Re: [Qemu-devel] [RFC PATCH v2] hw/s390x/virtio-ccw: Consolidate calls to virtio_ccw_unrealize()

[Cornelia Huck]

On Tue, 24 Jul 2018 10:43:59 +0200
Thomas Huth <address@hidden> wrote:

> Currently, every virtio-ccw device explicitely sets its unrealize
> function to virtio_ccw_unrealize() in its class_init function.
> We can simplify this by using a common unrealize function, just like
> it is already done for the realize functions.
> 
> Signed-off-by: Thomas Huth <address@hidden>
> ---
>  hw/s390x/virtio-ccw.c | 22 +++++++---------------
>  1 file changed, 7 insertions(+), 15 deletions(-)
> 
> diff --git a/hw/s390x/virtio-ccw.c b/hw/s390x/virtio-ccw.c
> index 7ddb378..f3ad7aa 100644
> --- a/hw/s390x/virtio-ccw.c
> +++ b/hw/s390x/virtio-ccw.c
> @@ -750,8 +750,9 @@ out_err:
>      g_free(sch);
>  }
>  
> -static void virtio_ccw_unrealize(VirtioCcwDevice *dev, Error **errp)
> +static void virtio_ccw_device_unrealize(VirtioCcwDevice *dev, Error **errp)
>  {
> +    VirtIOCCWDeviceClass *dc = VIRTIO_CCW_DEVICE_GET_CLASS(dev);
>      CcwDevice *ccw_dev = CCW_DEVICE(dev);
>      SubchDev *sch = ccw_dev->sch;
>  
> @@ -764,6 +765,10 @@ static void virtio_ccw_unrealize(VirtioCcwDevice *dev, 
> Error **errp)
>          release_indicator(&dev->routes.adapter, dev->indicators);
>          dev->indicators = NULL;
>      }
> +
> +    if (dc->unrealize) {
> +        dc->unrealize(dev, errp);
> +    }
>  }
>  

Hm... if any device type should have the need to do something in the
unrealize path, it would probably want to do it before the common
unrealize handling, I think. For example, it might want a valid sch.