Re: [libvirt] [PATCHv2 06/15] Add virtio-related options to interfaces

On Tue, Jun 06, 2017 at 01:36:20PM +0200, Ján Tomko wrote: > <interface type='user'> > <mac address='52:54:56:5a:5c:5e'/> > <model type='virtio'/> > <driver iommu_platform='on' ats='on'/> > </interface> > > https://bugzilla.redhat.com/show_bug.cgi?id=1283251 > --- > docs/formatdomain.html.in | 19 +++++++ > docs/schemas/domaincommon.rng | 12 +++++ > src/conf/domain_conf.c | 63 ++++++++++++++++++++++ > src/conf/domain_conf.h | 19 +++++++ > .../qemuxml2argv-virtio-options.xml | 2 + > .../qemuxml2xmlout-virtio-options.xml | 2 + > 6 files changed, 117 insertions(+) I would spit this patch into two patches, one that introduces the virtio related options and second one that introduces this element to the interfaces.

> diff --git a/docs/formatdomain.html.in b/docs/formatdomain.html.in > index 2f1e030..dcc2e5e 100644 > --- a/docs/formatdomain.html.in > +++ b/docs/formatdomain.html.in > @@ -3451,6 +3451,19 @@ > </dd> > </dl> > > + <h4><a name="elementsVirtio">Virtio-related options</a></h4> > + > + <p> > + QEMU's virtio devices have some attributes related to the virtio transport under > + the <code>driver</code> element: > + The <code>iommu_platform</code> attribute enables the use of emulated IOMMU I think that we tend to use camelCase for attributes and elements. Isn't "iommuEnabled" or just "iommu", we don't need to follow QEMU naming, it should be something generic.