qemu/hw/acpi
Markus Armbruster b69c3c21a5 qdev: Unrealize must not fail
Devices may have component devices and buses.

Device realization may fail.  Realization is recursive: a device's
realize() method realizes its components, and device_set_realized()
realizes its buses (which should in turn realize the devices on that
bus, except bus_set_realized() doesn't implement that, yet).

When realization of a component or bus fails, we need to roll back:
unrealize everything we realized so far.  If any of these unrealizes
failed, the device would be left in an inconsistent state.  Must not
happen.

device_set_realized() lets it happen: it ignores errors in the roll
back code starting at label child_realize_fail.

Since realization is recursive, unrealization must be recursive, too.
But how could a partly failed unrealize be rolled back?  We'd have to
re-realize, which can fail.  This design is fundamentally broken.

device_set_realized() does not roll back at all.  Instead, it keeps
unrealizing, ignoring further errors.

It can screw up even for a device with no buses: if the lone
dc->unrealize() fails, it still unregisters vmstate, and calls
listeners' unrealize() callback.

bus_set_realized() does not roll back either.  Instead, it stops
unrealizing.

Fortunately, no unrealize method can fail, as we'll see below.

To fix the design error, drop parameter @errp from all the unrealize
methods.

Any unrealize method that uses @errp now needs an update.  This leads
us to unrealize() methods that can fail.  Merely passing it to another
unrealize method cannot cause failure, though.  Here are the ones that
do other things with @errp:

* virtio_serial_device_unrealize()

  Fails when qbus_set_hotplug_handler() fails, but still does all the
  other work.  On failure, the device would stay realized with its
  resources completely gone.  Oops.  Can't happen, because
  qbus_set_hotplug_handler() can't actually fail here.  Pass
  &error_abort to qbus_set_hotplug_handler() instead.

* hw/ppc/spapr_drc.c's unrealize()

  Fails when object_property_del() fails, but all the other work is
  already done.  On failure, the device would stay realized with its
  vmstate registration gone.  Oops.  Can't happen, because
  object_property_del() can't actually fail here.  Pass &error_abort
  to object_property_del() instead.

* spapr_phb_unrealize()

  Fails and bails out when remove_drcs() fails, but other work is
  already done.  On failure, the device would stay realized with some
  of its resources gone.  Oops.  remove_drcs() fails only when
  chassis_from_bus()'s object_property_get_uint() fails, and it can't
  here.  Pass &error_abort to remove_drcs() instead.

Therefore, no unrealize method can fail before this patch.

device_set_realized()'s recursive unrealization via bus uses
object_property_set_bool().  Can't drop @errp there, so pass
&error_abort.

We similarly unrealize with object_property_set_bool() elsewhere,
always ignoring errors.  Pass &error_abort instead.

Several unrealize methods no longer handle errors from other unrealize
methods: virtio_9p_device_unrealize(),
virtio_input_device_unrealize(), scsi_qdev_unrealize(), ...
Much of the deleted error handling looks wrong anyway.

One unrealize methods no longer ignore such errors:
usb_ehci_pci_exit().

Several realize methods no longer ignore errors when rolling back:
v9fs_device_realize_common(), pci_qdev_unrealize(),
spapr_phb_realize(), usb_qdev_realize(), vfio_ccw_realize(),
virtio_device_realize().

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20200505152926.18877-17-armbru@redhat.com>
2020-05-15 07:08:14 +02:00
..
acpi_interface.c
acpi-stub.c
acpi-x86-stub.c acpi: move PC stubs out of stubs/ 2019-12-17 19:33:50 +01:00
aml-build-stub.c acpi: add aml builder stubs 2020-05-04 10:25:02 -04:00
aml-build.c ACPI: Build related register address fields via hardware error fw_cfg blob 2020-05-14 15:03:09 +01:00
bios-linker-loader.c Include qemu-common.h exactly where needed 2019-06-12 13:20:20 +02:00
core.c sysemu: Split sysemu/runstate.h off sysemu/sysemu.h 2019-08-16 13:37:36 +02:00
cpu_hotplug.c hw/acpi/cpu_hotplug: Include "hw/pci/pci.h" 2020-03-09 15:59:31 +01:00
cpu.c Typo: Correct the name of CPU hotplug memory region 2020-04-30 11:52:28 +01:00
generic_event_device.c ACPI: Record the Generic Error Status Block address 2020-05-14 15:03:09 +01:00
ghes.c ACPI: Record Generic Error Status Block(GESB) table 2020-05-14 15:03:09 +01:00
hmat.c hmat acpi: Build Memory Side Cache Information Structure(s) 2020-01-05 07:03:03 -05:00
hmat.h hmat acpi: Build Memory Proximity Domain Attributes Structure(s) 2020-01-05 07:03:03 -05:00
ich9.c Drop more @errp parameters after previous commit 2020-05-15 07:08:14 +02:00
ipmi-stub.c ipmi: Fix SSIF ACPI handling to use the right CRS 2019-09-20 14:08:10 -05:00
ipmi.c ipmi: Fix SSIF ACPI handling to use the right CRS 2019-09-20 14:08:10 -05:00
Kconfig ACPI: Build related register address fields via hardware error fw_cfg blob 2020-05-14 15:03:09 +01:00
Makefile.objs ACPI: Build related register address fields via hardware error fw_cfg blob 2020-05-14 15:03:09 +01:00
memory_hotplug.c hw/acpi: Do not create memory hotplug method when handler is not defined 2019-10-05 17:12:08 -04:00
nvdimm.c acpi: nvdimm: change NVDIMM_UUID_LE to a common macro 2020-05-14 15:03:08 +01:00
pci.c acpi: pci: use build_append_foo() API to construct MCFG 2019-05-29 18:00:57 -04:00
pcihp.c qdev: Unrealize must not fail 2020-05-15 07:08:14 +02:00
piix4.c qom: Drop parameter @errp of object_property_add() & friends 2020-05-15 07:07:58 +02:00
tco.c Include migration/vmstate.h less 2019-08-16 13:31:52 +02:00
tpm.c acpi: add ACPI memory clear interface 2019-01-17 21:10:57 -05:00
trace-events acpi: cpuhp: introduce 'Command data 2' field 2020-01-22 00:23:07 -05:00
utils.c hw/acpi: extract acpi_add_rom_blob() 2019-06-16 16:16:52 -04:00
vmgenid.c qdev: set properties with device_class_set_props() 2020-01-24 20:59:15 +01:00