On 1/8/2018 5:45 PM, Stephen Hemminger wrote:
> Use the new comon code in ethdev to handle link status update.
>
> Signed-off-by: Stephen Hemminger <step...@networkplumber.org>
<...>
> static int
> virtio_dev_link_update(struct rte_eth_dev *dev, __rte_unused int
> wait_to_complete)
> {
> - struct rte_eth_link link, old;
> - uint16_t status;
> struct virtio_hw *hw = dev->data->dev_private;
> - memset(&link, 0, sizeof(link));
> - virtio_dev_atomic_read_link_status(dev, &link);
> - old = link;
> - link.link_duplex = ETH_LINK_FULL_DUPLEX;
> - link.link_speed = SPEED_10G;
> + uint16_t status;
> + struct rte_eth_link link = {
> + .link_speed = ETH_SPEED_NUM_10G,
> + .link_duplex = ETH_LINK_FULL_DUPLEX,
> + .link_autoneg = ETH_LINK_SPEED_FIXED,
With a new cleanup commit [1] ETH_LINK_FIXED used for autoneg value (instead of
ETH_LINK_SPEED_FIXED). And since is ETH_LINK_FIXED is 0, setting autoneg to it
can be omitted.
[1]
93dd7b328a9b ("ethdev: fix link autonegotiation value")
This commit is in next-net.
Unlike Linux "net-next" sub-tree which is for next release development tree,
dpdk next-net is for current release network drivers which this patch seems fits
into.
I already rebased this patchset to next-net, I will share it to continue
discussion from that point.
> + };
>
> if (hw->started == 0) {
> link.link_status = ETH_LINK_DOWN;
> @@ -1957,9 +1916,8 @@ virtio_dev_link_update(struct rte_eth_dev *dev,
> __rte_unused int wait_to_complet
> } else {
> link.link_status = ETH_LINK_UP;
> }
> - virtio_dev_atomic_write_link_status(dev, &link);
>
> - return (old.link_status == link.link_status) ? -1 : 0;
> + return _rte_eth_linkstatus_set(dev, &link);
> }
>
> static int
>
