@ncafferkey

B&C - expected at this point. It's an Ivy Bridge chipset, so the ports were all switched to xHCI now and pcixhci.device is not functional on those boards yet

A - is this a regression then for WiFi dongle? I was under impression that it worked correctly with 0331 build.

@deadwood

A: Yes, I think it's a regression since then. Would it be useful if I test the other builds you released in between? Did you manage to get a WiFi dongle yet?

C: Shouldn't the USB2-only port still work with this mode? The laptop has an EHCI controller, which I presume is attached to this port.

A: Hmm, that's weird. The only thing that should impact pciusb.device was included in 0409 build and it should only trigger with 4GB RAM or more. WiFi dongle -> not yet, focusing on pen drives for now.

C: Boards with dual host controllers (EHCI/xHCI) starting from Ivy Bridge have the ports connected to both of them. The changes I added in 0411 build now correctly take over xHCI controller from BIOS and make the xHCI take over all the ports it can (same is happening in Linux codes). This now makes these ports "dead", but once pcixhci.device start working on those boards, all the ports will work with it.

@deadwood

A: I'll double check with the 0331 build. I have 8GB RAM FWIW.

C: You're probably right, but two ports on this laptop are labelled as SS (superspeed), while the last one is labelled with a lightning bolt (and no SS), so I thought that last one might not be under the control of xHCI.

@deadwood

A: I found the problem with WiFi. This GRUB entry has debug=serial set. When I change it to just "debug", the WiFi dongle works fine.

Also, when I remove "debug" completely, I can still boot normally, i.e. the problem I had with the 2026-03-31 build of getting stuck on the media screen unless I added "debug" is resolved.

BTW, Bifteck didn't work for me after I set debug=memory.

Ok, great to hear the problem was located

Hi All,

New build and some visible progress:

https://build.axrt.org/download/builds/Experimental/exp-usb3-pc-x86_64-20260414-143634.tar.gz

In TEST B (USB3) mode:
- pendrive partition is now visible in boot menu
- booting from pendrive starts
- booting from pendrive crashes with filesystem error

This has been confirmed on folowing hardware:
- GA-H81M-S2V
- DELL OPTIPLEX 9020
- DELL LATITUDE E6430
- MSI H81M-E34

If you experience blank screen after selecting TEST B in grub, reboot, go to bootmenu, check if the DU0 partition is visible and if so, "boot without startup sequence". This should show you then the filesystem error.

@ncafferkey
Turns out E6430 has the same chipset as your DELL. What I have on that machine is a single USB2 port on the left side and two USB3 ports (with SS icon) on the right side. If I plug pendrive into USB2 port and select TEST C, then it turns out on this hardware the USB2 port is not take over by xHCI so the booting finishes and then I can use both USB3 ports for USB3 data transfers.

Hi,

Thanks, is this for Intel only USB3?

No, fixes in this build are generic and should work on any xHCI host. It's possible they will fix GET_DESCRIPTOR error on your Lenovo V155.

Hi,

Thanks. Option B, now fixed GET_DESCRIPTOR error on Lenovo V155. Able to detect usb stick, format usb and install to it. DU0: present. But boot stops and blank screen after choosing option B from usb stick.

Ah, good to hear that detection is also fixed on your hardware.