Jason Wang
2022-Feb-23 07:50 UTC
[PATCH] virtio_ring: aovid reading flag from the descriptor ring
On Wed, Feb 23, 2022 at 3:34 PM Jason Wang <jasowang at redhat.com> wrote:> > On Wed, Feb 23, 2022 at 3:08 PM Michael S. Tsirkin <mst at redhat.com> wrote: > > > > On Wed, Feb 23, 2022 at 11:19:03AM +0800, Jason Wang wrote: > > > On Mon, Nov 8, 2021 at 4:13 PM Jason Wang <jasowang at redhat.com> wrote: > > > > > > > > Commit 72b5e8958738 ("virtio-ring: store DMA metadata in desc_extra > > > > for split virtqueue") tries to make it possible for the driver to not > > > > read from the descriptor ring to prevent the device from corrupting > > > > the descriptor ring. But it still read the descriptor flag from the > > > > descriptor ring during buffer detach. > > > > > > > > This patch fixes by always store the descriptor flag no matter whether > > > > DMA API is used and then we can avoid reading descriptor flag from the > > > > descriptor ring. This eliminates the possibly of unexpected next > > > > descriptor caused by the wrong flag (e.g the next flag). > > > > > > > > Signed-off-by: Jason Wang <jasowang at redhat.com> > > > > > > Michael, any comment for this? > > > > > > Thanks > > > > I don't exactly see why we should care without DMA API, it seems > > cleaner not to poke at the array one extra time. > > I think the answer is that we have any special care about the DMA APII meant "we haven't had" actually. Thanks> for all other places that are using desc_extra. > > Thanks > > > > > > > > --- > > > > drivers/virtio/virtio_ring.c | 4 ++-- > > > > 1 file changed, 2 insertions(+), 2 deletions(-) > > > > > > > > diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c > > > > index 00f64f2f8b72..28734f4e57d3 100644 > > > > --- a/drivers/virtio/virtio_ring.c > > > > +++ b/drivers/virtio/virtio_ring.c > > > > @@ -583,7 +583,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq, > > > > } > > > > /* Last one doesn't continue. */ > > > > desc[prev].flags &= cpu_to_virtio16(_vq->vdev, ~VRING_DESC_F_NEXT); > > > > - if (!indirect && vq->use_dma_api) > > > > + if (!indirect) > > > > vq->split.desc_extra[prev & (vq->split.vring.num - 1)].flags &> > > > ~VRING_DESC_F_NEXT; > > > > > > > > @@ -713,7 +713,7 @@ static void detach_buf_split(struct vring_virtqueue *vq, unsigned int head, > > > > /* Put back on free list: unmap first-level descriptors and find end */ > > > > i = head; > > > > > > > > - while (vq->split.vring.desc[i].flags & nextflag) { > > > > + while (vq->split.desc_extra[i].flags & nextflag) { > > > > vring_unmap_one_split(vq, i); > > > > i = vq->split.desc_extra[i].next; > > > > vq->vq.num_free++; > > > > -- > > > > 2.25.1 > > > > > >
Michael S. Tsirkin
2022-Feb-24 17:26 UTC
[PATCH] virtio_ring: aovid reading flag from the descriptor ring
On Wed, Feb 23, 2022 at 03:50:07PM +0800, Jason Wang wrote:> On Wed, Feb 23, 2022 at 3:34 PM Jason Wang <jasowang at redhat.com> wrote: > > > > On Wed, Feb 23, 2022 at 3:08 PM Michael S. Tsirkin <mst at redhat.com> wrote: > > > > > > On Wed, Feb 23, 2022 at 11:19:03AM +0800, Jason Wang wrote: > > > > On Mon, Nov 8, 2021 at 4:13 PM Jason Wang <jasowang at redhat.com> wrote: > > > > > > > > > > Commit 72b5e8958738 ("virtio-ring: store DMA metadata in desc_extra > > > > > for split virtqueue") tries to make it possible for the driver to not > > > > > read from the descriptor ring to prevent the device from corrupting > > > > > the descriptor ring. But it still read the descriptor flag from the > > > > > descriptor ring during buffer detach. > > > > > > > > > > This patch fixes by always store the descriptor flag no matter whether > > > > > DMA API is used and then we can avoid reading descriptor flag from the > > > > > descriptor ring. This eliminates the possibly of unexpected next > > > > > descriptor caused by the wrong flag (e.g the next flag). > > > > > > > > > > Signed-off-by: Jason Wang <jasowang at redhat.com> > > > > > > > > Michael, any comment for this? > > > > > > > > Thanks > > > > > > I don't exactly see why we should care without DMA API, it seems > > > cleaner not to poke at the array one extra time. > > > > I think the answer is that we have any special care about the DMA API > > I meant "we haven't had" actually. > > ThanksI'm just asking what's better for performance. An extra write in the first chunk has a cost. Want to test and see?> > for all other places that are using desc_extra. > > > > Thanks > > > > > > > > > > > > --- > > > > > drivers/virtio/virtio_ring.c | 4 ++-- > > > > > 1 file changed, 2 insertions(+), 2 deletions(-) > > > > > > > > > > diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c > > > > > index 00f64f2f8b72..28734f4e57d3 100644 > > > > > --- a/drivers/virtio/virtio_ring.c > > > > > +++ b/drivers/virtio/virtio_ring.c > > > > > @@ -583,7 +583,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq, > > > > > } > > > > > /* Last one doesn't continue. */ > > > > > desc[prev].flags &= cpu_to_virtio16(_vq->vdev, ~VRING_DESC_F_NEXT); > > > > > - if (!indirect && vq->use_dma_api) > > > > > + if (!indirect) > > > > > vq->split.desc_extra[prev & (vq->split.vring.num - 1)].flags &> > > > > ~VRING_DESC_F_NEXT; > > > > > > > > > > @@ -713,7 +713,7 @@ static void detach_buf_split(struct vring_virtqueue *vq, unsigned int head, > > > > > /* Put back on free list: unmap first-level descriptors and find end */ > > > > > i = head; > > > > > > > > > > - while (vq->split.vring.desc[i].flags & nextflag) { > > > > > + while (vq->split.desc_extra[i].flags & nextflag) { > > > > > vring_unmap_one_split(vq, i); > > > > > i = vq->split.desc_extra[i].next; > > > > > vq->vq.num_free++; > > > > > -- > > > > > 2.25.1 > > > > > > > >