diff options
| author | Rusty Russell <rusty@rustcorp.com.au> | 2012-01-12 15:44:42 +1030 |
|---|---|---|
| committer | Rusty Russell <rusty@rustcorp.com.au> | 2012-01-12 15:44:42 +1030 |
| commit | 7b21e34fd1c272e3a8c3846168f2f6287a4cd72b (patch) | |
| tree | 0f94c9f834f5b7cd8ba87168df892ed17b09cb8f /tools/virtio | |
| parent | e343a895a9f342f239c5e3c5ffc6c0b1707e6244 (diff) | |
virtio: harsher barriers for rpmsg.
We were cheating with our barriers; using the smp ones rather than the
real device ones. That was fine, until rpmsg came along, which is
used to talk to a real device (a non-SMP CPU).
Unfortunately, just putting back the real barriers (reverting
d57ed95d) causes a performance regression on virtio-pci. In
particular, Amos reports netbench's TCP_RR over virtio_net CPU
utilization increased up to 35% while throughput went down by up to
14%.
By comparison, this branch is in the noise.
Reference: https://lkml.org/lkml/2011/12/11/22
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Diffstat (limited to 'tools/virtio')
| -rw-r--r-- | tools/virtio/linux/virtio.h | 1 | ||||
| -rw-r--r-- | tools/virtio/virtio_test.c | 3 |
2 files changed, 3 insertions, 1 deletions
diff --git a/tools/virtio/linux/virtio.h b/tools/virtio/linux/virtio.h index 669bcdd4580..953db2abf6b 100644 --- a/tools/virtio/linux/virtio.h +++ b/tools/virtio/linux/virtio.h @@ -214,6 +214,7 @@ void *virtqueue_detach_unused_buf(struct virtqueue *vq); struct virtqueue *vring_new_virtqueue(unsigned int num, unsigned int vring_align, struct virtio_device *vdev, + bool weak_barriers, void *pages, void (*notify)(struct virtqueue *vq), void (*callback)(struct virtqueue *vq), diff --git a/tools/virtio/virtio_test.c b/tools/virtio/virtio_test.c index 74d3331bdaf..0740284396c 100644 --- a/tools/virtio/virtio_test.c +++ b/tools/virtio/virtio_test.c @@ -92,7 +92,8 @@ static void vq_info_add(struct vdev_info *dev, int num) assert(r >= 0); memset(info->ring, 0, vring_size(num, 4096)); vring_init(&info->vring, num, info->ring, 4096); - info->vq = vring_new_virtqueue(info->vring.num, 4096, &dev->vdev, info->ring, + info->vq = vring_new_virtqueue(info->vring.num, 4096, &dev->vdev, + true, info->ring, vq_notify, vq_callback, "test"); assert(info->vq); info->vq->priv = info; |