perf_test_result(struct evt_test *test, struct evt_options *opt)
{
RTE_SET_USED(opt);
+ int i;
+ uint64_t total = 0;
struct test_perf *t = evt_test_priv(test);
+ printf("Packet distribution across worker cores :\n");
+ for (i = 0; i < t->nb_workers; i++)
+ total += t->worker[i].processed_pkts;
+ for (i = 0; i < t->nb_workers; i++)
+ printf("Worker %d packets: "CLGRN"%"PRIx64" "CLNRM"percentage:"
+ CLGRN" %3.2f\n"CLNRM, i,
+ t->worker[i].processed_pkts,
+ (((double)t->worker[i].processed_pkts)/total)
+ * 100);
+
return t->result;
}
return 0;
}
+static inline int
+perf_event_timer_producer(void *arg)
+{
+ struct prod_data *p = arg;
+ struct test_perf *t = p->t;
+ struct evt_options *opt = t->opt;
+ uint32_t flow_counter = 0;
+ uint64_t count = 0;
+ uint64_t arm_latency = 0;
+ const uint8_t nb_timer_adptrs = opt->nb_timer_adptrs;
+ const uint32_t nb_flows = t->nb_flows;
+ const uint64_t nb_timers = opt->nb_timers;
+ struct rte_mempool *pool = t->pool;
+ struct perf_elt *m;
+ struct rte_event_timer_adapter **adptr = t->timer_adptr;
+ struct rte_event_timer tim;
+ uint64_t timeout_ticks = opt->expiry_nsec / opt->timer_tick_nsec;
+
+ memset(&tim, 0, sizeof(struct rte_event_timer));
+ timeout_ticks = opt->optm_timer_tick_nsec ?
+ (timeout_ticks * opt->timer_tick_nsec)
+ / opt->optm_timer_tick_nsec : timeout_ticks;
+ timeout_ticks += timeout_ticks ? 0 : 1;
+ tim.ev.event_type = RTE_EVENT_TYPE_TIMER;
+ tim.ev.op = RTE_EVENT_OP_NEW;
+ tim.ev.sched_type = t->opt->sched_type_list[0];
+ tim.ev.queue_id = p->queue_id;
+ tim.ev.priority = RTE_EVENT_DEV_PRIORITY_NORMAL;
+ tim.state = RTE_EVENT_TIMER_NOT_ARMED;
+ tim.timeout_ticks = timeout_ticks;
+
+ if (opt->verbose_level > 1)
+ printf("%s(): lcore %d\n", __func__, rte_lcore_id());
+
+ while (count < nb_timers && t->done == false) {
+ if (rte_mempool_get(pool, (void **)&m) < 0)
+ continue;
+
+ m->tim = tim;
+ m->tim.ev.flow_id = flow_counter++ % nb_flows;
+ m->tim.ev.event_ptr = m;
+ m->timestamp = rte_get_timer_cycles();
+ while (rte_event_timer_arm_burst(
+ adptr[flow_counter % nb_timer_adptrs],
+ (struct rte_event_timer **)&m, 1) != 1) {
+ if (t->done)
+ break;
+ rte_pause();
+ m->timestamp = rte_get_timer_cycles();
+ }
+ arm_latency += rte_get_timer_cycles() - m->timestamp;
+ count++;
+ }
+ fflush(stdout);
+ rte_delay_ms(1000);
+ printf("%s(): lcore %d Average event timer arm latency = %.3f us\n",
+ __func__, rte_lcore_id(), (float)(arm_latency / count) /
+ (rte_get_timer_hz() / 1000000));
+ return 0;
+}
+
+static inline int
+perf_event_timer_producer_burst(void *arg)
+{
+ int i;
+ struct prod_data *p = arg;
+ struct test_perf *t = p->t;
+ struct evt_options *opt = t->opt;
+ uint32_t flow_counter = 0;
+ uint64_t count = 0;
+ uint64_t arm_latency = 0;
+ const uint8_t nb_timer_adptrs = opt->nb_timer_adptrs;
+ const uint32_t nb_flows = t->nb_flows;
+ const uint64_t nb_timers = opt->nb_timers;
+ struct rte_mempool *pool = t->pool;
+ struct perf_elt *m[BURST_SIZE + 1] = {NULL};
+ struct rte_event_timer_adapter **adptr = t->timer_adptr;
+ struct rte_event_timer tim;
+ uint64_t timeout_ticks = opt->expiry_nsec / opt->timer_tick_nsec;
+
+ memset(&tim, 0, sizeof(struct rte_event_timer));
+ timeout_ticks = opt->optm_timer_tick_nsec ?
+ (timeout_ticks * opt->timer_tick_nsec)
+ / opt->optm_timer_tick_nsec : timeout_ticks;
+ timeout_ticks += timeout_ticks ? 0 : 1;
+ tim.ev.event_type = RTE_EVENT_TYPE_TIMER;
+ tim.ev.op = RTE_EVENT_OP_NEW;
+ tim.ev.sched_type = t->opt->sched_type_list[0];
+ tim.ev.queue_id = p->queue_id;
+ tim.ev.priority = RTE_EVENT_DEV_PRIORITY_NORMAL;
+ tim.state = RTE_EVENT_TIMER_NOT_ARMED;
+ tim.timeout_ticks = timeout_ticks;
+
+ if (opt->verbose_level > 1)
+ printf("%s(): lcore %d\n", __func__, rte_lcore_id());
+
+ while (count < nb_timers && t->done == false) {
+ if (rte_mempool_get_bulk(pool, (void **)m, BURST_SIZE) < 0)
+ continue;
+ for (i = 0; i < BURST_SIZE; i++) {
+ rte_prefetch0(m[i + 1]);
+ m[i]->tim = tim;
+ m[i]->tim.ev.flow_id = flow_counter++ % nb_flows;
+ m[i]->tim.ev.event_ptr = m[i];
+ m[i]->timestamp = rte_get_timer_cycles();
+ }
+ rte_event_timer_arm_tmo_tick_burst(
+ adptr[flow_counter % nb_timer_adptrs],
+ (struct rte_event_timer **)m,
+ tim.timeout_ticks,
+ BURST_SIZE);
+ arm_latency += rte_get_timer_cycles() - m[i - 1]->timestamp;
+ count += BURST_SIZE;
+ }
+ fflush(stdout);
+ rte_delay_ms(1000);
+ printf("%s(): lcore %d Average event timer arm latency = %.3f us\n",
+ __func__, rte_lcore_id(), (float)(arm_latency / count) /
+ (rte_get_timer_hz() / 1000000));
+ return 0;
+}
+
+static int
+perf_producer_wrapper(void *arg)
+{
+ struct prod_data *p = arg;
+ struct test_perf *t = p->t;
+ /* Launch the producer function only in case of synthetic producer. */
+ if (t->opt->prod_type == EVT_PROD_TYPE_SYNT)
+ return perf_producer(arg);
+ else if (t->opt->prod_type == EVT_PROD_TYPE_EVENT_TIMER_ADPTR &&
+ !t->opt->timdev_use_burst)
+ return perf_event_timer_producer(arg);
+ else if (t->opt->prod_type == EVT_PROD_TYPE_EVENT_TIMER_ADPTR &&
+ t->opt->timdev_use_burst)
+ return perf_event_timer_producer_burst(arg);
+ return 0;
+}
+
static inline uint64_t
processed_pkts(struct test_perf *t)
{
if (!(opt->plcores[lcore_id]))
continue;
- ret = rte_eal_remote_launch(perf_producer, &t->prod[port_idx],
- lcore_id);
+ ret = rte_eal_remote_launch(perf_producer_wrapper,
+ &t->prod[port_idx], lcore_id);
if (ret) {
evt_err("failed to launch perf_producer %d", lcore_id);
return ret;
port_idx++;
}
- const uint64_t total_pkts = opt->nb_pkts *
- evt_nr_active_lcores(opt->plcores);
+ const uint64_t total_pkts = t->outstand_pkts;
uint64_t dead_lock_cycles = rte_get_timer_cycles();
int64_t dead_lock_remaining = total_pkts;
fflush(stdout);
if (remaining <= 0) {
- t->done = true;
t->result = EVT_TEST_SUCCESS;
- rte_smp_wmb();
- break;
+ if (opt->prod_type == EVT_PROD_TYPE_SYNT ||
+ opt->prod_type ==
+ EVT_PROD_TYPE_EVENT_TIMER_ADPTR) {
+ t->done = true;
+ rte_smp_wmb();
+ break;
+ }
}
}
- if (new_cycles - dead_lock_cycles > dead_lock_sample) {
+ if (new_cycles - dead_lock_cycles > dead_lock_sample &&
+ (opt->prod_type == EVT_PROD_TYPE_SYNT ||
+ opt->prod_type == EVT_PROD_TYPE_EVENT_TIMER_ADPTR)) {
remaining = t->outstand_pkts - processed_pkts(t);
if (dead_lock_remaining == remaining) {
rte_event_dev_dump(opt->dev_id, stdout);
return 0;
}
+static int
+perf_event_rx_adapter_setup(struct evt_options *opt, uint8_t stride,
+ struct rte_event_port_conf prod_conf)
+{
+ int ret = 0;
+ uint16_t prod;
+ struct rte_event_eth_rx_adapter_queue_conf queue_conf;
+
+ memset(&queue_conf, 0,
+ sizeof(struct rte_event_eth_rx_adapter_queue_conf));
+ queue_conf.ev.sched_type = opt->sched_type_list[0];
+ RTE_ETH_FOREACH_DEV(prod) {
+ uint32_t cap;
+
+ ret = rte_event_eth_rx_adapter_caps_get(opt->dev_id,
+ prod, &cap);
+ if (ret) {
+ evt_err("failed to get event rx adapter[%d]"
+ " capabilities",
+ opt->dev_id);
+ return ret;
+ }
+ queue_conf.ev.queue_id = prod * stride;
+ ret = rte_event_eth_rx_adapter_create(prod, opt->dev_id,
+ &prod_conf);
+ if (ret) {
+ evt_err("failed to create rx adapter[%d]", prod);
+ return ret;
+ }
+ ret = rte_event_eth_rx_adapter_queue_add(prod, prod, -1,
+ &queue_conf);
+ if (ret) {
+ evt_err("failed to add rx queues to adapter[%d]", prod);
+ return ret;
+ }
+
+ if (!(cap & RTE_EVENT_ETH_RX_ADAPTER_CAP_INTERNAL_PORT)) {
+ uint32_t service_id;
+
+ rte_event_eth_rx_adapter_service_id_get(prod,
+ &service_id);
+ ret = evt_service_setup(service_id);
+ if (ret) {
+ evt_err("Failed to setup service core"
+ " for Rx adapter\n");
+ return ret;
+ }
+ }
+ }
+
+ return ret;
+}
+
+static int
+perf_event_timer_adapter_setup(struct test_perf *t)
+{
+ int i;
+ int ret;
+ struct rte_event_timer_adapter_info adapter_info;
+ struct rte_event_timer_adapter *wl;
+ uint8_t nb_producers = evt_nr_active_lcores(t->opt->plcores);
+ uint8_t flags = RTE_EVENT_TIMER_ADAPTER_F_ADJUST_RES;
+
+ if (nb_producers == 1)
+ flags |= RTE_EVENT_TIMER_ADAPTER_F_SP_PUT;
+
+ for (i = 0; i < t->opt->nb_timer_adptrs; i++) {
+ struct rte_event_timer_adapter_conf config = {
+ .event_dev_id = t->opt->dev_id,
+ .timer_adapter_id = i,
+ .timer_tick_ns = t->opt->timer_tick_nsec,
+ .max_tmo_ns = t->opt->max_tmo_nsec,
+ .nb_timers = t->opt->pool_sz,
+ .flags = flags,
+ };
+
+ wl = rte_event_timer_adapter_create(&config);
+ if (wl == NULL) {
+ evt_err("failed to create event timer ring %d", i);
+ return rte_errno;
+ }
+
+ memset(&adapter_info, 0,
+ sizeof(struct rte_event_timer_adapter_info));
+ rte_event_timer_adapter_get_info(wl, &adapter_info);
+ t->opt->optm_timer_tick_nsec = adapter_info.min_resolution_ns;
+
+ if (!(adapter_info.caps &
+ RTE_EVENT_TIMER_ADAPTER_CAP_INTERNAL_PORT)) {
+ uint32_t service_id;
+
+ rte_event_timer_adapter_service_id_get(wl,
+ &service_id);
+ ret = evt_service_setup(service_id);
+ if (ret) {
+ evt_err("Failed to setup service core"
+ " for timer adapter\n");
+ return ret;
+ }
+ rte_service_runstate_set(service_id, 1);
+ }
+ t->timer_adptr[i] = wl;
+ }
+ return 0;
+}
+
int
perf_event_dev_port_setup(struct evt_test *test, struct evt_options *opt,
- uint8_t stride, uint8_t nb_queues)
+ uint8_t stride, uint8_t nb_queues,
+ const struct rte_event_port_conf *port_conf)
{
struct test_perf *t = evt_test_priv(test);
- uint8_t port, prod;
+ uint16_t port, prod;
int ret = -1;
- /* port configuration */
- const struct rte_event_port_conf wkr_p_conf = {
- .dequeue_depth = opt->wkr_deq_dep,
- .enqueue_depth = 64,
- .new_event_threshold = 4096,
- };
-
/* setup one port per worker, linking to all queues */
for (port = 0; port < evt_nr_active_lcores(opt->wlcores);
port++) {
w->processed_pkts = 0;
w->latency = 0;
- ret = rte_event_port_setup(opt->dev_id, port, &wkr_p_conf);
+ ret = rte_event_port_setup(opt->dev_id, port, port_conf);
if (ret) {
evt_err("failed to setup port %d", port);
return ret;
}
/* port for producers, no links */
- const struct rte_event_port_conf prod_conf = {
- .dequeue_depth = 8,
- .enqueue_depth = 32,
- .new_event_threshold = 1200,
- };
- prod = 0;
- for ( ; port < perf_nb_event_ports(opt); port++) {
- struct prod_data *p = &t->prod[port];
+ if (opt->prod_type == EVT_PROD_TYPE_ETH_RX_ADPTR) {
+ for ( ; port < perf_nb_event_ports(opt); port++) {
+ struct prod_data *p = &t->prod[port];
+ p->t = t;
+ }
- p->dev_id = opt->dev_id;
- p->port_id = port;
- p->queue_id = prod * stride;
- p->t = t;
+ ret = perf_event_rx_adapter_setup(opt, stride, *port_conf);
+ if (ret)
+ return ret;
+ } else if (opt->prod_type == EVT_PROD_TYPE_EVENT_TIMER_ADPTR) {
+ prod = 0;
+ for ( ; port < perf_nb_event_ports(opt); port++) {
+ struct prod_data *p = &t->prod[port];
+ p->queue_id = prod * stride;
+ p->t = t;
+ prod++;
+ }
- ret = rte_event_port_setup(opt->dev_id, port, &prod_conf);
- if (ret) {
- evt_err("failed to setup port %d", port);
+ ret = perf_event_timer_adapter_setup(t);
+ if (ret)
return ret;
+ } else {
+ prod = 0;
+ for ( ; port < perf_nb_event_ports(opt); port++) {
+ struct prod_data *p = &t->prod[port];
+
+ p->dev_id = opt->dev_id;
+ p->port_id = port;
+ p->queue_id = prod * stride;
+ p->t = t;
+
+ ret = rte_event_port_setup(opt->dev_id, port,
+ port_conf);
+ if (ret) {
+ evt_err("failed to setup port %d", port);
+ return ret;
+ }
+ prod++;
}
- prod++;
}
return ret;
}
/* Fixups */
- if (opt->nb_stages == 1 && opt->fwd_latency) {
+ if ((opt->nb_stages == 1 &&
+ opt->prod_type != EVT_PROD_TYPE_EVENT_TIMER_ADPTR) &&
+ opt->fwd_latency) {
evt_info("fwd_latency is valid when nb_stages > 1, disabling");
opt->fwd_latency = 0;
}
+
if (opt->fwd_latency && !opt->q_priority) {
evt_info("enabled queue priority for latency measurement");
opt->q_priority = 1;
void
perf_eventdev_destroy(struct evt_test *test, struct evt_options *opt)
{
- RTE_SET_USED(test);
+ int i;
+ struct test_perf *t = evt_test_priv(test);
+ if (opt->prod_type == EVT_PROD_TYPE_EVENT_TIMER_ADPTR) {
+ for (i = 0; i < opt->nb_timer_adptrs; i++)
+ rte_event_timer_adapter_stop(t->timer_adptr[i]);
+ }
rte_event_dev_stop(opt->dev_id);
rte_event_dev_close(opt->dev_id);
}
memset(obj, 0, mp->elt_size);
}
+#define NB_RX_DESC 128
+#define NB_TX_DESC 512
+int
+perf_ethdev_setup(struct evt_test *test, struct evt_options *opt)
+{
+ uint16_t i;
+ struct test_perf *t = evt_test_priv(test);
+ struct rte_eth_conf port_conf = {
+ .rxmode = {
+ .mq_mode = ETH_MQ_RX_RSS,
+ .max_rx_pkt_len = RTE_ETHER_MAX_LEN,
+ .split_hdr_size = 0,
+ },
+ .rx_adv_conf = {
+ .rss_conf = {
+ .rss_key = NULL,
+ .rss_hf = ETH_RSS_IP,
+ },
+ },
+ };
+
+ if (opt->prod_type == EVT_PROD_TYPE_SYNT ||
+ opt->prod_type == EVT_PROD_TYPE_EVENT_TIMER_ADPTR)
+ return 0;
+
+ if (!rte_eth_dev_count_avail()) {
+ evt_err("No ethernet ports found.");
+ return -ENODEV;
+ }
+
+ RTE_ETH_FOREACH_DEV(i) {
+ struct rte_eth_dev_info dev_info;
+ struct rte_eth_conf local_port_conf = port_conf;
+
+ rte_eth_dev_info_get(i, &dev_info);
+
+ local_port_conf.rx_adv_conf.rss_conf.rss_hf &=
+ dev_info.flow_type_rss_offloads;
+ if (local_port_conf.rx_adv_conf.rss_conf.rss_hf !=
+ port_conf.rx_adv_conf.rss_conf.rss_hf) {
+ evt_info("Port %u modified RSS hash function based on hardware support,"
+ "requested:%#"PRIx64" configured:%#"PRIx64"\n",
+ i,
+ port_conf.rx_adv_conf.rss_conf.rss_hf,
+ local_port_conf.rx_adv_conf.rss_conf.rss_hf);
+ }
+
+ if (rte_eth_dev_configure(i, 1, 1, &local_port_conf) < 0) {
+ evt_err("Failed to configure eth port [%d]", i);
+ return -EINVAL;
+ }
+
+ if (rte_eth_rx_queue_setup(i, 0, NB_RX_DESC,
+ rte_socket_id(), NULL, t->pool) < 0) {
+ evt_err("Failed to setup eth port [%d] rx_queue: %d.",
+ i, 0);
+ return -EINVAL;
+ }
+
+ if (rte_eth_tx_queue_setup(i, 0, NB_TX_DESC,
+ rte_socket_id(), NULL) < 0) {
+ evt_err("Failed to setup eth port [%d] tx_queue: %d.",
+ i, 0);
+ return -EINVAL;
+ }
+
+ rte_eth_promiscuous_enable(i);
+ }
+
+ return 0;
+}
+
+void perf_ethdev_destroy(struct evt_test *test, struct evt_options *opt)
+{
+ uint16_t i;
+ RTE_SET_USED(test);
+
+ if (opt->prod_type == EVT_PROD_TYPE_ETH_RX_ADPTR) {
+ RTE_ETH_FOREACH_DEV(i) {
+ rte_event_eth_rx_adapter_stop(i);
+ rte_eth_dev_stop(i);
+ }
+ }
+}
+
int
perf_mempool_setup(struct evt_test *test, struct evt_options *opt)
{
struct test_perf *t = evt_test_priv(test);
- t->pool = rte_mempool_create(test->name, /* mempool name */
+ if (opt->prod_type == EVT_PROD_TYPE_SYNT ||
+ opt->prod_type == EVT_PROD_TYPE_EVENT_TIMER_ADPTR) {
+ t->pool = rte_mempool_create(test->name, /* mempool name */
opt->pool_sz, /* number of elements*/
sizeof(struct perf_elt), /* element size*/
512, /* cache size*/
0, NULL, NULL,
perf_elt_init, /* obj constructor */
NULL, opt->socket_id, 0); /* flags */
+ } else {
+ t->pool = rte_pktmbuf_pool_create(test->name, /* mempool name */
+ opt->pool_sz, /* number of elements*/
+ 512, /* cache size*/
+ 0,
+ RTE_MBUF_DEFAULT_BUF_SIZE,
+ opt->socket_id); /* flags */
+
+ }
+
if (t->pool == NULL) {
evt_err("failed to create mempool");
return -ENOMEM;
struct test_perf *t = evt_test_priv(test);
- t->outstand_pkts = opt->nb_pkts * evt_nr_active_lcores(opt->plcores);
+ if (opt->prod_type == EVT_PROD_TYPE_EVENT_TIMER_ADPTR) {
+ t->outstand_pkts = opt->nb_timers *
+ evt_nr_active_lcores(opt->plcores);
+ t->nb_pkts = opt->nb_timers;
+ } else {
+ t->outstand_pkts = opt->nb_pkts *
+ evt_nr_active_lcores(opt->plcores);
+ t->nb_pkts = opt->nb_pkts;
+ }
+
t->nb_workers = evt_nr_active_lcores(opt->wlcores);
t->done = false;
- t->nb_pkts = opt->nb_pkts;
t->nb_flows = opt->nb_flows;
t->result = EVT_TEST_FAILED;
t->opt = opt;
git.droids-corp.org / dpdk.git / blobdiff