Re: [virtio] [OASIS Issue Tracker] (VIRTIO-156) add VIRTIO_SCSI_F_FC_HOST feature

[Cornelia Huck <cornelia.huck@de.ibm.com>]

On Tue, 24 Jan 2017 13:01:03 +0000
Stefan Hajnoczi <stefanha@redhat.com> wrote:

> On Mon, Jan 23, 2017 at 01:52:59PM +0100, Cornelia Huck wrote:
> > On Sun, 22 Jan 2017 07:32:08 +0000 (UTC)
> > OASIS Issues Tracker  <workgroup_mailer@lists.oasis-open.org> wrote:
> > 
> > > 
> > >     [ https://issues.oasis-open.org/browse/VIRTIO-156?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=65000#comment-65000 ] 
> > > 
> > > Michael S. Tsirkin commented on VIRTIO-156:
> > > -------------------------------------------
> > > 
> > > BALLOT CREATED AT URL: https://www.oasis-open.org/apps/org/workgroup/virtio/ballot.php?id=3033
> > > 
> > > > add VIRTIO_SCSI_F_FC_HOST feature
> > > > ---------------------------------
> > > >
> > > >                 Key: VIRTIO-156
> > > >                 URL: https://issues.oasis-open.org/browse/VIRTIO-156
> > > >             Project: OASIS Virtual I/O Device (VIRTIO) TC
> > > >          Issue Type: New Feature
> > > >    Affects Versions: virtio 1.0 cs04
> > > >         Environment: Reported-by: Paolo Bonzini <pbonzini@redhat.com>
> > > >            Reporter: Paolo Bonzini
> > > >            Assignee: Paolo Bonzini
> > > >              Labels: Extension, request_tc_discussion
> > > >             Fix For: virtio 1.1 cs01
> > > >
> > > >
> > > > The new feature bit, and related configuration fields, allow showing a virtio-scsi device as a Fibre Channel host, including support for live migration.
> > > > The feature is modeled on Hyper-V's SCSI device Fibre Channel host support and should be similarly easy to implement in Linux.
> > 
> > I'm a bit surprised that this is up for vote, since there seemed to be
> > a bit of discussion on the virtualization list about this approach, and
> > it did not yet seem everybody liked this (although discussion seems to
> > have wound down).
> > 
> > I'm not familiar with this topic, and don't oppose this change, but
> > should we wait until all of the involved devs are happy with the general
> > design? Or is the current state good enough?
> 
> Agreed.  I voted "no" for the time being and would like to see the
> upstream driver discussion finish first.

Yes, I saw your vote and did the same.

Should the ballot be withdrawn? It's not that we know that the change
is bad, but just that more discussion is needed...