X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fenic%2Fenic_main.c;h=77eb34ce459ded8a681e38945d8db34c31aff7b7;hb=0f872d31295df39cedf9edfe134f48f3fdcefcbf;hp=e7117c1e7079c025eececd6172d2cdf7d22b5cf7;hpb=e8a71c46055a81f7d0b43515ebaf4763a6b419b9;p=dpdk.git diff --git a/drivers/net/enic/enic_main.c b/drivers/net/enic/enic_main.c index e7117c1e70..77eb34ce45 100644 --- a/drivers/net/enic/enic_main.c +++ b/drivers/net/enic/enic_main.c @@ -1,35 +1,6 @@ -/* - * Copyright 2008-2014 Cisco Systems, Inc. All rights reserved. +/* SPDX-License-Identifier: BSD-3-Clause + * Copyright 2008-2017 Cisco Systems, Inc. All rights reserved. * Copyright 2007 Nuova Systems, Inc. All rights reserved. - * - * Copyright (c) 2014, Cisco Systems, Inc. - * All rights reserved. - * - * 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 HOLDER 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 @@ -40,11 +11,12 @@ #include #include +#include #include #include #include #include -#include +#include #include "enic_compat.h" #include "enic.h" @@ -97,11 +69,6 @@ enic_rxmbuf_queue_release(__rte_unused struct enic *enic, struct vnic_rq *rq) } } -void enic_set_hdr_split_size(struct enic *enic, u16 split_hdr_size) -{ - vnic_set_hdr_split_size(enic->vdev, split_hdr_size); -} - static void enic_free_wq_buf(struct vnic_wq_buf *buf) { struct rte_mbuf *mbuf = (struct rte_mbuf *)buf->mb; @@ -137,6 +104,7 @@ static void enic_clear_soft_stats(struct enic *enic) struct enic_soft_stats *soft_stats = &enic->soft_stats; rte_atomic64_clear(&soft_stats->rx_nombuf); rte_atomic64_clear(&soft_stats->rx_packet_errors); + rte_atomic64_clear(&soft_stats->tx_oversized); } static void enic_init_soft_stats(struct enic *enic) @@ -144,6 +112,7 @@ static void enic_init_soft_stats(struct enic *enic) struct enic_soft_stats *soft_stats = &enic->soft_stats; rte_atomic64_init(&soft_stats->rx_nombuf); rte_atomic64_init(&soft_stats->rx_packet_errors); + rte_atomic64_init(&soft_stats->tx_oversized); enic_clear_soft_stats(enic); } @@ -154,16 +123,17 @@ void enic_dev_stats_clear(struct enic *enic) enic_clear_soft_stats(enic); } -void enic_dev_stats_get(struct enic *enic, struct rte_eth_stats *r_stats) +int enic_dev_stats_get(struct enic *enic, struct rte_eth_stats *r_stats) { struct vnic_stats *stats; struct enic_soft_stats *soft_stats = &enic->soft_stats; int64_t rx_truncated; uint64_t rx_packet_errors; + int ret = vnic_dev_stats_dump(enic->vdev, &stats); - if (vnic_dev_stats_dump(enic->vdev, &stats)) { + if (ret) { dev_err(enic, "Error in getting stats\n"); - return; + return ret; } /* The number of truncated packets can only be calculated by @@ -174,8 +144,7 @@ void enic_dev_stats_get(struct enic *enic, struct rte_eth_stats *r_stats) * which can make ibytes be slightly higher than it should be. */ rx_packet_errors = rte_atomic64_read(&soft_stats->rx_packet_errors); - rx_truncated = rx_packet_errors - stats->rx.rx_errors - - stats->rx.rx_no_bufs; + rx_truncated = rx_packet_errors - stats->rx.rx_errors; r_stats->ipackets = stats->rx.rx_frames_ok - rx_truncated; r_stats->opackets = stats->tx.tx_frames_ok; @@ -184,42 +153,37 @@ void enic_dev_stats_get(struct enic *enic, struct rte_eth_stats *r_stats) r_stats->obytes = stats->tx.tx_bytes_ok; r_stats->ierrors = stats->rx.rx_errors + stats->rx.rx_drop; - r_stats->oerrors = stats->tx.tx_errors; + r_stats->oerrors = stats->tx.tx_errors + + rte_atomic64_read(&soft_stats->tx_oversized); r_stats->imissed = stats->rx.rx_no_bufs + rx_truncated; r_stats->rx_nombuf = rte_atomic64_read(&soft_stats->rx_nombuf); + return 0; } -void enic_del_mac_address(struct enic *enic) +void enic_del_mac_address(struct enic *enic, int mac_index) { - if (vnic_dev_del_addr(enic->vdev, enic->mac_addr)) + struct rte_eth_dev *eth_dev = enic->rte_dev; + uint8_t *mac_addr = eth_dev->data->mac_addrs[mac_index].addr_bytes; + + if (vnic_dev_del_addr(enic->vdev, mac_addr)) dev_err(enic, "del mac addr failed\n"); } -void enic_set_mac_address(struct enic *enic, uint8_t *mac_addr) +int enic_set_mac_address(struct enic *enic, uint8_t *mac_addr) { int err; if (!is_eth_addr_valid(mac_addr)) { dev_err(enic, "invalid mac address\n"); - return; - } - - err = vnic_dev_del_addr(enic->vdev, enic->mac_addr); - if (err) { - dev_err(enic, "del mac addr failed\n"); - return; + return -EINVAL; } - ether_addr_copy((struct ether_addr *)mac_addr, - (struct ether_addr *)enic->mac_addr); - err = vnic_dev_add_addr(enic->vdev, mac_addr); - if (err) { + if (err) dev_err(enic, "add mac addr failed\n"); - return; - } + return err; } static void @@ -229,26 +193,32 @@ enic_free_rq_buf(struct rte_mbuf **mbuf) return; rte_pktmbuf_free(*mbuf); - mbuf = NULL; + *mbuf = NULL; } void enic_init_vnic_resources(struct enic *enic) { unsigned int error_interrupt_enable = 1; unsigned int error_interrupt_offset = 0; + unsigned int rxq_interrupt_enable = 0; + unsigned int rxq_interrupt_offset; unsigned int index = 0; unsigned int cq_idx; struct vnic_rq *data_rq; + if (enic->rte_dev->data->dev_conf.intr_conf.rxq) { + rxq_interrupt_enable = 1; + rxq_interrupt_offset = ENICPMD_RXQ_INTR_OFFSET; + } for (index = 0; index < enic->rq_count; index++) { - cq_idx = enic_cq_rq(enic, enic_sop_rq(index)); + cq_idx = enic_cq_rq(enic, enic_rte_rq_idx_to_sop_idx(index)); - vnic_rq_init(&enic->rq[enic_sop_rq(index)], + vnic_rq_init(&enic->rq[enic_rte_rq_idx_to_sop_idx(index)], cq_idx, error_interrupt_enable, error_interrupt_offset); - data_rq = &enic->rq[enic_data_rq(index)]; + data_rq = &enic->rq[enic_rte_rq_idx_to_data_idx(index)]; if (data_rq->in_use) vnic_rq_init(data_rq, cq_idx, @@ -261,11 +231,13 @@ void enic_init_vnic_resources(struct enic *enic) 0 /* cq_head */, 0 /* cq_tail */, 1 /* cq_tail_color */, - 0 /* interrupt_enable */, + rxq_interrupt_enable, 1 /* cq_entry_enable */, 0 /* cq_message_enable */, - 0 /* interrupt offset */, + rxq_interrupt_offset, 0 /* cq_message_addr */); + if (rxq_interrupt_enable) + rxq_interrupt_offset++; } for (index = 0; index < enic->wq_count; index++) { @@ -285,13 +257,15 @@ void enic_init_vnic_resources(struct enic *enic) 0 /* cq_entry_enable */, 1 /* cq_message_enable */, 0 /* interrupt offset */, - (u64)enic->wq[index].cqmsg_rz->phys_addr); + (u64)enic->wq[index].cqmsg_rz->iova); } - vnic_intr_init(&enic->intr, - enic->config.intr_timer_usec, - enic->config.intr_timer_type, - /*mask_on_assertion*/1); + for (index = 0; index < enic->intr_count; index++) { + vnic_intr_init(&enic->intr[index], + enic->config.intr_timer_usec, + enic->config.intr_timer_type, + /*mask_on_assertion*/1); + } } @@ -302,6 +276,8 @@ enic_alloc_rx_queue_mbufs(struct enic *enic, struct vnic_rq *rq) struct rq_enet_desc *rqd = rq->ring.descs; unsigned i; dma_addr_t dma_addr; + uint32_t max_rx_pkt_len; + uint16_t rq_buf_len; if (!rq->in_use) return 0; @@ -309,6 +285,18 @@ enic_alloc_rx_queue_mbufs(struct enic *enic, struct vnic_rq *rq) dev_debug(enic, "queue %u, allocating %u rx queue mbufs\n", rq->index, rq->ring.desc_count); + /* + * If *not* using scatter and the mbuf size is smaller than the + * requested max packet size (max_rx_pkt_len), then reduce the + * posted buffer size to max_rx_pkt_len. HW still receives packets + * larger than max_rx_pkt_len, but they will be truncated, which we + * drop in the rx handler. Not ideal, but better than returning + * large packets when the user is not expecting them. + */ + max_rx_pkt_len = enic->rte_dev->data->dev_conf.rxmode.max_rx_pkt_len; + rq_buf_len = rte_pktmbuf_data_room_size(rq->mp) - RTE_PKTMBUF_HEADROOM; + if (max_rx_pkt_len < rq_buf_len && !rq->data_queue_enable) + rq_buf_len = max_rx_pkt_len; for (i = 0; i < rq->ring.desc_count; i++, rqd++) { mb = rte_mbuf_raw_alloc(rq->mp); if (mb == NULL) { @@ -318,12 +306,12 @@ enic_alloc_rx_queue_mbufs(struct enic *enic, struct vnic_rq *rq) } mb->data_off = RTE_PKTMBUF_HEADROOM; - dma_addr = (dma_addr_t)(mb->buf_physaddr + dma_addr = (dma_addr_t)(mb->buf_iova + RTE_PKTMBUF_HEADROOM); rq_enet_desc_enc(rqd, dma_addr, (rq->is_sop ? RQ_ENET_TYPE_ONLY_SOP : RQ_ENET_TYPE_NOT_SOP), - mb->buf_len - RTE_PKTMBUF_HEADROOM); + rq_buf_len); rq->mbuf_ring[i] = mb; } @@ -364,7 +352,7 @@ enic_alloc_consistent(void *priv, size_t size, } vaddr = rz->addr; - *dma_handle = (dma_addr_t)rz->phys_addr; + *dma_handle = (dma_addr_t)rz->iova; mze = rte_malloc("enic memzone entry", sizeof(struct enic_memzone_entry), 0); @@ -373,6 +361,7 @@ enic_alloc_consistent(void *priv, size_t size, pr_err("%s : Failed to allocate memory for memzone list\n", __func__); rte_memzone_free(rz); + return NULL; } mze->rz = rz; @@ -396,7 +385,7 @@ enic_free_consistent(void *priv, rte_spinlock_lock(&enic->memzone_list_lock); LIST_FOREACH(mze, &enic->memzone_list, entries) { if (mze->rz->addr == vaddr && - mze->rz->phys_addr == dma_handle) + mze->rz->iova == dma_handle) break; } if (mze == NULL) { @@ -427,19 +416,67 @@ int enic_link_update(struct enic *enic) } static void -enic_intr_handler(__rte_unused struct rte_intr_handle *handle, - void *arg) +enic_intr_handler(void *arg) { struct rte_eth_dev *dev = (struct rte_eth_dev *)arg; struct enic *enic = pmd_priv(dev); - vnic_intr_return_all_credits(&enic->intr); + vnic_intr_return_all_credits(&enic->intr[ENICPMD_LSC_INTR_OFFSET]); enic_link_update(enic); _rte_eth_dev_callback_process(dev, RTE_ETH_EVENT_INTR_LSC, NULL); enic_log_q_error(enic); } +static int enic_rxq_intr_init(struct enic *enic) +{ + struct rte_intr_handle *intr_handle; + uint32_t rxq_intr_count, i; + int err; + + intr_handle = enic->rte_dev->intr_handle; + if (!enic->rte_dev->data->dev_conf.intr_conf.rxq) + return 0; + /* + * Rx queue interrupts only work when we have MSI-X interrupts, + * one per queue. Sharing one interrupt is technically + * possible with VIC, but it is not worth the complications it brings. + */ + if (!rte_intr_cap_multiple(intr_handle)) { + dev_err(enic, "Rx queue interrupts require MSI-X interrupts" + " (vfio-pci driver)\n"); + return -ENOTSUP; + } + rxq_intr_count = enic->intr_count - ENICPMD_RXQ_INTR_OFFSET; + err = rte_intr_efd_enable(intr_handle, rxq_intr_count); + if (err) { + dev_err(enic, "Failed to enable event fds for Rx queue" + " interrupts\n"); + return err; + } + intr_handle->intr_vec = rte_zmalloc("enic_intr_vec", + rxq_intr_count * sizeof(int), 0); + if (intr_handle->intr_vec == NULL) { + dev_err(enic, "Failed to allocate intr_vec\n"); + return -ENOMEM; + } + for (i = 0; i < rxq_intr_count; i++) + intr_handle->intr_vec[i] = i + ENICPMD_RXQ_INTR_OFFSET; + return 0; +} + +static void enic_rxq_intr_deinit(struct enic *enic) +{ + struct rte_intr_handle *intr_handle; + + intr_handle = enic->rte_dev->intr_handle; + rte_intr_efd_disable(intr_handle); + if (intr_handle->intr_vec != NULL) { + rte_free(intr_handle->intr_vec); + intr_handle->intr_vec = NULL; + } +} + int enic_enable(struct enic *enic) { unsigned int index; @@ -456,23 +493,26 @@ int enic_enable(struct enic *enic) if (eth_dev->data->dev_conf.intr_conf.lsc) vnic_dev_notify_set(enic->vdev, 0); + err = enic_rxq_intr_init(enic); + if (err) + return err; if (enic_clsf_init(enic)) dev_warning(enic, "Init of hash table for clsf failed."\ "Flow director feature will not work\n"); for (index = 0; index < enic->rq_count; index++) { err = enic_alloc_rx_queue_mbufs(enic, - &enic->rq[enic_sop_rq(index)]); + &enic->rq[enic_rte_rq_idx_to_sop_idx(index)]); if (err) { dev_err(enic, "Failed to alloc sop RX queue mbufs\n"); return err; } err = enic_alloc_rx_queue_mbufs(enic, - &enic->rq[enic_data_rq(index)]); + &enic->rq[enic_rte_rq_idx_to_data_idx(index)]); if (err) { /* release the allocated mbufs for the sop rq*/ enic_rxmbuf_queue_release(enic, - &enic->rq[enic_sop_rq(index)]); + &enic->rq[enic_rte_rq_idx_to_sop_idx(index)]); dev_err(enic, "Failed to alloc data RX queue mbufs\n"); return err; @@ -493,7 +533,8 @@ int enic_enable(struct enic *enic) enic_intr_handler, (void *)enic->rte_dev); rte_intr_enable(&(enic->pdev->intr_handle)); - vnic_intr_unmask(&enic->intr); + /* Unmask LSC interrupt */ + vnic_intr_unmask(&enic->intr[ENICPMD_LSC_INTR_OFFSET]); return 0; } @@ -501,17 +542,21 @@ int enic_enable(struct enic *enic) int enic_alloc_intr_resources(struct enic *enic) { int err; + unsigned int i; dev_info(enic, "vNIC resources used: "\ "wq %d rq %d cq %d intr %d\n", enic->wq_count, enic_vnic_rq_count(enic), enic->cq_count, enic->intr_count); - err = vnic_intr_alloc(enic->vdev, &enic->intr, 0); - if (err) - enic_free_vnic_resources(enic); - - return err; + for (i = 0; i < enic->intr_count; i++) { + err = vnic_intr_alloc(enic->vdev, &enic->intr[i], i); + if (err) { + enic_free_vnic_resources(enic); + return err; + } + } + return 0; } void enic_free_rq(void *rxq) @@ -541,7 +586,7 @@ void enic_free_rq(void *rxq) if (rq_data->in_use) vnic_rq_free(rq_data); - vnic_cq_free(&enic->cq[rq_sop->index]); + vnic_cq_free(&enic->cq[enic_sop_rq_idx_to_cq_idx(rq_sop->index)]); rq_sop->in_use = 0; rq_data->in_use = 0; @@ -569,8 +614,10 @@ int enic_stop_wq(struct enic *enic, uint16_t queue_idx) void enic_start_rq(struct enic *enic, uint16_t queue_idx) { - struct vnic_rq *rq_sop = &enic->rq[enic_sop_rq(queue_idx)]; - struct vnic_rq *rq_data = &enic->rq[rq_sop->data_queue_idx]; + struct vnic_rq *rq_sop; + struct vnic_rq *rq_data; + rq_sop = &enic->rq[enic_rte_rq_idx_to_sop_idx(queue_idx)]; + rq_data = &enic->rq[rq_sop->data_queue_idx]; struct rte_eth_dev *eth_dev = enic->rte_dev; if (rq_data->in_use) @@ -584,8 +631,10 @@ int enic_stop_rq(struct enic *enic, uint16_t queue_idx) { int ret1 = 0, ret2 = 0; struct rte_eth_dev *eth_dev = enic->rte_dev; - struct vnic_rq *rq_sop = &enic->rq[enic_sop_rq(queue_idx)]; - struct vnic_rq *rq_data = &enic->rq[rq_sop->data_queue_idx]; + struct vnic_rq *rq_sop; + struct vnic_rq *rq_data; + rq_sop = &enic->rq[enic_rte_rq_idx_to_sop_idx(queue_idx)]; + rq_data = &enic->rq[rq_sop->data_queue_idx]; ret2 = vnic_rq_disable(rq_sop); rte_mb(); @@ -603,17 +652,17 @@ int enic_stop_rq(struct enic *enic, uint16_t queue_idx) int enic_alloc_rq(struct enic *enic, uint16_t queue_idx, unsigned int socket_id, struct rte_mempool *mp, - uint16_t nb_desc) + uint16_t nb_desc, uint16_t free_thresh) { int rc; - uint16_t sop_queue_idx = enic_sop_rq(queue_idx); - uint16_t data_queue_idx = enic_data_rq(queue_idx); + uint16_t sop_queue_idx = enic_rte_rq_idx_to_sop_idx(queue_idx); + uint16_t data_queue_idx = enic_rte_rq_idx_to_data_idx(queue_idx); struct vnic_rq *rq_sop = &enic->rq[sop_queue_idx]; struct vnic_rq *rq_data = &enic->rq[data_queue_idx]; unsigned int mbuf_size, mbufs_per_pkt; unsigned int nb_sop_desc, nb_data_desc; uint16_t min_sop, max_sop, min_data, max_data; - uint16_t mtu = enic->rte_dev->data->mtu; + uint32_t max_rx_pkt_len; rq_sop->is_sop = 1; rq_sop->data_queue_idx = data_queue_idx; @@ -624,24 +673,49 @@ int enic_alloc_rq(struct enic *enic, uint16_t queue_idx, rq_data->socket_id = socket_id; rq_data->mp = mp; rq_sop->in_use = 1; + rq_sop->rx_free_thresh = free_thresh; + rq_data->rx_free_thresh = free_thresh; + dev_debug(enic, "Set queue_id:%u free thresh:%u\n", queue_idx, + free_thresh); mbuf_size = (uint16_t)(rte_pktmbuf_data_room_size(mp) - RTE_PKTMBUF_HEADROOM); + /* max_rx_pkt_len includes the ethernet header and CRC. */ + max_rx_pkt_len = enic->rte_dev->data->dev_conf.rxmode.max_rx_pkt_len; - if (enic->rte_dev->data->dev_conf.rxmode.enable_scatter) { + if (enic->rte_dev->data->dev_conf.rxmode.offloads & + DEV_RX_OFFLOAD_SCATTER) { dev_info(enic, "Rq %u Scatter rx mode enabled\n", queue_idx); - /* ceil((mtu + ETHER_HDR_LEN + 4)/mbuf_size) */ - mbufs_per_pkt = ((mtu + ETHER_HDR_LEN + 4) + - (mbuf_size - 1)) / mbuf_size; + /* ceil((max pkt len)/mbuf_size) */ + mbufs_per_pkt = (max_rx_pkt_len + mbuf_size - 1) / mbuf_size; } else { dev_info(enic, "Scatter rx mode disabled\n"); mbufs_per_pkt = 1; + if (max_rx_pkt_len > mbuf_size) { + dev_warning(enic, "The maximum Rx packet size (%u) is" + " larger than the mbuf size (%u), and" + " scatter is disabled. Larger packets will" + " be truncated.\n", + max_rx_pkt_len, mbuf_size); + } } if (mbufs_per_pkt > 1) { dev_info(enic, "Rq %u Scatter rx mode in use\n", queue_idx); rq_sop->data_queue_enable = 1; rq_data->in_use = 1; + /* + * HW does not directly support rxmode.max_rx_pkt_len. HW always + * receives packet sizes up to the "max" MTU. + * If not using scatter, we can achieve the effect of dropping + * larger packets by reducing the size of posted buffers. + * See enic_alloc_rx_queue_mbufs(). + */ + if (max_rx_pkt_len < + enic_mtu_to_max_rx_pktlen(enic->rte_dev->data->mtu)) { + dev_warning(enic, "rxmode.max_rx_pkt_len is ignored" + " when scatter rx mode is in use.\n"); + } } else { dev_info(enic, "Rq %u Scatter rx mode not being used\n", queue_idx); @@ -681,8 +755,9 @@ int enic_alloc_rq(struct enic *enic, uint16_t queue_idx, nb_data_desc = max_data; } if (mbufs_per_pkt > 1) { - dev_info(enic, "For mtu %d and mbuf size %d valid rx descriptor range is %d to %d\n", - mtu, mbuf_size, min_sop + min_data, + dev_info(enic, "For max packet size %u and mbuf size %u valid" + " rx descriptor range is %u to %u\n", + max_rx_pkt_len, mbuf_size, min_sop + min_data, max_sop + max_data); } dev_info(enic, "Using %d rx descriptors (sop %d, data %d)\n", @@ -829,8 +904,11 @@ int enic_disable(struct enic *enic) unsigned int i; int err; - vnic_intr_mask(&enic->intr); - (void)vnic_intr_masked(&enic->intr); /* flush write */ + for (i = 0; i < enic->intr_count; i++) { + vnic_intr_mask(&enic->intr[i]); + (void)vnic_intr_masked(&enic->intr[i]); /* flush write */ + } + enic_rxq_intr_deinit(enic); rte_intr_disable(&enic->pdev->intr_handle); rte_intr_callback_unregister(&enic->pdev->intr_handle, enic_intr_handler, @@ -873,7 +951,8 @@ int enic_disable(struct enic *enic) vnic_rq_clean(&enic->rq[i], enic_free_rq_buf); for (i = 0; i < enic->cq_count; i++) vnic_cq_clean(&enic->cq[i]); - vnic_intr_clean(&enic->intr); + for (i = 0; i < enic->intr_count; i++) + vnic_intr_clean(&enic->intr[i]); return 0; } @@ -916,44 +995,42 @@ static int enic_dev_open(struct enic *enic) return err; } -static int enic_set_rsskey(struct enic *enic) +static int enic_set_rsskey(struct enic *enic, uint8_t *user_key) { dma_addr_t rss_key_buf_pa; union vnic_rss_key *rss_key_buf_va = NULL; - static union vnic_rss_key rss_key = { - .key = { - [0] = {.b = {85, 67, 83, 97, 119, 101, 115, 111, 109, 101}}, - [1] = {.b = {80, 65, 76, 79, 117, 110, 105, 113, 117, 101}}, - [2] = {.b = {76, 73, 78, 85, 88, 114, 111, 99, 107, 115}}, - [3] = {.b = {69, 78, 73, 67, 105, 115, 99, 111, 111, 108}}, - } - }; - int err; + int err, i; u8 name[NAME_MAX]; + RTE_ASSERT(user_key != NULL); snprintf((char *)name, NAME_MAX, "rss_key-%s", enic->bdf_name); rss_key_buf_va = enic_alloc_consistent(enic, sizeof(union vnic_rss_key), &rss_key_buf_pa, name); if (!rss_key_buf_va) return -ENOMEM; - rte_memcpy(rss_key_buf_va, &rss_key, sizeof(union vnic_rss_key)); + for (i = 0; i < ENIC_RSS_HASH_KEY_SIZE; i++) + rss_key_buf_va->key[i / 10].b[i % 10] = user_key[i]; err = enic_set_rss_key(enic, rss_key_buf_pa, sizeof(union vnic_rss_key)); + /* Save for later queries */ + if (!err) { + rte_memcpy(&enic->rss_key, rss_key_buf_va, + sizeof(union vnic_rss_key)); + } enic_free_consistent(enic, sizeof(union vnic_rss_key), rss_key_buf_va, rss_key_buf_pa); return err; } -static int enic_set_rsscpu(struct enic *enic, u8 rss_hash_bits) +int enic_set_rss_reta(struct enic *enic, union vnic_rss_cpu *rss_cpu) { dma_addr_t rss_cpu_buf_pa; union vnic_rss_cpu *rss_cpu_buf_va = NULL; - int i; int err; u8 name[NAME_MAX]; @@ -963,9 +1040,7 @@ static int enic_set_rsscpu(struct enic *enic, u8 rss_hash_bits) if (!rss_cpu_buf_va) return -ENOMEM; - for (i = 0; i < (1 << rss_hash_bits); i++) - (*rss_cpu_buf_va).cpu[i / 4].b[i % 4] = - enic_sop_rq(i % enic->rq_count); + rte_memcpy(rss_cpu_buf_va, rss_cpu, sizeof(union vnic_rss_cpu)); err = enic_set_rss_cpu(enic, rss_cpu_buf_pa, @@ -974,6 +1049,9 @@ static int enic_set_rsscpu(struct enic *enic, u8 rss_hash_bits) enic_free_consistent(enic, sizeof(union vnic_rss_cpu), rss_cpu_buf_va, rss_cpu_buf_pa); + /* Save for later queries */ + if (!err) + rte_memcpy(&enic->rss_cpu, rss_cpu, sizeof(union vnic_rss_cpu)); return err; } @@ -983,8 +1061,6 @@ static int enic_set_niccfg(struct enic *enic, u8 rss_default_cpu, const u8 tso_ipid_split_en = 0; int err; - /* Enable VLAN tag stripping */ - err = enic_set_nic_cfg(enic, rss_default_cpu, rss_hash_type, rss_hash_bits, rss_base_cpu, @@ -994,47 +1070,50 @@ static int enic_set_niccfg(struct enic *enic, u8 rss_default_cpu, return err; } -int enic_set_rss_nic_cfg(struct enic *enic) +/* Initialize RSS with defaults, called from dev_configure */ +int enic_init_rss_nic_cfg(struct enic *enic) { - const u8 rss_default_cpu = 0; - const u8 rss_hash_type = NIC_CFG_RSS_HASH_TYPE_IPV4 | - NIC_CFG_RSS_HASH_TYPE_TCP_IPV4 | - NIC_CFG_RSS_HASH_TYPE_IPV6 | - NIC_CFG_RSS_HASH_TYPE_TCP_IPV6; - const u8 rss_hash_bits = 7; - const u8 rss_base_cpu = 0; - u8 rss_enable = ENIC_SETTING(enic, RSS) && (enic->rq_count > 1); - - if (rss_enable) { - if (!enic_set_rsskey(enic)) { - if (enic_set_rsscpu(enic, rss_hash_bits)) { - rss_enable = 0; - dev_warning(enic, "RSS disabled, "\ - "Failed to set RSS cpu indirection table."); - } - } else { - rss_enable = 0; - dev_warning(enic, - "RSS disabled, Failed to set RSS key.\n"); + static uint8_t default_rss_key[] = { + 85, 67, 83, 97, 119, 101, 115, 111, 109, 101, + 80, 65, 76, 79, 117, 110, 105, 113, 117, 101, + 76, 73, 78, 85, 88, 114, 111, 99, 107, 115, + 69, 78, 73, 67, 105, 115, 99, 111, 111, 108, + }; + struct rte_eth_rss_conf rss_conf; + union vnic_rss_cpu rss_cpu; + int ret, i; + + rss_conf = enic->rte_dev->data->dev_conf.rx_adv_conf.rss_conf; + /* + * If setting key for the first time, and the user gives us none, then + * push the default key to NIC. + */ + if (rss_conf.rss_key == NULL) { + rss_conf.rss_key = default_rss_key; + rss_conf.rss_key_len = ENIC_RSS_HASH_KEY_SIZE; + } + ret = enic_set_rss_conf(enic, &rss_conf); + if (ret) { + dev_err(enic, "Failed to configure RSS\n"); + return ret; + } + if (enic->rss_enable) { + /* If enabling RSS, use the default reta */ + for (i = 0; i < ENIC_RSS_RETA_SIZE; i++) { + rss_cpu.cpu[i / 4].b[i % 4] = + enic_rte_rq_idx_to_sop_idx(i % enic->rq_count); } + ret = enic_set_rss_reta(enic, &rss_cpu); + if (ret) + dev_err(enic, "Failed to set RSS indirection table\n"); } - - return enic_set_niccfg(enic, rss_default_cpu, rss_hash_type, - rss_hash_bits, rss_base_cpu, rss_enable); + return ret; } int enic_setup_finish(struct enic *enic) { - int ret; - enic_init_soft_stats(enic); - ret = enic_set_rss_nic_cfg(enic); - if (ret) { - dev_err(enic, "Failed to config nic, aborting.\n"); - return -1; - } - /* Default conf */ vnic_dev_packet_filter(enic->vdev, 1 /* directed */, @@ -1049,6 +1128,98 @@ int enic_setup_finish(struct enic *enic) return 0; } +static int enic_rss_conf_valid(struct enic *enic, + struct rte_eth_rss_conf *rss_conf) +{ + /* RSS is disabled per VIC settings. Ignore rss_conf. */ + if (enic->flow_type_rss_offloads == 0) + return 0; + if (rss_conf->rss_key != NULL && + rss_conf->rss_key_len != ENIC_RSS_HASH_KEY_SIZE) { + dev_err(enic, "Given rss_key is %d bytes, it must be %d\n", + rss_conf->rss_key_len, ENIC_RSS_HASH_KEY_SIZE); + return -EINVAL; + } + if (rss_conf->rss_hf != 0 && + (rss_conf->rss_hf & enic->flow_type_rss_offloads) == 0) { + dev_err(enic, "Given rss_hf contains none of the supported" + " types\n"); + return -EINVAL; + } + return 0; +} + +/* Set hash type and key according to rss_conf */ +int enic_set_rss_conf(struct enic *enic, struct rte_eth_rss_conf *rss_conf) +{ + struct rte_eth_dev *eth_dev; + uint64_t rss_hf; + u8 rss_hash_type; + u8 rss_enable; + int ret; + + RTE_ASSERT(rss_conf != NULL); + ret = enic_rss_conf_valid(enic, rss_conf); + if (ret) { + dev_err(enic, "RSS configuration (rss_conf) is invalid\n"); + return ret; + } + + eth_dev = enic->rte_dev; + rss_hash_type = 0; + rss_hf = rss_conf->rss_hf & enic->flow_type_rss_offloads; + if (enic->rq_count > 1 && + (eth_dev->data->dev_conf.rxmode.mq_mode & ETH_MQ_RX_RSS_FLAG) && + rss_hf != 0) { + rss_enable = 1; + if (rss_hf & ETH_RSS_IPV4) + rss_hash_type |= NIC_CFG_RSS_HASH_TYPE_IPV4; + if (rss_hf & ETH_RSS_NONFRAG_IPV4_TCP) + rss_hash_type |= NIC_CFG_RSS_HASH_TYPE_TCP_IPV4; + if (rss_hf & ETH_RSS_IPV6) + rss_hash_type |= NIC_CFG_RSS_HASH_TYPE_IPV6; + if (rss_hf & ETH_RSS_NONFRAG_IPV6_TCP) + rss_hash_type |= NIC_CFG_RSS_HASH_TYPE_TCP_IPV6; + if (rss_hf & ETH_RSS_IPV6_EX) + rss_hash_type |= NIC_CFG_RSS_HASH_TYPE_IPV6_EX; + if (rss_hf & ETH_RSS_IPV6_TCP_EX) + rss_hash_type |= NIC_CFG_RSS_HASH_TYPE_TCP_IPV6_EX; + } else { + rss_enable = 0; + rss_hf = 0; + } + + /* Set the hash key if provided */ + if (rss_enable && rss_conf->rss_key) { + ret = enic_set_rsskey(enic, rss_conf->rss_key); + if (ret) { + dev_err(enic, "Failed to set RSS key\n"); + return ret; + } + } + + ret = enic_set_niccfg(enic, ENIC_RSS_DEFAULT_CPU, rss_hash_type, + ENIC_RSS_HASH_BITS, ENIC_RSS_BASE_CPU, + rss_enable); + if (!ret) { + enic->rss_hf = rss_hf; + enic->rss_hash_type = rss_hash_type; + enic->rss_enable = rss_enable; + } + return 0; +} + +int enic_set_vlan_strip(struct enic *enic) +{ + /* + * Unfortunately, VLAN strip on/off and RSS on/off are configured + * together. So, re-do niccfg, preserving the current RSS settings. + */ + return enic_set_niccfg(enic, ENIC_RSS_DEFAULT_CPU, enic->rss_hash_type, + ENIC_RSS_HASH_BITS, ENIC_RSS_BASE_CPU, + enic->rss_enable); +} + void enic_add_packet_filter(struct enic *enic) { /* Args -> directed, multicast, broadcast, promisc, allmulti */ @@ -1069,6 +1240,10 @@ static void enic_dev_deinit(struct enic *enic) vnic_dev_notify_unset(enic->vdev); rte_free(eth_dev->data->mac_addrs); + rte_free(enic->cq); + rte_free(enic->intr); + rte_free(enic->rq); + rte_free(enic->wq); } @@ -1076,27 +1251,38 @@ int enic_set_vnic_res(struct enic *enic) { struct rte_eth_dev *eth_dev = enic->rte_dev; int rc = 0; + unsigned int required_rq, required_wq, required_cq, required_intr; + + /* Always use two vNIC RQs per eth_dev RQ, regardless of Rx scatter. */ + required_rq = eth_dev->data->nb_rx_queues * 2; + required_wq = eth_dev->data->nb_tx_queues; + required_cq = eth_dev->data->nb_rx_queues + eth_dev->data->nb_tx_queues; + required_intr = 1; /* 1 for LSC even if intr_conf.lsc is 0 */ + if (eth_dev->data->dev_conf.intr_conf.rxq) { + required_intr += eth_dev->data->nb_rx_queues; + } - /* With Rx scatter support, two RQs are now used per RQ used by - * the application. - */ - if (enic->conf_rq_count < eth_dev->data->nb_rx_queues) { + if (enic->conf_rq_count < required_rq) { dev_err(dev, "Not enough Receive queues. Requested:%u which uses %d RQs on VIC, Configured:%u\n", eth_dev->data->nb_rx_queues, - eth_dev->data->nb_rx_queues * 2, enic->conf_rq_count); + required_rq, enic->conf_rq_count); rc = -EINVAL; } - if (enic->conf_wq_count < eth_dev->data->nb_tx_queues) { + if (enic->conf_wq_count < required_wq) { dev_err(dev, "Not enough Transmit queues. Requested:%u, Configured:%u\n", eth_dev->data->nb_tx_queues, enic->conf_wq_count); rc = -EINVAL; } - if (enic->conf_cq_count < (eth_dev->data->nb_rx_queues + - eth_dev->data->nb_tx_queues)) { + if (enic->conf_cq_count < required_cq) { dev_err(dev, "Not enough Completion queues. Required:%u, Configured:%u\n", - (eth_dev->data->nb_rx_queues + - eth_dev->data->nb_tx_queues), enic->conf_cq_count); + required_cq, enic->conf_cq_count); + rc = -EINVAL; + } + if (enic->conf_intr_count < required_intr) { + dev_err(dev, "Not enough Interrupts to support Rx queue" + " interrupts. Required:%u, Configured:%u\n", + required_intr, enic->conf_intr_count); rc = -EINVAL; } @@ -1104,6 +1290,7 @@ int enic_set_vnic_res(struct enic *enic) enic->rq_count = eth_dev->data->nb_rx_queues; enic->wq_count = eth_dev->data->nb_tx_queues; enic->cq_count = enic->rq_count + enic->wq_count; + enic->intr_count = required_intr; } return rc; @@ -1114,11 +1301,12 @@ static int enic_reinit_rq(struct enic *enic, unsigned int rq_idx) { struct vnic_rq *sop_rq, *data_rq; - unsigned int cq_idx = enic_cq_rq(enic, rq_idx); + unsigned int cq_idx; int rc = 0; - sop_rq = &enic->rq[enic_sop_rq(rq_idx)]; - data_rq = &enic->rq[enic_data_rq(rq_idx)]; + sop_rq = &enic->rq[enic_rte_rq_idx_to_sop_idx(rq_idx)]; + data_rq = &enic->rq[enic_rte_rq_idx_to_data_idx(rq_idx)]; + cq_idx = rq_idx; vnic_cq_clean(&enic->cq[cq_idx]); vnic_cq_init(&enic->cq[cq_idx], @@ -1134,12 +1322,14 @@ enic_reinit_rq(struct enic *enic, unsigned int rq_idx) 0 /* cq_message_addr */); - vnic_rq_init_start(sop_rq, enic_cq_rq(enic, enic_sop_rq(rq_idx)), - 0, sop_rq->ring.desc_count - 1, 1, 0); + vnic_rq_init_start(sop_rq, enic_cq_rq(enic, + enic_rte_rq_idx_to_sop_idx(rq_idx)), 0, + sop_rq->ring.desc_count - 1, 1, 0); if (data_rq->in_use) { vnic_rq_init_start(data_rq, - enic_cq_rq(enic, enic_data_rq(rq_idx)), - 0, data_rq->ring.desc_count - 1, 1, 0); + enic_cq_rq(enic, + enic_rte_rq_idx_to_data_idx(rq_idx)), 0, + data_rq->ring.desc_count - 1, 1, 0); } rc = enic_alloc_rx_queue_mbufs(enic, sop_rq); @@ -1176,6 +1366,9 @@ int enic_set_mtu(struct enic *enic, uint16_t new_mtu) old_mtu = eth_dev->data->mtu; config_mtu = enic->config.mtu; + if (rte_eal_process_type() != RTE_PROC_PRIMARY) + return -E_RTE_SECONDARY; + if (new_mtu > enic->max_mtu) { dev_err(enic, "MTU not updated: requested (%u) greater than max (%u)\n", @@ -1196,7 +1389,8 @@ int enic_set_mtu(struct enic *enic, uint16_t new_mtu) /* The easy case is when scatter is disabled. However if the MTU * becomes greater than the mbuf data size, packet drops will ensue. */ - if (!enic->rte_dev->data->dev_conf.rxmode.enable_scatter) { + if (!(enic->rte_dev->data->dev_conf.rxmode.offloads & + DEV_RX_OFFLOAD_SCATTER)) { eth_dev->data->mtu = new_mtu; goto set_mtu_done; } @@ -1212,7 +1406,8 @@ int enic_set_mtu(struct enic *enic, uint16_t new_mtu) for (rq_idx = 0; rq_idx < enic->rq_count * 2; rq_idx++) { rq = &enic->rq[rq_idx]; if (rq->is_sop && rq->in_use) { - rc = enic_stop_rq(enic, enic_rq_sop(rq_idx)); + rc = enic_stop_rq(enic, + enic_sop_rq_idx_to_rte_idx(rq_idx)); if (rc) { dev_err(enic, "Failed to stop Rq %u\n", rq_idx); goto set_mtu_done; @@ -1220,7 +1415,7 @@ int enic_set_mtu(struct enic *enic, uint16_t new_mtu) } } - /* replace Rx funciton with a no-op to avoid getting stale pkts */ + /* replace Rx function with a no-op to avoid getting stale pkts */ eth_dev->rx_pkt_burst = enic_dummy_recv_pkts; rte_mb(); @@ -1234,11 +1429,11 @@ int enic_set_mtu(struct enic *enic, uint16_t new_mtu) /* free and reallocate RQs with the new MTU */ for (rq_idx = 0; rq_idx < enic->rq_count; rq_idx++) { - rq = &enic->rq[enic_sop_rq(rq_idx)]; + rq = &enic->rq[enic_rte_rq_idx_to_sop_idx(rq_idx)]; enic_free_rq(rq); rc = enic_alloc_rq(enic, rq_idx, rq->socket_id, rq->mp, - rq->tot_nb_desc); + rq->tot_nb_desc, rq->rx_free_thresh); if (rc) { dev_err(enic, "Fatal MTU alloc error- No traffic will pass\n"); @@ -1260,7 +1455,7 @@ int enic_set_mtu(struct enic *enic, uint16_t new_mtu) /* restart Rx traffic */ for (rq_idx = 0; rq_idx < enic->rq_count; rq_idx++) { - rq = &enic->rq[enic_sop_rq(rq_idx)]; + rq = &enic->rq[enic_rte_rq_idx_to_sop_idx(rq_idx)]; if (rq->is_sop && rq->in_use) enic_start_rq(enic, rq_idx); } @@ -1294,20 +1489,49 @@ static int enic_dev_init(struct enic *enic) dev_err(enic, "See the ENIC PMD guide for more information.\n"); return -EINVAL; } + /* Queue counts may be zeros. rte_zmalloc returns NULL in that case. */ + enic->cq = rte_zmalloc("enic_vnic_cq", sizeof(struct vnic_cq) * + enic->conf_cq_count, 8); + enic->intr = rte_zmalloc("enic_vnic_intr", sizeof(struct vnic_intr) * + enic->conf_intr_count, 8); + enic->rq = rte_zmalloc("enic_vnic_rq", sizeof(struct vnic_rq) * + enic->conf_rq_count, 8); + enic->wq = rte_zmalloc("enic_vnic_wq", sizeof(struct vnic_wq) * + enic->conf_wq_count, 8); + if (enic->conf_cq_count > 0 && enic->cq == NULL) { + dev_err(enic, "failed to allocate vnic_cq, aborting.\n"); + return -1; + } + if (enic->conf_intr_count > 0 && enic->intr == NULL) { + dev_err(enic, "failed to allocate vnic_intr, aborting.\n"); + return -1; + } + if (enic->conf_rq_count > 0 && enic->rq == NULL) { + dev_err(enic, "failed to allocate vnic_rq, aborting.\n"); + return -1; + } + if (enic->conf_wq_count > 0 && enic->wq == NULL) { + dev_err(enic, "failed to allocate vnic_wq, aborting.\n"); + return -1; + } /* Get the supported filters */ enic_fdir_info(enic); - eth_dev->data->mac_addrs = rte_zmalloc("enic_mac_addr", ETH_ALEN, 0); + eth_dev->data->mac_addrs = rte_zmalloc("enic_mac_addr", ETH_ALEN + * ENIC_MAX_MAC_ADDR, 0); if (!eth_dev->data->mac_addrs) { dev_err(enic, "mac addr storage alloc failed, aborting.\n"); return -1; } ether_addr_copy((struct ether_addr *) enic->mac_addr, - ð_dev->data->mac_addrs[0]); + eth_dev->data->mac_addrs); vnic_dev_set_reset_flag(enic->vdev, 0); + LIST_INIT(&enic->flows); + rte_spinlock_init(&enic->flows_lock); + /* set up link status checking */ vnic_dev_notify_set(enic->vdev, -1); /* No Intr for notify */ @@ -1322,6 +1546,10 @@ int enic_probe(struct enic *enic) dev_debug(enic, " Initializing ENIC PMD\n"); + /* if this is a secondary process the hardware is already initialized */ + if (rte_eal_process_type() != RTE_PROC_PRIMARY) + return 0; + enic->bar0.vaddr = (void *)pdev->mem_resource[0].addr; enic->bar0.len = pdev->mem_resource[0].len; @@ -1339,6 +1567,15 @@ int enic_probe(struct enic *enic) enic_alloc_consistent, enic_free_consistent); + /* + * Allocate the consistent memory for stats upfront so both primary and + * secondary processes can dump stats. + */ + err = vnic_dev_alloc_stats_mem(enic->vdev); + if (err) { + dev_err(enic, "Failed to allocate cmd memory, aborting\n"); + goto err_out_unregister; + } /* Issue device open to get device in known state */ err = enic_dev_open(enic); if (err) {