Re: [Xen-devel] [PATCH] Add a device_model_machine option for HVM domains to libxl.

[Paul Durrant <Paul.Durrant@xxxxxxxxxx>]

> -----Original Message-----
> From: Ian Campbell
> Sent: 19 June 2013 09:48
> To: Paul Durrant
> Cc: Stefano Stabellini; xen-devel@xxxxxxxxxxxxx
> Subject: Re: [Xen-devel] [PATCH] Add a device_model_machine option for
> HVM domains to libxl.
> 
> On Wed, 2013-06-19 at 09:29 +0100, Paul Durrant wrote:
> > > > +        b_info->u.hvm.machine =
> > > LIBXL_DEVICE_MODEL_HVM_MACHINE_DEFAULT;
> > >
> > > This isn't needed, libxl_domain_build_info_init() will DTRT here.
> > >
> >
> > That's what I thought. But without that line I seem to get some random
> value.
> 
> Weird. The autogenerated function "libxl_domain_build_info_init_type"
> should have coped with this, can you post its content?
> 

No sign of initialization here... Perhaps I needed to delete the fail to cause 
re-generation, or maybe running 'make tools' is not enough?

void libxl_domain_build_info_init_type(libxl_domain_build_info *p, 
libxl_domain_type type)
{
    assert(p->type == LIBXL_DOMAIN_TYPE_INVALID);
    p->type = type;
    switch (p->type) {
    case LIBXL_DOMAIN_TYPE_HVM:
        p->u.hvm.timer_mode = LIBXL_TIMER_MODE_DEFAULT;
        libxl_vga_interface_info_init(&p->u.hvm.vga);
        libxl_vnc_info_init(&p->u.hvm.vnc);
        libxl_sdl_info_init(&p->u.hvm.sdl);
        libxl_spice_info_init(&p->u.hvm.spice);
        break;
    case LIBXL_DOMAIN_TYPE_PV:
        p->u.pv.slack_memkb = LIBXL_MEMKB_DEFAULT;
        break;
    case LIBXL_DOMAIN_TYPE_INVALID:
        break;
    }
}

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel