• acts as a KernelSU metamodule
• works on APatch and Magisk too
• CONFIG_OVERLAY_FS=y is required
• CONFIG_TMPFS_XATTR=y is highly encouraged
• tries to mimic an OEM mount, like /mnt/vendor/my_bigball
• for module devs, you can also use this standalone script

1. touch /data/adb/modules/module_id/skip_mount
2. copies contents of /data/adb/modules/module_id to /mnt/vendor/fake_folder_name
3. mirrors SELinux context of every file from /data/adb/modules/module_id to /mnt/vendor/fake_folder_name
4. loops 2 and 3 for all modules
5. overlays /mnt/vendor/fake_folder_name/system/bin to /system/bin and other folders

1. touch /data/adb/modules/module_id/skip_mount
2. create an ext4 sparse image, mount it on /mnt/vendor/fake_folder_name
3. copies contents of /data/adb/modules/module_id to /mnt/vendor/fake_folder_name
4. mirrors SELinux context of every file from /data/adb/modules/module_id to /mnt/vendor/fake_folder_name
5. loops 3 and 4 for all modules
6. unmounts, resizes and remounts sparse image to /mnt/vendor/fake_folder_name
7. overlays /mnt/vendor/fake_folder_name/system/bin to /system/bin and other folders

• Magic mount drastically increases mount count, making detection possible (zimperium)
• OverlayFS mounting with ext4 image upperdir is detectable due to it creating device nodes on /proc/fs, while yes ext4 /data as overlay source is possible, this is rare nowadays.
• F2FS /data as overlay source fails with native casefolding (ovl_dentry_weird), so only sdcardfs users can use /data as overlay source.
• Frankly, I dont see a way to this module mounting situation, this shit is more of a shitty band-aid

• this is added to accomodate something like GPU drivers
• this causes detections but YMMV.
• this is not my problem, this is a fallback, not the main recommendation.
• and yes this is basically how Official KernelSU used to do it.
• if you're on GKI 5.10+, theres an experimental LKM that nukes these nodes.
• if you're on KernelSU 22105+ this is automatically handled.

• user-friendly config editing is available on the WebUI
• otherwise you can modify /data/adb/mountify/config.sh

• by default, mountify mounts all modules with a system folder. mountify_mounts=2
• to mount specific modules only, edit config.sh, mountify_mounts=1 then modify modules.txt to list modules you want mounted

module_id
Adreno_Gpu_Driver
DisplayFeatures
ViPER4Android-RE-Fork
mountify_whiteouts

• FAKE_MOUNT_NAME="mountify" to set a custom fake folder name
• mountify_stop_start=1 to restart android at service (needed for certain modules)

• test_decoy_mount=1 to enable testing for decoy mounts on tmpfs mode

• use_ext4_sparse=1 to force using ext4 mode if your setup is tmpfs_xattr capable
• spoof_sparse=1 to try spoof sparse mount as an android service
• FAKE_APEX_NAME="com.android.mntservice" to customize that android service spoofed name
• sparse_size="2048" to set your sparse size (in MB) to whatever you want
• enable_lkm_nuke=1 to try load an experimental LKM.
• lkm_filename="nuke.ko" to define LKM's filename

• use either NeoZygisk, NoHello, ReZygisk, Zygisk Assistant
• if you use Zygisk Next, then set Denylist Policy to "Enforced" or "Unmount Only"
• then edit config.sh
  • MOUNT_DEVICE_NAME="APatch" if you're on APatch
  • MOUNT_DEVICE_NAME="KSU" if you're on KernelSU forks
  • MOUNT_DEVICE_NAME="magisk" if you're on Magisk
• mountify_custom_umount=0 modify this value to enable known in-kernel umount methods.
  • NOTE: zygisk provider umount is still better, this is here as a second choice.

• this is easy, add skip_mountify to your module's folder.
• mountify checks this on /data/adb/modules/module_name
• [ -f /data/adb/modules/module_name/skip_mountify ]

• fails with De-Bloater, as it uses dummy text, NOT proper whiteouts
• I recommend System App Nuker instead. It uses proper whiteouts.

• module is on The Unlicense
• LKM is on GPLv2
• WebUI is on MIT

• None, none at all. I am handing you a sharp knife, it is not on me if you stab yourself with it.

Download