On Thu, 7 Jun 2018 12:49:23 +0200 David Hildenbrand <da...@redhat.com> wrote:
> On 04.06.2018 14:14, Cornelia Huck wrote: > > On Mon, 4 Jun 2018 14:00:43 +0200 > > David Hildenbrand <da...@redhat.com> wrote: > > > >> On 04.06.2018 13:38, Cornelia Huck wrote: > >>> On Fri, 25 May 2018 13:37:02 +0200 > >>> David Hildenbrand <da...@redhat.com> wrote: > >>> > >>>> Let's treat this like a separate device. TCG will have to store the > >>>> actual state/time later on. > >>>> > >>>> Include cpu-qom.h in kvm_s390x.h (due to S390CPU) to compile tod-kvm.c. > >>>> > >>> > >>> Can't you include it in tod-kvm.c instead? Why was it not needed before? > >>> > >> > >> Guess I can, kvm_s390x.h looks to me like the right place, as S390CPU is > >> used in there. > >> > >> All other users include "cpu.h" already, so it isn't needed. > > > > Hm, and for kvm_s390_irq, this file is using a forward declaration, > > which would likely be more awkward for this structure. > > I remember that we wanted to use cpu-qom.h for this typedef fun, that's > why I decided to do it this way. > > > > > Header dependencies are always fun :) > > > > So what do you suggest? Thanks! > Let's go with cpu-qom.h for this, then.
