Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Received-SPF: Pass (protection.outlook.com: domain of nvidia.com designates
 216.228.112.34 as permitted sender) receiver=protection.outlook.com;
 client-ip=216.228.112.34; helo=mail.nvidia.com;
Date:   Wed, 2 Jun 2021 10:42:05 +0300
From:   Eli Cohen <elic@nvidia.com>
To:     Jason Wang <jasowang@redhat.com>
CC:     <mst@redhat.com>, <virtualization@lists.linux-foundation.org>,
        <linux-kernel@vger.kernel.org>, <kvm@vger.kernel.org>,
        <netdev@vger.kernel.org>, <eli@mellanox.com>
Subject: Re: [PATCH V2 3/4] vp_vdpa: allow set vq state to initial state
 after reset
Message-ID: <20210602074205.GA12498@mtl-vdi-166.wap.labs.mlnx>
References: <20210602021043.39201-1-jasowang@redhat.com>
 <20210602021043.39201-4-jasowang@redhat.com>
 <20210602061324.GA8662@mtl-vdi-166.wap.labs.mlnx>
 <091dc6d0-8754-7b2a-64ec-985ef9db6329@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <091dc6d0-8754-7b2a-64ec-985ef9db6329@redhat.com>
User-Agent: Mutt/1.9.5 (bf161cf53efb) (2018-04-13)
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 02 Jun 2021 07:42:13.0643
 (UTC)
X-MS-Exchange-CrossTenant-Network-Message-Id: 83b53868-fbb8-4693-bd75-08d92599efc3
X-MS-Exchange-CrossTenant-Id: 43083d15-7273-40c1-b7db-39efd9ccc17a
X-MS-Exchange-CrossTenant-OriginalAttributedTenantConnectingIp: TenantId=43083d15-7273-40c1-b7db-39efd9ccc17a;Ip=[216.228.112.34];Helo=[mail.nvidia.com]
X-MS-Exchange-CrossTenant-AuthSource: BN8NAM11FT029.eop-nam11.prod.protection.outlook.com
X-MS-Exchange-CrossTenant-AuthAs: Anonymous
X-MS-Exchange-CrossTenant-FromEntityHeader: HybridOnPrem
X-MS-Exchange-Transport-CrossTenantHeadersStamped: DM6PR12MB2748
Precedence: bulk
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

On Wed, Jun 02, 2021 at 03:05:35PM +0800, Jason Wang wrote:
> 
> 在 2021/6/2 下午2:13, Eli Cohen 写道:
> > On Wed, Jun 02, 2021 at 10:10:42AM +0800, Jason Wang wrote:
> > > We used to fail the set_vq_state() since it was not supported yet by
> > > the virtio spec. But if the bus tries to set the state which is equal
> > > to the device initial state after reset, we can let it go.
> > > 
> > > This is a must for virtio_vdpa() to set vq state during probe which is
> > > required for some vDPA parents.
> > > 
> > > Signed-off-by: Jason Wang <jasowang@redhat.com>
> > > ---
> > >   drivers/vdpa/virtio_pci/vp_vdpa.c | 42 ++++++++++++++++++++++++++++---
> > >   1 file changed, 39 insertions(+), 3 deletions(-)
> > > 
> > > diff --git a/drivers/vdpa/virtio_pci/vp_vdpa.c b/drivers/vdpa/virtio_pci/vp_vdpa.c
> > > index c76ebb531212..18bf4a422772 100644
> > > --- a/drivers/vdpa/virtio_pci/vp_vdpa.c
> > > +++ b/drivers/vdpa/virtio_pci/vp_vdpa.c
> > > @@ -210,13 +210,49 @@ static int vp_vdpa_get_vq_state(struct vdpa_device *vdpa, u16 qid,
> > >   	return -EOPNOTSUPP;
> > >   }
> > > +static int vp_vdpa_set_vq_state_split(struct vdpa_device *vdpa,
> > > +				      const struct vdpa_vq_state *state)
> > > +{
> > > +	const struct vdpa_vq_state_split *split = &state->split;
> > > +
> > > +	if (split->avail_index == 0)
> > > +		return 0;
> > > +
> > > +	return -EOPNOTSUPP;
> > > +}
> > > +
> > > +static int vp_vdpa_set_vq_state_packed(struct vdpa_device *vdpa,
> > > +				       const struct vdpa_vq_state *state)
> > > +{
> > > +	const struct vdpa_vq_state_packed *packed = &state->packed;
> > > +
> > > +	if (packed->last_avail_counter == 1 &&
> > Can you elaborate on the requirement on last_avail_counter and
> > last_used_counter?
> 
> 
> This is required by the virtio spec:
> 
> "
> 2.7.1 Driver and Device Ring Wrap Counters
> Each of the driver and the device are expected to maintain, internally, a
> single-bit ring wrap counter initialized to 1.
> "
> 
> For virtio-pci device, since there's no way to assign the value of those
> counters, the counters will be reset to 1 after reset, otherwise the driver
> can't work.

I see, thanks for the explanation.

> 
> Thanks
> 
> 
> > 
> > > +	    packed->last_avail_idx == 0 &&
> > > +	    packed->last_used_counter == 1 &&
> > > +	    packed->last_used_idx == 0)
> > > +		return 0;
> > > +
> > > +	return -EOPNOTSUPP;
> > > +}
> > > +
> > >   static int vp_vdpa_set_vq_state(struct vdpa_device *vdpa, u16 qid,
> > >   				const struct vdpa_vq_state *state)
> > >   {
> > > -	/* Note that this is not supported by virtio specification, so
> > > -	 * we return -ENOPOTSUPP here. This means we can't support live
> > > -	 * migration, vhost device start/stop.
> > > +	struct virtio_pci_modern_device *mdev = vdpa_to_mdev(vdpa);
> > > +
> > > +	/* Note that this is not supported by virtio specification.
> > > +	 * But if the state is by chance equal to the device initial
> > > +	 * state, we can let it go.
> > >   	 */
> > > +	if ((vp_modern_get_status(mdev) & VIRTIO_CONFIG_S_FEATURES_OK) &&
> > > +	    !vp_modern_get_queue_enable(mdev, qid)) {
> > > +		if (vp_modern_get_driver_features(mdev) &
> > > +		    BIT_ULL(VIRTIO_F_RING_PACKED))
> > > +			return vp_vdpa_set_vq_state_packed(vdpa, state);
> > > +		else
> > > +			return vp_vdpa_set_vq_state_split(vdpa,	state);
> > > +	}
> > > +
> > >   	return -EOPNOTSUPP;
> > >   }
> > > -- 
> > > 2.25.1
> > > 
>