Bring back a touchpad that turned into a plain PS/2 mouse
omacom/omarchy#12181 supports the symptom (about 42 'lost sync at byte 1', 'synaptics: Unable to query device', fallback to PS/2 Generic Mouse) and the recovery 'modprobe -r psmouse && modprobe psmouse', both reported by the user. It also supports the separate boot-time non-enumeration (3 of 31 boots on linux-omarchy 7.2.5-3, 0 of 26 on stock 7.2.3), still open. Kernel source confirms the log strings, and that the InterTouch hint reads 'If i2c-hid and hid-rmi are not used, you might want to try setting psmouse.synaptics_intertouch to 1'. Three defects. (1) The cause says the more robust route is SMBus/RMI4 via the I2C-HID node, but DELL0A6F is I2C-HID, not RMI4, and the reporter saw the same desyncs over I2C-HID and under Windows. (2) Step 2 sends the user to synaptics_intertouch=1 without the kernel's own condition that i2c-hid is not already driving the pad, which is exactly the reporter's machine. (3) i8042.nomux=1 is presented as the persistent mitigation, but the Arch wiki offers it only for cursor jumps from i8042 IRQ loss, and nobody in the cited issue tried it for this failure. Also omarchy-restart-trackpad (the Omarchy menu's trackpad reset) only rebinds i2c_hid_acpi and intel_quicki2c, so it does not help a PS/2 pad, which is worth saying. Confirmed locally: i8042 is in modules.builtin for both kernels, psmouse is a loadable module, and the limine drop-in form is correct. The cross-reference to the unaudited linux-omarchy-7-2-5-hdmi-audio-underruns record is dropped in favour of picking the kernel at the Limine menu. Nothing was exercised on a touchpad.
i8042 parameters act on the controller that also runs the built-in keyboard. If the keyboard misbehaves after the change, edit the command line at the Limine menu for one boot or use an external keyboard, then delete /etc/limine-entry-tool.d/99-i8042.conf and run sudo limine-update. Blacklisting i2c_hid_acpi in step 5 also disables I2C touchscreens.
Symptom
The touchpad suddenly behaves like a cheap mouse: the pointer moves but two-finger scrolling, tap-to-click and gestures are gone. Or it is missing entirely on some boots. journalctl -k -b | grep -i psmouse shows a burst like psmouse serio1: TouchPad at isa0060/serio1/input0 lost sync at byte 1 repeated dozens of times, then synaptics: Unable to query device, and the device comes back as PS/2 Generic Mouse. hyprctl devices lists a generic mouse where synps/2-synaptics-touchpad used to be.
Cause
The PS/2 link to the touchpad lost protocol sync. The kernel's psmouse driver logs lost sync at byte 1, tries to reconnect, fails to re-identify the device as Synaptics (synaptics: Unable to query device) and falls back to the bare PS/2 mouse protocol, which carries no multi-finger data. Why the sync is lost is not established. The reporter on an Alienware m15 Ryzen Edition R5 saw the same desyncs under Windows and with the pad bound over I2C-HID (DELL0A6F, 06CB:CE62), which points at the touchpad firmware or the embedded controller rather than at the kernel. The Arch wiki names IRQ loss on the i8042 controller as one general cause of erratic PS/2 touchpad input. A separate failure on the same machine, the touchpad not being probed at all on 3 of 31 boots under linux-omarchy 7.2.5-3 against 0 of 26 under stock linux 7.2.3, was still open upstream on 2026-10-05.
Fix
**1. Recover now, without rebooting.**
sudo modprobe -r psmouse && sudo modprobe psmouse
hyprctl devices | grep -i -A2 touchpad
The Omarchy menu's trackpad reset (omarchy-restart-trackpad) only rebinds i2c_hid_acpi and intel_quicki2c, so it does not help a pad that is on PS/2.
**2. Check whether the pad has another bus, and whether it is already in use.**
grep -E 'Name=|Sysfs=' /proc/bus/input/devices | grep -i -B1 -E 'touchpad|i2c|rmi'
journalctl -k -b | grep -iE 'different bus|rmi4|i2c_hid'
If the log says Your touchpad (...) says it can support a different bus and no i2c_hid or rmi4 touchpad device is present, the kernel suggests psmouse.synaptics_intertouch=1. See the synaptics-touchpad-dead-needs-intertouch record for that procedure. Do not do this when the pad is also exposed through I2C-HID (for example a DELL0A6F node): the kernel's own message says the option is for machines where i2c-hid and hid-rmi are not used, and on the reported machine the desyncs also happened over I2C-HID.
**3. Optional: the Arch wiki's mitigation for i8042 IRQ loss, i8042.nomux=1.** It is documented for jumping cursors, not specifically for this fallback, and nobody in the cited report tested it. i8042 is built into the kernel (listed in modules.builtin for both 7.2.5-3-omarchy and 7.2.3-arch1-3), so it goes on the kernel command line, not in modprobe.d.
Omarchy 4:
printf 'KERNEL_CMDLINE[default]+=" i8042.nomux=1"\n' \
| sudo tee /etc/limine-entry-tool.d/99-i8042.conf
sudo limine-update
Plain Arch: add i8042.nomux=1 to GRUB_CMDLINE_LINUX_DEFAULT in /etc/default/grub and run sudo grub-mkconfig -o /boot/grub/grub.cfg, or to the options line in /boot/loader/entries/*.conf.
**4. If the pad is missing at boot only on linux-omarchy**, pick the stock linux kernel at the Limine menu for a number of boots and compare. The reported failure rate is about 1 boot in 10, so a few good boots prove little.
**5. Last resort, as the reporter did:** stop the kernel driving the pad and use an external mouse.
printf 'blacklist psmouse\nblacklist i2c_hid_acpi\n' | sudo tee /etc/modprobe.d/no-touchpad.conf
i2c_hid_acpi also drives I2C touchscreens and other I2C-HID devices, so leave that line out if the machine has any. Delete the file and reboot to undo it.
Verify
journalctl -k -b | grep -i synaptics shows the touchpad identified (for example Touchpad model: 1, fw: 9.16), hyprctl devices lists the touchpad by its Synaptics name, and two-finger scrolling works. After a reboot with the drop-in, cat /proc/cmdline contains i8042.nomux=1.