Hi all,
Cisco AnyConnect VPN for us as well. We’ve also had it occur for some users when not on VPN, I've seen it happen myself, but it’s very hard to reproduce consistently outside of the VPN.
Dell have also asked for one of our affected devices, but I’ve made it clear that by the time it gets to them it will just be a wiped OEM build anyway. To be fair, they haven’t said outright that they’re unaware of others having this issue, but they’re definitely not giving much away either. We’re putting a lot of pressure on them, and I’m hearing word of Intel engineers into the mix too, though they’re asking for even more logs.
Completely agree, this issue has consumed a huge amount of time. Gathering logs and testing repeatedly while still maintaining the rest of the environment is a real challenge.
We’re also working closely with Microsoft and seeing some strange behaviour during logging. The NPU driver appears to somehow be involved in a DNS lookup process, and when the issue occurs the NPU seems to lock up entirely, getting stuck while being involved in some weird DNS call while other processes start making unusual DNS requests to a long, unknown SID string that can’t be resolved. It could be related to the freeze and might also explain why it’s worse on Cisco AnyConnect VPN, since the VPN client intercepts DNS and may be exposing a deadlock or timeout condition in the NPU driver or Windows AI stack. Although, it could also be unrelated to the actual freeze, but it is suspicious.
Is anyone else seeing DNS-related errors or unusual lookups around the time of the fault?
We’re still using the NPU workaround for now as it remains the only reliable fix we’ve found.
For those who haven’t seen it, here’s another Dell thread on the issue (noting that the Dell accepted answer there doesn’t actually resolve it). https://www.dell.com/community/en/conversations/dell-pro-laptop-computers/pro-14-plus-pb14250-freezes-when-using-webcam-with-teams-or-zoom/68f17ba5a3935b753dc08618