X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Ftap%2Frte_eth_tap.c;h=7cfa0a85d375b19a140074bda12d30e1924b0e90;hb=98a7ea332ba3da0f74ec951595d36a616165b255;hp=d8322e8e8fd85c66bd704070ed19f4077c707062;hpb=f279f6f2bd90d3cff1eca09cf60342ea33bb3ec6;p=dpdk.git diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c index d8322e8e8f..7cfa0a85d3 100644 --- a/drivers/net/tap/rte_eth_tap.c +++ b/drivers/net/tap/rte_eth_tap.c @@ -1,7 +1,7 @@ /*- * BSD LICENSE * - * Copyright(c) 2016 Intel Corporation. All rights reserved. + * Copyright(c) 2016-2017 Intel Corporation. All rights reserved. * All rights reserved. * * Redistribution and use in source and binary forms, with or without @@ -33,13 +33,17 @@ #include #include +#include #include #include #include +#include #include #include #include #include +#include +#include #include #include @@ -71,12 +75,8 @@ #define ETH_TAP_IFACE_ARG "iface" #define ETH_TAP_SPEED_ARG "speed" #define ETH_TAP_REMOTE_ARG "remote" - -#ifdef IFF_MULTI_QUEUE -#define RTE_PMD_TAP_MAX_QUEUES 16 -#else -#define RTE_PMD_TAP_MAX_QUEUES 1 -#endif +#define ETH_TAP_MAC_ARG "mac" +#define ETH_TAP_MAC_FIXED "fixed" #define FLOWER_KERNEL_VERSION KERNEL_VERSION(4, 2, 0) #define FLOWER_VLAN_KERNEL_VERSION KERNEL_VERSION(4, 9, 0) @@ -87,6 +87,7 @@ static const char *valid_arguments[] = { ETH_TAP_IFACE_ARG, ETH_TAP_SPEED_ARG, ETH_TAP_REMOTE_ARG, + ETH_TAP_MAC_ARG, NULL }; @@ -108,9 +109,12 @@ tap_trigger_cb(int sig __rte_unused) tap_trigger = (tap_trigger + 1) | 0x80000000; } -static int -tap_ioctl(struct pmd_internals *pmd, unsigned long request, - struct ifreq *ifr, int set); +/* Specifies on what netdevices the ioctl should be applied */ +enum ioctl_mode { + LOCAL_AND_REMOTE, + LOCAL_ONLY, + REMOTE_ONLY, +}; static int tap_intr_handle_set(struct rte_eth_dev *dev, int set); @@ -120,7 +124,7 @@ static int tap_intr_handle_set(struct rte_eth_dev *dev, int set); * supplied name. */ static int -tun_alloc(struct pmd_internals *pmd, uint16_t qid) +tun_alloc(struct pmd_internals *pmd) { struct ifreq ifr; #ifdef IFF_MULTI_QUEUE @@ -219,70 +223,6 @@ tun_alloc(struct pmd_internals *pmd, uint16_t qid) strerror(errno)); } - if (qid == 0) { - struct ifreq ifr; - - if (tap_ioctl(pmd, SIOCGIFHWADDR, &ifr, 0) < 0) - goto error; - rte_memcpy(&pmd->eth_addr, ifr.ifr_hwaddr.sa_data, - ETHER_ADDR_LEN); - - pmd->if_index = if_nametoindex(pmd->name); - if (!pmd->if_index) { - RTE_LOG(ERR, PMD, - "Could not find ifindex for %s: rte_flow won't be usable.\n", - pmd->name); - return fd; - } - if (!pmd->flower_support) - return fd; - if (qdisc_create_multiq(pmd->nlsk_fd, pmd->if_index) < 0) { - RTE_LOG(ERR, PMD, - "Could not create multiq qdisc for %s: rte_flow won't be usable.\n", - pmd->name); - return fd; - } - if (qdisc_create_ingress(pmd->nlsk_fd, pmd->if_index) < 0) { - RTE_LOG(ERR, PMD, - "Could not create multiq qdisc for %s: rte_flow won't be usable.\n", - pmd->name); - return fd; - } - if (pmd->remote_if_index) { - /* - * Flush usually returns negative value because it tries - * to delete every QDISC (and on a running device, one - * QDISC at least is needed). Ignore negative return - * value. - */ - qdisc_flush(pmd->nlsk_fd, pmd->remote_if_index); - if (qdisc_create_ingress(pmd->nlsk_fd, - pmd->remote_if_index) < 0) - goto remote_fail; - LIST_INIT(&pmd->implicit_flows); - if (tap_flow_implicit_create( - pmd, TAP_REMOTE_LOCAL_MAC) < 0) - goto remote_fail; - if (tap_flow_implicit_create( - pmd, TAP_REMOTE_BROADCAST) < 0) - goto remote_fail; - if (tap_flow_implicit_create( - pmd, TAP_REMOTE_BROADCASTV6) < 0) - goto remote_fail; - if (tap_flow_implicit_create( - pmd, TAP_REMOTE_TX) < 0) - goto remote_fail; - } - } - - return fd; - -remote_fail: - RTE_LOG(ERR, PMD, - "Could not set up remote flow rules for %s: remote disabled.\n", - pmd->name); - pmd->remote_if_index = 0; - tap_flow_implicit_flush(pmd, NULL); return fd; error: @@ -291,6 +231,60 @@ error: return -1; } +static void +tap_verify_csum(struct rte_mbuf *mbuf) +{ + uint32_t l2 = mbuf->packet_type & RTE_PTYPE_L2_MASK; + uint32_t l3 = mbuf->packet_type & RTE_PTYPE_L3_MASK; + uint32_t l4 = mbuf->packet_type & RTE_PTYPE_L4_MASK; + unsigned int l2_len = sizeof(struct ether_hdr); + unsigned int l3_len; + uint16_t cksum = 0; + void *l3_hdr; + void *l4_hdr; + + if (l2 == RTE_PTYPE_L2_ETHER_VLAN) + l2_len += 4; + else if (l2 == RTE_PTYPE_L2_ETHER_QINQ) + l2_len += 8; + /* Don't verify checksum for packets with discontinuous L2 header */ + if (unlikely(l2_len + sizeof(struct ipv4_hdr) > + rte_pktmbuf_data_len(mbuf))) + return; + l3_hdr = rte_pktmbuf_mtod_offset(mbuf, void *, l2_len); + if (l3 == RTE_PTYPE_L3_IPV4 || l3 == RTE_PTYPE_L3_IPV4_EXT) { + struct ipv4_hdr *iph = l3_hdr; + + /* ihl contains the number of 4-byte words in the header */ + l3_len = 4 * (iph->version_ihl & 0xf); + if (unlikely(l2_len + l3_len > rte_pktmbuf_data_len(mbuf))) + return; + + cksum = ~rte_raw_cksum(iph, l3_len); + mbuf->ol_flags |= cksum ? + PKT_RX_IP_CKSUM_BAD : + PKT_RX_IP_CKSUM_GOOD; + } else if (l3 == RTE_PTYPE_L3_IPV6) { + l3_len = sizeof(struct ipv6_hdr); + } else { + /* IPv6 extensions are not supported */ + return; + } + if (l4 == RTE_PTYPE_L4_UDP || l4 == RTE_PTYPE_L4_TCP) { + l4_hdr = rte_pktmbuf_mtod_offset(mbuf, void *, l2_len + l3_len); + /* Don't verify checksum for multi-segment packets. */ + if (mbuf->nb_segs > 1) + return; + if (l3 == RTE_PTYPE_L3_IPV4) + cksum = ~rte_ipv4_udptcp_cksum(l3_hdr, l4_hdr); + else if (l3 == RTE_PTYPE_L3_IPV6) + cksum = ~rte_ipv6_udptcp_cksum(l3_hdr, l4_hdr); + mbuf->ol_flags |= cksum ? + PKT_RX_L4_CKSUM_BAD : + PKT_RX_L4_CKSUM_GOOD; + } +} + /* Callback to handle the rx burst of packets to the correct interface and * file descriptor(s) in a multi-queue setup. */ @@ -371,6 +365,8 @@ pmd_rx_burst(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts) seg->next = NULL; mbuf->packet_type = rte_net_get_ptype(mbuf, NULL, RTE_PTYPE_ALL_MASK); + if (rxq->rxmode->hw_ip_checksum) + tap_verify_csum(mbuf); /* account for the receive frame */ bufs[num_rx++] = mbuf; @@ -383,6 +379,56 @@ end: return num_rx; } +static void +tap_tx_offload(char *packet, uint64_t ol_flags, unsigned int l2_len, + unsigned int l3_len) +{ + void *l3_hdr = packet + l2_len; + + if (ol_flags & (PKT_TX_IP_CKSUM | PKT_TX_IPV4)) { + struct ipv4_hdr *iph = l3_hdr; + uint16_t cksum; + + iph->hdr_checksum = 0; + cksum = rte_raw_cksum(iph, l3_len); + iph->hdr_checksum = (cksum == 0xffff) ? cksum : ~cksum; + } + if (ol_flags & PKT_TX_L4_MASK) { + uint16_t l4_len; + uint32_t cksum; + uint16_t *l4_cksum; + void *l4_hdr; + + l4_hdr = packet + l2_len + l3_len; + if ((ol_flags & PKT_TX_L4_MASK) == PKT_TX_UDP_CKSUM) + l4_cksum = &((struct udp_hdr *)l4_hdr)->dgram_cksum; + else if ((ol_flags & PKT_TX_L4_MASK) == PKT_TX_TCP_CKSUM) + l4_cksum = &((struct tcp_hdr *)l4_hdr)->cksum; + else + return; + *l4_cksum = 0; + if (ol_flags & PKT_TX_IPV4) { + struct ipv4_hdr *iph = l3_hdr; + + l4_len = rte_be_to_cpu_16(iph->total_length) - l3_len; + cksum = rte_ipv4_phdr_cksum(l3_hdr, 0); + } else { + struct ipv6_hdr *ip6h = l3_hdr; + + /* payload_len does not include ext headers */ + l4_len = rte_be_to_cpu_16(ip6h->payload_len) - + l3_len + sizeof(struct ipv6_hdr); + cksum = rte_ipv6_phdr_cksum(l3_hdr, 0); + } + cksum += rte_raw_cksum(l4_hdr, l4_len); + cksum = ((cksum & 0xffff0000) >> 16) + (cksum & 0xffff); + cksum = (~cksum) & 0xffff; + if (cksum == 0) + cksum = 0xffff; + *l4_cksum = cksum; + } +} + /* Callback to handle sending packets from the tap interface */ static uint16_t @@ -403,6 +449,7 @@ pmd_tx_burst(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts) struct iovec iovecs[mbuf->nb_segs + 1]; struct tun_pi pi = { .flags = 0 }; struct rte_mbuf *seg = mbuf; + char m_copy[mbuf->data_len]; int n; int j; @@ -418,6 +465,19 @@ pmd_tx_burst(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts) rte_pktmbuf_mtod(seg, void *); seg = seg->next; } + if (mbuf->ol_flags & (PKT_TX_IP_CKSUM | PKT_TX_IPV4) || + (mbuf->ol_flags & PKT_TX_L4_MASK) == PKT_TX_UDP_CKSUM || + (mbuf->ol_flags & PKT_TX_L4_MASK) == PKT_TX_TCP_CKSUM) { + /* Support only packets with all data in the same seg */ + if (mbuf->nb_segs > 1) + break; + /* To change checksums, work on a copy of data. */ + rte_memcpy(m_copy, rte_pktmbuf_mtod(mbuf, void *), + rte_pktmbuf_data_len(mbuf)); + tap_tx_offload(m_copy, mbuf->ol_flags, + mbuf->l2_len, mbuf->l3_len); + iovecs[1].iov_base = m_copy; + } /* copy the tx frame data */ n = writev(txq->fd, iovecs, mbuf->nb_segs + 1); if (n <= 0) @@ -435,29 +495,42 @@ pmd_tx_burst(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts) return num_tx; } +static const char * +tap_ioctl_req2str(unsigned long request) +{ + switch (request) { + case SIOCSIFFLAGS: + return "SIOCSIFFLAGS"; + case SIOCGIFFLAGS: + return "SIOCGIFFLAGS"; + case SIOCGIFHWADDR: + return "SIOCGIFHWADDR"; + case SIOCSIFHWADDR: + return "SIOCSIFHWADDR"; + case SIOCSIFMTU: + return "SIOCSIFMTU"; + } + return "UNKNOWN"; +} + static int tap_ioctl(struct pmd_internals *pmd, unsigned long request, - struct ifreq *ifr, int set) + struct ifreq *ifr, int set, enum ioctl_mode mode) { short req_flags = ifr->ifr_flags; - int remote = !!pmd->remote_if_index; + int remote = pmd->remote_if_index && + (mode == REMOTE_ONLY || mode == LOCAL_AND_REMOTE); + if (!pmd->remote_if_index && mode == REMOTE_ONLY) + return 0; /* * If there is a remote netdevice, apply ioctl on it, then apply it on * the tap netdevice. */ - if (request == SIOCGIFFLAGS && !set) { - /* - * Special case for getting flags. If set is given, - * then return the flags from the remote netdevice only. - * Otherwise return the flags from the tap netdevice. - */ - remote = 0; - } apply: if (remote) snprintf(ifr->ifr_name, IFNAMSIZ, "%s", pmd->remote_iface); - else + else if (mode == LOCAL_ONLY || mode == LOCAL_AND_REMOTE) snprintf(ifr->ifr_name, IFNAMSIZ, "%s", pmd->name); switch (request) { case SIOCSIFFLAGS: @@ -470,33 +543,22 @@ apply: ifr->ifr_flags &= ~req_flags; break; case SIOCGIFFLAGS: - if (remote && set) - remote = 0; /* don't loop */ - break; case SIOCGIFHWADDR: - /* Set remote MAC on the tap netdevice */ - if (!remote && pmd->remote_if_index) { - request = SIOCSIFHWADDR; - goto apply; - } - break; case SIOCSIFHWADDR: case SIOCSIFMTU: break; default: - RTE_LOG(WARNING, PMD, "%s: ioctl() called with wrong arg\n", - pmd->name); - return -EINVAL; + RTE_ASSERT(!"unsupported request type: must not happen"); } if (ioctl(pmd->ioctl_sock, request, ifr) < 0) goto error; - if (remote--) + if (remote-- && mode == LOCAL_AND_REMOTE) goto apply; return 0; error: - RTE_LOG(ERR, PMD, "%s: ioctl(%lu) failed with error: %s\n", - ifr->ifr_name, request, strerror(errno)); + RTE_LOG(DEBUG, PMD, "%s: %s(%s) failed: %s(%d)\n", ifr->ifr_name, + __func__, tap_ioctl_req2str(request), strerror(errno), errno); return -errno; } @@ -507,7 +569,7 @@ tap_link_set_down(struct rte_eth_dev *dev) struct ifreq ifr = { .ifr_flags = IFF_UP }; dev->data->dev_link.link_status = ETH_LINK_DOWN; - return tap_ioctl(pmd, SIOCSIFFLAGS, &ifr, 0); + return tap_ioctl(pmd, SIOCSIFFLAGS, &ifr, 0, LOCAL_AND_REMOTE); } static int @@ -517,7 +579,7 @@ tap_link_set_up(struct rte_eth_dev *dev) struct ifreq ifr = { .ifr_flags = IFF_UP }; dev->data->dev_link.link_status = ETH_LINK_UP; - return tap_ioctl(pmd, SIOCSIFFLAGS, &ifr, 1); + return tap_ioctl(pmd, SIOCSIFFLAGS, &ifr, 1, LOCAL_AND_REMOTE); } static int @@ -593,6 +655,13 @@ tap_dev_info(struct rte_eth_dev *dev, struct rte_eth_dev_info *dev_info) dev_info->min_rx_bufsize = 0; dev_info->pci_dev = NULL; dev_info->speed_capa = tap_dev_speed_capa(); + dev_info->rx_offload_capa = (DEV_RX_OFFLOAD_IPV4_CKSUM | + DEV_RX_OFFLOAD_UDP_CKSUM | + DEV_RX_OFFLOAD_TCP_CKSUM); + dev_info->tx_offload_capa = + (DEV_TX_OFFLOAD_IPV4_CKSUM | + DEV_TX_OFFLOAD_UDP_CKSUM | + DEV_TX_OFFLOAD_TCP_CKSUM); } static void @@ -651,7 +720,7 @@ tap_stats_reset(struct rte_eth_dev *dev) } static void -tap_dev_close(struct rte_eth_dev *dev __rte_unused) +tap_dev_close(struct rte_eth_dev *dev) { int i; struct pmd_internals *internals = dev->data->dev_private; @@ -702,14 +771,14 @@ tap_link_update(struct rte_eth_dev *dev, int wait_to_complete __rte_unused) struct ifreq ifr = { .ifr_flags = 0 }; if (pmd->remote_if_index) { - tap_ioctl(pmd, SIOCGIFFLAGS, &ifr, 1); + tap_ioctl(pmd, SIOCGIFFLAGS, &ifr, 0, REMOTE_ONLY); if (!(ifr.ifr_flags & IFF_UP) || !(ifr.ifr_flags & IFF_RUNNING)) { dev_link->link_status = ETH_LINK_DOWN; return 0; } } - tap_ioctl(pmd, SIOCGIFFLAGS, &ifr, 0); + tap_ioctl(pmd, SIOCGIFFLAGS, &ifr, 0, LOCAL_ONLY); dev_link->link_status = ((ifr.ifr_flags & IFF_UP) && (ifr.ifr_flags & IFF_RUNNING) ? ETH_LINK_UP : @@ -724,7 +793,7 @@ tap_promisc_enable(struct rte_eth_dev *dev) struct ifreq ifr = { .ifr_flags = IFF_PROMISC }; dev->data->promiscuous = 1; - tap_ioctl(pmd, SIOCSIFFLAGS, &ifr, 1); + tap_ioctl(pmd, SIOCSIFFLAGS, &ifr, 1, LOCAL_AND_REMOTE); if (pmd->remote_if_index) tap_flow_implicit_create(pmd, TAP_REMOTE_PROMISC); } @@ -736,7 +805,7 @@ tap_promisc_disable(struct rte_eth_dev *dev) struct ifreq ifr = { .ifr_flags = IFF_PROMISC }; dev->data->promiscuous = 0; - tap_ioctl(pmd, SIOCSIFFLAGS, &ifr, 0); + tap_ioctl(pmd, SIOCSIFFLAGS, &ifr, 0, LOCAL_AND_REMOTE); if (pmd->remote_if_index) tap_flow_implicit_destroy(pmd, TAP_REMOTE_PROMISC); } @@ -748,7 +817,7 @@ tap_allmulti_enable(struct rte_eth_dev *dev) struct ifreq ifr = { .ifr_flags = IFF_ALLMULTI }; dev->data->all_multicast = 1; - tap_ioctl(pmd, SIOCSIFFLAGS, &ifr, 1); + tap_ioctl(pmd, SIOCSIFFLAGS, &ifr, 1, LOCAL_AND_REMOTE); if (pmd->remote_if_index) tap_flow_implicit_create(pmd, TAP_REMOTE_ALLMULTI); } @@ -760,16 +829,16 @@ tap_allmulti_disable(struct rte_eth_dev *dev) struct ifreq ifr = { .ifr_flags = IFF_ALLMULTI }; dev->data->all_multicast = 0; - tap_ioctl(pmd, SIOCSIFFLAGS, &ifr, 0); + tap_ioctl(pmd, SIOCSIFFLAGS, &ifr, 0, LOCAL_AND_REMOTE); if (pmd->remote_if_index) tap_flow_implicit_destroy(pmd, TAP_REMOTE_ALLMULTI); } - static void tap_mac_set(struct rte_eth_dev *dev, struct ether_addr *mac_addr) { struct pmd_internals *pmd = dev->data->dev_private; + enum ioctl_mode mode = LOCAL_ONLY; struct ifreq ifr; if (is_zero_ether_addr(mac_addr)) { @@ -777,12 +846,36 @@ tap_mac_set(struct rte_eth_dev *dev, struct ether_addr *mac_addr) dev->data->name); return; } - + /* Check the actual current MAC address on the tap netdevice */ + if (tap_ioctl(pmd, SIOCGIFHWADDR, &ifr, 0, LOCAL_ONLY) < 0) + return; + if (is_same_ether_addr((struct ether_addr *)&ifr.ifr_hwaddr.sa_data, + mac_addr)) + return; + /* Check the current MAC address on the remote */ + if (tap_ioctl(pmd, SIOCGIFHWADDR, &ifr, 0, REMOTE_ONLY) < 0) + return; + if (!is_same_ether_addr((struct ether_addr *)&ifr.ifr_hwaddr.sa_data, + mac_addr)) + mode = LOCAL_AND_REMOTE; ifr.ifr_hwaddr.sa_family = AF_LOCAL; rte_memcpy(ifr.ifr_hwaddr.sa_data, mac_addr, ETHER_ADDR_LEN); - if (tap_ioctl(pmd, SIOCSIFHWADDR, &ifr, 1) < 0) + if (tap_ioctl(pmd, SIOCSIFHWADDR, &ifr, 1, mode) < 0) return; rte_memcpy(&pmd->eth_addr, mac_addr, ETHER_ADDR_LEN); + if (pmd->remote_if_index) { + /* Replace MAC redirection rule after a MAC change */ + if (tap_flow_implicit_destroy(pmd, TAP_REMOTE_LOCAL_MAC) < 0) { + RTE_LOG(ERR, PMD, + "%s: Couldn't delete MAC redirection rule\n", + dev->data->name); + return; + } + if (tap_flow_implicit_create(pmd, TAP_REMOTE_LOCAL_MAC) < 0) + RTE_LOG(ERR, PMD, + "%s: Couldn't add MAC redirection rule\n", + dev->data->name); + } } static int @@ -793,29 +886,16 @@ tap_setup_queue(struct rte_eth_dev *dev, struct pmd_internals *pmd = dev->data->dev_private; struct rx_queue *rx = &internals->rxq[qid]; struct tx_queue *tx = &internals->txq[qid]; - int fd; + int fd = rx->fd == -1 ? tx->fd : rx->fd; - fd = rx->fd; - if (fd < 0) { - fd = tx->fd; + if (fd == -1) { + RTE_LOG(INFO, PMD, "Add queue to TAP %s for qid %d\n", + pmd->name, qid); + fd = tun_alloc(pmd); if (fd < 0) { - RTE_LOG(INFO, PMD, "Add queue to TAP %s for qid %d\n", - pmd->name, qid); - fd = tun_alloc(pmd, qid); - if (fd < 0) { - RTE_LOG(ERR, PMD, "tun_alloc(%s, %d) failed\n", - pmd->name, qid); - return -1; - } - if (qid == 0) { - struct ifreq ifr; - - ifr.ifr_mtu = dev->data->mtu; - if (tap_ioctl(pmd, SIOCSIFMTU, &ifr, 1) < 0) { - close(fd); - return -1; - } - } + RTE_LOG(ERR, PMD, "%s: tun_alloc() failed.\n", + pmd->name); + return -1; } } @@ -827,26 +907,6 @@ tap_setup_queue(struct rte_eth_dev *dev, return fd; } -static int -rx_setup_queue(struct rte_eth_dev *dev, - struct pmd_internals *internals, - uint16_t qid) -{ - dev->data->rx_queues[qid] = &internals->rxq[qid]; - - return tap_setup_queue(dev, internals, qid); -} - -static int -tx_setup_queue(struct rte_eth_dev *dev, - struct pmd_internals *internals, - uint16_t qid) -{ - dev->data->tx_queues[qid] = &internals->txq[qid]; - - return tap_setup_queue(dev, internals, qid); -} - static int tap_rx_queue_setup(struct rte_eth_dev *dev, uint16_t rx_queue_id, @@ -858,7 +918,9 @@ tap_rx_queue_setup(struct rte_eth_dev *dev, struct pmd_internals *internals = dev->data->dev_private; struct rx_queue *rxq = &internals->rxq[rx_queue_id]; struct rte_mbuf **tmp = &rxq->pool; - struct iovec (*iovecs)[nb_rx_desc + 1]; + long iov_max = sysconf(_SC_IOV_MAX); + uint16_t nb_desc = RTE_MIN(nb_rx_desc, iov_max - 1); + struct iovec (*iovecs)[nb_desc + 1]; int data_off = RTE_PKTMBUF_HEADROOM; int ret = 0; int fd; @@ -874,18 +936,19 @@ tap_rx_queue_setup(struct rte_eth_dev *dev, rxq->mp = mp; rxq->trigger_seen = 1; /* force initial burst */ rxq->in_port = dev->data->port_id; - rxq->nb_rx_desc = nb_rx_desc; + rxq->nb_rx_desc = nb_desc; iovecs = rte_zmalloc_socket(dev->data->name, sizeof(*iovecs), 0, socket_id); if (!iovecs) { RTE_LOG(WARNING, PMD, "%s: Couldn't allocate %d RX descriptors\n", - dev->data->name, nb_rx_desc); + dev->data->name, nb_desc); return -ENOMEM; } rxq->iovecs = iovecs; - fd = rx_setup_queue(dev, internals, rx_queue_id); + dev->data->rx_queues[rx_queue_id] = rxq; + fd = tap_setup_queue(dev, internals, rx_queue_id); if (fd == -1) { ret = fd; goto error; @@ -894,7 +957,7 @@ tap_rx_queue_setup(struct rte_eth_dev *dev, (*rxq->iovecs)[0].iov_len = sizeof(struct tun_pi); (*rxq->iovecs)[0].iov_base = &rxq->pi; - for (i = 1; i <= nb_rx_desc; i++) { + for (i = 1; i <= nb_desc; i++) { *tmp = rte_pktmbuf_alloc(rxq->mp); if (!*tmp) { RTE_LOG(WARNING, PMD, @@ -936,7 +999,8 @@ tap_tx_queue_setup(struct rte_eth_dev *dev, if (tx_queue_id >= internals->nb_queues) return -1; - ret = tx_setup_queue(dev, internals, tx_queue_id); + dev->data->tx_queues[tx_queue_id] = &internals->txq[tx_queue_id]; + ret = tap_setup_queue(dev, internals, tx_queue_id); if (ret == -1) return -1; @@ -953,7 +1017,7 @@ tap_mtu_set(struct rte_eth_dev *dev, uint16_t mtu) struct ifreq ifr = { .ifr_mtu = mtu }; int err = 0; - err = tap_ioctl(pmd, SIOCSIFMTU, &ifr, 1); + err = tap_ioctl(pmd, SIOCSIFMTU, &ifr, 1, LOCAL_AND_REMOTE); if (!err) dev->data->mtu = mtu; @@ -987,8 +1051,7 @@ tap_nl_msg_handler(struct nlmsghdr *nh, void *arg) } static void -tap_dev_intr_handler(struct rte_intr_handle *intr_handle __rte_unused, - void *cb_arg) +tap_dev_intr_handler(void *cb_arg) { struct rte_eth_dev *dev = cb_arg; struct pmd_internals *pmd = dev->data->dev_private; @@ -1116,12 +1179,14 @@ tap_kernel_support(struct pmd_internals *pmd) } static int -eth_dev_tap_create(const char *name, char *tap_name, char *remote_iface) +eth_dev_tap_create(struct rte_vdev_device *vdev, char *tap_name, + char *remote_iface, int fixed_mac_type) { int numa_node = rte_socket_id(); - struct rte_eth_dev *dev = NULL; - struct pmd_internals *pmd = NULL; - struct rte_eth_dev_data *data = NULL; + struct rte_eth_dev *dev; + struct pmd_internals *pmd; + struct rte_eth_dev_data *data; + struct ifreq ifr; int i; RTE_LOG(DEBUG, PMD, " TAP device on numa %u\n", rte_socket_id()); @@ -1132,22 +1197,14 @@ eth_dev_tap_create(const char *name, char *tap_name, char *remote_iface) goto error_exit; } - pmd = rte_zmalloc_socket(tap_name, sizeof(*pmd), 0, numa_node); - if (!pmd) { - RTE_LOG(ERR, PMD, "TAP Unable to allocate internal struct\n"); - goto error_exit; - } - - /* name in allocation and data->name must be consistent */ - snprintf(data->name, sizeof(data->name), "%s", name); - dev = rte_eth_dev_allocate(name); + dev = rte_eth_vdev_allocate(vdev, sizeof(*pmd)); if (!dev) { RTE_LOG(ERR, PMD, "TAP Unable to allocate device struct\n"); goto error_exit; } + pmd = dev->data->dev_private; snprintf(pmd->name, sizeof(pmd->name), "%s", tap_name); - pmd->nb_queues = RTE_PMD_TAP_MAX_QUEUES; pmd->ioctl_sock = socket(AF_INET, SOCK_DGRAM, 0); @@ -1159,12 +1216,9 @@ eth_dev_tap_create(const char *name, char *tap_name, char *remote_iface) } /* Setup some default values */ + rte_memcpy(data, dev->data, sizeof(*data)); data->dev_private = pmd; - data->port_id = dev->data->port_id; - data->mtu = dev->data->mtu; data->dev_flags = RTE_ETH_DEV_DETACHABLE | RTE_ETH_DEV_INTR_LSC; - data->kdrv = RTE_KDRV_NONE; - data->drv_name = pmd_tap_drv.driver.name; data->numa_node = numa_node; data->dev_link = pmd_link; @@ -1174,7 +1228,6 @@ eth_dev_tap_create(const char *name, char *tap_name, char *remote_iface) dev->data = data; dev->dev_ops = &ops; - dev->driver = NULL; dev->rx_pkt_burst = pmd_rx_burst; dev->tx_pkt_burst = pmd_tx_burst; @@ -1187,35 +1240,146 @@ eth_dev_tap_create(const char *name, char *tap_name, char *remote_iface) pmd->txq[i].fd = -1; } + if (fixed_mac_type) { + /* fixed mac = 00:64:74:61:70: */ + static int iface_idx; + char mac[ETHER_ADDR_LEN] = "\0dtap"; + + mac[ETHER_ADDR_LEN - 1] = iface_idx++; + rte_memcpy(&pmd->eth_addr, mac, ETHER_ADDR_LEN); + } else { + eth_random_addr((uint8_t *)&pmd->eth_addr); + } + + /* Immediately create the netdevice (this will create the 1st queue). */ + if (tap_setup_queue(dev, pmd, 0) == -1) + goto error_exit; + + ifr.ifr_mtu = dev->data->mtu; + if (tap_ioctl(pmd, SIOCSIFMTU, &ifr, 1, LOCAL_AND_REMOTE) < 0) + goto error_exit; + + memset(&ifr, 0, sizeof(struct ifreq)); + ifr.ifr_hwaddr.sa_family = AF_LOCAL; + rte_memcpy(ifr.ifr_hwaddr.sa_data, &pmd->eth_addr, ETHER_ADDR_LEN); + if (tap_ioctl(pmd, SIOCSIFHWADDR, &ifr, 0, LOCAL_ONLY) < 0) + goto error_exit; + tap_kernel_support(pmd); - if (!pmd->flower_support) - return 0; - LIST_INIT(&pmd->flows); + if (!pmd->flower_support) { + if (remote_iface[0]) { + RTE_LOG(ERR, PMD, + "%s: kernel does not support TC rules, required for remote feature.", + pmd->name); + goto error_exit; + } else { + RTE_LOG(INFO, PMD, + "%s: kernel too old for Flow API support.\n", + pmd->name); + return 0; + } + } + /* - * If no netlink socket can be created, then it will fail when - * creating/destroying flow rules. + * Set up everything related to rte_flow: + * - netlink socket + * - tap / remote if_index + * - mandatory QDISCs + * - rte_flow actual/implicit lists + * - implicit rules */ pmd->nlsk_fd = nl_init(0); + if (pmd->nlsk_fd == -1) { + RTE_LOG(WARNING, PMD, "%s: failed to create netlink socket.", + pmd->name); + goto disable_rte_flow; + } + pmd->if_index = if_nametoindex(pmd->name); + if (!pmd->if_index) { + RTE_LOG(ERR, PMD, "%s: failed to get if_index.", pmd->name); + goto disable_rte_flow; + } + if (qdisc_create_multiq(pmd->nlsk_fd, pmd->if_index) < 0) { + RTE_LOG(ERR, PMD, "%s: failed to create multiq qdisc.", + pmd->name); + goto disable_rte_flow; + } + if (qdisc_create_ingress(pmd->nlsk_fd, pmd->if_index) < 0) { + RTE_LOG(ERR, PMD, "%s: failed to create ingress qdisc.", + pmd->name); + goto disable_rte_flow; + } + LIST_INIT(&pmd->flows); + if (strlen(remote_iface)) { pmd->remote_if_index = if_nametoindex(remote_iface); + if (!pmd->remote_if_index) { + RTE_LOG(ERR, PMD, "%s: failed to get %s if_index.", + pmd->name, remote_iface); + goto error_remote; + } snprintf(pmd->remote_iface, RTE_ETH_NAME_MAX_LEN, "%s", remote_iface); - if (!pmd->remote_if_index) - RTE_LOG(ERR, PMD, "Could not find %s ifindex: " - "remote interface will remain unconfigured\n", - remote_iface); + if (tap_ioctl(pmd, SIOCGIFHWADDR, &ifr, 0, REMOTE_ONLY) < 0) { + RTE_LOG(ERR, PMD, "%s: failed to get %s MAC address.", + pmd->name, pmd->remote_iface); + goto error_remote; + } + rte_memcpy(&pmd->eth_addr, ifr.ifr_hwaddr.sa_data, + ETHER_ADDR_LEN); + /* The desired MAC is already in ifreq after SIOCGIFHWADDR. */ + if (tap_ioctl(pmd, SIOCSIFHWADDR, &ifr, 0, LOCAL_ONLY) < 0) { + RTE_LOG(ERR, PMD, "%s: failed to get %s MAC address.", + pmd->name, remote_iface); + goto error_remote; + } + + /* + * Flush usually returns negative value because it tries to + * delete every QDISC (and on a running device, one QDISC at + * least is needed). Ignore negative return value. + */ + qdisc_flush(pmd->nlsk_fd, pmd->remote_if_index); + if (qdisc_create_ingress(pmd->nlsk_fd, + pmd->remote_if_index) < 0) { + RTE_LOG(ERR, PMD, "%s: failed to create ingress qdisc.", + pmd->remote_iface); + goto error_remote; + } + LIST_INIT(&pmd->implicit_flows); + if (tap_flow_implicit_create(pmd, TAP_REMOTE_TX) < 0 || + tap_flow_implicit_create(pmd, TAP_REMOTE_LOCAL_MAC) < 0 || + tap_flow_implicit_create(pmd, TAP_REMOTE_BROADCAST) < 0 || + tap_flow_implicit_create(pmd, TAP_REMOTE_BROADCASTV6) < 0) { + RTE_LOG(ERR, PMD, + "%s: failed to create implicit rules.", + pmd->name); + goto error_remote; + } } return 0; -error_exit: - RTE_LOG(DEBUG, PMD, "TAP Unable to initialize %s\n", name); +disable_rte_flow: + RTE_LOG(ERR, PMD, " Disabling rte flow support: %s(%d)\n", + strerror(errno), errno); + if (strlen(remote_iface)) { + RTE_LOG(ERR, PMD, "Remote feature requires flow support.\n"); + goto error_exit; + } + pmd->flower_support = 0; + return 0; - rte_free(data); - rte_free(pmd); +error_remote: + RTE_LOG(ERR, PMD, " Can't set up remote feature: %s(%d)\n", + strerror(errno), errno); + tap_flow_implicit_flush(pmd, NULL); - rte_eth_dev_release_port(dev); +error_exit: + RTE_LOG(ERR, PMD, "TAP Unable to initialize %s\n", + rte_vdev_device_name(vdev)); + rte_free(data); return -EINVAL; } @@ -1258,16 +1422,32 @@ set_remote_iface(const char *key __rte_unused, return 0; } +static int +set_mac_type(const char *key __rte_unused, + const char *value, + void *extra_args) +{ + if (value && + !strncasecmp(ETH_TAP_MAC_FIXED, value, strlen(ETH_TAP_MAC_FIXED))) + *(int *)extra_args = 1; + return 0; +} + /* Open a TAP interface device. */ static int -rte_pmd_tap_probe(const char *name, const char *params) +rte_pmd_tap_probe(struct rte_vdev_device *dev) { + const char *name, *params; int ret; struct rte_kvargs *kvlist = NULL; int speed; char tap_name[RTE_ETH_NAME_MAX_LEN]; char remote_iface[RTE_ETH_NAME_MAX_LEN]; + int fixed_mac_type = 0; + + name = rte_vdev_device_name(dev); + params = rte_vdev_device_args(dev); speed = ETH_SPEED_NUM_10G; snprintf(tap_name, sizeof(tap_name), "%s%d", @@ -1275,7 +1455,7 @@ rte_pmd_tap_probe(const char *name, const char *params) memset(remote_iface, 0, RTE_ETH_NAME_MAX_LEN); if (params && (params[0] != '\0')) { - RTE_LOG(DEBUG, PMD, "paramaters (%s)\n", params); + RTE_LOG(DEBUG, PMD, "parameters (%s)\n", params); kvlist = rte_kvargs_parse(params, valid_arguments); if (kvlist) { @@ -1305,6 +1485,15 @@ rte_pmd_tap_probe(const char *name, const char *params) if (ret == -1) goto leave; } + + if (rte_kvargs_count(kvlist, ETH_TAP_MAC_ARG) == 1) { + ret = rte_kvargs_process(kvlist, + ETH_TAP_MAC_ARG, + &set_mac_type, + &fixed_mac_type); + if (ret == -1) + goto leave; + } } } pmd_link.link_speed = speed; @@ -1312,7 +1501,7 @@ rte_pmd_tap_probe(const char *name, const char *params) RTE_LOG(NOTICE, PMD, "Initializing pmd_tap for %s as %s\n", name, tap_name); - ret = eth_dev_tap_create(name, tap_name, remote_iface); + ret = eth_dev_tap_create(dev, tap_name, remote_iface, fixed_mac_type); leave: if (ret == -1) { @@ -1328,7 +1517,7 @@ leave: /* detach a TAP device. */ static int -rte_pmd_tap_remove(const char *name) +rte_pmd_tap_remove(struct rte_vdev_device *dev) { struct rte_eth_dev *eth_dev = NULL; struct pmd_internals *internals; @@ -1338,7 +1527,7 @@ rte_pmd_tap_remove(const char *name) rte_socket_id()); /* find the ethdev entry */ - eth_dev = rte_eth_dev_allocated(name); + eth_dev = rte_eth_dev_allocated(rte_vdev_device_name(dev)); if (!eth_dev) return 0; @@ -1367,4 +1556,8 @@ static struct rte_vdev_driver pmd_tap_drv = { }; RTE_PMD_REGISTER_VDEV(net_tap, pmd_tap_drv); RTE_PMD_REGISTER_ALIAS(net_tap, eth_tap); -RTE_PMD_REGISTER_PARAM_STRING(net_tap, "iface=,speed=N"); +RTE_PMD_REGISTER_PARAM_STRING(net_tap, + ETH_TAP_IFACE_ARG "= " + ETH_TAP_SPEED_ARG "= " + ETH_TAP_MAC_ARG "=" ETH_TAP_MAC_FIXED " " + ETH_TAP_REMOTE_ARG "=");