This patch series introduces funcionality for waiting for a job from a
different context, for now only for dump jobs, plus testing.
This can be useful in situations where the original requester of the job
crashes/restarts and then needs to continue waiting for that potentially
ongoing job. To avoid waiting for the wrong job type, the caller must
specify the required job type. For now I've only implemented this for
dump jobs; if the logic is correct I can extended it for all other job
types.
The first patch introduces the core functionality. I see that there is a
function specifically for waiting for a dump job:
qemuDumpWaitForCompletion. Ideally we should only use
qemuDomainObjWait, which would work for any job type, however I'm not
sure it's correct to use it instead of qemuDumpWaitForCompletion because
the latter checks dumpCompleted. It's not obvious to me what is the
purpose of dumpCompleted, is it because qemuDomainObjWait can return
because some other job, running in parallel with the dump job, has
completed, so it's effectively a false alarm?
The second patch is mainly for testing the wait functionality via virsh.
The remaining patches extend the test hypervisor's functionality on dump
testing.