On 2/2/23 14:54, Gregory Etelson wrote:
diff --git a/lib/ethdev/rte_flow.c b/lib/ethdev/rte_flow.c
index 7d0c24366c..4554c8f021 100644
--- a/lib/ethdev/rte_flow.c
+++ b/lib/ethdev/rte_flow.c
@@ -1883,3 +1883,85 @@ rte_flow_async_action_handle_query(uint16_t port_id,
action_handle, data, user_data,
error);
return flow_err(port_id, ret, error);
}
+
+int
+rte_flow_action_handle_query_update(uint16_t port_id,
+ struct rte_flow_action_handle *handle,
+ const void *update, void *query,
+ enum rte_flow_query_update_mode mode,
+ struct rte_flow_error *error)
+{
+ int ret;
+ struct rte_eth_dev *dev;
+ const struct rte_flow_ops *ops;
+
+ RTE_ETH_VALID_PORTID_OR_ERR_RET(port_id, -ENODEV);
+ if (!handle)
+ return -EINVAL;
+ dev = &rte_eth_devices[port_id];
+ ops = rte_flow_ops_get(port_id, error);
+ if (!ops)
+ return -ENOTSUP;
+ if (update && query) {
+ if (!ops->action_handle_query_update)
+ return -ENOTSUP;
+ if (mode != RTE_FLOW_QU_QUERY_FIRST &&
+ mode != RTE_FLOW_QU_UPDATE_FIRST)
+ return -EINVAL;
Shouldn't it be checked in any case?
Also I'd initialize RTE_FLOW_QU_QUERY_FIRST to 1 on enum
definition to ensure that just 0 is not used as a mode value.
Also "Required if both *update* and *query* are not NULL." does
not make sense in the description since you have no way to skip
it.
+ ret = ops->action_handle_query_update(dev, handle, update,
+ query, mode, error);
+ } else if (!update && query) {
+ ret = rte_flow_action_handle_query(port_id, handle, query,
+ error);
+ } else if (update && !query) {
+ ret = rte_flow_action_handle_update(port_id, handle, update,
+ error);
+ } else {
+ return -EINVAL;
+ }
IMHO logic is wrong above since it should be sufficient to
implement just one action_handle_query_update callback instead
of all three.
I.e. if action_handle_query_update is available, it should be
used in any case.
+ return flow_err(port_id, ret, error);
+}
};
INTERNAL {
same for async API
