#include <stdint.h>
#include <inttypes.h>
+#include <getopt.h>
#include <rte_eal.h>
#include <rte_ethdev.h>
#include <rte_cycles.h>
#include <rte_lcore.h>
#include <rte_mbuf.h>
+#include <rte_mbuf_dyn.h>
#define RX_RING_SIZE 1024
#define TX_RING_SIZE 1024
#define MBUF_CACHE_SIZE 250
#define BURST_SIZE 32
+static int hwts_dynfield_offset = -1;
+
+static inline rte_mbuf_timestamp_t *
+hwts_field(struct rte_mbuf *mbuf)
+{
+ return RTE_MBUF_DYNFIELD(mbuf,
+ hwts_dynfield_offset, rte_mbuf_timestamp_t *);
+}
+
+typedef uint64_t tsc_t;
+static int tsc_dynfield_offset = -1;
+
+static inline tsc_t *
+tsc_field(struct rte_mbuf *mbuf)
+{
+ return RTE_MBUF_DYNFIELD(mbuf, tsc_dynfield_offset, tsc_t *);
+}
+
+static const char usage[] =
+ "%s EAL_ARGS -- [-t]\n";
+
static const struct rte_eth_conf port_conf_default = {
.rxmode = {
- .max_rx_pkt_len = ETHER_MAX_LEN,
- .ignore_offload_bitfield = 1,
+ .max_rx_pkt_len = RTE_ETHER_MAX_LEN,
},
};
static struct {
uint64_t total_cycles;
+ uint64_t total_queue_cycles;
uint64_t total_pkts;
} latency_numbers;
+int hw_timestamping;
+
+#define TICKS_PER_CYCLE_SHIFT 16
+static uint64_t ticks_per_cycle_mult;
+/* Callback added to the RX port and applied to packets. 8< */
static uint16_t
add_timestamps(uint16_t port __rte_unused, uint16_t qidx __rte_unused,
struct rte_mbuf **pkts, uint16_t nb_pkts,
uint64_t now = rte_rdtsc();
for (i = 0; i < nb_pkts; i++)
- pkts[i]->udata64 = now;
+ *tsc_field(pkts[i]) = now;
return nb_pkts;
}
+/* >8 End of callback addition and application. */
+/* Callback is added to the TX port. 8< */
static uint16_t
-calc_latency(uint16_t port __rte_unused, uint16_t qidx __rte_unused,
+calc_latency(uint16_t port, uint16_t qidx __rte_unused,
struct rte_mbuf **pkts, uint16_t nb_pkts, void *_ __rte_unused)
{
uint64_t cycles = 0;
+ uint64_t queue_ticks = 0;
uint64_t now = rte_rdtsc();
+ uint64_t ticks;
unsigned i;
- for (i = 0; i < nb_pkts; i++)
- cycles += now - pkts[i]->udata64;
+ if (hw_timestamping)
+ rte_eth_read_clock(port, &ticks);
+
+ for (i = 0; i < nb_pkts; i++) {
+ cycles += now - *tsc_field(pkts[i]);
+ if (hw_timestamping)
+ queue_ticks += ticks - *hwts_field(pkts[i]);
+ }
+
latency_numbers.total_cycles += cycles;
+ if (hw_timestamping)
+ latency_numbers.total_queue_cycles += (queue_ticks
+ * ticks_per_cycle_mult) >> TICKS_PER_CYCLE_SHIFT;
+
latency_numbers.total_pkts += nb_pkts;
if (latency_numbers.total_pkts > (100 * 1000 * 1000ULL)) {
printf("Latency = %"PRIu64" cycles\n",
latency_numbers.total_cycles / latency_numbers.total_pkts);
- latency_numbers.total_cycles = latency_numbers.total_pkts = 0;
+ if (hw_timestamping) {
+ printf("Latency from HW = %"PRIu64" cycles\n",
+ latency_numbers.total_queue_cycles
+ / latency_numbers.total_pkts);
+ }
+ latency_numbers.total_cycles = 0;
+ latency_numbers.total_queue_cycles = 0;
+ latency_numbers.total_pkts = 0;
}
return nb_pkts;
}
+/* >8 End of callback addition. */
/*
* Initialises a given port using global settings and with the rx buffers
* coming from the mbuf_pool passed as parameter
*/
+
+ /* Port initialization. 8< */
static inline int
port_init(uint16_t port, struct rte_mempool *mbuf_pool)
{
int retval;
uint16_t q;
struct rte_eth_dev_info dev_info;
+ struct rte_eth_rxconf rxconf;
struct rte_eth_txconf txconf;
- if (port >= rte_eth_dev_count())
+ if (!rte_eth_dev_is_valid_port(port))
return -1;
- rte_eth_dev_info_get(port, &dev_info);
+ retval = rte_eth_dev_info_get(port, &dev_info);
+ if (retval != 0) {
+ printf("Error during getting device (port %u) info: %s\n",
+ port, strerror(-retval));
+
+ return retval;
+ }
+
if (dev_info.tx_offload_capa & DEV_TX_OFFLOAD_MBUF_FAST_FREE)
port_conf.txmode.offloads |=
DEV_TX_OFFLOAD_MBUF_FAST_FREE;
+ if (hw_timestamping) {
+ if (!(dev_info.rx_offload_capa & DEV_RX_OFFLOAD_TIMESTAMP)) {
+ printf("\nERROR: Port %u does not support hardware timestamping\n"
+ , port);
+ return -1;
+ }
+ port_conf.rxmode.offloads |= DEV_RX_OFFLOAD_TIMESTAMP;
+ rte_mbuf_dyn_rx_timestamp_register(&hwts_dynfield_offset, NULL);
+ if (hwts_dynfield_offset < 0) {
+ printf("ERROR: Failed to register timestamp field\n");
+ return -rte_errno;
+ }
+ }
+
retval = rte_eth_dev_configure(port, rx_rings, tx_rings, &port_conf);
if (retval != 0)
return retval;
if (retval != 0)
return retval;
+ rxconf = dev_info.default_rxconf;
+
for (q = 0; q < rx_rings; q++) {
retval = rte_eth_rx_queue_setup(port, q, nb_rxd,
- rte_eth_dev_socket_id(port), NULL, mbuf_pool);
+ rte_eth_dev_socket_id(port), &rxconf, mbuf_pool);
if (retval < 0)
return retval;
}
txconf = dev_info.default_txconf;
- txconf.txq_flags = ETH_TXQ_FLAGS_IGNORE;
txconf.offloads = port_conf.txmode.offloads;
for (q = 0; q < tx_rings; q++) {
retval = rte_eth_tx_queue_setup(port, q, nb_txd,
if (retval < 0)
return retval;
- struct ether_addr addr;
+ if (hw_timestamping && ticks_per_cycle_mult == 0) {
+ uint64_t cycles_base = rte_rdtsc();
+ uint64_t ticks_base;
+ retval = rte_eth_read_clock(port, &ticks_base);
+ if (retval != 0)
+ return retval;
+ rte_delay_ms(100);
+ uint64_t cycles = rte_rdtsc();
+ uint64_t ticks;
+ rte_eth_read_clock(port, &ticks);
+ uint64_t c_freq = cycles - cycles_base;
+ uint64_t t_freq = ticks - ticks_base;
+ double freq_mult = (double)c_freq / t_freq;
+ printf("TSC Freq ~= %" PRIu64
+ "\nHW Freq ~= %" PRIu64
+ "\nRatio : %f\n",
+ c_freq * 10, t_freq * 10, freq_mult);
+ /* TSC will be faster than internal ticks so freq_mult is > 0
+ * We convert the multiplication to an integer shift & mult
+ */
+ ticks_per_cycle_mult = (1 << TICKS_PER_CYCLE_SHIFT) / freq_mult;
+ }
- rte_eth_macaddr_get(port, &addr);
+ struct rte_ether_addr addr;
+
+ retval = rte_eth_macaddr_get(port, &addr);
+ if (retval < 0) {
+ printf("Failed to get MAC address on port %u: %s\n",
+ port, rte_strerror(-retval));
+ return retval;
+ }
printf("Port %u MAC: %02"PRIx8" %02"PRIx8" %02"PRIx8
" %02"PRIx8" %02"PRIx8" %02"PRIx8"\n",
(unsigned)port,
- addr.addr_bytes[0], addr.addr_bytes[1],
- addr.addr_bytes[2], addr.addr_bytes[3],
- addr.addr_bytes[4], addr.addr_bytes[5]);
+ RTE_ETHER_ADDR_BYTES(&addr));
- rte_eth_promiscuous_enable(port);
+ retval = rte_eth_promiscuous_enable(port);
+ if (retval != 0)
+ return retval;
+
+ /* RX and TX callbacks are added to the ports. 8< */
rte_eth_add_rx_callback(port, 0, add_timestamps, NULL);
rte_eth_add_tx_callback(port, 0, calc_latency, NULL);
+ /* >8 End of RX and TX callbacks. */
return 0;
}
+/* >8 End of port initialization. */
/*
* Main thread that does the work, reading from INPUT_PORT
* and writing to OUTPUT_PORT
*/
-static __attribute__((noreturn)) void
+static __rte_noreturn void
lcore_main(void)
{
uint16_t port;
struct rte_mempool *mbuf_pool;
uint16_t nb_ports;
uint16_t portid;
+ struct option lgopts[] = {
+ { NULL, 0, 0, 0 }
+ };
+ int opt, option_index;
+
+ static const struct rte_mbuf_dynfield tsc_dynfield_desc = {
+ .name = "example_bbdev_dynfield_tsc",
+ .size = sizeof(tsc_t),
+ .align = __alignof__(tsc_t),
+ };
/* init EAL */
int ret = rte_eal_init(argc, argv);
argc -= ret;
argv += ret;
- nb_ports = rte_eth_dev_count();
+ while ((opt = getopt_long(argc, argv, "t", lgopts, &option_index))
+ != EOF)
+ switch (opt) {
+ case 't':
+ hw_timestamping = 1;
+ break;
+ default:
+ printf(usage, argv[0]);
+ return -1;
+ }
+ optind = 1; /* reset getopt lib */
+
+ nb_ports = rte_eth_dev_count_avail();
if (nb_ports < 2 || (nb_ports & 1))
rte_exit(EXIT_FAILURE, "Error: number of ports must be even\n");
if (mbuf_pool == NULL)
rte_exit(EXIT_FAILURE, "Cannot create mbuf pool\n");
+ tsc_dynfield_offset =
+ rte_mbuf_dynfield_register(&tsc_dynfield_desc);
+ if (tsc_dynfield_offset < 0)
+ rte_exit(EXIT_FAILURE, "Cannot register mbuf field\n");
+
/* initialize all ports */
RTE_ETH_FOREACH_DEV(portid)
if (port_init(portid, mbuf_pool) != 0)
- rte_exit(EXIT_FAILURE, "Cannot init port %"PRIu8"\n",
+ rte_exit(EXIT_FAILURE, "Cannot init port %"PRIu16"\n",
portid);
if (rte_lcore_count() > 1)
printf("\nWARNING: Too much enabled lcores - "
"App uses only 1 lcore\n");
- /* call lcore_main on master core only */
+ /* call lcore_main on main core only */
lcore_main();
+
+ /* clean up the EAL */
+ rte_eal_cleanup();
+
return 0;
}
git.droids-corp.org / dpdk.git / blobdiff