Help Needed: USB 3.0 Ports Dropping Connection on Gigabyte X670 AORUS ELITE AX

[MightyRufo]

Hello Unraid Community,

 

I'm experiencing an issue with the USB 3.0 ports on my Gigabyte X670 AORUS ELITE AX motherboard and seeking your expertise. The ports in question are the two USB 3.0 ports located at the back, just below the USB 2.0 ports closest to the exhaust fan. This issue has led to intermittent disconnections of connected devices, affecting only the USB 3.0 ports currently, although I've noticed similar behavior with the USB 2.0 ports in the past. This is indeed a new build that has suffered from this issue since day one. (About 4 months ago)

 

usb 9-1-port1: disabled by hub (EMI?), re-enabling...

and

xhci_hcd 0000:18:00.3: xHCI host controller not responding, assume dead

xhci_hcd 0000:18:00.3: HC died; cleaning up

 

System Details:

• Motherboard: Gigabyte X670 AORUS ELITE AX, BIOS Version F22b (Dated 02/06/2024)
• CPU: AMD Ryzen 9 7950X 16-Core @ 4500 MHz
• RAM: Corsair Dual Channel 64 GiB DDR5
• OS: Unraid 6.12.8, Kernel Linux 6.1.74-Unraid x86_64
• Others: C-States Enabled, HVM Enabled, IOMMU Enabled, Network: bond0 fault-tolerance (active-backup), MTU 1500

 

Troubleshooting Steps Taken:

1. Isolated Devices: Disconnected USB devices one at a time to identify if any particular device was causing the issue. The problem persisted, eliminating device-specific faults.
2. BIOS Update: Updated the BIOS to the latest version (F22b). However, it's worth noting that the F22b version does not introduce any changes from the F22 version, as confirmed by Gigabyte.

 

The persistent nature of this issue across different types of USB ports suggests a more systemic problem, possibly related to the motherboard's USB controller or its interaction with Unraid.

 

I'm attaching my diagnostics file for a more detailed analysis but would appreciate any insights or recommendations from the community. Could this be a known issue with a potential workaround, or are there any settings within Unraid or the BIOS that I might have overlooked?

Thank you in advance for your help and support.

jasmasiserver-diagnostics-20240329-0022.zip

Edited March 31 by MightyRufo

[JorgeB]

Mar 28 18:08:39 JasmasiServer kernel: xhci_hcd 0000:18:00.3: xHCI host not responding to stop endpoint command
Mar 28 18:08:39 JasmasiServer kernel: xhci_hcd 0000:18:00.3: xHCI host controller not responding, assume dead
Mar 28 18:08:39 JasmasiServer kernel: xhci_hcd 0000:18:00.3: HC died; cleaning up

 

This looks more like a hardware issue, or at best some kernel compatibility issue, if you have already updated the BIOS and the issue remains, suggest re-trying with Unraid 6.13-beta once it's out, which should be soon, it uses a much more recent kernel, but doubt it will help, since many other users have boards with the same chipset and no issues, and if that doesn't help I'm afraid you will need to use a different board.

 

 

[MightyRufo]

I was afraid you'd say that. The interesting part is that the last time it happened. There was two USB ports on my motherboard that simply wouldn't respond to anything (not even power I believe) I also couldn't find any reports with x670.

 

If I do try the beta, can I downgrade if it causes other issues? Obviously I will make a backup of my flash.

[JorgeB]

1 hour ago, MightyRufo said:

If I do try the beta, can I downgrade if it causes other issues?

Yep, just restore the backup.

[MightyRufo]

I just encountered another error. I lost network connectivity with the server.. However the server still had a connection to the internet (I received notifications from a container via pushover having trouble connecting to another). The log shows as follows

Apr  5 19:29:21 JasmasiServer kernel: r8169 0000:10:00.0 eth0: Link is Up - 10Mbps/Full - flow control off
Apr  5 19:29:21 JasmasiServer kernel: bond0: (slave eth0): link status definitely up, 10 Mbps full duplex
Apr  5 19:29:21 JasmasiServer kernel: bond0: (slave eth0): making interface the new active one
Apr  5 19:29:21 JasmasiServer kernel: device eth0 entered promiscuous mode
Apr  5 19:29:21 JasmasiServer kernel: bond0: active interface up!
Apr  5 19:29:21 JasmasiServer kernel: br0: port 1(bond0) entered blocking state
Apr  5 19:29:21 JasmasiServer kernel: br0: port 1(bond0) entered forwarding state
Apr  5 19:29:22 JasmasiServer kernel: r8169 0000:10:00.0 eth0: Link is Down
Apr  5 19:29:22 JasmasiServer kernel: bond0: (slave eth0): link status definitely down, disabling slave
Apr  5 19:29:22 JasmasiServer kernel: device eth0 left promiscuous mode
Apr  5 19:29:22 JasmasiServer kernel: bond0: now running without any active interface!
Apr  5 19:29:22 JasmasiServer kernel: br0: port 1(bond0) entered disabled state
Apr  5 19:29:25 JasmasiServer kernel: r8169 0000:10:00.0 eth0: Link is Up - 1Gbps/Full - flow control off
Apr  5 19:29:25 JasmasiServer kernel: bond0: (slave eth0): link status definitely up, 1000 Mbps full duplex
Apr  5 19:29:25 JasmasiServer kernel: bond0: (slave eth0): making interface the new active one
Apr  5 19:29:25 JasmasiServer kernel: device eth0 entered promiscuous mode
Apr  5 19:29:25 JasmasiServer kernel: bond0: active interface up!
This continues to repeat. the log is just before the server lost network connectivity.

I think the log shows repeated fluctuations in the link status of my ethernet adaptor, until it gives up, I assume. Restarting the server restores it to normal working order. This was after almost 5 days of uptime. Even though I was able to press the power button on the server and it did shut down, I still got an unclean shutdown notification from unraid. No "shutdown information" was found in the log either. 

I think this also seems like a hardware issue? I'm not sure.. I wasn't able to connect to the webui or else I would of done a diagnostics. Let me know if there's any further information I can provide.

Edited April 6 by MightyRufo

[JorgeB]

9 hours ago, MightyRufo said:

10Mbps/Full

This could be a NIC problem, but can also be cable/switch, try a different cable and use another switch, or at least a different switch port, if it keeps happening it could be the NIC.

[MightyRufo]

14 hours ago, JorgeB said:

This could be a NIC problem, but can also be cable/switch, try a different cable and use another switch, or at least a different switch port, if it keeps happening it could be the NIC.

This issue arised from day 1 of building this server. The previous one was fine.. it has to be the motherboard. That really sucks.

I noticed the server was offline again.. this time it would NOT respond to me pressing the power button. I am having call traces.

Apr 6 02:20:34 JasmasiServer kernel: invalid opcode: 0000 [#1] PREEMPT SMP NOPTI Apr 6 02:20:34

JasmasiServer kernel: CPU: 19 PID: 16015 Comm: tdarr-ffmpeg Tainted: P O 6.1.79-Unraid #1 Apr 6 02:20:34

JasmasiServer kernel: Hardware name: Gigabyte Technology Co., Ltd. X670 AORUS ELITE AX/X670 AORUS ELITE AX, BIOS F22b 02/06/2024 Apr 6 02:20:34

JasmasiServer kernel: RIP: 0010:plist_add+0x9c/0x9e Apr 6 02:20:34 J

asmasiServer kernel: Code: 42 10 48 83 c2 08 48 89 50 08 48 89 48 10 4c 89 01 4c 89 ca 48 8b 4a 08 48 89 7a 08 48 89 50 18 48 89 48 20 48 89 39 e9 ca 08 <3b> 00 48 8b 47 08 48 8d 57 08 48 39 c2 74 41 48 8b 47 18 48 39 c6 Apr 6 02:20:34

JasmasiServer kernel: RSP: 0018:ffffc9002f72fd88 EFLAGS: 00010246 Apr 6 02:20:34

JasmasiServer kernel: RAX: ffffc9002f72fe20 RBX: ffffc9002f72fe20 RCX: ffff8881013ba908 Apr 6 02:20:34

JasmasiServer kernel: RDX: ffff8881013ba908 RSI: ffff8881013ba908 RDI: ffffc9002f72fe38 Apr 6 02:20:34

JasmasiServer kernel: RBP: ffff888763dcac80 R08: ffffc9002f72fe28 R09: ffff8881013ba908 Apr 6 02:20:34

JasmasiServer kernel: R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000 Apr 6 02:20:34

JasmasiServer kernel: R13: 0000000000000000 R14: ffff888763dcac80 R15: 000056254857e964 Apr 6 02:20:34

JasmasiServer kernel: FS: 0000150ee61f0700(0000) GS:ffff888ffe2c0000(0000) knlGS:0000000000000000 Apr 6 02:20:34

JasmasiServer kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 Apr 6 02:20:34

JasmasiServer kernel: CR2: 0000152bdedb2d20 CR3: 00000002dcb2c000 CR4: 0000000000750ee0 Apr 6 02:20:34

JasmasiServer kernel: PKRU: 55555554 Apr 6 02:20:34 JasmasiServer kernel: Call Trace:

And it continues.

I'm gonna run windows server and see if I encounter the same issues.

Edited April 7 by MightyRufo