On 08/15/2018 12:38 PM, Cornelia Huck wrote:
On Mon, 13 Aug 2018 17:48:15 -0400
Tony Krowiak <[email protected]> wrote:

From: Tony Krowiak <[email protected]>

Implements the VFIO_DEVICE_RESET ioctl. This ioctl zeroizes
all of the AP queues assigned to the guest.

Signed-off-by: Tony Krowiak <[email protected]>
Reviewed-by: Halil Pasic <[email protected]>
Tested-by: Michael Mueller <[email protected]>
Tested-by: Farhan Ali <[email protected]>
Reviewed-by: Pierre Morel <[email protected]>
Tested-by: Pierre Morel <[email protected]>
Signed-off-by: Christian Borntraeger <[email protected]>
---
  drivers/s390/crypto/vfio_ap_ops.c |    5 ++++-
  1 files changed, 4 insertions(+), 1 deletions(-)

diff --git a/drivers/s390/crypto/vfio_ap_ops.c 
b/drivers/s390/crypto/vfio_ap_ops.c
index cc90323..d4a065b 100644
--- a/drivers/s390/crypto/vfio_ap_ops.c
+++ b/drivers/s390/crypto/vfio_ap_ops.c
@@ -832,7 +832,7 @@ static int vfio_ap_mdev_get_device_info(unsigned long arg)
        if (info.argsz < minsz)
                return -EINVAL;
- info.flags = VFIO_DEVICE_FLAGS_AP;
+       info.flags = VFIO_DEVICE_FLAGS_AP | VFIO_DEVICE_FLAGS_RESET;
        info.num_regions = 0;
        info.num_irqs = 0;
@@ -848,6 +848,9 @@ static ssize_t vfio_ap_mdev_ioctl(struct mdev_device *mdev,
        case VFIO_DEVICE_GET_INFO:
                ret = vfio_ap_mdev_get_device_info(arg);
                break;
+       case VFIO_DEVICE_RESET:
+               ret = vfio_ap_mdev_reset_queues(mdev, true);
If I see it correctly, you call this function only ever with force ==
true (here and in the previous patch). Is that what you intended?

That does seem to be the case now; however, I think at one time there were
additional calls to this function. For some reason of which I am not aware,
those were removed, so there is probably no need for it now.


+               break;
        default:
                ret = -EOPNOTSUPP;
                break;


Reply via email to