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

Make USB autosuspend actually turn off, because Omarchy's modprobe.d option never applies

omarchyarchliminelaptopdesktopusbbluetooth
CORRECTED An auditor found the fix wrong or incomplete and rewrote it. · checked against Omarchy omarchy 4.0.4-1 on 2026-10-05
AUDIT NOTE

Core claim confirmed on this workstation: the file contains options usbcore autosuspend=-1, usbcore is in modules.builtin for 7.2.5-3-omarchy, the parameter reads 2, /proc/cmdline has no usbcore option. Quattro still ships the same file. Issues #8097, #12294, #13982 and PRs #10140 and #10111 (both open) support the cause and the symptoms, including the Thunderbolt hotplug case (#8097). The kernel USB power-management doc confirms that a built-in usbcore takes usbcore.autosuspend= on the command line and that a runtime write only affects devices detected afterwards. The drop-in approach is right: limine-common-functions loads /etc/limine-entry-tool.d/*.conf and the installer's /etc/default/limine appends with +=, and migration 1786482992.sh rebuilds with limine-mkinitcpio. Two errors. First, the verify step: the parameter is in seconds and the per-device autosuspend_delay_ms is in milliseconds (here a parameter of 2 gives 2000 on every device), so -1 produces -1000, not -1. Second, the cause says a device udev sets to power/control=auto through the hwdb keeps autosuspending regardless. 60-autosuspend.rules writes only power/control (and a delay only when the hwdb sets ID_AUTOSUSPEND_DELAY_MS), and the kernel doc says a negative delay means never autosuspend, so such a device should inherit -1000 and stay awake. The evidence in #12095 for the contrary wrote the parameter at runtime and reloaded btusb without re-enumerating the USB device, which by the kernel doc cannot change its delay. TLP is the confirmed exception because it writes the delay explicitly. Not exercised: no command line was changed and nothing rebuilt.

The cause above was rewritten on 2026-10-05 to match this note; the fix was corrected by the audit itself.
RISK

Disabling USB autosuspend for every device raises idle power draw on laptops. A syntax error in the limine-entry-tool.d drop-in can make limine-mkinitcpio fail, so read its output before rebooting and keep the Limine snapshot entries available as a way back.

Symptom

USB devices still drop out after sitting idle or after resume, even though Omarchy ships /etc/modprobe.d/omarchy-usb-autosuspend.conf containing options usbcore autosuspend=-1. cat /sys/module/usbcore/parameters/autosuspend prints 2, not -1. Reported forms: dock keyboard and mouse vanish for most of a minute after resume, a USB Bluetooth adapter drops an A2DP stream every few minutes (spa.bluez5.sink.media: Missing completion reports for packet), or a device plugged into a USB-C or Thunderbolt port after boot is never detected at all.

Cause

usbcore is built into the kernel (CONFIG_USB=y) in both linux-omarchy and stock Arch linux. Files in /etc/modprobe.d/ are only read when modprobe loads a module, and a built-in module is never loaded, so the option is silently ignored and the parameter stays at the compiled default of 2 seconds. The Arch wiki and the kernel's USB power-management documentation both say a built-in module's parameter has to go on the kernel command line. Confirmed on this workstation (omarchy-settings 4.0.4-1, linux-omarchy 7.2.5-3): the file is owned by omarchy-settings, usbcore is listed in /lib/modules/$(uname -r)/modules.builtin, the parameter reads 2, and /proc/cmdline carries no usbcore option. The upstream quattro branch still ships the same file as of 2026-10-04, and the PRs that move it to the command line (#10140, #10111) are open and unmerged. Limits on what the parameter does even when set correctly: it only sets the default idle delay for devices as they enumerate, so a runtime write does nothing for devices already present. A device that systemd's hwdb marks ID_AUTOSUSPEND=1 gets power/control=auto from 60-autosuspend.rules, but that rule does not write a delay unless the hwdb also sets ID_AUTOSUSPEND_DELAY_MS, so per the kernel documentation it should inherit the negative delay and not autosuspend. That case has not been tested with the parameter on the command line. TLP is different: it writes autosuspend_delay_ms explicitly, so TLP users keep autosuspend regardless of this parameter. That is a separate record.

Fix

**Check that you are affected.** Neither command needs root:


cat /sys/module/usbcore/parameters/autosuspend          # 2 means the modprobe.d file did nothing
grep -c usbcore /lib/modules/$(uname -r)/modules.builtin  # 1 means usbcore is built in
tr ' ' '\n' < /proc/cmdline | grep usbcore               # empty means no command-line override

**Omarchy 4 (Limine, UKI).** The command line is embedded in the UKI, so it goes in a limine-entry-tool drop-in and the image is rebuilt. Never edit /boot/limine.conf, which limine-entry-tool regenerates.


echo 'KERNEL_CMDLINE[default]+=" usbcore.autosuspend=-1"' | sudo tee /etc/limine-entry-tool.d/90-usb-autosuspend.conf
sudo limine-mkinitcpio

limine-mkinitcpio is the same rebuild Omarchy's own migration 1786482992.sh runs after a command-line change. Read its output for errors, then reboot.

**Plain Arch.** Add usbcore.autosuspend=-1 to your boot loader's kernel command line: GRUB_CMDLINE_LINUX_DEFAULT in /etc/default/grub followed by sudo grub-mkconfig -o /boot/grub/grub.cfg, or the options line in /boot/loader/entries/*.conf for systemd-boot, or /etc/kernel/cmdline followed by sudo mkinitcpio -P for a UKI built from presets.

**To test before rebooting.** The parameter is writable at runtime, but per the kernel documentation it only sets the delay for devices that enumerate afterwards, so replug the device after writing it:


echo -1 | sudo tee /sys/module/usbcore/parameters/autosuspend

**Pin one device instead of all of them.** If only one device misbehaves, a udev rule is narrower and keeps autosuspend for everything else. Writing on to power/control prevents autosuspend for that device:


# /etc/udev/rules.d/61-usb-no-autosuspend.rules
ACTION=="add", SUBSYSTEM=="usb", ATTR{idVendor}=="0bda", ATTR{idProduct}=="8156", TEST=="power/control", ATTR{power/control}="on"

Replace the IDs with the ones lsusb prints for your device. The 61- prefix makes it run after systemd's 60-autosuspend.rules.

Verify

After the reboot, cat /sys/module/usbcore/parameters/autosuspend prints -1 and tr ' ' '\n' < /proc/cmdline | grep usbcore prints usbcore.autosuspend=-1. The per-device delay is in milliseconds, so cat /sys/bus/usb/devices/*/power/autosuspend_delay_ms | sort | uniq -c shows -1000 for devices that enumerated with the new default (a value of 2 seconds shows as 2000). A device that still shows 2000 had its delay written by something else, such as TLP or a hwdb ID_AUTOSUSPEND_DELAY_MS entry.

Sources

omarchy-usb-autosuspend-modprobe-conf-has-no-effect