[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20231130050852-mutt-send-email-mst@kernel.org>
Date: Thu, 30 Nov 2023 05:10:42 -0500
From: "Michael S. Tsirkin" <mst@...hat.com>
To: Xuan Zhuo <xuanzhuo@...ux.alibaba.com>
Cc: virtualization@...ts.linux-foundation.org,
Jason Wang <jasowang@...hat.com>,
"David S. Miller" <davem@...emloft.net>,
Eric Dumazet <edumazet@...gle.com>,
Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
Alexei Starovoitov <ast@...nel.org>,
Daniel Borkmann <daniel@...earbox.net>,
Jesper Dangaard Brouer <hawk@...nel.org>,
John Fastabend <john.fastabend@...il.com>, netdev@...r.kernel.org,
bpf@...r.kernel.org, Christoph Hellwig <hch@...radead.org>
Subject: Re: [PATCH vhost v13 11/12] virtio_ring: introduce dma sync api for
virtqueue
On Thu, Nov 30, 2023 at 05:49:38PM +0800, Xuan Zhuo wrote:
> On Thu, 30 Nov 2023 04:45:58 -0500, "Michael S. Tsirkin" <mst@...hat.com> wrote:
> > On Thu, Aug 10, 2023 at 08:30:56PM +0800, Xuan Zhuo wrote:
> > > These API has been introduced:
> > >
> > > * virtqueue_dma_need_sync
> > > * virtqueue_dma_sync_single_range_for_cpu
> > > * virtqueue_dma_sync_single_range_for_device
> > >
> > > These APIs can be used together with the premapped mechanism to sync the
> > > DMA address.
> > >
> > > Signed-off-by: Xuan Zhuo <xuanzhuo@...ux.alibaba.com>
> > > ---
> > > drivers/virtio/virtio_ring.c | 76 ++++++++++++++++++++++++++++++++++++
> > > include/linux/virtio.h | 8 ++++
> > > 2 files changed, 84 insertions(+)
> > >
> > > diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
> > > index 916479c9c72c..81ecb29c88f1 100644
> > > --- a/drivers/virtio/virtio_ring.c
> > > +++ b/drivers/virtio/virtio_ring.c
> > > @@ -3175,4 +3175,80 @@ int virtqueue_dma_mapping_error(struct virtqueue *_vq, dma_addr_t addr)
> > > }
> > > EXPORT_SYMBOL_GPL(virtqueue_dma_mapping_error);
> > >
> > > +/**
> > > + * virtqueue_dma_need_sync - check a dma address needs sync
> > > + * @_vq: the struct virtqueue we're talking about.
> > > + * @addr: DMA address
> > > + *
> > > + * Check if the dma address mapped by the virtqueue_dma_map_* APIs needs to be
> > > + * synchronized
> > > + *
> > > + * return bool
> > > + */
> > > +bool virtqueue_dma_need_sync(struct virtqueue *_vq, dma_addr_t addr)
> > > +{
> > > + struct vring_virtqueue *vq = to_vvq(_vq);
> > > +
> > > + if (!vq->use_dma_api)
> > > + return false;
> > > +
> > > + return dma_need_sync(vring_dma_dev(vq), addr);
> > > +}
> > > +EXPORT_SYMBOL_GPL(virtqueue_dma_need_sync);
> > > +
> > > +/**
> > > + * virtqueue_dma_sync_single_range_for_cpu - dma sync for cpu
> > > + * @_vq: the struct virtqueue we're talking about.
> > > + * @addr: DMA address
> > > + * @offset: DMA address offset
> > > + * @size: buf size for sync
> > > + * @dir: DMA direction
> > > + *
> > > + * Before calling this function, use virtqueue_dma_need_sync() to confirm that
> > > + * the DMA address really needs to be synchronized
> > > + *
> > > + */
> > > +void virtqueue_dma_sync_single_range_for_cpu(struct virtqueue *_vq,
> > > + dma_addr_t addr,
> > > + unsigned long offset, size_t size,
> > > + enum dma_data_direction dir)
> > > +{
> > > + struct vring_virtqueue *vq = to_vvq(_vq);
> > > + struct device *dev = vring_dma_dev(vq);
> > > +
> > > + if (!vq->use_dma_api)
> > > + return;
> > > +
> > > + dma_sync_single_range_for_cpu(dev, addr, offset, size,
> > > + DMA_BIDIRECTIONAL);
> > > +}
> >
> >
> > Why did you use DMA_BIDIRECTIONAL here?
> > Why is "dir" ignored?
>
> This is a mistake.
>
> I see Jason has a fix patch.
the one he sent in response to the bug report? it's incomplete though.
> How can I help?
>
> Thanks.
develop a full patch with commit log, explanation of what
the result of the bug is, Fixes tag etc, test
in some environment where dir makes a difference and post.
>
> >
> >
> > > +EXPORT_SYMBOL_GPL(virtqueue_dma_sync_single_range_for_cpu);
> > > +
> > > +/**
> > > + * virtqueue_dma_sync_single_range_for_device - dma sync for device
> > > + * @_vq: the struct virtqueue we're talking about.
> > > + * @addr: DMA address
> > > + * @offset: DMA address offset
> > > + * @size: buf size for sync
> > > + * @dir: DMA direction
> > > + *
> > > + * Before calling this function, use virtqueue_dma_need_sync() to confirm that
> > > + * the DMA address really needs to be synchronized
> > > + */
> > > +void virtqueue_dma_sync_single_range_for_device(struct virtqueue *_vq,
> > > + dma_addr_t addr,
> > > + unsigned long offset, size_t size,
> > > + enum dma_data_direction dir)
> > > +{
> > > + struct vring_virtqueue *vq = to_vvq(_vq);
> > > + struct device *dev = vring_dma_dev(vq);
> > > +
> > > + if (!vq->use_dma_api)
> > > + return;
> > > +
> > > + dma_sync_single_range_for_device(dev, addr, offset, size,
> > > + DMA_BIDIRECTIONAL);
> > > +}
> > > +EXPORT_SYMBOL_GPL(virtqueue_dma_sync_single_range_for_device);
> > > +
> > > MODULE_LICENSE("GPL");
> >
> > same question here.
> >
> > > diff --git a/include/linux/virtio.h b/include/linux/virtio.h
> > > index 79e3c74391e0..1311a7fbe675 100644
> > > --- a/include/linux/virtio.h
> > > +++ b/include/linux/virtio.h
> > > @@ -220,4 +220,12 @@ void virtqueue_dma_unmap_single_attrs(struct virtqueue *_vq, dma_addr_t addr,
> > > size_t size, enum dma_data_direction dir,
> > > unsigned long attrs);
> > > int virtqueue_dma_mapping_error(struct virtqueue *_vq, dma_addr_t addr);
> > > +
> > > +bool virtqueue_dma_need_sync(struct virtqueue *_vq, dma_addr_t addr);
> > > +void virtqueue_dma_sync_single_range_for_cpu(struct virtqueue *_vq, dma_addr_t addr,
> > > + unsigned long offset, size_t size,
> > > + enum dma_data_direction dir);
> > > +void virtqueue_dma_sync_single_range_for_device(struct virtqueue *_vq, dma_addr_t addr,
> > > + unsigned long offset, size_t size,
> > > + enum dma_data_direction dir);
> > > #endif /* _LINUX_VIRTIO_H */
> > > --
> > > 2.32.0.3.g01195cf9f
> >
> >
Powered by blists - more mailing lists