reTerminal DM 10.1" GT9271 touch mapping wrong after Raspbian image flash

Remudo · June 26, 2026, 8:38am

Hi Seeed / community,

I need help with the GT9271 touch configuration for the reTerminal DM 10.1".

History:
After buying the device I initially had a black screen. I then followed the official flashing guide:

I flashed the image with integrated drivers:
Raspbian-reTerminal-DM-arm64-v1.1.8.tar.xz

After that the display turned on immediately, but the touch input is wrong. The touch controller responds, but the coordinates are not mapped correctly. It is not just a simple X/Y swap or axis inversion. or any guide here reTerminal DM Flash OS | Seeed Studio Wiki
Some areas appear mirrored or shifted, and the touch position does not match the screen position.

Current status:

Touch controller: Goodix GT9271
I2C address: 0x5d
Linux driver binds as: Goodix-TS
Firmware/config file used: /lib/firmware/goodix_9271_cfg.bin

I have been trying to build/tune a goodix_9271_cfg.bin myself. Changing this file clearly changes the raw touch range, so it looks like the GT9271 config is important. But so far I only had limited success.

Could Seeed or someone with a working reTerminal DM 10.1" please confirm:

Is a correct goodix_9271_cfg.bin required for this device?
If yes, where can I get the official factory config?
Can someone share a working goodix_9271_cfg.bin or register dump?
Is the mainline Goodix-TS driver expected to work, or is a Seeed-specific driver required?
Or is my approach wrong and this could indicate a defective touch panel?

Any known-good config or guidance would help a lot.

Thanks.

Remudo · June 26, 2026, 8:37pm

I tested multiple older OS versions, both with and without the Seeed drivers pre-installed. The touch behavior remains unpredictable in all cases.

After a fresh installation with the default config, the mirrored half issue is gone, but the touchpad is still uncontrollable overall. Please see the screenshot attached.

I traced the lines with my finger and tracked the mouse pointer to verify the behavior. The pointer movement is still not stable or consistent across the full screen.

testet on Raspbian-reTerminal-DM-arm64.zip V1.0.9

As of now, I don’t see a practical way to calibrate it for this; it should work right out of the box.

Remudo · June 26, 2026, 9:02pm

Update with new screenshots

The cyan line shows the target path drawn by the test script.
The red line shows the measured touch input
The screenshots from the raw I2C script show the direct sensor input mapped locally in the script.
The screenshots from the desktop script in post above show the touch input after the full desktop/Wayland input chain.

This post is only to show the difference between the raw input layer and the final desktop touch result.