X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fmlx4%2Fmlx4.c;h=90c3c07701288ce817ef85c5ff3b2d3ae58df863;hb=0880c40113ef2d69b6433d7dfa0b4032cc378b0d;hp=2f49ed57427c34bbaee0acfd8877e49c68159f7e;hpb=a421b86a4a0246b5e128365435a6a00645758d14;p=dpdk.git diff --git a/drivers/net/mlx4/mlx4.c b/drivers/net/mlx4/mlx4.c index 2f49ed5742..90c3c07701 100644 --- a/drivers/net/mlx4/mlx4.c +++ b/drivers/net/mlx4/mlx4.c @@ -54,25 +54,24 @@ #include #include #include -#include #include #include +#include /* Verbs header. */ /* ISO C doesn't support unnamed structs/unions, disabling -pedantic. */ #ifdef PEDANTIC -#pragma GCC diagnostic ignored "-pedantic" +#pragma GCC diagnostic ignored "-Wpedantic" #endif #include #ifdef PEDANTIC -#pragma GCC diagnostic error "-pedantic" +#pragma GCC diagnostic error "-Wpedantic" #endif /* DPDK headers don't like -pedantic. */ #ifdef PEDANTIC -#pragma GCC diagnostic ignored "-pedantic" +#pragma GCC diagnostic ignored "-Wpedantic" #endif -#include #include #include #include @@ -85,8 +84,10 @@ #include #include #include +#include +#include #ifdef PEDANTIC -#pragma GCC diagnostic error "-pedantic" +#pragma GCC diagnostic error "-Wpedantic" #endif /* Generated configuration header. */ @@ -130,15 +131,6 @@ typedef union { #define WR_ID(o) (((wr_id_t *)&(o))->data) -/* Compile-time check. */ -static inline void wr_id_t_check(void) -{ - wr_id_t check[1 + (2 * -!(sizeof(wr_id_t) == sizeof(uint64_t)))]; - - (void)check; - (void)wr_id_t_check; -} - /* Transpose flags. Useful to convert IBV to DPDK flags. */ #define TRANSPOSE(val, from, to) \ (((from) >= (to)) ? \ @@ -204,7 +196,6 @@ struct rxq { unsigned int sp:1; /* Use scattered RX elements. */ unsigned int csum:1; /* Enable checksum offloading. */ unsigned int csum_l2tun:1; /* Same for L2 tunnels. */ - uint32_t mb_len; /* Length of a mp-issued mbuf. */ struct mlx4_rxq_stats stats; /* RX queue counters. */ unsigned int socket; /* CPU socket ID for allocations. */ struct ibv_exp_res_domain *rd; /* Resource Domain. */ @@ -227,7 +218,7 @@ typedef uint8_t linear_t[16384]; struct txq { struct priv *priv; /* Back pointer to private data. */ struct { - struct rte_mempool *mp; /* Cached Memory Pool. */ + const struct rte_mempool *mp; /* Cached Memory Pool. */ struct ibv_mr *mr; /* Memory Region (for mp). */ uint32_t lkey; /* mr->lkey */ } mp2mr[MLX4_PMD_TX_MP_CACHE]; /* MP to MR translation table. */ @@ -282,6 +273,7 @@ struct priv { unsigned int hw_csum_l2tun:1; /* Same for L2 tunnels. */ unsigned int rss:1; /* RSS is enabled. */ unsigned int vf:1; /* This is a VF device. */ + unsigned int pending_alarm:1; /* An alarm is pending. */ #ifdef INLINE_RECV unsigned int inl_recv_size; /* Inline recv size */ #endif @@ -292,9 +284,50 @@ struct priv { unsigned int txqs_n; /* TX queues array size. */ struct rxq *(*rxqs)[]; /* RX queues. */ struct txq *(*txqs)[]; /* TX queues. */ + struct rte_intr_handle intr_handle; /* Interrupt handler. */ rte_spinlock_t lock; /* Lock for control functions. */ }; +/* Local storage for secondary process data. */ +struct mlx4_secondary_data { + struct rte_eth_dev_data data; /* Local device data. */ + struct priv *primary_priv; /* Private structure from primary. */ + struct rte_eth_dev_data *shared_dev_data; /* Shared device data. */ + rte_spinlock_t lock; /* Port configuration lock. */ +} mlx4_secondary_data[RTE_MAX_ETHPORTS]; + +/** + * Check if running as a secondary process. + * + * @return + * Nonzero if running as a secondary process. + */ +static inline int +mlx4_is_secondary(void) +{ + return rte_eal_process_type() != RTE_PROC_PRIMARY; +} + +/** + * Return private structure associated with an Ethernet device. + * + * @param dev + * Pointer to Ethernet device structure. + * + * @return + * Pointer to private structure. + */ +static struct priv * +mlx4_get_priv(struct rte_eth_dev *dev) +{ + struct mlx4_secondary_data *sd; + + if (!mlx4_is_secondary()) + return dev->data->dev_private; + sd = &mlx4_secondary_data[dev->data->port_id]; + return sd->data.dev_private; +} + /** * Lock private structure to protect it from concurrent access in the * control path. @@ -624,7 +657,15 @@ priv_get_mtu(struct priv *priv, uint16_t *mtu) static int priv_set_mtu(struct priv *priv, uint16_t mtu) { - return priv_set_sysfs_ulong(priv, "mtu", mtu); + uint16_t new_mtu; + + if (priv_set_sysfs_ulong(priv, "mtu", mtu) || + priv_get_mtu(priv, &new_mtu)) + return -1; + if (new_mtu == mtu) + return 0; + errno = EINVAL; + return -1; } /** @@ -648,15 +689,22 @@ priv_set_flags(struct priv *priv, unsigned int keep, unsigned int flags) if (priv_get_sysfs_ulong(priv, "flags", &tmp) == -1) return -1; tmp &= keep; - tmp |= flags; + tmp |= (flags & (~keep)); return priv_set_sysfs_ulong(priv, "flags", tmp); } /* Device configuration. */ +static int +txq_setup(struct rte_eth_dev *dev, struct txq *txq, uint16_t desc, + unsigned int socket, const struct rte_eth_txconf *conf); + +static void +txq_cleanup(struct txq *txq); + static int rxq_setup(struct rte_eth_dev *dev, struct rxq *rxq, uint16_t desc, - unsigned int socket, const struct rte_eth_rxconf *conf, + unsigned int socket, int inactive, const struct rte_eth_rxconf *conf, struct rte_mempool *mp); static void @@ -692,6 +740,15 @@ dev_configure(struct rte_eth_dev *dev) } if (rxqs_n == priv->rxqs_n) return 0; + if (!rte_is_power_of_2(rxqs_n)) { + unsigned n_active; + + n_active = rte_align32pow2(rxqs_n + 1) >> 1; + WARN("%p: number of RX queues must be a power" + " of 2: %u queues among %u will be active", + (void *)dev, n_active, rxqs_n); + } + INFO("%p: RX queues number update: %u -> %u", (void *)dev, priv->rxqs_n, rxqs_n); /* If RSS is enabled, disable it first. */ @@ -727,7 +784,7 @@ dev_configure(struct rte_eth_dev *dev) priv->rss = 1; tmp = priv->rxqs_n; priv->rxqs_n = rxqs_n; - ret = rxq_setup(dev, &priv->rxq_parent, 0, 0, NULL, NULL); + ret = rxq_setup(dev, &priv->rxq_parent, 0, 0, 0, NULL, NULL); if (!ret) return 0; /* Failure, rollback. */ @@ -752,6 +809,8 @@ mlx4_dev_configure(struct rte_eth_dev *dev) struct priv *priv = dev->data->dev_private; int ret; + if (mlx4_is_secondary()) + return -E_RTE_SECONDARY; priv_lock(priv); ret = dev_configure(dev); assert(ret >= 0); @@ -759,6 +818,160 @@ mlx4_dev_configure(struct rte_eth_dev *dev) return -ret; } +static uint16_t mlx4_tx_burst(void *, struct rte_mbuf **, uint16_t); +static uint16_t removed_rx_burst(void *, struct rte_mbuf **, uint16_t); + +/** + * Configure secondary process queues from a private data pointer (primary + * or secondary) and update burst callbacks. Can take place only once. + * + * All queues must have been previously created by the primary process to + * avoid undefined behavior. + * + * @param priv + * Private data pointer from either primary or secondary process. + * + * @return + * Private data pointer from secondary process, NULL in case of error. + */ +static struct priv * +mlx4_secondary_data_setup(struct priv *priv) +{ + unsigned int port_id = 0; + struct mlx4_secondary_data *sd; + void **tx_queues; + void **rx_queues; + unsigned int nb_tx_queues; + unsigned int nb_rx_queues; + unsigned int i; + + /* priv must be valid at this point. */ + assert(priv != NULL); + /* priv->dev must also be valid but may point to local memory from + * another process, possibly with the same address and must not + * be dereferenced yet. */ + assert(priv->dev != NULL); + /* Determine port ID by finding out where priv comes from. */ + while (1) { + sd = &mlx4_secondary_data[port_id]; + rte_spinlock_lock(&sd->lock); + /* Primary process? */ + if (sd->primary_priv == priv) + break; + /* Secondary process? */ + if (sd->data.dev_private == priv) + break; + rte_spinlock_unlock(&sd->lock); + if (++port_id == RTE_DIM(mlx4_secondary_data)) + port_id = 0; + } + /* Switch to secondary private structure. If private data has already + * been updated by another thread, there is nothing else to do. */ + priv = sd->data.dev_private; + if (priv->dev->data == &sd->data) + goto end; + /* Sanity checks. Secondary private structure is supposed to point + * to local eth_dev, itself still pointing to the shared device data + * structure allocated by the primary process. */ + assert(sd->shared_dev_data != &sd->data); + assert(sd->data.nb_tx_queues == 0); + assert(sd->data.tx_queues == NULL); + assert(sd->data.nb_rx_queues == 0); + assert(sd->data.rx_queues == NULL); + assert(priv != sd->primary_priv); + assert(priv->dev->data == sd->shared_dev_data); + assert(priv->txqs_n == 0); + assert(priv->txqs == NULL); + assert(priv->rxqs_n == 0); + assert(priv->rxqs == NULL); + nb_tx_queues = sd->shared_dev_data->nb_tx_queues; + nb_rx_queues = sd->shared_dev_data->nb_rx_queues; + /* Allocate local storage for queues. */ + tx_queues = rte_zmalloc("secondary ethdev->tx_queues", + sizeof(sd->data.tx_queues[0]) * nb_tx_queues, + RTE_CACHE_LINE_SIZE); + rx_queues = rte_zmalloc("secondary ethdev->rx_queues", + sizeof(sd->data.rx_queues[0]) * nb_rx_queues, + RTE_CACHE_LINE_SIZE); + if (tx_queues == NULL || rx_queues == NULL) + goto error; + /* Lock to prevent control operations during setup. */ + priv_lock(priv); + /* TX queues. */ + for (i = 0; i != nb_tx_queues; ++i) { + struct txq *primary_txq = (*sd->primary_priv->txqs)[i]; + struct txq *txq; + + if (primary_txq == NULL) + continue; + txq = rte_calloc_socket("TXQ", 1, sizeof(*txq), 0, + primary_txq->socket); + if (txq != NULL) { + if (txq_setup(priv->dev, + txq, + primary_txq->elts_n * MLX4_PMD_SGE_WR_N, + primary_txq->socket, + NULL) == 0) { + txq->stats.idx = primary_txq->stats.idx; + tx_queues[i] = txq; + continue; + } + rte_free(txq); + } + while (i) { + txq = tx_queues[--i]; + txq_cleanup(txq); + rte_free(txq); + } + goto error; + } + /* RX queues. */ + for (i = 0; i != nb_rx_queues; ++i) { + struct rxq *primary_rxq = (*sd->primary_priv->rxqs)[i]; + + if (primary_rxq == NULL) + continue; + /* Not supported yet. */ + rx_queues[i] = NULL; + } + /* Update everything. */ + priv->txqs = (void *)tx_queues; + priv->txqs_n = nb_tx_queues; + priv->rxqs = (void *)rx_queues; + priv->rxqs_n = nb_rx_queues; + sd->data.rx_queues = rx_queues; + sd->data.tx_queues = tx_queues; + sd->data.nb_rx_queues = nb_rx_queues; + sd->data.nb_tx_queues = nb_tx_queues; + sd->data.dev_link = sd->shared_dev_data->dev_link; + sd->data.mtu = sd->shared_dev_data->mtu; + memcpy(sd->data.rx_queue_state, sd->shared_dev_data->rx_queue_state, + sizeof(sd->data.rx_queue_state)); + memcpy(sd->data.tx_queue_state, sd->shared_dev_data->tx_queue_state, + sizeof(sd->data.tx_queue_state)); + sd->data.dev_flags = sd->shared_dev_data->dev_flags; + /* Use local data from now on. */ + rte_mb(); + priv->dev->data = &sd->data; + rte_mb(); + priv->dev->tx_pkt_burst = mlx4_tx_burst; + priv->dev->rx_pkt_burst = removed_rx_burst; + priv_unlock(priv); +end: + /* More sanity checks. */ + assert(priv->dev->tx_pkt_burst == mlx4_tx_burst); + assert(priv->dev->rx_pkt_burst == removed_rx_burst); + assert(priv->dev->data == &sd->data); + rte_spinlock_unlock(&sd->lock); + return priv; +error: + priv_unlock(priv); + rte_free(tx_queues); + rte_free(rx_queues); + rte_spinlock_unlock(&sd->lock); + return NULL; +} + /* TX queues handling. */ /** @@ -791,7 +1004,7 @@ txq_alloc_elts(struct txq *txq, unsigned int elts_n) } mr_linear = ibv_reg_mr(txq->priv->pd, elts_linear, sizeof(*elts_linear), - (IBV_ACCESS_LOCAL_WRITE | IBV_ACCESS_REMOTE_WRITE)); + IBV_ACCESS_LOCAL_WRITE); if (mr_linear == NULL) { ERROR("%p: unable to configure MR, ibv_reg_mr() failed", (void *)txq); @@ -840,14 +1053,20 @@ error: static void txq_free_elts(struct txq *txq) { - unsigned int i; unsigned int elts_n = txq->elts_n; + unsigned int elts_head = txq->elts_head; + unsigned int elts_tail = txq->elts_tail; struct txq_elt (*elts)[elts_n] = txq->elts; linear_t (*elts_linear)[elts_n] = txq->elts_linear; struct ibv_mr *mr_linear = txq->mr_linear; DEBUG("%p: freeing WRs", (void *)txq); txq->elts_n = 0; + txq->elts_head = 0; + txq->elts_tail = 0; + txq->elts_comp = 0; + txq->elts_comp_cd = 0; + txq->elts_comp_cd_init = 0; txq->elts = NULL; txq->elts_linear = NULL; txq->mr_linear = NULL; @@ -857,12 +1076,17 @@ txq_free_elts(struct txq *txq) rte_free(elts_linear); if (elts == NULL) return; - for (i = 0; (i != elemof(*elts)); ++i) { - struct txq_elt *elt = &(*elts)[i]; + while (elts_tail != elts_head) { + struct txq_elt *elt = &(*elts)[elts_tail]; - if (elt->buf == NULL) - continue; + assert(elt->buf != NULL); rte_pktmbuf_free(elt->buf); +#ifndef NDEBUG + /* Poisoning. */ + memset(elt, 0x77, sizeof(*elt)); +#endif + if (++elts_tail == elts_n) + elts_tail = 0; } rte_free(elts); } @@ -980,6 +1204,139 @@ txq_complete(struct txq *txq) return 0; } +struct mlx4_check_mempool_data { + int ret; + char *start; + char *end; +}; + +/* Called by mlx4_check_mempool() when iterating the memory chunks. */ +static void mlx4_check_mempool_cb(struct rte_mempool *mp, + void *opaque, struct rte_mempool_memhdr *memhdr, + unsigned mem_idx) +{ + struct mlx4_check_mempool_data *data = opaque; + + (void)mp; + (void)mem_idx; + + /* It already failed, skip the next chunks. */ + if (data->ret != 0) + return; + /* It is the first chunk. */ + if (data->start == NULL && data->end == NULL) { + data->start = memhdr->addr; + data->end = data->start + memhdr->len; + return; + } + if (data->end == memhdr->addr) { + data->end += memhdr->len; + return; + } + if (data->start == (char *)memhdr->addr + memhdr->len) { + data->start -= memhdr->len; + return; + } + /* Error, mempool is not virtually contigous. */ + data->ret = -1; +} + +/** + * Check if a mempool can be used: it must be virtually contiguous. + * + * @param[in] mp + * Pointer to memory pool. + * @param[out] start + * Pointer to the start address of the mempool virtual memory area + * @param[out] end + * Pointer to the end address of the mempool virtual memory area + * + * @return + * 0 on success (mempool is virtually contiguous), -1 on error. + */ +static int mlx4_check_mempool(struct rte_mempool *mp, uintptr_t *start, + uintptr_t *end) +{ + struct mlx4_check_mempool_data data; + + memset(&data, 0, sizeof(data)); + rte_mempool_mem_iter(mp, mlx4_check_mempool_cb, &data); + *start = (uintptr_t)data.start; + *end = (uintptr_t)data.end; + + return data.ret; +} + +/* For best performance, this function should not be inlined. */ +static struct ibv_mr *mlx4_mp2mr(struct ibv_pd *, struct rte_mempool *) + __attribute__((noinline)); + +/** + * Register mempool as a memory region. + * + * @param pd + * Pointer to protection domain. + * @param mp + * Pointer to memory pool. + * + * @return + * Memory region pointer, NULL in case of error. + */ +static struct ibv_mr * +mlx4_mp2mr(struct ibv_pd *pd, struct rte_mempool *mp) +{ + const struct rte_memseg *ms = rte_eal_get_physmem_layout(); + uintptr_t start; + uintptr_t end; + unsigned int i; + + if (mlx4_check_mempool(mp, &start, &end) != 0) { + ERROR("mempool %p: not virtually contiguous", + (void *)mp); + return NULL; + } + + DEBUG("mempool %p area start=%p end=%p size=%zu", + (void *)mp, (void *)start, (void *)end, + (size_t)(end - start)); + /* Round start and end to page boundary if found in memory segments. */ + for (i = 0; (i < RTE_MAX_MEMSEG) && (ms[i].addr != NULL); ++i) { + uintptr_t addr = (uintptr_t)ms[i].addr; + size_t len = ms[i].len; + unsigned int align = ms[i].hugepage_sz; + + if ((start > addr) && (start < addr + len)) + start = RTE_ALIGN_FLOOR(start, align); + if ((end > addr) && (end < addr + len)) + end = RTE_ALIGN_CEIL(end, align); + } + DEBUG("mempool %p using start=%p end=%p size=%zu for MR", + (void *)mp, (void *)start, (void *)end, + (size_t)(end - start)); + return ibv_reg_mr(pd, + (void *)start, + end - start, + IBV_ACCESS_LOCAL_WRITE); +} + +/** + * Get Memory Pool (MP) from mbuf. If mbuf is indirect, the pool from which + * the cloned mbuf is allocated is returned instead. + * + * @param buf + * Pointer to mbuf. + * + * @return + * Memory pool where data is located for given mbuf. + */ +static struct rte_mempool * +txq_mb2mp(struct rte_mbuf *buf) +{ + if (unlikely(RTE_MBUF_INDIRECT(buf))) + return rte_mbuf_from_indirect(buf)->pool; + return buf->pool; +} + /** * Get Memory Region (MR) <-> Memory Pool (MP) association from txq->mp2mr[]. * Add MP to txq->mp2mr[] if it's not registered yet. If mp2mr[] is full, @@ -1011,11 +1368,9 @@ txq_mp2mr(struct txq *txq, struct rte_mempool *mp) } } /* Add a new entry, register MR first. */ - DEBUG("%p: discovered new memory pool %p", (void *)txq, (void *)mp); - mr = ibv_reg_mr(txq->priv->pd, - (void *)mp->elt_va_start, - (mp->elt_va_end - mp->elt_va_start), - (IBV_ACCESS_LOCAL_WRITE | IBV_ACCESS_REMOTE_WRITE)); + DEBUG("%p: discovered new memory pool \"%s\" (%p)", + (void *)txq, mp->name, (void *)mp); + mr = mlx4_mp2mr(txq->priv->pd, mp); if (unlikely(mr == NULL)) { DEBUG("%p: unable to configure MR, ibv_reg_mr() failed.", (void *)txq); @@ -1026,7 +1381,7 @@ txq_mp2mr(struct txq *txq, struct rte_mempool *mp) DEBUG("%p: MR <-> MP table full, dropping oldest entry.", (void *)txq); --i; - claim_zero(ibv_dereg_mr(txq->mp2mr[i].mr)); + claim_zero(ibv_dereg_mr(txq->mp2mr[0].mr)); memmove(&txq->mp2mr[0], &txq->mp2mr[1], (sizeof(txq->mp2mr) - sizeof(txq->mp2mr[0]))); } @@ -1034,11 +1389,66 @@ txq_mp2mr(struct txq *txq, struct rte_mempool *mp) txq->mp2mr[i].mp = mp; txq->mp2mr[i].mr = mr; txq->mp2mr[i].lkey = mr->lkey; - DEBUG("%p: new MR lkey for MP %p: 0x%08" PRIu32, - (void *)txq, (void *)mp, txq->mp2mr[i].lkey); + DEBUG("%p: new MR lkey for MP \"%s\" (%p): 0x%08" PRIu32, + (void *)txq, mp->name, (void *)mp, txq->mp2mr[i].lkey); return txq->mp2mr[i].lkey; } +struct txq_mp2mr_mbuf_check_data { + int ret; +}; + +/** + * Callback function for rte_mempool_obj_iter() to check whether a given + * mempool object looks like a mbuf. + * + * @param[in] mp + * The mempool pointer + * @param[in] arg + * Context data (struct txq_mp2mr_mbuf_check_data). Contains the + * return value. + * @param[in] obj + * Object address. + * @param index + * Object index, unused. + */ +static void +txq_mp2mr_mbuf_check(struct rte_mempool *mp, void *arg, void *obj, + uint32_t index __rte_unused) +{ + struct txq_mp2mr_mbuf_check_data *data = arg; + struct rte_mbuf *buf = obj; + + /* Check whether mbuf structure fits element size and whether mempool + * pointer is valid. */ + if (sizeof(*buf) > mp->elt_size || buf->pool != mp) + data->ret = -1; +} + +/** + * Iterator function for rte_mempool_walk() to register existing mempools and + * fill the MP to MR cache of a TX queue. + * + * @param[in] mp + * Memory Pool to register. + * @param *arg + * Pointer to TX queue structure. + */ +static void +txq_mp2mr_iter(struct rte_mempool *mp, void *arg) +{ + struct txq *txq = arg; + struct txq_mp2mr_mbuf_check_data data = { + .ret = 0, + }; + + /* Register mempool only if the first element looks like a mbuf. */ + if (rte_mempool_obj_iter(mp, txq_mp2mr_mbuf_check, &data) == 0 || + data.ret == -1) + return; + txq_mp2mr(txq, mp); +} + #if MLX4_PMD_SGE_WR_N > 1 /** @@ -1120,7 +1530,7 @@ tx_burst_sg(struct txq *txq, unsigned int segs, struct txq_elt *elt, uint32_t lkey; /* Retrieve Memory Region key for this memory pool. */ - lkey = txq_mp2mr(txq, buf->pool); + lkey = txq_mp2mr(txq, txq_mb2mp(buf)); if (unlikely(lkey == (uint32_t)-1)) { /* MR does not exist. */ DEBUG("%p: unable to get MP <-> MR association", @@ -1173,6 +1583,8 @@ tx_burst_sg(struct txq *txq, unsigned int segs, struct txq_elt *elt, sge->length = size; sge->lkey = txq->mr_linear->lkey; sent_size += size; + /* Include last segment. */ + segs++; } return (struct tx_burst_sg_ret){ .length = sent_size, @@ -1205,7 +1617,6 @@ mlx4_tx_burst(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n) { struct txq *txq = (struct txq *)dpdk_txq; unsigned int elts_head = txq->elts_head; - const unsigned int elts_tail = txq->elts_tail; const unsigned int elts_n = txq->elts_n; unsigned int elts_comp_cd = txq->elts_comp_cd; unsigned int elts_comp = 0; @@ -1215,7 +1626,7 @@ mlx4_tx_burst(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n) assert(elts_comp_cd != 0); txq_complete(txq); - max = (elts_n - (elts_head - elts_tail)); + max = (elts_n - (elts_head - txq->elts_tail)); if (max > elts_n) max -= elts_n; assert(max >= 1); @@ -1242,6 +1653,10 @@ mlx4_tx_burst(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n) if (likely(elt->buf != NULL)) { struct rte_mbuf *tmp = elt->buf; +#ifndef NDEBUG + /* Poisoning. */ + memset(elt, 0x66, sizeof(*elt)); +#endif /* Faster than rte_pktmbuf_free(). */ do { struct rte_mbuf *next = NEXT(tmp); @@ -1276,7 +1691,7 @@ mlx4_tx_burst(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n) addr = rte_pktmbuf_mtod(buf, uintptr_t); length = DATA_LEN(buf); /* Retrieve Memory Region key for this memory pool. */ - lkey = txq_mp2mr(txq, buf->pool); + lkey = txq_mp2mr(txq, txq_mb2mp(buf)); if (unlikely(lkey == (uint32_t)-1)) { /* MR does not exist. */ DEBUG("%p: unable to get MP <-> MR" @@ -1367,6 +1782,46 @@ stop: return i; } +/** + * DPDK callback for TX in secondary processes. + * + * This function configures all queues from primary process information + * if necessary before reverting to the normal TX burst callback. + * + * @param dpdk_txq + * Generic pointer to TX queue structure. + * @param[in] pkts + * Packets to transmit. + * @param pkts_n + * Number of packets in array. + * + * @return + * Number of packets successfully transmitted (<= pkts_n). + */ +static uint16_t +mlx4_tx_burst_secondary_setup(void *dpdk_txq, struct rte_mbuf **pkts, + uint16_t pkts_n) +{ + struct txq *txq = dpdk_txq; + struct priv *priv = mlx4_secondary_data_setup(txq->priv); + struct priv *primary_priv; + unsigned int index; + + if (priv == NULL) + return 0; + primary_priv = + mlx4_secondary_data[priv->dev->data->port_id].primary_priv; + /* Look for queue index in both private structures. */ + for (index = 0; index != priv->txqs_n; ++index) + if (((*primary_priv->txqs)[index] == txq) || + ((*priv->txqs)[index] == txq)) + break; + if (index == priv->txqs_n) + return 0; + txq = (*priv->txqs)[index]; + return priv->dev->tx_pkt_burst(txq, pkts, pkts_n); +} + /** * Configure a TX queue. * @@ -1388,7 +1843,7 @@ static int txq_setup(struct rte_eth_dev *dev, struct txq *txq, uint16_t desc, unsigned int socket, const struct rte_eth_txconf *conf) { - struct priv *priv = dev->data->dev_private; + struct priv *priv = mlx4_get_priv(dev); struct txq tmpl = { .priv = priv, .socket = socket @@ -1404,6 +1859,8 @@ txq_setup(struct rte_eth_dev *dev, struct txq *txq, uint16_t desc, int ret = 0; (void)conf; /* Thresholds configuration (ignored). */ + if (priv == NULL) + return EINVAL; if ((desc == 0) || (desc % MLX4_PMD_SGE_WR_N)) { ERROR("%p: invalid number of TX descriptors (must be a" " multiple of %d)", (void *)dev, MLX4_PMD_SGE_WR_N); @@ -1547,6 +2004,8 @@ txq_setup(struct rte_eth_dev *dev, struct txq *txq, uint16_t desc, txq_cleanup(txq); *txq = tmpl; DEBUG("%p: txq updated with %p", (void *)txq, (void *)&tmpl); + /* Pre-register known mempools. */ + rte_mempool_walk(txq_mp2mr_iter, txq); assert(ret == 0); return 0; error: @@ -1580,6 +2039,8 @@ mlx4_tx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc, struct txq *txq = (*priv->txqs)[idx]; int ret; + if (mlx4_is_secondary()) + return -E_RTE_SECONDARY; priv_lock(priv); DEBUG("%p: configuring queue %u for %u descriptors", (void *)dev, idx, desc); @@ -1635,6 +2096,8 @@ mlx4_tx_queue_release(void *dpdk_txq) struct priv *priv; unsigned int i; + if (mlx4_is_secondary()) + return; if (txq == NULL) return; priv = txq->priv; @@ -2485,6 +2948,8 @@ rxq_cleanup(struct rxq *rxq) * @param flags * RX completion flags returned by poll_length_flags(). * + * @note: fix mlx4_dev_supported_ptypes_get() if any change here. + * * @return * Packet type for struct rte_mbuf. */ @@ -2530,25 +2995,20 @@ rxq_cq_to_ol_flags(const struct rxq *rxq, uint32_t flags) if (rxq->csum) ol_flags |= - TRANSPOSE(~flags, + TRANSPOSE(flags, IBV_EXP_CQ_RX_IP_CSUM_OK, - PKT_RX_IP_CKSUM_BAD) | - TRANSPOSE(~flags, + PKT_RX_IP_CKSUM_GOOD) | + TRANSPOSE(flags, IBV_EXP_CQ_RX_TCP_UDP_CSUM_OK, - PKT_RX_L4_CKSUM_BAD); - /* - * PKT_RX_IP_CKSUM_BAD and PKT_RX_L4_CKSUM_BAD are used in place - * of PKT_RX_EIP_CKSUM_BAD because the latter is not functional - * (its value is 0). - */ + PKT_RX_L4_CKSUM_GOOD); if ((flags & IBV_EXP_CQ_RX_TUNNEL_PACKET) && (rxq->csum_l2tun)) ol_flags |= - TRANSPOSE(~flags, + TRANSPOSE(flags, IBV_EXP_CQ_RX_OUTER_IP_CSUM_OK, - PKT_RX_IP_CKSUM_BAD) | - TRANSPOSE(~flags, + PKT_RX_IP_CKSUM_GOOD) | + TRANSPOSE(flags, IBV_EXP_CQ_RX_OUTER_TCP_UDP_CSUM_OK, - PKT_RX_L4_CKSUM_BAD); + PKT_RX_L4_CKSUM_GOOD); return ol_flags; } @@ -2664,7 +3124,7 @@ mlx4_rx_burst_sp(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n) * cacheline while allocating rep. */ rte_prefetch0(seg); - rep = __rte_mbuf_raw_alloc(rxq->mp); + rep = rte_mbuf_raw_alloc(rxq->mp); if (unlikely(rep == NULL)) { /* * Unable to allocate a replacement mbuf, @@ -2693,7 +3153,6 @@ mlx4_rx_burst_sp(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n) rep->ol_flags = -1; #endif assert(rep->buf_len == seg->buf_len); - assert(rep->buf_len == rxq->mb_len); /* Reconfigure sge to use rep instead of seg. */ assert(sge->lkey == rxq->mr->lkey); sge->addr = ((uintptr_t)rep->buf_addr + seg_headroom); @@ -2735,6 +3194,7 @@ mlx4_rx_burst_sp(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n) PORT(pkt_buf) = rxq->port_id; PKT_LEN(pkt_buf) = pkt_buf_len; pkt_buf->packet_type = rxq_cq_to_pkt_type(flags); + pkt_buf->ol_flags = rxq_cq_to_ol_flags(rxq, flags); /* Return packet. */ *(pkts++) = pkt_buf; @@ -2819,6 +3279,12 @@ mlx4_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n) assert(wr->num_sge == 1); assert(elts_head < rxq->elts_n); assert(rxq->elts_head < rxq->elts_n); + /* + * Fetch initial bytes of packet descriptor into a + * cacheline while allocating rep. + */ + rte_mbuf_prefetch_part1(seg); + rte_mbuf_prefetch_part2(seg); ret = rxq->if_cq->poll_length_flags(rxq->cq, NULL, NULL, &flags); if (unlikely(ret < 0)) { @@ -2856,12 +3322,7 @@ mlx4_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n) if (ret == 0) break; len = ret; - /* - * Fetch initial bytes of packet descriptor into a - * cacheline while allocating rep. - */ - rte_prefetch0(seg); - rep = __rte_mbuf_raw_alloc(rxq->mp); + rep = rte_mbuf_raw_alloc(rxq->mp); if (unlikely(rep == NULL)) { /* * Unable to allocate a replacement mbuf, @@ -2931,6 +3392,46 @@ repost: return pkts_ret; } +/** + * DPDK callback for RX in secondary processes. + * + * This function configures all queues from primary process information + * if necessary before reverting to the normal RX burst callback. + * + * @param dpdk_rxq + * Generic pointer to RX queue structure. + * @param[out] pkts + * Array to store received packets. + * @param pkts_n + * Maximum number of packets in array. + * + * @return + * Number of packets successfully received (<= pkts_n). + */ +static uint16_t +mlx4_rx_burst_secondary_setup(void *dpdk_rxq, struct rte_mbuf **pkts, + uint16_t pkts_n) +{ + struct rxq *rxq = dpdk_rxq; + struct priv *priv = mlx4_secondary_data_setup(rxq->priv); + struct priv *primary_priv; + unsigned int index; + + if (priv == NULL) + return 0; + primary_priv = + mlx4_secondary_data[priv->dev->data->port_id].primary_priv; + /* Look for queue index in both private structures. */ + for (index = 0; index != priv->rxqs_n; ++index) + if (((*primary_priv->rxqs)[index] == rxq) || + ((*priv->rxqs)[index] == rxq)) + break; + if (index == priv->rxqs_n) + return 0; + rxq = (*priv->rxqs)[index]; + return priv->dev->rx_pkt_burst(rxq, pkts, pkts_n); +} + /** * Allocate a Queue Pair. * Optionally setup inline receive if supported. @@ -3033,7 +3534,8 @@ rxq_setup_qp_rss(struct priv *priv, struct ibv_cq *cq, uint16_t desc, attr.qpg.qpg_type = IBV_EXP_QPG_PARENT; /* TSS isn't necessary. */ attr.qpg.parent_attrib.tss_child_count = 0; - attr.qpg.parent_attrib.rss_child_count = priv->rxqs_n; + attr.qpg.parent_attrib.rss_child_count = + rte_align32pow2(priv->rxqs_n + 1) >> 1; DEBUG("initializing parent RSS queue"); } else { attr.qpg.qpg_type = IBV_EXP_QPG_CHILD_RX; @@ -3071,6 +3573,7 @@ rxq_rehash(struct rte_eth_dev *dev, struct rxq *rxq) unsigned int i, k; struct ibv_exp_qp_attr mod; struct ibv_recv_wr *bad_wr; + unsigned int mb_len; int err; int parent = (rxq == &priv->rxq_parent); @@ -3079,6 +3582,7 @@ rxq_rehash(struct rte_eth_dev *dev, struct rxq *rxq) (void *)dev, (void *)rxq); return EINVAL; } + mb_len = rte_pktmbuf_data_room_size(rxq->mp); DEBUG("%p: rehashing queue %p", (void *)dev, (void *)rxq); /* Number of descriptors and mbufs currently allocated. */ desc_n = (tmpl.elts_n * (tmpl.sp ? MLX4_PMD_SGE_WR_N : 1)); @@ -3093,9 +3597,10 @@ rxq_rehash(struct rte_eth_dev *dev, struct rxq *rxq) rxq->csum_l2tun = tmpl.csum_l2tun; } /* Enable scattered packets support for this queue if necessary. */ + assert(mb_len >= RTE_PKTMBUF_HEADROOM); if ((dev->data->dev_conf.rxmode.jumbo_frame) && (dev->data->dev_conf.rxmode.max_rx_pkt_len > - (tmpl.mb_len - RTE_PKTMBUF_HEADROOM))) { + (mb_len - RTE_PKTMBUF_HEADROOM))) { tmpl.sp = 1; desc_n /= MLX4_PMD_SGE_WR_N; } else @@ -3256,6 +3761,9 @@ skip_rtr: * Number of descriptors to configure in queue. * @param socket * NUMA socket on which memory must be allocated. + * @param inactive + * If true, the queue is disabled because its index is higher or + * equal to the real number of queues, which must be a power of 2. * @param[in] conf * Thresholds parameters. * @param mp @@ -3266,7 +3774,7 @@ skip_rtr: */ static int rxq_setup(struct rte_eth_dev *dev, struct rxq *rxq, uint16_t desc, - unsigned int socket, const struct rte_eth_rxconf *conf, + unsigned int socket, int inactive, const struct rte_eth_rxconf *conf, struct rte_mempool *mp) { struct priv *priv = dev->data->dev_private; @@ -3283,7 +3791,7 @@ rxq_setup(struct rte_eth_dev *dev, struct rxq *rxq, uint16_t desc, } attr; enum ibv_exp_query_intf_status status; struct ibv_recv_wr *bad_wr; - struct rte_mbuf *buf; + unsigned int mb_len; int ret = 0; int parent = (rxq == &priv->rxq_parent); @@ -3299,42 +3807,29 @@ rxq_setup(struct rte_eth_dev *dev, struct rxq *rxq, uint16_t desc, desc = 1; goto skip_mr; } + mb_len = rte_pktmbuf_data_room_size(mp); if ((desc == 0) || (desc % MLX4_PMD_SGE_WR_N)) { ERROR("%p: invalid number of RX descriptors (must be a" " multiple of %d)", (void *)dev, MLX4_PMD_SGE_WR_N); return EINVAL; } - /* Get mbuf length. */ - buf = rte_pktmbuf_alloc(mp); - if (buf == NULL) { - ERROR("%p: unable to allocate mbuf", (void *)dev); - return ENOMEM; - } - tmpl.mb_len = buf->buf_len; - assert((rte_pktmbuf_headroom(buf) + - rte_pktmbuf_tailroom(buf)) == tmpl.mb_len); - assert(rte_pktmbuf_headroom(buf) == RTE_PKTMBUF_HEADROOM); - rte_pktmbuf_free(buf); /* Toggle RX checksum offload if hardware supports it. */ if (priv->hw_csum) tmpl.csum = !!dev->data->dev_conf.rxmode.hw_ip_checksum; if (priv->hw_csum_l2tun) tmpl.csum_l2tun = !!dev->data->dev_conf.rxmode.hw_ip_checksum; /* Enable scattered packets support for this queue if necessary. */ + assert(mb_len >= RTE_PKTMBUF_HEADROOM); if ((dev->data->dev_conf.rxmode.jumbo_frame) && (dev->data->dev_conf.rxmode.max_rx_pkt_len > - (tmpl.mb_len - RTE_PKTMBUF_HEADROOM))) { + (mb_len - RTE_PKTMBUF_HEADROOM))) { tmpl.sp = 1; desc /= MLX4_PMD_SGE_WR_N; } DEBUG("%p: %s scattered packets support (%u WRs)", (void *)dev, (tmpl.sp ? "enabling" : "disabling"), desc); /* Use the entire RX mempool as the memory region. */ - tmpl.mr = ibv_reg_mr(priv->pd, - (void *)mp->elt_va_start, - (mp->elt_va_end - mp->elt_va_start), - (IBV_ACCESS_LOCAL_WRITE | - IBV_ACCESS_REMOTE_WRITE)); + tmpl.mr = mlx4_mp2mr(priv->pd, mp); if (tmpl.mr == NULL) { ret = EINVAL; ERROR("%p: MR creation failure: %s", @@ -3371,7 +3866,7 @@ skip_mr: DEBUG("priv->device_attr.max_sge is %d", priv->device_attr.max_sge); #ifdef RSS_SUPPORT - if (priv->rss) + if (priv->rss && !inactive) tmpl.qp = rxq_setup_qp_rss(priv, tmpl.cq, desc, parent, tmpl.rd); else @@ -3507,8 +4002,11 @@ mlx4_rx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc, { struct priv *priv = dev->data->dev_private; struct rxq *rxq = (*priv->rxqs)[idx]; + int inactive = 0; int ret; + if (mlx4_is_secondary()) + return -E_RTE_SECONDARY; priv_lock(priv); DEBUG("%p: configuring queue %u for %u descriptors", (void *)dev, idx, desc); @@ -3536,7 +4034,9 @@ mlx4_rx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc, return -ENOMEM; } } - ret = rxq_setup(dev, rxq, desc, socket, conf, mp); + if (idx >= rte_align32pow2(priv->rxqs_n + 1) >> 1) + inactive = 1; + ret = rxq_setup(dev, rxq, desc, socket, inactive, conf, mp); if (ret) rte_free(rxq); else { @@ -3567,6 +4067,8 @@ mlx4_rx_queue_release(void *dpdk_rxq) struct priv *priv; unsigned int i; + if (mlx4_is_secondary()) + return; if (rxq == NULL) return; priv = rxq->priv; @@ -3584,6 +4086,9 @@ mlx4_rx_queue_release(void *dpdk_rxq) priv_unlock(priv); } +static void +priv_dev_interrupt_handler_install(struct priv *, struct rte_eth_dev *); + /** * DPDK callback to start the device. * @@ -3603,6 +4108,8 @@ mlx4_dev_start(struct rte_eth_dev *dev) unsigned int r; struct rxq *rxq; + if (mlx4_is_secondary()) + return -E_RTE_SECONDARY; priv_lock(priv); if (priv->started) { priv_unlock(priv); @@ -3643,8 +4150,10 @@ mlx4_dev_start(struct rte_eth_dev *dev) } } priv->started = 0; + priv_unlock(priv); return -ret; } while ((--r) && ((rxq = (*priv->rxqs)[++i]), i)); + priv_dev_interrupt_handler_install(priv, dev); priv_unlock(priv); return 0; } @@ -3665,6 +4174,8 @@ mlx4_dev_stop(struct rte_eth_dev *dev) unsigned int r; struct rxq *rxq; + if (mlx4_is_secondary()) + return; priv_lock(priv); if (!priv->started) { priv_unlock(priv); @@ -3741,6 +4252,9 @@ removed_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n) return 0; } +static void +priv_dev_interrupt_handler_uninstall(struct priv *, struct rte_eth_dev *); + /** * DPDK callback to close the device. * @@ -3752,10 +4266,12 @@ removed_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n) static void mlx4_dev_close(struct rte_eth_dev *dev) { - struct priv *priv = dev->data->dev_private; + struct priv *priv = mlx4_get_priv(dev); void *tmp; unsigned int i; + if (priv == NULL) + return; priv_lock(priv); DEBUG("%p: closing device \"%s\"", (void *)dev, @@ -3801,10 +4317,95 @@ mlx4_dev_close(struct rte_eth_dev *dev) claim_zero(ibv_close_device(priv->ctx)); } else assert(priv->ctx == NULL); + priv_dev_interrupt_handler_uninstall(priv, dev); priv_unlock(priv); memset(priv, 0, sizeof(*priv)); } +/** + * Change the link state (UP / DOWN). + * + * @param priv + * Pointer to Ethernet device private data. + * @param up + * Nonzero for link up, otherwise link down. + * + * @return + * 0 on success, errno value on failure. + */ +static int +priv_set_link(struct priv *priv, int up) +{ + struct rte_eth_dev *dev = priv->dev; + int err; + unsigned int i; + + if (up) { + err = priv_set_flags(priv, ~IFF_UP, IFF_UP); + if (err) + return err; + for (i = 0; i < priv->rxqs_n; i++) + if ((*priv->rxqs)[i]->sp) + break; + /* Check if an sp queue exists. + * Note: Some old frames might be received. + */ + if (i == priv->rxqs_n) + dev->rx_pkt_burst = mlx4_rx_burst; + else + dev->rx_pkt_burst = mlx4_rx_burst_sp; + dev->tx_pkt_burst = mlx4_tx_burst; + } else { + err = priv_set_flags(priv, ~IFF_UP, ~IFF_UP); + if (err) + return err; + dev->rx_pkt_burst = removed_rx_burst; + dev->tx_pkt_burst = removed_tx_burst; + } + return 0; +} + +/** + * DPDK callback to bring the link DOWN. + * + * @param dev + * Pointer to Ethernet device structure. + * + * @return + * 0 on success, errno value on failure. + */ +static int +mlx4_set_link_down(struct rte_eth_dev *dev) +{ + struct priv *priv = dev->data->dev_private; + int err; + + priv_lock(priv); + err = priv_set_link(priv, 0); + priv_unlock(priv); + return err; +} + +/** + * DPDK callback to bring the link UP. + * + * @param dev + * Pointer to Ethernet device structure. + * + * @return + * 0 on success, errno value on failure. + */ +static int +mlx4_set_link_up(struct rte_eth_dev *dev) +{ + struct priv *priv = dev->data->dev_private; + int err; + + priv_lock(priv); + err = priv_set_link(priv, 1); + priv_unlock(priv); + return err; +} /** * DPDK callback to get information about the device. * @@ -3816,9 +4417,12 @@ mlx4_dev_close(struct rte_eth_dev *dev) static void mlx4_dev_infos_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *info) { - struct priv *priv = dev->data->dev_private; + struct priv *priv = mlx4_get_priv(dev); unsigned int max; + char ifname[IF_NAMESIZE]; + if (priv == NULL) + return; priv_lock(priv); /* FIXME: we should ask the device for these values. */ info->min_rx_bufsize = 32; @@ -3834,7 +4438,8 @@ mlx4_dev_infos_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *info) max = 65535; info->max_rx_queues = max; info->max_tx_queues = max; - info->max_mac_addrs = elemof(priv->mac); + /* Last array entry is reserved for broadcast. */ + info->max_mac_addrs = (elemof(priv->mac) - 1); info->rx_offload_capa = (priv->hw_csum ? (DEV_RX_OFFLOAD_IPV4_CKSUM | @@ -3847,9 +4452,35 @@ mlx4_dev_infos_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *info) DEV_TX_OFFLOAD_UDP_CKSUM | DEV_TX_OFFLOAD_TCP_CKSUM) : 0); + if (priv_get_ifname(priv, &ifname) == 0) + info->if_index = if_nametoindex(ifname); + info->speed_capa = + ETH_LINK_SPEED_1G | + ETH_LINK_SPEED_10G | + ETH_LINK_SPEED_20G | + ETH_LINK_SPEED_40G | + ETH_LINK_SPEED_56G; priv_unlock(priv); } +static const uint32_t * +mlx4_dev_supported_ptypes_get(struct rte_eth_dev *dev) +{ + static const uint32_t ptypes[] = { + /* refers to rxq_cq_to_pkt_type() */ + RTE_PTYPE_L3_IPV4, + RTE_PTYPE_L3_IPV6, + RTE_PTYPE_INNER_L3_IPV4, + RTE_PTYPE_INNER_L3_IPV6, + RTE_PTYPE_UNKNOWN + }; + + if (dev->rx_pkt_burst == mlx4_rx_burst || + dev->rx_pkt_burst == mlx4_rx_burst_sp) + return ptypes; + return NULL; +} + /** * DPDK callback to get device statistics. * @@ -3861,11 +4492,13 @@ mlx4_dev_infos_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *info) static void mlx4_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats) { - struct priv *priv = dev->data->dev_private; + struct priv *priv = mlx4_get_priv(dev); struct rte_eth_stats tmp = {0}; unsigned int i; unsigned int idx; + if (priv == NULL) + return; priv_lock(priv); /* Add software counters. */ for (i = 0; (i != priv->rxqs_n); ++i) { @@ -3924,10 +4557,12 @@ mlx4_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats) static void mlx4_stats_reset(struct rte_eth_dev *dev) { - struct priv *priv = dev->data->dev_private; + struct priv *priv = mlx4_get_priv(dev); unsigned int i; unsigned int idx; + if (priv == NULL) + return; priv_lock(priv); for (i = 0; (i != priv->rxqs_n); ++i) { if ((*priv->rxqs)[i] == NULL) @@ -3939,7 +4574,7 @@ mlx4_stats_reset(struct rte_eth_dev *dev) for (i = 0; (i != priv->txqs_n); ++i) { if ((*priv->txqs)[i] == NULL) continue; - idx = (*priv->rxqs)[i]->stats.idx; + idx = (*priv->txqs)[i]->stats.idx; (*priv->txqs)[i]->stats = (struct mlx4_txq_stats){ .idx = idx }; } @@ -3962,14 +4597,13 @@ mlx4_mac_addr_remove(struct rte_eth_dev *dev, uint32_t index) { struct priv *priv = dev->data->dev_private; + if (mlx4_is_secondary()) + return; priv_lock(priv); DEBUG("%p: removing MAC address from index %" PRIu32, (void *)dev, index); - if (index >= MLX4_MAX_MAC_ADDRESSES) - goto end; - /* Refuse to remove the broadcast address, this one is special. */ - if (!memcmp(priv->mac[index].addr_bytes, "\xff\xff\xff\xff\xff\xff", - ETHER_ADDR_LEN)) + /* Last array entry is reserved for broadcast. */ + if (index >= (elemof(priv->mac) - 1)) goto end; priv_mac_addr_del(priv, index); end: @@ -3994,15 +4628,14 @@ mlx4_mac_addr_add(struct rte_eth_dev *dev, struct ether_addr *mac_addr, { struct priv *priv = dev->data->dev_private; + if (mlx4_is_secondary()) + return; (void)vmdq; priv_lock(priv); DEBUG("%p: adding MAC address at index %" PRIu32, (void *)dev, index); - if (index >= MLX4_MAX_MAC_ADDRESSES) - goto end; - /* Refuse to add the broadcast address, this one is special. */ - if (!memcmp(mac_addr->addr_bytes, "\xff\xff\xff\xff\xff\xff", - ETHER_ADDR_LEN)) + /* Last array entry is reserved for broadcast. */ + if (index >= (elemof(priv->mac) - 1)) goto end; priv_mac_addr_add(priv, index, (const uint8_t (*)[ETHER_ADDR_LEN]) @@ -4011,6 +4644,22 @@ end: priv_unlock(priv); } +/** + * DPDK callback to set the primary MAC address. + * + * @param dev + * Pointer to Ethernet device structure. + * @param mac_addr + * MAC address to register. + */ +static void +mlx4_mac_addr_set(struct rte_eth_dev *dev, struct ether_addr *mac_addr) +{ + DEBUG("%p: setting primary MAC address", (void *)dev); + mlx4_mac_addr_remove(dev, 0); + mlx4_mac_addr_add(dev, mac_addr, 0, 0); +} + /** * DPDK callback to enable promiscuous mode. * @@ -4024,6 +4673,8 @@ mlx4_promiscuous_enable(struct rte_eth_dev *dev) unsigned int i; int ret; + if (mlx4_is_secondary()) + return; priv_lock(priv); if (priv->promisc) { priv_unlock(priv); @@ -4070,6 +4721,8 @@ mlx4_promiscuous_disable(struct rte_eth_dev *dev) struct priv *priv = dev->data->dev_private; unsigned int i; + if (mlx4_is_secondary()) + return; priv_lock(priv); if (!priv->promisc) { priv_unlock(priv); @@ -4100,6 +4753,8 @@ mlx4_allmulticast_enable(struct rte_eth_dev *dev) unsigned int i; int ret; + if (mlx4_is_secondary()) + return; priv_lock(priv); if (priv->allmulti) { priv_unlock(priv); @@ -4146,6 +4801,8 @@ mlx4_allmulticast_disable(struct rte_eth_dev *dev) struct priv *priv = dev->data->dev_private; unsigned int i; + if (mlx4_is_secondary()) + return; priv_lock(priv); if (!priv->allmulti) { priv_unlock(priv); @@ -4174,7 +4831,7 @@ end: static int mlx4_link_update_unlocked(struct rte_eth_dev *dev, int wait_to_complete) { - struct priv *priv = dev->data->dev_private; + struct priv *priv = mlx4_get_priv(dev); struct ethtool_cmd edata = { .cmd = ETHTOOL_GSET }; @@ -4182,6 +4839,8 @@ mlx4_link_update_unlocked(struct rte_eth_dev *dev, int wait_to_complete) struct rte_eth_link dev_link; int link_speed = 0; + if (priv == NULL) + return -EINVAL; (void)wait_to_complete; if (priv_ifreq(priv, SIOCGIFFLAGS, &ifr)) { WARN("ioctl(SIOCGIFFLAGS) failed: %s", strerror(errno)); @@ -4190,7 +4849,7 @@ mlx4_link_update_unlocked(struct rte_eth_dev *dev, int wait_to_complete) memset(&dev_link, 0, sizeof(dev_link)); dev_link.link_status = ((ifr.ifr_flags & IFF_UP) && (ifr.ifr_flags & IFF_RUNNING)); - ifr.ifr_data = &edata; + ifr.ifr_data = (void *)&edata; if (priv_ifreq(priv, SIOCETHTOOL, &ifr)) { WARN("ioctl(SIOCETHTOOL, ETHTOOL_GSET) failed: %s", strerror(errno)); @@ -4203,6 +4862,8 @@ mlx4_link_update_unlocked(struct rte_eth_dev *dev, int wait_to_complete) dev_link.link_speed = link_speed; dev_link.link_duplex = ((edata.duplex == DUPLEX_HALF) ? ETH_LINK_HALF_DUPLEX : ETH_LINK_FULL_DUPLEX); + dev_link.link_autoneg = !(dev->data->dev_conf.link_speeds & + ETH_LINK_SPEED_FIXED); if (memcmp(&dev_link, &dev->data->dev_link, sizeof(dev_link))) { /* Link status changed. */ dev->data->dev_link = dev_link; @@ -4223,9 +4884,11 @@ mlx4_link_update_unlocked(struct rte_eth_dev *dev, int wait_to_complete) static int mlx4_link_update(struct rte_eth_dev *dev, int wait_to_complete) { - struct priv *priv = dev->data->dev_private; + struct priv *priv = mlx4_get_priv(dev); int ret; + if (priv == NULL) + return -EINVAL; priv_lock(priv); ret = mlx4_link_update_unlocked(dev, wait_to_complete); priv_unlock(priv); @@ -4258,6 +4921,8 @@ mlx4_dev_set_mtu(struct rte_eth_dev *dev, uint16_t mtu) uint16_t (*rx_func)(void *, struct rte_mbuf **, uint16_t) = mlx4_rx_burst; + if (mlx4_is_secondary()) + return -E_RTE_SECONDARY; priv_lock(priv); /* Set kernel interface MTU first. */ if (priv_set_mtu(priv, mtu)) { @@ -4278,6 +4943,7 @@ mlx4_dev_set_mtu(struct rte_eth_dev *dev, uint16_t mtu) /* Reconfigure each RX queue. */ for (i = 0; (i != priv->rxqs_n); ++i) { struct rxq *rxq = (*priv->rxqs)[i]; + unsigned int mb_len; unsigned int max_frame_len; int sp; @@ -4287,7 +4953,9 @@ mlx4_dev_set_mtu(struct rte_eth_dev *dev, uint16_t mtu) * toggle scattered support (sp) if necessary. */ max_frame_len = (priv->mtu + ETHER_HDR_LEN + (ETHER_MAX_VLAN_FRAME_LEN - ETHER_MAX_LEN)); - sp = (max_frame_len > (rxq->mb_len - RTE_PKTMBUF_HEADROOM)); + mb_len = rte_pktmbuf_data_room_size(rxq->mp); + assert(mb_len >= RTE_PKTMBUF_HEADROOM); + sp = (max_frame_len > (mb_len - RTE_PKTMBUF_HEADROOM)); /* Provide new values to rxq_setup(). */ dev->data->dev_conf.rxmode.jumbo_frame = sp; dev->data->dev_conf.rxmode.max_rx_pkt_len = max_frame_len; @@ -4341,7 +5009,9 @@ mlx4_dev_get_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf) }; int ret; - ifr.ifr_data = ðpause; + if (mlx4_is_secondary()) + return -E_RTE_SECONDARY; + ifr.ifr_data = (void *)ðpause; priv_lock(priv); if (priv_ifreq(priv, SIOCETHTOOL, &ifr)) { ret = errno; @@ -4389,7 +5059,9 @@ mlx4_dev_set_flow_ctrl(struct rte_eth_dev *dev, struct rte_eth_fc_conf *fc_conf) }; int ret; - ifr.ifr_data = ðpause; + if (mlx4_is_secondary()) + return -E_RTE_SECONDARY; + ifr.ifr_data = (void *)ðpause; ethpause.autoneg = fc_conf->autoneg; if (((fc_conf->mode & RTE_FC_FULL) == RTE_FC_FULL) || (fc_conf->mode & RTE_FC_RX_PAUSE)) @@ -4528,6 +5200,8 @@ mlx4_vlan_filter_set(struct rte_eth_dev *dev, uint16_t vlan_id, int on) struct priv *priv = dev->data->dev_private; int ret; + if (mlx4_is_secondary()) + return -E_RTE_SECONDARY; priv_lock(priv); ret = vlan_filter_set(dev, vlan_id, on); priv_unlock(priv); @@ -4539,6 +5213,8 @@ static const struct eth_dev_ops mlx4_dev_ops = { .dev_configure = mlx4_dev_configure, .dev_start = mlx4_dev_start, .dev_stop = mlx4_dev_stop, + .dev_set_link_down = mlx4_set_link_down, + .dev_set_link_up = mlx4_set_link_up, .dev_close = mlx4_dev_close, .promiscuous_enable = mlx4_promiscuous_enable, .promiscuous_disable = mlx4_promiscuous_disable, @@ -4549,6 +5225,7 @@ static const struct eth_dev_ops mlx4_dev_ops = { .stats_reset = mlx4_stats_reset, .queue_stats_mapping_set = NULL, .dev_infos_get = mlx4_dev_infos_get, + .dev_supported_ptypes_get = mlx4_dev_supported_ptypes_get, .vlan_filter_set = mlx4_vlan_filter_set, .vlan_tpid_set = NULL, .vlan_strip_queue_set = NULL, @@ -4564,9 +5241,8 @@ static const struct eth_dev_ops mlx4_dev_ops = { .priority_flow_ctrl_set = NULL, .mac_addr_remove = mlx4_mac_addr_remove, .mac_addr_add = mlx4_mac_addr_add, + .mac_addr_set = mlx4_mac_addr_set, .mtu_set = mlx4_dev_set_mtu, - .udp_tunnel_add = NULL, - .udp_tunnel_del = NULL, }; /** @@ -4694,6 +5370,158 @@ mlx4_getenv_int(const char *name) return atoi(val); } +static void +mlx4_dev_link_status_handler(void *); +static void +mlx4_dev_interrupt_handler(struct rte_intr_handle *, void *); + +/** + * Link status handler. + * + * @param priv + * Pointer to private structure. + * @param dev + * Pointer to the rte_eth_dev structure. + * + * @return + * Nonzero if the callback process can be called immediately. + */ +static int +priv_dev_link_status_handler(struct priv *priv, struct rte_eth_dev *dev) +{ + struct ibv_async_event event; + int port_change = 0; + int ret = 0; + + /* Read all message and acknowledge them. */ + for (;;) { + if (ibv_get_async_event(priv->ctx, &event)) + break; + + if (event.event_type == IBV_EVENT_PORT_ACTIVE || + event.event_type == IBV_EVENT_PORT_ERR) + port_change = 1; + else + DEBUG("event type %d on port %d not handled", + event.event_type, event.element.port_num); + ibv_ack_async_event(&event); + } + + if (port_change ^ priv->pending_alarm) { + struct rte_eth_link *link = &dev->data->dev_link; + + priv->pending_alarm = 0; + mlx4_link_update_unlocked(dev, 0); + if (((link->link_speed == 0) && link->link_status) || + ((link->link_speed != 0) && !link->link_status)) { + /* Inconsistent status, check again later. */ + priv->pending_alarm = 1; + rte_eal_alarm_set(MLX4_ALARM_TIMEOUT_US, + mlx4_dev_link_status_handler, + dev); + } else + ret = 1; + } + return ret; +} + +/** + * Handle delayed link status event. + * + * @param arg + * Registered argument. + */ +static void +mlx4_dev_link_status_handler(void *arg) +{ + struct rte_eth_dev *dev = arg; + struct priv *priv = dev->data->dev_private; + int ret; + + priv_lock(priv); + assert(priv->pending_alarm == 1); + ret = priv_dev_link_status_handler(priv, dev); + priv_unlock(priv); + if (ret) + _rte_eth_dev_callback_process(dev, RTE_ETH_EVENT_INTR_LSC, NULL); +} + +/** + * Handle interrupts from the NIC. + * + * @param[in] intr_handle + * Interrupt handler. + * @param cb_arg + * Callback argument. + */ +static void +mlx4_dev_interrupt_handler(struct rte_intr_handle *intr_handle, void *cb_arg) +{ + struct rte_eth_dev *dev = cb_arg; + struct priv *priv = dev->data->dev_private; + int ret; + + (void)intr_handle; + priv_lock(priv); + ret = priv_dev_link_status_handler(priv, dev); + priv_unlock(priv); + if (ret) + _rte_eth_dev_callback_process(dev, RTE_ETH_EVENT_INTR_LSC, NULL); +} + +/** + * Uninstall interrupt handler. + * + * @param priv + * Pointer to private structure. + * @param dev + * Pointer to the rte_eth_dev structure. + */ +static void +priv_dev_interrupt_handler_uninstall(struct priv *priv, struct rte_eth_dev *dev) +{ + if (!dev->data->dev_conf.intr_conf.lsc) + return; + rte_intr_callback_unregister(&priv->intr_handle, + mlx4_dev_interrupt_handler, + dev); + if (priv->pending_alarm) + rte_eal_alarm_cancel(mlx4_dev_link_status_handler, dev); + priv->pending_alarm = 0; + priv->intr_handle.fd = 0; + priv->intr_handle.type = RTE_INTR_HANDLE_UNKNOWN; +} + +/** + * Install interrupt handler. + * + * @param priv + * Pointer to private structure. + * @param dev + * Pointer to the rte_eth_dev structure. + */ +static void +priv_dev_interrupt_handler_install(struct priv *priv, struct rte_eth_dev *dev) +{ + int rc, flags; + + if (!dev->data->dev_conf.intr_conf.lsc) + return; + assert(priv->ctx->async_fd > 0); + flags = fcntl(priv->ctx->async_fd, F_GETFL); + rc = fcntl(priv->ctx->async_fd, F_SETFL, flags | O_NONBLOCK); + if (rc < 0) { + INFO("failed to change file descriptor async event queue"); + dev->data->dev_conf.intr_conf.lsc = 0; + } else { + priv->intr_handle.fd = priv->ctx->async_fd; + priv->intr_handle.type = RTE_INTR_HANDLE_EXT; + rte_intr_callback_register(&priv->intr_handle, + mlx4_dev_interrupt_handler, + dev); + } +} + static struct eth_driver mlx4_driver; /** @@ -4711,7 +5539,7 @@ static struct eth_driver mlx4_driver; * 0 on success, negative errno value on failure. */ static int -mlx4_pci_devinit(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev) +mlx4_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev) { struct ibv_device **list; struct ibv_device *ibv_dev; @@ -4795,7 +5623,7 @@ mlx4_pci_devinit(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev) struct ibv_port_attr port_attr; struct ibv_pd *pd = NULL; struct priv *priv = NULL; - struct rte_eth_dev *eth_dev; + struct rte_eth_dev *eth_dev = NULL; #ifdef HAVE_EXP_QUERY_DEVICE struct ibv_exp_device_attr exp_device_attr; #endif /* HAVE_EXP_QUERY_DEVICE */ @@ -4820,10 +5648,17 @@ mlx4_pci_devinit(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev) ERROR("port query failed: %s", strerror(err)); goto port_error; } + + if (port_attr.link_layer != IBV_LINK_LAYER_ETHERNET) { + ERROR("port %d is not configured in Ethernet mode", + port); + goto port_error; + } + if (port_attr.state != IBV_PORT_ACTIVE) - WARN("bad state for port %d: \"%s\" (%d)", - port, ibv_port_state_str(port_attr.state), - port_attr.state); + DEBUG("port %d is not active: \"%s\" (%d)", + port, ibv_port_state_str(port_attr.state), + port_attr.state); /* Allocate protection domain. */ pd = ibv_alloc_pd(ctx); @@ -4939,7 +5774,7 @@ mlx4_pci_devinit(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev) claim_zero(priv_mac_addr_add(priv, 0, (const uint8_t (*)[ETHER_ADDR_LEN]) mac.addr_bytes)); - claim_zero(priv_mac_addr_add(priv, 1, + claim_zero(priv_mac_addr_add(priv, (elemof(priv->mac) - 1), &(const uint8_t [ETHER_ADDR_LEN]) { "\xff\xff\xff\xff\xff\xff" })); #ifndef NDEBUG @@ -4963,7 +5798,7 @@ mlx4_pci_devinit(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev) snprintf(name, sizeof(name), "%s port %u", ibv_get_device_name(ibv_dev), port); - eth_dev = rte_eth_dev_allocate(name, RTE_ETH_DEV_PCI); + eth_dev = rte_eth_dev_allocate(name); } if (eth_dev == NULL) { ERROR("can not allocate rte ethdev"); @@ -4971,15 +5806,47 @@ mlx4_pci_devinit(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev) goto port_error; } - eth_dev->data->dev_private = priv; + /* Secondary processes have to use local storage for their + * private data as well as a copy of eth_dev->data, but this + * pointer must not be modified before burst functions are + * actually called. */ + if (mlx4_is_secondary()) { + struct mlx4_secondary_data *sd = + &mlx4_secondary_data[eth_dev->data->port_id]; + + sd->primary_priv = eth_dev->data->dev_private; + if (sd->primary_priv == NULL) { + ERROR("no private data for port %u", + eth_dev->data->port_id); + err = EINVAL; + goto port_error; + } + sd->shared_dev_data = eth_dev->data; + rte_spinlock_init(&sd->lock); + memcpy(sd->data.name, sd->shared_dev_data->name, + sizeof(sd->data.name)); + sd->data.dev_private = priv; + sd->data.rx_mbuf_alloc_failed = 0; + sd->data.mtu = ETHER_MTU; + sd->data.port_id = sd->shared_dev_data->port_id; + sd->data.mac_addrs = priv->mac; + eth_dev->tx_pkt_burst = mlx4_tx_burst_secondary_setup; + eth_dev->rx_pkt_burst = mlx4_rx_burst_secondary_setup; + } else { + eth_dev->data->dev_private = priv; + eth_dev->data->rx_mbuf_alloc_failed = 0; + eth_dev->data->mtu = ETHER_MTU; + eth_dev->data->mac_addrs = priv->mac; + } eth_dev->pci_dev = pci_dev; + + rte_eth_copy_pci_info(eth_dev, pci_dev); + eth_dev->driver = &mlx4_driver; - eth_dev->data->rx_mbuf_alloc_failed = 0; - eth_dev->data->mtu = ETHER_MTU; priv->dev = eth_dev; eth_dev->dev_ops = &mlx4_dev_ops; - eth_dev->data->mac_addrs = priv->mac; + TAILQ_INIT(ð_dev->link_intr_cbs); /* Bring Ethernet device up. */ DEBUG("forcing Ethernet interface up"); @@ -4992,6 +5859,8 @@ port_error: claim_zero(ibv_dealloc_pd(pd)); if (ctx) claim_zero(ibv_close_device(ctx)); + if (eth_dev) + rte_eth_dev_release_port(eth_dev); break; } @@ -5019,22 +5888,16 @@ error: static const struct rte_pci_id mlx4_pci_id_map[] = { { - .vendor_id = PCI_VENDOR_ID_MELLANOX, - .device_id = PCI_DEVICE_ID_MELLANOX_CONNECTX3, - .subsystem_vendor_id = PCI_ANY_ID, - .subsystem_device_id = PCI_ANY_ID + RTE_PCI_DEVICE(PCI_VENDOR_ID_MELLANOX, + PCI_DEVICE_ID_MELLANOX_CONNECTX3) }, { - .vendor_id = PCI_VENDOR_ID_MELLANOX, - .device_id = PCI_DEVICE_ID_MELLANOX_CONNECTX3PRO, - .subsystem_vendor_id = PCI_ANY_ID, - .subsystem_device_id = PCI_ANY_ID + RTE_PCI_DEVICE(PCI_VENDOR_ID_MELLANOX, + PCI_DEVICE_ID_MELLANOX_CONNECTX3PRO) }, { - .vendor_id = PCI_VENDOR_ID_MELLANOX, - .device_id = PCI_DEVICE_ID_MELLANOX_CONNECTX3VF, - .subsystem_vendor_id = PCI_ANY_ID, - .subsystem_device_id = PCI_ANY_ID + RTE_PCI_DEVICE(PCI_VENDOR_ID_MELLANOX, + PCI_DEVICE_ID_MELLANOX_CONNECTX3VF) }, { .vendor_id = 0 @@ -5043,9 +5906,12 @@ static const struct rte_pci_id mlx4_pci_id_map[] = { static struct eth_driver mlx4_driver = { .pci_drv = { - .name = MLX4_DRIVER_NAME, + .driver = { + .name = MLX4_DRIVER_NAME + }, .id_table = mlx4_pci_id_map, - .devinit = mlx4_pci_devinit, + .probe = mlx4_pci_probe, + .drv_flags = RTE_PCI_DRV_INTR_LSC, }, .dev_private_size = sizeof(struct priv) }; @@ -5053,11 +5919,11 @@ static struct eth_driver mlx4_driver = { /** * Driver initialization routine. */ -static int -rte_mlx4_pmd_init(const char *name, const char *args) +RTE_INIT(rte_mlx4_pmd_init); +static void +rte_mlx4_pmd_init(void) { - (void)name; - (void)args; + RTE_BUILD_BUG_ON(sizeof(wr_id_t) != sizeof(uint64_t)); /* * RDMAV_HUGEPAGES_SAFE tells ibv_fork_init() we intend to use * huge pages. Calling ibv_fork_init() during init allows @@ -5067,13 +5933,9 @@ rte_mlx4_pmd_init(const char *name, const char *args) setenv("RDMAV_HUGEPAGES_SAFE", "1", 1); ibv_fork_init(); rte_eal_pci_register(&mlx4_driver.pci_drv); - return 0; } -static struct rte_driver rte_mlx4_driver = { - .type = PMD_PDEV, - .name = MLX4_DRIVER_NAME, - .init = rte_mlx4_pmd_init, -}; - -PMD_REGISTER_DRIVER(rte_mlx4_driver) +RTE_PMD_EXPORT_NAME(net_mlx4, __COUNTER__); +RTE_PMD_REGISTER_PCI_TABLE(net_mlx4, mlx4_pci_id_map); +RTE_PMD_REGISTER_KMOD_DEP(net_mlx4, + "* ib_uverbs & mlx4_en & mlx4_core & mlx4_ib");