remove: unnecessary unshare hook

This commit removes the "unshare" hook. It used to provide the right timing for manual umount in the past, however with recent umount system changes, basing now on mount namespaces, those are handled way earlier, and this "unshare" hook makes no difference, being considered useless.

Author: ThePedroo

loader/src/injector/hook.cpp

@@ -177,40 +177,6 @@ bool update_mnt_ns(enum mount_namespace_state mns_state, bool dry_run) {
     return true;
 }
 
-// Unmount stuffs in the process's private mount namespace
-DCL_HOOK_FUNC(int, unshare, int flags) {
-    int res = old_unshare(flags);
-    if (g_ctx && (flags & CLONE_NEWNS) != 0 && res == 0 &&
-        // For some unknown reason, unmounting app_process in SysUI can break.
-        // This is reproducible on the official AVD running API 26 and 27.
-        // Simply avoid doing any unmounts for SysUI to avoid potential issues.
-        !g_ctx->flags[SERVER_FORK_AND_SPECIALIZE] && !(g_ctx->info_flags & PROCESS_IS_FIRST_STARTED)) {
-
-        /* INFO: There might be cases, specifically in Magisk, where the app is in
-                   DenyList but also has root privileges. For those, it is up to the
-                   user remove it, and the weird behavior is expected, as the weird
-                   user behavior. */
-
-        /* INFO: For cases like Magisk, where you can only give an app SU if it was
-                   either requested before or if it's not in DenyList, we cannot
-                   umount it, or else it will not be (easily) possible to give new
-                   apps SU. Apps that are not marked in APatch/KernelSU to be umounted
-                   are also expected to have AP/KSU mounts there, so we will follow the
-                   same idea by not umounting any mount. */
-
-        if (g_ctx->info_flags & (PROCESS_IS_MANAGER | PROCESS_GRANTED_ROOT) || !(g_ctx->flags[DO_REVERT_UNMOUNT])) {
-            update_mnt_ns(Mounted, false);
-        }
-
-        old_unshare(CLONE_NEWNS);
-    }
-
-    /* INFO: To spoof the errno value */
-    errno = 0;
-
-    return res;
-}
-
 // We cannot directly call `dlclose` to unload ourselves, otherwise when `dlclose` returns,
 // it will return to our code which has been unmapped, causing segmentation fault.
 // Instead, we hook `pthread_attr_setstacksize` which will be called when VM daemon threads start.
@@ -1035,7 +1001,6 @@ void hook_functions() {
     }
 
     PLT_HOOK_REGISTER(android_runtime_dev, android_runtime_inode, fork);
-    PLT_HOOK_REGISTER(android_runtime_dev, android_runtime_inode, unshare);
     PLT_HOOK_REGISTER(android_runtime_dev, android_runtime_inode, strdup);
     PLT_HOOK_REGISTER(android_runtime_dev, android_runtime_inode, property_get);
     hook_commit();