In HWS template API, the aging mechanism doesn't support shared host
mode now. When the guest's counter is set to 0, the aging won't be
initialized.
The current implementation didn't prevent the user from querying the
aged flows on the uninitialized port. The access of invalid pointers
would cause a crash.
With this commit, the flag of the per port aging initialization will
be checked. This would help to get rid of the invalid accessing.
Fixes: 04a4de756e14 ("net/mlx5: support flow age action with HWS")
Cc: michae...@nvidia.com
Cc: sta...@dpdk.org
Signed-off-by: Bing Zhao <bi...@nvidia.com>
Acked-by: Suanming Mou <suanmi...@nvidia.com>
---
drivers/net/mlx5/mlx5_flow_hw.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/drivers/net/mlx5/mlx5_flow_hw.c b/drivers/net/mlx5/mlx5_flow_hw.c
index 3bb3a9a178..0e6d6e02dd 100644
--- a/drivers/net/mlx5/mlx5_flow_hw.c
+++ b/drivers/net/mlx5/mlx5_flow_hw.c
@@ -10937,6 +10937,10 @@ flow_hw_get_q_aged_flows(struct rte_eth_dev *dev,
uint32_t queue_id,
return rte_flow_error_set(error, EINVAL,
RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
NULL, "empty context");
+ if (!priv->hws_age_req)
+ return rte_flow_error_set(error, ENOENT,
+ RTE_FLOW_ERROR_TYPE_UNSPECIFIED,
+ NULL, "No aging initialized");
if (priv->hws_strict_queue) {
if (queue_id >= age_info->hw_q_age->nb_rings)
return rte_flow_error_set(error, EINVAL,
--
2.34.1
