On Thu, Mar 23, 2023 at 11:46 AM Stefano Garzarella <sgarzare@xxxxxxxxxx> wrote:
>
> On Thu, Mar 23, 2023 at 09:09:14AM +0100, Eugenio Perez Martin wrote:
> >On Tue, Mar 21, 2023 at 4:43 PM Stefano Garzarella <sgarzare@xxxxxxxxxx> wrote:
> >>
> >> vDPA supports the possibility to use user VA in the iotlb messages.
> >> So, let's add support for user VA in vringh to use it in the vDPA
> >> simulators.
> >>
> >> Signed-off-by: Stefano Garzarella <sgarzare@xxxxxxxxxx>
> >> ---
> >>
> >> Notes:
> >> v3:
> >> - refactored avoiding code duplication [Eugenio]
> >> v2:
> >> - replace kmap_atomic() with kmap_local_page() [see previous patch]
> >> - fix cast warnings when build with W=1 C=1
> >>
> >> include/linux/vringh.h | 5 +-
> >> drivers/vdpa/mlx5/net/mlx5_vnet.c | 2 +-
> >> drivers/vdpa/vdpa_sim/vdpa_sim.c | 4 +-
> >> drivers/vhost/vringh.c | 153 +++++++++++++++++++++++-------
> >> 4 files changed, 127 insertions(+), 37 deletions(-)
> >>
> >> diff --git a/include/linux/vringh.h b/include/linux/vringh.h
> >> index 1991a02c6431..d39b9f2dcba0 100644
> >> --- a/include/linux/vringh.h
> >> +++ b/include/linux/vringh.h
> >> @@ -32,6 +32,9 @@ struct vringh {
> >> /* Can we get away with weak barriers? */
> >> bool weak_barriers;
> >>
> >> + /* Use user's VA */
> >> + bool use_va;
> >> +
> >> /* Last available index we saw (ie. where we're up to). */
> >> u16 last_avail_idx;
> >>
> >> @@ -279,7 +282,7 @@ void vringh_set_iotlb(struct vringh *vrh, struct vhost_iotlb *iotlb,
> >> spinlock_t *iotlb_lock);
> >>
> >> int vringh_init_iotlb(struct vringh *vrh, u64 features,
> >> - unsigned int num, bool weak_barriers,
> >> + unsigned int num, bool weak_barriers, bool use_va,
> >> struct vring_desc *desc,
> >> struct vring_avail *avail,
> >> struct vring_used *used);
> >> diff --git a/drivers/vdpa/mlx5/net/mlx5_vnet.c b/drivers/vdpa/mlx5/net/mlx5_vnet.c
> >> index 520646ae7fa0..dfd0e000217b 100644
> >> --- a/drivers/vdpa/mlx5/net/mlx5_vnet.c
> >> +++ b/drivers/vdpa/mlx5/net/mlx5_vnet.c
> >> @@ -2537,7 +2537,7 @@ static int setup_cvq_vring(struct mlx5_vdpa_dev *mvdev)
> >>
> >> if (mvdev->actual_features & BIT_ULL(VIRTIO_NET_F_CTRL_VQ))
> >> err = vringh_init_iotlb(&cvq->vring, mvdev->actual_features,
> >> - MLX5_CVQ_MAX_ENT, false,
> >> + MLX5_CVQ_MAX_ENT, false, false,
> >> (struct vring_desc *)(uintptr_t)cvq->desc_addr,
> >> (struct vring_avail *)(uintptr_t)cvq->driver_addr,
> >> (struct vring_used *)(uintptr_t)cvq->device_addr);
> >> diff --git a/drivers/vdpa/vdpa_sim/vdpa_sim.c b/drivers/vdpa/vdpa_sim/vdpa_sim.c
> >> index eea23c630f7c..47cdf2a1f5b8 100644
> >> --- a/drivers/vdpa/vdpa_sim/vdpa_sim.c
> >> +++ b/drivers/vdpa/vdpa_sim/vdpa_sim.c
> >> @@ -60,7 +60,7 @@ static void vdpasim_queue_ready(struct vdpasim *vdpasim, unsigned int idx)
> >> struct vdpasim_virtqueue *vq = &vdpasim->vqs[idx];
> >> uint16_t last_avail_idx = vq->vring.last_avail_idx;
> >>
> >> - vringh_init_iotlb(&vq->vring, vdpasim->features, vq->num, true,
> >> + vringh_init_iotlb(&vq->vring, vdpasim->features, vq->num, true, false,
> >> (struct vring_desc *)(uintptr_t)vq->desc_addr,
> >> (struct vring_avail *)
> >> (uintptr_t)vq->driver_addr,
> >> @@ -92,7 +92,7 @@ static void vdpasim_vq_reset(struct vdpasim *vdpasim,
> >> vq->cb = NULL;
> >> vq->private = NULL;
> >> vringh_init_iotlb(&vq->vring, vdpasim->dev_attr.supported_features,
> >> - VDPASIM_QUEUE_MAX, false, NULL, NULL, NULL);
> >> + VDPASIM_QUEUE_MAX, false, false, NULL, NULL, NULL);
> >>
> >> vq->vring.notify = NULL;
> >> }
> >> diff --git a/drivers/vhost/vringh.c b/drivers/vhost/vringh.c
> >> index 0ba3ef809e48..72c88519329a 100644
> >> --- a/drivers/vhost/vringh.c
> >> +++ b/drivers/vhost/vringh.c
> >> @@ -1094,10 +1094,18 @@ EXPORT_SYMBOL(vringh_need_notify_kern);
> >>
> >> #if IS_REACHABLE(CONFIG_VHOST_IOTLB)
> >>
> >> +struct iotlb_vec {
> >> + union {
> >> + struct iovec *iovec;
> >> + struct bio_vec *bvec;
> >> + } iov;
> >> + size_t count;
> >> + bool is_iovec;
> >> +};
> >> +
> >> static int iotlb_translate(const struct vringh *vrh,
> >> u64 addr, u64 len, u64 *translated,
> >> - struct bio_vec iov[],
> >> - int iov_size, u32 perm)
> >> + struct iotlb_vec *ivec, u32 perm)
> >> {
> >> struct vhost_iotlb_map *map;
> >> struct vhost_iotlb *iotlb = vrh->iotlb;
> >> @@ -1107,9 +1115,9 @@ static int iotlb_translate(const struct vringh *vrh,
> >> spin_lock(vrh->iotlb_lock);
> >>
> >> while (len > s) {
> >> - u64 size, pa, pfn;
> >> + u64 size;
> >>
> >> - if (unlikely(ret >= iov_size)) {
> >> + if (unlikely(ret >= ivec->count)) {
> >> ret = -ENOBUFS;
> >> break;
> >> }
> >> @@ -1124,10 +1132,22 @@ static int iotlb_translate(const struct vringh *vrh,
> >> }
> >>
> >> size = map->size - addr + map->start;
> >> - pa = map->addr + addr - map->start;
> >> - pfn = pa >> PAGE_SHIFT;
> >> - bvec_set_page(&iov[ret], pfn_to_page(pfn), min(len - s, size),
> >> - pa & (PAGE_SIZE - 1));
> >> + if (ivec->is_iovec) {
> >> + struct iovec *iovec = ivec->iov.iovec;
> >> +
> >> + iovec[ret].iov_len = min(len - s, size);
> >> + iovec[ret].iov_base = (void __user *)(unsigned long)
> >
> >s/unsigned long/uintptr_t ?
> >
>
> yep, good catch!
>
> As I wrote to Jason, I think I'll take it out of the if and just declare
> an uintptr_t variable, since I'm using it also in the else branch.
>
> >
> >
> >> + (map->addr + addr - map->start);
> >> + } else {
> >> + u64 pa = map->addr + addr - map->start;
> >> + u64 pfn = pa >> PAGE_SHIFT;
> >> + struct bio_vec *bvec = ivec->iov.bvec;
> >> +
> >> + bvec_set_page(&bvec[ret], pfn_to_page(pfn),
> >> + min(len - s, size),
> >> + pa & (PAGE_SIZE - 1));
> >> + }
> >> +
> >> s += size;
> >> addr += size;
> >> ++ret;
> >> @@ -1141,26 +1161,42 @@ static int iotlb_translate(const struct vringh *vrh,
> >> return ret;
> >> }
> >>
> >> +#define IOTLB_IOV_SIZE 16
> >
> >I'm fine with defining here, but maybe it is better to isolate the
> >change in a previous patch or reuse another well known macro?
>
> Yep, good point!
>
> Do you have any well known macro to suggest?
>
Not really, 16 seems like a convenience value here actually :). Maybe
replace _SIZE with _STRIDE or similar?
I keep the Acked-by even if the final name is IOTLB_IOV_SIZE though.
Thanks!
