15. Packet copying using Intel® QuickData Technology

15.1. Overview

This sample is intended as a demonstration of the basic components of a DPDK forwarding application and example of how to use IOAT driver API to make packets copies.

Also while forwarding, the MAC addresses are affected as follows:

• The source MAC address is replaced by the TX port MAC address
• The destination MAC address is replaced by 02:00:00:00:00:TX_PORT_ID

This application can be used to compare performance of using software packet copy with copy done using a DMA device for different sizes of packets. The example will print out statistics each second. The stats shows received/send packets and packets dropped or failed to copy.

15.2. Compiling the Application

To compile the sample application see Compiling the Sample Applications.

The application is located in the ioat sub-directory.

15.3. Running the Application

In order to run the hardware copy application, the copying device needs to be bound to user-space IO driver.

Refer to the “IOAT Rawdev Driver” chapter in the “Rawdev Drivers” document for information on using the driver.

The application requires a number of command line options:

./<build_dir>/examples/dpdk-ioat [EAL options] -- [-p MASK] [-q NQ] [-s RS] [-c <sw|hw>]
    [--[no-]mac-updating]

where,

• p MASK: A hexadecimal bitmask of the ports to configure (default is all)
• q NQ: Number of Rx queues used per port equivalent to CBDMA channels per port (default is 1)
• c CT: Performed packet copy type: software (sw) or hardware using DMA (hw) (default is hw)
• s RS: Size of IOAT rawdev ring for hardware copy mode or rte_ring for software copy mode (default is 2048)
• –[no-]mac-updating: Whether MAC address of packets should be changed or not (default is mac-updating)

The application can be launched in various configurations depending on provided parameters. The app can use up to 2 lcores: one of them receives incoming traffic and makes a copy of each packet. The second lcore then updates MAC address and sends the copy. If one lcore per port is used, both operations are done sequentially. For each configuration an additional lcore is needed since the main lcore does not handle traffic but is responsible for configuration, statistics printing and safe shutdown of all ports and devices.

The application can use a maximum of 8 ports.

To run the application in a Linux environment with 3 lcores (the main lcore, plus two forwarding cores), a single port (port 0), software copying and MAC updating issue the command:

$ ./<build_dir>/examples/dpdk-ioat -l 0-2 -n 2 -- -p 0x1 --mac-updating -c sw

To run the application in a Linux environment with 2 lcores (the main lcore, plus one forwarding core), 2 ports (ports 0 and 1), hardware copying and no MAC updating issue the command:

$ ./<build_dir>/examples/dpdk-ioat -l 0-1 -n 1 -- -p 0x3 --no-mac-updating -c hw

Refer to the DPDK Getting Started Guide for general information on running applications and the Environment Abstraction Layer (EAL) options.

15.4. Explanation

The following sections provide an explanation of the main components of the code.

All DPDK library functions used in the sample code are prefixed with rte_ and are explained in detail in the DPDK API Documentation.

15.4.1. The Main Function

The main() function performs the initialization and calls the execution threads for each lcore.

The first task is to initialize the Environment Abstraction Layer (EAL). The argc and argv arguments are provided to the rte_eal_init() function. The value returned is the number of parsed arguments:

ret = rte_eal_init(argc, argv);
if (ret < 0)
	rte_exit(EXIT_FAILURE, "Invalid EAL arguments\n");

The main() also allocates a mempool to hold the mbufs (Message Buffers) used by the application:

nb_mbufs = RTE_MAX(nb_ports * (nb_queues * (nb_rxd + nb_txd +
	4 * MAX_PKT_BURST) + rte_lcore_count() * MEMPOOL_CACHE_SIZE),
	MIN_POOL_SIZE);

/* Create the mbuf pool */
ioat_pktmbuf_pool = rte_pktmbuf_pool_create("mbuf_pool", nb_mbufs,
	MEMPOOL_CACHE_SIZE, 0, RTE_MBUF_DEFAULT_BUF_SIZE,
	rte_socket_id());
if (ioat_pktmbuf_pool == NULL)
	rte_exit(EXIT_FAILURE, "Cannot init mbuf pool\n");

Mbufs are the packet buffer structure used by DPDK. They are explained in detail in the “Mbuf Library” section of the DPDK Programmer’s Guide.

The main() function also initializes the ports:

cfg.nb_ports = 0;
RTE_ETH_FOREACH_DEV(portid)
	port_init(portid, ioat_pktmbuf_pool, nb_queues);

Each port is configured using port_init() function. The Ethernet ports are configured with local settings using the rte_eth_dev_configure() function and the port_conf struct. The RSS is enabled so that multiple Rx queues could be used for packet receiving and copying by multiple CBDMA channels per port:

static const struct rte_eth_conf port_conf = {
	.rxmode = {
		.mq_mode = ETH_MQ_RX_RSS,
		.max_rx_pkt_len = RTE_ETHER_MAX_LEN
	},
	.rx_adv_conf = {
		.rss_conf = {
			.rss_key = NULL,
			.rss_hf = ETH_RSS_PROTO_MASK,
		}
	}
};

For this example the ports are set up with the number of Rx queues provided with -q option and 1 Tx queue using the rte_eth_rx_queue_setup() and rte_eth_tx_queue_setup() functions.

The Ethernet port is then started:

ret = rte_eth_dev_start(portid);
if (ret < 0)
	rte_exit(EXIT_FAILURE,
		"rte_eth_dev_start:err=%d, port=%u\n",
		ret, portid);

Finally the Rx port is set in promiscuous mode:

rte_eth_promiscuous_enable(portid);

After that each port application assigns resources needed.

while (!check_link_status(ioat_enabled_port_mask) && !force_quit)
	sleep(1);

/* Check if there is enough lcores for all ports. */
cfg.nb_lcores = rte_lcore_count() - 1;
if (cfg.nb_lcores < 1)
	rte_exit(EXIT_FAILURE,
		"There should be at least one worker lcore.\n");

if (copy_mode == COPY_MODE_IOAT_NUM)
	assign_rawdevs();
else /* copy_mode == COPY_MODE_SW_NUM */
	assign_rings();

Depending on mode set (whether copy should be done by software or by hardware) special structures are assigned to each port. If software copy was chosen, application have to assign ring structures for packet exchanging between lcores assigned to ports.

static void
assign_rings(void)
{
	uint32_t i;

	for (i = 0; i < cfg.nb_ports; i++) {
		char ring_name[RTE_RING_NAMESIZE];

		snprintf(ring_name, sizeof(ring_name), "rx_to_tx_ring_%u", i);
		/* Create ring for inter core communication */
		cfg.ports[i].rx_to_tx_ring = rte_ring_create(
			ring_name, ring_size,
			rte_socket_id(), RING_F_SP_ENQ | RING_F_SC_DEQ);

		if (cfg.ports[i].rx_to_tx_ring == NULL)
			rte_exit(EXIT_FAILURE, "Ring create failed: %s\n",
				rte_strerror(rte_errno));
	}
}

When using hardware copy each Rx queue of the port is assigned an IOAT device (assign_rawdevs()) using IOAT Rawdev Driver API functions:

static void
assign_rawdevs(void)
{
	uint16_t nb_rawdev = 0, rdev_id = 0;
	uint32_t i, j;

	for (i = 0; i < cfg.nb_ports; i++) {
		for (j = 0; j < cfg.ports[i].nb_queues; j++) {
			struct rte_rawdev_info rdev_info = { 0 };

			do {
				if (rdev_id == rte_rawdev_count())
					goto end;
				rte_rawdev_info_get(rdev_id++, &rdev_info, 0);
			} while (rdev_info.driver_name == NULL ||
					strcmp(rdev_info.driver_name,
						IOAT_PMD_RAWDEV_NAME_STR) != 0);

			cfg.ports[i].ioat_ids[j] = rdev_id - 1;
			configure_rawdev_queue(cfg.ports[i].ioat_ids[j]);
			++nb_rawdev;
		}
	}
end:
	if (nb_rawdev < cfg.nb_ports * cfg.ports[0].nb_queues)
		rte_exit(EXIT_FAILURE,
			"Not enough IOAT rawdevs (%u) for all queues (%u).\n",
			nb_rawdev, cfg.nb_ports * cfg.ports[0].nb_queues);
	RTE_LOG(INFO, IOAT, "Number of used rawdevs: %u.\n", nb_rawdev);
}

The initialization of hardware device is done by rte_rawdev_configure() function using rte_rawdev_info struct. After configuration the device is started using rte_rawdev_start() function. Each of the above operations is done in configure_rawdev_queue().

static void
configure_rawdev_queue(uint32_t dev_id)
{
	struct rte_ioat_rawdev_config dev_config = {
			.ring_size = ring_size,
			.no_prefetch_completions = (cfg.nb_lcores > 1),
	};
	struct rte_rawdev_info info = { .dev_private = &dev_config };

	if (rte_rawdev_configure(dev_id, &info, sizeof(dev_config)) != 0) {
		rte_exit(EXIT_FAILURE,
			"Error with rte_rawdev_configure()\n");
	}
	if (rte_rawdev_start(dev_id) != 0) {
		rte_exit(EXIT_FAILURE,
			"Error with rte_rawdev_start()\n");
	}
}

If initialization is successful, memory for hardware device statistics is allocated.

Finally main() function starts all packet handling lcores and starts printing stats in a loop on the main lcore. The application can be interrupted and closed using Ctrl-C. The main lcore waits for all worker lcores to finish, deallocates resources and exits.

The processing lcores launching function are described below.

15.4.2. The Lcores Launching Functions

As described above, main() function invokes start_forwarding_cores() function in order to start processing for each lcore:

static void start_forwarding_cores(void)
{
	uint32_t lcore_id = rte_lcore_id();

	RTE_LOG(INFO, IOAT, "Entering %s on lcore %u\n",
		__func__, rte_lcore_id());

	if (cfg.nb_lcores == 1) {
		lcore_id = rte_get_next_lcore(lcore_id, true, true);
		rte_eal_remote_launch((lcore_function_t *)rxtx_main_loop,
			NULL, lcore_id);
	} else if (cfg.nb_lcores > 1) {
		lcore_id = rte_get_next_lcore(lcore_id, true, true);
		rte_eal_remote_launch((lcore_function_t *)rx_main_loop,
			NULL, lcore_id);

		lcore_id = rte_get_next_lcore(lcore_id, true, true);
		rte_eal_remote_launch((lcore_function_t *)tx_main_loop, NULL,
			lcore_id);
	}
}

The function launches Rx/Tx processing functions on configured lcores using rte_eal_remote_launch(). The configured ports, their number and number of assigned lcores are stored in user-defined rxtx_transmission_config struct:

struct rxtx_transmission_config {
	struct rxtx_port_config ports[RTE_MAX_ETHPORTS];
	uint16_t nb_ports;
	uint16_t nb_lcores;
};

The structure is initialized in ‘main()’ function with the values corresponding to ports and lcores configuration provided by the user.

15.4.3. The Lcores Processing Functions

For receiving packets on each port, the ioat_rx_port() function is used. The function receives packets on each configured Rx queue. Depending on the mode the user chose, it will enqueue packets to IOAT rawdev channels and then invoke copy process (hardware copy), or perform software copy of each packet using pktmbuf_sw_copy() function and enqueue them to an rte_ring:

static void
ioat_rx_port(struct rxtx_port_config *rx_config)
{
	uint32_t nb_rx, nb_enq, i, j;
	struct rte_mbuf *pkts_burst[MAX_PKT_BURST];

	for (i = 0; i < rx_config->nb_queues; i++) {

		nb_rx = rte_eth_rx_burst(rx_config->rxtx_port, i,
			pkts_burst, MAX_PKT_BURST);

		if (nb_rx == 0)
			continue;

		port_statistics.rx[rx_config->rxtx_port] += nb_rx;

		if (copy_mode == COPY_MODE_IOAT_NUM) {
			/* Perform packet hardware copy */
			nb_enq = ioat_enqueue_packets(pkts_burst,
				nb_rx, rx_config->ioat_ids[i]);
			if (nb_enq > 0)
				rte_ioat_perform_ops(rx_config->ioat_ids[i]);
		} else {
			/* Perform packet software copy, free source packets */
			int ret;
			struct rte_mbuf *pkts_burst_copy[MAX_PKT_BURST];

			ret = rte_mempool_get_bulk(ioat_pktmbuf_pool,
				(void *)pkts_burst_copy, nb_rx);

			if (unlikely(ret < 0))
				rte_exit(EXIT_FAILURE,
					"Unable to allocate memory.\n");

			for (j = 0; j < nb_rx; j++)
				pktmbuf_sw_copy(pkts_burst[j],
					pkts_burst_copy[j]);

			rte_mempool_put_bulk(ioat_pktmbuf_pool,
				(void *)pkts_burst, nb_rx);

			nb_enq = rte_ring_enqueue_burst(
				rx_config->rx_to_tx_ring,
				(void *)pkts_burst_copy, nb_rx, NULL);

			/* Free any not enqueued packets. */
			rte_mempool_put_bulk(ioat_pktmbuf_pool,
				(void *)&pkts_burst_copy[nb_enq],
				nb_rx - nb_enq);
		}

		port_statistics.copy_dropped[rx_config->rxtx_port] +=
			(nb_rx - nb_enq);
	}
}

The packets are received in burst mode using rte_eth_rx_burst() function. When using hardware copy mode the packets are enqueued in copying device’s buffer using ioat_enqueue_packets() which calls rte_ioat_enqueue_copy(). When all received packets are in the buffer the copy operations are started by calling rte_ioat_perform_ops(). Function rte_ioat_enqueue_copy() operates on physical address of the packet. Structure rte_mbuf contains only physical address to start of the data buffer (buf_iova). Thus the address is adjusted by addr_offset value in order to get the address of rearm_data member of rte_mbuf. That way both the packet data and metadata can be copied in a single operation. This method can be used because the mbufs are direct mbufs allocated by the apps. If another app uses external buffers, or indirect mbufs, then multiple copy operations must be used.

static void
ioat_rx_port(struct rxtx_port_config *rx_config)
{
	uint32_t nb_rx, nb_enq, i, j;
	struct rte_mbuf *pkts_burst[MAX_PKT_BURST];

	for (i = 0; i < rx_config->nb_queues; i++) {

		nb_rx = rte_eth_rx_burst(rx_config->rxtx_port, i,
			pkts_burst, MAX_PKT_BURST);

		if (nb_rx == 0)
			continue;

		port_statistics.rx[rx_config->rxtx_port] += nb_rx;

		if (copy_mode == COPY_MODE_IOAT_NUM) {
			/* Perform packet hardware copy */
			nb_enq = ioat_enqueue_packets(pkts_burst,
				nb_rx, rx_config->ioat_ids[i]);
			if (nb_enq > 0)
				rte_ioat_perform_ops(rx_config->ioat_ids[i]);
		} else {
			/* Perform packet software copy, free source packets */
			int ret;
			struct rte_mbuf *pkts_burst_copy[MAX_PKT_BURST];

			ret = rte_mempool_get_bulk(ioat_pktmbuf_pool,
				(void *)pkts_burst_copy, nb_rx);

			if (unlikely(ret < 0))
				rte_exit(EXIT_FAILURE,
					"Unable to allocate memory.\n");

			for (j = 0; j < nb_rx; j++)
				pktmbuf_sw_copy(pkts_burst[j],
					pkts_burst_copy[j]);

			rte_mempool_put_bulk(ioat_pktmbuf_pool,
				(void *)pkts_burst, nb_rx);

			nb_enq = rte_ring_enqueue_burst(
				rx_config->rx_to_tx_ring,
				(void *)pkts_burst_copy, nb_rx, NULL);

			/* Free any not enqueued packets. */
			rte_mempool_put_bulk(ioat_pktmbuf_pool,
				(void *)&pkts_burst_copy[nb_enq],
				nb_rx - nb_enq);
		}

		port_statistics.copy_dropped[rx_config->rxtx_port] +=
			(nb_rx - nb_enq);
	}
}

All completed copies are processed by ioat_tx_port() function. When using hardware copy mode the function invokes rte_ioat_completed_ops() on each assigned IOAT channel to gather copied packets. If software copy mode is used the function dequeues copied packets from the rte_ring. Then each packet MAC address is changed if it was enabled. After that copies are sent in burst mode using `` rte_eth_tx_burst()``.

static void
ioat_tx_port(struct rxtx_port_config *tx_config)
{
	uint32_t i, j, nb_dq = 0;
	struct rte_mbuf *mbufs_src[MAX_PKT_BURST];
	struct rte_mbuf *mbufs_dst[MAX_PKT_BURST];

	for (i = 0; i < tx_config->nb_queues; i++) {
		if (copy_mode == COPY_MODE_IOAT_NUM) {
			/* Dequeue the mbufs from IOAT device. Since all memory
			 * is DPDK pinned memory and therefore all addresses should
			 * be valid, we don't check for copy errors
			 */
			nb_dq = rte_ioat_completed_ops(
				tx_config->ioat_ids[i], MAX_PKT_BURST, NULL, NULL,
				(void *)mbufs_src, (void *)mbufs_dst);
		} else {
			/* Dequeue the mbufs from rx_to_tx_ring. */
			nb_dq = rte_ring_dequeue_burst(
				tx_config->rx_to_tx_ring, (void *)mbufs_dst,
				MAX_PKT_BURST, NULL);
		}

		if ((int32_t) nb_dq <= 0)
			return;

		if (copy_mode == COPY_MODE_IOAT_NUM)
			rte_mempool_put_bulk(ioat_pktmbuf_pool,
				(void *)mbufs_src, nb_dq);

		/* Update macs if enabled */
		if (mac_updating) {
			for (j = 0; j < nb_dq; j++)
				update_mac_addrs(mbufs_dst[j],
					tx_config->rxtx_port);
		}

		const uint16_t nb_tx = rte_eth_tx_burst(
			tx_config->rxtx_port, 0,
			(void *)mbufs_dst, nb_dq);

		port_statistics.tx[tx_config->rxtx_port] += nb_tx;

		/* Free any unsent packets. */
		if (unlikely(nb_tx < nb_dq))
			rte_mempool_put_bulk(ioat_pktmbuf_pool,
			(void *)&mbufs_dst[nb_tx],
				nb_dq - nb_tx);
	}
}

15.4.4. The Packet Copying Functions

In order to perform packet copy there is a user-defined function pktmbuf_sw_copy() used. It copies a whole packet by copying metadata from source packet to new mbuf, and then copying a data chunk of source packet. Both memory copies are done using rte_memcpy():

static inline void
pktmbuf_sw_copy(struct rte_mbuf *src, struct rte_mbuf *dst)
{
	/* Copy packet metadata */
	rte_memcpy(&dst->rearm_data,
		&src->rearm_data,
		offsetof(struct rte_mbuf, cacheline1)
		- offsetof(struct rte_mbuf, rearm_data));

	/* Copy packet data */
	rte_memcpy(rte_pktmbuf_mtod(dst, char *),
		rte_pktmbuf_mtod(src, char *), src->data_len);
}

The metadata in this example is copied from rearm_data member of rte_mbuf struct up to cacheline1.

In order to understand why software packet copying is done as shown above please refer to the “Mbuf Library” section of the DPDK Programmer’s Guide.