Fix root loss issue when MagiskHide is enabled

In previous implementations, proc_monitor checks whether the mount namespace of an app is actually separated from zygote using a list generated at startup.
However, for some unknown reason, some devices (e.g. Samsung) has multiple zygote servers running in the background.
This means that app processes spawned from the unlisted zygotes are not checked whether the separation is done or not, causing MagiskHide unmount stuffs in the namespace of zygote, and since zygote is the "mother" of all apps, all apps will no longer have root access.

Since I'm not sure of the reason why multiple zygotes exists, so instead of checking the namespace against a list, compare the current namespace against the parent process's namespace.
This will make sure the namespace is NOT the same as the parent process, which is supposed to be the zygote server.

native/jni/magiskhide/proc_monitor.c

@@ -19,8 +19,7 @@
 #include "utils.h"
 #include "magiskhide.h"
 
-static char init_ns[32], zygote_ns[2][32], cache_block[256];
-static int zygote_num, has_cache = 1, pipefd[2] = { -1, -1 };
+static int pipefd[2] = { -1, -1 };
 
 // Workaround for the lack of pthread_cancel
 static void term_thread(int sig) {
@@ -40,27 +39,30 @@ static void term_thread(int sig) {
 
 static int read_namespace(const int pid, char* target, const size_t size) {
 	char path[32];
-	snprintf(path, sizeof(path), "/proc/%d/ns/mnt", pid);
+	sprintf(path, "/proc/%d/ns/mnt", pid);
 	if (access(path, R_OK) == -1)
 		return 1;
 	xreadlink(path, target, size);
 	return 0;
 }
 
-static void store_zygote_ns(int pid) {
-	if (zygote_num == 2) return;
-	do {
-		usleep(500);
-		read_namespace(pid, zygote_ns[zygote_num], 32);
-	} while (strcmp(zygote_ns[zygote_num], init_ns) == 0);
-	++zygote_num;
-}
-
 static void lazy_unmount(const char* mountpoint) {
 	if (umount2(mountpoint, MNT_DETACH) != -1)
 		LOGD("hide_daemon: Unmounted (%s)\n", mountpoint);
 }
 
+static int parse_ppid(int pid) {
+	char stat[512], path[32];
+	int fd, ppid;
+	sprintf(path, "/proc/%d/stat", pid);
+	fd = xopen(path, O_RDONLY);
+	xread(fd, stat, sizeof(stat));
+	close(fd);
+	/* PID COMM STATE PPID ..... */
+	sscanf(stat, "%*d %*s %*c %d", &ppid);
+	return ppid;
+}
+
 static void hide_daemon(int pid) {
 	LOGD("hide_daemon: start unmount for pid=[%d]\n", pid);
 
@@ -77,18 +79,6 @@ static void hide_daemon(int pid) {
 	vec_init(&mount_list);
 	file_to_vector(buffer, &mount_list);
 
-	// Find the cache block name if not found yet
-	if (has_cache && cache_block[0] == '\0') {
-		vec_for_each(&mount_list, line) {
-			if (strstr(line, " /cache ")) {
-				sscanf(line, "%256s", cache_block);
-				break;
-			}
-		}
-		if (strlen(cache_block) == 0)
-			has_cache = 0;
-	}
-
 	// Unmount dummy skeletons and /sbin links
 	vec_for_each(&mount_list, line) {
 		if (strstr(line, "tmpfs /system/") || strstr(line, "tmpfs /vendor/") || strstr(line, "tmpfs /sbin")) {
@@ -133,32 +123,10 @@ void proc_monitor() {
 	act.sa_handler = term_thread;
 	sigaction(TERM_THREAD, &act, NULL);
 
-	cache_block[0] = '\0';
-
-	// Get the mount namespace of init
-	if (read_namespace(1, init_ns, 32)) {
+	if (access("/proc/1/ns/mnt", F_OK) != 0) {
 		LOGE("proc_monitor: Your kernel doesn't support mount namespace :(\n");
 		term_thread(TERM_THREAD);
 	}
-	LOGI("proc_monitor: init ns=%s\n", init_ns);
-
-	// Get the mount namespace of zygote
-	zygote_num = 0;
-	while(!zygote_num) {
-		// Check zygote every 10 ms
-		usleep(10000);
-		ps_filter_proc_name("zygote", store_zygote_ns);
-	}
-	ps_filter_proc_name("zygote64", store_zygote_ns);
-
-	switch(zygote_num) {
-	case 1:
-		LOGI("proc_monitor: zygote ns=%s\n", zygote_ns[0]);
-		break;
-	case 2:
-		LOGI("proc_monitor: zygote ns=%s zygote64 ns=%s\n", zygote_ns[0], zygote_ns[1]);
-		break;
-	}
 
 	// Register our listener to logcat monitor
 	xpipe2(pipefd, O_CLOEXEC);
@@ -171,8 +139,8 @@ void proc_monitor() {
 			continue;
 		}
 		char *ss = strchr(log, '[');
-		int pid, ret, comma = 0;
-		char *pos = ss, processName[256], ns[32];
+		int pid, ppid, ret, comma = 0;
+		char *pos = ss, proc[256], ns[32], pns[32];
 
 		while(1) {
 			pos = strchr(pos, ',');
@@ -183,42 +151,45 @@ void proc_monitor() {
 		}
 
 		if (comma == 6)
-			ret = sscanf(ss, "[%*d %d %*d %*d %256s", &pid, processName);
+			ret = sscanf(ss, "[%*d %d %*d %*d %256s", &pid, proc);
 		else
-			ret = sscanf(ss, "[%*d %d %*d %256s", &pid, processName);
+			ret = sscanf(ss, "[%*d %d %*d %256s", &pid, proc);
 
 		if(ret != 2)
 			continue;
 
+		ppid = parse_ppid(pid);
+
 		// Allow hiding sub-services of applications
-		char *colon = strchr(processName, ':');
+		char *colon = strchr(proc, ':');
 		if (colon)
 			*colon = '\0';
 
 		// Critical region
 		pthread_mutex_lock(&hide_lock);
 		vec_for_each(hide_list, line) {
-			if (strcmp(processName, line) == 0) {
-				while(1) {
-					ret = 1;
-					for (int i = 0; i < zygote_num; ++i) {
-						read_namespace(pid, ns, sizeof(ns));
-						if (strcmp(ns, zygote_ns[i]) == 0) {
-							usleep(50);
-							ret = 0;
-							break;
-						}
-					}
-					if (ret) break;
-				}
+			if (strcmp(proc, line) == 0) {
+				read_namespace(ppid, pns, sizeof(pns));
+				do {
+					read_namespace(pid, ns, sizeof(ns));
+					if (strcmp(ns, pns) == 0)
+						usleep(50);
+					else
+						break;
+				} while (1);
 
 				// Send pause signal ASAP
-				if (kill(pid, SIGSTOP) == -1) continue;
+				if (kill(pid, SIGSTOP) == -1)
+					continue;
 
 				// Restore the colon so we can log the actual process name
 				if (colon)
 					*colon = ':';
-				LOGI("proc_monitor: %s (PID=%d ns=%s)\n", processName, pid, ns);
+#ifdef MAGISK_DEBUG
+				LOGI("proc_monitor: %s (PID=[%d] ns=%s)(PPID=[%d] ns=%s)\n", proc, pid, ns + 4, ppid, pns + 4);
+#else
+				LOGI("proc_monitor: %s\n", proc);
+#endif
 
 				/*
 				 * The setns system call do not support multithread processes