OPINIONATED OMARCHY
TROUBLESHOOTING CORPUS
790
RECORDS
790
AUDITED
0
UNCHECKED
2388
SOURCES
LIVE
power-suspend high rare DISPUTE

Fix a kernel panic on resume on Intel laptops with an I2C touchpad

archomarchyendeavouroscachyosmanjarolaptopintelmkinitcpio
CORRECTED An auditor found the fix wrong or incomplete and rewrote it.
AUDIT NOTE

Checked on this workstation (omarchy 4.0.2-1, omarchy-settings 4.0.2-1, Hyprland 0.56.2-1, kernel 7.1.9-arch1-2), which is a DESKTOP with an NVIDIA card, so no Intel LPSS touchpad is present and the panic itself, the fix and the suspend cycles could NOT be exercised. lsmod | grep intel_lpss returns nothing here and modinfo intel_lpss_pci shows the module exists at /lib/modules/7.1.9-arch1-2/kernel/drivers/mfd/intel-lpss-pci.ko.zst with no module parameters, so the lsmod step in the fix is sound but unexercised. Four things were wrong. First, the boot command: /etc/mkinitcpio.d/ is empty on this machine and neither the linux nor the mkinitcpio package ships a preset into it, so sudo mkinitcpio -P is a no-op on Omarchy 4, and /usr/local/bin/mkinitcpio (from limine-mkinitcpio-hook 1.37.1-1) is a wrapper that warns about exactly this and offers limine-mkinitcpio. Second, the file: Omarchy assigns MODULES and HOOKS from /etc/mkinitcpio.conf.d/ (nvidia.conf, omarchy_hooks.conf, omarchy_resume.conf, thunderbolt_module.conf are present), /etc/mkinitcpio.conf is package stock, and /usr/bin/mkinitcpio lines 1120 to 1137 concatenate the main file then the drop-ins in sort -V order, so a drop-in using += is the correct shape. Third, the verify: there is no /boot/initramfs-linux.img on a UKI install, the image is inside /boot/EFI/Linux/omarchy_linux.efi (derived from CUSTOM_UKI_NAME="omarchy" in /etc/limine-entry-tool.d/omarchy-defaults.conf and the UKI_PATH line in /usr/share/libalpm/scripts/limine-mkinitcpio-install), /boot is dmask=0077 so the read needs root, and grep intel_lpss would miss the file because it is stored as intel-lpss-pci.ko.zst with hyphens. lsinitcpio 41.1 does unpack a UKI (detect_uki/unpack_uki). Fourth, the danger told the reader to keep a fallback boot entry, and Omarchy 4 has none: MKINITCPIO_FALLBACK is unset in /etc/default/limine and in every /etc/limine-entry-tool.d/*.conf, and limine-list shows only Omarchy > linux, a Snapshots submenu and the EFI fallback bootloader, so Snapshots is the real recovery path. On sources: the cited Arch wiki section 'Touchpad causes a kernel panic on resume' does say what the record claims, but its single citation is https://bbs.archlinux.org/viewtopic.php?id=231881, which I read in full. It is titled 'Stuck during resume from hibernation', dated 2017, and the whole fix is one reply of 2017-12-23 saying intel_lpss_pci in MODULES solved it for that person. There is no evidence for suspend to RAM anywhere, and the initramfs is not re-entered on an S3 or s2idle resume, which is why the symptom now names hibernation. The cause was rewritten because presence in the image is not the mechanism: /usr/lib/initcpio/init line 38 runs modprobe -qab $MODULES, and the autodetect hook already pulls the module in on affected hardware. Frequency was moved from occasional to rare because the whole evidence base is one report from kernel 4.14 era, and current mainline drivers/mfd/intel-lpss.c has a full LATE_SYSTEM_SLEEP_PM_OPS with an explicit resume-children-before-sleep step, so the 2017 failure may no longer exist. I could not test whether it does. The record carried no kernel version floor and no fixed-in claim, so there was none to correct.

The cause above was rewritten on 2026-09-12 to match this note; the fix was corrected by the audit itself.
RISK

This is the opposite of the fix for a black screen on hibernate resume, where modules have to be removed from MODULES instead. Change one thing at a time.

Omarchy 4 builds no fallback boot entry. MKINITCPIO_FALLBACK is unset in /etc/default/limine and in every /etc/limine-entry-tool.d/*.conf, so limine-list shows only Omarchy > linux, a Snapshots submenu and the EFI fallback bootloader. If a rebuilt initramfs stops the machine booting, recover from a Snapshots entry, not from a fallback kernel entry that does not exist. On plain Arch, keep the -fallback entry your bootloader already generates.

Symptom

Resuming gives a dead machine with the Caps Lock LED blinking, which is a kernel panic. Nothing is written to the journal because the panic happens before the disk is writable again. Intel laptop with an I2C/LPSS touchpad.

The only reported case is resume from hibernation, where the initramfs runs again to restore the image. Resume from suspend to RAM never re-enters the initramfs, so if your machine dies coming back from an ordinary suspend, this record is probably not your problem.

Cause

MODULES= in mkinitcpio does more than copy a module into the image. /usr/lib/initcpio/init runs modprobe -qab $MODULES before anything else, so naming intel_lpss_pci there force-loads the driver for the LPSS controller behind the touchpad early, before the hibernation image is restored. The autodetect hook already copies that module into the image on a laptop that uses it, so mere presence is not the fix, the early load is.

The evidence is thin and old. The Arch wiki's only citation is one 2017 forum reply about a hibernation resume hang, and current mainline drivers/mfd/intel-lpss.c carries a full LATE_SYSTEM_SLEEP_PM_OPS with an explicit resume-children-before-sleep step, so the original bug may well be gone on kernel 7.1.

Fix

Force the module to load early in the initramfs. On Omarchy 4 and on plain Arch the file you edit and the command you run are different.

**On Omarchy 4.** /etc/mkinitcpio.conf is package stock and Omarchy assigns MODULES and HOOKS from drop-ins in /etc/mkinitcpio.conf.d/. Add your own drop-in and use +=, so it cannot clobber the MODULES that nvidia.conf and thunderbolt_module.conf set:


printf 'MODULES+=(intel_lpss_pci)\n' | sudo tee /etc/mkinitcpio.conf.d/intel_lpss.conf
sudo limine-mkinitcpio

Do not use sudo mkinitcpio -P here. Omarchy 4 boots a UKI that limine-mkinitcpio-hook builds from /usr/lib/modules/*/pkgbase, and /etc/mkinitcpio.d/ is empty, so -P finds no preset and silently does nothing. /usr/local/bin/mkinitcpio is a wrapper that spots -P or -p, warns that it did not update the Limine entries, and offers to run limine-mkinitcpio for you. sudo limine-update also works and additionally reinstalls the bootloader.

**On plain Arch** with a conventional initramfs and no Limine UKI:


# /etc/mkinitcpio.conf
MODULES=(intel_lpss_pci)

sudo mkinitcpio -P

If MODULES=() already has entries, append rather than replace.

Either way, confirm the module is the one your touchpad sits behind:


lsmod | grep intel_lpss

Verify

On Omarchy 4 the initramfs lives inside the UKI and /boot is mounted dmask=0077, so the check needs root. lsinitcpio 41 unpacks a UKI, and the file inside the image is named with hyphens, so grep for intel-lpss and not intel_lpss:


sudo lsinitcpio -l /boot/EFI/Linux/omarchy_linux.efi | grep intel-lpss

On plain Arch: sudo lsinitcpio -l /boot/initramfs-linux.img | grep intel-lpss.

Then five hibernate and resume cycles complete without a panic.

Sources

touchpad-kernel-panic-on-resume-intel-lpss