You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Can You add Support to 138a:00ab Validity Sensors, Inc. #238
Builds for Ubuntu 22.04, 24.04, 25.10, and 26.04. After install, enroll via Settings → Users → Fingerprint Login, or sudo fprintd-enroll -f right-index-finger $USER.
Hi small issue, on every reboot and poweroff the thing get vanished, I've to manually enter
this From Gemini
# 1. Stop the services to release the busy USB lock
sudo systemctl stop python3-validity open-fprintd
# 2. Tell the kernel to reset the specific USB port your sensor is plugged into
sudo udevadm trigger --attr-match=idVendor=138a --attr-match=idProduct=00ab
# 3. Start the services back up so they can freshly initialize the chip
sudo systemctl start python3-validity open-fprintd
then only the OS detect the fp and on restart or power off it'll get vanish need to run this again.
Hey @Killersparrow1 — really glad it works on Fedora 44, thanks for confirming it crosses distros.
The reboot issue is fixed in commit cd0def0 on the branch. The fix is a defensive USB reset at the start of Usb.open_dev() — same effect as your manual udevadm trigger workaround, but built into the driver so it runs automatically every time the daemon opens the chip. No more "vanishes on reboot."
To update on Fedora, rebuild from the latest branch tip:
If you used git clone + setup.py install, do git pull in the working tree then re-run install.
If you used pip install git+https://github.com/SimpleX-T/python-validity.git@feat/sensor-type-0xd51, just rerun that command.
Then sudo systemctl restart python3-validity.service and it should survive reboots cleanly. Let me know if anything else comes up!
Heads-up for other 138a:00ab owners: that USB ID covers at least two different sensor silicon. @SimpleX-T's branch handles the HP EliteBook 840 G5 variant, which reports sensor type 0xd51. The HP ZBook Studio x360 G5 uses the same 138a:00ab PID but a different chip reporting sensor type 0x969 (identifies as "57K0 ..."), so the 0xd51 aliasing doesn't pick it up.
It's a 57K0-family part, so it aliases cleanly to the same 0x199 profile. Three-line mirror of the existing 0xd51 handling:
add 0x969 to line_update_type1_devices
alias 0x969 -> 0x199 in Sensor.open()
accept the b[0]==3 finger interrupt when real_device_type is 0x969
Verified end-to-end on an HP ZBook Studio x360 G5: init + calibration + enroll (5 stages, enroll-completed) + verify (verify-match), and Mint's fingwit GUI drives it via open-fprintd. PR against @SimpleX-T's branch: SimpleX-T#2
Quick update for this thread since a lot converged here:
@ggiesen — thanks for filing SimpleX-T/python-validity#2. Merged into the PR branch (0c00f0a); the 0x969 handling for HP ZBook Studio x360 G5 is now upstream in PR #256 along with your test evidence.
@Killersparrow1 — your "vanishes on reboot" issue has been fixed on the branch since cd0def0 (defensive USB reset in open_dev()), which is in the current PPA build. On any of your machines that still need the manual udevadm trigger workaround, updating from the PPA should resolve it.
PR #256 just got four more commits covering post-resume recovery for 0x969, friendlier errors when the reset_blob is rejected on this chip family (0xd51 / 0x969 don't accept the reset_blob we ship — no fix yet, but a real error message + tracking pointer), and Karloss1234's flash-management diagnosis. Details on PR #256.
Next: rolling all of this into PPA build 0.16~hp8 for jammy/noble/questing/resolute.
i got the validity sensor to work on my EliteBook 840 G5 with the help of claude and the original pull from git. i did ask the ai to create an .sh file to try to speed up the process for anyone with this specfic issue and the 138a:00ab sensor. but im fairly new to git and i dont know exactly what im doing. all i know is that i spent alot time coming back to claude and seeing if getting the sensor to work on my fedora workstation 44 gnome setup would work. so far it works perfectly and i am now able to unlock the laptop with multiple fingerprints and i did have to troubleshoot briefly to get the fingerprint to only be used during unlocking the laptop and not for any CLI function
Activity
SimpleX-T commentedon May 21, 2026
Hey @Killersparrow1 —
138a:00abon the EliteBook 840 G5 is now supported:Builds for Ubuntu 22.04, 24.04, 25.10, and 26.04. After install, enroll via Settings → Users → Fingerprint Login, or
sudo fprintd-enroll -f right-index-finger $USER.Killersparrow1 commentedon May 23, 2026
Super Great work sir tested on Fedora 44 working.
SimpleX-T commentedon May 23, 2026
I'm glad it worked for you.
Killersparrow1 commentedon May 24, 2026
Hi small issue, on every reboot and poweroff the thing get vanished, I've to manually enter
this From Gemini
then only the OS detect the fp and on restart or power off it'll get vanish need to run this again.
any fixes sir.
usb.py: defensive USB reset at open_dev() entry
SimpleX-T commentedon May 24, 2026
Hey @Killersparrow1 — really glad it works on Fedora 44, thanks for confirming it crosses distros.
The reboot issue is fixed in commit cd0def0 on the branch. The fix is a defensive USB reset at the start of
Usb.open_dev()— same effect as your manualudevadm triggerworkaround, but built into the driver so it runs automatically every time the daemon opens the chip. No more "vanishes on reboot."To update on Fedora, rebuild from the latest branch tip:
git clone + setup.py install, dogit pullin the working tree then re-run install.pip install git+https://github.com/SimpleX-T/python-validity.git@feat/sensor-type-0xd51, just rerun that command.Then
sudo systemctl restart python3-validity.serviceand it should survive reboots cleanly. Let me know if anything else comes up!ggiesen commentedon Jul 6, 2026
Heads-up for other
138a:00abowners: that USB ID covers at least two different sensor silicon. @SimpleX-T's branch handles the HP EliteBook 840 G5 variant, which reports sensor type0xd51. The HP ZBook Studio x360 G5 uses the same138a:00abPID but a different chip reporting sensor type0x969(identifies as "57K0 ..."), so the0xd51aliasing doesn't pick it up.It's a 57K0-family part, so it aliases cleanly to the same
0x199profile. Three-line mirror of the existing0xd51handling:0x969toline_update_type1_devices0x969->0x199inSensor.open()b[0]==3finger interrupt whenreal_device_typeis0x969Verified end-to-end on an HP ZBook Studio x360 G5: init + calibration + enroll (5 stages,
enroll-completed) + verify (verify-match), and Mint'sfingwitGUI drives it via open-fprintd. PR against @SimpleX-T's branch: SimpleX-T#2SimpleX-T commentedon Jul 7, 2026
Quick update for this thread since a lot converged here:
@ggiesen — thanks for filing SimpleX-T/python-validity#2. Merged into the PR branch (
0c00f0a); the0x969handling for HP ZBook Studio x360 G5 is now upstream in PR #256 along with your test evidence.@Killersparrow1 — your "vanishes on reboot" issue has been fixed on the branch since
cd0def0(defensive USB reset inopen_dev()), which is in the current PPA build. On any of your machines that still need the manualudevadm triggerworkaround, updating from the PPA should resolve it.PR #256 just got four more commits covering post-resume recovery for 0x969, friendlier errors when the reset_blob is rejected on this chip family (0xd51 / 0x969 don't accept the reset_blob we ship — no fix yet, but a real error message + tracking pointer), and Karloss1234's flash-management diagnosis. Details on PR #256.
Next: rolling all of this into PPA build
0.16~hp8for jammy/noble/questing/resolute.crawlmonkey27 commentedon Sep 26, 2026
setup-validity-00ab.sh
i got the validity sensor to work on my EliteBook 840 G5 with the help of claude and the original pull from git. i did ask the ai to create an .sh file to try to speed up the process for anyone with this specfic issue and the 138a:00ab sensor. but im fairly new to git and i dont know exactly what im doing. all i know is that i spent alot time coming back to claude and seeing if getting the sensor to work on my fedora workstation 44 gnome setup would work. so far it works perfectly and i am now able to unlock the laptop with multiple fingerprints and i did have to troubleshoot briefly to get the fingerprint to only be used during unlocking the laptop and not for any CLI function
saatvik333 commentedon Oct 13, 2026
https://github.com/saatvik333/fp-840G5