mirror of
https://github.com/topjohnwu/Magisk.git
synced 2024-11-24 10:35:26 +00:00
321 lines
7.6 KiB
C++
321 lines
7.6 KiB
C++
/* proc_monitor.cpp - Monitor am_proc_start events and unmount
|
|
*
|
|
* We monitor the listed APK files from /data/app until they get opened
|
|
* via inotify to detect a new app launch.
|
|
*
|
|
* If it's a target we pause it ASAP, and fork a new process to join
|
|
* its mount namespace and do all the unmounting/mocking.
|
|
*/
|
|
|
|
#include <stdlib.h>
|
|
#include <stdio.h>
|
|
#include <string.h>
|
|
#include <unistd.h>
|
|
#include <fcntl.h>
|
|
#include <signal.h>
|
|
#include <pthread.h>
|
|
#include <sys/inotify.h>
|
|
#include <sys/types.h>
|
|
#include <sys/wait.h>
|
|
#include <sys/mount.h>
|
|
#include <vector>
|
|
#include <string>
|
|
#include <map>
|
|
|
|
#include <magisk.h>
|
|
#include <utils.h>
|
|
|
|
#include "magiskhide.h"
|
|
|
|
using namespace std;
|
|
|
|
extern char *system_block, *vendor_block, *data_block;
|
|
|
|
static int inotify_fd = -1;
|
|
|
|
#define EVENT_SIZE sizeof(struct inotify_event)
|
|
#define EVENT_BUF_LEN (1024 * (EVENT_SIZE + 16))
|
|
#define __ALIGN_EVENT __attribute__ ((aligned(__alignof__(struct inotify_event))))
|
|
|
|
// Workaround for the lack of pthread_cancel
|
|
static void term_thread(int) {
|
|
LOGD("proc_monitor: running cleanup\n");
|
|
hide_list.clear();
|
|
hide_uid.clear();
|
|
hide_enabled = false;
|
|
pthread_mutex_destroy(&list_lock);
|
|
close(inotify_fd);
|
|
inotify_fd = -1;
|
|
LOGD("proc_monitor: terminating\n");
|
|
pthread_exit(nullptr);
|
|
}
|
|
|
|
static inline int read_ns(const int pid, struct stat *st) {
|
|
char path[32];
|
|
sprintf(path, "/proc/%d/ns/mnt", pid);
|
|
return stat(path, st);
|
|
}
|
|
|
|
static inline void lazy_unmount(const char* mountpoint) {
|
|
if (umount2(mountpoint, MNT_DETACH) != -1)
|
|
LOGD("hide_daemon: Unmounted (%s)\n", mountpoint);
|
|
}
|
|
|
|
/* APK monitoring doesn't seem to require checking namespace
|
|
* separation from PPID. Preserve this function just in case */
|
|
#if 0
|
|
static inline int parse_ppid(const int pid) {
|
|
char path[32];
|
|
int ppid;
|
|
|
|
sprintf(path, "/proc/%d/stat", pid);
|
|
FILE *stat = fopen(path, "re");
|
|
if (stat == nullptr)
|
|
return -1;
|
|
|
|
/* PID COMM STATE PPID ..... */
|
|
fscanf(stat, "%*d %*s %*c %d", &ppid);
|
|
fclose(stat);
|
|
|
|
return ppid;
|
|
}
|
|
#endif
|
|
|
|
static bool is_pid_safetynet_process(const int pid) {
|
|
char path[32];
|
|
char buf[64];
|
|
int fd;
|
|
ssize_t len;
|
|
|
|
sprintf(path, "/proc/%d/cmdline", pid);
|
|
fd = open(path, O_RDONLY);
|
|
if (fd == -1)
|
|
return false;
|
|
|
|
len = read(fd, buf, sizeof(buf));
|
|
close(fd);
|
|
if (len == -1)
|
|
return false;
|
|
|
|
return !strcmp(buf, SAFETYNET_PROCESS);
|
|
}
|
|
|
|
static void hide_daemon(int pid) {
|
|
LOGD("hide_daemon: handling PID=[%d]\n", pid);
|
|
|
|
char buffer[4096];
|
|
vector<string> mounts;
|
|
|
|
manage_selinux();
|
|
clean_magisk_props();
|
|
|
|
if (switch_mnt_ns(pid))
|
|
goto exit;
|
|
|
|
snprintf(buffer, sizeof(buffer), "/proc/%d", pid);
|
|
chdir(buffer);
|
|
|
|
mounts = file_to_vector("mounts");
|
|
// Unmount dummy skeletons and /sbin links
|
|
for (auto &s : mounts) {
|
|
if (str_contains(s, "tmpfs /system/") || str_contains(s, "tmpfs /vendor/") ||
|
|
str_contains(s, "tmpfs /sbin")) {
|
|
sscanf(s.c_str(), "%*s %4096s", buffer);
|
|
lazy_unmount(buffer);
|
|
}
|
|
}
|
|
|
|
// Re-read mount infos
|
|
mounts = file_to_vector("mounts");
|
|
|
|
// Unmount everything under /system, /vendor, and data mounts
|
|
for (auto &s : mounts) {
|
|
if ((str_contains(s, " /system/") || str_contains(s, " /vendor/")) &&
|
|
(str_contains(s, system_block) || str_contains(s, vendor_block) || str_contains(s, data_block))) {
|
|
sscanf(s.c_str(), "%*s %4096s", buffer);
|
|
lazy_unmount(buffer);
|
|
}
|
|
}
|
|
|
|
exit:
|
|
// Send resume signal
|
|
kill(pid, SIGCONT);
|
|
_exit(0);
|
|
}
|
|
|
|
// A mapping from pid to namespace inode to avoid time-consuming GC
|
|
static map<int, uint64_t> pid_ns_map;
|
|
|
|
static bool process_pid(int pid) {
|
|
// We're only interested in PIDs > 1000
|
|
if (pid <= 1000)
|
|
return true;
|
|
|
|
struct stat ns;
|
|
int uid = get_uid(pid);
|
|
if (hide_uid.count(uid)) {
|
|
// Make sure we can read mount namespace
|
|
if (read_ns(pid, &ns))
|
|
return true;
|
|
|
|
// Check if it's a process we haven't already hijacked
|
|
auto pos = pid_ns_map.find(pid);
|
|
if (pos != pid_ns_map.end() && pos->second == ns.st_ino)
|
|
return true;
|
|
|
|
if (uid == gms_uid) {
|
|
// Check /proc/uid/cmdline to see if it's SAFETYNET_PROCESS
|
|
if (!is_pid_safetynet_process(pid))
|
|
return true;
|
|
|
|
LOGD("proc_monitor: " SAFETYNET_PROCESS "\n");
|
|
}
|
|
|
|
// Send pause signal ASAP
|
|
if (kill(pid, SIGSTOP) == -1)
|
|
return true;
|
|
|
|
pid_ns_map[pid] = ns.st_ino;
|
|
LOGI("proc_monitor: UID=[%d] PID=[%d] ns=[%llu]\n", uid, pid, ns.st_ino);
|
|
|
|
/*
|
|
* The setns system call do not support multithread processes
|
|
* We have to fork a new process, setns, then do the unmounts
|
|
*/
|
|
if (fork_dont_care() == 0)
|
|
hide_daemon(pid);
|
|
}
|
|
return true;
|
|
}
|
|
|
|
static int xinotify_add_watch(int fd, const char* path, uint32_t mask) {
|
|
int ret = inotify_add_watch(fd, path, mask);
|
|
if (ret >= 0) {
|
|
LOGI("proc_monitor: Monitoring %s\n", path);
|
|
} else {
|
|
PLOGE("proc_monitor: Monitor %s", path);
|
|
}
|
|
return ret;
|
|
}
|
|
|
|
static char *append_path(char *eof, const char *name) {
|
|
*(eof++) = '/';
|
|
char c;
|
|
while ((c = *(name++)))
|
|
*(eof++) = c;
|
|
*eof = '\0';
|
|
return eof;
|
|
}
|
|
|
|
#define DATA_APP "/data/app"
|
|
|
|
static void find_apks(char *path, char *eof) {
|
|
DIR *dir = opendir(path);
|
|
if (dir == nullptr)
|
|
return;
|
|
|
|
// assert(*eof == '\0');
|
|
|
|
struct dirent *entry;
|
|
while ((entry = xreaddir(dir))) {
|
|
if (strcmp(entry->d_name, ".") == 0 || strcmp(entry->d_name, "..") == 0)
|
|
continue;
|
|
if (entry->d_type == DT_DIR) {
|
|
find_apks(path, append_path(eof, entry->d_name));
|
|
*eof = '\0';
|
|
} else if (strend(entry->d_name, ".apk") == 0) {
|
|
// Path will be in this format: /data/app/[pkg]-[hash or 1 or 2]
|
|
char *dash = strchr(path, '-');
|
|
*dash = '\0';
|
|
for (auto &s : hide_list) {
|
|
if (s == path + sizeof(DATA_APP)) {
|
|
*dash = '-';
|
|
append_path(eof, entry->d_name);
|
|
xinotify_add_watch(inotify_fd, path, IN_OPEN | IN_DELETE);
|
|
*eof = '\0';
|
|
break;
|
|
}
|
|
}
|
|
*dash = '-';
|
|
break;
|
|
}
|
|
}
|
|
closedir(dir);
|
|
}
|
|
|
|
// Iterate through /data/app and search all .apk files
|
|
void update_inotify_mask() {
|
|
char buf[4096];
|
|
|
|
if (inotify_fd >= 0)
|
|
close(inotify_fd);
|
|
|
|
inotify_fd = inotify_init();
|
|
if (inotify_fd < 0) {
|
|
LOGE("proc_monitor: Cannot initialize inotify: %s\n", strerror(errno));
|
|
term_thread(TERM_THREAD);
|
|
}
|
|
|
|
LOGI("proc_monitor: Updating APK list\n");
|
|
strcpy(buf, DATA_APP);
|
|
pthread_mutex_lock(&list_lock);
|
|
find_apks(buf, buf + sizeof(DATA_APP) - 1);
|
|
pthread_mutex_unlock(&list_lock);
|
|
|
|
// Add /data/app itself to the watch list to detect app (un)installations/updates
|
|
xinotify_add_watch(inotify_fd, DATA_APP, IN_CLOSE_WRITE | IN_MOVED_TO | IN_DELETE);
|
|
}
|
|
|
|
void proc_monitor() {
|
|
// Unblock user signals
|
|
sigset_t block_set;
|
|
sigemptyset(&block_set);
|
|
sigaddset(&block_set, TERM_THREAD);
|
|
pthread_sigmask(SIG_UNBLOCK, &block_set, nullptr);
|
|
|
|
// Register the cancel signal
|
|
struct sigaction act{};
|
|
act.sa_handler = term_thread;
|
|
sigaction(TERM_THREAD, &act, nullptr);
|
|
|
|
if (access("/proc/1/ns/mnt", F_OK) != 0) {
|
|
LOGE("proc_monitor: Your kernel doesn't support mount namespace :(\n");
|
|
term_thread(TERM_THREAD);
|
|
}
|
|
|
|
// Read inotify events
|
|
struct inotify_event *event;
|
|
ssize_t len;
|
|
char *p;
|
|
char buffer[EVENT_BUF_LEN] __ALIGN_EVENT;
|
|
for (;;) {
|
|
len = read(inotify_fd, buffer, EVENT_BUF_LEN);
|
|
if (len == -1) {
|
|
PLOGE("proc_monitor: read inotify");
|
|
sleep(1);
|
|
continue;
|
|
}
|
|
|
|
for (p = buffer; p < buffer + len; ) {
|
|
event = (struct inotify_event *)p;
|
|
|
|
if (event->mask & IN_OPEN) {
|
|
// Since we're just watching files,
|
|
// extracting file name is not possible from querying event
|
|
pthread_mutex_lock(&list_lock);
|
|
crawl_procfs(process_pid);
|
|
pthread_mutex_unlock(&list_lock);
|
|
} else {
|
|
LOGI("proc_monitor: inotify: /data/app change detected\n");
|
|
pthread_mutex_lock(&list_lock);
|
|
refresh_uid();
|
|
pthread_mutex_unlock(&list_lock);
|
|
update_inotify_mask();
|
|
break;
|
|
}
|
|
|
|
p += EVENT_SIZE + event->len;
|
|
}
|
|
}
|
|
}
|