qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Paolo Bonzini <pbonzini@redhat.com>
To: Eduardo Habkost <ehabkost@redhat.com>,
	Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Cc: qemu-devel <qemu-devel@nongnu.org>,
	Markus Armbruster <armbru@redhat.com>
Subject: Re: QOM address space handling
Date: Sun, 20 Dec 2020 10:25:25 +0100	[thread overview]
Message-ID: <750e98b8-4e7b-f912-f06b-d2b835cc6113@redhat.com> (raw)
In-Reply-To: <20201218223250.GW3140057@habkost.net>

On 18/12/20 23:32, Eduardo Habkost wrote:
> Who owns the FlatView reference, exactly?

The AddressSpace.  The device creates the AddressSpace, which holds a 
reference to the MemoryRegion through FlatView and AddressSpaceDispatch, 
which holds a reference to the device.

By destroying the address space that it created, the device can break 
the reference loop.

> If the FlatView reference is owned by the MemoryRegion, we have a
> reference loop: the device holds a reference to the MemoryRegion,
> which owns the FlatView, which holds a reference to the device.
> In this case, who owns the reference loop and is responsible for
> breaking it?

The reference loop is owned by the device, which breaks it through 
unrealize (called by unparent).

instance_finalize by definition cannot break reference loops, so this 
means that my suggestion of using address_space_init in instance_init 
was wrong.

Thanks,

Paolo

> If the FlatView reference is not owned by the MemoryRegion, who
> owns it?


>>>> There are 2 possible solutions here: 1) ensure QOM objects that add
>>>> address spaces during instance init have a corresponding instance
>>>> finalize function to remove them or 2) move the creation of address
>>>> spaces from instance init to realize.
>>>>
>>>> Does anyone have any arguments for which solution is preferred?
>>> 
>>> I slightly prefer (1) because there could be cases where you also create
>>> subdevices using that address space, and in order to set properties of
>>> subdevices before realize, you would have to create the subdevices in
>>> instance_init as well.



  reply	other threads:[~2020-12-20  9:26 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-11-10 11:14 QOM address space handling Mark Cave-Ayland
2020-11-10 11:40 ` Paolo Bonzini
2020-12-18  7:49   ` Mark Cave-Ayland
2020-12-18 22:32     ` Eduardo Habkost
2020-12-20  9:25       ` Paolo Bonzini [this message]
2020-12-21 18:54         ` Eduardo Habkost
2020-12-21 19:16           ` Paolo Bonzini
2020-12-21 19:28             ` Eduardo Habkost
2020-11-10 15:03 ` Eduardo Habkost
2020-11-10 15:08   ` Paolo Bonzini
2020-11-10 17:46     ` Eduardo Habkost
2020-11-10 18:36       ` Eduardo Habkost

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=750e98b8-4e7b-f912-f06b-d2b835cc6113@redhat.com \
    --to=pbonzini@redhat.com \
    --cc=armbru@redhat.com \
    --cc=ehabkost@redhat.com \
    --cc=mark.cave-ayland@ilande.co.uk \
    --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).