On 12/06/2018 12:44, Gerd Hoffmann wrote: > From: Marc-André Lureau <marcandre.lur...@redhat.com> > > qbus_create_inplace() creates a new reference in realize(), it must be > released in unrealize(). > > Signed-off-by: Marc-André Lureau <marcandre.lur...@redhat.com> > Message-id: 20180531195119.22021-4-marcandre.lur...@redhat.com > Signed-off-by: Gerd Hoffmann <kra...@redhat.com>
Nope, that's only happening because patch 3 was wrong---and if we did keep patch 3, you would have to do it in all devices that produce buses; not just this one. Paolo > --- > hw/usb/dev-smartcard-reader.c | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/hw/usb/dev-smartcard-reader.c b/hw/usb/dev-smartcard-reader.c > index 2131e33d27..f7c91230d5 100644 > --- a/hw/usb/dev-smartcard-reader.c > +++ b/hw/usb/dev-smartcard-reader.c > @@ -1147,6 +1147,7 @@ static void ccid_unrealize(USBDevice *dev, Error **errp) > USBCCIDState *s = USB_CCID_DEV(dev); > > ccid_bulk_in_clear(s); > + object_unref(OBJECT(&s->bus)); > } > > static void ccid_flush_pending_answers(USBCCIDState *s) >
