Vadim Galitsyn <vadim.galit...@profitbricks.com> writes: > Hi Markus, > > Thank you for the input. > >> However, your query-memory looks like it could fail. > > With the latest version of a patch ( > http://lists.nongnu.org/archive/html/qemu-devel/2017-06/msg03475.html) > "query-memory" can fail only in two cases: > > a) if in QOM there is an object of type of TYPE_PC_DIMM which has no > property of type PC_DIMM_SIZE_PROP, or > b) PC_DIMM_SIZE_PROP is not represented as an integer. > > As far as I understand both (a) and (b) can happen only in case if QOM > somehow corrupted which is not a normal case > (please correct me if I am wrong).
These feel like programming errors. They are impossible to test without screwing up the program. Untestable error recovery is commonly wrong. I recommend to assert instead. Igor, what do you think? > Please also note, this is not the last version of the patch since new > functionality was suggested > to be included (NUMA information). > > If patch will be accepted, I think we would need a test case for it since > command output differs once memory > was hot-added/removed per given NUMA node. When final functionality will be > negotiated, I will come up > with test scenario and test case itself. Thanks!
