isa: Clean up use of cannot_instantiate_with_device_add_yet
commitf3b176402fa92149320dcd5479916ccb39cfa0a8
authorMarkus Armbruster <armbru@redhat.com>
Thu, 28 Nov 2013 16:27:02 +0000 (28 17:27 +0100)
committerAndreas Färber <afaerber@suse.de>
Sun, 22 Dec 2013 23:27:23 +0000 (23 00:27 +0100)
tree297212eb1153bf3f1846f141dcfdcbcacb99bdee
parent04916ee9fb8ff4be4d9f612b85aef009ec57b0b4
isa: Clean up use of cannot_instantiate_with_device_add_yet

Drop it when there's no obvious reason why device_add could not work.
Else keep and document why.

* isa-fdc: drop

* i8042: drop, even though its I/O base is hardcoded (because you
  could conceivably still add one to a board that has none), and even
  though PC board code wires up the A20 line (because that wiring is
  optional)

* port92: keep because it needs additional wiring by port92_init()

* mc146818rtc: keep because it needs to be wired up by rtc_init()

* m48t59_isa: keep because needs to be wired up by m48t59_init_isa()

* isa-pit, kvm-pit: keep (in their abstract base pic-common) because
  the PIT needs additional wiring by board code, depending on HPET
  presence

* pcspk: keep because of pointer property pit, and because realize
  sets global pcspk_state

* vmmouse: keep because of pointer property ps2_mouse

* vmport: keep because realize sets global port_state

* isa-i8259, kvm-i8259: keep (in their abstract base pic-common),
  because the PICs' IRQ input lines are set up by board code, and the
  wiring of the slave to the master is hard-coded in device model code

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
hw/audio/pcspk.c
hw/block/fdc.c
hw/i386/pc.c
hw/input/pckbd.c
hw/input/vmmouse.c
hw/intc/i8259_common.c
hw/misc/vmport.c
hw/timer/i8254_common.c
hw/timer/m48t59.c
hw/timer/mc146818rtc.c