From: Anthony Liguori <anthony@codemonkey.ws>
To: Paolo Bonzini <pbonzini@redhat.com>
Cc: Peter Maydell <peter.maydell@linaro.org>,
Jan Kiszka <jan.kiszka@siemens.com>,
qemu-devel <qemu-devel@nongnu.org>,
Markus Armbruster <armbru@redhat.com>,
Gerd Hoffmann <kraxel@redhat.com>,
"Edgar E. Iglesias" <edgar.iglesias@gmail.com>
Subject: Re: [Qemu-devel] [RFC] Plan for moving forward with QOM
Date: Thu, 15 Sep 2011 08:26:49 -0500 [thread overview]
Message-ID: <4E71FD19.6050606@codemonkey.ws> (raw)
In-Reply-To: <4E719F7C.10700@redhat.com>
On 09/15/2011 01:47 AM, Paolo Bonzini wrote:
> On 09/14/2011 08:04 PM, Anthony Liguori wrote:
>> The concept of busses are implemented as an
>> interface that a device implements.
>
> I noticed that you haven't written in the document how to make devices reside on
> a particular bus (PCI, ISA, I2C, ...).
>
> The three possibilities for this are:
>
> * a device implements an interface. I would rule this out because for most buses
> the devices will need to store some data (PCI: configuration data, pointer to
> the parent bus; ISA: pointer to the parent bus). Interfaces are
> implementation-only, so you have to place the data in each device and cause
> massive code duplication.
I agree.
>
> * a device inherits from an abstract class, e.g. PCIDevice. It is useful to see
> how the inheritance tree would look like for two devices with a common chipset
> and multiple interfaces:
>
> Device
> NE2000
> PCIDevice
> PCI_NE2000 ------> includes a NE2000
> ISA_NE2000 ------> includes a NE2000
I think this model is the closest to what we have today and is the most obvious.
For something like ne2k, I would expect:
class NE2000 : public Device
{
// ne2k public functions
};
class PCI_NE2000 : public PciDevice
{
// implement PCI functions by calling ne2k public functions
NE2000 ne2k;
};
class ISA_NE2000 : public IsaDevice
{
// implement ISA functions by calling ne2k public functions
NE2000 ne2k;
};
> * a device is composed with a connector object. There is no PCIDevice class
> anymore, so the bus has an array of socket<PCIConnector> instead. The case above
> would look like this
>
> Device
> NE2000 (abstract)
> PCI_NE2000 ------> includes a PCIConnector
> ISA_NE2000 ------> includes an ISAConnector
>
> Or, if you insist on a closer mapping of real hardware, where there are no
> abstract classes, it would look like this:
>
> Device
> NE2000
> PCI_NE2000 ------> includes an NE2000 and a PCIConnector
> ISA_NE2000 ------> includes an NE2000 and an ISAConnector
I think there are two ways to view this:
class PciDevice : public Device
{
PciConnector connector;
// init function registers closures with connector that dispatch
// to abstract functions
};
Or:
class PciConnector : public PciDevice
{
// provides interfaces to register closures which implement
// PCI abstract functions
};
I personally lean toward the later as I don't think the PciConnector model
really does map all that well to hardware (normally, at least). I think this is
much closer to how real hardware actually works.
>
> Advantages of abstract classes are pretty obvious, so I will just list them: it
> is more similar to what is done in QDev, and perhaps it is more intuitive.
>
>
> Advantages of connectors include:
>
> * it is more flexible: it lets you choose between a more abstract and a more
> low-level representation (the two hierarchies above);
>
> * you have the option of showing a simpler device tree to the user, without the
> internal composition. This is important because, unlike QDev, composition in QOM
> is explicit. So QOM would place NIC properties in NE2000, not in *_NE2000 (right?).
>
> * related to this, it keeps property names more stable. If I understand
> correctly, if the device starts as ISA-only or PCI-only, i.e.:
>
> Device
> PCIDevice
> PCI_NE2000
>
> and later you change it to support multiple buses, suddenly properties would
> have to be addressed differently to account for the composition of NE2000 inside
> PCI_NE2000. You could I guess implement "forwarder properties", but that would
> also lead to more boilerplate code.
>
> Any other opinions?
The properties thing is definitely an interesting point, but I'm not sure how
far you can push it. If you start out with a NE2000 device that is ISA and you
decide to abstract it to a shared model, all you need to do is keep the ISA
NE2000 device named NE2000 and call the common chip and PCI bridge something else.
I really think it's important to keep the simple cases simple. I think any
model where you don't do:
class E1000 : public PciDevice
{
};
Is unnecessarily complicated. If it's too complicated, conversions will be much
slower to do and will be more likely to be done wrong.
Regards,
Anthony Liguori
> Paolo
>
next prev parent reply other threads:[~2011-09-15 13:26 UTC|newest]
Thread overview: 88+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-09-14 18:04 [Qemu-devel] [RFC] Plan for moving forward with QOM Anthony Liguori
2011-09-14 18:49 ` Anthony Liguori
2011-09-14 19:30 ` Jan Kiszka
2011-09-14 19:42 ` Anthony Liguori
2011-09-14 21:15 ` Jan Kiszka
2011-09-14 22:11 ` Anthony Liguori
2011-09-15 13:43 ` Jan Kiszka
2011-09-15 14:11 ` Anthony Liguori
2011-09-15 16:38 ` Jan Kiszka
2011-09-15 18:01 ` Anthony Liguori
2011-09-16 10:12 ` Kevin Wolf
2011-09-16 13:00 ` Anthony Liguori
2011-09-14 20:00 ` Edgar E. Iglesias
2011-09-14 20:22 ` Anthony Liguori
2011-09-14 20:27 ` Edgar E. Iglesias
2011-09-14 20:37 ` Blue Swirl
2011-09-14 21:25 ` Anthony Liguori
2011-09-15 6:31 ` Gleb Natapov
2011-09-15 10:49 ` Stefan Hajnoczi
2011-09-15 13:08 ` Anthony Liguori
2011-09-15 13:17 ` Anthony Liguori
2011-09-15 14:23 ` Gleb Natapov
2011-09-16 14:46 ` John Williams
2011-09-16 16:10 ` Anthony Liguori
2011-09-17 1:11 ` Edgar E. Iglesias
2011-09-17 2:12 ` Anthony Liguori
2011-09-17 2:35 ` Edgar E. Iglesias
2011-09-15 13:57 ` Anthony Liguori
2011-09-15 14:14 ` Paolo Bonzini
2011-09-15 14:25 ` Gleb Natapov
2011-09-15 15:28 ` Anthony Liguori
2011-09-15 15:38 ` Gleb Natapov
2011-09-15 16:33 ` Anthony Liguori
2011-09-15 16:59 ` Gleb Natapov
2011-09-15 17:51 ` Anthony Liguori
2011-09-15 20:29 ` Gleb Natapov
2011-09-15 20:45 ` Peter Maydell
2011-09-15 21:15 ` Anthony Liguori
2011-09-16 16:33 ` Gleb Natapov
2011-09-16 17:47 ` Peter Maydell
2011-09-16 18:08 ` Anthony Liguori
2011-09-16 18:22 ` Gleb Natapov
2011-09-16 18:42 ` Anthony Liguori
2011-09-16 19:13 ` Gleb Natapov
2011-09-16 19:29 ` Anthony Liguori
2011-09-16 20:48 ` Gleb Natapov
2011-09-16 21:03 ` Anthony Liguori
2011-09-17 0:01 ` Edgar E. Iglesias
2011-09-16 18:18 ` Gleb Natapov
2011-09-15 20:50 ` Anthony Liguori
2011-09-16 16:47 ` Gleb Natapov
2011-09-17 0:48 ` Edgar E. Iglesias
2011-09-17 2:17 ` Anthony Liguori
2011-09-17 2:29 ` Anthony Liguori
2011-09-17 2:41 ` Edgar E. Iglesias
2011-09-15 6:47 ` Paolo Bonzini
2011-09-15 13:26 ` Anthony Liguori [this message]
2011-09-15 13:35 ` Paolo Bonzini
2011-09-15 13:54 ` Peter Maydell
2011-09-15 14:18 ` Anthony Liguori
2011-09-15 14:33 ` Paolo Bonzini
2011-09-15 14:48 ` Peter Maydell
2011-09-15 15:31 ` Anthony Liguori
2011-09-15 15:47 ` Paolo Bonzini
2011-09-15 20:23 ` Avi Kivity
2011-09-15 20:52 ` Anthony Liguori
2011-09-18 7:56 ` Avi Kivity
2011-09-18 14:00 ` Avi Kivity
2011-09-16 9:36 ` Gerd Hoffmann
2011-12-13 4:47 ` Paul Brook
2011-12-13 13:22 ` Anthony Liguori
2011-12-13 17:40 ` Paul Brook
2011-12-13 18:00 ` Anthony Liguori
2011-12-13 20:36 ` Paul Brook
2011-12-13 21:53 ` Anthony Liguori
2011-12-14 0:39 ` Paul Brook
2011-12-14 13:53 ` Anthony Liguori
2011-12-14 14:01 ` Avi Kivity
2011-12-14 14:11 ` Anthony Liguori
2011-12-14 14:35 ` Avi Kivity
2011-12-14 14:46 ` Anthony Liguori
2011-12-14 14:50 ` Avi Kivity
2011-12-15 18:59 ` Paul Brook
2011-12-15 19:12 ` Anthony Liguori
2011-12-15 21:28 ` Paul Brook
2011-12-16 2:08 ` Anthony Liguori
2011-12-16 5:11 ` Paul Brook
2011-12-14 9:11 ` Andreas Färber
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=4E71FD19.6050606@codemonkey.ws \
--to=anthony@codemonkey.ws \
--cc=armbru@redhat.com \
--cc=edgar.iglesias@gmail.com \
--cc=jan.kiszka@siemens.com \
--cc=kraxel@redhat.com \
--cc=pbonzini@redhat.com \
--cc=peter.maydell@linaro.org \
--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).