Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20230118164359.1523760-1-eperezma@redhat.com> <20230118164359.1523760-2-eperezma@redhat.com>
In-Reply-To: <20230118164359.1523760-2-eperezma@redhat.com>
From:   Jason Wang <jasowang@redhat.com>
Date:   Thu, 19 Jan 2023 11:16:13 +0800
Message-ID: <CACGkMEt=po+FMEikgj3OXC7_QC=yhRP2Sx0NU=2sFPGgQ4bujA@mail.gmail.com>
Subject: Re: [PATCH 1/2] vdpa_sim: not reset state in vdpasim_queue_ready
To:     =?UTF-8?Q?Eugenio_P=C3=A9rez?= <eperezma@redhat.com>
Cc:     mst@redhat.com, leiyang@redhat.com,
        Laurent Vivier <lvivier@redhat.com>, sgarzare@redhat.com,
        Zhu Lingshan <lingshan.zhu@intel.com>,
        virtualization@lists.linux-foundation.org, si-wei.liu@oracle.com,
        linux-kernel@vger.kernel.org, lulu@redhat.com,
        Gautam Dawar <gdawar@xilinx.com>, alvaro.karsz@solid-run.com
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Thu, Jan 19, 2023 at 12:44 AM Eugenio P=C3=A9rez <eperezma@redhat.com> w=
rote:
>
> vdpasim_queue_ready calls vringh_init_iotlb, which resets split indexes.
> But it can be called after setting a ring base with
> vdpasim_set_vq_state.
>
> Fix it by stashing them. They're still resetted in vdpasim_vq_reset.
>
> This was discovered and tested live migrating the vdpa_sim_net device.
>
> Fixes: 2c53d0f64c06 ("vdpasim: vDPA device simulator")
> Signed-off-by: Eugenio P=C3=A9rez <eperezma@redhat.com>
> ---
>  drivers/vdpa/vdpa_sim/vdpa_sim.c | 2 ++
>  1 file changed, 2 insertions(+)
>
> diff --git a/drivers/vdpa/vdpa_sim/vdpa_sim.c b/drivers/vdpa/vdpa_sim/vdp=
a_sim.c
> index cb88891b44a8..8839232a3fcb 100644
> --- a/drivers/vdpa/vdpa_sim/vdpa_sim.c
> +++ b/drivers/vdpa/vdpa_sim/vdpa_sim.c
> @@ -66,6 +66,7 @@ static void vdpasim_vq_notify(struct vringh *vring)
>  static void vdpasim_queue_ready(struct vdpasim *vdpasim, unsigned int id=
x)
>  {
>         struct vdpasim_virtqueue *vq =3D &vdpasim->vqs[idx];
> +       uint16_t last_avail_idx =3D vq->vring.last_avail_idx;
>
>         vringh_init_iotlb(&vq->vring, vdpasim->features, vq->num, false,
>                           (struct vring_desc *)(uintptr_t)vq->desc_addr,
> @@ -74,6 +75,7 @@ static void vdpasim_queue_ready(struct vdpasim *vdpasim=
, unsigned int idx)
>                           (struct vring_used *)
>                           (uintptr_t)vq->device_addr);
>
> +       vq->vring.last_avail_idx =3D last_avail_idx;

Does this need to be serialized with the datapath?

E.g in set_vq_state() we do:

spin_lock(&vdpasim->lock);
vrh->last_avail_idx =3D state->split.avail_index;
spin_unlock(&vdpasim->lock);

Thanks

>         vq->vring.notify =3D vdpasim_vq_notify;
>  }
>
> --
> 2.31.1
>