Re: [REGRESSION] USB-A devices not working on boot after recent USB merge

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

On Wed, 8 Oct 2025 14:16:43 +0200, Arisa Snowbell wrote:
> I am back with the results, kinda struggled honestly.
> 
> So I tried finding the PCIe address from the good kernel and that
> didn't work out, it disappeared and on the good kernel I have 5x 2.0
> root hubs and on the bad one I have 4 of them.

Looks like xhci_hcd fails to register the USB2-only bus at all.
You probably have some noise about it in the log, try

dmesg |grep xhci_hcd

shortly after booting and you should see stuff like:

[ 2.843799] xhci_hcd 0000:0d:00.0: xHCI Host Controller
[ 2.843811] xhci_hcd 0000:0d:00.0: new USB bus registered, assigned bus number 1
[ 2.899808] xhci_hcd 0000:0d:00.0: hcc params 0x0200ef81 hci version 0x110 quirks 0x0000000000000010
[ 2.900068] xhci_hcd 0000:0d:00.0: xHCI Host Controller
[ 2.900070] xhci_hcd 0000:0d:00.0: new USB bus registered, assigned bus number 2
[ 2.900072] xhci_hcd 0000:0d:00.0: Host supports USB 3.2 Enhanced SuperSpeed

followed by a bunch of "new USB device", repeating several times with
different PCI addresses. One of them will be the broken controller not
showing in lsusb and maybe something interesting will be printed there.

Once you know its address you can also try to bind the driver again
with dynamic debug enabled, which will print a lot more info.