WPA3-Personal bug report for devices with Intel 9560 chipset on Windows 10/11

I've been banging my head against a wall dealing with Intel L1 tech support and just remembered this incredible base of folks who are passionate about Wi-Fi. I don't know why I didn't ask here first. :slight_smile:

Could someone, anyone, please run the following test:

  1. Collect an 8th gen Intel Core series machine with a 9560 PCI module running Windows 10 or 11. I say 8th gen, because my Lenovo L13 Gen 1 with a 10th Gen i3 does not experience this problem.

  2. Roll back the Wi-Fi driver to 22.150.1.1 (or older)

  3. From a hotspot or wireless access point, create a WPA3-Personal network.

  • If you follow these steps precisely, it will not matter if this new network includes WPA2 fallback or not (ie. Transitional mode flag)
  1. Connect to the newly created WPA3 network

  2. Observe the Wi-Fi profile is indeed using WPA3SAE authentication by running the following commands in the command prompt or powershell:

  • netsh wlan export profile MyNewSSID
  • notepad.exe .\Wi-Fi-MyNewSSID.xml
  1. Update the driver to anything newer than 22.190.0.4.
  • Test with 22.240.0.6 if you prefer.
  • Reboot if you feel it’s necessary
  1. Watch that you can no longer connect to this network!
  • Windows does not honour the transitional mode flag after the Wi-Fi profile has already been created.

For good measure, I added the following steps when reporting this issue to Intel:

  1. Inform Dell, Lenovo, HP, Microsoft et. all and have them revoke affected drivers from the Microsoft catalog
  2. Advise Microsoft that their code to handle the transitional mode flag could be more resilient to handle this sort of scenario
  3. Pay Jason a bug bounty :blush:

Bonus poll question for this forum. Do you think I'm deserving of a bug bounty for discovering this?
I've certainly been busy at work due to it. No paid overtime, just busier.

  • Reproduction Instructions deserving of a bug bounty
  • Nah, only security issues are deserving of a bug bounty
0 voters

Problem devices so far are:
Lenovo T490
Dell OptiPlex 5070
Dell Latitude 3310 (education customer specific)

What does this have to do with OpenWrt specifically? This sounds like it is a Windows 10/11 test on specific PC hardware that could be done with any wifi APs+firmware that has WPA3 support, not specificially related to interactions with OpenWrt. Can you please clarify why this should be here in these forums where we uniquely focus on OpenWrt specific issues?

1 Like

Are you serious? This forum does not have an Off-Topic category, so I put it here.

I thought it would be great to inform the community of a bug in Intel products, which they may run up against testing the latest OpenWrt release candidates....while simultaneously getting someone to replicate this bug to show Intel that I'm not the only one. They have yet to even confirm a reproduction attempt since the issue was reported on Sept. 9th.

Someone here may even be able to help Intel fix the problem! What a relief that would be.

Yup... I am.

Because this forum isn't for things that are not OpenWrt or very closely related.

What you are asking here sounds more like asking for confirmation (or refutation) of a bug in the Intel wifi stack under Windows. I say this because you specifically say that this has not been replicated by the support teams you've been working with thus far. Therefore, this next statement sounds premature:

And then this may or may not actually be an issue if the claimed bug doesn't actually exist:

further, this would not be something that would be addressed by OpenWrt except insofar as to know that those Intel products may produce misleading results. While this could be frustrating (i.e. problems would appear during testing that are not actually the fault of OpenWrt, but it may be hard to determine the actual cause), it would really only lead to a list of devices to avoid/exclude or to use with known limitations during testing.

No. That will need to be intel, if it's really a bug with their hardware or drivers.

All of that said, if you are able to demonstrate that the problems do not occur with a specific version of OpenWrt, but that they do manifest with a different (more recent) version, at that point there could be something that would be addressable by the OpenWrt project.

A little chicken and egg scenario.
Who should know about this issue? How to know if no one reports it?

Anyway, I've changed the title to limit who might venture into this thread.

Well, you are asserting that there is an Intel/Windows bug here (and to be clear: I'm not refuting your claim), but you don't have any evidence except your own experience with a very limited number of devices. While there can be a chicken-or-egg scenario, often driver bugs tend to be fairly well reported and/or reproducible when they occur.

This doesn't change the fact that it is entirely off-topic for the OpenWrt forums. There are probably many other forums that will be more appropriate for this topic/investigation -- consider PC hardware forums, Windows forums, or even sites that offer broad/general network related advice (as compared to ones that are focused on specific networking topics such as this one devoted to OpenWrt).

1 Like

This topic was automatically closed after 12 hours. New replies are no longer allowed.