From: Anthony Liguori <anthony@codemonkey.ws>
To: "Paolo Bonzini" <pbonzini@redhat.com>,
"Andreas Färber" <afaerber@suse.de>
Cc: Liu Ping Fan <pingfank@linux.vnet.ibm.com>,
liu ping fan <qemulist@gmail.com>,
qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] [RFC PATCH 0/9] qom: improve reference counting and hotplug
Date: Mon, 27 Aug 2012 08:15:08 -0500 [thread overview]
Message-ID: <877gskh3xf.fsf@codemonkey.ws> (raw)
In-Reply-To: <503B637A.7060905@redhat.com>
Paolo Bonzini <pbonzini@redhat.com> writes:
> Il 27/08/2012 13:46, Andreas Färber ha scritto:
>>>> >>
>>>> >> Since link and child properties all hold references, in order to actually free
>>>> >> an object, we need to break those links. User created devices end up as
>>>> >> children of a container. But child properties cannot be removed which means
>>>> >> there's no obvious way to remove the reference and ultimately free the object.
>>>> >>
>>> > Why? Since we call _add_child() in qdev_device_add(), why can not we
>>> > call object_property_del_child() for qmp_device_del(). Could you
>>> > explain it more detail?
>> Seconded. If we hot-unplug a device, we should surely remove its child<>
>> property from /machine/unassigned or parent bus or whatever.
That's exactly what is happening in this series. qmp_device_del adds an
ejection notifier that unparents the device to remove the last reference count.
>
> Sure, as soon as the device is ejected by the guest. But until that
> point we need to keep the device in the QOM tree so that: 1) it has a
> canonical path; 2) it can be examined; 3) it keeps children alive.
>
>> Why is it that child<> properties cannot be removed?
>
> Yeah, I didn't quite understand the difference between unparenting and
> setting the child property to NULL.
They are exactly the same thing. Setting the child property to NULL is
unparenting.
Unparenting is essentially deleting. This series makes it such that
there is a white list of devices that are capable of being deleted.
Regards,
Anthony Liguori
>
> Paolo
next prev parent reply other threads:[~2012-08-27 13:15 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-08-26 15:51 [Qemu-devel] [RFC PATCH 0/9] qom: improve reference counting and hotplug Anthony Liguori
2012-08-26 15:51 ` [Qemu-devel] [PATCH 1/9] savevm: don't rely on paths if we can store a DeviceState object Anthony Liguori
2012-08-26 15:51 ` [Qemu-devel] [PATCH 2/9] object: automatically free objects based on a release function Anthony Liguori
2012-08-27 13:31 ` Andreas Färber
2012-08-26 15:51 ` [Qemu-devel] [PATCH 3/9] qbus: remove glib_allocated/qom_allocated and use release hook to free memory Anthony Liguori
2012-08-27 14:02 ` Andreas Färber
2012-08-27 14:22 ` Anthony Liguori
2012-08-27 14:43 ` Andreas Färber
2012-08-26 15:51 ` [Qemu-devel] [PATCH 4/9] object: remove object_finalize Anthony Liguori
2012-08-27 15:01 ` Andreas Färber
2012-08-27 15:49 ` Anthony Liguori
2012-08-26 15:51 ` [Qemu-devel] [PATCH 5/9] object: add support for nullable child properties Anthony Liguori
2012-08-26 15:51 ` [Qemu-devel] [PATCH 6/9] qdev: make devices created with device_add nullable so they can be deleted Anthony Liguori
2012-08-26 15:51 ` [Qemu-devel] [PATCH 7/9] qdev: add notifier for when the device loses its parent bus (eject) Anthony Liguori
2012-08-26 15:51 ` [Qemu-devel] [PATCH 8/9] qdev: make qdev_set_parent_bus() just set a link property Anthony Liguori
2012-08-27 7:22 ` liu ping fan
2012-08-27 13:12 ` Anthony Liguori
2012-08-26 15:51 ` [Qemu-devel] [PATCH 9/9] hotplug: refactor hotplug to leverage new QOM functions Anthony Liguori
2012-08-27 7:22 ` liu ping fan
2012-08-27 7:22 ` [Qemu-devel] [RFC PATCH 0/9] qom: improve reference counting and hotplug liu ping fan
2012-08-27 11:46 ` Andreas Färber
2012-08-27 12:09 ` Paolo Bonzini
2012-08-27 13:15 ` Anthony Liguori [this message]
2012-10-09 17:15 ` Vasilis Liaskovitis
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=877gskh3xf.fsf@codemonkey.ws \
--to=anthony@codemonkey.ws \
--cc=afaerber@suse.de \
--cc=pbonzini@redhat.com \
--cc=pingfank@linux.vnet.ibm.com \
--cc=qemu-devel@nongnu.org \
--cc=qemulist@gmail.com \
/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).