Re: [Qemu-devel] [PATCH 3/3] virtio-blk: note optional features

[Anthony Liguori <anthony@codemonkey.ws>]

On 03/06/2012 06:22 AM, Paolo Bonzini wrote:
> The guest must already be prepared to see SG_IO support
> disappear from under its feet, for example if migration
> refers to a block device on the source and file-based
> storage on the destination; or more likely, if the source
> kernel allows (gasp) SG_IO on a partition and the destination
> does not.  So, we can migrate safely even if the source
> had VIRTIO_BLK_F_SCSI and the destination does not.

I don't know how comfortable I feel about this.

You can't just remove a feature in flight.  The guest is going to behave 
differently in such a way that the host isn't expecting.  Yes, it should fail 
gracefully, but nonetheless it will fail.

Aren't you just delaying the inevitable?  Instead of having migration fail, the 
guest workload is going to fail.  How is this an improvement?

Regards,

Anthony Liguorig

>
> Signed-off-by: Paolo Bonzini<pbonzini@redhat.com>
> ---
>   hw/virtio-blk.c |    2 +-
>   1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/hw/virtio-blk.c b/hw/virtio-blk.c
> index c95f8fc..9a4158a 100644
> --- a/hw/virtio-blk.c
> +++ b/hw/virtio-blk.c
> @@ -542,7 +542,7 @@ static int virtio_blk_load(QEMUFile *f, void *opaque, int version_id)
>       if (version_id != 2)
>           return -EINVAL;
>
> -    ret = virtio_load(&s->vdev, f, 0);
> +    ret = virtio_load(&s->vdev, f, VIRTIO_BLK_F_SCSI);
>       if (ret) {
>           return ret;
>       }