mtd_name_to_number(source + 4);
if (n < 0) {
return -1;
}
snprintf(tmp, sizeof(tmp), "/dev/block/mtdblock%d", n);
if (wait)
wait_for_file(tmp, COMMAND_RETRY_TIMEOUT);
if (mount(tmp, target, system, flags, options) < 0) {
return -1;
}
goto exit_success;
} else if (!strncmp(source, "loop@", 5)) {
int mode, loop, fd;
struct loop_info info;
mode = (flags & MS_RDONLY) ? O_RDONLY : O_RDWR;
fd = open(source + 5, mode | O_CLOEXEC);
if (fd < 0) {
return -1;
}
for (n = 0; ; n++) {
snprintf(tmp, sizeof(tmp), "/dev/block/loop%d", n);
loop = open(tmp, mode | O_CLOEXEC);
if (loop < 0) {
close(fd);
return -1;
}
/* if it is a blank loop device */
if (ioctl(loop, LOOP_GET_STATUS, &info) < 0 && errno == ENXIO) {
/* if it becomes our loop device */
if (ioctl(loop, LOOP_SET_FD, fd) >= 0) {
close(fd);
if (mount(tmp, target, system, flags, options) < 0) {
ioctl(loop, LOOP_CLR_FD, 0);
close(loop);
return -1;
}
close(loop);
goto exit_success;
}
}
close(loop);
}
close(fd);
ERROR("out of loopback devices");
return -1;
} else {
if (wait)
wait_for_file(source, COMMAND_RETRY_TIMEOUT);
if (mount(source, target, system, flags, options) < 0) {
return -1;
}
}
exit_success:
return 0;
}
static int wipe_data_via_recovery()
{
mkdir("/cache/recovery", 0700);
int fd = open("/cache/recovery/command", O_RDWR|O_CREAT|O_TRUNC|O_CLOEXEC, 0600);
if (fd >= 0) {
write(fd, "--wipe_data\n", strlen("--wipe_data\n") + 1);
write(fd, "--reason=wipe_data_via_recovery\n", strlen("--reason=wipe_data_via_recovery\n") + 1);
close(fd);
} else {
ERROR("could not open /cache/recovery/command\n");
return -1;
}
android_reboot(ANDROID_RB_RESTART2, 0, "recovery");
while (1) { pause(); } // never reached
}
/*
* This function might request a reboot, in which case it will
* not return.
*/
int do_mount_all(int nargs, char **args)
{
pid_t pid;
int ret = -1;
int child_ret = -1;
int status;
struct fstab *fstab;
if (nargs != 2) {
return -1;
}
/*
* Call fs_mgr_mount_all() to mount all filesystems. We fork(2) and
* do the call in the child to provide protection to the main init
* process if anything goes wrong (crash or memory leak), and wait for
* the child to finish in the parent.
*/
pid = fork();
if (pid > 0) {
/* Parent. Wait for the child to return */
int wp_ret = TEMP_FAILURE_RETRY(waitpid(pid, &status, 0));
if (wp_ret < 0) {
/* Unexpected error code. We will continue anyway. */
NOTICE("waitpid failed rc=%d: %s\n", wp_ret, strerror(errno));
}
if (WIFEXITED(status)) {
ret = WEXITSTATUS(status);
} else {
ret = -1;
}
} else if (pid == 0) {
/* child, call fs_mgr_mount_all() */
klog_set_level(6); /* So we can see what fs_mgr_mount_all() does */
fstab = fs_mgr_read_fstab(args[1]);
child_ret = fs_mgr_mount_all(fstab);
fs_mgr_free_fstab(fstab);
if (child_ret == -1) {
ERROR("fs_mgr_mount_all returned an error\n");
}
_exit(child_ret);
} else {
/* fork failed, return an error */
return -1;
}
if (ret == FS_MGR_MNTALL_DEV_NEEDS_ENCRYPTION) {
property_set("vold.decrypt", "trigger_encryption");
} else if (ret == FS_MGR_MNTALL_DEV_MIGHT_BE_ENCRYPTED) {
property_set("ro.crypto.state", "encrypted");
property_set("ro.crypto.type", "block");
property_set("vold.decrypt", "trigger_default_encryption");
} else if (ret == FS_MGR_MNTALL_DEV_NOT_ENCRYPTED) {
property_set("ro.crypto.state", "unencrypted");
/* If fs_mgr determined this is an unencrypted device, then trigger
* that action.
*/
action_for_each_trigger("nonencrypted", action_add_queue_tail);
} else if (ret == FS_MGR_MNTALL_DEV_NEEDS_RECOVERY) {
/* Setup a wipe via recovery, and reboot into recovery */
ERROR("fs_mgr_mount_all suggested recovery, so wiping data via recovery.\n");
ret = wipe_data_via_recovery();
/* If reboot worked, there is no return. */
} else if (ret == FS_MGR_MNTALL_DEV_DEFAULT_FILE_ENCRYPTED) {
if (e4crypt_install_keyring()) {
return -1;
}
property_set("ro.crypto.state", "encrypted");
property_set("ro.crypto.type", "file");
// Although encrypted, we have device key, so we do not need to
// do anything different from the nonencrypted case.
action_for_each_trigger("nonencrypted", action_add_queue_tail);
} else if (ret == FS_MGR_MNTALL_DEV_NON_DEFAULT_FILE_ENCRYPTED) {
if (e4crypt_install_keyring()) {
return -1;
}
property_set("ro.crypto.state", "encrypted");
property_set("ro.crypto.type", "file");
property_set("vold.decrypt", "trigger_restart_min_framework");
} else if (ret > 0) {
ERROR("fs_mgr_mount_all returned unexpected error %d\n", ret);
}
/* else ... < 0: error */
return ret;
}
int do_swapon_all(int nargs, char **args)
{
struct fstab *fstab;
int ret;
fstab = fs_mgr_read_fstab(args[1]);
ret = fs_mgr_swapon_all(fstab);
fs_mgr_free_fstab(fstab);
return ret;
}
int do_setprop(int nargs, char **args)
{
const char *name = args[1];
const char *value = args[2];
char prop_val[PROP_VALUE_MAX];
int ret;
ret = expand_props(prop_val, value, sizeof(prop_val));
if (ret) {
ERROR("cannot expand '%s' while assigning to '%s'\n", value, name);
return -EINVAL;
}
property_set(name, prop_val);
return 0;
}
int do_setrlimit(int nargs, char **args)
{
struct rlimit limit;
int resource;
resource = atoi(args[1]);
limit.rlim_cur = atoi(args[2]);
limit.rlim_max = atoi(args[3]);
return setrlimit(resource, &limit);
}
int do_start(int nargs, char **args)
{
struct service *svc;
svc = service_find_by_name(args[1]);
if (svc) {
service_start(svc, NULL);
}
return 0;
}
int do_stop(int nargs, char **args)
{
struct service *svc;
svc = service_find_by_name(args[1]);
if (svc) {
service_stop(svc);
}
return 0;
}
int do_restart(int nargs, char **args)
{
struct service *svc;
svc = service_find_by_name(args[1]);
if (svc) {
service_restart(svc);
}
return 0;
}
int do_powerctl(int nargs, char **args)
{
char command[PROP_VALUE_MAX];
int res;
int len = 0;
int cmd = 0;
const char *reboot_target;
res = expand_props(command, args[1], sizeof(command));
if (res) {
ERROR("powerctl: cannot expand '%s'\n", args[1]);
return -EINVAL;
}
if (strncmp(command, "shutdown", 8) == 0) {
cmd = ANDROID_RB_POWEROFF;
len = 8;
} else if (strncmp(command, "reboot", 6) == 0) {
cmd = ANDROID_RB_RESTART2;
len = 6;
} else {
ERROR("powerctl: unrecognized command '%s'\n", command);
return -EINVAL;
}
if (command[len] == ',') {
reboot_target = &command[len + 1];
} else if (command[len] == '\0') {
reboot_target = "";
} else {
ERROR("powerctl: unrecognized reboot target '%s'\n", &command[len]);
return -EINVAL;
}
return android_reboot(cmd, 0, reboot_target);
}
int do_trigger(int nargs, char **args)
{
action_for_each_trigger(args[1], action_add_queue_tail);
return 0;
}
int do_symlink(int nargs, char **args)
{
return symlink(args[1], args[2]);
}
int do_rm(int nargs, char **args)
{
return unlink(args[1]);
}
int do_rmdir(int nargs, char **args)
{
return rmdir(args[1]);
}
int do_sysclktz(int nargs, char **args)
{
struct timezone tz;
if (nargs != 2)
return -1;
memset(&tz, 0, sizeof(tz));
tz.tz_minuteswest = atoi(args[1]);
if (settimeofday(NULL, &tz))
return -1;
return 0;
}
int do_verity_load_state(int nargs, char **args) {
int mode = -1;
int rc = fs_mgr_load_verity_state(&mode);
if (rc == 0 && mode == VERITY_MODE_LOGGING) {
action_for_each_trigger("verity-logging", action_add_queue_tail);
}
return rc;
}
static void verity_update_property(fstab_rec *fstab, const char *mount_point, int mode, int status) {
property_set(android::base::StringPrintf("partition.%s.verified", mount_point).c_str(),
android::base::StringPrintf("%d", mode).c_str());
}
int do_verity_update_state(int nargs, char** args) {
return fs_mgr_update_verity_state(verity_update_property);
}
int do_write(int nargs, char **args)
{
const char *path = args[1];
const char *value = args[2];
char expanded_value[256];
if (expand_props(expanded_value, value, sizeof(expanded_value))) {
ERROR("cannot expand '%s' while writing to '%s'\n", value, path);
return -EINVAL;
}
return write_file(path, expanded_value);
}
int do_copy(int nargs, char **args)
{
char *buffer = NULL;
int rc = 0;
int fd1 = -1, fd2 = -1;
struct stat info;
int brtw, brtr;
char *p;
if (nargs != 3)
return -1;
if (stat(args[1], &info) < 0)
return -1;
if ((fd1 = open(args[1], O_RDONLY|O_CLOEXEC)) < 0)
goto out_err;
if ((fd2 = open(args[2], O_WRONLY|O_CREAT|O_TRUNC|O_CLOEXEC, 0660)) < 0)
goto out_err;
if (!(buffer = (char*) malloc(info.st_size)))
goto out_err;
p = buffer;
brtr = info.st_size;
while(brtr) {
rc = read(fd1, p, brtr);
if (rc < 0)
goto out_err;
if (rc == 0)
break;
p += rc;
brtr -= rc;
}
p = buffer;
brtw = info.st_size;
while(brtw) {
rc = write(fd2, p, brtw);
if (rc < 0)
goto out_err;
if (rc == 0)
break;
p += rc;
brtw -= rc;
}
rc = 0;
goto out;
out_err:
rc = -1;
out:
if (buffer)
free(buffer);
if (fd1 >= 0)
close(fd1);
if (fd2 >= 0)
close(fd2);
return rc;
}
int do_chown(int nargs, char **args) {
/* GID is optional. */
if (nargs == 3) {
if (lchown(args[2], decode_uid(args[1]), -1) == -1)
return -errno;
} else if (nargs == 4) {
if (lchown(args[3], decode_uid(args[1]), decode_uid(args[2])) == -1)
return -errno;
} else {
return -1;
}
return 0;
}
static mode_t get_mode(const char *s) {
mode_t mode = 0;
while (*s) {
if (*s >= '0' && *s <= '7') {
mode = (mode<<3) | (*s-'0');
} else {
return -1;
}
s++;
}
return mode;
}
int do_chmod(int nargs, char **args) {
mode_t mode = get_mode(args[1]);
if (fchmodat(AT_FDCWD, args[2], mode, AT_SYMLINK_NOFOLLOW) < 0) {
return -errno;
}
return 0;
}
int do_restorecon(int nargs, char **args) {
int i;
int ret = 0;
for (i = 1; i < nargs; i++) {
if (restorecon(args[i]) < 0)
ret = -errno;
}
return ret;
}
int do_restorecon_recursive(int nargs, char **args) {
int i;
int ret = 0;
for (i = 1; i < nargs; i++) {
if (restorecon_recursive(args[i]) < 0)
ret = -errno;
}
return ret;
}
int do_loglevel(int nargs, char **args) {
int log_level;
char log_level_str[PROP_VALUE_MAX] = "";
if (nargs != 2) {
ERROR("loglevel: missing argument\n");
return -EINVAL;
}
if (expand_props(log_level_str, args[1], sizeof(log_level_str))) {
ERROR("loglevel: cannot expand '%s'\n", args[1]);
return -EINVAL;
}
log_level = atoi(log_level_str);
if (log_level < KLOG_ERROR_LEVEL || log_level > KLOG_DEBUG_LEVEL) {
ERROR("loglevel: invalid log level'%d'\n", log_level);
return -EINVAL;
}
klog_set_level(log_level);
return 0;
}
int do_load_persist_props(int nargs, char **args) {
if (nargs == 1) {
load_persist_props();
return 0;
}
return -1;
}
int do_load_system_props(int nargs, char **args) {
if (nargs == 1) {
load_system_props();
return 0;
}
return -1;
}
int do_wait(int nargs, char **args)
{
if (nargs == 2) {
return wait_for_file(args[1], COMMAND_RETRY_TIMEOUT);
} else if (nargs == 3) {
return wait_for_file(args[1], atoi(args[2]));
} else
return -1;
}
/*
* Callback to make a directory from the ext4 code
*/
static int do_installkeys_ensure_dir_exists(const char* dir)
{
if (make_dir(dir, 0700) && errno != EEXIST) {
return -1;
}
return 0;
}
static bool is_file_crypto() {
char prop_value[PROP_VALUE_MAX] = {0};
property_get("ro.crypto.type", prop_value);
return strcmp(prop_value, "file") == 0;
}
int do_installkey(int nargs, char **args)
{
if (nargs != 2) {
return -1;
}
if (!is_file_crypto()) {
return 0;
}
return e4crypt_create_device_key(args[1],
do_installkeys_ensure_dir_exists);
}
int do_setusercryptopolicies(int nargs, char **args)
{
if (nargs != 2) {
return -1;
}
if (!is_file_crypto()) {
return 0;
}
return e4crypt_set_user_crypto_policies(args[1]);
}