At this point, this is not a problem I can fix. It's very much tied to their hardware, and to the kernel on that hardware. This is the only instance of this I have ever seen, anywhere, over literally thousands of test runs, and it is very easily reproduced manually by simply catting the efi variable for SecureBoot, so it exists outside the test suite.
I've added a kernel task, if there's any fix to be had here, it would have to occur there.
At this point, this is not a problem I can fix. It's very much tied to their hardware, and to the kernel on that hardware. This is the only instance of this I have ever seen, anywhere, over literally thousands of test runs, and it is very easily reproduced manually by simply catting the efi variable for SecureBoot, so it exists outside the test suite.
I've added a kernel task, if there's any fix to be had here, it would have to occur there.