Xuan Zhuo
2023-Mar-08 06:44 UTC
[PATCH vhost v2 04/12] virtio_ring: split: support premapped
virtio core only supports virtual addresses, dma is completed in virtio
core.
In some scenarios (such as the AF_XDP), the memory is allocated
and DMA is completed in advance, so it is necessary for us to support
passing the DMA address to virtio core.
Drives can use sg->dma_address to pass the mapped dma address to virtio
core. If one sg->dma_address is used then all sgs must use sg->
dma_address, otherwise all dma_address must be null.
On the non-indirect path, if dma_address is used, extra.addr will be
set to DMA_MAPPING_ERROR. So when do unmap, we can pass it.
Signed-off-by: Xuan Zhuo <xuanzhuo at linux.alibaba.com>
---
drivers/virtio/virtio_ring.c | 22 ++++++++++++++--------
1 file changed, 14 insertions(+), 8 deletions(-)
diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
index 221ff54fe58b..61deaf0a4faf 100644
--- a/drivers/virtio/virtio_ring.c
+++ b/drivers/virtio/virtio_ring.c
@@ -457,6 +457,9 @@ static unsigned int vring_unmap_one_split(const struct
vring_virtqueue *vq,
(flags & VRING_DESC_F_WRITE) ?
DMA_FROM_DEVICE : DMA_TO_DEVICE);
} else {
+ if (extra[i].addr == DMA_MAPPING_ERROR)
+ goto out;
+
dma_unmap_page(vring_dma_dev(vq),
extra[i].addr,
extra[i].len,
@@ -497,7 +500,8 @@ static inline unsigned int virtqueue_add_desc_split(struct
virtqueue *vq,
dma_addr_t addr,
unsigned int len,
u16 flags,
- bool indirect)
+ bool indirect,
+ bool do_map)
{
struct vring_virtqueue *vring = to_vvq(vq);
struct vring_desc_extra *extra = vring->split.desc_extra;
@@ -511,7 +515,7 @@ static inline unsigned int virtqueue_add_desc_split(struct
virtqueue *vq,
next = extra[i].next;
desc[i].next = cpu_to_virtio16(vq->vdev, next);
- extra[i].addr = addr;
+ extra[i].addr = do_map ? addr : DMA_MAPPING_ERROR;
extra[i].len = len;
extra[i].flags = flags;
} else
@@ -604,7 +608,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq,
struct scatterlist *sg;
struct vring_desc *desc;
unsigned int i, n, avail, descs_used, prev;
- bool indirect;
+ bool indirect, do_map;
int head;
START_USE(vq);
@@ -657,7 +661,8 @@ static inline int virtqueue_add_split(struct virtqueue *_vq,
return -ENOSPC;
}
- if (virtqueue_map_sgs(vq, sgs, total_sg, out_sgs, in_sgs))
+ do_map = !sgs[0]->dma_address;
+ if (do_map && virtqueue_map_sgs(vq, sgs, total_sg, out_sgs, in_sgs))
return -ENOMEM;
for (n = 0; n < out_sgs; n++) {
@@ -670,7 +675,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq,
sg->dma_address,
sg->length,
VRING_DESC_F_NEXT,
- indirect);
+ indirect, do_map);
}
}
for (; n < (out_sgs + in_sgs); n++) {
@@ -684,7 +689,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq,
sg->length,
VRING_DESC_F_NEXT |
VRING_DESC_F_WRITE,
- indirect);
+ indirect, do_map);
}
}
/* Last one doesn't continue. */
@@ -705,7 +710,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq,
head, addr,
total_sg * sizeof(struct vring_desc),
VRING_DESC_F_INDIRECT,
- false);
+ false, true);
}
/* We're using some buffers from the free list. */
@@ -748,7 +753,8 @@ static inline int virtqueue_add_split(struct virtqueue *_vq,
return 0;
unmap_release:
- virtqueue_unmap_sgs(vq, sgs, total_sg, out_sgs, in_sgs);
+ if (do_map)
+ virtqueue_unmap_sgs(vq, sgs, total_sg, out_sgs, in_sgs);
if (indirect)
kfree(desc);
--
2.32.0.3.g01195cf9f
Jason Wang
2023-Mar-14 07:32 UTC
[PATCH vhost v2 04/12] virtio_ring: split: support premapped
On Wed, Mar 8, 2023 at 2:44?PM Xuan Zhuo <xuanzhuo at linux.alibaba.com> wrote:> > virtio core only supports virtual addresses, dma is completed in virtio > core. > > In some scenarios (such as the AF_XDP), the memory is allocated > and DMA is completed in advance, so it is necessary for us to supportShould be "DMA mapping is completed".> passing the DMA address to virtio core. > > Drives can use sg->dma_address to pass the mapped dma address to virtio > core. If one sg->dma_address is used then all sgs must use sg-> > dma_address, otherwise all dma_address must be null."must be null when passing it to virtqueue_add_sgs()"?> > On the non-indirect path,Should it be "direct desc path"?> if dma_address is used, extra.addr will be > set to DMA_MAPPING_ERROR. So when do unmap, we can pass it.I don't get the meaning of "pass" here. Or do you mean DMA_MAPING_ERROR is a hint that the desc is mapped by the upper layer instead of the virtio core?> > Signed-off-by: Xuan Zhuo <xuanzhuo at linux.alibaba.com> > --- > drivers/virtio/virtio_ring.c | 22 ++++++++++++++-------- > 1 file changed, 14 insertions(+), 8 deletions(-) > > diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c > index 221ff54fe58b..61deaf0a4faf 100644 > --- a/drivers/virtio/virtio_ring.c > +++ b/drivers/virtio/virtio_ring.c > @@ -457,6 +457,9 @@ static unsigned int vring_unmap_one_split(const struct vring_virtqueue *vq, > (flags & VRING_DESC_F_WRITE) ? > DMA_FROM_DEVICE : DMA_TO_DEVICE); > } else { > + if (extra[i].addr == DMA_MAPPING_ERROR) > + goto out; > + > dma_unmap_page(vring_dma_dev(vq), > extra[i].addr, > extra[i].len, > @@ -497,7 +500,8 @@ static inline unsigned int virtqueue_add_desc_split(struct virtqueue *vq, > dma_addr_t addr, > unsigned int len, > u16 flags, > - bool indirect) > + bool indirect, > + bool do_map) > { > struct vring_virtqueue *vring = to_vvq(vq); > struct vring_desc_extra *extra = vring->split.desc_extra; > @@ -511,7 +515,7 @@ static inline unsigned int virtqueue_add_desc_split(struct virtqueue *vq, > next = extra[i].next; > desc[i].next = cpu_to_virtio16(vq->vdev, next); > > - extra[i].addr = addr; > + extra[i].addr = do_map ? addr : DMA_MAPPING_ERROR;Any reason we don't need to do the same for indirect descriptors? Thanks> extra[i].len = len; > extra[i].flags = flags; > } else > @@ -604,7 +608,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq, > struct scatterlist *sg; > struct vring_desc *desc; > unsigned int i, n, avail, descs_used, prev; > - bool indirect; > + bool indirect, do_map; > int head; > > START_USE(vq); > @@ -657,7 +661,8 @@ static inline int virtqueue_add_split(struct virtqueue *_vq, > return -ENOSPC; > } > > - if (virtqueue_map_sgs(vq, sgs, total_sg, out_sgs, in_sgs)) > + do_map = !sgs[0]->dma_address; > + if (do_map && virtqueue_map_sgs(vq, sgs, total_sg, out_sgs, in_sgs)) > return -ENOMEM; > > for (n = 0; n < out_sgs; n++) { > @@ -670,7 +675,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq, > sg->dma_address, > sg->length, > VRING_DESC_F_NEXT, > - indirect); > + indirect, do_map); > } > } > for (; n < (out_sgs + in_sgs); n++) { > @@ -684,7 +689,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq, > sg->length, > VRING_DESC_F_NEXT | > VRING_DESC_F_WRITE, > - indirect); > + indirect, do_map); > } > } > /* Last one doesn't continue. */ > @@ -705,7 +710,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq, > head, addr, > total_sg * sizeof(struct vring_desc), > VRING_DESC_F_INDIRECT, > - false); > + false, true); > } > > /* We're using some buffers from the free list. */ > @@ -748,7 +753,8 @@ static inline int virtqueue_add_split(struct virtqueue *_vq, > return 0; > > unmap_release: > - virtqueue_unmap_sgs(vq, sgs, total_sg, out_sgs, in_sgs); > + if (do_map) > + virtqueue_unmap_sgs(vq, sgs, total_sg, out_sgs, in_sgs); > > if (indirect) > kfree(desc); > -- > 2.32.0.3.g01195cf9f >