Query the link status can end up in an inconsist state where the port is
down but it is reporting speed. For that another query is scheduled for a
later time.
A race condition is possible between the scheduled call and other link
status interrupt handlers. When the scheduled query by-pass those handlers,
the link status will be stuck in an in-consist state.
This patch addresses the race condition by not blocking link status
queries in case delayed query is on the flight.

Fixes: 198a3c339a8f ("mlx5: handle link status interrupts")
CC: sta...@dpdk.org

Signed-off-by: Shahaf Shuler <shah...@mellanox.com>
---
on v3:
* allow only single alarm on the flight.

on v2:
* change the commit log to better explain the race
* instead of blocking other interrupt handlers allow everyone to query the link 
status
---
 drivers/net/mlx5/mlx5_ethdev.c | 26 +++++++++++---------------
 1 file changed, 11 insertions(+), 15 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_ethdev.c b/drivers/net/mlx5/mlx5_ethdev.c
index e77238f..fcbe273 100644
--- a/drivers/net/mlx5/mlx5_ethdev.c
+++ b/drivers/net/mlx5/mlx5_ethdev.c
@@ -1128,7 +1128,7 @@ struct priv *
 priv_dev_link_status_handler(struct priv *priv, struct rte_eth_dev *dev)
 {
        struct ibv_async_event event;
-       int port_change = 0;
+       struct rte_eth_link *link = &dev->data->dev_link;
        int ret = 0;
 
        /* Read all message and acknowledge them. */
@@ -1136,29 +1136,24 @@ struct priv *
                if (ibv_get_async_event(priv->ctx, &event))
                        break;
 
-               if (event.event_type == IBV_EVENT_PORT_ACTIVE ||
-                   event.event_type == IBV_EVENT_PORT_ERR)
-                       port_change = 1;
-               else
+               if (event.event_type != IBV_EVENT_PORT_ACTIVE &&
+                   event.event_type != IBV_EVENT_PORT_ERR)
                        DEBUG("event type %d on port %d not handled",
                              event.event_type, event.element.port_num);
                ibv_ack_async_event(&event);
        }
-
-       if (port_change ^ priv->pending_alarm) {
-               struct rte_eth_link *link = &dev->data->dev_link;
-
-               priv->pending_alarm = 0;
-               mlx5_link_update(dev, 0);
-               if (((link->link_speed == 0) && link->link_status) ||
-                   ((link->link_speed != 0) && !link->link_status)) {
+       mlx5_link_update(dev, 0);
+       if (((link->link_speed == 0) && link->link_status) ||
+           ((link->link_speed != 0) && !link->link_status)) {
+               if (!priv->pending_alarm) {
                        /* Inconsistent status, check again later. */
                        priv->pending_alarm = 1;
                        rte_eal_alarm_set(MLX5_ALARM_TIMEOUT_US,
                                          mlx5_dev_link_status_handler,
                                          dev);
-               } else
-                       ret = 1;
+               }
+       } else {
+               ret = 1;
        }
        return ret;
 }
@@ -1178,6 +1173,7 @@ struct priv *
 
        priv_lock(priv);
        assert(priv->pending_alarm == 1);
+       priv->pending_alarm = 0;
        ret = priv_dev_link_status_handler(priv, dev);
        priv_unlock(priv);
        if (ret)
-- 
1.8.3.1

Reply via email to