DPDK failed to configure eth device with err No: -22, tx_queues is greater than one (Ethdev port_id=0 nb_tx_queues=2 > 1) - dpdk

Question

I have an application that uses DPDK for Fast Path. During DPDK initialization, am trying to configure two TX queues for a port, but it failed to configure the eth device.
I am using Intel IGB driver(I350 NIC) on a Bare Metal setup. As per DPDK documentation IGB Poll Mode Driver (https://doc.dpdk.org/guides/nics/igb.html) should support multiple RX/TX queues for a port. Am trying to configure two TX queues for a port(port_id = 0), when invoking the API "rte_eth_dev_configure(portid = 0, nb_rx_queue = 1, nb_tx_queue = 2, &local_port_conf)", it is returning error code: -22, "Ethdev port_id=0 nb_tx_queues=2 > 1".
Does the IGB PMD driver support multiple TX queues for a port? Or do I need to do any configuration changes to support multiple TX queues?

Answer 1

For virtual functions (VFs), the NIC model in question supports only single-queue mode (source).
To test multi-queue support, consider passing a physical function (PF) to the setup instead.

Answer 2

Intel Foundational NIC i350 assign Up to eight queues per port for Physical Funtion driver as per PRODUCT BRIEF Intel® Ethernet Controller I350. With respect to virtual function per port 1 queue for each VF with a total of 8 VF max per port defined under Intel® Ethernet Controller I350 Datasheet.
In Linux, these can be validated by
ethtool: One can check per queue instance with option -S, as statistics is supported per queue.
dpdk: using rte_eth_dev_info_get to get attributes max_rx_queues and max_tx_queues
Note: as VF gets allocated with 1 RX-TX queue pair, the error mentioned -22, Ethdev port_id=0 nb_tx_queues=2 > 1 can be of VF in use. Hence the right way of configuration is first get the DPDK port max rx-tx queues and ensure it is in bounds.
Sample Code Snippet:
/* Initialise each port */
RTE_ETH_FOREACH_DEV(portid) {
struct rte_eth_rxconf rxq_conf;
struct rte_eth_txconf txq_conf;
struct rte_eth_conf local_port_conf = port_conf;
struct rte_eth_dev_info dev_info;
/* skip ports that are not enabled */
if ((l2fwd_enabled_port_mask & (1 << portid)) == 0) {
printf("Skipping disabled port %u\n", portid);
continue;
}
nb_ports_available++;
/* init port */
printf("Initializing port %u... ", portid);
fflush(stdout);
ret = rte_eth_dev_info_get(portid, &dev_info);
if (ret != 0)
rte_exit(EXIT_FAILURE,
"Error during getting device (port %u) info: %s\n",
portid, strerror(-ret));
if (dev_info.tx_offload_capa & RTE_ETH_TX_OFFLOAD_MBUF_FAST_FREE)
local_port_conf.txmode.offloads |=
RTE_ETH_TX_OFFLOAD_MBUF_FAST_FREE;
/* Configure the number of queues for a port. */
if ((dev_info.max_rx_queues >= user_rx_queues) && (dev_info.max_tx_queues >= user_tx_queues)) {
ret = rte_eth_dev_configure(portid, user_rx_queues, user_tx_queues, &local_port_conf);
if (ret < 0)
rte_exit(EXIT_FAILURE, "Cannot configure device: err=%d, port=%u\n",
ret, portid);
/* >8 End of configuration of the number of queues for a port. */
}