* Markus Armbruster (arm...@redhat.com) wrote: > "Dr. David Alan Gilbert" <dgilb...@redhat.com> writes: > > > * Markus Armbruster (arm...@redhat.com) wrote: > >> We compile pci-hmp-cmds.c always, but pci-qmp-cmds.c only when > >> CONFIG_PCI. hw/pci/pci-stub.c keeps the linker happy when > >> !CONFIG_PCI. Build pci-hmp-cmds.c that way, too. > >> > >> Signed-off-by: Markus Armbruster <arm...@redhat.com> > > > > Had you considered wrapping the hmp-commands-info.hx entry > > with a #if defined instead? > > No. Would you prefer that?
It seemed a bit simpler to me, but I'm not too fussed. I kind of preferred the idea of the command giving an error if there's no PCI built in. > Code containing #ifdef CONFIG_PCI is target-dependent. Looks like the > affected monitor code already is, so no new headaches. > Aside: splitting off its target-independent parts could be nice. Not > today. Yeh. Dave > [...] > -- Dr. David Alan Gilbert / dgilb...@redhat.com / Manchester, UK
