Michael S. Tsirkin
2023-Apr-08 06:37 UTC
[PATCH] virtio-vdpa: add VIRTIO_F_NOTIFICATION_DATA feature support
On Sun, Apr 02, 2023 at 11:10:34AM +0300, Alvaro Karsz wrote:> Add VIRTIO_F_NOTIFICATION_DATA support for vDPA transport. > If this feature is negotiated, the driver passes extra data when kicking > a virtqueue. > > A device that offers this feature needs to implement the > kick_vq_with_data callback. > > kick_vq_with_data receives the vDPA device and data. > data includes the vqn, next_off and next_wrap for packed virtqueues. > > This patch follows a patch [1] by Viktor Prutyanov which adds support > for the MMIO, channel I/O and modern PCI transports. > > This patch needs to be applied on top of Viktor's patch. > > [1] https://lore.kernel.org/lkml/20230324195029.2410503-1-viktor at daynix.com/ > > Signed-off-by: Alvaro Karsz <alvaro.karsz at solid-run.com>Hmm so I conclude that drivers without kick_vq_with_data should not accept VIRTIO_F_NOTIFICATION_DATA then?> --- > drivers/virtio/virtio_vdpa.c | 20 ++++++++++++++++++-- > include/linux/vdpa.h | 6 ++++++ > 2 files changed, 24 insertions(+), 2 deletions(-) > > diff --git a/drivers/virtio/virtio_vdpa.c b/drivers/virtio/virtio_vdpa.c > index d7f5af62dda..bdaf30f7fbf 100644 > --- a/drivers/virtio/virtio_vdpa.c > +++ b/drivers/virtio/virtio_vdpa.c > @@ -112,6 +112,17 @@ static bool virtio_vdpa_notify(struct virtqueue *vq) > return true; > } > > +static bool virtio_vdpa_notify_with_data(struct virtqueue *vq) > +{ > + struct vdpa_device *vdpa = vd_get_vdpa(vq->vdev); > + const struct vdpa_config_ops *ops = vdpa->config; > + u32 data = vring_notification_data(vq); > + > + ops->kick_vq_with_data(vdpa, data); > + > + return true; > +} > + > static irqreturn_t virtio_vdpa_config_cb(void *private) > { > struct virtio_vdpa_device *vd_dev = private; > @@ -138,6 +149,7 @@ virtio_vdpa_setup_vq(struct virtio_device *vdev, unsigned int index, > struct device *dma_dev; > const struct vdpa_config_ops *ops = vdpa->config; > struct virtio_vdpa_vq_info *info; > + bool (*notify)(struct virtqueue *vq); > struct vdpa_callback cb; > struct virtqueue *vq; > u64 desc_addr, driver_addr, device_addr; > @@ -154,6 +166,11 @@ virtio_vdpa_setup_vq(struct virtio_device *vdev, unsigned int index, > if (index >= vdpa->nvqs) > return ERR_PTR(-ENOENT); > > + if (__virtio_test_bit(vdev, VIRTIO_F_NOTIFICATION_DATA)) > + notify = virtio_vdpa_notify_with_data; > + else > + notify = virtio_vdpa_notify; > + > /* Queue shouldn't already be set up. */ > if (ops->get_vq_ready(vdpa, index)) > return ERR_PTR(-ENOENT); > @@ -183,8 +200,7 @@ virtio_vdpa_setup_vq(struct virtio_device *vdev, unsigned int index, > dma_dev = vdpa_get_dma_dev(vdpa); > vq = vring_create_virtqueue_dma(index, max_num, align, vdev, > true, may_reduce_num, ctx, > - virtio_vdpa_notify, callback, > - name, dma_dev); > + notify, callback, name, dma_dev); > if (!vq) { > err = -ENOMEM; > goto error_new_virtqueue; > diff --git a/include/linux/vdpa.h b/include/linux/vdpa.h > index 43f59ef10cc..a83bb0501c5 100644 > --- a/include/linux/vdpa.h > +++ b/include/linux/vdpa.h > @@ -143,6 +143,11 @@ struct vdpa_map_file { > * @kick_vq: Kick the virtqueue > * @vdev: vdpa device > * @idx: virtqueue index > + * @kick_vq_with_data: Kick the virtqueue and supply extra data > + * (only if VIRTIO_F_NOTIFICATION_DATA is negotiated) > + * @vdev: vdpa device > + * @data: includes vqn, next_off and next_wrap for > + * packed virtqueues > * @set_vq_cb: Set the interrupt callback function for > * a virtqueue > * @vdev: vdpa device > @@ -300,6 +305,7 @@ struct vdpa_config_ops { > u64 device_area); > void (*set_vq_num)(struct vdpa_device *vdev, u16 idx, u32 num); > void (*kick_vq)(struct vdpa_device *vdev, u16 idx); > + void (*kick_vq_with_data)(struct vdpa_device *vdev, u32 data); > void (*set_vq_cb)(struct vdpa_device *vdev, u16 idx, > struct vdpa_callback *cb); > void (*set_vq_ready)(struct vdpa_device *vdev, u16 idx, bool ready); > -- > 2.34.1
Alvaro Karsz
2023-Apr-08 07:32 UTC
[PATCH] virtio-vdpa: add VIRTIO_F_NOTIFICATION_DATA feature support
> Hmm so I conclude that drivers without kick_vq_with_data > should not accept VIRTIO_F_NOTIFICATION_DATA then?Yes, vDPA drivers without kick_vq_with_data should not offer VIRTIO_F_NOTIFICATION_DATA in the first place. We can be more forgiving in such cases and clear the feature bit + print a warning. It may be easier to debug when creating a new vDPA driver.
Seemingly Similar Threads
- [PATCH] virtio-vdpa: add VIRTIO_F_NOTIFICATION_DATA feature support
- [PATCH] virtio-vdpa: add VIRTIO_F_NOTIFICATION_DATA feature support
- [PATCH] virtio-vdpa: add VIRTIO_F_NOTIFICATION_DATA feature support
- [PATCH] virtio-vdpa: add VIRTIO_F_NOTIFICATION_DATA feature support
- [PATCH v6] virtio: add VIRTIO_F_NOTIFICATION_DATA feature support