On Tue, Apr 11, 2017 at 11:42:28AM +0200, Markus Armbruster wrote: > Eric Blake <ebl...@redhat.com> writes: > > > On 04/04/2017 08:28 AM, Kashyap Chamarthy wrote: > > > >>> Minor or not, it is a useful viewpoint. Either way, as long as the new > >>> way of getting a transactional non-pivot successful completion is > >>> something that libvirt can learn via introspection, > >> > >> Can you elaborate a little more on the above, for my own edification -- > >> how might it be possible for "libvirt can learn via introspection"? Is > >> it via some method using the QMP 'query-commands' / > >> 'query-command-line-options'? > > > > Those, and query-qmp-schema. > > Avoid query-command-line-options if you can. If you can't, know its > limitations and quirks.
Thanks; I actually meant to mention 'query-qmp-schema' (thanks to Eric for correcting). And I didn't fortet your reminder of the incompleteness / limitations of 'query-command-line-options' from your KVMForum 2015 presentation :-) > > > If the change includes the addition of > > something new in the .json files describing QMP, then query-qmp-schema > > will let libvirt probe whether the version of qemu it it talking to has > > that new field or not (presumably, in this case it would be a new > > optional bool member to select the mode when creating a job, as well as > > a new command visible through query-commands to change the mode of a > > running job). -- /kashyap
