diff options
author | Rusty Russell <rusty@rustcorp.com.au> | 2008-02-04 23:50:07 -0500 |
---|---|---|
committer | Rusty Russell <rusty@rustcorp.com.au> | 2008-02-04 23:50:07 +1100 |
commit | a48bd8f67003c342e147309a331f656a5e75a5e4 (patch) | |
tree | 30ab7e3b1763a526dbc1610960bcb5aa9c88dec5 | |
parent | e70f2f1bb8c8cf08ba0d836466f77eea27dd3474 (diff) | |
download | op-kernel-dev-a48bd8f67003c342e147309a331f656a5e75a5e4.zip op-kernel-dev-a48bd8f67003c342e147309a331f656a5e75a5e4.tar.gz |
virtio: flush buffers on open
Fix bug found by Christian Borntraeger: if the other side fills all
the registered network buffers before we enable NAPI, we will never
get an interrupt. The simplest fix is to process the input queue once
on open.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
-rw-r--r-- | drivers/net/virtio_net.c | 7 |
1 files changed, 7 insertions, 0 deletions
diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c index be5688f..bd4d26a 100644 --- a/drivers/net/virtio_net.c +++ b/drivers/net/virtio_net.c @@ -283,6 +283,13 @@ static int virtnet_open(struct net_device *dev) struct virtnet_info *vi = netdev_priv(dev); napi_enable(&vi->napi); + + /* If all buffers were filled by other side before we napi_enabled, we + * won't get another interrupt, so process any outstanding packets + * now. virtnet_poll wants re-enable the queue, so we disable here. */ + vi->rvq->vq_ops->disable_cb(vi->rvq); + netif_rx_schedule(vi->dev, &vi->napi); + return 0; } |