[6.6.0-rc1] xHCI dead, incomplete boot Threadripper 1950

Diagnostics [from previous version 6.5.3] attached.

tower-diagnostics-20180904-1836.zip

Can you please try upgrading to BIOS 1402 (released on 8/10/18) and retrying Unraid 6.6.0-rc1?

Having array drives attached via USB is not ideal. It will work when everything is healthy, but error handling and recovery is sub par, and smart reporting can be problematic.

43 minutes ago, eschultz said:

Can you please try upgrading to BIOS 1402 (released on 8/10/18) and retrying Unraid 6.6.0-rc1?

Yessir, am working on that now.  Will follow up after that's done.  Actually thought I had this update but apparently downloaded it and didn't apply it.

20 minutes ago, jonathanm said:

Having array drives attached via USB is not ideal. It will work when everything is healthy, but error handling and recovery is sub par, and smart reporting can be problematic.

Thank you for your concern, but I'm well aware of this.  There is no way to fit them inside the chassis with the other hardware in there.  I actually was going to mention this in the original post, but figured it didn't move the conversation forward or have anything to do with the bug.  My apologies for the omission.

I was able to get a little farther this time, in that the front panel USB worked, meaning I had a functional keyboard.  I was able to get a USB key to mount, and copied off the contents of the /var/log directory, figuring that'd be the most useful.

 

I then tried booting with the UNRAID drive in this port, but unfortunately that failed in the exact same manner.  I've PM'd the full zip to eschultz and am posting the dmesg output here for all eyes.

dmesg.txt

XHCI should work but failed at device side, XHCI driver update issue in 6.6rc1 ?

 

[   30.917042] usb usb6-port4: attempt power cycle
[   32.189079] usb 6-1: new SuperSpeed USB device number 5 using xhci_hcd
[   32.202160] usb-storage 6-1:1.0: USB Mass Storage device detected
[   32.202696] scsi host3: usb-storage 6-1:1.0
 

 

[  102.443495] usb 1-5: USB disconnect, device number 2
[  109.781040] usb 1-5: new high-speed USB device number 5 using xhci_hcd
[  110.008870] usb-storage 1-5:1.0: USB Mass Storage device detected
 

01:00.0 USB controller [0c03]: Advanced Micro Devices, Inc. [AMD] X399 Series Chipset USB 3.1 xHCI Controller [1022:43ba] (rev 02)
    Subsystem: ASMedia Technology Inc. Device [1b21:1142]
    Kernel driver in use: xhci_hcd

 

06:00.0 USB controller [0c03]: ASMedia Technology Inc. Device [1b21:2142]
    Subsystem: ASUSTeK Computer Inc. Device [1043:8756]
    Kernel driver in use: xhci_hcd

 

You have several XHCI controller in your system, if I am correct TR / Ryzen should have a USB come from CPU ( but MFG may not use it ). Does all have same problem ? Suggest try differnet USB port too.

Edited September 5, 2018 by Benson

Since you have a Zenith Extreme we have the same board.. Mine works yours doesn't so lets do a few experiments. Do you have a chasis usb3.0 cable that can connect the onboard USB3.0 boards to external ports. In your manual #16 for motherboard list. If you do then plug the usb drive for unraid into a usb plug off of that. Further if your chasis has usb2.0 ports you can use header #26 like I did.

9 hours ago, Benson said:

01:00.0 USB controller [0c03]: Advanced Micro Devices, Inc. [AMD] X399 Series Chipset USB 3.1 xHCI Controller [1022:43ba] (rev 02)

    Subsystem: ASMedia Technology Inc. Device [1b21:1142]
    Kernel driver in use: xhci_hcd

 

06:00.0 USB controller [0c03]: ASMedia Technology Inc. Device [1b21:2142]
    Subsystem: ASUSTeK Computer Inc. Device [1043:8756]
    Kernel driver in use: xhci_hcd

 

You have several XHCI controller in your system, if I am correct TR / Ryzen should have a USB come from CPU ( but MFG may not use it ). Does all have same problem ? Suggest try differnet USB port too.

The ASMedia USB header requires a specific plug that I don't have available, so it is enabled but not currently used.

The rest of the USBs (minus my Sonnet USB card, which is for VM use only and isolated at boot) I believe stem from the AMD chipset.  All available USBs seem to exhibit the problem.

 

 

2 hours ago, Jerky_san said:

Since you have a Zenith Extreme we have the same board.. Mine works yours doesn't so lets do a few experiments. Do you have a chasis usb3.0 cable that can connect the onboard USB3.0 boards to external ports. In your manual #16 for motherboard list. If you do then plug the usb drive for unraid into a usb plug off of that. Further if your chasis has usb2.0 ports you can use header #26 like I did.

In my previous reply, I mention that the front panel USB works for M/KB with new bios revision.  While this isn't terribly obvious (sorry about that), front panel connections are always via MB header.  While I haven't attempted to use my front panel USB 2.0 ports to boot from, I'd argue that if that does in fact solve the problem it's not a solution, as if you don't have USB 2.0 header ports available it won't work, and it's probably a symptom of a bigger problem.

 

Also, what TR do you have and what board revision?  Use:

 dmidecode --type 2

And you should get back something like mine:

Manufacturer: ASUSTeK COMPUTER INC.
        Product Name: ROG ZENITH EXTREME
        Version: Rev 1.xx
        Serial Number: 170706217200585

 

Edited September 5, 2018 by Rhynri

I have a 2990WX

 

Base Board Information
        Manufacturer: ASUSTeK COMPUTER INC.
        Product Name: ROG ZENITH EXTREME
        Version: Rev 1.xx

 

Only BIOS changes made are I turned on " Enumerate all iommu in IVRs " and SVM.. Also I used to us my USB 3.0 ports I had to stop as when I passed my controller it passed two USB controllers instead of one.. The onboard one and the 3.1 in the back are the only two that don't pass when I do the passing command even though I only told Unraid to grab 1 of the IOMMUs groups and not both. Was a very strange thing. Also if I turned off my wireless/bluetooth that USB controller no longer passes.

I have a 1950.  I’m using all the back USB ports for the Unraid OS itself and then I pass individual controllers off a Sonnet card for the VMs.  I have the wireless and Bluetooth disabled because it cleans up the pass through for the rest, although I’d love for unraid to be able to use the wireless for additional network redundancy.  

 

In the bios I have NUMA set up and then additional tweaks to the pcie setup because I’m splitting the bottom slot between the sonnet card and a U.2. 

Edited September 5, 2018 by Rhynri

Changed Status to Retest

Hello!  The new RC2 now boots correctly.  Thanks!  I've run into some other weirdness but it's not the same issue so I've closed this one.

Edited September 8, 2018 by Rhynri