Re: [Qemu-devel] [PATCH 07/12] qdev-core: isolate reset register/unregister code

[Eduardo Habkost]

On Tue, Oct 23, 2012 at 12:56:18PM -0200, Eduardo Habkost wrote:
> On Wed, Oct 17, 2012 at 01:08:23PM -0500, Anthony Liguori wrote:
> > Eduardo Habkost <address@hidden> writes:
> > 
> > > The reset register/unregister code is specific to qemu-system-*, so
> > > isolate it so it can be moved to qdev-system.c.
> > >
> > > Signed-off-by: Eduardo Habkost <address@hidden>
> > > ---
> > >  hw/qdev-core.c | 24 ++++++++++++++++++------
> > >  1 file changed, 18 insertions(+), 6 deletions(-)
> > >
> > > diff --git a/hw/qdev-core.c b/hw/qdev-core.c
> > > index af0af52..a105679 100644
> > > --- a/hw/qdev-core.c
> > > +++ b/hw/qdev-core.c
> > > @@ -47,6 +47,21 @@ void GCC_WEAK qdev_finalize_vmstate(DeviceState *dev)
> > >  {
> > >  }
> > >  
> > > +static void qbus_register_reset(BusState *bus)
> > > +{
> > > +    if (bus != sysbus_get_default()) {
> > > +        /* TODO: once all bus devices are qdevified,
> > > +           only reset handler for main_system_bus should be registered 
> > > here. */
> > > +        qemu_register_reset(qbus_reset_all_fn, bus);
> > > +    }
> > > +}
> > > +
> > > +static void qbus_unregister_reset(BusState *bus)
> > > +{
> > > +    assert(bus != sysbus_get_default()); /* main_system_bus is never 
> > > freed */
> > > +    qemu_unregister_reset(qbus_reset_all_fn, bus);
> > > +}
> > > +
> > 
> > Again, I'd suggest stubbing out qemu_[un]register_reset instead of
> > trying to cope with it's users.
> 
> I was going to implement it the way you suggested. But then I noticed
> that I _ *-user will need to request the devices to be reset once, too
> (even if the only devices available are the CPU objects).
> 
> On the next version, I will move the reset-function list from vl.c to
> qdev-core.c. Code that uses qdev will need to take care of calling
> qemu_devices_reset() eventually (qemu-system does that on vl.c *-user
> will need to that somewhere else).

I don't usually reply to myself to correct typos, but the message above
was almost impossible to parse. So, rewriting it:

  I was going to implement it the way you suggested. But then I noticed
  that *-user will need to request the devices to be reset once, too
  (even if the only devices available are the CPU objects).

  On the next version, I will move the reset-function list from vl.c to
  qdev-core.c. Code that uses qdev will need to take care of calling
  qemu_devices_reset() eventually (qemu-system does that on vl.c, and
  *-user will need to that somewhere else).


> 
> > 
> > Regards,
> > 
> > Anthony Liguori
> > 
> > >  const char *qdev_fw_name(DeviceState *dev)
> > >  {
> > >      DeviceClass *dc = DEVICE_GET_CLASS(dev);
> > > @@ -355,10 +370,8 @@ static void qbus_realize(BusState *bus)
> > >          QLIST_INSERT_HEAD(&bus->parent->child_bus, bus, sibling);
> > >          bus->parent->num_child_bus++;
> > >          object_property_add_child(OBJECT(bus->parent), bus->name, 
> > > OBJECT(bus), NULL);
> > > -    } else if (bus != sysbus_get_default()) {
> > > -        /* TODO: once all bus devices are qdevified,
> > > -           only reset handler for main_system_bus should be registered 
> > > here. */
> > > -        qemu_register_reset(qbus_reset_all_fn, bus);
> > > +    } else {
> > > +        qbus_register_reset(bus);
> > >      }
> > >  }
> > >  
> > > @@ -692,8 +705,7 @@ static void qbus_finalize(Object *obj)
> > >          QLIST_REMOVE(bus, sibling);
> > >          bus->parent->num_child_bus--;
> > >      } else {
> > > -        assert(bus != sysbus_get_default()); /* main_system_bus is never 
> > > freed */
> > > -        qemu_unregister_reset(qbus_reset_all_fn, bus);
> > > +        qbus_unregister_reset(bus);
> > >      }
> > >      g_free((char *)bus->name);
> > >  }
> > > -- 
> > > 1.7.11.7
> > 
> 
> -- 
> Eduardo
> 

-- 
Eduardo