<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Carsten Otte wrote:
<blockquote cite="mid:474BF157.3080709@de.ibm.com" type="cite">
<pre wrap="">Avi Kivity wrote:
</pre>
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">We intend to bind our virtio devices to PCI too, so that they look the
same in Linux userland across architectures.
</pre>
</blockquote>
<pre wrap="">Ouch.
</pre>
</blockquote>
<pre wrap=""><!---->That was my initial opinion too, but HPA has come up with a lean and
clean PCI binding for lguest. I think we should seriously consider
using that over the current qemu device emulation based thing.
</pre>
</blockquote>
There are two solutions for this problem:<br>
1. Use hypercalls and supply mechanism for hypercall patching for qemu.<br>
This way we can make s390 & qemu/xen happy.<br>
2. Have two transport mechanism for virtio.<br>
Actually this is what we have today (but not yet merged) - lguest
uses the pci config space<br>
but without using Anthony's pci module.<br>
We'll have virtio host i(qemu/kernel) implementation for the shared
memory and interface.<br>
We'll have pci transport for x86 that glues the above and a virtual
transport for s390 and paravirt_ops.<br>
Both transports will be based on Rusty's config space.<br>
This is the idea I suggested in Tuscon:<br>
<br>
----- ------------ ---------<br>
| 9p | | network | | block |<br>
------ ------------ ---------<br>
| virtio interface|<br>
----------------<br>
-----------
--------------------------------------------------<br>
| virtio_pci| OR | virtio_vbus (includes
configs & hypercall/portio) |<br>
-----------
--------------------------------------------------<br>
----------- -------------<br>
| virtio_ring| |virtio_config|<br>
----------- -------------<br>
<br>
Regards,<br>
Dor<br>
<blockquote cite="mid:474BF157.3080709@de.ibm.com" type="cite">
<pre wrap="">-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
<a class="moz-txt-link-freetext" href="http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/">http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/</a>
_______________________________________________
kvm-devel mailing list
<a class="moz-txt-link-abbreviated" href="mailto:kvm-devel@lists.sourceforge.net">kvm-devel@lists.sourceforge.net</a>
<a class="moz-txt-link-freetext" href="https://lists.sourceforge.net/lists/listinfo/kvm-devel">https://lists.sourceforge.net/lists/listinfo/kvm-devel</a>
</pre>
</blockquote>
<br>
</body>
</html>