Comment 2 for bug 1848492

------- Comment From <email address hidden> 2019-10-17 09:58 EDT-------
(In reply to comment #7)
> @IBM: For a kernel SRU request we would need a test case.
> Since the terminology between Linux, HMC in DPM mode and HMC in classic mode
> is different,
> am I right assuming that we talk (in terms of classic mode, that's what we
> have) about:
> Activation Profile --> Storage
> and then
> Central Storage:
> Initial n GB
> and
> Reserved m GB
> and after a re-activation of the LPAR, incl. IPL, free must only show n GB,
> right?

Yes, that's correct. We are talking about the "Reserved" field with "Central Storage".
I wouldn't know how to write a test case here, since the kernel brought up all "reserved"/"standby" memory already before user space can figure out anything.

The only thing that works: _if_ you know that an LPAR is configured with "Reserved" memory then the last line of the output of lsmem must _not_ look this:

Total offline memory: 0B

But instead contain a value that is not zero.

We may play some tricks about to which (Linux) memory zone got added, but that changed in the past - and I really wouldn't rely on that for a test case.