| Message ID | 20240130114224.86536-5-xuanzhuo@linux.alibaba.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | virtio: drivers maintain dma info for premapped vq | expand |
On Tue, Jan 30, 2024 at 7:42 PM Xuan Zhuo <xuanzhuo@linux.alibaba.com> wrote: > > In the functions vring_unmap_one_split and > vring_unmap_one_split_indirect, > multiple checks are made whether unmap is performed and whether it is > INDIRECT. > > These two functions are usually called in a loop, and we should put the > check outside the loop. > > And we unmap the descs with VRING_DESC_F_INDIRECT on the same path with > other descs, that make the thing more complex. If we distinguish the > descs with VRING_DESC_F_INDIRECT before unmap, thing will be clearer. > > 1. only one desc of the desc table is used, we do not need the loop > 2. the called unmap api is difference from the other desc > 3. the vq->premapped is not needed to check > 4. the vq->indirect is not needed to check > 5. the state->indir_desc must not be null > > Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com> > --- > drivers/virtio/virtio_ring.c | 80 ++++++++++++++++++------------------ > 1 file changed, 39 insertions(+), 41 deletions(-) > > diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c > index dd03bc5a81fe..2b41fdbce975 100644 > --- a/drivers/virtio/virtio_ring.c > +++ b/drivers/virtio/virtio_ring.c > @@ -452,9 +452,6 @@ static void vring_unmap_one_split_indirect(const struct vring_virtqueue *vq, > { > u16 flags; > > - if (!vring_need_unmap_buffer(vq)) > - return; > - > flags = virtio16_to_cpu(vq->vq.vdev, desc->flags); > > dma_unmap_page(vring_dma_dev(vq), > @@ -472,27 +469,12 @@ static unsigned int vring_unmap_one_split(const struct vring_virtqueue *vq, > > flags = extra[i].flags; > > - if (flags & VRING_DESC_F_INDIRECT) { > - if (!vq->use_dma_api) > - goto out; > - > - dma_unmap_single(vring_dma_dev(vq), > - extra[i].addr, > - extra[i].len, > - (flags & VRING_DESC_F_WRITE) ? > - DMA_FROM_DEVICE : DMA_TO_DEVICE); > - } else { > - if (!vring_need_unmap_buffer(vq)) > - goto out; > - > - dma_unmap_page(vring_dma_dev(vq), > - extra[i].addr, > - extra[i].len, > - (flags & VRING_DESC_F_WRITE) ? > - DMA_FROM_DEVICE : DMA_TO_DEVICE); > - } > + dma_unmap_page(vring_dma_dev(vq), > + extra[i].addr, > + extra[i].len, > + (flags & VRING_DESC_F_WRITE) ? > + DMA_FROM_DEVICE : DMA_TO_DEVICE); > > -out: > return extra[i].next; > } > > @@ -660,7 +642,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq, > vq, desc, total_sg * sizeof(struct vring_desc), > DMA_TO_DEVICE); > if (vring_mapping_error(vq, addr)) { > - if (vq->premapped) > + if (!vring_need_unmap_buffer(vq)) > goto free_indirect; > > goto unmap_release; > @@ -713,6 +695,9 @@ static inline int virtqueue_add_split(struct virtqueue *_vq, > return 0; > > unmap_release: > + > + WARN_ON(!vring_need_unmap_buffer(vq)); > + > err_idx = i; > > if (indirect) > @@ -774,34 +759,42 @@ static void detach_buf_split(struct vring_virtqueue *vq, unsigned int head, > { > unsigned int i, j; > __virtio16 nextflag = cpu_to_virtio16(vq->vq.vdev, VRING_DESC_F_NEXT); > + u16 flags; > > /* Clear data ptr. */ > vq->split.desc_state[head].data = NULL; > + flags = vq->split.desc_extra[head].flags; > > /* Put back on free list: unmap first-level descriptors and find end */ > i = head; > > - while (vq->split.vring.desc[i].flags & nextflag) { > - vring_unmap_one_split(vq, i); > - i = vq->split.desc_extra[i].next; > - vq->vq.num_free++; > - } > - > - vring_unmap_one_split(vq, i); > - vq->split.desc_extra[i].next = vq->free_head; > - vq->free_head = head; > + if (!(flags & VRING_DESC_F_INDIRECT)) { So during add we do: if (!indirect && vring_need_unmap_buffer(vq)) vq->split.desc_extra[prev & (vq->split.vring.num - 1)].flags &= ~VRING_DESC_F_NEXT; Then using flags here unconditionally seems not reliable. I post a patch to store flags unconditionally at: https://lore.kernel.org/all/20220224122655-mutt-send-email-mst@kernel.org/ > + while (vq->split.vring.desc[i].flags & nextflag) { > + if (vring_need_unmap_buffer(vq)) > + vring_unmap_one_split(vq, i); > + i = vq->split.desc_extra[i].next; > + vq->vq.num_free++; > + } > > - /* Plus final descriptor */ > - vq->vq.num_free++; > + if (vring_need_unmap_buffer(vq)) > + vring_unmap_one_split(vq, i); > > - if (vq->indirect) { > + if (ctx) > + *ctx = vq->split.desc_state[head].indir_desc; > + } else { > struct vring_desc *indir_desc = > vq->split.desc_state[head].indir_desc; > u32 len; > > - /* Free the indirect table, if any, now that it's unmapped. */ > - if (!indir_desc) > - return; > + if (vq->use_dma_api) { > + struct vring_desc_extra *extra = vq->split.desc_extra; > + > + dma_unmap_single(vring_dma_dev(vq), > + extra[i].addr, > + extra[i].len, > + (flags & VRING_DESC_F_WRITE) ? > + DMA_FROM_DEVICE : DMA_TO_DEVICE); > + } Note that there's a following BUG_ON(!(vq->split.desc_extra[head].flags & VRING_DESC_F_INDIRECT)); Which I think we can remove. Thanks
On Wed, Jan 31, 2024 at 05:12:22PM +0800, Jason Wang wrote: > I post a patch to store flags unconditionally at: > > https://lore.kernel.org/all/20220224122655-mutt-send-email-mst@kernel.org/ what happened to it btw?
On Wed, Jan 31, 2024 at 5:43 PM Michael S. Tsirkin <mst@redhat.com> wrote: > > On Wed, Jan 31, 2024 at 05:12:22PM +0800, Jason Wang wrote: > > I post a patch to store flags unconditionally at: > > > > https://lore.kernel.org/all/20220224122655-mutt-send-email-mst@kernel.org/ > > what happened to it btw? Haven't got time for a benchmark. We can evaluate whether we still need it here and do the benchmark if yes. Thanks > > -- > MST >
On Wed, 31 Jan 2024 17:12:22 +0800, Jason Wang <jasowang@redhat.com> wrote: > On Tue, Jan 30, 2024 at 7:42 PM Xuan Zhuo <xuanzhuo@linux.alibaba.com> wrote: > > > > In the functions vring_unmap_one_split and > > vring_unmap_one_split_indirect, > > multiple checks are made whether unmap is performed and whether it is > > INDIRECT. > > > > These two functions are usually called in a loop, and we should put the > > check outside the loop. > > > > And we unmap the descs with VRING_DESC_F_INDIRECT on the same path with > > other descs, that make the thing more complex. If we distinguish the > > descs with VRING_DESC_F_INDIRECT before unmap, thing will be clearer. > > > > 1. only one desc of the desc table is used, we do not need the loop > > 2. the called unmap api is difference from the other desc > > 3. the vq->premapped is not needed to check > > 4. the vq->indirect is not needed to check > > 5. the state->indir_desc must not be null > > > > Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com> > > --- > > drivers/virtio/virtio_ring.c | 80 ++++++++++++++++++------------------ > > 1 file changed, 39 insertions(+), 41 deletions(-) > > > > diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c > > index dd03bc5a81fe..2b41fdbce975 100644 > > --- a/drivers/virtio/virtio_ring.c > > +++ b/drivers/virtio/virtio_ring.c > > @@ -452,9 +452,6 @@ static void vring_unmap_one_split_indirect(const struct vring_virtqueue *vq, > > { > > u16 flags; > > > > - if (!vring_need_unmap_buffer(vq)) > > - return; > > - > > flags = virtio16_to_cpu(vq->vq.vdev, desc->flags); > > > > dma_unmap_page(vring_dma_dev(vq), > > @@ -472,27 +469,12 @@ static unsigned int vring_unmap_one_split(const struct vring_virtqueue *vq, > > > > flags = extra[i].flags; > > > > - if (flags & VRING_DESC_F_INDIRECT) { > > - if (!vq->use_dma_api) > > - goto out; > > - > > - dma_unmap_single(vring_dma_dev(vq), > > - extra[i].addr, > > - extra[i].len, > > - (flags & VRING_DESC_F_WRITE) ? > > - DMA_FROM_DEVICE : DMA_TO_DEVICE); > > - } else { > > - if (!vring_need_unmap_buffer(vq)) > > - goto out; > > - > > - dma_unmap_page(vring_dma_dev(vq), > > - extra[i].addr, > > - extra[i].len, > > - (flags & VRING_DESC_F_WRITE) ? > > - DMA_FROM_DEVICE : DMA_TO_DEVICE); > > - } > > + dma_unmap_page(vring_dma_dev(vq), > > + extra[i].addr, > > + extra[i].len, > > + (flags & VRING_DESC_F_WRITE) ? > > + DMA_FROM_DEVICE : DMA_TO_DEVICE); > > > > -out: > > return extra[i].next; > > } > > > > @@ -660,7 +642,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq, > > vq, desc, total_sg * sizeof(struct vring_desc), > > DMA_TO_DEVICE); > > if (vring_mapping_error(vq, addr)) { > > - if (vq->premapped) > > + if (!vring_need_unmap_buffer(vq)) > > goto free_indirect; > > > > goto unmap_release; > > @@ -713,6 +695,9 @@ static inline int virtqueue_add_split(struct virtqueue *_vq, > > return 0; > > > > unmap_release: > > + > > + WARN_ON(!vring_need_unmap_buffer(vq)); > > + > > err_idx = i; > > > > if (indirect) > > @@ -774,34 +759,42 @@ static void detach_buf_split(struct vring_virtqueue *vq, unsigned int head, > > { > > unsigned int i, j; > > __virtio16 nextflag = cpu_to_virtio16(vq->vq.vdev, VRING_DESC_F_NEXT); > > + u16 flags; > > > > /* Clear data ptr. */ > > vq->split.desc_state[head].data = NULL; > > + flags = vq->split.desc_extra[head].flags; > > > > /* Put back on free list: unmap first-level descriptors and find end */ > > i = head; > > > > - while (vq->split.vring.desc[i].flags & nextflag) { > > - vring_unmap_one_split(vq, i); > > - i = vq->split.desc_extra[i].next; > > - vq->vq.num_free++; > > - } > > - > > - vring_unmap_one_split(vq, i); > > - vq->split.desc_extra[i].next = vq->free_head; > > - vq->free_head = head; > > + if (!(flags & VRING_DESC_F_INDIRECT)) { > > So during add we do: > > if (!indirect && vring_need_unmap_buffer(vq)) > vq->split.desc_extra[prev & (vq->split.vring.num - 1)].flags &= > ~VRING_DESC_F_NEXT; This does not affect this patch. 1. this just considers the VRING_DESC_F_NEXT of desc_extra.flags 2. the desc_extra.flags is updated by virtqueue_add_desc_split() So for desc_extra.flags & VRING_DESC_F_INDIRECT, that is right. Thanks. > > Then using flags here unconditionally seems not reliable. > > I post a patch to store flags unconditionally at: > > https://lore.kernel.org/all/20220224122655-mutt-send-email-mst@kernel.org/ > > > + while (vq->split.vring.desc[i].flags & nextflag) { > > + if (vring_need_unmap_buffer(vq)) > > + vring_unmap_one_split(vq, i); > > + i = vq->split.desc_extra[i].next; > > + vq->vq.num_free++; > > + } > > > > - /* Plus final descriptor */ > > - vq->vq.num_free++; > > + if (vring_need_unmap_buffer(vq)) > > + vring_unmap_one_split(vq, i); > > > > - if (vq->indirect) { > > + if (ctx) > > + *ctx = vq->split.desc_state[head].indir_desc; > > + } else { > > struct vring_desc *indir_desc = > > vq->split.desc_state[head].indir_desc; > > u32 len; > > > > - /* Free the indirect table, if any, now that it's unmapped. */ > > - if (!indir_desc) > > - return; > > + if (vq->use_dma_api) { > > + struct vring_desc_extra *extra = vq->split.desc_extra; > > + > > + dma_unmap_single(vring_dma_dev(vq), > > + extra[i].addr, > > + extra[i].len, > > + (flags & VRING_DESC_F_WRITE) ? > > + DMA_FROM_DEVICE : DMA_TO_DEVICE); > > + } > > Note that there's a following > > BUG_ON(!(vq->split.desc_extra[head].flags & > VRING_DESC_F_INDIRECT)); > > Which I think we can remove. > > Thanks >
diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c index dd03bc5a81fe..2b41fdbce975 100644 --- a/drivers/virtio/virtio_ring.c +++ b/drivers/virtio/virtio_ring.c @@ -452,9 +452,6 @@ static void vring_unmap_one_split_indirect(const struct vring_virtqueue *vq, { u16 flags; - if (!vring_need_unmap_buffer(vq)) - return; - flags = virtio16_to_cpu(vq->vq.vdev, desc->flags); dma_unmap_page(vring_dma_dev(vq), @@ -472,27 +469,12 @@ static unsigned int vring_unmap_one_split(const struct vring_virtqueue *vq, flags = extra[i].flags; - if (flags & VRING_DESC_F_INDIRECT) { - if (!vq->use_dma_api) - goto out; - - dma_unmap_single(vring_dma_dev(vq), - extra[i].addr, - extra[i].len, - (flags & VRING_DESC_F_WRITE) ? - DMA_FROM_DEVICE : DMA_TO_DEVICE); - } else { - if (!vring_need_unmap_buffer(vq)) - goto out; - - dma_unmap_page(vring_dma_dev(vq), - extra[i].addr, - extra[i].len, - (flags & VRING_DESC_F_WRITE) ? - DMA_FROM_DEVICE : DMA_TO_DEVICE); - } + dma_unmap_page(vring_dma_dev(vq), + extra[i].addr, + extra[i].len, + (flags & VRING_DESC_F_WRITE) ? + DMA_FROM_DEVICE : DMA_TO_DEVICE); -out: return extra[i].next; } @@ -660,7 +642,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq, vq, desc, total_sg * sizeof(struct vring_desc), DMA_TO_DEVICE); if (vring_mapping_error(vq, addr)) { - if (vq->premapped) + if (!vring_need_unmap_buffer(vq)) goto free_indirect; goto unmap_release; @@ -713,6 +695,9 @@ static inline int virtqueue_add_split(struct virtqueue *_vq, return 0; unmap_release: + + WARN_ON(!vring_need_unmap_buffer(vq)); + err_idx = i; if (indirect) @@ -774,34 +759,42 @@ static void detach_buf_split(struct vring_virtqueue *vq, unsigned int head, { unsigned int i, j; __virtio16 nextflag = cpu_to_virtio16(vq->vq.vdev, VRING_DESC_F_NEXT); + u16 flags; /* Clear data ptr. */ vq->split.desc_state[head].data = NULL; + flags = vq->split.desc_extra[head].flags; /* Put back on free list: unmap first-level descriptors and find end */ i = head; - while (vq->split.vring.desc[i].flags & nextflag) { - vring_unmap_one_split(vq, i); - i = vq->split.desc_extra[i].next; - vq->vq.num_free++; - } - - vring_unmap_one_split(vq, i); - vq->split.desc_extra[i].next = vq->free_head; - vq->free_head = head; + if (!(flags & VRING_DESC_F_INDIRECT)) { + while (vq->split.vring.desc[i].flags & nextflag) { + if (vring_need_unmap_buffer(vq)) + vring_unmap_one_split(vq, i); + i = vq->split.desc_extra[i].next; + vq->vq.num_free++; + } - /* Plus final descriptor */ - vq->vq.num_free++; + if (vring_need_unmap_buffer(vq)) + vring_unmap_one_split(vq, i); - if (vq->indirect) { + if (ctx) + *ctx = vq->split.desc_state[head].indir_desc; + } else { struct vring_desc *indir_desc = vq->split.desc_state[head].indir_desc; u32 len; - /* Free the indirect table, if any, now that it's unmapped. */ - if (!indir_desc) - return; + if (vq->use_dma_api) { + struct vring_desc_extra *extra = vq->split.desc_extra; + + dma_unmap_single(vring_dma_dev(vq), + extra[i].addr, + extra[i].len, + (flags & VRING_DESC_F_WRITE) ? + DMA_FROM_DEVICE : DMA_TO_DEVICE); + } len = vq->split.desc_extra[head].len; @@ -816,9 +809,14 @@ static void detach_buf_split(struct vring_virtqueue *vq, unsigned int head, kfree(indir_desc); vq->split.desc_state[head].indir_desc = NULL; - } else if (ctx) { - *ctx = vq->split.desc_state[head].indir_desc; + } + + vq->split.desc_extra[i].next = vq->free_head; + vq->free_head = head; + + /* Plus final descriptor */ + vq->vq.num_free++; } static bool more_used_split(const struct vring_virtqueue *vq)
In the functions vring_unmap_one_split and vring_unmap_one_split_indirect, multiple checks are made whether unmap is performed and whether it is INDIRECT. These two functions are usually called in a loop, and we should put the check outside the loop. And we unmap the descs with VRING_DESC_F_INDIRECT on the same path with other descs, that make the thing more complex. If we distinguish the descs with VRING_DESC_F_INDIRECT before unmap, thing will be clearer. 1. only one desc of the desc table is used, we do not need the loop 2. the called unmap api is difference from the other desc 3. the vq->premapped is not needed to check 4. the vq->indirect is not needed to check 5. the state->indir_desc must not be null Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com> --- drivers/virtio/virtio_ring.c | 80 ++++++++++++++++++------------------ 1 file changed, 39 insertions(+), 41 deletions(-)