Re: [PATCH v11 13/26] s390: vfio-ap: zeroize the AP queues

[Cornelia Huck]

On Tue, 25 Sep 2018 19:16:28 -0400
Tony Krowiak <akrowiak@xxxxxxxxxxxxxxxxxx> wrote:

> From: Tony Krowiak <akrowiak@xxxxxxxxxxxxx>
> 
> Let's call PAPQ(ZAPQ) to zeroize a queue for each queue configured
> for a mediated matrix device when it is released.
> 
> Zeroizing a queue resets the queue, clears all pending
> messages for the queue entries and disables adapter interruptions
> associated with the queue.
> 
> Signed-off-by: Tony Krowiak <akrowiak@xxxxxxxxxxxxx>
> Reviewed-by: Halil Pasic <pasic@xxxxxxxxxxxxx>
> Tested-by: Michael Mueller <mimu@xxxxxxxxxxxxx>
> Tested-by: Farhan Ali <alifm@xxxxxxxxxxxxx>
> ---
>  drivers/s390/crypto/vfio_ap_ops.c | 44 +++++++++++++++++++++++++++++++
>  1 file changed, 44 insertions(+)
> 
> diff --git a/drivers/s390/crypto/vfio_ap_ops.c b/drivers/s390/crypto/vfio_ap_ops.c
> index b4f67d547493..1943e47d7d78 100644
> --- a/drivers/s390/crypto/vfio_ap_ops.c
> +++ b/drivers/s390/crypto/vfio_ap_ops.c
> @@ -816,6 +816,49 @@ static int vfio_ap_mdev_group_notifier(struct notifier_block *nb,
>  	return NOTIFY_OK;
>  }
>  
> +static int vfio_ap_mdev_reset_queue(unsigned int apid, unsigned int apqi,
> +				    unsigned int retry)
> +{
> +	struct ap_queue_status status;
> +
> +	do {
> +		status = ap_zapq(AP_MKQID(apid, apqi));
> +		switch (status.response_code) {
> +		case AP_RESPONSE_NORMAL:
> +			return 0;
> +		case AP_RESPONSE_RESET_IN_PROGRESS:
> +		case AP_RESPONSE_BUSY:
> +			msleep(20);
> +			break;
> +		default:
> +			/* things are really broken, give up */
> +			return -EIO;
> +		}
> +	} while (retry--);
> +
> +	return -EBUSY;
> +}
> +
> +static int vfio_ap_mdev_reset_queues(struct mdev_device *mdev)
> +{
> +	int ret;
> +	int rc = 0;
> +	unsigned long apid, apqi;
> +	struct ap_matrix_mdev *matrix_mdev = mdev_get_drvdata(mdev);
> +
> +	for_each_set_bit_inv(apid, matrix_mdev->matrix.apm,
> +			     matrix_mdev->matrix.apm_max + 1) {
> +		for_each_set_bit_inv(apqi, matrix_mdev->matrix.aqm,
> +				     matrix_mdev->matrix.aqm_max + 1) {
> +			ret = vfio_ap_mdev_reset_queue(apid, apqi, 1);

I'd still like to see a comment here:

/*
 * Regardless whether a queue turns out to be busy, or is not
 * operational, we need to continue resetting the remaining
 * queues.
 */

or so.

> +			if (ret)
> +				rc = ret;
> +		}
> +	}
> +
> +	return rc;
> +}
> +
>  static int vfio_ap_mdev_open(struct mdev_device *mdev)
>  {
>  	struct ap_matrix_mdev *matrix_mdev = mdev_get_drvdata(mdev);
> @@ -846,6 +889,7 @@ static void vfio_ap_mdev_release(struct mdev_device *mdev)
>  	if (matrix_mdev->kvm)
>  		kvm_arch_crypto_clear_masks(matrix_mdev->kvm);
>  
> +	vfio_ap_mdev_reset_queues(mdev);
>  	vfio_unregister_notifier(mdev_dev(mdev), VFIO_GROUP_NOTIFY,
>  				 &matrix_mdev->group_notifier);
>  	matrix_mdev->kvm = NULL;