[bugreport] JMicron JMS567 kills AMD Renoir xHCI

[bugreport] JMicron JMS567 kills AMD Renoir xHCI

[Tobias Jakobi]

Greetings,

I recently got a new notebook, a Dell G5 5505, mostly composed of AMD 
hardware. After the initial hiccups most things work smoothly now, until 
I tried to connect my backup drive (2TB HDD in a USB3 enclosure with 
JMicron bridge):
Nov  8 23:39:09 leena kernel: usb 1-2: new low-speed USB device number 2 
using xhci_hcd
Nov  8 23:39:19 leena kernel: xhci_hcd 0000:08:00.3: xHCI host not 
responding to stop endpoint command.
Nov  8 23:39:19 leena kernel: xhci_hcd 0000:08:00.3: USBSTS:
Nov  8 23:39:19 leena kernel: xhci_hcd 0000:08:00.3: xHCI host 
controller not responding, assume dead
Nov  8 23:39:19 leena kernel: xhci_hcd 0000:08:00.3: HC died; cleaning up
Nov  8 23:39:19 leena kernel: usb usb1-port2: couldn't allocate usb_device
Nov  8 23:39:19 leena kernel: usb usb2-port2: couldn't allocate usb_device

Sadly that's not a random hiccup, but reproducable. I did some research 
and some sources recommended to add iommu=pt, but not changes here. The 
BIOS of the G5 is kinda sparse, the only option looking relevant was 
"USB emulation", but toggling it also did not change anything.

I uploaded several logs:
kernel_g5.log (full kernel log from the G5)
lsusb_g5. log (lsusb log, JMicron not plugged in)
lspci_g5.log (lspci log)
lsusb_enclosure (lsusb excerpt for the JMicron, taken via a different 
system)

You can find the files here: 
https://gist.github.com/tobiasjakobi/84b6d58d5c387cfb4bbba87e7a589ca8

The "different system" is another AMD based one (desktop, AMD Athlon 
3000G, ASRock B450M-HDV R4.0). lspci says the following about the USB hw:
01:00.0 USB controller: Advanced Micro Devices, Inc. [AMD] 400 Series 
Chipset USB 3.1 XHCI Controller (rev 01)
09:00.3 USB controller: Advanced Micro Devices, Inc. [AMD] Raven2 USB 3.1

Anything I can do to help triage this problem?

With best wishes,
Tobias