/*- * BSD LICENSE * * Copyright (c) 2016-2017 Solarflare Communications Inc. * All rights reserved. * * This software was jointly developed between OKTET Labs (under contract * for Solarflare) and Solarflare Communications, Inc. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions are met: * * 1. Redistributions of source code must retain the above copyright notice, * this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright notice, * this list of conditions and the following disclaimer in the documentation * and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, * THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, * EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include #include #include #include #include "efx.h" #include "sfc.h" #include "sfc_debug.h" #include "sfc_log.h" #include "sfc_ev.h" #include "sfc_rx.h" #include "sfc_tx.h" #include "sfc_kvargs.h" /* Initial delay when waiting for event queue init complete event */ #define SFC_EVQ_INIT_BACKOFF_START_US (1) /* Maximum delay between event queue polling attempts */ #define SFC_EVQ_INIT_BACKOFF_MAX_US (10 * 1000) /* Event queue init approx timeout */ #define SFC_EVQ_INIT_TIMEOUT_US (2 * US_PER_S) /* Management event queue polling period in microseconds */ #define SFC_MGMT_EV_QPOLL_PERIOD_US (US_PER_S) static boolean_t sfc_ev_initialized(void *arg) { struct sfc_evq *evq = arg; /* Init done events may be duplicated on SFN7xxx (SFC bug 31631) */ SFC_ASSERT(evq->init_state == SFC_EVQ_STARTING || evq->init_state == SFC_EVQ_STARTED); evq->init_state = SFC_EVQ_STARTED; return B_FALSE; } static boolean_t sfc_ev_rx(void *arg, __rte_unused uint32_t label, uint32_t id, uint32_t size, uint16_t flags) { struct sfc_evq *evq = arg; struct sfc_rxq *rxq; unsigned int stop; unsigned int pending_id; unsigned int delta; unsigned int i; struct sfc_rx_sw_desc *rxd; if (unlikely(evq->exception)) goto done; rxq = evq->rxq; SFC_ASSERT(rxq != NULL); SFC_ASSERT(rxq->evq == evq); SFC_ASSERT(rxq->state & SFC_RXQ_STARTED); stop = (id + 1) & rxq->ptr_mask; pending_id = rxq->pending & rxq->ptr_mask; delta = (stop >= pending_id) ? (stop - pending_id) : (rxq->ptr_mask + 1 - pending_id + stop); if (delta == 0) { /* * Rx event with no new descriptors done and zero length * is used to abort scattered packet when there is no room * for the tail. */ if (unlikely(size != 0)) { evq->exception = B_TRUE; sfc_err(evq->sa, "EVQ %u RxQ %u invalid RX abort " "(id=%#x size=%u flags=%#x); needs restart", evq->evq_index, sfc_rxq_sw_index(rxq), id, size, flags); goto done; } /* Add discard flag to the first fragment */ rxq->sw_desc[pending_id].flags |= EFX_DISCARD; /* Remove continue flag from the last fragment */ rxq->sw_desc[id].flags &= ~EFX_PKT_CONT; } else if (unlikely(delta > rxq->batch_max)) { evq->exception = B_TRUE; sfc_err(evq->sa, "EVQ %u RxQ %u completion out of order " "(id=%#x delta=%u flags=%#x); needs restart", evq->evq_index, sfc_rxq_sw_index(rxq), id, delta, flags); goto done; } for (i = pending_id; i != stop; i = (i + 1) & rxq->ptr_mask) { rxd = &rxq->sw_desc[i]; rxd->flags = flags; SFC_ASSERT(size < (1 << 16)); rxd->size = (uint16_t)size; } rxq->pending += delta; done: return B_FALSE; } static boolean_t sfc_ev_tx(void *arg, __rte_unused uint32_t label, uint32_t id) { struct sfc_evq *evq = arg; struct sfc_txq *txq; unsigned int stop; unsigned int delta; txq = evq->txq; SFC_ASSERT(txq != NULL); SFC_ASSERT(txq->evq == evq); if (unlikely((txq->state & SFC_TXQ_STARTED) == 0)) goto done; stop = (id + 1) & txq->ptr_mask; id = txq->pending & txq->ptr_mask; delta = (stop >= id) ? (stop - id) : (txq->ptr_mask + 1 - id + stop); txq->pending += delta; done: return B_FALSE; } static boolean_t sfc_ev_exception(void *arg, __rte_unused uint32_t code, __rte_unused uint32_t data) { struct sfc_evq *evq = arg; if (code == EFX_EXCEPTION_UNKNOWN_SENSOREVT) return B_FALSE; evq->exception = B_TRUE; sfc_warn(evq->sa, "hardware exception %s (code=%u, data=%#x) on EVQ %u;" " needs recovery", (code == EFX_EXCEPTION_RX_RECOVERY) ? "RX_RECOVERY" : (code == EFX_EXCEPTION_RX_DSC_ERROR) ? "RX_DSC_ERROR" : (code == EFX_EXCEPTION_TX_DSC_ERROR) ? "TX_DSC_ERROR" : (code == EFX_EXCEPTION_FWALERT_SRAM) ? "FWALERT_SRAM" : (code == EFX_EXCEPTION_UNKNOWN_FWALERT) ? "UNKNOWN_FWALERT" : (code == EFX_EXCEPTION_RX_ERROR) ? "RX_ERROR" : (code == EFX_EXCEPTION_TX_ERROR) ? "TX_ERROR" : (code == EFX_EXCEPTION_EV_ERROR) ? "EV_ERROR" : "UNKNOWN", code, data, evq->evq_index); return B_TRUE; } static boolean_t sfc_ev_rxq_flush_done(void *arg, __rte_unused uint32_t rxq_hw_index) { struct sfc_evq *evq = arg; struct sfc_rxq *rxq; rxq = evq->rxq; SFC_ASSERT(rxq != NULL); SFC_ASSERT(rxq->hw_index == rxq_hw_index); SFC_ASSERT(rxq->evq == evq); sfc_rx_qflush_done(rxq); return B_FALSE; } static boolean_t sfc_ev_rxq_flush_failed(void *arg, __rte_unused uint32_t rxq_hw_index) { struct sfc_evq *evq = arg; struct sfc_rxq *rxq; rxq = evq->rxq; SFC_ASSERT(rxq != NULL); SFC_ASSERT(rxq->hw_index == rxq_hw_index); SFC_ASSERT(rxq->evq == evq); sfc_rx_qflush_failed(rxq); return B_FALSE; } static boolean_t sfc_ev_txq_flush_done(void *arg, __rte_unused uint32_t txq_hw_index) { struct sfc_evq *evq = arg; struct sfc_txq *txq; txq = evq->txq; SFC_ASSERT(txq != NULL); SFC_ASSERT(txq->hw_index == txq_hw_index); SFC_ASSERT(txq->evq == evq); sfc_tx_qflush_done(txq); return B_FALSE; } static boolean_t sfc_ev_software(void *arg, uint16_t magic) { struct sfc_evq *evq = arg; sfc_err(evq->sa, "EVQ %u unexpected software event magic=%#.4x", evq->evq_index, magic); return B_TRUE; } static boolean_t sfc_ev_sram(void *arg, uint32_t code) { struct sfc_evq *evq = arg; sfc_err(evq->sa, "EVQ %u unexpected SRAM event code=%u", evq->evq_index, code); return B_TRUE; } static boolean_t sfc_ev_wake_up(void *arg, uint32_t index) { struct sfc_evq *evq = arg; sfc_err(evq->sa, "EVQ %u unexpected wake up event index=%u", evq->evq_index, index); return B_TRUE; } static boolean_t sfc_ev_timer(void *arg, uint32_t index) { struct sfc_evq *evq = arg; sfc_err(evq->sa, "EVQ %u unexpected timer event index=%u", evq->evq_index, index); return B_TRUE; } static boolean_t sfc_ev_link_change(void *arg, efx_link_mode_t link_mode) { struct sfc_evq *evq = arg; struct sfc_adapter *sa = evq->sa; struct rte_eth_link *dev_link = &sa->eth_dev->data->dev_link; struct rte_eth_link new_link; uint64_t new_link_u64; uint64_t old_link_u64; EFX_STATIC_ASSERT(sizeof(*dev_link) == sizeof(rte_atomic64_t)); sfc_port_link_mode_to_info(link_mode, &new_link); new_link_u64 = *(uint64_t *)&new_link; do { old_link_u64 = rte_atomic64_read((rte_atomic64_t *)dev_link); if (old_link_u64 == new_link_u64) break; if (rte_atomic64_cmpset((volatile uint64_t *)dev_link, old_link_u64, new_link_u64)) { evq->sa->port.lsc_seq++; break; } } while (B_TRUE); return B_FALSE; } static const efx_ev_callbacks_t sfc_ev_callbacks = { .eec_initialized = sfc_ev_initialized, .eec_rx = sfc_ev_rx, .eec_tx = sfc_ev_tx, .eec_exception = sfc_ev_exception, .eec_rxq_flush_done = sfc_ev_rxq_flush_done, .eec_rxq_flush_failed = sfc_ev_rxq_flush_failed, .eec_txq_flush_done = sfc_ev_txq_flush_done, .eec_software = sfc_ev_software, .eec_sram = sfc_ev_sram, .eec_wake_up = sfc_ev_wake_up, .eec_timer = sfc_ev_timer, .eec_link_change = sfc_ev_link_change, }; void sfc_ev_qpoll(struct sfc_evq *evq) { SFC_ASSERT(evq->init_state == SFC_EVQ_STARTED || evq->init_state == SFC_EVQ_STARTING); /* Synchronize the DMA memory for reading not required */ efx_ev_qpoll(evq->common, &evq->read_ptr, &sfc_ev_callbacks, evq); if (unlikely(evq->exception) && sfc_adapter_trylock(evq->sa)) { struct sfc_adapter *sa = evq->sa; int rc; if ((evq->rxq != NULL) && (evq->rxq->state & SFC_RXQ_RUNNING)) { unsigned int rxq_sw_index = sfc_rxq_sw_index(evq->rxq); sfc_warn(sa, "restart RxQ %u because of exception on its EvQ %u", rxq_sw_index, evq->evq_index); sfc_rx_qstop(sa, rxq_sw_index); rc = sfc_rx_qstart(sa, rxq_sw_index); if (rc != 0) sfc_err(sa, "cannot restart RxQ %u", rxq_sw_index); } if (evq->txq != NULL) { unsigned int txq_sw_index = sfc_txq_sw_index(evq->txq); sfc_warn(sa, "restart TxQ %u because of exception on its EvQ %u", txq_sw_index, evq->evq_index); sfc_tx_qstop(sa, txq_sw_index); rc = sfc_tx_qstart(sa, txq_sw_index); if (rc != 0) sfc_err(sa, "cannot restart TxQ %u", txq_sw_index); } if (evq->exception) sfc_panic(sa, "unrecoverable exception on EvQ %u", evq->evq_index); sfc_adapter_unlock(sa); } /* Poll-mode driver does not re-prime the event queue for interrupts */ } void sfc_ev_mgmt_qpoll(struct sfc_adapter *sa) { if (rte_spinlock_trylock(&sa->mgmt_evq_lock)) { struct sfc_evq *mgmt_evq = sa->evq_info[sa->mgmt_evq_index].evq; if (mgmt_evq->init_state == SFC_EVQ_STARTED) sfc_ev_qpoll(mgmt_evq); rte_spinlock_unlock(&sa->mgmt_evq_lock); } } int sfc_ev_qprime(struct sfc_evq *evq) { SFC_ASSERT(evq->init_state == SFC_EVQ_STARTED); return efx_ev_qprime(evq->common, evq->read_ptr); } int sfc_ev_qstart(struct sfc_adapter *sa, unsigned int sw_index) { const struct sfc_evq_info *evq_info; struct sfc_evq *evq; efsys_mem_t *esmp; unsigned int total_delay_us; unsigned int delay_us; int rc; sfc_log_init(sa, "sw_index=%u", sw_index); evq_info = &sa->evq_info[sw_index]; evq = evq_info->evq; esmp = &evq->mem; /* Clear all events */ (void)memset((void *)esmp->esm_base, 0xff, EFX_EVQ_SIZE(evq_info->entries)); /* Create the common code event queue */ rc = efx_ev_qcreate(sa->nic, sw_index, esmp, evq_info->entries, 0 /* unused on EF10 */, 0, evq_info->flags, &evq->common); if (rc != 0) goto fail_ev_qcreate; evq->init_state = SFC_EVQ_STARTING; /* Wait for the initialization event */ total_delay_us = 0; delay_us = SFC_EVQ_INIT_BACKOFF_START_US; do { (void)sfc_ev_qpoll(evq); /* Check to see if the initialization complete indication * posted by the hardware. */ if (evq->init_state == SFC_EVQ_STARTED) goto done; /* Give event queue some time to init */ rte_delay_us(delay_us); total_delay_us += delay_us; /* Exponential backoff */ delay_us *= 2; if (delay_us > SFC_EVQ_INIT_BACKOFF_MAX_US) delay_us = SFC_EVQ_INIT_BACKOFF_MAX_US; } while (total_delay_us < SFC_EVQ_INIT_TIMEOUT_US); rc = ETIMEDOUT; goto fail_timedout; done: return 0; fail_timedout: evq->init_state = SFC_EVQ_INITIALIZED; efx_ev_qdestroy(evq->common); fail_ev_qcreate: sfc_log_init(sa, "failed %d", rc); return rc; } void sfc_ev_qstop(struct sfc_adapter *sa, unsigned int sw_index) { const struct sfc_evq_info *evq_info; struct sfc_evq *evq; sfc_log_init(sa, "sw_index=%u", sw_index); SFC_ASSERT(sw_index < sa->evq_count); evq_info = &sa->evq_info[sw_index]; evq = evq_info->evq; if (evq == NULL || evq->init_state != SFC_EVQ_STARTED) return; evq->init_state = SFC_EVQ_INITIALIZED; evq->read_ptr = 0; evq->exception = B_FALSE; efx_ev_qdestroy(evq->common); } static void sfc_ev_mgmt_periodic_qpoll(void *arg) { struct sfc_adapter *sa = arg; int rc; sfc_ev_mgmt_qpoll(sa); rc = rte_eal_alarm_set(SFC_MGMT_EV_QPOLL_PERIOD_US, sfc_ev_mgmt_periodic_qpoll, sa); if (rc == -ENOTSUP) { sfc_warn(sa, "alarms are not supported"); sfc_warn(sa, "management EVQ must be polled indirectly using no-wait link status update"); } else if (rc != 0) { sfc_err(sa, "cannot rearm management EVQ polling alarm (rc=%d)", rc); } } static void sfc_ev_mgmt_periodic_qpoll_start(struct sfc_adapter *sa) { sfc_ev_mgmt_periodic_qpoll(sa); } static void sfc_ev_mgmt_periodic_qpoll_stop(struct sfc_adapter *sa) { rte_eal_alarm_cancel(sfc_ev_mgmt_periodic_qpoll, sa); } int sfc_ev_start(struct sfc_adapter *sa) { int rc; sfc_log_init(sa, "entry"); rc = efx_ev_init(sa->nic); if (rc != 0) goto fail_ev_init; /* Start management EVQ used for global events */ rte_spinlock_lock(&sa->mgmt_evq_lock); rc = sfc_ev_qstart(sa, sa->mgmt_evq_index); if (rc != 0) goto fail_mgmt_evq_start; if (sa->intr.lsc_intr) { rc = sfc_ev_qprime(sa->evq_info[sa->mgmt_evq_index].evq); if (rc != 0) goto fail_evq0_prime; } rte_spinlock_unlock(&sa->mgmt_evq_lock); /* * Start management EVQ polling. If interrupts are disabled * (not used), it is required to process link status change * and other device level events to avoid unrecoverable * error because the event queue overflow. */ sfc_ev_mgmt_periodic_qpoll_start(sa); /* * Rx/Tx event queues are started/stopped when corresponding * Rx/Tx queue is started/stopped. */ return 0; fail_evq0_prime: sfc_ev_qstop(sa, 0); fail_mgmt_evq_start: rte_spinlock_unlock(&sa->mgmt_evq_lock); efx_ev_fini(sa->nic); fail_ev_init: sfc_log_init(sa, "failed %d", rc); return rc; } void sfc_ev_stop(struct sfc_adapter *sa) { unsigned int sw_index; sfc_log_init(sa, "entry"); sfc_ev_mgmt_periodic_qpoll_stop(sa); /* Make sure that all event queues are stopped */ sw_index = sa->evq_count; while (sw_index-- > 0) { if (sw_index == sa->mgmt_evq_index) { /* Locks are required for the management EVQ */ rte_spinlock_lock(&sa->mgmt_evq_lock); sfc_ev_qstop(sa, sa->mgmt_evq_index); rte_spinlock_unlock(&sa->mgmt_evq_lock); } else { sfc_ev_qstop(sa, sw_index); } } efx_ev_fini(sa->nic); } int sfc_ev_qinit(struct sfc_adapter *sa, unsigned int sw_index, unsigned int entries, int socket_id) { struct sfc_evq_info *evq_info; struct sfc_evq *evq; int rc; sfc_log_init(sa, "sw_index=%u", sw_index); evq_info = &sa->evq_info[sw_index]; SFC_ASSERT(rte_is_power_of_2(entries)); SFC_ASSERT(entries <= evq_info->max_entries); evq_info->entries = entries; evq = rte_zmalloc_socket("sfc-evq", sizeof(*evq), RTE_CACHE_LINE_SIZE, socket_id); if (evq == NULL) return ENOMEM; evq->sa = sa; evq->evq_index = sw_index; /* Allocate DMA space */ rc = sfc_dma_alloc(sa, "evq", sw_index, EFX_EVQ_SIZE(evq_info->entries), socket_id, &evq->mem); if (rc != 0) return rc; evq->init_state = SFC_EVQ_INITIALIZED; evq_info->evq = evq; return 0; } void sfc_ev_qfini(struct sfc_adapter *sa, unsigned int sw_index) { struct sfc_evq *evq; sfc_log_init(sa, "sw_index=%u", sw_index); evq = sa->evq_info[sw_index].evq; SFC_ASSERT(evq->init_state == SFC_EVQ_INITIALIZED); sa->evq_info[sw_index].evq = NULL; sfc_dma_free(sa, &evq->mem); rte_free(evq); } static int sfc_ev_qinit_info(struct sfc_adapter *sa, unsigned int sw_index) { struct sfc_evq_info *evq_info = &sa->evq_info[sw_index]; unsigned int max_entries; sfc_log_init(sa, "sw_index=%u", sw_index); max_entries = sfc_evq_max_entries(sa, sw_index); SFC_ASSERT(rte_is_power_of_2(max_entries)); evq_info->max_entries = max_entries; evq_info->flags = sa->evq_flags | ((sa->intr.lsc_intr && sw_index == sa->mgmt_evq_index) ? EFX_EVQ_FLAGS_NOTIFY_INTERRUPT : EFX_EVQ_FLAGS_NOTIFY_DISABLED); return 0; } static int sfc_kvarg_perf_profile_handler(__rte_unused const char *key, const char *value_str, void *opaque) { uint64_t *value = opaque; if (strcasecmp(value_str, SFC_KVARG_PERF_PROFILE_THROUGHPUT) == 0) *value = EFX_EVQ_FLAGS_TYPE_THROUGHPUT; else if (strcasecmp(value_str, SFC_KVARG_PERF_PROFILE_LOW_LATENCY) == 0) *value = EFX_EVQ_FLAGS_TYPE_LOW_LATENCY; else if (strcasecmp(value_str, SFC_KVARG_PERF_PROFILE_AUTO) == 0) *value = EFX_EVQ_FLAGS_TYPE_AUTO; else return -EINVAL; return 0; } static void sfc_ev_qfini_info(struct sfc_adapter *sa, unsigned int sw_index) { sfc_log_init(sa, "sw_index=%u", sw_index); /* Nothing to cleanup */ } int sfc_ev_init(struct sfc_adapter *sa) { int rc; unsigned int sw_index; sfc_log_init(sa, "entry"); sa->evq_flags = EFX_EVQ_FLAGS_TYPE_THROUGHPUT; rc = sfc_kvargs_process(sa, SFC_KVARG_PERF_PROFILE, sfc_kvarg_perf_profile_handler, &sa->evq_flags); if (rc != 0) { sfc_err(sa, "invalid %s parameter value", SFC_KVARG_PERF_PROFILE); goto fail_kvarg_perf_profile; } sa->evq_count = sfc_ev_qcount(sa); sa->mgmt_evq_index = 0; rte_spinlock_init(&sa->mgmt_evq_lock); /* Allocate EVQ info array */ rc = ENOMEM; sa->evq_info = rte_calloc_socket("sfc-evqs", sa->evq_count, sizeof(struct sfc_evq_info), 0, sa->socket_id); if (sa->evq_info == NULL) goto fail_evqs_alloc; for (sw_index = 0; sw_index < sa->evq_count; ++sw_index) { rc = sfc_ev_qinit_info(sa, sw_index); if (rc != 0) goto fail_ev_qinit_info; } rc = sfc_ev_qinit(sa, sa->mgmt_evq_index, SFC_MGMT_EVQ_ENTRIES, sa->socket_id); if (rc != 0) goto fail_mgmt_evq_init; /* * Rx/Tx event queues are created/destroyed when corresponding * Rx/Tx queue is created/destroyed. */ return 0; fail_mgmt_evq_init: fail_ev_qinit_info: while (sw_index-- > 0) sfc_ev_qfini_info(sa, sw_index); rte_free(sa->evq_info); sa->evq_info = NULL; fail_evqs_alloc: sa->evq_count = 0; fail_kvarg_perf_profile: sfc_log_init(sa, "failed %d", rc); return rc; } void sfc_ev_fini(struct sfc_adapter *sa) { int sw_index; sfc_log_init(sa, "entry"); /* Cleanup all event queues */ sw_index = sa->evq_count; while (--sw_index >= 0) { if (sa->evq_info[sw_index].evq != NULL) sfc_ev_qfini(sa, sw_index); sfc_ev_qfini_info(sa, sw_index); } rte_free(sa->evq_info); sa->evq_info = NULL; sa->evq_count = 0; }