On 06/12/2012 15:21, Peter Maydell wrote:
On 6 December 2012 13:58, KONRAD Frédéric <fred.kon...@greensocs.com> wrote:
On 06/12/2012 11:13, Peter Maydell wrote:
It can't just be a command line alias, or we will break migration.
It has to be a simple device that composes together the virtio-pci
and virtio-blk devices, plus legacy support for properties and
migration state, I think.
Can we do virtio-blk refactoring and virtio-blk-pci at the same time for not
breaking anything ?
Not breaking things is a key part of the requirements here.
Agree with that.
It's ok to say "I haven't converted virtio-net or the s390
transport in this patchset and therefore they are broken" as
an initial RFC (because we can look at how PCI/blk is done
and check it works before we expand the same thing out to
other transports/devices). But you need to show how the
virtio-blk / virtio-pci refactoring works and leaves you with
a virtio-blk-pci that isn't broken (either at the end or at
any step along the way).
And if virtio-blk-pci is broken can we refactor it in the same "patch" ?
-- PMM
