USB hardware issues and a kernel panic

TomHu · 2026-02-12 14:30:44

Hi, since the last kernel update (6.18.9.arch1-2) I experience seemingly hardware related issues, that I am unable to follow up on. I am looking for directions, really.

The system is an AMD Ryzen 5 8500G on a Gigabyte B650I AX. Turned off EXPO settings; no overclocking. The temperatures seem fine. There’s 1 ram module installed, Crucial CT32G52C42U5.M16GB.

This is an issue that recurs during audio playback. It takes a few minutes, and the USB DAC stops and produces this error. I added the kernel parameter `iommu=pt`, to no avail.

Feb 12 15:01:26 gulliverkli kernel: xhci_hcd 0000:0f:00.0: ERROR Transfer event for disabled endpoint slot 8 ep 14
Feb 12 15:01:26 gulliverkli kernel: xhci_hcd 0000:0f:00.0: @000000010e384b50 16306730 00320001 011a0000 080f8001
Feb 12 15:01:26 gulliverkli kernel: xhci_hcd 0000:0f:00.0: ERROR Transfer event for disabled endpoint slot 8 ep 7
Feb 12 15:01:26 gulliverkli kernel: xhci_hcd 0000:0f:00.0: @000000010e384b60 16306740 00320001 011d0000 08088001
Feb 12 15:01:26 gulliverkli kernel: xhci_hcd 0000:0f:00.0: Event dma 0x0000000116156750 for ep 1 status 1 not part of TD at 0000000116156730 - 0000000116156730
repeats…

sensors output:

gigabyte_wmi-virtual-0
Adapter: Virtual device
temp1:        +49.0°C  
temp2:        +64.0°C  
temp3:        +53.0°C  
temp4:        +58.0°C  
temp5:        +50.0°C  
temp6:        +49.0°C

Also today I had a kernel panic. But I don’t understand the message.
panic.archlinux.org…

OS: Arch Linux x86_64
Host: B650I AX
BIOS (UEFI): F37 (5.35)
Bootmgr: Arch iommu=pt - vmlinuz-linux
Board: B650I AX (x.x)
Chassis: Desktop
Kernel: Linux 6.18.9-arch1-2
Init System: systemd 259.1-1-arch
Uptime: 1 hour, 29 mins
Loadavg: 0.48, 0.36, 0.40
Processes: 383
Packages: 1781 (pacman)
Shell: bash 5.3.9
Editor: vim 9.1
Display (U24E590): 3840x2160 in 23", 60 Hz [External]
Monitor (U24E590): 3840x2160 px @ 60 Hz - 520x290 mm (23.44 inches, 187.95 ppi)
LM: greetd (Wayland)
WM: Sway 1.11 (Wayland)
Theme: Fusion [Qt], Breeze [GTK2]
Icons: oxygen [Qt], breeze-dark [GTK2]
Font: Sans Serif (10pt, Regular) [Qt], Fira Sans (9pt) [GTK2]
Cursor: capitaine (48px)
Terminal: kitty 0.45.0
Terminal Font: FiraCodeNF-Reg (20pt)
Terminal Size: 117 columns x 65 rows (1872px x 2145px)
Terminal Theme: #EBDBB2 (FG) - #272727 (BG) [Dark]
CPU: AMD Ryzen 5 8500G (12) @ 5.08 GHz - 50.5°C
CPU Cache (L1): 6x32.00 KiB (D), 6x32.00 KiB (I)
CPU Cache (L2): 6x1.00 MiB (U)
CPU Cache (L3): 16.00 MiB (U)
CPU Usage: 0%
GPU: AMD Radeon 740M Graphics (4) @ 2.80 GHz - 50.0°C (1.79 GiB / 3.85 GiB, 46%) [Integrated]
Memory: 3.19 GiB / 27.02 GiB (12%)
Disk (/): 480.47 GiB / 915.31 GiB (52%) - ext4
Locale: de_DE.UTF-8
Vulkan: 1.4.328 - radv [Mesa 25.3.5-arch1.1]
OpenGL: 4.6 (Compatibility Profile) Mesa 25.3.5-arch1.1
OpenCL: 2.1 AMD-APP.dbg (3581.0)
Mouse 1: Logitech ERGO K860
Mouse 2: Logitech MX Master 3
Mouse 3: Microsoft Microsoft® 2.4GHz Transceiver v9.0 Mouse
Mouse 4: Logitech MX Keys
Keyboard 1: Microsoft Microsoft® 2.4GHz Transceiver v9.0 Consumer Control
Keyboard 2: Microsoft Microsoft® 2.4GHz Transceiver v9.0
Network IO (wlan0): 917 B/s (IN) - 450 B/s (OUT)
Disk IO (WD Blue SN580 1TB): 0 B/s (R) - 0 B/s (W)
Physical Disk (WD Blue SN580 1TB): 931.51 GiB [SSD, Fixed] - 59.9°C
TPM: TPM 2.0 Device
Version: fastfetch 2.58.0 (x86_64)

seth · 2026-02-12 15:54:50

[11151.164475] Oops: general protection fault, probably for non-canonical address 0xff088cd5c11f7cd0: 0000 [#1] SMP NOPTI
[11151.164484] CPU: 10 UID: 0 PID: 0 Comm: swapper/10 Not tainted 6.18.9-arch1-2 #1 PREEMPT(full)  acc0345693780c55f83aaf9a0c9efd10d8da358f

What does your swap config look like?
zswap? zram? swap partition or file?
Do not combine zswap and zram and zswap is designed to augment a physical swap device (partition or file doesn't matter)

CPU: AMD Ryzen 5 8500G (12) @ 5.08 GHz - 50.5°C
[11151.164603]  ? cpuidle_enter_state+0xac/0x410
[11151.164607]  cpuidle_enter+0x31/0x50
[11151.164612]  do_idle+0x1ae/0x210
[11151.164618]  cpu_startup_entry+0x29/0x30

https://wiki.archlinux.org/title/Ryzen#Troubleshooting - try to limit the c-states

Feb 12 15:01:26 gulliverkli kernel: xhci_hcd 0000:0f:00.0: ERROR Transfer event for disabled endpoint slot 8 ep 14

Please provide more context, in doubt the system journal after such event.

sudo journalctl -b | curl -F 'file=@-' 0x0.st

TomHu · 2026-02-12 17:08:45

I have no swap config. The zram module isnt loaded. `/proc/swaps` is empty. There’s only `systemd-tmpfiles-setup.service`.

Here’s the journalctl output of `-b -1`. That boot just panic’ed too.
https://0x0.st/PmZg.txt

This boot message seems relevant; I don’t understand it though.

kernel: ACPI Warning: SystemIO range 0x0000000000000B00-0x0000000000000B08 conflicts with OpRegion 0x0000000000000B00-0x0000000000000B0F (\GSA1.SMBI) (20250807/utaddress-204)

I’ve found a BIOS setting: “DDR5 Auto Booster”, which was set to “Auto”. I then set it to disabled. I don’t understand it though.

EDIT: `systemd-coredump[20120]: [?] Process 20071 (WebExtensions) of user 1000 dumped core.`

There are tiny graphics glitches.

EDIT2: added `processor.max_cstate=5`

Last edited by TomHu (2026-02-12 18:15:01)

seth · 2026-02-12 21:04:12

Feb 12 13:56:56 archlinux kernel: zswap: loaded using pool zstd

https://wiki.archlinux.org/title/Zswap#Toggling_zswap

The xhci errors come out of nowhere but seem to be one of

Feb 12 13:56:59 archlinux kernel: input: Microsoft Microsoft® 2.4GHz Transceiver v9.0 as /devices/pci0000:00/0000:00:02.1/0000:02:00.0/0000:03:0c.0/0000:0f:00.0/usb1/1-7/1-7:1.0/0003:045E:07A5.0008/input/input26
Feb 12 13:56:59 archlinux kernel: hid-generic 0003:045E:07A5.0008: input,hidraw5: USB HID v1.11 Keyboard [Microsoft Microsoft® 2.4GHz Transceiver v9.0] on usb-0000:0f:00.0-7/input0
Feb 12 13:56:59 archlinux kernel: input: Microsoft Microsoft® 2.4GHz Transceiver v9.0 Mouse as /devices/pci0000:00/0000:00:02.1/0000:02:00.0/0000:03:0c.0/0000:0f:00.0/usb1/1-7/1-7:1.1/0003:045E:07A5.0009/input/input27
Feb 12 13:56:59 archlinux kernel: input: Microsoft Microsoft® 2.4GHz Transceiver v9.0 Consumer Control as /devices/pci0000:00/0000:00:02.1/0000:02:00.0/0000:03:0c.0/0000:0f:00.0/usb1/1-7/1-7:1.1/0003:045E:07A5.0009/input/input28
Feb 12 13:56:59 archlinux kernel: hid-generic 0003:045E:07A5.0009: input,hidraw6: USB HID v1.11 Mouse [Microsoft Microsoft® 2.4GHz Transceiver v9.0] on usb-0000:0f:00.0-7/input1
Feb 12 13:56:59 archlinux kernel: input: Microsoft Microsoft® 2.4GHz Transceiver v9.0 Consumer Control as /devices/pci0000:00/0000:00:02.1/0000:02:00.0/0000:03:0c.0/0000:0f:00.0/usb1/1-7/1-7:1.2/0003:045E:07A5.000A/input/input29
Feb 12 13:56:59 archlinux kernel: input: Microsoft Microsoft® 2.4GHz Transceiver v9.0 System Control as /devices/pci0000:00/0000:00:02.1/0000:02:00.0/0000:03:0c.0/0000:0f:00.0/usb1/1-7/1-7:1.2/0003:045E:07A5.000A/input/input30
Feb 12 13:56:59 archlinux kernel: hid-generic 0003:045E:07A5.000A: input,hidraw7: USB HID v1.11 Device [Microsoft Microsoft® 2.4GHz Transceiver v9.0] on usb-0000:0f:00.0-7/input2
Feb 12 13:57:00 archlinux kernel: input: Yubico YubiKey OTP+FIDO+CCID as /devices/pci0000:00/0000:00:02.1/0000:02:00.0/0000:03:0c.0/0000:0f:00.0/usb1/1-8/1-8:1.0/0003:1050:0407.000B/input/input31
Feb 12 13:57:00 archlinux kernel: hid-generic 0003:1050:0407.000B: input,hidraw8: USB HID v1.10 Keyboard [Yubico YubiKey OTP+FIDO+CCID] on usb-0000:0f:00.0-8/input0
Feb 12 13:57:00 archlinux kernel: hid-generic 0003:1050:0407.000C: hiddev97,hidraw9: USB HID v1.10 Device [Yubico YubiKey OTP+FIDO+CCID] on usb-0000:0f:00.0-8/input1
Feb 12 13:57:00 archlinux kernel: hid-generic 0003:048D:5702.000D: hiddev98,hidraw10: USB HID v1.12 Device [ITE Tech. Inc. ITE Device] on usb-0000:0f:00.0-11/input0

Most likely the mickeysoft dongle, possibly power saving related?

There are tiny graphics glitches.

Of what exact nature?

mmy8x · 2026-02-12 23:14:03

Does it mean some earlier kernel used to run the same software on the same hardware without issue?
Which version it was, what happens if you restore the previous kernel packages?

The xhci stuff looks like a bug in the driver. If not that then some serious data corruption or CPU bug.

edit:
On second thought, it may be data corruption. Those invalid events have pointers ending with 6730 and 6740 and the next event ends with 6750 while the driver still waits for a pointer ending in 6730. Chances are that the invalid events were mangled, i.e. wrong endpoint/slot IDs and also partly wrong pointer, and that's why the driver missed them.

So RAM went bad or the new kernel triggered a CPU bug. You could try other RAM (if available) or downgrading.
I would be cautious with using a machine in such state because getting corrupted data written do disk may be not fun.

Last edited by mmy8x (2026-02-12 23:31:38)

TomHu · 2026-02-21 16:54:09

Hi both of you,

a week has gone by, and I have been using `appstresstest` for hours. Not one memory problem reported though.

Removed the Microsoft dongle, replaced for Logitech. The Logitech dongle doesn’t report anything – excluding a known (fixed) mouse bug, which was the initial reason I switched to the Microsoft mouse:
https://bbs.archlinux.org/viewtopic.php?id=285653

There have been applications crashing and system freezes, now, but it’s been hard to provoke them. Running audio through the usb DAC seems fine. I had the odd impression that the crashes may be provoked by GPU load. For example:

browser tab crashes with video-ads running
sdl-freerdp sessions with the screen content being hardware decoded h264
the browser crashes are foreboded by graphics glitches, where medium sized, rectangular areas of screen are disturbed by what seems to be pixel static for a fraction of a second, hardly visible

This is the journal message from a firefox tab crash. Otherwise there is nothing.

Feb 21 14:01:26  kernel: traps: Isolated Web Co[4544] general protection fault ip:7fd60166fc7e sp:7ffc790>
Feb 21 14:01:26  systemd-coredump[5392]: Process 4544 (Isolated Web Co) of user 1000 terminated abnormall>

I attribute the improvement to a BIOS setting for something called `CPU Voltage Load Line Line` to `high` (I have no clue how much that changes voltage). Also dropping the ram frequency from 5200Mhz (default) to 4800Mhz made the glitches occur less frequently. I have the odd feeling, there’s something amiss with the CPU, either the memory controller oder some graphics portion. But before it seemed USB related.

I’ll downgrade the kernel this weekend and report.

Current kernel cmdline:

initrd=\amd-ucode.img initrd=\initramfs-linux.img rd.luks.name=[…]=ROOT rd.luks.options=[…] root=UUID=[…] iommu=pt processor.max_cstate=5 acpi_enforce_resources=lax rw

I have no spare CPU nor memory to test… guess I dont need to tell you about ram prices…

EDIT:

Yet another segfault:
https://0x0.st/s/XxJIHm_cGrmXvVKxvBtBDg/PS9k.txt

Last edited by TomHu (2026-02-21 21:06:36)

seth · 2026-02-21 21:39:03

a) are the xhci errors from the OP still there?
b) have you disabled zswap or provided it w/ a physical swap device (swap file)?
c) do you have the full backtrace for the firefox crash(es)?
https://wiki.archlinux.org/title/Core_d … _core_dump

TomHu · 2026-02-22 18:34:50

a) Just now an xhci error returned while playing music with the usb dac: https://0x0.st/s/GhLfDPNMpH5iW8EzJgqgVA/PSYQ.txt
b) disabled via cmdline
c) yes, here’s the info, the dump itself is ~80MiB https://0x0.st/s/8UVDVEVU1I0nrohAT3zCPg/PSY2.txt

seth · 2026-02-22 21:26:17

Firefox crashes somewhere deep in XUL - if this is frequent enough, does it happen w/ https://wiki.archlinux.org/title/Firefo … shoot_Mode (resp. all extensions disabled)?

Did the xhci error still go along

takes a few minutes, and the USB DAC stops

or kernel panics?
Does disabling https://wiki.archlinux.org/title/Power_ … utosuspend help?

TomHu · 2026-02-22 23:07:08

I’ll try w/o addons, but 20 minutes after I rebooted with `usbcore.autosuspend=-1` kernel parameter I got a wayland/sway crash shortly after the graphics glitches were visible.

MCE: Killing WaylandProxy:2704 due to hardware memory corruption fault at 7fa8c605f008

http://0x0.st/s/sP9dptWVOlyK2uMBeDOKgQ/PS6j.txt

Now there’s a defunct firefox process zoo dangling around.

Last edited by TomHu (2026-02-22 23:08:28)

seth · Yesterday 08:32:18

This is after complete failure of the AMD gpu, please post the complete journal of the boot.

TomHu · Yesterday 11:03:04

At 00:37h the system shutdown logs alot of "stuff".

https://0x0.st/s/UR1xpBAOw2QxzvxfvpLMHQ/PS0t.txt

seth · Yesterday 13:36:42

That's one way to put it

Warning to others, the journal is fairly sizable, w/ the last 100.000 lines or so being

Feb 23 00:37:09 gulliverkli kernel: RIP: 0010:__list_del_entry_valid_or_report+0xfa/0x10a

related kernel warnings because of

Feb 23 00:37:09 gulliverkli kernel:  non-paged memory

Safe to say that at least after

Feb 22 23:35:10 gulliverkli kernel: amdgpu 0000:11:00.0: amdgpu: [mmhub] page fault (src_id:0 ring:24 vmid:15 pasid:32780)
Feb 22 23:35:10 gulliverkli kernel: amdgpu 0000:11:00.0: amdgpu:  Process RDD Process pid 2850 thread browser 4 :cs0 pid 3104
Feb 22 23:35:10 gulliverkli kernel: amdgpu 0000:11:00.0: amdgpu:   in page starting at address 0x0000800100bf5000 from client 18
Feb 22 23:35:10 gulliverkli kernel: amdgpu 0000:11:00.0: amdgpu: MMVM_L2_PROTECTION_FAULT_STATUS:0x00F03830
Feb 22 23:35:10 gulliverkli kernel: amdgpu 0000:11:00.0: amdgpu:          Faulty UTCL2 client ID: VCNU (0x1c)
Feb 22 23:35:10 gulliverkli kernel: amdgpu 0000:11:00.0: amdgpu:          MORE_FAULTS: 0x0
Feb 22 23:35:10 gulliverkli kernel: amdgpu 0000:11:00.0: amdgpu:          WALKER_ERROR: 0x0
Feb 22 23:35:10 gulliverkli kernel: amdgpu 0000:11:00.0: amdgpu:          PERMISSION_FAULTS: 0x3
Feb 22 23:35:10 gulliverkli kernel: amdgpu 0000:11:00.0: amdgpu:          MAPPING_ERROR: 0x0
Feb 22 23:35:10 gulliverkli kernel: amdgpu 0000:11:00.0: amdgpu:          RW: 0x0

the kernel memory is severely corrupted.

iommu=pt

Why?
Try

iommu=soft

instead and for good measure also rather

processor.max_cstate=1

Do you have PBO control in the UEFI?
https://wiki.archlinux.org/title/Ryzen#Troubleshooting

Make sure the RAM isn't undervolted or overclocked, no XMP or whatever performance mode is being used.

mmy8x · Yesterday 23:33:28

Does this machine pass memtester? Or mprime?
Too many different problems that nobody else is having, I really feel it's just random data corruption.

Arch Linux

#1 2026-02-12 14:30:44

USB hardware issues and a kernel panic

#2 2026-02-12 15:54:50

Re: USB hardware issues and a kernel panic

#3 2026-02-12 17:08:45

Re: USB hardware issues and a kernel panic

#4 2026-02-12 21:04:12

Re: USB hardware issues and a kernel panic

#5 2026-02-12 23:14:03

Re: USB hardware issues and a kernel panic

#6 2026-02-21 16:54:09

Re: USB hardware issues and a kernel panic

#7 2026-02-21 21:39:03

Re: USB hardware issues and a kernel panic

#8 2026-02-22 18:34:50

Re: USB hardware issues and a kernel panic

#9 2026-02-22 21:26:17

Re: USB hardware issues and a kernel panic

#10 2026-02-22 23:07:08

Re: USB hardware issues and a kernel panic

#11 Yesterday 08:32:18

Re: USB hardware issues and a kernel panic

#12 Yesterday 11:03:04

Re: USB hardware issues and a kernel panic

#13 Yesterday 13:36:42

Re: USB hardware issues and a kernel panic

#14 Yesterday 23:33:28

Re: USB hardware issues and a kernel panic

Board footer