On Fri, 2018-03-16 at 11:33 -0400, John Ferlan wrote: > This looks particularly useful for other commands, but... > [...] > > It's only generated/built for one - seems like a lot of work for little > gain unless of course there's a plan to add in all the others. Without > adding in others, one is left to wonder the "best" way to add things in > the future. The RFC versions of the series added three more features, which made very good use of the generic code introduced here. I've later decided to skip those features, at least initially, because QEMU added mitigated machine type variants that cover 99% of what people need. I'm still unconvinced there is a need to have more fine-grained knobs there, so those three features are probably not going to be introduced any time soon. Moreover, between the time I posted this series and today, I've learned that work on a POWER9-compatible version of HTM is progressing much faster than initially anticipated, so this last feature too might end up being unnecessary in the medium run. I'm still not sure about this last part, but there is enough uncertainty there that I would probably not push this before 4.2.0 is out even if it were fully R-b'd. Finally, it was always my intention, assuming the approach taken here had been considered good, to convert existing features to use it: I guess at this point the tables have turned and I should *start* with that instead :) -- Andrea Bolognani / Red Hat / Virtualization -- libvir-list mailing list libvir-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/libvir-list
