[PATCH v2] UBUNTU: [Config] Mark CONFIG_SCSI_VIRTIO built-in

[Dimitri John Ledkov]

On Wed, 29 Mar 2023 at 09:36, Roxana Nicolescu
<roxana.nicolescu at canonical.com> wrote:
>
>
> On 28/03/2023 16:51, Dimitri John Ledkov wrote:
> > BugLink: https://bugs.launchpad.net/bugs/1685291
> > BugLink: https://bugs.launchpad.net/bugs/1931841
> > BugLink: https://bugs.launchpad.net/bugs/2006488
> >
> > This enables initrd-less boot on generic kernel, enabling any Ubuntu
> > installation to be lifted from any hardware and converted into a VM
> > which will, even with corrupted initrd and missing modules, be able to
> > boot by default. This is an extremly useful property to have on all
> > ubuntu kernels. Separately, this also removes the need for linux-kvm
> > kernel flavour, as all the changes it has are now incorporated into
> > generic flavour & cloud-images.
> >
> > Signed-off-by: Dimitri John Ledkov <dimitri.ledkov at canonical.com>
> > ---
> >
> >   Changes since v1:
> >   - add buglinks
> >   - improve annotations comment
> >
> >   debian.master/config/annotations | 4 ++--
> >   1 file changed, 2 insertions(+), 2 deletions(-)
> >
> > diff --git a/debian.master/config/annotations b/debian.master/config/annotations
> > index bc466d7318..4017f1f651 100644
> > --- a/debian.master/config/annotations
> > +++ b/debian.master/config/annotations
> > @@ -528,8 +528,8 @@ CONFIG_SCSI_IPR_DUMP                            note<'LP: #1343109'>
> >   CONFIG_SCSI_IPR_TRACE                           policy<{'amd64': 'y', 'arm64': 'y', 'armhf': 'y', 'ppc64el': 'y', 'riscv64': 'y'}>
> >   CONFIG_SCSI_IPR_TRACE                           note<'LP: #1343109'>
> >
> > -CONFIG_SCSI_VIRTIO                              policy<{'amd64': 'm', 'arm64': 'm', 'armhf': 'm', 'ppc64el': 'm', 'riscv64': 'm', 's390x': 'm'}>
> > -CONFIG_SCSI_VIRTIO                              note<'tech preview of new feature'>
> > +CONFIG_SCSI_VIRTIO                              policy<{'amd64': 'y', 'arm64': 'y', 'armhf': 'y', 'ppc64el': 'y', 'riscv64': 'y', 's390x': 'y'}>
> > +CONFIG_SCSI_VIRTIO                              note<'enable initrdless boot in VMs for ease of use, cloud-kernel parity and disaster recovery'>
> >
> >   CONFIG_SECURITY_SAFESETID                       policy<{'amd64': 'y', 'arm64': 'y', 'armhf': 'y', 'ppc64el': 'y', 'riscv64': 'y', 's390x': 'y'}>
> >   CONFIG_SECURITY_SAFESETID                       note<'LP: #1845391'>
> >
> Q: From my understanding this is intended for lunar only. Is there a
> plan to propagate this to other series?
> Just thinking out loud, it would be nice to get rid of one derivative
> (even though I would have to find others to crank;)), but given that it
> is a derivative that people may already use, it may break that, so I am
> curious about the strategy here.

If  / when this lands in lunar v6.2, and if it works out fine to
transition linux-kvm to linux-virtual then we can think about doing a
similar transition in a stable series.
Kinetic will drop off by itself quick enough.
For jammy - i wonder if we will be able to roll linux-kvm to
linux-virtual provided by v6.2 / hwe.
For focal - hopefully we will be able to do nothing / just crank linux-kvm.
Alternative for jammy & focal, is to basically inherit the full
generic config and maybe that will be enough.
The transition in stable series is risky, as there might be users that
rely on built-in or -modules things, that are in -modules-extra. I
haven't done analysis on stable series to ensure that everything that
used to be linux-kvm builtin or modules, isn't left behind in
-modules-extras.

-- 
okurrr,

Dimitri