A community hub for driver code, patches, and setup notes for fingerprint sensors that are not yet supported in upstream libfprint.
If your sensor shows up in lsusb but fprintd-enroll fails or the device is
unknown to libfprint, there may be a working fix here, or you can contribute one.
Browsable as a website, one page per sensor: https://fprint.jedbillyb.com
Check upstream first: if your reader is on libfprint's own supported devices list, use your distro's libfprint. This repo is only for the ones that are not.
- Find your laptop by model - start here if you do not know your USB ID
- Find your device
- Working fixes
- In progress (unmerged upstream MRs)
- Open upstream MRs with no entry here yet
- Recently merged upstream
- Stale (MR closed without merging)
- Experimental community forks
- Vendor blob and non-LGPL routes
- Claimed upstream, but unreliable
- No known fix yet
- Building: docs/BUILD.md - build, install, PAM, troubleshooting
- Contributing
Know your laptop but not your sensor? Look your machine up in docs/laptops.md, an index of every laptop model named anywhere in this repo or on the libfprint wiki, mapped to its USB ID and the route that works. Adding your own model there is the most useful small contribution you can make.
Otherwise, clone the repo and let the helper script match your hardware against everything catalogued here:
git clone https://github.com/jedbillyb/linux-fingerprint-drivers.git
cd linux-fingerprint-drivers
./tools/detect.sh # or: ./tools/detect.sh 27c6:55b4It scans lsusb for likely readers and tells you, per device, whether there is
an entry here, whether it is on the known-unsupported list, or whether it is
unknown to this repo entirely.
Doing it by hand: run lsusb and find the reader:
$ lsusb
...
Bus 003 Device 002: ID 27c6:55b4 Shenzhen Goodix Technology Co.,Ltd. FingerPrint
...
The 27c6:55b4 part is the USB ID (vendor:product), and it is the key this
repo is organised by: see devices/27c6:55b4/. If nothing
is obviously a reader, compare against the known sensor vendors: Goodix 27c6,
Synaptics/Validity 06cb and 138a, ELAN 04f3, FocalTech 2808, FPC 10a5,
Broadcom 0a5c, EgisTec 1c7a. Many newer sensors are on SPI or I2C and never
appear in lsusb at all.
| Device ID | Chip | Status | Distros tested |
|---|---|---|---|
| 27c6:55b4 | Goodix GF3268 | Working | Arch, Debian, Ubuntu, Fedora |
| 27c6:5385 (5335/5395) | Goodix HTK32 | Working | Arch, Fedora, Ubuntu/Debian |
| 04f3:0c4c (0c00) | ELAN match-on-chip | Working | Debian/Ubuntu |
| 138a:0090 (0097) | Validity/Synaptics VFS0090 | Working | Ubuntu, Arch, Fedora, NixOS |
| 06cb:009a (138a:009d) | Synaptics Prometheus | Working | Ubuntu, Arch, Fedora |
| 06cb:008a | Synaptics Viper | Working | Ubuntu, Debian |
| 04f3:0c6c | ELAN Match-on-Chip 2 | Working | Ubuntu/Debian, Fedora, Arch |
| 10a5:9800 | FPC fpcmoh (match-on-host) | Working | Fedora, Arch, Debian, Ubuntu |
| 2541:0236 (9711) | Chipsailing CS9711 | Working | Arch (GPD Win Max 2, AYANEO 2) |
| 06cb:00ff (+00c9/00d1/00e7/0124/0169) | Synaptics Tudor MiS | Working | Arch, Fedora, Ubuntu (community routes) |
| 06cb:00fd | Synaptics Tudor MiS (IdeaPad pkg, v139) | Working | Ubuntu 26.04 |
| 06cb:00eb | Synaptics Tudor MiS (Dynabook, synatlsmoc) | Working | Kali rolling |
| 1b55:0124 | ZKTeco ZK9500 (optical) | Working | Arch / CachyOS |
| 347d:0302 (0304) | Blestech Betterlife btlfp | Working | Arch / CachyOS |
| 2009:7638 | iStorage ADST / BYD BF5325 | Working | Arch / CachyOS |
| 2808:9e65 | FocalTech | Partial | Undocumented (unconfirmed) |
| gxfp5187 (SPI) | Goodix GXFP5187 | Working | Ubuntu 24.04 (entry) |
Status legend: Working (enroll + verify reliable), Partial (works with caveats), WIP (in progress). Statuses only move when someone reports back, so please open a device report with your distro and hardware either way.
Every entry needs libfprint built from source. The build, install, PAM, package pinning, troubleshooting and revert steps are the same for all of them and live in docs/BUILD.md; each device entry covers only what is specific to that sensor.
Note: 06cb:008a (Synaptics Viper, see entry) and 06cb:009a
(Synaptics Prometheus, see entry) are handled by
python-validity, a userspace driver daemon that plugs into fprintd via
open-fprintd, rather than a libfprint driver.
Some very common sensors have no open driver at all: the only thing that works is a proprietary vendor driver loaded through libfprint's TOD (Touch OEM Driver) mechanism, or a project under a license this hub cannot host.
This repo hosts only LGPL-2.1 code. But leaving these devices uncatalogued just means their owners find nothing and conclude Linux cannot use their reader, so they get entries that point at the vendor route and state the trade-offs plainly. Nothing proprietary is mirrored here, and a pointer is not an endorsement: read the caveats in each entry before putting a closed binary in your authentication path.
| Device ID | Chip | Route | Entry |
|---|---|---|---|
| 27c6:550a | Goodix | Lenovo TOD blob (libfprint-2-tod1-goodix); ThinkPad E14/E15, ThinkBook |
entry |
| 27c6:533c | Goodix | Dell OEM TOD blob; XPS 13 9300, XPS 15 9500 | entry |
| 04f3:0c4b | ELAN | Lenovo TOD blob (libfprint-2-tod1-elan); ThinkPad E14 Gen 4 |
entry |
| 0a5c:58xx / 586x | Broadcom | Dell/Canonical TOD blob, alongside the open MR | entry |
| 10a5:9201 | FPC | fingerprint-ocv, standalone AGPL-3.0 daemon |
entry |
| 2808:9348 | FocalTech FT9348 | standalone GPL-2.0 capture tool + FocalTech's proprietary matcher .so, PAM; DIGMA Pro Magnus M |
entry |
| 06cb:00ff family, 06cb:00fd | Synaptics Tudor | relinked vendor Windows driver, or a fully open native driver | 00ff entry, 00fd entry |
Still genuinely out of scope: sensors where nobody has published any working route at all (those live in the gap-map), and routes we cannot verify (see Contributing on the vetting bar).
Drivers that have been submitted to the official libfprint project but are
not merged yet, so they are not in any released libfprint. Each has an entry
under devices/ with the merge-request link and build notes. Status
changes to Working here once someone confirms it, and the row moves once the MR
lands upstream.
MR states were last verified against GitLab on 2026-11-17; the merge request itself is always the authority.
| Device ID | Chip / sensor | MR | Notes |
|---|---|---|---|
| 147e:1002 | UPEK Touchstrip (upeksonly) | !585 | Draft |
| 0a5c:xxxx (Dell) | Broadcom ControlVault3 (fp + NFC) | !620 | Dell family; needs vendor firmware; see entry |
| 138a:0097 / 009d | Validity/Synaptics VFS0097 (VCSFW) | !626, !579, !619 | !626 is the active continuation of !579 (now closed); also covers 0090 + 06cb:009a |
| 138a:00ab | Validity VCSFW 0x969/0xd51 | !626 | hardware-validated; HP ZBook Studio x360 G5 |
| 06cb:00cb | Validity VCSFW 0x969 | !626 | hardware-validated; HP Pavilion x360 14-dh |
| 06cb:00b7 | Validity VCSFW 0xd51 | !626 | registered, untested; HP G6 series |
| 2808:c652 | FocalTech FT9362 (match-on-host) | !588 | |
| 2808:9338 / 93a9 | FocalTech FT9201 | !572 | |
| 10a5:9200 | FPC1022 (FPC Disum) | !570 | |
| 10a5:9924 | FPC match-on-host (Honor) | !611, !666 | !666 picks up !611 and fixes the USB reset that broke enrollment |
| 04e8:7305 | Samsung 7305 | !586 | |
| 04e8:730b | Samsung 730B (image) | !556 | |
| 1c7a:0576 | EgisTec 0576 | !571 | |
| 1c7a:0575 | EgisTec EGIS0575 (swipe) | !695, !357 | !695 (2026-11) rebases the stalled !357 and fixes its reported bugs; author uses it daily on an Acer Swift 3 SF314-43 |
| 298d:2033 | NextBiometrics NB-2033-U | !574 | |
| crfpmoc (cros_ec) | Chromium OS EC FPMCU | !512 | not USB |
| mafp8800 (SPI) | Microarray MAFP8800 | !580 | not USB (SPI) |
Sensors that got a driver or a new product ID posted upstream since the last refresh, and that have no entry in this repo yet. None is merged, so build the MR branch to try one (see docs/BUILD.md). If one works for you, an entry with your laptop model is a welcome contribution. MRs for sensors on the gap-map are listed in its Lead column instead.
| Device ID | Chip / sensor | MR | Notes |
|---|---|---|---|
| 04f3:0ca2 | ELAN match-on-chip (LG laptops) | !693 | adds the PID to elanmoc |
| 04f3:0c3d | ELAN press sensor | !654 | new elanpress driver |
| 06cb:0082 | Synaptics | !664, !665 | new driver; !665 (draft) drops its OpenCV dependency |
| 1162:2370 | SecuGen U20-AL (external) | !633 | new driver |
| 1c7a:05b1 | Egis ETU906Axx-E | !671 | draft, egismoc |
| 2541:fa10 | Realtek (GPD Win 5) | !677 | adds the PID to realtek |
| 27c6:6005 | Goodix match-on-chip | !678 | adds the PID to goodixmoc |
| 27c6:6f94 | Goodix match-on-chip | !667 | adds the PID to goodixmoc |
| 2808:a573 | FocalTech FT9365 ESS | !657 | adds the PID to focaltech_moc |
| 28e9:5151 | Petaic match-on-chip | !692 | new petaicmoc driver |
| 2c0a:8132 | Generic Realtek USB2.0 bridge | !661 | draft, realtek |
| 3274:9011 | Realtek-based sensor | !694 | adds the PID to realtek |
| 33a7:2388 | Jiangxi OMS match-on-chip module | !659 | draft, new driver |
| 047d:8228 | Kensington VeriMark Desktop 2.0 (external) | !663 | realtek |
These were WIP here and their merge requests have since merged into libfprint (2026-06/07). They shipped in libfprint 1.94.100 (2026-07-26), so a distro with that release or newer supports them out of the box. Entries are kept for people still on an older libfprint.
| Device ID | Chip / sensor | MR | Merged |
|---|---|---|---|
| 2808:6553 | FocalTech FT9365 ESS (not the capture function on dual-function FT9201 modules) | !554 | 2026-06-18 |
| 04f3:0c9c | ELAN ARM-M4 (0c9c) | !568 | 2026-06-18 |
| 3274:8012 | Microarray match-on-chip | !492 | 2026-06-18 |
| 298d:2020 | NextBiometrics NB-2020-U | !569 | 2026-07-02 |
Merged after 1.94.100, so only in libfprint git master until the next release. None of these ever had an entry here; they are product IDs added to existing drivers. If your reader is one of them, build libfprint from git.
| Device ID | Driver | MR | Merged |
|---|---|---|---|
| 04f3:0cb6 | elanmoc |
!635 | 2026-07 |
| 27c6:65ba, 6515, 66ac, 660a | goodixmoc |
!645 | 2026-08 |
| 27c6:6082 | goodixmoc |
!652 | 2026-09 |
| 06cb:010b | synaptics |
!653 | 2026-09 |
| 04f3:0cb5 | elanmoc |
!696 | 2026-11 |
Entries kept as leads for anyone wanting to pick the reverse-engineering back up.
| Device ID | Chip / sensor | MR | State |
|---|---|---|---|
| 04f3:0c8e | ELAN elanmoc2 | !560 | Closed 2026-06-22; no maintained upstream work |
| 04f3:310d | ELAN ELAN7006 | !383 | Closed; the draft never enrolled or verified |
Drivers from community forks rather than upstream merge requests. These are not in libfprint and may be unstable; see each device entry for details.
| Device ID | Chip / sensor | Source | Notes |
|---|---|---|---|
| 27c6:521d (538d) | Goodix 521d/538d | goodix-fp-linux-dev, lbssousa/libfprint | Works via AUR libfprint-goodix-521d; 27c6:538d verified on a Dell Inspiron 5405 with the lbssousa SIGFM fork plus a one-time goodix-fp-dump init (see entry) |
| 27c6:5110 + family | Goodix (newer) | goodix-fp-linux-dev/libfprint | Experimental, not for daily use |
| 27c6:5201 | Goodix GF5288 / HTK32 (ASUS ZenBook S UX391) | J0UH/goodix-5201-linux | Enroll, verify and sudo work on one UX391UA. Patched libfprint with SIGFM (!530) |
| 27c6:550c | Goodix GF5288 / GM168SEC (Lenovo Yoga 9 14IAP7) | goodix-550c-linux-driver, Debian packages | Enroll, verify, sudo and lock screen on stock firmware 13021; needs the PSK from the original Windows install. Reflash alternative: brianporeilly/goodix-550c-driver |
| 2df0:0003 | CanvasBio CB2000 (Samsung Galaxy Book2/Book3 360) | four independent drivers | Two report daily use. 2df0:0007 is a different part and is not supported by any of them; an experimental patch on the upstream Realtek driver logs it in, with a security caveat. Upstream MRs now open: !687, !670 (:0003) and !675 (:0007, SDCP) |
| 04f3:0c80 | ELAN ARM-M4 (Surface Laptop Go 2) | patch on Depau's libfprint | Enroll + verify work; clearing sensor storage still times out, and re-enroll wipes all prints |
libfprint listing your sensor as supported does not always mean it works. These have in-tree drivers that fail in practice, with open MRs that fix them. If your reader is detected but you cannot finish an enrollment, start here.
| Device ID | Chip | Problem | Fix MR | Entry |
|---|---|---|---|---|
| 04f3:0c28 (0c0x-0c4x family) | ELAN image sensors | enroll fails every second capture; thresholds and swipe-mode assumptions wrong for the hardware | !217, !530 | entry |
| 1c7a:0570 (0571) | EgisTec egis0570 | calibration not implemented, erratic captures | !548 | entry |
| 147e:1002 and upektc/upeksonly | UPEK swipe sensors | 144x384 captures yield too few minutiae to match reliably | !576 | entry |
| 1c7a:05ae, 1c7a:9201 | EgisTec etu905 | no firmware template update after verify | !610 | - |
115 more USB sensors are on the libfprint wiki's unsupported list with no working driver here or upstream. They are catalogued, with any known reverse-engineering leads, in docs/unsupported-devices.md - the contributor gap-map. If you own one of those, a protocol dump or a driver is very welcome.
SPI/I2C sensors (ELAN7001/7002/079C, GXFP5187/51B7, GDIX51C0, fpc1020, ...) are
also largely unsupported and never appear in lsusb; see the wiki's SPI Devices
section.
New device entries, status reports, and corrections are all welcome. See
CONTRIBUTING.md for the folder layout, required files,
attribution, and license rules, and run python3 tools/check-repo.py before
opening a PR.
Two things are worth as much as a driver: a report that an entry works (or stopped working) on your distro, and a protocol dump for anything in the gap-map.
Hosting vs cataloguing. Code and patches in this repo are LGPL-2.1 only. Cataloguing is broader: a device entry may point at a proprietary TOD blob or a differently-licensed project when that is the only thing that works, as long as the entry says so plainly and mirrors none of it.
Vetting bar. There are many small per-device fingerprint repos on GitHub, and
a fair number are abandoned, unlicensed, or generated. An entry gets added when
the route can be checked: real driver code, a license, and some evidence it
worked on hardware. Unverifiable leads belong in an issue, not in devices/.
Where a fix belongs upstream, please also push it there: this hub exists to make unmerged work findable, not to replace libfprint. Entries whose merge requests land upstream get moved to the merged table and eventually dropped.
All driver code and patches hosted here are derived from libfprint and are licensed under LGPL-2.1. See LICENSE. Contributed code must be LGPL-2.1-compatible.
Device entries may link to routes under other licenses, including proprietary vendor drivers, where that is the only thing that works for a sensor. Those entries say so explicitly and mirror no code. See Vendor blob and non-LGPL routes.