qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Markus Armbruster <armbru@redhat.com>
To: Alistair Francis <alistair23@gmail.com>
Cc: "Daniel P. Berrange" <berrange@redhat.com>,
	Eduardo Habkost <ehabkost@redhat.com>,
	"Michael S . Tsirkin" <mst@redhat.com>,
	Alistair Francis <alistair@alistair23.me>,
	Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>,
	"qemu-devel@nongnu.org Developers" <qemu-devel@nongnu.org>,
	Gerd Hoffmann <kraxel@redhat.com>,
	Paolo Bonzini <pbonzini@redhat.com>,
	David Gibson <david@gibson.dropbear.id.au>
Subject: Re: [PATCH 03/55] qdev: New qdev_new(), qdev_realize(), etc.
Date: Wed, 20 May 2020 06:26:13 +0200	[thread overview]
Message-ID: <87mu63ut96.fsf@dusky.pond.sub.org> (raw)
In-Reply-To: <CAKmqyKP6ccwxc+6DoxJ3kH1uA-PLL47OMxw9RjBzcMXHo3S9Fw@mail.gmail.com> (Alistair Francis's message of "Tue, 19 May 2020 14:02:47 -0700")

Alistair Francis <alistair23@gmail.com> writes:

> On Tue, May 19, 2020 at 8:11 AM Markus Armbruster <armbru@redhat.com> wrote:
>>
>> We commonly plug devices into their bus right when we create them,
>> like this:
>>
>>     dev = qdev_create(bus, type_name);
>>
>> Note that @dev is a weak reference.  The reference from @bus to @dev
>> is the only strong one.
>>
>> We realize at some later time, either with
>>
>>     object_property_set_bool(OBJECT(dev), true, "realized", errp);
>>
>> or its convenience wrapper
>>
>>     qdev_init_nofail(dev);
>>
>> If @dev still has no QOM parent then, realizing makes the
>> /machine/unattached/ orphanage its QOM parent.
>>
>> Note that the device returned by qdev_create() is plugged into a bus,
>> but doesn't have a QOM parent, yet.  Until it acquires one,
>> unrealizing the bus will hang in bus_unparent():
>>
>>     while ((kid = QTAILQ_FIRST(&bus->children)) != NULL) {
>>         DeviceState *dev = kid->child;
>>         object_unparent(OBJECT(dev));
>>     }
>>
>> object_unparent() does nothing when its argument has no QOM parent,
>> and the loop spins forever.
>>
>> Device state "no QOM parent, but plugged into bus" is dangerous.
>>
>> Paolo suggested to delay plugging into the bus until realize.  We need
>> to plug into the parent bus before we call the device's realize
>> method, in case it uses the parent bus.  So the dangerous state still
>> exists, but only within realization, where we can manage it safely.
>>
>> This commit creates infrastructure to do this:
>>
>>     dev = qdev_new(type_name);
>>     ...
>>     qdev_realize_and_unref(dev, bus, errp)
>>
>> Note that @dev becomes a strong reference here.
>> qdev_realize_and_unref() drops it.  There is also plain
>> qdev_realize(), which doesn't drop it.
>>
>> The remainder of this series will convert all users to this new
>> interface.
>>
>> Cc: Michael S. Tsirkin <mst@redhat.com>
>> Cc: Marcel Apfelbaum <marcel.apfelbaum@gmail.com>
>> Cc: Alistair Francis <alistair@alistair23.me>
>> Cc: Gerd Hoffmann <kraxel@redhat.com>
>> Cc: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
>> Cc: David Gibson <david@gibson.dropbear.id.au>
>> Signed-off-by: Markus Armbruster <armbru@redhat.com>
>> ---
>>  include/hw/qdev-core.h | 11 ++++-
>>  hw/core/bus.c          | 14 +++++++
>>  hw/core/qdev.c         | 94 ++++++++++++++++++++++++++++++++++++++++++
>>  3 files changed, 118 insertions(+), 1 deletion(-)
>>
>> diff --git a/include/hw/qdev-core.h b/include/hw/qdev-core.h
>> index b870b27966..fba29308f7 100644
>> --- a/include/hw/qdev-core.h
>> +++ b/include/hw/qdev-core.h
>> @@ -57,7 +57,7 @@ typedef void (*BusUnrealize)(BusState *bus);
>>   * After successful realization, setting static properties will fail.
>>   *
>>   * As an interim step, the #DeviceState:realized property can also be
>> - * set with qdev_init_nofail().
>> + * set with qdev_realize() or qdev_init_nofail().
>>   * In the future, devices will propagate this state change to their children
>>   * and along busses they expose.
>>   * The point in time will be deferred to machine creation, so that values
>> @@ -322,7 +322,13 @@ compat_props_add(GPtrArray *arr,
>>
>>  DeviceState *qdev_create(BusState *bus, const char *name);
>>  DeviceState *qdev_try_create(BusState *bus, const char *name);
>> +DeviceState *qdev_new(const char *name);
>> +DeviceState *qdev_try_new(const char *name);
>>  void qdev_init_nofail(DeviceState *dev);
>> +bool qdev_realize(DeviceState *dev, BusState *bus, Error **errp);
>> +bool qdev_realize_and_unref(DeviceState *dev, BusState *bus, Error **errp);
>> +void qdev_unrealize(DeviceState *dev);
>> +
>>  void qdev_set_legacy_instance_id(DeviceState *dev, int alias_id,
>>                                   int required_for_version);
>>  HotplugHandler *qdev_get_bus_hotplug_handler(DeviceState *dev);
>> @@ -411,6 +417,9 @@ typedef int (qdev_walkerfn)(DeviceState *dev, void *opaque);
>>  void qbus_create_inplace(void *bus, size_t size, const char *typename,
>>                           DeviceState *parent, const char *name);
>>  BusState *qbus_create(const char *typename, DeviceState *parent, const char *name);
>> +bool qbus_realize(BusState *bus, Error **errp);
>> +void qbus_unrealize(BusState *bus);
>> +
>>  /* Returns > 0 if either devfn or busfn skip walk somewhere in cursion,
>>   *         < 0 if either devfn or busfn terminate walk somewhere in cursion,
>>   *           0 otherwise. */
>> diff --git a/hw/core/bus.c b/hw/core/bus.c
>> index 08c5eab24a..bf622604a3 100644
>> --- a/hw/core/bus.c
>> +++ b/hw/core/bus.c
>> @@ -169,6 +169,20 @@ BusState *qbus_create(const char *typename, DeviceState *parent, const char *nam
>>      return bus;
>>  }
>>
>> +bool qbus_realize(BusState *bus, Error **errp)
>> +{
>> +    Error *err = NULL;
>> +
>> +    object_property_set_bool(OBJECT(bus), true, "realized", &err);
>> +    error_propagate(errp, err);
>> +    return !err;
>> +}
>> +
>> +void qbus_unrealize(BusState *bus)
>> +{
>> +    object_property_set_bool(OBJECT(bus), true, "realized", &error_abort);
>
> Not false?
>
> Alistair

Reasons it's &error_abort:

1. PATCH 06 and 07 transform variations of

      object_property_set_bool(..., false, "realized", &error_abort);

   to

      qdev_unrealize(...);

   No untransformed unrealization remain.  Thus, we always abort on
   unrealization error before this series.

2. If unrealize could fail, we'd be in deep trouble.  Recent commit
   b69c3c21a5 "qdev: Unrealize must not fail" explains:

   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.

Clearer now?

With any luck, people will use the simpler qdev_unrealize() and
qbus_unrealize(), which is the form that doesn't let them get the error
handling wrong.  I like it when interfaces make misuse hard :)



  reply	other threads:[~2020-05-20  4:27 UTC|newest]

Thread overview: 125+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-05-19 14:54 [PATCH 00/55] qdev: Rework how we plug into the parent bus Markus Armbruster
2020-05-19 14:54 ` [PATCH 01/55] qdev: Rename qbus_realize() to qbus_init() Markus Armbruster
2020-05-19 14:54 ` [PATCH 02/55] qdev: Drop redundant bus realization Markus Armbruster
2020-05-20 12:00   ` Philippe Mathieu-Daudé
2020-05-20 14:25     ` Markus Armbruster
2020-05-19 14:54 ` [PATCH 03/55] qdev: New qdev_new(), qdev_realize(), etc Markus Armbruster
2020-05-19 21:02   ` Alistair Francis
2020-05-20  4:26     ` Markus Armbruster [this message]
2020-05-20  4:51       ` Alistair Francis
2020-05-20  7:29         ` Markus Armbruster
2020-05-20  6:22   ` Paolo Bonzini
2020-05-20  8:11     ` Markus Armbruster
2020-05-20  8:17       ` Paolo Bonzini
2020-05-20 14:42         ` Markus Armbruster
2020-05-20 16:28           ` Paolo Bonzini
2020-05-25  6:30             ` Markus Armbruster
2020-05-25  6:40               ` Paolo Bonzini
2020-05-29 12:22           ` Markus Armbruster
2020-05-20  8:49   ` Gerd Hoffmann
2020-05-19 14:55 ` [PATCH 04/55] qdev: Put qdev_new() to use with Coccinelle Markus Armbruster
2020-05-19 14:55 ` [PATCH 05/55] qdev: Convert to qbus_realize(), qbus_unrealize() Markus Armbruster
2020-05-19 14:55 ` [PATCH 06/55] qdev: Convert to qdev_unrealize() with Coccinelle Markus Armbruster
2020-05-19 14:55 ` [PATCH 07/55] qdev: Convert to qdev_unrealize() manually Markus Armbruster
2020-05-20  6:25   ` Paolo Bonzini
2020-05-20  8:12     ` Markus Armbruster
2020-05-19 14:55 ` [PATCH 08/55] qdev: Convert uses of qdev_create() with Coccinelle Markus Armbruster
2020-05-20  6:30   ` Paolo Bonzini
2020-05-20  8:16     ` Markus Armbruster
2020-05-19 14:55 ` [PATCH 09/55] qdev: Convert uses of qdev_create() manually Markus Armbruster
2020-05-19 14:55 ` [PATCH 10/55] qdev: Convert uses of qdev_set_parent_bus() with Coccinelle Markus Armbruster
2020-05-19 14:55 ` [PATCH 11/55] qdev: Convert uses of qdev_set_parent_bus() manually Markus Armbruster
2020-05-19 14:55 ` [PATCH 12/55] pci: New pci_new(), pci_realize_and_unref() etc Markus Armbruster
2020-05-19 14:55 ` [PATCH 13/55] hw/ppc: Eliminate two superfluous QOM casts Markus Armbruster
2020-05-26 11:56   ` Philippe Mathieu-Daudé
2020-05-19 14:55 ` [PATCH 14/55] pci: Convert uses of pci_create() etc. with Coccinelle Markus Armbruster
2020-05-19 14:55 ` [PATCH 15/55] pci: Convert uses of pci_create() etc. manually Markus Armbruster
2020-05-19 14:55 ` [PATCH 16/55] pci: pci_create(), pci_create_multifunction() are now unused, drop Markus Armbruster
2020-05-19 14:55 ` [PATCH 17/55] isa: New isa_new(), isa_realize_and_unref() etc Markus Armbruster
2020-05-19 14:55 ` [PATCH 18/55] isa: Convert uses of isa_create() with Coccinelle Markus Armbruster
2020-05-19 14:55 ` [PATCH 19/55] isa: Convert uses of isa_create(), isa_try_create() manually Markus Armbruster
2020-05-19 14:55 ` [PATCH 20/55] isa: isa_create(), isa_try_create() are now unused, drop Markus Armbruster
2020-05-19 14:55 ` [PATCH 21/55] ssi: ssi_auto_connect_slaves() never does anything, drop Markus Armbruster
2020-05-19 21:08   ` Alistair Francis
2020-05-19 14:55 ` [PATCH 22/55] ssi: Convert uses of ssi_create_slave_no_init() with Coccinelle Markus Armbruster
2020-05-19 21:07   ` Alistair Francis
2020-05-19 14:55 ` [PATCH 23/55] ssi: Convert last use of ssi_create_slave_no_init() manually Markus Armbruster
2020-05-19 20:58   ` Alistair Francis
2020-05-19 14:55 ` [PATCH 24/55] ssi: ssi_create_slave_no_init() is now unused, drop Markus Armbruster
2020-05-19 21:11   ` Alistair Francis
2020-05-19 14:55 ` [PATCH 25/55] usb: New usb_new(), usb_realize_and_unref() Markus Armbruster
2020-05-20  8:44   ` Gerd Hoffmann
2020-05-19 14:55 ` [PATCH 26/55] usb: Convert uses of usb_create() Markus Armbruster
2020-05-20  8:45   ` Gerd Hoffmann
2020-05-19 14:55 ` [PATCH 27/55] usb: usb_create() is now unused, drop Markus Armbruster
2020-05-20  8:46   ` Gerd Hoffmann
2020-05-19 14:55 ` [PATCH 28/55] usb: Eliminate usb_try_create_simple() Markus Armbruster
2020-05-20  8:46   ` Gerd Hoffmann
2020-05-19 14:55 ` [PATCH 29/55] qdev: qdev_create(), qdev_try_create() are now unused, drop Markus Armbruster
2020-05-19 14:55 ` [PATCH 30/55] auxbus: New aux_realize_bus(), pairing with aux_init_bus() Markus Armbruster
2020-05-26 11:54   ` Philippe Mathieu-Daudé
2020-05-27  4:39     ` Markus Armbruster
2020-05-19 14:55 ` [PATCH 31/55] auxbus: Convert a use of qdev_set_parent_bus() Markus Armbruster
2020-05-19 14:55 ` [PATCH 32/55] auxbus: Eliminate aux_create_slave() Markus Armbruster
2020-05-20 11:52   ` Philippe Mathieu-Daudé
2020-05-19 14:55 ` [PATCH 33/55] qom: Tidy up a few object_initialize_child() calls Markus Armbruster
2020-05-19 21:14   ` Alistair Francis
2020-05-26 11:51   ` Philippe Mathieu-Daudé
2020-05-19 14:55 ` [PATCH 34/55] qom: Less verbose object_initialize_child() Markus Armbruster
2020-05-19 21:16   ` Alistair Francis
2020-05-19 14:55 ` [PATCH 35/55] macio: Convert use of qdev_set_parent_bus() Markus Armbruster
2020-05-19 14:55 ` [PATCH 36/55] macio: Eliminate macio_init_child_obj() Markus Armbruster
2020-05-19 14:55 ` [PATCH 37/55] sysbus: Drop useless OBJECT() in sysbus_init_child_obj() calls Markus Armbruster
2020-05-20 12:02   ` Philippe Mathieu-Daudé
2020-05-19 14:55 ` [PATCH 38/55] microbit: Tidy up sysbus_init_child_obj() @child argument Markus Armbruster
2020-05-20 12:06   ` Philippe Mathieu-Daudé
2020-05-20 14:49     ` Markus Armbruster
2020-05-20 14:54       ` Philippe Mathieu-Daudé
2020-05-19 14:55 ` [PATCH 39/55] sysbus: Tidy up sysbus_init_child_obj()'s @childsize arg, part 1 Markus Armbruster
2020-05-19 14:55 ` [PATCH 40/55] hw/arm/armsse: Pass correct child size to sysbus_init_child_obj() Markus Armbruster
2020-05-20 11:51   ` Philippe Mathieu-Daudé
2020-05-20 14:54     ` Markus Armbruster
2020-05-19 14:55 ` [PATCH 41/55] sysbus: Tidy up sysbus_init_child_obj()'s @childsize arg, part 2 Markus Armbruster
2020-05-19 14:55 ` [PATCH 42/55] sysbus: New sysbus_realize(), sysbus_realize_and_unref() Markus Armbruster
2020-05-19 14:55 ` [PATCH 43/55] sysbus: Convert to sysbus_realize() etc. with Coccinelle Markus Armbruster
2020-05-19 21:18   ` Alistair Francis
2020-05-19 14:55 ` [PATCH 44/55] qdev: Drop qdev_realize() support for null bus Markus Armbruster
2020-05-19 14:55 ` [PATCH 45/55] sysbus: Convert qdev_set_parent_bus() use with Coccinelle, part 1 Markus Armbruster
2020-05-19 21:25   ` Alistair Francis
2020-05-19 14:55 ` [PATCH 46/55] sysbus: Convert qdev_set_parent_bus() use with Coccinelle, part 2 Markus Armbruster
2020-05-19 21:26   ` Alistair Francis
2020-05-19 14:55 ` [PATCH 47/55] sysbus: Convert qdev_set_parent_bus() use with Coccinelle, part 3 Markus Armbruster
2020-05-19 14:55 ` [PATCH 48/55] sysbus: Convert qdev_set_parent_bus() use with Coccinelle, part 4 Markus Armbruster
2020-05-19 14:55 ` [PATCH 49/55] sysbus: sysbus_init_child_obj() is now unused, drop Markus Armbruster
2020-05-19 14:55 ` [PATCH 50/55] s390x/event-facility: Simplify creation of SCLP event devices Markus Armbruster
2020-05-20  8:09   ` David Hildenbrand
2020-05-21  8:44     ` David Hildenbrand
2020-05-25  7:01       ` Markus Armbruster
2020-05-25  8:26         ` Paolo Bonzini
2020-05-26  6:27           ` Markus Armbruster
2020-05-26  7:51             ` Paolo Bonzini
2020-05-26  8:59               ` Markus Armbruster
2020-05-29 13:45         ` Markus Armbruster
2020-05-26  9:45   ` Cornelia Huck
2020-05-26 11:23     ` Paolo Bonzini
2020-05-26 11:38       ` Cornelia Huck
2020-05-26  9:59   ` David Hildenbrand
2020-05-19 14:55 ` [PATCH 51/55] qdev: Make qdev_realize() support bus-less devices Markus Armbruster
2020-05-20  6:43   ` Paolo Bonzini
2020-05-20 15:02     ` Markus Armbruster
2020-05-20 16:24       ` Paolo Bonzini
2020-05-25  6:38         ` Markus Armbruster
2020-05-25 10:11           ` Paolo Bonzini
2020-05-26  5:14             ` Markus Armbruster
2020-05-26  7:54               ` Paolo Bonzini
2020-05-19 14:55 ` [PATCH 52/55] qdev: Use qdev_realize() in qdev_device_add() Markus Armbruster
2020-05-19 14:55 ` [PATCH 53/55] qdev: Convert bus-less devices to qdev_realize() with Coccinelle Markus Armbruster
2020-05-19 21:28   ` Alistair Francis
2020-05-19 14:55 ` [PATCH 54/55] qdev: qdev_init_nofail() is now unused, drop Markus Armbruster
2020-05-19 14:55 ` [PATCH 55/55] MAINTAINERS: Make section QOM cover hw/core/*bus.c as well Markus Armbruster
2020-05-20  6:46 ` [PATCH 00/55] qdev: Rework how we plug into the parent bus Paolo Bonzini
2020-06-08 10:56   ` Markus Armbruster
2020-06-08 10:59     ` Paolo Bonzini
2020-06-09  6:41       ` Markus Armbruster
2020-06-09  6:55         ` Paolo Bonzini
2020-06-09  9:34           ` Markus Armbruster

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=87mu63ut96.fsf@dusky.pond.sub.org \
    --to=armbru@redhat.com \
    --cc=alistair23@gmail.com \
    --cc=alistair@alistair23.me \
    --cc=berrange@redhat.com \
    --cc=david@gibson.dropbear.id.au \
    --cc=ehabkost@redhat.com \
    --cc=kraxel@redhat.com \
    --cc=mark.cave-ayland@ilande.co.uk \
    --cc=mst@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).