Public bug reported:
Summary: Synaptics I2C-HID touchpad (06CB:CE26) silently loses
multitouch/gesture
data after HID re-initialization, falling back to plain relative "Mouse" HID
interface with correct-looking but non-functional capability descriptor
## Hardware
- System: Alienware 16 Aurora AC16250
- BIOS: 1.15.0 (2026-07-08) — bug present on both 1.10.0 and 1.15.0
- Touchpad: Synaptics, Bus 0018 (I2C-HID), Vendor 06cb, Product ce26, Version
0100
- ACPI path: i2c-VEN_06CB:00
- Kernel: 7.0.0-31-generic (Ubuntu 24.04.4 HWE)
- Drivers involved: i2c_hid_acpi, hid-multitouch, hid-generic
## Summary of the bug
This touchpad exposes two HID interfaces from the same physical chip: a
multitouch "Touchpad" interface (handled by hid-multitouch, supports
gestures) and a plain relative "Mouse" interface (basic pointer movement
only, no gestures, different/uncalibrated pointer acceleration).
Under certain HID (re-)initialization events, the Touchpad interface's
capability descriptor continues to report full multitouch support
correctly, but the interface stops delivering any actual input events.
The desktop environment (GNOME/libinput/X11 in our case) then silently
falls back to processing pointer input through the secondary Mouse
interface instead. Symptoms as experienced by the user:
- Cursor movement still works (via the Mouse interface)
- All gestures (2-finger scroll, pinch, 3-finger swipe) stop working
- Pointer "sensitivity" feels different/higher (Mouse interface uses a
different acceleration curve than the Touchpad interface's tuned one)
This is confirmed NOT to be a simple capability-descriptor loss: the
`ABS` bitmap read from `/proc/bus/input/devices` for the Touchpad node
remains `2e0800000000003` (ABS_MT_SLOT, ABS_MT_POSITION_X/Y,
ABS_MT_TRACKING_ID, ABS_MT_TOOL_TYPE all present) even while the
interface is fully dead. We confirmed the dead state directly by running
`xinput test <touchpad-x11-id>` while actively swiping the trackpad for
5+ seconds and observing zero events, then running
`xinput test <mouse-x11-id>` during the same physical action and seeing
a continuous stream of `motion` events there instead.
## Two distinct triggers observed for the same dead-endpoint state
### Trigger 1: System suspend/resume (S0ix / s2idle)
On some (not all) resumes from suspend, the kernel logs:
```
i2c_hid_acpi i2c-VEN_06CB:00: failed to set a report to device: -121
i2c_hid_acpi i2c-VEN_06CB:00: failed to set a report to device: -121
i2c_hid_acpi i2c-VEN_06CB:00: failed to change power setting.
```
(-121 is -EREMOTEIO — the I2C slave did not ACK part of the message.)
Immediately after, the Touchpad interface is in the dead state described
above.
### Trigger 2: Cold boot, with no suspend/resume involved and NO kernel
error logged at all
On a fresh boot, we observed the same physical HID instance number
(`0018:06CB:CE26.0001`) claimed twice in immediate succession:
```
kernel: input: VEN_06CB:00 06CB:CE26 Mouse as
.../0018:06CB:CE26.0001/input/input6
kernel: input: VEN_06CB:00 06CB:CE26 Touchpad as
.../0018:06CB:CE26.0001/input/input7
kernel: hid-generic 0018:06CB:CE26.0001: input,hidraw0: I2C HID v1.00 Mouse
[VEN_06CB:00 06CB:CE26] on i2c-VEN_06CB:00
kernel: input: VEN_06CB:00 06CB:CE26 Mouse as
.../0018:06CB:CE26.0001/input/input9
kernel: input: VEN_06CB:00 06CB:CE26 Touchpad as
.../0018:06CB:CE26.0001/input/input10
kernel: hid-multitouch 0018:06CB:CE26.0001: input,hidraw0: I2C HID v1.00 Mouse
[VEN_06CB:00 06CB:CE26] on i2c-VEN_06CB:00
```
i.e. hid-generic claims the device first, then hid-multitouch reclaims
it moments later — suggesting the initial HID report descriptor fetch is
flaky/inconsistent on this specific chip, independent of ACPI resume.
Sometime after this boot (no suspend occurred), the Touchpad interface
was found dead by the same xinput test method above, with nothing
whatsoever in the kernel log to indicate a failure.
## Attempted fixes and their results
1. **BIOS update 1.10.0 → 1.15.0** (via fwupd/LVFS): did not fix the
issue. The suspend/resume failure recurred on the very next
suspend/resume cycle after updating, and the no-suspend cold-boot
variant recurred the following day.
2. **Unbind/rebind the i2c_hid_acpi driver instance**
(`echo i2c-VEN_06CB:00 > /sys/bus/i2c/drivers/i2c_hid_acpi/unbind`,
then `> .../bind`) sometimes restores the Touchpad interface to a
working state, and sometimes does not — inconsistent across
otherwise-identical attempts. This strongly suggests a hardware/
firmware timing race rather than a deterministic software state,
since retrying the same recovery action does not reliably converge
on success.
## Reproduction steps
1. Suspend the system (`systemctl suspend`) and resume, OR simply
observe behavior shortly after a cold boot.
2. Check `journalctl -k | grep i2c_hid_acpi` for the power-setting
failure (may or may not be present — see Trigger 2).
3. Attempt a two-finger scroll/pinch on the trackpad. If gestures don't
work but the cursor still moves, identify the touchpad's two X11
device IDs via `xinput list | grep 06CB`.
4. Run `xinput test <touchpad-id>` while swiping — compare against
`xinput test <mouse-id>` during the same physical action. The
dead/live split confirms the fallback.
## Environment
- `uname -r`: 7.0.0-31-generic
- Ubuntu 24.04.4 LTS (Noble Numbat)
- Session type: X11 (GNOME)
- libinput correctly classifies the Touchpad node as `cap:pg`
(pointer+gesture) even while it is in the dead state — i.e. libinput's
own device classification is not at fault; the kernel HID layer is not
delivering data to that device node at all during the dead periods.
ProblemType: Bug
DistroRelease: Ubuntu 24.04
Package: linux-image-7.0.0-31-generic 7.0.0-31.31~24.04.1
ProcVersionSignature: Ubuntu 7.0.0-31.31~24.04.1-generic 7.0.14
Uname: Linux 7.0.0-31-generic x86_64
ApportVersion: 2.28.3-0ubuntu0.1
Architecture: amd64
AudioDevicesInUse:
USER PID ACCESS COMMAND
/dev/snd/controlC1: will 3116 F.... wireplumber
/dev/snd/controlC0: will 3116 F.... wireplumber
/dev/snd/seq: will 3110 F.... pipewire
CRDA: N/A
CasperMD5CheckResult: pass
CurrentDesktop: ubuntu:GNOME
Date: Wed Sep 23 10:19:03 2026
InstallationDate: Installed on 2026-09-06 (17 days ago)
InstallationMedia: Ubuntu 24.04.4 LTS "Noble Numbat" - Release amd64 (20260210)
MachineType: Alienware Alienware 16 Aurora AC16250
ProcEnviron:
LANG=en_US.UTF-8
PATH=(custom, no user)
SHELL=/bin/bash
TERM=xterm-256color
XDG_RUNTIME_DIR=<set>
ProcFB: 0 i915drmfb
ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-7.0.0-31-generic
root=/dev/mapper/ubuntu--vg-ubuntu--lv ro quiet splash vt.handoff=7
RelatedPackageVersions:
linux-restricted-modules-7.0.0-31-generic N/A
linux-backports-modules-7.0.0-31-generic N/A
linux-firmware 20240318.git3b128b60.0ubuntu3.1
SourcePackage: linux-hwe-7.0
UpgradeStatus: No upgrade log present (probably fresh install)
dmi.bios.date: 07/08/2026
dmi.bios.release: 1.15
dmi.bios.vendor: Alienware
dmi.bios.version: 1.15.0
dmi.board.name: 0MWC59
dmi.board.vendor: Alienware
dmi.board.version: A00
dmi.chassis.type: 10
dmi.chassis.vendor: Alienware
dmi.ec.firmware.release: 1.18
dmi.modalias:
dmi:bvnAlienware:bvr1.15.0:bd07/08/2026:br1.15:efr1.18:svnAlienware:pnAlienware16AuroraAC16250:pvr:rvnAlienware:rn0MWC59:rvrA00:cvnAlienware:ct10:cvr:sku0D35:pfaAlienware:
dmi.product.family: Alienware
dmi.product.name: Alienware 16 Aurora AC16250
dmi.product.sku: 0D35
dmi.sys.vendor: Alienware
** Affects: linux-hwe-7.0 (Ubuntu)
Importance: Undecided
Status: New
** Tags: amd64 apport-bug noble
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2168019
Title:
Synaptics I2C-HID touchpad (06CB:CE26) silently loses
multitouch/gesture data after HID re-initialization
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux-hwe-7.0/+bug/2168019/+subscriptions
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
