On Tue, Aug 24, 2021, 2:20 PM Jan Beulich <jbeul...@suse.com> wrote: > On 18.08.2021 22:29, Bobby Eshleman wrote: > > --- /dev/null > > +++ b/xen/include/asm-x86/gdbsx.h > > @@ -0,0 +1,17 @@ > > +#ifndef __X86_GDBX_H > > +#define __X86_GDBX_H__ > > + > > +#ifdef CONFIG_GDBSX > > + > > +int gdbsx_guest_mem_io(domid_t domid, struct xen_domctl_gdbsx_memio > *iop); > > + > > +#else > > + > > +static inline int gdbsx_guest_mem_io(domid_t domid, struct > xen_domctl_gdbsx_memio *iop) > > +{ > > + return -EOPNOTSUPP; > > +} > > + > > +#endif > > In addition to what Andrew has said, you also want to make sure > - domid_t is actually declared (need to include public/xen.h, I think), > - struct xen_domctl_gdbsx_memio has a forward declaration (ahead of the > #ifdef). > This is so the header can actually be #include-d without needing to > worry about prereq #include-s. > > Jan >
Roger that. I'll be away from work on vacation until mid September but I'll get to these changes and the others when I get back, if there are some pending / still uncommitted. Writing this from my phone, so sorry for gmail's email strangeness. >