X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fmlx5%2Fmlx5.c;h=7d58d66bb92b48c869cf78bbcf975bf7e2d1cca1;hb=d8dd31652cf41e1ab0f4e23aab605a1d2114ac8d;hp=5a959f7c525afe2982538a92b0946622a421333c;hpb=4a984153430cbdafe2b971e999bba8d96610b299;p=dpdk.git diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c index 5a959f7c52..7d58d66bb9 100644 --- a/drivers/net/mlx5/mlx5.c +++ b/drivers/net/mlx5/mlx5.c @@ -1,40 +1,13 @@ -/*- - * BSD LICENSE - * - * Copyright 2015 6WIND S.A. - * Copyright 2015 Mellanox. - * - * Redistribution and use in source and binary forms, with or without - * modification, are permitted provided that the following conditions - * are met: - * - * * Redistributions of source code must retain the above copyright - * notice, this list of conditions and the following disclaimer. - * * 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. - * * Neither the name of 6WIND S.A. nor the names of its - * contributors may be used to endorse or promote products derived - * from this software without specific prior written permission. - * - * 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. +/* SPDX-License-Identifier: BSD-3-Clause + * Copyright 2015 6WIND S.A. + * Copyright 2015 Mellanox. */ #include #include #include #include +#include #include #include #include @@ -57,6 +30,7 @@ #include #include #include +#include #include #include @@ -65,6 +39,7 @@ #include "mlx5_rxtx.h" #include "mlx5_autoconf.h" #include "mlx5_defs.h" +#include "mlx5_glue.h" /* Device parameter to enable RX completion queue compression. */ #define MLX5_RXQ_CQE_COMP_EN "rxq_cqe_comp_en" @@ -102,6 +77,9 @@ #define MLX5DV_CONTEXT_FLAGS_CQE_128B_COMP (1 << 4) #endif +/** Driver-specific log messages type. */ +int mlx5_logtype; + /** * Retrieve integer value from environment variable. * @@ -133,7 +111,7 @@ mlx5_getenv_int(const char *name) * A pointer to the callback data. * * @return - * a pointer to the allocate space. + * Allocated buffer, NULL otherwise and rte_errno is set. */ static void * mlx5_alloc_verbs_buf(size_t size, void *data) @@ -155,7 +133,8 @@ mlx5_alloc_verbs_buf(size_t size, void *data) } assert(data != NULL); ret = rte_malloc_socket(__func__, size, alignment, socket); - DEBUG("Extern alloc size: %lu, align: %lu: %p", size, alignment, ret); + if (!ret && size) + rte_errno = ENOMEM; return ret; } @@ -171,7 +150,6 @@ static void mlx5_free_verbs_buf(void *ptr, void *data __rte_unused) { assert(data != NULL); - DEBUG("Extern free request: %p", ptr); rte_free(ptr); } @@ -190,13 +168,12 @@ mlx5_dev_close(struct rte_eth_dev *dev) unsigned int i; int ret; - priv_lock(priv); - DEBUG("%p: closing device \"%s\"", - (void *)dev, - ((priv->ctx != NULL) ? priv->ctx->device->name : "")); + DRV_LOG(DEBUG, "port %u closing device \"%s\"", + dev->data->port_id, + ((priv->ctx != NULL) ? priv->ctx->device->name : "")); /* In case mlx5_dev_stop() has not been called. */ - priv_dev_interrupt_handler_uninstall(priv, dev); - priv_dev_traffic_disable(priv, dev); + mlx5_dev_interrupt_handler_uninstall(dev); + mlx5_traffic_disable(dev); /* Prevent crashes when queues are still in use. */ dev->rx_pkt_burst = removed_rx_burst; dev->tx_pkt_burst = removed_tx_burst; @@ -204,7 +181,7 @@ mlx5_dev_close(struct rte_eth_dev *dev) /* XXX race condition if mlx5_rx_burst() is still running. */ usleep(1000); for (i = 0; (i != priv->rxqs_n); ++i) - mlx5_priv_rxq_release(priv, i); + mlx5_rxq_release(dev, i); priv->rxqs_n = 0; priv->rxqs = NULL; } @@ -212,46 +189,54 @@ mlx5_dev_close(struct rte_eth_dev *dev) /* XXX race condition if mlx5_tx_burst() is still running. */ usleep(1000); for (i = 0; (i != priv->txqs_n); ++i) - mlx5_priv_txq_release(priv, i); + mlx5_txq_release(dev, i); priv->txqs_n = 0; priv->txqs = NULL; } if (priv->pd != NULL) { assert(priv->ctx != NULL); - claim_zero(ibv_dealloc_pd(priv->pd)); - claim_zero(ibv_close_device(priv->ctx)); + claim_zero(mlx5_glue->dealloc_pd(priv->pd)); + claim_zero(mlx5_glue->close_device(priv->ctx)); } else assert(priv->ctx == NULL); if (priv->rss_conf.rss_key != NULL) rte_free(priv->rss_conf.rss_key); if (priv->reta_idx != NULL) rte_free(priv->reta_idx); - priv_socket_uninit(priv); - ret = mlx5_priv_hrxq_ibv_verify(priv); + if (priv->primary_socket) + mlx5_socket_uninit(dev); + ret = mlx5_hrxq_ibv_verify(dev); if (ret) - WARN("%p: some Hash Rx queue still remain", (void *)priv); - ret = mlx5_priv_ind_table_ibv_verify(priv); + DRV_LOG(WARNING, "port %u some hash Rx queue still remain", + dev->data->port_id); + ret = mlx5_ind_table_ibv_verify(dev); if (ret) - WARN("%p: some Indirection table still remain", (void *)priv); - ret = mlx5_priv_rxq_ibv_verify(priv); + DRV_LOG(WARNING, "port %u some indirection table still remain", + dev->data->port_id); + ret = mlx5_rxq_ibv_verify(dev); if (ret) - WARN("%p: some Verbs Rx queue still remain", (void *)priv); - ret = mlx5_priv_rxq_verify(priv); + DRV_LOG(WARNING, "port %u some Verbs Rx queue still remain", + dev->data->port_id); + ret = mlx5_rxq_verify(dev); if (ret) - WARN("%p: some Rx Queues still remain", (void *)priv); - ret = mlx5_priv_txq_ibv_verify(priv); + DRV_LOG(WARNING, "port %u some Rx queues still remain", + dev->data->port_id); + ret = mlx5_txq_ibv_verify(dev); if (ret) - WARN("%p: some Verbs Tx queue still remain", (void *)priv); - ret = mlx5_priv_txq_verify(priv); + DRV_LOG(WARNING, "port %u some Verbs Tx queue still remain", + dev->data->port_id); + ret = mlx5_txq_verify(dev); if (ret) - WARN("%p: some Tx Queues still remain", (void *)priv); - ret = priv_flow_verify(priv); + DRV_LOG(WARNING, "port %u some Tx queues still remain", + dev->data->port_id); + ret = mlx5_flow_verify(dev); if (ret) - WARN("%p: some flows still remain", (void *)priv); - ret = priv_mr_verify(priv); + DRV_LOG(WARNING, "port %u some flows still remain", + dev->data->port_id); + ret = mlx5_mr_verify(dev); if (ret) - WARN("%p: some Memory Region still remain", (void *)priv); - priv_unlock(priv); + DRV_LOG(WARNING, "port %u some memory region still remain", + dev->data->port_id); memset(priv, 0, sizeof(*priv)); } @@ -391,7 +376,7 @@ mlx5_dev_idx(struct rte_pci_addr *pci_addr) * User data. * * @return - * 0 on success, negative errno value on failure. + * 0 on success, a negative errno value otherwise and rte_errno is set. */ static int mlx5_args_check(const char *key, const char *val, void *opaque) @@ -402,8 +387,9 @@ mlx5_args_check(const char *key, const char *val, void *opaque) errno = 0; tmp = strtoul(val, NULL, 0); if (errno) { - WARN("%s: \"%s\" is not a valid integer", key, val); - return errno; + rte_errno = errno; + DRV_LOG(WARNING, "%s: \"%s\" is not a valid integer", key, val); + return -rte_errno; } if (strcmp(MLX5_RXQ_CQE_COMP_EN, key) == 0) { config->cqe_comp = !!tmp; @@ -422,8 +408,9 @@ mlx5_args_check(const char *key, const char *val, void *opaque) } else if (strcmp(MLX5_RX_VEC_EN, key) == 0) { config->rx_vec_en = !!tmp; } else { - WARN("%s: unknown parameter", key); - return -EINVAL; + DRV_LOG(WARNING, "%s: unknown parameter", key); + rte_errno = EINVAL; + return -rte_errno; } return 0; } @@ -437,7 +424,7 @@ mlx5_args_check(const char *key, const char *val, void *opaque) * Device arguments structure. * * @return - * 0 on success, errno value on failure. + * 0 on success, a negative errno value otherwise and rte_errno is set. */ static int mlx5_args(struct mlx5_dev_config *config, struct rte_devargs *devargs) @@ -468,9 +455,10 @@ mlx5_args(struct mlx5_dev_config *config, struct rte_devargs *devargs) if (rte_kvargs_count(kvlist, params[i])) { ret = rte_kvargs_process(kvlist, params[i], mlx5_args_check, config); - if (ret != 0) { + if (ret) { + rte_errno = EINVAL; rte_kvargs_free(kvlist); - return ret; + return -rte_errno; } } } @@ -492,19 +480,19 @@ static void *uar_base; /** * Reserve UAR address space for primary process. * - * @param[in] priv - * Pointer to private structure. + * @param[in] dev + * Pointer to Ethernet device. * * @return - * 0 on success, errno value on failure. + * 0 on success, a negative errno value otherwise and rte_errno is set. */ static int -priv_uar_init_primary(struct priv *priv) +mlx5_uar_init_primary(struct rte_eth_dev *dev) { + struct priv *priv = dev->data->dev_private; void *addr = (void *)0; int i; const struct rte_mem_config *mcfg; - int ret; if (uar_base) { /* UAR address space mapped. */ priv->uar_base = uar_base; @@ -524,15 +512,18 @@ priv_uar_init_primary(struct priv *priv) addr = mmap(addr, MLX5_UAR_SIZE, PROT_NONE, MAP_PRIVATE | MAP_ANONYMOUS, -1, 0); if (addr == MAP_FAILED) { - ERROR("Failed to reserve UAR address space, please adjust " - "MLX5_UAR_SIZE or try --base-virtaddr"); - ret = ENOMEM; - return ret; + DRV_LOG(ERR, + "port %u failed to reserve UAR address space, please" + " adjust MLX5_UAR_SIZE or try --base-virtaddr", + dev->data->port_id); + rte_errno = ENOMEM; + return -rte_errno; } /* Accept either same addr or a new addr returned from mmap if target * range occupied. */ - INFO("Reserved UAR address space: %p", addr); + DRV_LOG(INFO, "port %u reserved UAR address space: %p", + dev->data->port_id, addr); priv->uar_base = addr; /* for primary and secondary UAR re-mmap. */ uar_base = addr; /* process local, don't reserve again. */ return 0; @@ -542,17 +533,17 @@ priv_uar_init_primary(struct priv *priv) * Reserve UAR address space for secondary process, align with * primary process. * - * @param[in] priv - * Pointer to private structure. + * @param[in] dev + * Pointer to Ethernet device. * * @return - * 0 on success, errno value on failure. + * 0 on success, a negative errno value otherwise and rte_errno is set. */ static int -priv_uar_init_secondary(struct priv *priv) +mlx5_uar_init_secondary(struct rte_eth_dev *dev) { + struct priv *priv = dev->data->dev_private; void *addr; - int ret; assert(priv->uar_base); if (uar_base) { /* already reserved. */ @@ -563,20 +554,23 @@ priv_uar_init_secondary(struct priv *priv) addr = mmap(priv->uar_base, MLX5_UAR_SIZE, PROT_NONE, MAP_PRIVATE | MAP_ANONYMOUS, -1, 0); if (addr == MAP_FAILED) { - ERROR("UAR mmap failed: %p size: %llu", - priv->uar_base, MLX5_UAR_SIZE); - ret = ENXIO; - return ret; + DRV_LOG(ERR, "port %u UAR mmap failed: %p size: %llu", + dev->data->port_id, priv->uar_base, MLX5_UAR_SIZE); + rte_errno = ENXIO; + return -rte_errno; } if (priv->uar_base != addr) { - ERROR("UAR address %p size %llu occupied, please adjust " - "MLX5_UAR_OFFSET or try EAL parameter --base-virtaddr", - priv->uar_base, MLX5_UAR_SIZE); - ret = ENXIO; - return ret; + DRV_LOG(ERR, + "port %u UAR address %p size %llu occupied, please" + " adjust MLX5_UAR_OFFSET or try EAL parameter" + " --base-virtaddr", + dev->data->port_id, priv->uar_base, MLX5_UAR_SIZE); + rte_errno = ENXIO; + return -rte_errno; } uar_base = addr; /* process local, don't reserve again */ - INFO("Reserved UAR address space: %p", addr); + DRV_LOG(INFO, "port %u reserved UAR address space: %p", + dev->data->port_id, addr); return 0; } @@ -592,45 +586,46 @@ priv_uar_init_secondary(struct priv *priv) * PCI device information. * * @return - * 0 on success, negative errno value on failure. + * 0 on success, a negative errno value otherwise and rte_errno is set. */ static int -mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev) +mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused, + struct rte_pci_device *pci_dev) { - struct ibv_device **list; + struct ibv_device **list = NULL; struct ibv_device *ibv_dev; int err = 0; struct ibv_context *attr_ctx = NULL; struct ibv_device_attr_ex device_attr; - unsigned int sriov; unsigned int mps; unsigned int cqe_comp; unsigned int tunnel_en = 0; int idx; int i; - struct mlx5dv_context attrs_out; + struct mlx5dv_context attrs_out = {0}; #ifdef HAVE_IBV_DEVICE_COUNTERS_SET_SUPPORT struct ibv_counter_set_description cs_desc; #endif - (void)pci_drv; assert(pci_drv == &mlx5_driver); /* Get mlx5_dev[] index. */ idx = mlx5_dev_idx(&pci_dev->addr); if (idx == -1) { - ERROR("this driver cannot support any more adapters"); - return -ENOMEM; + DRV_LOG(ERR, "this driver cannot support any more adapters"); + err = ENOMEM; + goto error; } - DEBUG("using driver device index %d", idx); - + DRV_LOG(DEBUG, "using driver device index %d", idx); /* Save PCI address. */ mlx5_dev[idx].pci_addr = pci_dev->addr; - list = ibv_get_device_list(&i); + list = mlx5_glue->get_device_list(&i); if (list == NULL) { assert(errno); + err = errno; if (errno == ENOSYS) - ERROR("cannot list devices, is ib_uverbs loaded?"); - return -errno; + DRV_LOG(ERR, + "cannot list devices, is ib_uverbs loaded?"); + goto error; } assert(i >= 0); /* @@ -641,7 +636,7 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev) struct rte_pci_addr pci_addr; --i; - DEBUG("checking device \"%s\"", list[i]->name); + DRV_LOG(DEBUG, "checking device \"%s\"", list[i]->name); if (mlx5_ibv_device_to_pci_addr(list[i], &pci_addr)) continue; if ((pci_dev->addr.domain != pci_addr.domain) || @@ -649,67 +644,47 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev) (pci_dev->addr.devid != pci_addr.devid) || (pci_dev->addr.function != pci_addr.function)) continue; - sriov = ((pci_dev->id.device_id == - PCI_DEVICE_ID_MELLANOX_CONNECTX4VF) || - (pci_dev->id.device_id == - PCI_DEVICE_ID_MELLANOX_CONNECTX4LXVF) || - (pci_dev->id.device_id == - PCI_DEVICE_ID_MELLANOX_CONNECTX5VF) || - (pci_dev->id.device_id == - PCI_DEVICE_ID_MELLANOX_CONNECTX5EXVF)); - switch (pci_dev->id.device_id) { - case PCI_DEVICE_ID_MELLANOX_CONNECTX4: - tunnel_en = 1; - break; - case PCI_DEVICE_ID_MELLANOX_CONNECTX4LX: - case PCI_DEVICE_ID_MELLANOX_CONNECTX5: - case PCI_DEVICE_ID_MELLANOX_CONNECTX5VF: - case PCI_DEVICE_ID_MELLANOX_CONNECTX5EX: - case PCI_DEVICE_ID_MELLANOX_CONNECTX5EXVF: - tunnel_en = 1; - break; - default: - break; - } - INFO("PCI information matches, using device \"%s\"" - " (SR-IOV: %s)", - list[i]->name, - sriov ? "true" : "false"); - attr_ctx = ibv_open_device(list[i]); - err = errno; + DRV_LOG(INFO, "PCI information matches, using device \"%s\"", + list[i]->name); + attr_ctx = mlx5_glue->open_device(list[i]); + rte_errno = errno; + err = rte_errno; break; } if (attr_ctx == NULL) { - ibv_free_device_list(list); + mlx5_glue->free_device_list(list); switch (err) { case 0: - ERROR("cannot access device, is mlx5_ib loaded?"); - return -ENODEV; + DRV_LOG(ERR, + "cannot access device, is mlx5_ib loaded?"); + err = ENODEV; + goto error; case EINVAL: - ERROR("cannot use device, are drivers up to date?"); - return -EINVAL; + DRV_LOG(ERR, + "cannot use device, are drivers up to date?"); + goto error; } - assert(err > 0); - return -err; } ibv_dev = list[i]; - - DEBUG("device opened"); + DRV_LOG(DEBUG, "device opened"); /* * Multi-packet send is supported by ConnectX-4 Lx PF as well * as all ConnectX-5 devices. */ - mlx5dv_query_device(attr_ctx, &attrs_out); +#ifdef HAVE_IBV_DEVICE_TUNNEL_SUPPORT + attrs_out.comp_mask |= MLX5DV_CONTEXT_MASK_TUNNEL_OFFLOADS; +#endif + mlx5_glue->dv_query_device(attr_ctx, &attrs_out); if (attrs_out.flags & MLX5DV_CONTEXT_FLAGS_MPW_ALLOWED) { if (attrs_out.flags & MLX5DV_CONTEXT_FLAGS_ENHANCED_MPW) { - DEBUG("Enhanced MPW is supported"); + DRV_LOG(DEBUG, "enhanced MPW is supported"); mps = MLX5_MPW_ENHANCED; } else { - DEBUG("MPW is supported"); + DRV_LOG(DEBUG, "MPW is supported"); mps = MLX5_MPW; } } else { - DEBUG("MPW isn't supported"); + DRV_LOG(DEBUG, "MPW isn't supported"); mps = MLX5_MPW_DISABLED; } if (RTE_CACHE_LINE_SIZE == 128 && @@ -717,10 +692,25 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev) cqe_comp = 0; else cqe_comp = 1; - if (ibv_query_device_ex(attr_ctx, NULL, &device_attr)) +#ifdef HAVE_IBV_DEVICE_TUNNEL_SUPPORT + if (attrs_out.comp_mask & MLX5DV_CONTEXT_MASK_TUNNEL_OFFLOADS) { + tunnel_en = ((attrs_out.tunnel_offloads_caps & + MLX5DV_RAW_PACKET_CAP_TUNNELED_OFFLOAD_VXLAN) && + (attrs_out.tunnel_offloads_caps & + MLX5DV_RAW_PACKET_CAP_TUNNELED_OFFLOAD_GRE)); + } + DRV_LOG(DEBUG, "tunnel offloading is %ssupported", + tunnel_en ? "" : "not "); +#else + DRV_LOG(WARNING, + "tunnel offloading disabled due to old OFED/rdma-core version"); +#endif + if (mlx5_glue->query_device_ex(attr_ctx, NULL, &device_attr)) { + err = errno; goto error; - INFO("%u port(s) detected", device_attr.orig_attr.phys_port_cnt); - + } + DRV_LOG(INFO, "%u port(s) detected", + device_attr.orig_attr.phys_port_cnt); for (i = 0; i < device_attr.orig_attr.phys_port_cnt; i++) { char name[RTE_ETH_NAME_MAX_LEN]; int len; @@ -730,11 +720,9 @@ mlx5_pci_probe(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; struct ibv_device_attr_ex device_attr_ex; struct ether_addr mac; - uint16_t num_vfs = 0; - struct ibv_device_attr_ex device_attr; struct mlx5_dev_config config = { .cqe_comp = cqe_comp, .mps = mps, @@ -752,32 +740,26 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev) pci_dev->addr.devid, pci_dev->addr.function); if (device_attr.orig_attr.phys_port_cnt > 1) snprintf(name + len, sizeof(name), " port %u", i); - mlx5_dev[idx].ports |= test; - if (rte_eal_process_type() == RTE_PROC_SECONDARY) { eth_dev = rte_eth_dev_attach_secondary(name); if (eth_dev == NULL) { - ERROR("can not attach rte ethdev"); - err = ENOMEM; + DRV_LOG(ERR, "can not attach rte ethdev"); + rte_errno = ENOMEM; + err = rte_errno; goto error; } eth_dev->device = &pci_dev->device; eth_dev->dev_ops = &mlx5_dev_sec_ops; - priv = eth_dev->data->dev_private; - err = priv_uar_init_secondary(priv); - if (err < 0) { - err = -err; + err = mlx5_uar_init_secondary(eth_dev); + if (err) goto error; - } /* Receive command fd from primary process */ - err = priv_socket_connect(priv); - if (err < 0) { - err = -err; + err = mlx5_socket_connect(eth_dev); + if (err) goto error; - } /* Remap UAR for Tx queues. */ - err = priv_tx_uar_remap(priv, err); + err = mlx5_tx_uar_remap(eth_dev, err); if (err) goto error; /* @@ -786,60 +768,52 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev) * secondary process. */ eth_dev->rx_pkt_burst = - priv_select_rx_function(priv, eth_dev); + mlx5_select_rx_function(eth_dev); eth_dev->tx_pkt_burst = - priv_select_tx_function(priv, eth_dev); + mlx5_select_tx_function(eth_dev); continue; } - - DEBUG("using port %u (%08" PRIx32 ")", port, test); - - ctx = ibv_open_device(ibv_dev); + DRV_LOG(DEBUG, "using port %u (%08" PRIx32 ")", port, test); + ctx = mlx5_glue->open_device(ibv_dev); if (ctx == NULL) { err = ENODEV; goto port_error; } - - ibv_query_device_ex(ctx, NULL, &device_attr); /* Check port status. */ - err = ibv_query_port(ctx, port, &port_attr); + err = mlx5_glue->query_port(ctx, port, &port_attr); if (err) { - ERROR("port query failed: %s", strerror(err)); + DRV_LOG(ERR, "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); + DRV_LOG(ERR, + "port %d is not configured in Ethernet mode", + port); err = EINVAL; goto port_error; } - if (port_attr.state != IBV_PORT_ACTIVE) - DEBUG("port %d is not active: \"%s\" (%d)", - port, ibv_port_state_str(port_attr.state), - port_attr.state); - + DRV_LOG(DEBUG, "port %d is not active: \"%s\" (%d)", + port, + mlx5_glue->port_state_str(port_attr.state), + port_attr.state); /* Allocate protection domain. */ - pd = ibv_alloc_pd(ctx); + pd = mlx5_glue->alloc_pd(ctx); if (pd == NULL) { - ERROR("PD allocation failure"); + DRV_LOG(ERR, "PD allocation failure"); err = ENOMEM; goto port_error; } - mlx5_dev[idx].ports |= test; - /* from rte_ethdev.c */ priv = rte_zmalloc("ethdev private structure", sizeof(*priv), RTE_CACHE_LINE_SIZE); if (priv == NULL) { - ERROR("priv allocation failure"); + DRV_LOG(ERR, "priv allocation failure"); err = ENOMEM; goto port_error; } - priv->ctx = ctx; strncpy(priv->ibdev_path, priv->ctx->device->ibdev_path, sizeof(priv->ibdev_path)); @@ -849,34 +823,26 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev) priv->mtu = ETHER_MTU; err = mlx5_args(&config, pci_dev->device.devargs); if (err) { - ERROR("failed to process device arguments: %s", - strerror(err)); + DRV_LOG(ERR, "failed to process device arguments: %s", + strerror(err)); goto port_error; } - if (ibv_query_device_ex(ctx, NULL, &device_attr_ex)) { - ERROR("ibv_query_device_ex() failed"); + if (mlx5_glue->query_device_ex(ctx, NULL, &device_attr_ex)) { + DRV_LOG(ERR, "ibv_query_device_ex() failed"); + err = errno; goto port_error; } - config.hw_csum = !!(device_attr_ex.device_cap_flags_ex & IBV_DEVICE_RAW_IP_CSUM); - DEBUG("checksum offloading is %ssupported", - (config.hw_csum ? "" : "not ")); - -#ifdef HAVE_IBV_DEVICE_VXLAN_SUPPORT - config.hw_csum_l2tun = - !!(exp_device_attr.exp_device_cap_flags & - IBV_DEVICE_VXLAN_SUPPORT); -#endif - DEBUG("Rx L2 tunnel checksum offloads are %ssupported", - (config.hw_csum_l2tun ? "" : "not ")); - + DRV_LOG(DEBUG, "checksum offloading is %ssupported", + (config.hw_csum ? "" : "not ")); #ifdef HAVE_IBV_DEVICE_COUNTERS_SET_SUPPORT config.flow_counter_en = !!(device_attr.max_counter_sets); - ibv_describe_counter_set(ctx, 0, &cs_desc); - DEBUG("counter type = %d, num of cs = %ld, attributes = %d", - cs_desc.counter_type, cs_desc.num_of_cs, - cs_desc.attributes); + mlx5_glue->describe_counter_set(ctx, 0, &cs_desc); + DRV_LOG(DEBUG, + "counter type = %d, num of cs = %ld, attributes = %d", + cs_desc.counter_type, cs_desc.num_of_cs, + cs_desc.attributes); #endif config.ind_table_max_size = device_attr_ex.rss_caps.max_rwq_indirection_table_size; @@ -885,27 +851,24 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev) if (config.ind_table_max_size > (unsigned int)ETH_RSS_RETA_SIZE_512) config.ind_table_max_size = ETH_RSS_RETA_SIZE_512; - DEBUG("maximum RX indirection table size is %u", - config.ind_table_max_size); + DRV_LOG(DEBUG, "maximum Rx indirection table size is %u", + config.ind_table_max_size); config.hw_vlan_strip = !!(device_attr_ex.raw_packet_caps & IBV_RAW_PACKET_CAP_CVLAN_STRIPPING); - DEBUG("VLAN stripping is %ssupported", - (config.hw_vlan_strip ? "" : "not ")); + DRV_LOG(DEBUG, "VLAN stripping is %ssupported", + (config.hw_vlan_strip ? "" : "not ")); - config.hw_fcs_strip = - !!(device_attr_ex.orig_attr.device_cap_flags & - IBV_WQ_FLAGS_SCATTER_FCS); - DEBUG("FCS stripping configuration is %ssupported", - (config.hw_fcs_strip ? "" : "not ")); + config.hw_fcs_strip = !!(device_attr_ex.raw_packet_caps & + IBV_RAW_PACKET_CAP_SCATTER_FCS); + DRV_LOG(DEBUG, "FCS stripping configuration is %ssupported", + (config.hw_fcs_strip ? "" : "not ")); #ifdef HAVE_IBV_WQ_FLAG_RX_END_PADDING config.hw_padding = !!device_attr_ex.rx_pad_end_addr_align; #endif - DEBUG("hardware RX end alignment padding is %ssupported", - (config.hw_padding ? "" : "not ")); - - priv_get_num_vfs(priv, &num_vfs); - config.sriov = (num_vfs || sriov); + DRV_LOG(DEBUG, + "hardware Rx end alignment padding is %ssupported", + (config.hw_padding ? "" : "not ")); config.tso = ((device_attr_ex.tso_caps.max_tso > 0) && (device_attr_ex.tso_caps.supported_qpts & (1 << IBV_QPT_RAW_PACKET))); @@ -913,117 +876,124 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev) config.tso_max_payload_sz = device_attr_ex.tso_caps.max_tso; if (config.mps && !mps) { - ERROR("multi-packet send not supported on this device" - " (" MLX5_TXQ_MPW_EN ")"); + DRV_LOG(ERR, + "multi-packet send not supported on this device" + " (" MLX5_TXQ_MPW_EN ")"); err = ENOTSUP; goto port_error; } - INFO("%sMPS is %s", - config.mps == MLX5_MPW_ENHANCED ? "Enhanced " : "", - config.mps != MLX5_MPW_DISABLED ? "enabled" : "disabled"); + DRV_LOG(INFO, "%s MPS is %s", + config.mps == MLX5_MPW_ENHANCED ? "enhanced " : "", + config.mps != MLX5_MPW_DISABLED ? "enabled" : + "disabled"); if (config.cqe_comp && !cqe_comp) { - WARN("Rx CQE compression isn't supported"); + DRV_LOG(WARNING, "Rx CQE compression isn't supported"); config.cqe_comp = 0; } - err = priv_uar_init_primary(priv); + eth_dev = rte_eth_dev_allocate(name); + if (eth_dev == NULL) { + DRV_LOG(ERR, "can not allocate rte ethdev"); + err = ENOMEM; + goto port_error; + } + eth_dev->data->dev_private = priv; + priv->dev = eth_dev; + eth_dev->data->mac_addrs = priv->mac; + eth_dev->device = &pci_dev->device; + rte_eth_copy_pci_info(eth_dev, pci_dev); + eth_dev->device->driver = &mlx5_driver.driver; + err = mlx5_uar_init_primary(eth_dev); if (err) goto port_error; /* Configure the first MAC address by default. */ - if (priv_get_mac(priv, &mac.addr_bytes)) { - ERROR("cannot get MAC address, is mlx5_en loaded?" - " (errno: %s)", strerror(errno)); + if (mlx5_get_mac(eth_dev, &mac.addr_bytes)) { + DRV_LOG(ERR, + "port %u cannot get MAC address, is mlx5_en" + " loaded? (errno: %s)", + eth_dev->data->port_id, strerror(errno)); err = ENODEV; goto port_error; } - INFO("port %u MAC address is %02x:%02x:%02x:%02x:%02x:%02x", - priv->port, - mac.addr_bytes[0], mac.addr_bytes[1], - mac.addr_bytes[2], mac.addr_bytes[3], - mac.addr_bytes[4], mac.addr_bytes[5]); + DRV_LOG(INFO, + "port %u MAC address is %02x:%02x:%02x:%02x:%02x:%02x", + eth_dev->data->port_id, + mac.addr_bytes[0], mac.addr_bytes[1], + mac.addr_bytes[2], mac.addr_bytes[3], + mac.addr_bytes[4], mac.addr_bytes[5]); #ifndef NDEBUG { char ifname[IF_NAMESIZE]; - if (priv_get_ifname(priv, &ifname) == 0) - DEBUG("port %u ifname is \"%s\"", - priv->port, ifname); + if (mlx5_get_ifname(eth_dev, &ifname) == 0) + DRV_LOG(DEBUG, "port %u ifname is \"%s\"", + eth_dev->data->port_id, ifname); else - DEBUG("port %u ifname is unknown", priv->port); + DRV_LOG(DEBUG, "port %u ifname is unknown", + eth_dev->data->port_id); } #endif /* Get actual MTU if possible. */ - priv_get_mtu(priv, &priv->mtu); - DEBUG("port %u MTU is %u", priv->port, priv->mtu); - - eth_dev = rte_eth_dev_allocate(name); - if (eth_dev == NULL) { - ERROR("can not allocate rte ethdev"); - err = ENOMEM; + err = mlx5_get_mtu(eth_dev, &priv->mtu); + if (err) goto port_error; - } - eth_dev->data->dev_private = priv; - eth_dev->data->mac_addrs = priv->mac; - eth_dev->device = &pci_dev->device; - rte_eth_copy_pci_info(eth_dev, pci_dev); - eth_dev->device->driver = &mlx5_driver.driver; + DRV_LOG(DEBUG, "port %u MTU is %u", eth_dev->data->port_id, + priv->mtu); /* * Initialize burst functions to prevent crashes before link-up. */ eth_dev->rx_pkt_burst = removed_rx_burst; eth_dev->tx_pkt_burst = removed_tx_burst; - priv->dev = eth_dev; eth_dev->dev_ops = &mlx5_dev_ops; /* Register MAC address. */ claim_zero(mlx5_mac_addr_add(eth_dev, &mac, 0, 0)); TAILQ_INIT(&priv->flows); TAILQ_INIT(&priv->ctrl_flows); - /* Hint libmlx5 to use PMD allocator for data plane resources */ struct mlx5dv_ctx_allocators alctr = { .alloc = &mlx5_alloc_verbs_buf, .free = &mlx5_free_verbs_buf, .data = priv, }; - mlx5dv_set_context_attr(ctx, MLX5DV_CTX_ATTR_BUF_ALLOCATORS, - (void *)((uintptr_t)&alctr)); - + mlx5_glue->dv_set_context_attr(ctx, + MLX5DV_CTX_ATTR_BUF_ALLOCATORS, + (void *)((uintptr_t)&alctr)); /* Bring Ethernet device up. */ - DEBUG("forcing Ethernet interface up"); - priv_set_flags(priv, ~IFF_UP, IFF_UP); + DRV_LOG(DEBUG, "port %u forcing Ethernet interface up", + eth_dev->data->port_id); + mlx5_set_link_up(eth_dev); /* Store device configuration on private structure. */ priv->config = config; continue; - port_error: if (priv) rte_free(priv); if (pd) - claim_zero(ibv_dealloc_pd(pd)); + claim_zero(mlx5_glue->dealloc_pd(pd)); if (ctx) - claim_zero(ibv_close_device(ctx)); + claim_zero(mlx5_glue->close_device(ctx)); break; } - /* * XXX if something went wrong in the loop above, there is a resource * leak (ctx, pd, priv, dpdk ethdev) but we can do nothing about it as * long as the dpdk does not provide a way to deallocate a ethdev and a * way to enumerate the registered ethdevs to free the previous ones. */ - /* no port found, complain */ if (!mlx5_dev[idx].ports) { - err = ENODEV; - goto error; + rte_errno = ENODEV; + err = rte_errno; } - error: if (attr_ctx) - claim_zero(ibv_close_device(attr_ctx)); + claim_zero(mlx5_glue->close_device(attr_ctx)); if (list) - ibv_free_device_list(list); - assert(err >= 0); - return -err; + mlx5_glue->free_device_list(list); + if (err) { + rte_errno = err; + return -rte_errno; + } + return 0; } static const struct rte_pci_id mlx5_pci_id_map[] = { @@ -1073,6 +1043,138 @@ static struct rte_pci_driver mlx5_driver = { .drv_flags = RTE_PCI_DRV_INTR_LSC | RTE_PCI_DRV_INTR_RMV, }; +#ifdef RTE_LIBRTE_MLX5_DLOPEN_DEPS + +/** + * Suffix RTE_EAL_PMD_PATH with "-glue". + * + * This function performs a sanity check on RTE_EAL_PMD_PATH before + * suffixing its last component. + * + * @param buf[out] + * Output buffer, should be large enough otherwise NULL is returned. + * @param size + * Size of @p out. + * + * @return + * Pointer to @p buf or @p NULL in case suffix cannot be appended. + */ +static char * +mlx5_glue_path(char *buf, size_t size) +{ + static const char *const bad[] = { "/", ".", "..", NULL }; + const char *path = RTE_EAL_PMD_PATH; + size_t len = strlen(path); + size_t off; + int i; + + while (len && path[len - 1] == '/') + --len; + for (off = len; off && path[off - 1] != '/'; --off) + ; + for (i = 0; bad[i]; ++i) + if (!strncmp(path + off, bad[i], (int)(len - off))) + goto error; + i = snprintf(buf, size, "%.*s-glue", (int)len, path); + if (i == -1 || (size_t)i >= size) + goto error; + return buf; +error: + DRV_LOG(ERR, + "unable to append \"-glue\" to last component of" + " RTE_EAL_PMD_PATH (\"" RTE_EAL_PMD_PATH "\")," + " please re-configure DPDK"); + return NULL; +} + +/** + * Initialization routine for run-time dependency on rdma-core. + */ +static int +mlx5_glue_init(void) +{ + char glue_path[sizeof(RTE_EAL_PMD_PATH) - 1 + sizeof("-glue")]; + const char *path[] = { + /* + * A basic security check is necessary before trusting + * MLX5_GLUE_PATH, which may override RTE_EAL_PMD_PATH. + */ + (geteuid() == getuid() && getegid() == getgid() ? + getenv("MLX5_GLUE_PATH") : NULL), + /* + * When RTE_EAL_PMD_PATH is set, use its glue-suffixed + * variant, otherwise let dlopen() look up libraries on its + * own. + */ + (*RTE_EAL_PMD_PATH ? + mlx5_glue_path(glue_path, sizeof(glue_path)) : ""), + }; + unsigned int i = 0; + void *handle = NULL; + void **sym; + const char *dlmsg; + + while (!handle && i != RTE_DIM(path)) { + const char *end; + size_t len; + int ret; + + if (!path[i]) { + ++i; + continue; + } + end = strpbrk(path[i], ":;"); + if (!end) + end = path[i] + strlen(path[i]); + len = end - path[i]; + ret = 0; + do { + char name[ret + 1]; + + ret = snprintf(name, sizeof(name), "%.*s%s" MLX5_GLUE, + (int)len, path[i], + (!len || *(end - 1) == '/') ? "" : "/"); + if (ret == -1) + break; + if (sizeof(name) != (size_t)ret + 1) + continue; + DRV_LOG(DEBUG, "looking for rdma-core glue as \"%s\"", + name); + handle = dlopen(name, RTLD_LAZY); + break; + } while (1); + path[i] = end + 1; + if (!*end) + ++i; + } + if (!handle) { + rte_errno = EINVAL; + dlmsg = dlerror(); + if (dlmsg) + DRV_LOG(WARNING, "cannot load glue library: %s", dlmsg); + goto glue_error; + } + sym = dlsym(handle, "mlx5_glue"); + if (!sym || !*sym) { + rte_errno = EINVAL; + dlmsg = dlerror(); + if (dlmsg) + DRV_LOG(ERR, "cannot resolve glue symbol: %s", dlmsg); + goto glue_error; + } + mlx5_glue = *sym; + return 0; +glue_error: + if (handle) + dlclose(handle); + DRV_LOG(WARNING, + "cannot initialize PMD due to missing run-time dependency on" + " rdma-core libraries (libibverbs, libmlx5)"); + return -rte_errno; +} + +#endif + /** * Driver initialization routine. */ @@ -1092,10 +1194,38 @@ rte_mlx5_pmd_init(void) /* Match the size of Rx completion entry to the size of a cacheline. */ if (RTE_CACHE_LINE_SIZE == 128) setenv("MLX5_CQE_SIZE", "128", 0); - ibv_fork_init(); +#ifdef RTE_LIBRTE_MLX5_DLOPEN_DEPS + if (mlx5_glue_init()) + return; + assert(mlx5_glue); +#endif +#ifndef NDEBUG + /* Glue structure must not contain any NULL pointers. */ + { + unsigned int i; + + for (i = 0; i != sizeof(*mlx5_glue) / sizeof(void *); ++i) + assert(((const void *const *)mlx5_glue)[i]); + } +#endif + if (strcmp(mlx5_glue->version, MLX5_GLUE_VERSION)) { + DRV_LOG(ERR, + "rdma-core glue \"%s\" mismatch: \"%s\" is required", + mlx5_glue->version, MLX5_GLUE_VERSION); + return; + } + mlx5_glue->fork_init(); rte_pci_register(&mlx5_driver); } RTE_PMD_EXPORT_NAME(net_mlx5, __COUNTER__); RTE_PMD_REGISTER_PCI_TABLE(net_mlx5, mlx5_pci_id_map); RTE_PMD_REGISTER_KMOD_DEP(net_mlx5, "* ib_uverbs & mlx5_core & mlx5_ib"); + +/** Initialize driver log type. */ +RTE_INIT(vdev_netvsc_init_log) +{ + mlx5_logtype = rte_log_register("pmd.net.mlx5"); + if (mlx5_logtype >= 0) + rte_log_set_level(mlx5_logtype, RTE_LOG_NOTICE); +}