USB port issues when waking from S4 (hibernate)

Thunderbolt TBT3-UDZ
1

Same problem. It is actually all over the place - when restoring from S4 it could be just USB ports not working, could be USB ports + both display ports not working, etc.

Below is the outtake from your debug tool:

Device Name	Manufacturer	Service	Device Type Code	Device Type Name	Device Instance ID	Location	Capabilities	Config Flags	Disabled	Connected	Device Registry Time 1	Device Registry Time 2	Drive Letter	Container ID	Driver Description	Driver Version	.inf File	.inf Section	Driver Date	Driver Registry Time
Unknown USB Device (Port Reset Failed)	(Standard USB Host Controller)		USB		USB\VID_0000&PID_0001\8&29f545bd&0&1	Port_#0001.Hub_#0003	0x00000064	0x00000000	No	No	12/14/2020 20:55	12/14/2020 20:55		{ae159ecd-3dc6-11eb-98e0-813aa3b89e08}	Unknown USB Device (Port Reset Failed)	10.0.19041.488	usb.inf	BADDEVICE.Dev.NT	6/21/2006	12/14/2020 20:55

Device Name	Description	Device Type	Connected	Safe To Unplug	Disabled	USB Hub	Drive Letter	Serial Number	Created Date	Last Plug/Unplug Date	VendorID	ProductID	Firmware Revision	USB Class	USB SubClass	USB Protocol	Hub / Port	Computer Name	Vendor Name	Product Name	ParentId Prefix	Service Name	Service Description	Driver Filename	Device Class	Device Mfg	Power	USB Version	Driver Description	Driver Version	Driver InfSection	Driver InfPath	Instance ID	Capabilities
Port_#0001.Hub_#0003	Unknown USB Device (Port Reset Failed)	Unknown	No	No	No	No			12/14/2020 21:06	12/14/2020 20:55	0	1	0	0	0	0										(Standard USB Host Controller)			Unknown USB Device (Port Reset Failed)	10.0.19041.488	BADDEVICE.Dev.NT	usb.inf	USB\VID_0000&PID_0001\8&29f545bd&0&1	Removable, SilentInstall, RawDeviceOK

If you fix privacy issues with your tool, I may share the full zip, until then I can share the information of interest after I understand it’s purpose and value for troubleshooting USB ports not working.

To be honest, I don’t see this dock being stable with any laptop I tried it with - there is always a case where something doesn’t work and requires plug/unplug or reboot. Surface Pro 7, for example, plugging in “hot” (device up and running) - got USB ports working. but no screens. Had to unplug, plug into different laptop and re-plug into Surface to make it work.
I guess we can blame it on MSFT for Thunderbolt issues after restoring from S5, but there are no known issue for hibernate at this point.

Return window on Amazon ends on Jan 31, if we don’t get this resolved - I’ll probably just return until there is a more stable dock available. :frowning:

2

Thanks for contacting Plugable support. Sorry to hear about these issues. I’ll be happy to assist if I can.

Note: I have separated your post into a new topic.

Unfortunately, I’m unsure if we’ll be able to assist much further without gathering the diagnostics, we use many different aspects of the files collected to help understand what is happening with a device like a docking station when it isn’t working correctly. If you do not wish to send us the diagnostics I understand, and a return for a refund is certainly no problem. Please note, if purchased from us via Amazon, we can accept a return for a refund past the date Amazon has provided you. So there isn’t a time limit should you choose to troubleshoot with us and if you decide to return it later.

For privacy concerns, we speak to what is collected on the diagnostics page link at the bottom, I’ve copied it below for your reference. The data collected is designed to specifically not contain any personal information such as passwords, stored credit card information, addresses, etc.

At minimum, to diagnose an issue like what you’re experiencing, I would need the following files sent to us:

  • devman.csv (so I can see the hardware devices attached, driver versions for said hardware, etc)
  • dxdiag.txt (primarily for the USB device tree)
  • installed_apps.csv (check various software updates, driver software installations, etc)
  • msinfo32.txt (system details such as hardware, system components, and software environmen)
  • usbdev.csv (USB device details)

You’re free to examine these files yourself, there should be no sensitive data collected.

Please don’t hesitate to let us know of other questions.

Thanks again for contacting Plugable support and best wishes!

Joshua Henry
Senior Engineer | Product Owner
Plugable Technologies

As mentioned above, details for our diagnostics:

Background on System Information and Logs:

The Plugable USB and Bluetooth Debugging Tool (PlugDebug) gathers the following system information:

  • PC and OS Spec and other system information
  • USB devices currently connected to the PC
  • Bluetooth devices in range of the PC
  • Installed software
  • Recent history of devices plugged in to PC
  • Crash dumps
  • Driver installation history
  • Current driver store packages
  • Network adapter information
  • DirectX Ddiagnostic information
  • DisplayLink devices connected to the PC and status
  • Extended Display Identification Data (EDID)
  • Hard Drive SMART Data
  • DisplayLink debug logs
  • DisplayLink registry settings
  • Power configuration information

We use the following NirSoft utilities in the Windows version of Plugdebug:

  • BlueScreenView
  • BluetoothView
  • DevManView
  • InstalledDriversList
  • MonitorInfoView
  • MultiMonitorTool
  • MyUninstaller
  • USBDeview

We’d like to thank Nirsoft for their wonderful software. Without them this tool would have taken considerably longer to develop.

We also bundle a Windows binary of smartctl from the smartmontools project.

3

Hey Joshua,

The purposes you outline here do sound reasonable, but the information you collect in the files is more than is needed per your description.

  • devman.csv - if there are any particular devices you are interested - I can share that information. I cannot share full list of devices as there are proprietary devices connected to my environment, knowledge of which will reveal sensitive IP.
  • dxdiag.txt - cannot share - contains machine name => PII
  • installed_apps.csv - cannot share, path to installed apps contains user name => PII
  • msinfo32.txt - cannot share - contains machine name and user name => PII
  • usbdev.csv - same as devman.csv

For what sake you are collecting users.csv and bunch of other data is beyond me and none of the PII I mentioned is listed on the download page or your list. Also, please note that crash dumps contain passwords and other sensitive information. You are saying:

The data collected is designed to specifically not contain any personal information such as passwords, stored credit card information, addresses, etc.

And then in your list there is bullet point saying “Crash dumps”.

Thanks for a great suggestion to return the device, it is always a pleasure to work with a knowledgeable and technical support team!

4

I understand you will not be able to send us the diagnostics required due to your system having sensitive IP (proprietary devices, etc).

For what sake you are collecting users.csv and bunch of other data is beyond me and none of the PII I mentioned is listed on the download page or your list.

We collect data such as users.csv because several of our products involve migrating customer data, when being contacted for help, we need to know how to walk them through the process. Many customers are not knowledgeable in such areas and need help to even know what their user account name is. A user account name is not usually considered a private piece of information to most customers.

Also, please note that crash dumps contain passwords and other sensitive information.

Crash dumps are vital to solve some issues, hence why they’re collected. Generally speaking, crash dumps do not contain plain text passwords or other usable identifiable information. It would require significant knowledge and purposeful malicious intent to extract such data from a crash dump.

My best suggestion at this time would be to ensure your host system has all of the latest drivers, UEFI BIOS updates, firmware updates, etc installed. Also, we’d recommend to ensure you have the latest Intel Thunderbolt 3 and Intel Management Engine software installed.

If the aforementioned updates do not resolve the issues you’re experiencing, a refund (which you initially mentioned) "Return window on Amazon ends on Jan 31, if we don’t get this resolved - I’ll probably just return until there is a more stable dock available." would likely be the best option as we don’t have the ability to properly diagnose the issue you’re experiencing.

Thanks and best wishes,
Josh

5

I think what you are trying to say is “yes we collect PII, but we blindfold ourselves when looking at the collected data”. Usually this should be achieved by scrubbing before submission, not by telling engineers to turn their head around and forget last 10 minutes when they see user name being “AMZN\JeffB”, which is clearly a personally identifiable information. I’d suggest you to work on both - your EULA + privacy statement and your data collection, because as you mentioned - some information is not relevant for troubleshooting of this particular device, but it is still being collected. We should make it clear that it is not me not willing to submit troubleshooting information, but you asking me to submit PII that is irrelevant for troubleshooting.

Honestly, I don’t see any of the information collected being required to share for troubleshooting of the problem. If you know the list of incompatible devices, then you should simply publish it for the users to self-troubleshoot, which will reduce the burden on the support team. If you are just looking for a device that you are unfamiliar with and plan to blame it all on the third-party, well, then we are going nowhere. Same for software and the rest of the information that is being collected.

It appears to me that at this point we would need to kd over serial to laptop and JTAG into the dock. I’ll give you time to reproduce the issue in your QA lab and work on the fix until the return window. Keep in mind we are not talking about TongFang or Clevo here - this is pretty popular PC from the well-known brand, which I’d assume would definitely be in the test matrix. So far this seems to be just another device that was rushed into the market without proper testing/validation phase.

All drivers are at the latest from WU, of course.

6

Hello,

I’d suggest you to work on both - your EULA + privacy statement and your data collection, because as you mentioned - some information is not relevant for troubleshooting of this particular device, but it is still being collected.

I’ll be happy to share your privacy concerns with our development team.

If you are just looking for a device that you are unfamiliar with and plan to blame it all on the third-party, well, then we are going nowhere. Same for software and the rest of the information that is being collected.

Sometimes we do encounter new devices or software that can cause issues, but that’s not our primary goal (to point blame at a third party device or software).

As an example, with a different dock model, we had seen issues caused by Microsoft Teams and Zoom which would cause instability with DisplayLink USB Graphics earlier this year. This was only able to be discovered from the data points gathered from multiple customer reports of issues and finding these applications in common. We were able to isolate that issue and find a workaround in that case.

It appears to me that at this point we would need to kd over serial to laptop and JTAG into the dock.

This is not something that we’ll be able to do in our standard diagnostics process.

Keep in mind we are not talking about TongFang or Clevo here - this is pretty popular PC from the well-known brand, which I’d assume would definitely be in the test matrix.

If I may ask, what brand and exact model system do you have? If you can provide that information at a minimum, I can look into any known incompatibilities or commonalities from other customer reports. So far, you’ve only mentioned a Microsoft Surface Pro 7, but also called out that it hasn’t been stable with any laptop you’ve tried. So I’m curious what systems you are seeing issues with.

As to well known brands, if say HP or Dell for example, there are numerous different configurations for any given model. Sometimes a specific sub-model may have troubles, but not the entire model line. In addition, it is not possible for us to test all system models on the market with our products, there are tens of thousands of currently available system models/sub-models across major manufacturers.

All drivers are at the latest from WU, of course.

Unfortunately, Windows Update is not a reliable platform to ensure updated drivers (except for the Microsoft Surface as that is the primary method of update for that system line). We’d suggest to get the latest from the system manufacturer directly, or via chipset manufacturers if the system manufacturer has not published the latest drivers on their website (a common problem we encounter).

So far this seems to be just another device that was rushed into the market without proper testing/validation phase.

For a Thunderbolt 3 device to be released to market, it must pass Intel certification, which it has.

Thanks!
Josh

7

Doing data mining on un-scrubbed user data needs to be clearly called out as such. So far we are slowly uncovering personal and sensitive information that is leaking through the diagnostic tooling you provided.

There is nothing wrong with scenarios you have outlined, but that is not the (only) data you are collecting. Take this particular example:

… we had seen issues caused by Microsoft Teams and Zoom …

Do you really need to know the path (which includes PII) to the software installation location to come to a conclusion that all users with X installed have issues with Y?

I frankly don’t know why you decided to split up the threads, because this conversation is related to HP Spectre x360 13 (13t-aw100), and it was in the subject line of the original thread.

…you’ve only mentioned a Microsoft Surface Pro 7…

…Windows Update is not a reliable platform to ensure updated drivers (except for the Microsoft Surface…

¯\_(ツ)_/¯

For a Thunderbolt 3 device to be released to market, it must pass Intel certification, which it has.

It seems like you are mixing apples and oranges here. That is indeed a requirement from Intel Corp., but not from a manufacturer who would (I assume) want a high customer satisfaction ratio. There are device-specific functional requirements that are not covered by Intel Corp. simply allowing you to slap “Thunderbolt” trademark on the device. In simple words, Intel certification is not going to verify whether your USB mug heater is reaching the temperature that you claim it reaches in the Amazon description. If in your mind a sole requirement for a Plugable TBT3-UDZ dock to be released to the market is to pass Intel’s certification, then I’m sorry to disappoint you by saying that by us, the consumers, it is actually expected to perform functions that it claims to perform…

Using consumers as a free test resource is “great”, but there is a limited amount of time I can spend on this, as I have my job to do to make sure my customers don’t experience issues with my products. If you cannot provide compliant troubleshooting tooling that collects relevant “personally non-identifiable” (as your online privacy statement says) information and allows you to pinpoint the issue - then I think it’s a wrap.

Thanks.

8

Hello,

I frankly don’t know why you decided to split up the threads, because this conversation is related to HP Spectre x360 13 (13t-aw100), and it was in the subject line of the original thread.

We treat each report as a separate case to not cause confusion or create noise for other customers. Your system could be slightly different than what the other customer has. I cannot just assume it’s identical.

That aforementioned model number is a generic “Product Name” and not the exact sub model. Please see here: https://support.hp.com/us-en/document/c00033108

We need to know the “Product number (SKU, part number)” to know what exact system and configuration you’ve got on hand. I’d ask for the serial number, but I don’t expect you’ll be willing to provide that.

If in your mind a sole requirement for a Plugable TBT3-UDZ dock to be released to the market is to pass Intel’s certification, then I’m sorry to disappoint you by saying that by us, the consumers, it is actually expected to perform functions that it claims to perform…

We follow the guidelines and procedures as recommended to us by Intel certification. Beyond that, we do our best to perform testing on a variety of host systems and operating systems that is at a level considered as standard within this industry, but it’s impossible to validate every single combination for a product like this.

Using consumers as a free test resource is “great”, but there is a limited amount of time I can spend on this, as I have my job to do to make sure my customers don’t experience issues with my products. If you cannot provide compliant troubleshooting tooling that collects relevant “personally non-identifiable” (as your online privacy statement says) information and allows you to pinpoint the issue - then I think it’s a wrap.

I understand your concerns, and I understand you can’t expend more time into troubleshooting. Unfortunately, I cannot provide what you ask for at this time. We will not be able to revamp our diagnostics utility before your Amazon return window closes. (However, as mentioned we can assist in a refund past that date, but I understand if you don’t want to go that route.)

Best wishes,
Josh

9

Spectre SKU is 6XL30AV and you got it correct about the serial.

Since you did split up the threads, we could look into Surface, but that has even tighter compliance requirements than the HP, so not sure how much we will be able to advance there.

Indeed, I’m not in the mood to opt-in to something other than “1-click” return on Amazon, so I’ll give it 30-40 days and call it done.

11

Thanks for the model!

Looking on HP’s support website, I recommend to double check that these updates are installed as they won’t be provided via Windows Update:

  1. Newest BIOS for this model: https://support.hp.com/us-en/drivers/selfservice/swdetails/hp-spectre-13-aw0000-x360-convertible-pc-series/29100376/model/29100378/swItemId/ob-255002-1

  2. Latest Intel chipset drivers from HP: https://support.hp.com/us-en/drivers/selfservice/swdetails/hp-spectre-13-aw0000-x360-convertible-pc-series/29100376/model/29100378/swItemId/ob-246051-1

  3. Latest Intel Management Engine driver from HP: https://support.hp.com/us-en/drivers/selfservice/swdetails/hp-spectre-13-aw0000-x360-convertible-pc-series/29100376/model/29100378/swItemId/ob-240155-2

  4. *Latest Intel Thunderbolt 3 driver from HP: https://support.hp.com/us-en/drivers/selfservice/swdetails/hp-spectre-13-aw0000-x360-convertible-pc-series/29100376/model/29100378/swItemId/ob-240153-2

*The TBT3 drivers from HP are out of date, I suggest to get the newest direct from Intel: https://downloadmirror.intel.com/28735/eng/TBT_WIN10_64_1.41.1015.0.zip

To install the Intel provided Thunderbolt 3 drivers:

  1. Download this driver package and unzip it to a folder on the system.
  2. Connect a Thunderbolt device to the Thunderbolt port.
  3. Browse to and select the folder where you unzipped the driver package.
  4. Right-click on all four .inf files individually and select Install:
  • TbtHostController.inf
  • TbtHostControllerHsaComponent.inf
  • TbtHostControllerToastComponent.inf
  • TbtP2pNdisDrv.inf

You’re probably already aware, but just in case, you’ll also need the newest Thunderbolt Control Center software (available on the Microsoft Store) to approve and manage Thunderbolt devices. Though, I suspect your system will authenticate new devices automatically without the need for user intervention.

If you still encounter the USB issues when waking from hibernate, check in the usbdev.csv to see if there are three “Generic SuperSpeed USB Hub” devices showing as connected with the Vendor ID “05e3” (Genesys Logic). If they are not detected, we know the issue is from the hub ICs failing to wake and are not being enumerated. If they are detected, but are not recognizing devices, the hubs may be in a bad state. As to why, I am unsure at this moment.

As a test to see if there’s any change, we can try disabling power saving on those hubs in the device manager and see if there is any change. (On the “Power Management” tab of the “Generic SuperSpeed USB Hub Properties” and unchecking “Allow the computer to turn off this device to save power”.)

It may also be worth disabling hybrid sleep. (On the “Advanced settings” tab of “Power Options”, expand Sleep tree, then expand Allow hybrid sleep subtree.)

Best wishes,

Josh

12

Okay, first of all, I needed a consistent repro, as previously it was all over the place. I did some troubleshooting and the problem appears to be happening more often when switching dock between different laptops, especially when dock is disconnected after laptop was switched off/hibernated. If I keep the dock plugged in to the Spectre, then it goes into this limp mode occasionally after restart (1 out of 3 times shutdown /r /t 0) or hibernate.

(Spoiler alert: this turned out not to be true, but this whole troubleshooting took me about an hour to perform consistently, so by the time I realized that simple reboot does not yield consistent behavior - I already tried double-laptop scenario a few times)

The final test scenario I settled on (kind of time consuming, but was a consistent repro):

  1. Dock connected to SP7, up and running.
  2. Put SP7 to hibernation
  3. Disconnect the dock from SP7
  4. Plug the dock into HP Spectre, which is hibernated as well
  5. Power on HP Spectre and wait for it to boot
    Result: No USB devices are working, while external monitors are functional

It is worth mentioning that doing this procedure from Spectre to SP7 - you have USB devices working, but external displays not working.

  1. Newest BIOS for this model: …

The problem with what you (and HP) call the “newest” BIOS is that it’s not the newest and downgrade is not allowed due to security reasons.

image
image767×276 7.66 KB

  1. Latest Intel chipset drivers from HP: …

Installed this first, tested above steps - no difference.

  1. Latest Intel Management Engine driver from HP: …

Newest already installed: 2013.14.0.1529 (compared to 1910.13.0.1060 from HP). Skipped.

  1. *Latest Intel Thunderbolt 3 driver from HP: …

HP has: 1.41.648.5
My machine has: 1.41.987.0
Installed the suggested 1.41.1015.0 (for Intel NUCs) - no improvement (in fact, normal reboot right after the driver installation put dock in the limp mode, even though I didn’t plug/unplug anything).

After installing latest TB3 driver I don’t see dock showing up in the Thunderbolt Control Center at all, even when it is connected and appears to be in the working state (everything attached to it works):

<see image in the next post>

… try disabling power saving on those hubs …

No difference.

After I updated power settings and rebooted - dock went into a limp mode again. This time there were no hubs in devman at all and I got the following error message after boot (notice how it says “last USB device you connected”, even though I didn’t connect anything, it was plugged in all along):

<see image in the next post>

I’ve had this message on random occasions previously with restoring from hibernate as well, but more often it’s just USB ports not working without the message.

Unplug the dock and plug it back fixed the issue. This time dock did show up in the Thunderbolt Control Center.

These were two consecutive reboots (not hibernates) when dock didn’t come back online.

It may also be worth disabling hybrid sleep

Didn’t try, as I was getting it more often than not with a simple reboot.

I give up at this point and deem this device as not usable. Reboot should be the basic scenario, in my opinion. Unless there is a new firmware for the dock itself - I’m not going to install random drivers, putting the machine down the rabbit hole up to a no-recovery point.

(Lame forum doesn’t allow more than one media per post, to be continued…)

13

No dock in TCC, but everything works:

image
image749×568 24.8 KB

14

Error message right after reboot:

image

15

Thanks for providing all of this information. It is most appreciated.

This exact behavior is unlike anything we’ve encountered so far, even on other similar HP systems (like the issue from the original thread before I separated them). The closest we’ve seen is where the USB hubs aren’t redetected on wake, but we’ve not run into the dock no longer being detected in the Thunderbolt 3 utility or encountered this USB error. We have also encountered some display issues, but that’s been more intermittent and likely not directly tied to these specific issues.

Looking through our other support tickets for this dock and similar HP systems, we are definitely seeing some problems with Intel 10 and 11 series based HP systems more so than other brands, or earlier HP models with Intel 7th through 9th series. Speaking with my colleagues this morning about these issues we suspect that there’s something incorrect happening in the communication between the host system Thunderbolt 3 controller and the dock TBT3 controller.

We’re going to communicate with Intel to see if they’ve heard of anything like this, and if they have any engineering contacts at HP that we can work with. Unfortunately, at this moment I don’t have further suggestions for a path towards a resolution given what we’ve already tried.

I’ll keep you updated with any new information I’m able to gather.

Thanks,
Josh

16

Well, your Intel folks may not need to go that far.

This laptop is on F16 BIOS.

AFAIK, F16 updated CPU microcode, which I doubt was developed by HP. I’d expect Intel to be aware of what they have changed there. Given that it is an Ice Lake (integrated TB3 controller), it might as well be some changes to the host TB controller and incompatibility with the TB controller you are using in the dock. In that case it is one of two things - Intel rolls-back new microcode for Ice Lake or your dock vendor produces new firmware that is compatible with latest Intel changes…

17

This topic was automatically closed 20 days after the last reply. New replies are no longer allowed.