Re: [PATCH iwl-next v3 13/13] vfio/ice: Implement vfio_pci driver for E800 devices

[Jason Gunthorpe <jgg@nvidia.com>]

On Fri, Oct 13, 2023 at 08:52:07AM +0000, Tian, Kevin wrote:
> > From: Cao, Yahui <yahui.cao@intel.com>
> > Sent: Monday, September 18, 2023 2:26 PM
> >
> > +static struct file *
> > +ice_vfio_pci_step_device_state_locked(struct ice_vfio_pci_core_device
> > *ice_vdev,
> > +				      u32 new, u32 final)
> > +{
> > +	u32 cur = ice_vdev->mig_state;
> > +	int ret;
> > +
> > +	if (cur == VFIO_DEVICE_STATE_RUNNING && new ==
> > VFIO_DEVICE_STATE_RUNNING_P2P) {
> > +		ice_migration_suspend_dev(ice_vdev->pf, ice_vdev->vf_id);
> > +		return NULL;
> > +	}
> > +
> > +	if (cur == VFIO_DEVICE_STATE_RUNNING_P2P && new ==
> > VFIO_DEVICE_STATE_STOP)
> > +		return NULL;
> > +
> 
> Jason, above is one open which your clarification is appreciated.
> 
> From my talk with Yahui this device can drain/stop outgoing
> traffic but has no interface to stop incoming request.

> is it OK to do nothing for RUNNING_P2P->STOP transition like above?

Yes.

The purpose of RUNNING_P2P->STOP is to allow the device to do anything
it may need to stop internal autonomous operations prior to doing a
get_state. If the device does not have such a concept then a NOP is
fine.

Jason