Public bug reported:
Summary: ELAN0676 (04F3:3195) touchpad on ThinkPad T14 Gen 6 (Intel/Lunar Lake)
enumerates correctly but produces zero input events - GPIO IRQ never re-fires
after initial boot handshake.
Hardware: ThinkPad T14 Gen 6, Intel Core Ultra 200V (Lunar Lake)
Kernels tested (both affected, confirmed dead on both):
- 6.17.0-1033-oem (Ubuntu 24.04 Pro, as-shipped)
- 7.0.0-42-generic (Ubuntu mainline-kernel-ppa build)
Confirmed facts:
- Device probes and binds correctly: hid-generic grabs it first (~3.5s),
hid-multitouch takes over ~100ms later (0018:04F3:3195.0001) - normal.
- Resulting input device (/proc/bus/input/devices) has a healthy capability
set (PROP=5 buttonpad, EV=1b, populated KEY/ABS bitmaps) - driver parsed
the report descriptor fine.
- GPIO pin is correctly configured: `gpio-171 ( |TPD0 GpioInt(0) ) in hi IRQ
ACTIVE LOW` (gpiochip0, parent platform/INTC105E:00) - idle-high is correct
rest state for active-low.
- /proc/interrupts: IRQ for “intel-gpio 171 ELAN0676:00” fires exactly twice
at boot, then never again regardless of touching/clicking the pad.
- No I2C bus errors/timeouts/NAKs in dmesg at any verbosity we checked.
- sudo evtest on the Touchpad event node: zero events ever, even raw.
- No PowerResource/_PR0/_ON/_OFF anywhere in the device’s ACPI scope
(ruling out the IOAPIC-routing-disconnect bug class seen on some other
Lunar Lake boards, e.g. LP #2168946).
- TrackPoint + its 3 dedicated buttons work fine (separate PS/2 device);
only the capacitive touchpad surface itself is dead.
ACPI findings (decompiled DSDT, device TPD0 under \_SB.PC00.I2C0):
- This is a shared “universal touchpad” ACPI slot supporting multiple
vendor chips (Synaptics/Elan) via a lookup table (ITML) matched against
TDVI/TDPI globals (presumably sourced from EC/NVRAM at the factory).
- For our chip (04F3:3195 “ELAN0676”), the table sets I2C bus speed
(SPED) to 0x00061A80 = 400kHz.
- UNCONFIRMED HYPOTHESIS, not verified: there’s an active (2026) upstream
thread that this exact chip ID is unreliable at 400kHz and needs adding
to i2c_acpi_force_100khz_device_ids (already done for sibling ID
ELAN0678; a maintainer asked whether ELAN0676 needed it too - unclear if
merged). We saw no direct I2C-level errors, so this is NOT confirmed as
root cause, just a lead worth checking.
Related upstream context:
- Backport discussion for other Lunar Lake GPIO/pinctrl fixes into
6.11-oem/6.17-oem is already happening (platform/x86 int3472,
pinctrl-intel-platform INTC10B5) - may or may not be the same root cause,
flagging for correlation.
- i2c_acpi_force_100khz_device_ids ELAN0678 patch thread (100kHz quirk
for a sibling chip ID):
https://lists.openwall.net/linux-kernel/2026/05/12/2630
interrupts remain the same before and after touch and click on the pad:
32: 0 0 0 1158 0 0
0 0 0 0 0 0 0 0
IR-IO-APIC 32-fasteoi idma64.0, i2c_designware.0
33: 0 0 0 0 1103 0
0 0 0 0 0 0 0 0
IR-IO-APIC 33-fasteoi idma64.1, i2c_designware.1
160: 0 0 0 0 0 0
0 0 0 0 0 0 0 2
intel-gpio 171 ELAN0676:00
gpio:
gpiochip0: 451 GPIOs, parent: platform/INTC105E:00, INTC105E:00:
gpio-171 ( |TPD0 GpioInt(0) ) in hi IRQ ACTIVE LOW
Request: investigate why the GPIO interrupt (pin 171, intel-gpio,
INTC105E:00) stops delivering after the first two boot-time events, and
whether the 100kHz I2C quirk applies here too.
** Affects: linux (Ubuntu)
Importance: Undecided
Status: New
** Attachment added: "device-tpd0-block.txt"
https://bugs.launchpad.net/bugs/2174247/+attachment/6017293/+files/device-tpd0-block.txt
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2174247
Title:
ELAN0676 (04F3:3195) touchpad on ThinkPad T14 Gen 6 (Intel/Lunar Lake)
enumerates correctly but produces zero input events - GPIO IRQ never
re-fires after initial boot handshake.
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2174247/+subscriptions
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs