`KVM_EXIT_FAIL_ENTRY` on GitHub runners

[mkroening]

Starting at roughly the same time as #1129, we are now frequently seeing KVM_EXIT_FAIL_ENTRY panics on GitHub runners (see #1137):

$ uhyve --cpu-count=1 rusty_demo
thread '<unnamed>' panicked at src/linux/x86_64/kvm_cpu.rs:601:25:
not implemented: FailEntry(0, 2)
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
Error: The operation was canceled.

It would be good to fix this in Uhyve, since other VMMs such as QEMU and Firecracker do not have this issue.
Reproducing and figuring out the cause will be cumbersome, though.

[fogti]

Do we have any chance of decoding what that hardware_entry_failure_reason (presumably in his case 0) means? Or do we have just no direct way of finding out what even remotely went wrong here?