Politely terminating processes with kill hangs the VM backend

Using a QEMU backend on a amd64 host, using `kill -s TERM` or equivalent to kill a running process from the spread shell, a test script inside the VM (or anything else as far as I know) results in the whole VM hanging then displaying this message and exiting:

```
2023-06-06 11:39:17 Error running debug shell: wait: remote command exited without exit status or exit signal
```

Note that `kill -s KILL` still works as expected.

Start by reproducing the issue with the QEMU backend on an amd64 host, terminating a running process with TERM and then KILL for comparison. Trace how the VM backend handles the remote command's termination and verify that TERM no longer hangs or exits without status, while KILL continues to work.

Written by the indexing model from the issue text.