Re: [PATCH v11 11/26] s390: vfio-ap: implement mediated device open callback

[Tony Krowiak]

On 09/28/2018 06:14 AM, Cornelia Huck wrote:
> On Tue, 25 Sep 2018 19:16:26 -0400
> Tony Krowiak <akrowiak@xxxxxxxxxxxxxxxxxx> wrote:
>
> > From: Tony Krowiak <akrowiak@xxxxxxxxxxxxx>
> >
> > Implements the open callback on the mediated matrix device.
> > The function registers a group notifier to receive notification
> > of the VFIO_GROUP_NOTIFY_SET_KVM event. When notified,
> > the vfio_ap device driver will get access to the guest's
> > kvm structure. The open callback must ensure that only one
> > mediated device shall be opened per guest.
> >
> > Signed-off-by: Tony Krowiak <akrowiak@xxxxxxxxxxxxx>
> > Acked-by: Halil Pasic <pasic@xxxxxxxxxxxxx>
> > Tested-by: Michael Mueller <mimu@xxxxxxxxxxxxx>
> > Tested-by: Farhan Ali <alifm@xxxxxxxxxxxxx>
> > Tested-by: Pierre Morel <pmorel@xxxxxxxxxxxxx>
> > Acked-by: Pierre Morel <pmorel@xxxxxxxxxxxxx>
> > ---
> >   arch/s390/include/asm/kvm_host.h      |   1 +
> >   drivers/s390/crypto/vfio_ap_ops.c     | 159 ++++++++++++++++++++++++++
> >   drivers/s390/crypto/vfio_ap_private.h |   5 +
> >   3 files changed, 165 insertions(+)
>
> > +static void vfio_ap_mdev_copy_masks(struct ap_matrix_mdev *matrix_mdev)
> > +{
> > +	int nbytes;
> > +	unsigned long *apm, *aqm, *adm;
> > +	struct kvm_s390_crypto_cb *crycb = matrix_mdev->kvm->arch.crypto.crycb;
> > +
> > +	switch (matrix_mdev->kvm->arch.crypto.crycbd & CRYCB_FORMAT_MASK) {
> > +	case CRYCB_FORMAT2:
> > +		apm = (unsigned long *)crycb->apcb1.apm;
> > +		aqm = (unsigned long *)crycb->apcb1.aqm;
> > +		adm = (unsigned long *)crycb->apcb1.adm;
> > +		break;
> > +	case CRYCB_FORMAT1:
> > +	case CRYCB_FORMAT0:
> > +	default:
>
> That one looks a tad odd: the masking above should guarantee that we
> hit either the format 2, 1, or 0 cases, so the default case should
> never hit. Moreover, if the mask above will ever be changed, we don't
> know where we should look for the correct masks (what will a FORMAT<n>
> use: apcb1, apcb0, or something else?) Would it be better to moan in
> the default case and copy nothing?

It probably is a bit odd, but is not a problem now. If a new
format is added, one would hope the developer coding the change
would be smart enough to figure out what to do here.

> > +		apm = (unsigned long *)crycb->apcb0.apm;
> > +		aqm = (unsigned long *)crycb->apcb0.aqm;
> > +		adm = (unsigned long *)crycb->apcb0.adm;
> > +		break;
> > +	}
> > +
> > +	nbytes = DIV_ROUND_UP(matrix_mdev->matrix.apm_max + 1, BITS_PER_BYTE);
> > +	memcpy(apm, matrix_mdev->matrix.apm, nbytes);
> > +	nbytes = DIV_ROUND_UP(matrix_mdev->matrix.aqm_max + 1, BITS_PER_BYTE);
> > +	memcpy(aqm, matrix_mdev->matrix.aqm, nbytes);
> > +	nbytes = DIV_ROUND_UP(matrix_mdev->matrix.adm_max + 1, BITS_PER_BYTE);
> > +	memcpy(adm, matrix_mdev->matrix.adm, nbytes);
> > +}
>
> Anyway, that can be easily changed later on; in general,

true

> Reviewed-by: Cornelia Huck <cohuck@xxxxxxxxxx>