diff options
Diffstat (limited to 'drivers/vhost')
-rw-r--r-- | drivers/vhost/net.c | 37 | ||||
-rw-r--r-- | drivers/vhost/vhost.c | 75 | ||||
-rw-r--r-- | drivers/vhost/vhost.h | 4 |
3 files changed, 66 insertions, 50 deletions
diff --git a/drivers/vhost/net.c b/drivers/vhost/net.c index 986058a57917..eeaf6739215f 100644 --- a/drivers/vhost/net.c +++ b/drivers/vhost/net.c @@ -105,7 +105,9 @@ struct vhost_net_virtqueue { /* vhost zerocopy support fields below: */ /* last used idx for outstanding DMA zerocopy buffers */ int upend_idx; - /* first used idx for DMA done zerocopy buffers */ + /* For TX, first used idx for DMA done zerocopy buffers + * For RX, number of batched heads + */ int done_idx; /* an array of userspace buffers info */ struct ubuf_info *ubuf_info; @@ -626,6 +628,18 @@ static int sk_has_rx_data(struct sock *sk) return skb_queue_empty(&sk->sk_receive_queue); } +static void vhost_rx_signal_used(struct vhost_net_virtqueue *nvq) +{ + struct vhost_virtqueue *vq = &nvq->vq; + struct vhost_dev *dev = vq->dev; + + if (!nvq->done_idx) + return; + + vhost_add_used_and_signal_n(dev, vq, vq->heads, nvq->done_idx); + nvq->done_idx = 0; +} + static int vhost_net_rx_peek_head_len(struct vhost_net *net, struct sock *sk) { struct vhost_net_virtqueue *rvq = &net->vqs[VHOST_NET_VQ_RX]; @@ -635,6 +649,8 @@ static int vhost_net_rx_peek_head_len(struct vhost_net *net, struct sock *sk) int len = peek_head_len(rvq, sk); if (!len && vq->busyloop_timeout) { + /* Flush batched heads first */ + vhost_rx_signal_used(rvq); /* Both tx vq and rx socket were polled here */ mutex_lock_nested(&vq->mutex, 1); vhost_disable_notify(&net->dev, vq); @@ -762,7 +778,7 @@ static void handle_rx(struct vhost_net *net) }; size_t total_len = 0; int err, mergeable; - s16 headcount, nheads = 0; + s16 headcount; size_t vhost_hlen, sock_hlen; size_t vhost_len, sock_len; struct socket *sock; @@ -790,8 +806,8 @@ static void handle_rx(struct vhost_net *net) while ((sock_len = vhost_net_rx_peek_head_len(net, sock->sk))) { sock_len += sock_hlen; vhost_len = sock_len + vhost_hlen; - headcount = get_rx_bufs(vq, vq->heads + nheads, vhost_len, - &in, vq_log, &log, + headcount = get_rx_bufs(vq, vq->heads + nvq->done_idx, + vhost_len, &in, vq_log, &log, likely(mergeable) ? UIO_MAXIOV : 1); /* On error, stop handling until the next kick. */ if (unlikely(headcount < 0)) @@ -862,12 +878,9 @@ static void handle_rx(struct vhost_net *net) vhost_discard_vq_desc(vq, headcount); goto out; } - nheads += headcount; - if (nheads > VHOST_RX_BATCH) { - vhost_add_used_and_signal_n(&net->dev, vq, vq->heads, - nheads); - nheads = 0; - } + nvq->done_idx += headcount; + if (nvq->done_idx > VHOST_RX_BATCH) + vhost_rx_signal_used(nvq); if (unlikely(vq_log)) vhost_log_write(vq, vq_log, log, vhost_len); total_len += vhost_len; @@ -878,9 +891,7 @@ static void handle_rx(struct vhost_net *net) } vhost_net_enable_vq(net, vq); out: - if (nheads) - vhost_add_used_and_signal_n(&net->dev, vq, vq->heads, - nheads); + vhost_rx_signal_used(nvq); mutex_unlock(&vq->mutex); } diff --git a/drivers/vhost/vhost.c b/drivers/vhost/vhost.c index bec722e41f58..f0be5f35ab28 100644 --- a/drivers/vhost/vhost.c +++ b/drivers/vhost/vhost.c @@ -641,14 +641,14 @@ void vhost_dev_cleanup(struct vhost_dev *dev) } EXPORT_SYMBOL_GPL(vhost_dev_cleanup); -static int log_access_ok(void __user *log_base, u64 addr, unsigned long sz) +static bool log_access_ok(void __user *log_base, u64 addr, unsigned long sz) { u64 a = addr / VHOST_PAGE_SIZE / 8; /* Make sure 64 bit math will not overflow. */ if (a > ULONG_MAX - (unsigned long)log_base || a + (unsigned long)log_base > ULONG_MAX) - return 0; + return false; return access_ok(VERIFY_WRITE, log_base + a, (sz + VHOST_PAGE_SIZE * 8 - 1) / VHOST_PAGE_SIZE / 8); @@ -661,30 +661,30 @@ static bool vhost_overflow(u64 uaddr, u64 size) } /* Caller should have vq mutex and device mutex. */ -static int vq_memory_access_ok(void __user *log_base, struct vhost_umem *umem, - int log_all) +static bool vq_memory_access_ok(void __user *log_base, struct vhost_umem *umem, + int log_all) { struct vhost_umem_node *node; if (!umem) - return 0; + return false; list_for_each_entry(node, &umem->umem_list, link) { unsigned long a = node->userspace_addr; if (vhost_overflow(node->userspace_addr, node->size)) - return 0; + return false; if (!access_ok(VERIFY_WRITE, (void __user *)a, node->size)) - return 0; + return false; else if (log_all && !log_access_ok(log_base, node->start, node->size)) - return 0; + return false; } - return 1; + return true; } static inline void __user *vhost_vq_meta_fetch(struct vhost_virtqueue *vq, @@ -701,13 +701,13 @@ static inline void __user *vhost_vq_meta_fetch(struct vhost_virtqueue *vq, /* Can we switch to this memory table? */ /* Caller should have device mutex but not vq mutex */ -static int memory_access_ok(struct vhost_dev *d, struct vhost_umem *umem, - int log_all) +static bool memory_access_ok(struct vhost_dev *d, struct vhost_umem *umem, + int log_all) { int i; for (i = 0; i < d->nvqs; ++i) { - int ok; + bool ok; bool log; mutex_lock(&d->vqs[i]->mutex); @@ -717,12 +717,12 @@ static int memory_access_ok(struct vhost_dev *d, struct vhost_umem *umem, ok = vq_memory_access_ok(d->vqs[i]->log_base, umem, log); else - ok = 1; + ok = true; mutex_unlock(&d->vqs[i]->mutex); if (!ok) - return 0; + return false; } - return 1; + return true; } static int translate_desc(struct vhost_virtqueue *vq, u64 addr, u32 len, @@ -744,7 +744,7 @@ static int vhost_copy_to_user(struct vhost_virtqueue *vq, void __user *to, struct iov_iter t; void __user *uaddr = vhost_vq_meta_fetch(vq, (u64)(uintptr_t)to, size, - VHOST_ADDR_DESC); + VHOST_ADDR_USED); if (uaddr) return __copy_to_user(uaddr, from, size); @@ -959,21 +959,21 @@ static void vhost_iotlb_notify_vq(struct vhost_dev *d, spin_unlock(&d->iotlb_lock); } -static int umem_access_ok(u64 uaddr, u64 size, int access) +static bool umem_access_ok(u64 uaddr, u64 size, int access) { unsigned long a = uaddr; /* Make sure 64 bit math will not overflow. */ if (vhost_overflow(uaddr, size)) - return -EFAULT; + return false; if ((access & VHOST_ACCESS_RO) && !access_ok(VERIFY_READ, (void __user *)a, size)) - return -EFAULT; + return false; if ((access & VHOST_ACCESS_WO) && !access_ok(VERIFY_WRITE, (void __user *)a, size)) - return -EFAULT; - return 0; + return false; + return true; } static int vhost_process_iotlb_msg(struct vhost_dev *dev, @@ -981,6 +981,7 @@ static int vhost_process_iotlb_msg(struct vhost_dev *dev, { int ret = 0; + mutex_lock(&dev->mutex); vhost_dev_lock_vqs(dev); switch (msg->type) { case VHOST_IOTLB_UPDATE: @@ -988,7 +989,7 @@ static int vhost_process_iotlb_msg(struct vhost_dev *dev, ret = -EFAULT; break; } - if (umem_access_ok(msg->uaddr, msg->size, msg->perm)) { + if (!umem_access_ok(msg->uaddr, msg->size, msg->perm)) { ret = -EFAULT; break; } @@ -1016,6 +1017,8 @@ static int vhost_process_iotlb_msg(struct vhost_dev *dev, } vhost_dev_unlock_vqs(dev); + mutex_unlock(&dev->mutex); + return ret; } ssize_t vhost_chr_write_iter(struct vhost_dev *dev, @@ -1135,10 +1138,10 @@ static int vhost_iotlb_miss(struct vhost_virtqueue *vq, u64 iova, int access) return 0; } -static int vq_access_ok(struct vhost_virtqueue *vq, unsigned int num, - struct vring_desc __user *desc, - struct vring_avail __user *avail, - struct vring_used __user *used) +static bool vq_access_ok(struct vhost_virtqueue *vq, unsigned int num, + struct vring_desc __user *desc, + struct vring_avail __user *avail, + struct vring_used __user *used) { size_t s = vhost_has_feature(vq, VIRTIO_RING_F_EVENT_IDX) ? 2 : 0; @@ -1161,8 +1164,8 @@ static void vhost_vq_meta_update(struct vhost_virtqueue *vq, vq->meta_iotlb[type] = node; } -static int iotlb_access_ok(struct vhost_virtqueue *vq, - int access, u64 addr, u64 len, int type) +static bool iotlb_access_ok(struct vhost_virtqueue *vq, + int access, u64 addr, u64 len, int type) { const struct vhost_umem_node *node; struct vhost_umem *umem = vq->iotlb; @@ -1220,7 +1223,7 @@ EXPORT_SYMBOL_GPL(vq_iotlb_prefetch); /* Can we log writes? */ /* Caller should have device mutex but not vq mutex */ -int vhost_log_access_ok(struct vhost_dev *dev) +bool vhost_log_access_ok(struct vhost_dev *dev) { return memory_access_ok(dev, dev->umem, 1); } @@ -1228,8 +1231,8 @@ EXPORT_SYMBOL_GPL(vhost_log_access_ok); /* Verify access for write logging. */ /* Caller should have vq mutex and device mutex */ -static int vq_log_access_ok(struct vhost_virtqueue *vq, - void __user *log_base) +static bool vq_log_access_ok(struct vhost_virtqueue *vq, + void __user *log_base) { size_t s = vhost_has_feature(vq, VIRTIO_RING_F_EVENT_IDX) ? 2 : 0; @@ -1242,12 +1245,14 @@ static int vq_log_access_ok(struct vhost_virtqueue *vq, /* Can we start vq? */ /* Caller should have vq mutex and device mutex */ -int vhost_vq_access_ok(struct vhost_virtqueue *vq) +bool vhost_vq_access_ok(struct vhost_virtqueue *vq) { - int ret = vq_log_access_ok(vq, vq->log_base); + if (!vq_log_access_ok(vq, vq->log_base)) + return false; - if (ret || vq->iotlb) - return ret; + /* Access validation occurs at prefetch time with IOTLB */ + if (vq->iotlb) + return true; return vq_access_ok(vq, vq->num, vq->desc, vq->avail, vq->used); } diff --git a/drivers/vhost/vhost.h b/drivers/vhost/vhost.h index d8ee85ae8fdc..6c844b90a168 100644 --- a/drivers/vhost/vhost.h +++ b/drivers/vhost/vhost.h @@ -178,8 +178,8 @@ void vhost_dev_cleanup(struct vhost_dev *); void vhost_dev_stop(struct vhost_dev *); long vhost_dev_ioctl(struct vhost_dev *, unsigned int ioctl, void __user *argp); long vhost_vring_ioctl(struct vhost_dev *d, unsigned int ioctl, void __user *argp); -int vhost_vq_access_ok(struct vhost_virtqueue *vq); -int vhost_log_access_ok(struct vhost_dev *); +bool vhost_vq_access_ok(struct vhost_virtqueue *vq); +bool vhost_log_access_ok(struct vhost_dev *); int vhost_get_vq_desc(struct vhost_virtqueue *, struct iovec iov[], unsigned int iov_count, |