fprintd-enroll goes directly to enroll failed

[kp4djt]

Just installed the code on my Thinkpad X1 Carbon Gen 4. Running Mate Ubuntu 22.04. Here is what I get when I invoke fprintd-enrol kp4djt@kp4djt-carbon:~$ fprintd-enroll Using device /net/reactivated/Fprint/Device/0 Enrolling right-index-finger finger. Enroll result: enroll-failed kp4djt@kp4djt-carbon:~$

It goes immediately to the Enroll result not allowing me to swipe finger/ Results of sudo systemctl status python3-validity ● python3-validity.service - python-validity driver dbus service Loaded: loaded (/lib/systemd/system/python3-validity.service; enabled; vendor preset: enabled) Active: active (running) since Sun 2023-01-08 01:05:40 CST; 14min ago Main PID: 832 (python3) Tasks: 3 (limit: 9082) Memory: 25.1M CPU: 342ms CGroup: /system.slice/python3-validity.service └─832 python3 /usr/lib/python-validity/dbus-service --debug

Jan 08 01:07:03 kp4djt-carbon python3[832]: DEBUG:root:<tls< 17: 0104 Jan 08 01:07:03 kp4djt-carbon ERROR:root:Failed[832]: 0401 Traceback (most recent call last): File "/usr/lib/python-validity/dbus-service", line 157, in run sensor.enroll(usr, index, update_cb) File "/usr/lib/python3/dist-packages/validitysensor/sensor.py", line 828, in enroll self.create_enrollment() File "/usr/lib/python3/dist-packages/validitysensor/sensor.py", line 749, in create_enrollment assert_status(tls.app(pack('<BL', 0x69, 1))) File "/usr/lib/python3/dist-packages/validitysensor/util.py", line 12, in assert_status raise Exception('Failed: %04x' % s) Exception: Failed: 0401 Jan 08 01:07:03 kp4djt-carbon python3[832]: DEBUG:root:EnrollStatus

I have other code running Python3 so not sure what to think.

[Hansxsourse]

Same issue with same platform

[zgshen]

Same issue on T460p running Ubuntu 22.04.

[Tymoniasty]

The same error T460p with Ubuntu 22.04.02 LTS.

In the GUI I can see now the fingerprint section (it was not available before using this package) but the fprintd-enroll returns error.

The same if I try to add fingerprint in GUI: It does not even allow me to scan - just once I select scan and pick which finger it displays the above.

[StewartPolsky]

Same issue T460s Mint 21.2

[uunicorn]

Exception: Failed: 0401

This looks suspiciously similar to #71 . Can you check which device you have with lsusb?

[StewartPolsky]

@uunicorn thank you for the quick reply. Verified my device with lsusb I have 138a:0090. Looks like it is the same issue as #71 in that I can't enroll as it is match-on-host.

Just trying to figure out how to enroll my fingerprint now

[uunicorn]

Yeah, as explained in #71 , match-on-host devices are not supported by python-validity.

There is a branch of the official libfprint project which adds partial 138a:0090 support. "Partial" here means that this branch rely on your device to be paired by an external tool. This external tool itself is based on python-validity code to do the pairing. This is why python-validity sort of supports 138a:0090 but not really.

To get it working you can try removing python-validity, open-fprintd and follow instructions from the link above.

[StewartPolsky]

I had initially started with the branch you linked. Unfortunately the repo seems to be dead. The PPA has not been updated to support Ubuntu 22.04.

When I try to compile myself I receive some errors. There are issues opened for both of these but there doesn’t seem to have been an update in years.

[uunicorn]

Yeah, neither of these projects receive much love unfortunately. Ages ago I created a branch of python-validity to add a proper support for match-on-host devices (138a:0090 specifically). I wanted it to be compatible with Windows, so that fingers enrolled on Windows will work on Linux and vice versa. I got to a point where I reverse engineered the finger template format stored on the device. I also found a library which Synaptic / Validity software is using (based on?) to enroll / match finger templates. It appears to be a redistributable part of DigitalPersona SDK. I.e. if you buy DigitalPersona SDK, you can redistribute those libraries as part of your product. It appears to be present in a bunch of Android phones and tablets. There are also x86 Linux binaries available. But I don't think we can simply steal this library and put it into an open source project. Re-implementing the whole math behind template matching / generation is way too much work. Besides, it will probably still violate a bunch of IP rights and patents. So, I just gave up.