Moonlander and mouse disconnects
About
For some odd reason, my computer would sometimes ignore inputs from my keyboard and mouse when my moonlander is plugged in. This happened more frequently after some period of inactivity.
It did not happen when I have planck keyboard though.
Fix
Someone on the internet recommended going through the BIOS settings. The IO in BIOS stands for input/output, so it would make sense that keyboard and mouse activity progresses through there.
I found some configuration for xHCI. A search about xHCI on google highlights the following idea xHCI's wikipedia page:
The xHCI eliminates host memory based USB transaction schedules, enabling zero host memory activity when there is no USB data movement.
This sounds like what could be the cause for my issue!
Sure enough, I turned it off and I have successfully written this blog post without any jitters.
Update (2024-02-04)
The previous fix did not resolve my issue. This time around, I was increasingly frustrated and dared to challenge the kernel.
linux messages
Luckily the dmesg was filled with logs that highlighted my issue.
dmesg -T | grep 'reset full-speed USB device number' | tail -n 10[Sun Feb 4 00:39:16 2024] usb 1-1.2.4.4: reset full-speed USB device number 9 using xhci_hcd [Sun Feb 4 00:39:17 2024] usb 1-1.2.4.1: reset full-speed USB device number 7 using xhci_hcd [Sun Feb 4 00:39:18 2024] usb 1-1.2.4.4: reset full-speed USB device number 9 using xhci_hcd [Sun Feb 4 00:39:20 2024] usb 1-1.2.4.4: reset full-speed USB device number 9 using xhci_hcd [Sun Feb 4 00:39:21 2024] usb 1-1.2.4.1: reset full-speed USB device number 7 using xhci_hcd [Sun Feb 4 00:39:24 2024] usb 1-1.2.4.1: reset full-speed USB device number 7 using xhci_hcd [Sun Feb 4 00:39:25 2024] usb 1-1.2.4.4: reset full-speed USB device number 9 using xhci_hcd [Sun Feb 4 00:39:45 2024] usb 1-1.2.4.4: reset full-speed USB device number 9 using xhci_hcd [Sun Feb 4 00:39:49 2024] usb 1-1.2.4.4: reset full-speed USB device number 9 using xhci_hcd [Sun Feb 4 00:39:49 2024] usb 1-1.2.4.4: reset full-speed USB device number 9 using xhci_hcd
The device 7 and 9 are my moonlander and mouse.
lsusb | grep -e '007' -e '009'Bus 001 Device 009: ID 3297:1969 ZSA Technology Labs Moonlander Mark I Bus 001 Device 007: ID 046d:c53a Logitech, Inc. PowerPlay Wireless Charging System
This is a sigh of relief as I've finally discovered a lead.
hardware troubleshooting
My keyboard and mouse goes through a powered USB hub and is connected to my monitor's KVM which is connected to my desktop.
If I attach my keyboard directly to my desktop, there is no sign of the kernel resetting the USB device. This leaves me thinking that the fault lies in either the USB Hub or the KVM.
I eliminate the Monitor KVM node but the problem persists. This is where I stop for the night as it's 3 in the morning.
While resting on this problem in bed, I thought about connecting the Powered USB Hub to my laptop.
After I wake up, this is the first thing that I try and a success! There is no messages from the kernel about resetting the USB devices.
Motherboard
In the previous section, we've eliminated the Powered USB Hub and the Monitor KVM as the potenteial culprit. This leaves us with only the Desktop as the cause.
Sure enough, a search on the internet highlights that the chipsets for Zen 3 processors have a known issue with USB ports.
I have an older motherboard (B450) than the ones (B550) but I lumped myself into the group nevertheless.
The users report fixing the issue with configuring a variety of different things in the motherboard's settings:
- CPPC
- C-State
- XHCI Hands-off (what I did earlier)
- Power Supply Idle Control
- …
I disabled the CPPC right getting into the troubleshooting the hardware, and it did not work.
Finally, disabling the Global C-State Control and setting the Power Supply Idle Control to Typical did the trick!