search for: vhost_get_avail_size

...vhost/vhost.c b/drivers/vhost/vhost.c > index 2025543..1015464 100644 > --- a/drivers/vhost/vhost.c > +++ b/drivers/vhost/vhost.c > @@ -413,6 +413,27 @@ static void vhost_dev_free_iovecs(struct vhost_dev *dev) > vhost_vq_free_iovecs(dev->vqs[i]); > } > > +static size_t vhost_get_avail_size(struct vhost_virtqueue *vq, int num) Nit: Any reason not to make `num` unsigned or size_t? > +{ > + size_t event = vhost_has_feature(vq, VIRTIO_RING_F_EVENT_IDX) ? 2 : 0; > + > + return sizeof(*vq->avail) + > + sizeof(*vq->avail->ring) * num + event; > +} > +...

...vhost/vhost.c b/drivers/vhost/vhost.c > index 2025543..1015464 100644 > --- a/drivers/vhost/vhost.c > +++ b/drivers/vhost/vhost.c > @@ -413,6 +413,27 @@ static void vhost_dev_free_iovecs(struct vhost_dev *dev) > vhost_vq_free_iovecs(dev->vqs[i]); > } > > +static size_t vhost_get_avail_size(struct vhost_virtqueue *vq, int num) Nit: Any reason not to make `num` unsigned or size_t? > +{ > + size_t event = vhost_has_feature(vq, VIRTIO_RING_F_EVENT_IDX) ? 2 : 0; > + > + return sizeof(*vq->avail) + > + sizeof(*vq->avail->ring) * num + event; > +} > +...

...8 deletions(-) diff --git a/drivers/vhost/vhost.c b/drivers/vhost/vhost.c index 2025543..1015464 100644 --- a/drivers/vhost/vhost.c +++ b/drivers/vhost/vhost.c @@ -413,6 +413,27 @@ static void vhost_dev_free_iovecs(struct vhost_dev *dev) vhost_vq_free_iovecs(dev->vqs[i]); } +static size_t vhost_get_avail_size(struct vhost_virtqueue *vq, int num) +{ + size_t event = vhost_has_feature(vq, VIRTIO_RING_F_EVENT_IDX) ? 2 : 0; + + return sizeof(*vq->avail) + + sizeof(*vq->avail->ring) * num + event; +} + +static size_t vhost_get_used_size(struct vhost_virtqueue *vq, int num) +{ + size_t event =...

...ndex 2025543..1015464 100644 >> --- a/drivers/vhost/vhost.c >> +++ b/drivers/vhost/vhost.c >> @@ -413,6 +413,27 @@ static void vhost_dev_free_iovecs(struct vhost_dev *dev) >> vhost_vq_free_iovecs(dev->vqs[i]); >> } >> >> +static size_t vhost_get_avail_size(struct vhost_virtqueue *vq, int num) >> +{ >> + size_t event = vhost_has_feature(vq, VIRTIO_RING_F_EVENT_IDX) ? 2 : 0; >> + >> + return sizeof(*vq->avail) + >> + sizeof(*vq->avail->ring) * num + event; >> +} >> + >> +...

..._get_desc_size(vq, vq->num), + &addr)) + return -EINVAL; + + virtio_mdev_writel(mdev, VIRTIO_MDEV_QUEUE_DESC_LOW, addr); + virtio_mdev_writel(mdev, VIRTIO_MDEV_QUEUE_DESC_HIGH, + (addr >> 32)); + + if (!vhost_translate_ring_addr(vq, (u64)vq->avail, + vhost_get_avail_size(vq, vq->num), + &addr)) + return -EINVAL; + + virtio_mdev_writel(mdev, VIRTIO_MDEV_QUEUE_AVAIL_LOW, addr); + virtio_mdev_writel(mdev, VIRTIO_MDEV_QUEUE_AVAIL_HIGH, + (addr >> 32)); + + if (!vhost_translate_ring_addr(vq, (u64)vq->used, + vhost_get_use...

...eady an IOVA that has been map by VFIO. > + > + virtio_mdev_writel(mdev, VIRTIO_MDEV_QUEUE_DESC_LOW, addr); > + virtio_mdev_writel(mdev, VIRTIO_MDEV_QUEUE_DESC_HIGH, > + (addr >> 32)); > + > + if (!vhost_translate_ring_addr(vq, (u64)vq->avail, > + vhost_get_avail_size(vq, vq->num), > + &addr)) > + return -EINVAL; > + > + virtio_mdev_writel(mdev, VIRTIO_MDEV_QUEUE_AVAIL_LOW, addr); > + virtio_mdev_writel(mdev, VIRTIO_MDEV_QUEUE_AVAIL_HIGH, > + (addr >> 32)); > + > + if (!vhost_translate_ring_addr(vq, (u64...

This RFC is to demonstrate below ideas, a) Build vhost-mdev on top of the same abstraction defined in the virtio-mdev series [1]; b) Introduce /dev/vhost-mdev to do vhost ioctls and support setting mdev device as backend; Now the userspace API looks like this: - Userspace generates a compatible mdev device; - Userspace opens this mdev device with VFIO API (including doing IOMMU

This RFC is to demonstrate below ideas, a) Build vhost-mdev on top of the same abstraction defined in the virtio-mdev series [1]; b) Introduce /dev/vhost-mdev to do vhost ioctls and support setting mdev device as backend; Now the userspace API looks like this: - Userspace generates a compatible mdev device; - Userspace opens this mdev device with VFIO API (including doing IOMMU

This series tries to access virtqueue metadata through kernel virtual address instead of copy_user() friends since they had too much overheads like checks, spec barriers or even hardware feature toggling. This is done through setup kernel address through vmap() and resigter MMU notifier for invalidation. Test shows about 24% improvement on TX PPS. TCP_STREAM doesn't see obvious improvement.

This series tries to access virtqueue metadata through kernel virtual address instead of copy_user() friends since they had too much overheads like checks, spec barriers or even hardware feature toggling. This is done through setup kernel address through vmap() and resigter MMU notifier for invalidation. Test shows about 24% improvement on TX PPS. TCP_STREAM doesn't see obvious improvement.

This series tries to access virtqueue metadata through kernel virtual address instead of copy_user() friends since they had too much overheads like checks, spec barriers or even hardware feature toggling. Test shows about 24% improvement on TX PPS. It should benefit other cases as well. Changes from V2: - fix buggy range overlapping check - tear down MMU notifier during vhost ioctl to make sure

This series tries to access virtqueue metadata through kernel virtual address instead of copy_user() friends since they had too much overheads like checks, spec barriers or even hardware feature toggling. Test shows about 24% improvement on TX PPS. It should benefit other cases as well. Changes from V2: - fix buggy range overlapping check - tear down MMU notifier during vhost ioctl to make sure

...f(mn, struct vhost_dev, > + mmu_notifier); > + int i; > + > + for (i = 0; i < dev->nvqs; i++) { > + struct vhost_virtqueue *vq = dev->vqs[i]; > + > + if (vhost_invalidate_vmap(vq, &vq->avail_ring, > + (unsigned long)vq->avail, > + vhost_get_avail_size(vq, vq->num), > + start, end, blockable)) > + return -EAGAIN; > + if (vhost_invalidate_vmap(vq, &vq->desc_ring, > + (unsigned long)vq->desc, > + vhost_get_desc_size(vq, vq->num), > + start, end, blockable)) > + return -EAGAIN; > +...

...host.c >> index 2025543..1015464 100644 >> --- a/drivers/vhost/vhost.c >> +++ b/drivers/vhost/vhost.c >> @@ -413,6 +413,27 @@ static void vhost_dev_free_iovecs(struct vhost_dev *dev) >> vhost_vq_free_iovecs(dev->vqs[i]); >> } >> >> +static size_t vhost_get_avail_size(struct vhost_virtqueue *vq, int num) > Nit: Any reason not to make `num` unsigned or size_t? > Let me use unsigned int to match the definition of vq->num. Thanks

...ble) +{ + struct vhost_dev *dev = container_of(mn, struct vhost_dev, + mmu_notifier); + int i; + + for (i = 0; i < dev->nvqs; i++) { + struct vhost_virtqueue *vq = dev->vqs[i]; + + if (vhost_invalidate_vmap(vq, &vq->avail_ring, + (unsigned long)vq->avail, + vhost_get_avail_size(vq, vq->num), + start, end, blockable)) + return -EAGAIN; + if (vhost_invalidate_vmap(vq, &vq->desc_ring, + (unsigned long)vq->desc, + vhost_get_desc_size(vq, vq->num), + start, end, blockable)) + return -EAGAIN; + if (vhost_invalidate_vmap(vq, &vq-...

...vhost_invalidate(struct vhost_dev *dev, + unsigned long start, unsigned long end) +{ + int i; + + for (i = 0; i < dev->nvqs; i++) { + struct vhost_virtqueue *vq = dev->vqs[i]; + + vhost_invalidate_vmap(vq, &vq->avail_ring, + (unsigned long)vq->avail, + vhost_get_avail_size(vq, vq->num), + start, end); + vhost_invalidate_vmap(vq, &vq->desc_ring, + (unsigned long)vq->desc, + vhost_get_desc_size(vq, vq->num), + start, end); + vhost_invalidate_vmap(vq, &vq->used_ring, + (unsigned long)vq->used, +...

This series tries to access virtqueue metadata through kernel virtual address instead of copy_user() friends since they had too much overheads like checks, spec barriers or even hardware feature toggling. This is done through setup kernel address through direct mapping and co-opreate VM management with MMU notifiers. Test shows about 23% improvement on TX PPS. TCP_STREAM doesn't see obvious

...unsigned long start, unsigned long end) > +{ > + int i; > + > + for (i = 0; i < dev->nvqs; i++) { > + struct vhost_virtqueue *vq = dev->vqs[i]; > + > + vhost_invalidate_vmap(vq, &vq->avail_ring, > + (unsigned long)vq->avail, > + vhost_get_avail_size(vq, vq->num), > + start, end); > + vhost_invalidate_vmap(vq, &vq->desc_ring, > + (unsigned long)vq->desc, > + vhost_get_desc_size(vq, vq->num), > + start, end); > + vhost_invalidate_vmap(vq, &vq->used_ring, > +...

...unsigned long start, unsigned long end) > +{ > + int i; > + > + for (i = 0; i < dev->nvqs; i++) { > + struct vhost_virtqueue *vq = dev->vqs[i]; > + > + vhost_invalidate_vmap(vq, &vq->avail_ring, > + (unsigned long)vq->avail, > + vhost_get_avail_size(vq, vq->num), > + start, end); > + vhost_invalidate_vmap(vq, &vq->desc_ring, > + (unsigned long)vq->desc, > + vhost_get_desc_size(vq, vq->num), > + start, end); > + vhost_invalidate_vmap(vq, &vq->used_ring, > +...

Hi: This series tries to access virtqueue metadata through kernel virtual address instead of copy_user() friends since they had too much overheads like checks, spec barriers or even hardware feature toggling like SMAP. This is done through setup kernel address through direct mapping and co-opreate VM management with MMU notifiers. Test shows about 23% improvement on TX PPS. TCP_STREAM