Re: [libvirt] [PATCH 6/7] Introduce qemuDomainMachineIsVirt

On Tue, 2016-05-03 at 13:05 +0200, Peter Krempa wrote: > > > > +bool > > +qemuDomainMachineIsVirt(const virDomainDef *def) > > +{ > > +    return STREQ(def->os.machine, "virt") || > > +           STRPREFIX(def->os.machine, "virt-"); > > This could be transcribed as: > strcmp(def->os.machine, "virt") || > strncmp(def->os.machine, "virt-", strlen("virt-")) so > "STRPREFIX(def->os.machine, "virt")" should be equivalent. I think we want to make sure we only accept "virt" and its versioned variants here - if QEMU introuced a new machine type called "virtpc" we wouldn't want it to pass the test. (Then again, "virt-pc" would pass in any case. I think  the current check strikes a good balance, YMMV.)