Arch Linux

Mmdrza

Member

Registered: 2026-09-22

Posts: 15

Re: USB headset produces a periodic tick every second on Linux

ehci-hcd.park=3 / ehci-hcd.park=2 / ehci-hcd.park=1
https://wiki.archlinux.org/title/Kernel_parameters
(The module is built in so you cannot use a modprobe config)

making sure that it applied:

MmdrzaArch% sudo dmesg | grep -i park
cat /sys/module/ehci_hcd/parameters/park 2>/dev/null
[sudo] password for Mmdrza:
[    0.000000] Command line: BOOT_IMAGE=/vmlinuz-linux root=UUID=4c4a57d6-60cf-4f5e-80bd-b1ebbb8c215a rw loglevel=3 quiet ehci_hcd.park=3
[    0.043795] Kernel command line: BOOT_IMAGE=/vmlinuz-linux root=UUID=4c4a57d6-60cf-4f5e-80bd-b1ebbb8c215a rw loglevel=3 quiet ehci_hcd.park=3
3

tried all ehci-hcd.park=3 / ehci-hcd.park=2 / ehci-hcd.park=1
all of them tick the same

mmy8x

Member

Registered: 2025-03-02

Posts: 121

Re: USB headset produces a periodic tick every second on Linux

The hub is probably built into the motherboard, i don't have a external one and i don't think the headset has a hub in it

That's what it seems. Your USB 3.0 lsusb shows no hub inside the headset.

Is it just this one machine or any others with EHCI too? I tested something similar: full-speed audio dongle and a separate mouse instead of the HID interface, both behind the same hub, it works fine. And you say you even tried unbinding the HID interface and it still didn't help? Might be that it's not an EHCI scheduling bug but something wrong with this one machine.

Does anything show up in dmesg after running this?

echo 'module ehci_hcd +p' >/proc/dynamic_debug/control
echo 'module snd_usb_audio +p' >/proc/dynamic_debug/control

What if you load one or all CPUs to 100%? For example, open as many terminals as there are cores and in each:

while :; do :; done

I suppose this motherboard is too old for PCI passthrough, so you couldn't pass the EHCI controller to Windows in order to see if that works? It could still not work if the problem is due to some arcane chipset configuration rather than EHCI driver.

Last edited by mmy8x (Today 19:20:50)

seth

Member

From: Won't reply 2 private help req

Registered: 2012-09-03

Posts: 77,979

Re: USB headset produces a periodic tick every second on Linux

◉ modinfo ehci-hcd                    
name:           ehci_hcd
filename:       (builtin)
license:        GPL
file:           drivers/usb/host/ehci-hcd
author:         David Brownell
description:    USB 2.0 'Enhanced' Host Controller (EHCI) Driver
parm:           log2_irq_thresh:log2 IRQ latency, 1-64 microframes (int)
parm:           park:park setting; 1-3 back-to-back async packets (uint)
parm:           ignore_oc:ignore bogus hardware overcurrent indications (bool)

But I've even less hope for the other two parameters.

https://learn.microsoft.com/en-us/answe … -audio-dev blames HPET, if your board has no such option add "hpet=disable" to the kernel parameters…

---

How to upload text · How to boot w/o GUI · Disable Windows Fast-Start! · Fix your xinitrc

Mmdrza

Member

Registered: 2026-09-22

Posts: 15

Re: USB headset produces a periodic tick every second on Linux

Is it just this one machine or any others with EHCI too? I tested something similar: full-speed audio dongle and a separate mouse instead of the HID interface, both behind the same hub, it works fine. And you say you even tried unbinding the HID interface and it still didn't help? Might be that it's not an EHCI scheduling bug but something wrong with this one machine.

I also tested it in a Toshiba dynamic satelite B451/E laptop that is also EHCI, same tick happens there too

Does anything show up in dmesg after running this?

echo 'module ehci_hcd +p' >/proc/dynamic_debug/control
echo 'module snd_usb_audio +p' >/proc/dynamic_debug/control

Nothing appeared when i entered the command, i also tested playback which showed the normal playback start/stop lines in dmesg, nothing more:

[ 7955.811649] usb 1-1.2: Open EP 0x1, iface=1:2, idx=0
[ 7955.811656] usb 1-1.2:   channels=2, rate=48000, format=S24_3LE, period_bytes=3072, periods=64, implicit_fb=0
[ 7955.811661] usb 1-1.2: Setting params for data EP 0x1, pipe 0x8300
[ 7955.811664] usb 1-1.2: Set up 3 URBS, ret=0
[ 7955.811691] usb 1-1.2: Setting usb interface 1:2 for EP 0x1
[ 7955.812995] usb 1-1.2: 1:2 Set sample rate 48000, clock 0
[ 7955.813286] usb 1-1.2: Starting data EP 0x1 (running 0)
[ 7955.813293] usb 1-1.2: ep 01: reserve iso @ 4+8 (0.4+1) [7/0 us] mask 0030
[ 7955.813301] usb 1-1.2: 3 URBs submitted for EP 0x1
[ 7955.815409] usb 1-1.2: 1:2 Start Playback PCM
[ 7967.648468] usb 1-1.2: Stopping data EP 0x1 (running 1)
[ 7967.648476] usb 1-1.2: 1:2 Stop Playback PCM
[ 7967.662938] usb 1-1.2: Closing EP 0x1 (count 1)
[ 7967.662943] usb 1-1.2: Setting usb interface 1:0 for EP 0x1
[ 7967.663090] usb 1-1.2: ep 01: release iso @ 4+8 (0.4+1) [7/0 us] mask 0030
[ 7967.663096] usb 1-1.2: EP 0x1 closed

What if you load one or all CPUs to 100%? For example, open as many terminals as there are cores and in each:

while :; do :; done

Nothing appeared in dmesg and the ticking is still present

I suppose this motherboard is too old for PCI passthrough, so you couldn't pass the EHCI controller to Windows in order to see if that works? It could still not work if the problem is due to some arcane chipset configuration rather than EHCI driver.

the motherboard actually does support VT-d, i just tested it with EHCI passthrough to the VM and it does not tick

Last edited by Mmdrza (Today 22:01:31)

Mmdrza

Member

Registered: 2026-09-22

Posts: 15

Re: USB headset produces a periodic tick every second on Linux

◉ modinfo ehci-hcd                    
name:           ehci_hcd
filename:       (builtin)
license:        GPL
file:           drivers/usb/host/ehci-hcd
author:         David Brownell
description:    USB 2.0 'Enhanced' Host Controller (EHCI) Driver
parm:           log2_irq_thresh:log2 IRQ latency, 1-64 microframes (int)
parm:           park:park setting; 1-3 back-to-back async packets (uint)
parm:           ignore_oc:ignore bogus hardware overcurrent indications (bool)

But I've even less hope for the other two parameters.

https://learn.microsoft.com/en-us/answe … -audio-dev blames HPET, if your board has no such option add "hpet=disable" to the kernel parameters…

Tested the remaining two parameters plus hpet-disable, no change to the tick