Troubleshooting
Common issues and their fixes. If your problem isn't listed here, ask in the #modding channel on the Discord.
1. Jibo Not Detected / Can't Enter RCM
Symptoms: lsusb doesn't show 0955:7740,
or ShofEL2 reports it can't find the device.
Checklist
- The ports are behind the panel on the back of Jibo's head — make sure you have that panel open and are plugging into the micro-USB port inside it, not any other connector.
- Use a data cable, not a charge-only cable. Many micro-USB cables sold with devices are charge-only. Test by trying a different cable first.
- Try a different USB port directly on your computer — avoid USB hubs.
- Make sure you're holding the RCM button while pressing reset. The timing matters — hold RCM first, then press reset, then release both after you see the red LED.
-
If you see
0955:9134instead of0955:7740: this is an alternative PID that sometimes appears. It may work, but if ShofEL2 fails, try unplugging, waiting 10 seconds, and re-entering RCM.
Linux permissions
If lsusb shows the device but the tool fails to open it, run with
sudo or install the udev rule:
$ sudo cp 99-jibo-rcm.rules /etc/udev/rules.d/
$ sudo udevadm control --reload-rules
$ sudo udevadm trigger
Windows driver
On Windows, the APX device needs WinUSB as its driver. Use Zadig — select the "NVIDIA APX" or "APX" device and install WinUSB. Then run the installer as Administrator.
2. ShofEL2 Build Fails
Symptoms: make exits with errors about missing headers
or linker errors.
-
Make sure all dependencies are installed — especially
libusb-1.0-0-dev(Debian/Ubuntu) orlibusb(Arch). The package names differ by distro. -
If payload
.binfiles are missing during the build, you also need the ARM cross-compiler:arm-none-eabi-gcc(gcc-arm-none-eabion Debian,arm-none-eabi-gccon Arch). -
Check that you cloned the correct branch:
improvements/IncreasedUSBReadWriteSpeedfromdevsparx/ShofEL2-for-T124. Other branches or forks may not work with Jibo. -
"Error code 1" at the end of
makeis often harmless — check ifshofel2_t124was created anyway withls -la shofel2_t124.
3. Dump Crashes Near the End (98–99%)
The last partition on Jibo's eMMC is mostly empty space and sometimes causes the read to time out. If the dump reaches past 90%, the important data is captured.
- You can proceed with a ~99% dump — the missing sectors are from the end of the empty partition 6.
-
If you only need developer mode access, consider using fast mode
(
--mode-json-onlyin the automod tool) which only reads the/varpartition (~500 MB) and doesn't touch partition 6.
4. mode.json Edit Fails (Automod Tool)
The automod tool tries three methods to edit mode.json inside the
ext filesystem image:
- Linux loop mount (preferred on Linux)
- debugfs (preferred on Windows if installed)
- Raw byte-pattern overwrite (last resort)
If it falls back to raw patching and fails, the most reliable fix is to install
e2fsprogs (which includes debugfs):
Ubuntu/Debian:
$ sudo apt install e2fsprogs
Windows (MSYS2):
$ pacman -S e2fsprogs
5. Hash Mismatch After Write
Symptoms: md5sum returns different hashes for
var_partition.bin and verify_var.bin.
Causes:
- Unstable USB cable or connection
- Jibo dropped out of RCM mode mid-write
- Corrupted read-back (the write may have actually succeeded)
Fix: Re-enter RCM mode and re-run the write command. If you're using the automod tool:
$ python3 jibo_automod.py --mode-json-only --full-var-write
If hashes keep mismatching, try a different USB cable and port.
6. Jibo Doesn't Show the Checkmark on Boot
If Jibo boots normally (eye animation, etc.) instead of showing the checkmark, the mode.json change didn't take effect. Possible causes:
-
Forgot to unmount before flashing — if you edited while mounted
but didn't unmount, the changes weren't flushed to the file. Redo the edit and
unmount explicitly:
sudo umount jibo_var -
Wrong partition flashed — double-check your start sector hex
value. Re-run
fdisk -lon your dump and recalculate. -
Raw patch failed silently — if the automod tool used raw
patching and couldn't find the JSON pattern, it may have written garbage.
Use the
--mode-json-only --full-var-writeapproach with debugfs installed.
7. SSH Works But Updater Fails
Jibo is not reachable:
- Confirm Jibo is showing the checkmark boot and is on your network
- Try
ping <jibo-ip>andnmap -p 22 <jibo-ip> - Make sure you're using the correct IP (it may have changed via DHCP)
"Rootfs is read-only" error:
The updater tries to remount / as writable. If that fails, try
doing it manually first:
root@jibo:~# mount -o remount,rw /
Can't find build/ in archive:
$ python3 jibo_updater.py --ip <jibo-ip> --build-path path/to/build
8. Error Codes Reference
Jibo displays error codes when things go wrong. These can be triggered manually via port 10004. Here's what they mean:
| Error Code(s) | Shows Popup? | Notes |
|---|---|---|
| L2, L7, Q1 | No | Server connection errors (expected without cloud) |
| WIFI4, WIFI4a | Yes | WiFi errors — still show popup |
| Q4 | No | — |
| N1–N12 | No | Network errors |
| OTA11, R1 | No | Triggers "error within error": NOT HANDLED BY ERROR SKILL |
Still Stuck?
If none of the above fixes your issue, post in the #modding channel on the Jibo Revival Discord with:
- Your OS and kernel version
- The exact command you ran and its full output
- Your Jibo's approximate firmware version (check the Settings screen if accessible)
The community is active and someone will help.