Re: [BUG] xHCI: AMD Strix Halo [1022:158b] isochronous full-duplex USB audio instability (frame active: -18)

[Michal Pecio <michal.pecio@gmail.com>]

On Sun, 21 Jun 2026 18:30:23 +0200, Ingo Haenlein wrote:
> Hello,
> 
> I am reporting a reproducible isochronous full-duplex USB audio 
> instability on an AMD Strix Halo system (HP ZBook Ultra G1a) under 
> Linux. A detailed bug report including kernel logs, USB descriptors, and 
> endpoint data has been filed at:
> 
> https://bugzilla.kernel.org/show_bug.cgi?id=221650
> 
> This mail summarizes the findings and forwards the issue to linux-usb as 
> the problem appears to be in the xHCI host controller path rather than 
> in snd-usb-audio.
> 
> 
> Affected system
> ---------------
> 
> Machine:    HP ZBook Ultra G1a (AMD Strix Halo / Ryzen AI Max)
> Kernel:
> 7.0.12-zen1-1-zen (linux-zen, Arch Linux)
> 7.0.12.arch1-1 (linux, Arch Linux, vanilla kernel.org sources)
> Controller: c5:00.4 AMD Strix Halo USB 3.1 xHCI [1022:158b]
> Driver:     xhci_hcd, CONFIG_USB_XHCI_PCI=y
> 
> 
> Symptom
> -------
> 
> During continuous full-duplex USB audio operation, round-trip latency jumps
> discretely. Kernel logs show:
> 
>    frame active: -18
>    retire_capture_urb

Not sure what you mean and I can't find these strings in USB audio
drievr, but -18 is -EXDEV i.e. Missed Service Error in xHCI.

Will you see messages about it after enabling dynamic debug?

echo 'func handle_tx_event +p' >/proc/dynamic_debug/control

> The Audient iD44 MK2 works without this issue on a Lenovo ThinkPad X1
> Gen 2 using the same Linux audio workflow. This confirms a
> host-controller-specific problem.

And since it's unique to a recent AMD system, maybe this?
https://lore.kernel.org/linux-usb/18b4e4f7089aa4f1.da8dbe994ae3bb77.445e21b98b0b205b@GordonMsi/

Regards,
Michal