GitHub ↗
← all posts
bosgamestrix-haloryzen-ai-maxlinuxusb4

Bosgame M5 / AXB35 USB-C and rear USB not working on Linux: the fix

On the Bosgame M5 (BeyondMax, board AXB35-02, Ryzen AI Max+ 395), Linux sees only the front USB-A port. Three of the board's four USB controllers sleep in D3cold and never wake, so a drive in a USB-C or rear port produces no kernel event at all. A one-line udev rule keeps the controllers awake. With it, every external port worked in a plug test on the hardware on 14 November 2026 (BIOS 1.10).

Why don't the USB-C and rear USB ports work on Linux?#

The board has four xHCI USB controllers (PCI ids 1022:1587, 1588, 1589, 158b). On a stock Linux boot, three of them autosuspend to D3cold and stay there. A port whose controller is asleep does not report a plug, so the symptom looks electrical: no dmesg line, no device, and the port registers stay in RxDetect. The one controller that stays awake carries the front USB-A port, which is why that port works and nothing else does. The original report says every port works under Windows on the same box (engine#823); we have not tested Windows ourselves.

It is not a missing driver, and it is not a Type-C or UCSI problem. Writing on to each controller's power/control moves it to D0, and its ports come alive.

How do I fix it?#

Install this rule as /etc/udev/rules.d/91-1bit-xhci-keepawake.rules (one line):

ACTION=="add|bind", SUBSYSTEM=="pci", DRIVER=="xhci_hcd", ATTR{[dmi/id]board_name}=="AXB35-02", ATTR{vendor}=="0x1022", ATTR{device}=="0x1587|0x1588|0x1589|0x158b", TEST=="power/control", ATTR{power/control}="on"

Then reboot. The rule only binds on a board whose DMI name is AXB35-02, and it only writes power/control = on on those four controllers, so it does nothing on any other machine.

1bit OS ships this rule in its image. The same file is in the engine repository.

How do I check that it worked?#

The engine repository has a read-only check, tools/xhci_keepawake_check.sh:

tools/xhci_keepawake_check.sh                    # DMI match, each controller's power state, a verdict
tools/xhci_keepawake_check.sh --require-awake    # exit 1 unless every matched controller is in D0
sudo tools/xhci_keepawake_check.sh --apply       # the rule's one write, without a reboot (--revert undoes it)

The verdict all awake means a plug into any port is a valid test. N asleep means the rule is not installed or did not apply yet.

Which port is on which controller?#

Measured on the Bosgame M5 (AXB35-02), BIOS 1.10, kernel 7.4.0-rc1-1bit, with the four controllers kept awake:

physical port xHCI controller result
front USB-A c7:00.0 (1588) works with or without the rule
front USB-C c7:00.3 (1589) SuperSpeed 5 Gbps, USB storage
rear USB-C c7:00.4 (158b) SuperSpeed 5 Gbps, USB storage
rear USB-A c7:00.0 (1588), port 4 high-speed with a USB 2 stick (a USB 3 device not yet tried)
rear USB-A next to the power connector c5:00.4 (1587) SuperSpeed 5 Gbps, USB storage

One enclosure we tried, a Realtek RTL9210, enumerates at USB 2 speed on every port, including ports where another drive reaches 5 Gbps. That limit is the enclosure or its cable, not the board.

Does a BIOS update fix it?#

No. The same three controllers sleep in D3cold on BIOS 1.09 and on BIOS 1.10 (11/04/2026). The rule is needed on both.

What about USB4 and Thunderbolt?#

The Bosgame M5 uses the APU's native USB4: USB4 v1 at 40 Gbps, through two USB4 host routers. There is no discrete USB4 v2 (80 Gbps) controller on this board. USB 2 and USB 3 devices on the USB-C ports go through the xHCI controllers above, which is what the rule fixes. A USB4 or Thunderbolt device has not been tested yet, and neither has a 1bit OS boot with the rule loaded from the image (the plug test above ran on an Ubuntu host with the rule installed).

The full record, with the per-boot captures and the measurements behind each step, is on AXB35 USB ports. For booting 1bit OS from a stick on this board, see Install. The issue is engine#823. More from this week is in This week in the engine.

Frequently asked questions#

Why does only the front USB-A port work on the Bosgame M5 under Linux?#

Three of its four xHCI USB controllers autosuspend to D3cold and never wake. Only the controller behind the front USB-A port stays awake.

Is there a fix for the Bosgame M5 USB-C ports on Linux?#

Yes. A one-line udev rule sets power/control to on for the board's four xHCI controllers. With it, every external USB-A and USB-C port worked in a plug test on BIOS 1.10.

Does the udev rule affect other computers?#

No. It only binds when the DMI board name is AXB35-02, and only on the four AMD xHCI controller ids it names.

Does the Bosgame M5 have USB4 v2 at 80 Gbps?#

No. It uses the Ryzen AI Max+ 395's native USB4 v1 at 40 Gbps. USB4 and Thunderbolt devices have not been tested with the fix yet.