On Wed, Oct 09, 2013 at 11:27:24AM +0800, liu ping fan wrote: > On Mon, Sep 30, 2013 at 11:58 PM, Michael S. Tsirkin <m...@redhat.com> wrote: > > On Mon, Sep 30, 2013 at 05:48:03PM +0200, Paolo Bonzini wrote: > >> Il 30/09/2013 11:30, Michael S. Tsirkin ha scritto: > >> > I was really only talking about q35 here. > >> > I thought it's ugly that users can control intcap > >> > directly. Can object_set_property be used after > >> > qdev_try_create? > >> > >> Yes, after that and before qdev_init. This is how Ping Fan is doing > >> PIIX right now. > >> > >> > PIIX has another issue: > >> > the default value in hpet is really Q35 specific, > >> > that's also kind of ugly, isn't it? > >> > >> Yeah, perhaps it is. So are you suggesting the default to be 4 (GSI2 > >> only)? > >> > >> Paolo > > > > I suggest it fails unless caller set the property. > > > Sorry, out of office for a long time, and did not keep up with this > thread in time. > When letting the caller set the intcap, we should consider the > compatibility of q35. For pc-q35-1.7 or later, the caller should set > the property, otherwise not.
Set it always - just set it to a compatible value for 1.6. > But how can the caller tell that it runs > on q35-1.7? > The essential problem is that "set the property" will always overwrite > the property which is set up by compatible mechanism. So it is hard to > implement without breaking the current mechanism. Do you think so? > > Thanks and regards, > Ping fan Not that hard. Fail init if it wasn't set.
