X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fbonding%2Frte_eth_bond_pmd.c;h=057b1ada54c7d72d1a143ce5b1fb1b44a24e740a;hb=f30e69b41f94;hp=0f5ab09e38b5a7138aa9c2826fcf03c1dedf8b7c;hpb=617d1ac2fd22e9a82d305c0607853f4ee0f7277b;p=dpdk.git diff --git a/drivers/net/bonding/rte_eth_bond_pmd.c b/drivers/net/bonding/rte_eth_bond_pmd.c index 0f5ab09e38..057b1ada54 100644 --- a/drivers/net/bonding/rte_eth_bond_pmd.c +++ b/drivers/net/bonding/rte_eth_bond_pmd.c @@ -2,6 +2,7 @@ * Copyright(c) 2010-2017 Intel Corporation */ #include +#include #include #include @@ -20,8 +21,8 @@ #include #include "rte_eth_bond.h" -#include "rte_eth_bond_private.h" -#include "rte_eth_bond_8023ad_private.h" +#include "eth_bond_private.h" +#include "eth_bond_8023ad_private.h" #define REORDER_PERIOD_MS 10 #define DEFAULT_POLLING_INTERVAL_10_MS (10) @@ -33,21 +34,22 @@ static uint64_t tlb_last_obytets[RTE_MAX_ETHPORTS]; static inline size_t -get_vlan_offset(struct ether_hdr *eth_hdr, uint16_t *proto) +get_vlan_offset(struct rte_ether_hdr *eth_hdr, uint16_t *proto) { size_t vlan_offset = 0; - if (rte_cpu_to_be_16(ETHER_TYPE_VLAN) == *proto || - rte_cpu_to_be_16(ETHER_TYPE_QINQ) == *proto) { - struct vlan_hdr *vlan_hdr = (struct vlan_hdr *)(eth_hdr + 1); + if (rte_cpu_to_be_16(RTE_ETHER_TYPE_VLAN) == *proto || + rte_cpu_to_be_16(RTE_ETHER_TYPE_QINQ) == *proto) { + struct rte_vlan_hdr *vlan_hdr = + (struct rte_vlan_hdr *)(eth_hdr + 1); - vlan_offset = sizeof(struct vlan_hdr); + vlan_offset = sizeof(struct rte_vlan_hdr); *proto = vlan_hdr->eth_proto; - if (rte_cpu_to_be_16(ETHER_TYPE_VLAN) == *proto) { + if (rte_cpu_to_be_16(RTE_ETHER_TYPE_VLAN) == *proto) { vlan_hdr = vlan_hdr + 1; *proto = vlan_hdr->eth_proto; - vlan_offset += sizeof(struct vlan_hdr); + vlan_offset += sizeof(struct rte_vlan_hdr); } } return vlan_offset; @@ -58,28 +60,34 @@ bond_ethdev_rx_burst(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts) { struct bond_dev_private *internals; - uint16_t num_rx_slave = 0; uint16_t num_rx_total = 0; - + uint16_t slave_count; + uint16_t active_slave; int i; /* Cast to structure, containing bonded device's port id and queue id */ struct bond_rx_queue *bd_rx_q = (struct bond_rx_queue *)queue; - internals = bd_rx_q->dev_private; + slave_count = internals->active_slave_count; + active_slave = bd_rx_q->active_slave; + for (i = 0; i < slave_count && nb_pkts; i++) { + uint16_t num_rx_slave; - for (i = 0; i < internals->active_slave_count && nb_pkts; i++) { /* Offset of pointer to *bufs increases as packets are received * from other slaves */ - num_rx_slave = rte_eth_rx_burst(internals->active_slaves[i], - bd_rx_q->queue_id, bufs + num_rx_total, nb_pkts); - if (num_rx_slave) { - num_rx_total += num_rx_slave; - nb_pkts -= num_rx_slave; - } - } - + num_rx_slave = + rte_eth_rx_burst(internals->active_slaves[active_slave], + bd_rx_q->queue_id, + bufs + num_rx_total, nb_pkts); + num_rx_total += num_rx_slave; + nb_pkts -= num_rx_slave; + if (++active_slave == slave_count) + active_slave = 0; + } + + if (++bd_rx_q->active_slave >= slave_count) + bd_rx_q->active_slave = 0; return num_rx_total; } @@ -101,7 +109,8 @@ bond_ethdev_rx_burst_active_backup(void *queue, struct rte_mbuf **bufs, static inline uint8_t is_lacp_packets(uint16_t ethertype, uint8_t subtype, struct rte_mbuf *mbuf) { - const uint16_t ether_type_slow_be = rte_be_to_cpu_16(ETHER_TYPE_SLOW); + const uint16_t ether_type_slow_be = + rte_be_to_cpu_16(RTE_ETHER_TYPE_SLOW); return !((mbuf->ol_flags & PKT_RX_VLAN) ? mbuf->vlan_tci : 0) && (ethertype == ether_type_slow_be && @@ -115,7 +124,7 @@ is_lacp_packets(uint16_t ethertype, uint8_t subtype, struct rte_mbuf *mbuf) static struct rte_flow_item_eth flow_item_eth_type_8023ad = { .dst.addr_bytes = { 0 }, .src.addr_bytes = { 0 }, - .type = RTE_BE16(ETHER_TYPE_SLOW), + .type = RTE_BE16(RTE_ETHER_TYPE_SLOW), }; static struct rte_flow_item_eth flow_item_eth_mask_type_8023ad = { @@ -152,8 +161,7 @@ bond_ethdev_8023ad_flow_verify(struct rte_eth_dev *bond_dev, uint16_t slave_port) { struct rte_eth_dev_info slave_info; struct rte_flow_error error; - struct bond_dev_private *internals = (struct bond_dev_private *) - (bond_dev->data->dev_private); + struct bond_dev_private *internals = bond_dev->data->dev_private; const struct rte_flow_action_queue lacp_queue_conf = { .index = 0, @@ -178,7 +186,15 @@ bond_ethdev_8023ad_flow_verify(struct rte_eth_dev *bond_dev, return -1; } - rte_eth_dev_info_get(slave_port, &slave_info); + ret = rte_eth_dev_info_get(slave_port, &slave_info); + if (ret != 0) { + RTE_BOND_LOG(ERR, + "%s: Error during getting device (port %u) info: %s\n", + __func__, slave_port, strerror(-ret)); + + return ret; + } + if (slave_info.max_rx_queues < bond_dev->data->nb_rx_queues || slave_info.max_tx_queues < bond_dev->data->nb_tx_queues) { RTE_BOND_LOG(ERR, @@ -193,14 +209,22 @@ bond_ethdev_8023ad_flow_verify(struct rte_eth_dev *bond_dev, int bond_8023ad_slow_pkt_hw_filter_supported(uint16_t port_id) { struct rte_eth_dev *bond_dev = &rte_eth_devices[port_id]; - struct bond_dev_private *internals = (struct bond_dev_private *) - (bond_dev->data->dev_private); + struct bond_dev_private *internals = bond_dev->data->dev_private; struct rte_eth_dev_info bond_info; uint16_t idx; + int ret; /* Verify if all slaves in bonding supports flow director and */ if (internals->slave_count > 0) { - rte_eth_dev_info_get(bond_dev->data->port_id, &bond_info); + ret = rte_eth_dev_info_get(bond_dev->data->port_id, &bond_info); + if (ret != 0) { + RTE_BOND_LOG(ERR, + "%s: Error during getting device (port %u) info: %s\n", + __func__, bond_dev->data->port_id, + strerror(-ret)); + + return ret; + } internals->mode4.dedicated_queues.rx_qid = bond_info.nb_rx_queues; internals->mode4.dedicated_queues.tx_qid = bond_info.nb_tx_queues; @@ -219,9 +243,7 @@ int bond_ethdev_8023ad_flow_set(struct rte_eth_dev *bond_dev, uint16_t slave_port) { struct rte_flow_error error; - struct bond_dev_private *internals = (struct bond_dev_private *) - (bond_dev->data->dev_private); - + struct bond_dev_private *internals = bond_dev->data->dev_private; struct rte_flow_action_queue lacp_queue_conf = { .index = internals->mode4.dedicated_queues.rx_qid, }; @@ -249,165 +271,46 @@ bond_ethdev_8023ad_flow_set(struct rte_eth_dev *bond_dev, uint16_t slave_port) { return 0; } -static uint16_t -bond_ethdev_rx_burst_8023ad_fast_queue(void *queue, struct rte_mbuf **bufs, - uint16_t nb_pkts) -{ - struct bond_rx_queue *bd_rx_q = (struct bond_rx_queue *)queue; - struct bond_dev_private *internals = bd_rx_q->dev_private; - uint16_t num_rx_total = 0; /* Total number of received packets */ - uint16_t slaves[RTE_MAX_ETHPORTS]; - uint16_t slave_count; - - uint16_t i, idx; - - /* Copy slave list to protect against slave up/down changes during tx - * bursting */ - slave_count = internals->active_slave_count; - memcpy(slaves, internals->active_slaves, - sizeof(internals->active_slaves[0]) * slave_count); - - for (i = 0, idx = internals->active_slave; - i < slave_count && num_rx_total < nb_pkts; i++, idx++) { - idx = idx % slave_count; - - /* Read packets from this slave */ - num_rx_total += rte_eth_rx_burst(slaves[idx], bd_rx_q->queue_id, - &bufs[num_rx_total], nb_pkts - num_rx_total); - } - - internals->active_slave = idx; - - return num_rx_total; -} - -static uint16_t -bond_ethdev_tx_burst_8023ad_fast_queue(void *queue, struct rte_mbuf **bufs, - uint16_t nb_bufs) -{ - struct bond_tx_queue *bd_tx_q = (struct bond_tx_queue *)queue; - struct bond_dev_private *internals = bd_tx_q->dev_private; - - uint16_t slave_port_ids[RTE_MAX_ETHPORTS]; - uint16_t slave_count; - - uint16_t dist_slave_port_ids[RTE_MAX_ETHPORTS]; - uint16_t dist_slave_count; - - /* 2-D array to sort mbufs for transmission on each slave into */ - struct rte_mbuf *slave_bufs[RTE_MAX_ETHPORTS][nb_bufs]; - /* Number of mbufs for transmission on each slave */ - uint16_t slave_nb_bufs[RTE_MAX_ETHPORTS] = { 0 }; - /* Mapping array generated by hash function to map mbufs to slaves */ - uint16_t bufs_slave_port_idxs[RTE_MAX_ETHPORTS] = { 0 }; - - uint16_t slave_tx_count; - uint16_t total_tx_count = 0, total_tx_fail_count = 0; - - uint16_t i; - - if (unlikely(nb_bufs == 0)) - return 0; - - /* Copy slave list to protect against slave up/down changes during tx - * bursting */ - slave_count = internals->active_slave_count; - if (unlikely(slave_count < 1)) - return 0; - - memcpy(slave_port_ids, internals->active_slaves, - sizeof(slave_port_ids[0]) * slave_count); - - - dist_slave_count = 0; - for (i = 0; i < slave_count; i++) { - struct port *port = &mode_8023ad_ports[slave_port_ids[i]]; - - if (ACTOR_STATE(port, DISTRIBUTING)) - dist_slave_port_ids[dist_slave_count++] = - slave_port_ids[i]; - } - - if (unlikely(dist_slave_count < 1)) - return 0; - - /* - * Populate slaves mbuf with the packets which are to be sent on it - * selecting output slave using hash based on xmit policy - */ - internals->burst_xmit_hash(bufs, nb_bufs, dist_slave_count, - bufs_slave_port_idxs); - - for (i = 0; i < nb_bufs; i++) { - /* Populate slave mbuf arrays with mbufs for that slave. */ - uint8_t slave_idx = bufs_slave_port_idxs[i]; - - slave_bufs[slave_idx][slave_nb_bufs[slave_idx]++] = bufs[i]; - } - - - /* Send packet burst on each slave device */ - for (i = 0; i < dist_slave_count; i++) { - if (slave_nb_bufs[i] == 0) - continue; - - slave_tx_count = rte_eth_tx_burst(dist_slave_port_ids[i], - bd_tx_q->queue_id, slave_bufs[i], - slave_nb_bufs[i]); - - total_tx_count += slave_tx_count; - - /* If tx burst fails move packets to end of bufs */ - if (unlikely(slave_tx_count < slave_nb_bufs[i])) { - int slave_tx_fail_count = slave_nb_bufs[i] - - slave_tx_count; - total_tx_fail_count += slave_tx_fail_count; - memcpy(&bufs[nb_bufs - total_tx_fail_count], - &slave_bufs[i][slave_tx_count], - slave_tx_fail_count * sizeof(bufs[0])); - } - } - - return total_tx_count; -} - - -static uint16_t -bond_ethdev_rx_burst_8023ad(void *queue, struct rte_mbuf **bufs, - uint16_t nb_pkts) +static inline uint16_t +rx_burst_8023ad(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts, + bool dedicated_rxq) { /* Cast to structure, containing bonded device's port id and queue id */ struct bond_rx_queue *bd_rx_q = (struct bond_rx_queue *)queue; struct bond_dev_private *internals = bd_rx_q->dev_private; - struct ether_addr bond_mac; - - struct ether_hdr *hdr; + struct rte_eth_dev *bonded_eth_dev = + &rte_eth_devices[internals->port_id]; + struct rte_ether_addr *bond_mac = bonded_eth_dev->data->mac_addrs; + struct rte_ether_hdr *hdr; - const uint16_t ether_type_slow_be = rte_be_to_cpu_16(ETHER_TYPE_SLOW); + const uint16_t ether_type_slow_be = + rte_be_to_cpu_16(RTE_ETHER_TYPE_SLOW); uint16_t num_rx_total = 0; /* Total number of received packets */ uint16_t slaves[RTE_MAX_ETHPORTS]; uint16_t slave_count, idx; uint8_t collecting; /* current slave collecting status */ - const uint8_t promisc = internals->promiscuous_en; - uint8_t i, j, k; + const uint8_t promisc = rte_eth_promiscuous_get(internals->port_id); + const uint8_t allmulti = rte_eth_allmulticast_get(internals->port_id); uint8_t subtype; + uint16_t i; + uint16_t j; + uint16_t k; - rte_eth_macaddr_get(internals->port_id, &bond_mac); /* Copy slave list to protect against slave up/down changes during tx * bursting */ slave_count = internals->active_slave_count; memcpy(slaves, internals->active_slaves, sizeof(internals->active_slaves[0]) * slave_count); - idx = internals->active_slave; + idx = bd_rx_q->active_slave; if (idx >= slave_count) { - internals->active_slave = 0; + bd_rx_q->active_slave = 0; idx = 0; } for (i = 0; i < slave_count && num_rx_total < nb_pkts; i++) { j = num_rx_total; - collecting = ACTOR_STATE(&mode_8023ad_ports[slaves[idx]], + collecting = ACTOR_STATE(&bond_mode_8023ad_ports[slaves[idx]], COLLECTING); /* Read packets from this slave */ @@ -419,26 +322,31 @@ bond_ethdev_rx_burst_8023ad(void *queue, struct rte_mbuf **bufs, /* Handle slow protocol packets. */ while (j < num_rx_total) { - - /* If packet is not pure L2 and is known, skip it */ - if ((bufs[j]->packet_type & ~RTE_PTYPE_L2_ETHER) != 0) { - j++; - continue; - } - if (j + 3 < num_rx_total) rte_prefetch0(rte_pktmbuf_mtod(bufs[j + 3], void *)); - hdr = rte_pktmbuf_mtod(bufs[j], struct ether_hdr *); + hdr = rte_pktmbuf_mtod(bufs[j], struct rte_ether_hdr *); subtype = ((struct slow_protocol_frame *)hdr)->slow_protocol.subtype; - /* Remove packet from array if it is slow packet or slave is not - * in collecting state or bonding interface is not in promiscuous - * mode and packet address does not match. */ - if (unlikely(is_lacp_packets(hdr->ether_type, subtype, bufs[j]) || - !collecting || (!promisc && - !is_multicast_ether_addr(&hdr->d_addr) && - !is_same_ether_addr(&bond_mac, &hdr->d_addr)))) { + /* Remove packet from array if: + * - it is slow packet but no dedicated rxq is present, + * - slave is not in collecting state, + * - bonding interface is not in promiscuous mode: + * - packet is unicast and address does not match, + * - packet is multicast and bonding interface + * is not in allmulti, + */ + if (unlikely( + (!dedicated_rxq && + is_lacp_packets(hdr->ether_type, subtype, + bufs[j])) || + !collecting || + (!promisc && + ((rte_is_unicast_ether_addr(&hdr->d_addr) && + !rte_is_same_ether_addr(bond_mac, + &hdr->d_addr)) || + (!allmulti && + rte_is_multicast_ether_addr(&hdr->d_addr)))))) { if (hdr->ether_type == ether_type_slow_be) { bond_mode_8023ad_handle_slow_pkt( @@ -459,10 +367,26 @@ bond_ethdev_rx_burst_8023ad(void *queue, struct rte_mbuf **bufs, idx = 0; } - internals->active_slave = idx; + if (++bd_rx_q->active_slave >= slave_count) + bd_rx_q->active_slave = 0; + return num_rx_total; } +static uint16_t +bond_ethdev_rx_burst_8023ad(void *queue, struct rte_mbuf **bufs, + uint16_t nb_pkts) +{ + return rx_burst_8023ad(queue, bufs, nb_pkts, false); +} + +static uint16_t +bond_ethdev_rx_burst_8023ad_fast_queue(void *queue, struct rte_mbuf **bufs, + uint16_t nb_pkts) +{ + return rx_burst_8023ad(queue, bufs, nb_pkts, true); +} + #if defined(RTE_LIBRTE_BOND_DEBUG_ALB) || defined(RTE_LIBRTE_BOND_DEBUG_ALB_L1) uint32_t burstnumberRX; uint32_t burstnumberTX; @@ -470,35 +394,31 @@ uint32_t burstnumberTX; #ifdef RTE_LIBRTE_BOND_DEBUG_ALB static void -arp_op_name(uint16_t arp_op, char *buf) +arp_op_name(uint16_t arp_op, char *buf, size_t buf_len) { switch (arp_op) { - case ARP_OP_REQUEST: - snprintf(buf, sizeof("ARP Request"), "%s", "ARP Request"); + case RTE_ARP_OP_REQUEST: + strlcpy(buf, "ARP Request", buf_len); return; - case ARP_OP_REPLY: - snprintf(buf, sizeof("ARP Reply"), "%s", "ARP Reply"); + case RTE_ARP_OP_REPLY: + strlcpy(buf, "ARP Reply", buf_len); return; - case ARP_OP_REVREQUEST: - snprintf(buf, sizeof("Reverse ARP Request"), "%s", - "Reverse ARP Request"); + case RTE_ARP_OP_REVREQUEST: + strlcpy(buf, "Reverse ARP Request", buf_len); return; - case ARP_OP_REVREPLY: - snprintf(buf, sizeof("Reverse ARP Reply"), "%s", - "Reverse ARP Reply"); + case RTE_ARP_OP_REVREPLY: + strlcpy(buf, "Reverse ARP Reply", buf_len); return; - case ARP_OP_INVREQUEST: - snprintf(buf, sizeof("Peer Identify Request"), "%s", - "Peer Identify Request"); + case RTE_ARP_OP_INVREQUEST: + strlcpy(buf, "Peer Identify Request", buf_len); return; - case ARP_OP_INVREPLY: - snprintf(buf, sizeof("Peer Identify Reply"), "%s", - "Peer Identify Reply"); + case RTE_ARP_OP_INVREPLY: + strlcpy(buf, "Peer Identify Reply", buf_len); return; default: break; } - snprintf(buf, sizeof("Unknown"), "%s", "Unknown"); + strlcpy(buf, "Unknown", buf_len); return; } #endif @@ -569,12 +489,13 @@ update_client_stats(uint32_t addr, uint16_t port, uint32_t *TXorRXindicator) #endif static void -mode6_debug(const char __attribute__((unused)) *info, struct ether_hdr *eth_h, - uint16_t port, uint32_t __attribute__((unused)) *burstnumber) +mode6_debug(const char __rte_unused *info, + struct rte_ether_hdr *eth_h, uint16_t port, + uint32_t __rte_unused *burstnumber) { - struct ipv4_hdr *ipv4_h; + struct rte_ipv4_hdr *ipv4_h; #ifdef RTE_LIBRTE_BOND_DEBUG_ALB - struct arp_hdr *arp_h; + struct rte_arp_hdr *arp_h; char dst_ip[16]; char ArpOp[24]; char buf[16]; @@ -588,8 +509,8 @@ mode6_debug(const char __attribute__((unused)) *info, struct ether_hdr *eth_h, strlcpy(buf, info, 16); #endif - if (ether_type == rte_cpu_to_be_16(ETHER_TYPE_IPv4)) { - ipv4_h = (struct ipv4_hdr *)((char *)(eth_h + 1) + offset); + if (ether_type == rte_cpu_to_be_16(RTE_ETHER_TYPE_IPV4)) { + ipv4_h = (struct rte_ipv4_hdr *)((char *)(eth_h + 1) + offset); ipv4_addr_to_dot(ipv4_h->src_addr, src_ip, MaxIPv4String); #ifdef RTE_LIBRTE_BOND_DEBUG_ALB ipv4_addr_to_dot(ipv4_h->dst_addr, dst_ip, MaxIPv4String); @@ -598,11 +519,12 @@ mode6_debug(const char __attribute__((unused)) *info, struct ether_hdr *eth_h, update_client_stats(ipv4_h->src_addr, port, burstnumber); } #ifdef RTE_LIBRTE_BOND_DEBUG_ALB - else if (ether_type == rte_cpu_to_be_16(ETHER_TYPE_ARP)) { - arp_h = (struct arp_hdr *)((char *)(eth_h + 1) + offset); + else if (ether_type == rte_cpu_to_be_16(RTE_ETHER_TYPE_ARP)) { + arp_h = (struct rte_arp_hdr *)((char *)(eth_h + 1) + offset); ipv4_addr_to_dot(arp_h->arp_data.arp_sip, src_ip, MaxIPv4String); ipv4_addr_to_dot(arp_h->arp_data.arp_tip, dst_ip, MaxIPv4String); - arp_op_name(rte_be_to_cpu_16(arp_h->arp_op), ArpOp); + arp_op_name(rte_be_to_cpu_16(arp_h->arp_opcode), + ArpOp, sizeof(ArpOp)); MODE6_DEBUG(buf, src_ip, dst_ip, eth_h, ArpOp, port, *burstnumber); } #endif @@ -612,9 +534,9 @@ mode6_debug(const char __attribute__((unused)) *info, struct ether_hdr *eth_h, static uint16_t bond_ethdev_rx_burst_alb(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts) { - struct bond_tx_queue *bd_tx_q = (struct bond_tx_queue *)queue; - struct bond_dev_private *internals = bd_tx_q->dev_private; - struct ether_hdr *eth_h; + struct bond_rx_queue *bd_rx_q = (struct bond_rx_queue *)queue; + struct bond_dev_private *internals = bd_rx_q->dev_private; + struct rte_ether_hdr *eth_h; uint16_t ether_type, offset; uint16_t nb_recv_pkts; int i; @@ -622,18 +544,18 @@ bond_ethdev_rx_burst_alb(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts) nb_recv_pkts = bond_ethdev_rx_burst(queue, bufs, nb_pkts); for (i = 0; i < nb_recv_pkts; i++) { - eth_h = rte_pktmbuf_mtod(bufs[i], struct ether_hdr *); + eth_h = rte_pktmbuf_mtod(bufs[i], struct rte_ether_hdr *); ether_type = eth_h->ether_type; offset = get_vlan_offset(eth_h, ðer_type); - if (ether_type == rte_cpu_to_be_16(ETHER_TYPE_ARP)) { + if (ether_type == rte_cpu_to_be_16(RTE_ETHER_TYPE_ARP)) { #if defined(RTE_LIBRTE_BOND_DEBUG_ALB) || defined(RTE_LIBRTE_BOND_DEBUG_ALB_L1) mode6_debug("RX ARP:", eth_h, bufs[i]->port, &burstnumberRX); #endif bond_mode_alb_arp_recv(eth_h, offset, internals); } #if defined(RTE_LIBRTE_BOND_DEBUG_ALB) || defined(RTE_LIBRTE_BOND_DEBUG_ALB_L1) - else if (ether_type == rte_cpu_to_be_16(ETHER_TYPE_IPv4)) + else if (ether_type == rte_cpu_to_be_16(RTE_ETHER_TYPE_IPV4)) mode6_debug("RX IPv4:", eth_h, bufs[i]->port, &burstnumberRX); #endif } @@ -722,7 +644,7 @@ bond_ethdev_tx_burst_active_backup(void *queue, } static inline uint16_t -ether_hash(struct ether_hdr *eth_hdr) +ether_hash(struct rte_ether_hdr *eth_hdr) { unaligned_uint16_t *word_src_addr = (unaligned_uint16_t *)eth_hdr->s_addr.addr_bytes; @@ -735,13 +657,13 @@ ether_hash(struct ether_hdr *eth_hdr) } static inline uint32_t -ipv4_hash(struct ipv4_hdr *ipv4_hdr) +ipv4_hash(struct rte_ipv4_hdr *ipv4_hdr) { return ipv4_hdr->src_addr ^ ipv4_hdr->dst_addr; } static inline uint32_t -ipv6_hash(struct ipv6_hdr *ipv6_hdr) +ipv6_hash(struct rte_ipv6_hdr *ipv6_hdr) { unaligned_uint32_t *word_src_addr = (unaligned_uint32_t *)&(ipv6_hdr->src_addr[0]); @@ -757,14 +679,14 @@ ipv6_hash(struct ipv6_hdr *ipv6_hdr) void burst_xmit_l2_hash(struct rte_mbuf **buf, uint16_t nb_pkts, - uint8_t slave_count, uint16_t *slaves) + uint16_t slave_count, uint16_t *slaves) { - struct ether_hdr *eth_hdr; + struct rte_ether_hdr *eth_hdr; uint32_t hash; int i; for (i = 0; i < nb_pkts; i++) { - eth_hdr = rte_pktmbuf_mtod(buf[i], struct ether_hdr *); + eth_hdr = rte_pktmbuf_mtod(buf[i], struct rte_ether_hdr *); hash = ether_hash(eth_hdr); @@ -774,16 +696,16 @@ burst_xmit_l2_hash(struct rte_mbuf **buf, uint16_t nb_pkts, void burst_xmit_l23_hash(struct rte_mbuf **buf, uint16_t nb_pkts, - uint8_t slave_count, uint16_t *slaves) + uint16_t slave_count, uint16_t *slaves) { uint16_t i; - struct ether_hdr *eth_hdr; + struct rte_ether_hdr *eth_hdr; uint16_t proto; size_t vlan_offset; uint32_t hash, l3hash; for (i = 0; i < nb_pkts; i++) { - eth_hdr = rte_pktmbuf_mtod(buf[i], struct ether_hdr *); + eth_hdr = rte_pktmbuf_mtod(buf[i], struct rte_ether_hdr *); l3hash = 0; proto = eth_hdr->ether_type; @@ -791,13 +713,13 @@ burst_xmit_l23_hash(struct rte_mbuf **buf, uint16_t nb_pkts, vlan_offset = get_vlan_offset(eth_hdr, &proto); - if (rte_cpu_to_be_16(ETHER_TYPE_IPv4) == proto) { - struct ipv4_hdr *ipv4_hdr = (struct ipv4_hdr *) + if (rte_cpu_to_be_16(RTE_ETHER_TYPE_IPV4) == proto) { + struct rte_ipv4_hdr *ipv4_hdr = (struct rte_ipv4_hdr *) ((char *)(eth_hdr + 1) + vlan_offset); l3hash = ipv4_hash(ipv4_hdr); - } else if (rte_cpu_to_be_16(ETHER_TYPE_IPv6) == proto) { - struct ipv6_hdr *ipv6_hdr = (struct ipv6_hdr *) + } else if (rte_cpu_to_be_16(RTE_ETHER_TYPE_IPV6) == proto) { + struct rte_ipv6_hdr *ipv6_hdr = (struct rte_ipv6_hdr *) ((char *)(eth_hdr + 1) + vlan_offset); l3hash = ipv6_hash(ipv6_hdr); } @@ -812,26 +734,27 @@ burst_xmit_l23_hash(struct rte_mbuf **buf, uint16_t nb_pkts, void burst_xmit_l34_hash(struct rte_mbuf **buf, uint16_t nb_pkts, - uint8_t slave_count, uint16_t *slaves) + uint16_t slave_count, uint16_t *slaves) { - struct ether_hdr *eth_hdr; + struct rte_ether_hdr *eth_hdr; uint16_t proto; size_t vlan_offset; int i; - struct udp_hdr *udp_hdr; - struct tcp_hdr *tcp_hdr; + struct rte_udp_hdr *udp_hdr; + struct rte_tcp_hdr *tcp_hdr; uint32_t hash, l3hash, l4hash; for (i = 0; i < nb_pkts; i++) { - eth_hdr = rte_pktmbuf_mtod(buf[i], struct ether_hdr *); + eth_hdr = rte_pktmbuf_mtod(buf[i], struct rte_ether_hdr *); + size_t pkt_end = (size_t)eth_hdr + rte_pktmbuf_data_len(buf[i]); proto = eth_hdr->ether_type; vlan_offset = get_vlan_offset(eth_hdr, &proto); l3hash = 0; l4hash = 0; - if (rte_cpu_to_be_16(ETHER_TYPE_IPv4) == proto) { - struct ipv4_hdr *ipv4_hdr = (struct ipv4_hdr *) + if (rte_cpu_to_be_16(RTE_ETHER_TYPE_IPV4) == proto) { + struct rte_ipv4_hdr *ipv4_hdr = (struct rte_ipv4_hdr *) ((char *)(eth_hdr + 1) + vlan_offset); size_t ip_hdr_offset; @@ -841,32 +764,36 @@ burst_xmit_l34_hash(struct rte_mbuf **buf, uint16_t nb_pkts, if (likely(rte_ipv4_frag_pkt_is_fragmented(ipv4_hdr) == 0)) { ip_hdr_offset = (ipv4_hdr->version_ihl - & IPV4_HDR_IHL_MASK) * - IPV4_IHL_MULTIPLIER; + & RTE_IPV4_HDR_IHL_MASK) * + RTE_IPV4_IHL_MULTIPLIER; if (ipv4_hdr->next_proto_id == IPPROTO_TCP) { - tcp_hdr = (struct tcp_hdr *) + tcp_hdr = (struct rte_tcp_hdr *) ((char *)ipv4_hdr + ip_hdr_offset); - l4hash = HASH_L4_PORTS(tcp_hdr); + if ((size_t)tcp_hdr + sizeof(*tcp_hdr) + < pkt_end) + l4hash = HASH_L4_PORTS(tcp_hdr); } else if (ipv4_hdr->next_proto_id == IPPROTO_UDP) { - udp_hdr = (struct udp_hdr *) + udp_hdr = (struct rte_udp_hdr *) ((char *)ipv4_hdr + ip_hdr_offset); - l4hash = HASH_L4_PORTS(udp_hdr); + if ((size_t)udp_hdr + sizeof(*udp_hdr) + < pkt_end) + l4hash = HASH_L4_PORTS(udp_hdr); } } - } else if (rte_cpu_to_be_16(ETHER_TYPE_IPv6) == proto) { - struct ipv6_hdr *ipv6_hdr = (struct ipv6_hdr *) + } else if (rte_cpu_to_be_16(RTE_ETHER_TYPE_IPV6) == proto) { + struct rte_ipv6_hdr *ipv6_hdr = (struct rte_ipv6_hdr *) ((char *)(eth_hdr + 1) + vlan_offset); l3hash = ipv6_hash(ipv6_hdr); if (ipv6_hdr->proto == IPPROTO_TCP) { - tcp_hdr = (struct tcp_hdr *)(ipv6_hdr + 1); + tcp_hdr = (struct rte_tcp_hdr *)(ipv6_hdr + 1); l4hash = HASH_L4_PORTS(tcp_hdr); } else if (ipv6_hdr->proto == IPPROTO_UDP) { - udp_hdr = (struct udp_hdr *)(ipv6_hdr + 1); + udp_hdr = (struct rte_udp_hdr *)(ipv6_hdr + 1); l4hash = HASH_L4_PORTS(udp_hdr); } } @@ -882,7 +809,7 @@ burst_xmit_l34_hash(struct rte_mbuf **buf, uint16_t nb_pkts, struct bwg_slave { uint64_t bwg_left_int; uint64_t bwg_left_remainder; - uint8_t slave; + uint16_t slave; }; void @@ -919,8 +846,14 @@ bandwidth_left(uint16_t port_id, uint64_t load, uint8_t update_idx, struct bwg_slave *bwg_slave) { struct rte_eth_link link_status; + int ret; - rte_eth_link_get_nowait(port_id, &link_status); + ret = rte_eth_link_get_nowait(port_id, &link_status); + if (ret < 0) { + RTE_BOND_LOG(ERR, "Slave (port %u) link get failed: %s", + port_id, rte_strerror(-ret)); + return; + } uint64_t link_bwg = link_status.link_speed * 1000000ULL / 8; if (link_bwg == 0) return; @@ -935,11 +868,12 @@ bond_ethdev_update_tlb_slave_cb(void *arg) struct bond_dev_private *internals = arg; struct rte_eth_stats slave_stats; struct bwg_slave bwg_array[RTE_MAX_ETHPORTS]; - uint8_t slave_count; + uint16_t slave_count; uint64_t tx_bytes; uint8_t update_stats = 0; - uint8_t i, slave_id; + uint16_t slave_id; + uint16_t i; internals->slave_update_idx++; @@ -986,9 +920,9 @@ bond_ethdev_tx_burst_tlb(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts) uint16_t num_of_slaves = internals->active_slave_count; uint16_t slaves[RTE_MAX_ETHPORTS]; - struct ether_hdr *ether_hdr; - struct ether_addr primary_slave_addr; - struct ether_addr active_slave_addr; + struct rte_ether_hdr *ether_hdr; + struct rte_ether_addr primary_slave_addr; + struct rte_ether_addr active_slave_addr; if (num_of_slaves < 1) return num_tx_total; @@ -997,7 +931,7 @@ bond_ethdev_tx_burst_tlb(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts) sizeof(internals->tlb_slaves_order[0]) * num_of_slaves); - ether_addr_copy(primary_port->data->mac_addrs, &primary_slave_addr); + rte_ether_addr_copy(primary_port->data->mac_addrs, &primary_slave_addr); if (nb_pkts > 3) { for (i = 0; i < 3; i++) @@ -1010,9 +944,12 @@ bond_ethdev_tx_burst_tlb(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts) if (j + 3 < nb_pkts) rte_prefetch0(rte_pktmbuf_mtod(bufs[j+3], void*)); - ether_hdr = rte_pktmbuf_mtod(bufs[j], struct ether_hdr *); - if (is_same_ether_addr(ðer_hdr->s_addr, &primary_slave_addr)) - ether_addr_copy(&active_slave_addr, ðer_hdr->s_addr); + ether_hdr = rte_pktmbuf_mtod(bufs[j], + struct rte_ether_hdr *); + if (rte_is_same_ether_addr(ðer_hdr->s_addr, + &primary_slave_addr)) + rte_ether_addr_copy(&active_slave_addr, + ðer_hdr->s_addr); #if defined(RTE_LIBRTE_BOND_DEBUG_ALB) || defined(RTE_LIBRTE_BOND_DEBUG_ALB_L1) mode6_debug("TX IPv4:", ether_hdr, slaves[i], &burstnumberTX); #endif @@ -1046,7 +983,7 @@ bond_ethdev_tx_burst_alb(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts) struct bond_tx_queue *bd_tx_q = (struct bond_tx_queue *)queue; struct bond_dev_private *internals = bd_tx_q->dev_private; - struct ether_hdr *eth_h; + struct rte_ether_hdr *eth_h; uint16_t ether_type, offset; struct client_data *client_info; @@ -1076,11 +1013,11 @@ bond_ethdev_tx_burst_alb(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts) /* Search tx buffer for ARP packets and forward them to alb */ for (i = 0; i < nb_pkts; i++) { - eth_h = rte_pktmbuf_mtod(bufs[i], struct ether_hdr *); + eth_h = rte_pktmbuf_mtod(bufs[i], struct rte_ether_hdr *); ether_type = eth_h->ether_type; offset = get_vlan_offset(eth_h, ðer_type); - if (ether_type == rte_cpu_to_be_16(ETHER_TYPE_ARP)) { + if (ether_type == rte_cpu_to_be_16(RTE_ETHER_TYPE_ARP)) { slave_idx = bond_mode_alb_arp_xmit(eth_h, offset, internals); /* Change src mac in eth header */ @@ -1110,8 +1047,10 @@ bond_ethdev_tx_burst_alb(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts) "Failed to allocate ARP packet from pool"); continue; } - pkt_size = sizeof(struct ether_hdr) + sizeof(struct arp_hdr) - + client_info->vlan_count * sizeof(struct vlan_hdr); + pkt_size = sizeof(struct rte_ether_hdr) + + sizeof(struct rte_arp_hdr) + + client_info->vlan_count * + sizeof(struct rte_vlan_hdr); upd_pkt->data_len = pkt_size; upd_pkt->pkt_len = pkt_size; @@ -1142,7 +1081,8 @@ bond_ethdev_tx_burst_alb(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts) #if defined(RTE_LIBRTE_BOND_DEBUG_ALB) || defined(RTE_LIBRTE_BOND_DEBUG_ALB_L1) /* Print TX stats including update packets */ for (j = 0; j < slave_bufs_pkts[i]; j++) { - eth_h = rte_pktmbuf_mtod(slave_bufs[i][j], struct ether_hdr *); + eth_h = rte_pktmbuf_mtod(slave_bufs[i][j], + struct rte_ether_hdr *); mode6_debug("TX ARP:", eth_h, i, &burstnumberTX); } #endif @@ -1159,7 +1099,8 @@ bond_ethdev_tx_burst_alb(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts) } #if defined(RTE_LIBRTE_BOND_DEBUG_ALB) || defined(RTE_LIBRTE_BOND_DEBUG_ALB_L1) for (j = 0; j < update_bufs_pkts[i]; j++) { - eth_h = rte_pktmbuf_mtod(update_bufs[i][j], struct ether_hdr *); + eth_h = rte_pktmbuf_mtod(update_bufs[i][j], + struct rte_ether_hdr *); mode6_debug("TX ARPupd:", eth_h, i, &burstnumberTX); } #endif @@ -1183,16 +1124,13 @@ bond_ethdev_tx_burst_alb(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts) return num_tx_total; } -static uint16_t -bond_ethdev_tx_burst_balance(void *queue, struct rte_mbuf **bufs, - uint16_t nb_bufs) +static inline uint16_t +tx_burst_balance(void *queue, struct rte_mbuf **bufs, uint16_t nb_bufs, + uint16_t *slave_port_ids, uint16_t slave_count) { struct bond_tx_queue *bd_tx_q = (struct bond_tx_queue *)queue; struct bond_dev_private *internals = bd_tx_q->dev_private; - uint16_t slave_port_ids[RTE_MAX_ETHPORTS]; - uint16_t slave_count; - /* Array to sort mbufs for transmission on each slave into */ struct rte_mbuf *slave_bufs[RTE_MAX_ETHPORTS][nb_bufs]; /* Number of mbufs for transmission on each slave */ @@ -1205,18 +1143,6 @@ bond_ethdev_tx_burst_balance(void *queue, struct rte_mbuf **bufs, uint16_t i; - if (unlikely(nb_bufs == 0)) - return 0; - - /* Copy slave list to protect against slave up/down changes during tx - * bursting */ - slave_count = internals->active_slave_count; - if (unlikely(slave_count < 1)) - return 0; - - memcpy(slave_port_ids, internals->active_slaves, - sizeof(slave_port_ids[0]) * slave_count); - /* * Populate slaves mbuf with the packets which are to be sent on it * selecting output slave using hash based on xmit policy @@ -1226,7 +1152,7 @@ bond_ethdev_tx_burst_balance(void *queue, struct rte_mbuf **bufs, for (i = 0; i < nb_bufs; i++) { /* Populate slave mbuf arrays with mbufs for that slave. */ - uint8_t slave_idx = bufs_slave_port_idxs[i]; + uint16_t slave_idx = bufs_slave_port_idxs[i]; slave_bufs[slave_idx][slave_nb_bufs[slave_idx]++] = bufs[i]; } @@ -1257,7 +1183,7 @@ bond_ethdev_tx_burst_balance(void *queue, struct rte_mbuf **bufs, } static uint16_t -bond_ethdev_tx_burst_8023ad(void *queue, struct rte_mbuf **bufs, +bond_ethdev_tx_burst_balance(void *queue, struct rte_mbuf **bufs, uint16_t nb_bufs) { struct bond_tx_queue *bd_tx_q = (struct bond_tx_queue *)queue; @@ -1266,91 +1192,54 @@ bond_ethdev_tx_burst_8023ad(void *queue, struct rte_mbuf **bufs, uint16_t slave_port_ids[RTE_MAX_ETHPORTS]; uint16_t slave_count; - uint16_t dist_slave_port_ids[RTE_MAX_ETHPORTS]; - uint16_t dist_slave_count; - - /* 2-D array to sort mbufs for transmission on each slave into */ - struct rte_mbuf *slave_bufs[RTE_MAX_ETHPORTS][nb_bufs]; - /* Number of mbufs for transmission on each slave */ - uint16_t slave_nb_bufs[RTE_MAX_ETHPORTS] = { 0 }; - /* Mapping array generated by hash function to map mbufs to slaves */ - uint16_t bufs_slave_port_idxs[RTE_MAX_ETHPORTS] = { 0 }; - - uint16_t slave_tx_count; - uint16_t total_tx_count = 0, total_tx_fail_count = 0; - - uint16_t i; - if (unlikely(nb_bufs == 0)) return 0; /* Copy slave list to protect against slave up/down changes during tx - * bursting */ + * bursting + */ slave_count = internals->active_slave_count; if (unlikely(slave_count < 1)) return 0; memcpy(slave_port_ids, internals->active_slaves, sizeof(slave_port_ids[0]) * slave_count); + return tx_burst_balance(queue, bufs, nb_bufs, slave_port_ids, + slave_count); +} - dist_slave_count = 0; - for (i = 0; i < slave_count; i++) { - struct port *port = &mode_8023ad_ports[slave_port_ids[i]]; - - if (ACTOR_STATE(port, DISTRIBUTING)) - dist_slave_port_ids[dist_slave_count++] = - slave_port_ids[i]; - } - - if (likely(dist_slave_count > 1)) { - - /* - * Populate slaves mbuf with the packets which are to be sent - * on it, selecting output slave using hash based on xmit policy - */ - internals->burst_xmit_hash(bufs, nb_bufs, dist_slave_count, - bufs_slave_port_idxs); - - for (i = 0; i < nb_bufs; i++) { - /* - * Populate slave mbuf arrays with mbufs for that - * slave - */ - uint8_t slave_idx = bufs_slave_port_idxs[i]; +static inline uint16_t +tx_burst_8023ad(void *queue, struct rte_mbuf **bufs, uint16_t nb_bufs, + bool dedicated_txq) +{ + struct bond_tx_queue *bd_tx_q = (struct bond_tx_queue *)queue; + struct bond_dev_private *internals = bd_tx_q->dev_private; - slave_bufs[slave_idx][slave_nb_bufs[slave_idx]++] = - bufs[i]; - } + uint16_t slave_port_ids[RTE_MAX_ETHPORTS]; + uint16_t slave_count; + uint16_t dist_slave_port_ids[RTE_MAX_ETHPORTS]; + uint16_t dist_slave_count; - /* Send packet burst on each slave device */ - for (i = 0; i < dist_slave_count; i++) { - if (slave_nb_bufs[i] == 0) - continue; + uint16_t slave_tx_count; - slave_tx_count = rte_eth_tx_burst( - dist_slave_port_ids[i], - bd_tx_q->queue_id, slave_bufs[i], - slave_nb_bufs[i]); + uint16_t i; - total_tx_count += slave_tx_count; + /* Copy slave list to protect against slave up/down changes during tx + * bursting */ + slave_count = internals->active_slave_count; + if (unlikely(slave_count < 1)) + return 0; - /* If tx burst fails move packets to end of bufs */ - if (unlikely(slave_tx_count < slave_nb_bufs[i])) { - int slave_tx_fail_count = slave_nb_bufs[i] - - slave_tx_count; - total_tx_fail_count += slave_tx_fail_count; + memcpy(slave_port_ids, internals->active_slaves, + sizeof(slave_port_ids[0]) * slave_count); - memcpy(&bufs[nb_bufs - total_tx_fail_count], - &slave_bufs[i][slave_tx_count], - slave_tx_fail_count * sizeof(bufs[0])); - } - } - } + if (dedicated_txq) + goto skip_tx_ring; /* Check for LACP control packets and send if available */ for (i = 0; i < slave_count; i++) { - struct port *port = &mode_8023ad_ports[slave_port_ids[i]]; + struct port *port = &bond_mode_8023ad_ports[slave_port_ids[i]]; struct rte_mbuf *ctrl_pkt = NULL; if (likely(rte_ring_empty(port->tx_ring))) @@ -1369,7 +1258,38 @@ bond_ethdev_tx_burst_8023ad(void *queue, struct rte_mbuf **bufs, } } - return total_tx_count; +skip_tx_ring: + if (unlikely(nb_bufs == 0)) + return 0; + + dist_slave_count = 0; + for (i = 0; i < slave_count; i++) { + struct port *port = &bond_mode_8023ad_ports[slave_port_ids[i]]; + + if (ACTOR_STATE(port, DISTRIBUTING)) + dist_slave_port_ids[dist_slave_count++] = + slave_port_ids[i]; + } + + if (unlikely(dist_slave_count < 1)) + return 0; + + return tx_burst_balance(queue, bufs, nb_bufs, dist_slave_port_ids, + dist_slave_count); +} + +static uint16_t +bond_ethdev_tx_burst_8023ad(void *queue, struct rte_mbuf **bufs, + uint16_t nb_bufs) +{ + return tx_burst_8023ad(queue, bufs, nb_bufs, false); +} + +static uint16_t +bond_ethdev_tx_burst_8023ad_fast_queue(void *queue, struct rte_mbuf **bufs, + uint16_t nb_bufs) +{ + return tx_burst_8023ad(queue, bufs, nb_bufs, true); } static uint16_t @@ -1379,8 +1299,9 @@ bond_ethdev_tx_burst_broadcast(void *queue, struct rte_mbuf **bufs, struct bond_dev_private *internals; struct bond_tx_queue *bd_tx_q; - uint8_t tx_failed_flag = 0, num_of_slaves; uint16_t slaves[RTE_MAX_ETHPORTS]; + uint8_t tx_failed_flag = 0; + uint16_t num_of_slaves; uint16_t max_nb_of_tx_pkts = 0; @@ -1432,7 +1353,7 @@ bond_ethdev_tx_burst_broadcast(void *queue, struct rte_mbuf **bufs, return max_nb_of_tx_pkts; } -void +static void link_properties_set(struct rte_eth_dev *ethdev, struct rte_eth_link *slave_link) { struct bond_dev_private *bond_ctx = ethdev->data->dev_private; @@ -1457,7 +1378,7 @@ link_properties_set(struct rte_eth_dev *ethdev, struct rte_eth_link *slave_link) } } -int +static int link_properties_valid(struct rte_eth_dev *ethdev, struct rte_eth_link *slave_link) { @@ -1476,9 +1397,10 @@ link_properties_valid(struct rte_eth_dev *ethdev, } int -mac_address_get(struct rte_eth_dev *eth_dev, struct ether_addr *dst_mac_addr) +mac_address_get(struct rte_eth_dev *eth_dev, + struct rte_ether_addr *dst_mac_addr) { - struct ether_addr *mac_addr; + struct rte_ether_addr *mac_addr; if (eth_dev == NULL) { RTE_BOND_LOG(ERR, "NULL pointer eth_dev specified"); @@ -1492,14 +1414,15 @@ mac_address_get(struct rte_eth_dev *eth_dev, struct ether_addr *dst_mac_addr) mac_addr = eth_dev->data->mac_addrs; - ether_addr_copy(mac_addr, dst_mac_addr); + rte_ether_addr_copy(mac_addr, dst_mac_addr); return 0; } int -mac_address_set(struct rte_eth_dev *eth_dev, struct ether_addr *new_mac_addr) +mac_address_set(struct rte_eth_dev *eth_dev, + struct rte_ether_addr *new_mac_addr) { - struct ether_addr *mac_addr; + struct rte_ether_addr *mac_addr; if (eth_dev == NULL) { RTE_BOND_LOG(ERR, "NULL pointer eth_dev specified"); @@ -1520,7 +1443,7 @@ mac_address_set(struct rte_eth_dev *eth_dev, struct ether_addr *new_mac_addr) return 0; } -static const struct ether_addr null_mac_addr; +static const struct rte_ether_addr null_mac_addr; /* * Add additional MAC addresses to the slave @@ -1530,11 +1453,11 @@ slave_add_mac_addresses(struct rte_eth_dev *bonded_eth_dev, uint16_t slave_port_id) { int i, ret; - struct ether_addr *mac_addr; + struct rte_ether_addr *mac_addr; for (i = 1; i < BOND_MAX_MAC_ADDRS; i++) { mac_addr = &bonded_eth_dev->data->mac_addrs[i]; - if (is_same_ether_addr(mac_addr, &null_mac_addr)) + if (rte_is_same_ether_addr(mac_addr, &null_mac_addr)) break; ret = rte_eth_dev_mac_addr_add(slave_port_id, mac_addr, 0); @@ -1558,12 +1481,12 @@ slave_remove_mac_addresses(struct rte_eth_dev *bonded_eth_dev, uint16_t slave_port_id) { int i, rc, ret; - struct ether_addr *mac_addr; + struct rte_ether_addr *mac_addr; rc = 0; for (i = 1; i < BOND_MAX_MAC_ADDRS; i++) { mac_addr = &bonded_eth_dev->data->mac_addrs[i]; - if (is_same_ether_addr(mac_addr, &null_mac_addr)) + if (rte_is_same_ether_addr(mac_addr, &null_mac_addr)) break; ret = rte_eth_dev_mac_addr_remove(slave_port_id, mac_addr); @@ -1579,6 +1502,7 @@ int mac_address_slaves_update(struct rte_eth_dev *bonded_eth_dev) { struct bond_dev_private *internals = bonded_eth_dev->data->dev_private; + bool set; int i; /* Update slave devices MAC addresses */ @@ -1606,15 +1530,16 @@ mac_address_slaves_update(struct rte_eth_dev *bonded_eth_dev) case BONDING_MODE_TLB: case BONDING_MODE_ALB: default: + set = true; for (i = 0; i < internals->slave_count; i++) { if (internals->slaves[i].port_id == internals->current_primary_port) { if (rte_eth_dev_default_mac_addr_set( - internals->primary_port, + internals->current_primary_port, bonded_eth_dev->data->mac_addrs)) { RTE_BOND_LOG(ERR, "Failed to update port Id %d MAC address", internals->current_primary_port); - return -1; + set = false; } } else { if (rte_eth_dev_default_mac_addr_set( @@ -1622,10 +1547,11 @@ mac_address_slaves_update(struct rte_eth_dev *bonded_eth_dev) &internals->slaves[i].persisted_mac_addr)) { RTE_BOND_LOG(ERR, "Failed to update port Id %d MAC address", internals->slaves[i].port_id); - return -1; } } } + if (!set) + return -1; } return 0; @@ -1699,9 +1625,8 @@ slave_configure_slow_queue(struct rte_eth_dev *bonded_eth_dev, struct rte_eth_dev *slave_eth_dev) { int errval = 0; - struct bond_dev_private *internals = (struct bond_dev_private *) - bonded_eth_dev->data->dev_private; - struct port *port = &mode_8023ad_ports[slave_eth_dev->data->port_id]; + struct bond_dev_private *internals = bonded_eth_dev->data->dev_private; + struct port *port = &bond_mode_8023ad_ports[slave_eth_dev->data->port_id]; if (port->slow_pool == NULL) { char mem_name[256]; @@ -1766,11 +1691,13 @@ slave_configure(struct rte_eth_dev *bonded_eth_dev, uint16_t q_id; struct rte_flow_error flow_error; - struct bond_dev_private *internals = (struct bond_dev_private *) - bonded_eth_dev->data->dev_private; + struct bond_dev_private *internals = bonded_eth_dev->data->dev_private; /* Stop slave */ - rte_eth_dev_stop(slave_eth_dev->data->port_id); + errval = rte_eth_dev_stop(slave_eth_dev->data->port_id); + if (errval != 0) + RTE_BOND_LOG(ERR, "rte_eth_dev_stop: port %u, err (%d)", + slave_eth_dev->data->port_id, errval); /* Enable interrupts on slave device if supported */ if (slave_eth_dev->data->dev_flags & RTE_ETH_DEV_INTR_LSC) @@ -1778,12 +1705,11 @@ slave_configure(struct rte_eth_dev *bonded_eth_dev, /* If RSS is enabled for bonding, try to enable it for slaves */ if (bonded_eth_dev->data->dev_conf.rxmode.mq_mode & ETH_MQ_RX_RSS_FLAG) { - if (bonded_eth_dev->data->dev_conf.rx_adv_conf.rss_conf.rss_key_len - != 0) { + if (internals->rss_key_len != 0) { slave_eth_dev->data->dev_conf.rx_adv_conf.rss_conf.rss_key_len = - bonded_eth_dev->data->dev_conf.rx_adv_conf.rss_conf.rss_key_len; + internals->rss_key_len; slave_eth_dev->data->dev_conf.rx_adv_conf.rss_conf.rss_key = - bonded_eth_dev->data->dev_conf.rx_adv_conf.rss_conf.rss_key; + internals->rss_key; } else { slave_eth_dev->data->dev_conf.rx_adv_conf.rss_conf.rss_key = NULL; } @@ -1932,7 +1858,7 @@ void slave_remove(struct bond_dev_private *internals, struct rte_eth_dev *slave_eth_dev) { - uint8_t i; + uint16_t i; for (i = 0; i < internals->slave_count; i++) if (internals->slaves[i].port_id == @@ -1956,7 +1882,7 @@ slave_remove(struct bond_dev_private *internals, internals->slave_count--; /* force reconfiguration of slave interfaces */ - _rte_eth_dev_reset(slave_eth_dev); + rte_eth_dev_internal_reset(slave_eth_dev); } static void @@ -1982,7 +1908,7 @@ slave_add(struct bond_dev_private *internals, slave_details->link_status_wait_to_complete = 0; /* clean tlb_last_obytes when adding port for bonding device */ memcpy(&(slave_details->persisted_mac_addr), slave_eth_dev->data->mac_addrs, - sizeof(struct ether_addr)); + sizeof(struct rte_ether_addr)); } void @@ -2001,7 +1927,7 @@ bond_ethdev_primary_set(struct bond_dev_private *internals, } } -static void +static int bond_ethdev_promiscuous_enable(struct rte_eth_dev *eth_dev); static int @@ -2028,7 +1954,7 @@ bond_ethdev_start(struct rte_eth_dev *eth_dev) } if (internals->user_defined_mac == 0) { - struct ether_addr *new_mac_addr = NULL; + struct rte_ether_addr *new_mac_addr = NULL; for (i = 0; i < internals->slave_count; i++) if (internals->slaves[i].port_id == internals->primary_port) @@ -2044,10 +1970,6 @@ bond_ethdev_start(struct rte_eth_dev *eth_dev) } } - /* If bonded device is configure in promiscuous mode then re-apply config */ - if (internals->promiscuous_en) - bond_ethdev_promiscuous_enable(eth_dev); - if (internals->mode == BONDING_MODE_8023AD) { if (internals->mode4.dedicated_queues.enabled == 1) { internals->mode4.dedicated_queues.rx_qid = @@ -2108,7 +2030,7 @@ out_err: static void bond_ethdev_free_queues(struct rte_eth_dev *dev) { - uint8_t i; + uint16_t i; if (dev->data->rx_queues != NULL) { for (i = 0; i < dev->data->nb_rx_queues; i++) { @@ -2127,11 +2049,12 @@ bond_ethdev_free_queues(struct rte_eth_dev *dev) } } -void +int bond_ethdev_stop(struct rte_eth_dev *eth_dev) { struct bond_dev_private *internals = eth_dev->data->dev_private; - uint8_t i; + uint16_t i; + int ret; if (internals->mode == BONDING_MODE_8023AD) { struct port *port; @@ -2141,7 +2064,7 @@ bond_ethdev_stop(struct rte_eth_dev *eth_dev) /* Discard all messages to/from mode 4 state machines */ for (i = 0; i < internals->active_slave_count; i++) { - port = &mode_8023ad_ports[internals->active_slaves[i]]; + port = &bond_mode_8023ad_ports[internals->active_slaves[i]]; RTE_ASSERT(port->rx_ring != NULL); while (rte_ring_dequeue(port->rx_ring, &pkt) != -ENOENT) @@ -2160,27 +2083,49 @@ bond_ethdev_stop(struct rte_eth_dev *eth_dev) tlb_last_obytets[internals->active_slaves[i]] = 0; } - internals->link_status_polling_enabled = 0; - for (i = 0; i < internals->slave_count; i++) - internals->slaves[i].last_link_status = 0; - eth_dev->data->dev_link.link_status = ETH_LINK_DOWN; eth_dev->data->dev_started = 0; + + internals->link_status_polling_enabled = 0; + for (i = 0; i < internals->slave_count; i++) { + uint16_t slave_id = internals->slaves[i].port_id; + if (find_slave_by_id(internals->active_slaves, + internals->active_slave_count, slave_id) != + internals->active_slave_count) { + internals->slaves[i].last_link_status = 0; + ret = rte_eth_dev_stop(slave_id); + if (ret != 0) { + RTE_BOND_LOG(ERR, "Failed to stop device on port %u", + slave_id); + return ret; + } + deactivate_slave(eth_dev, slave_id); + } + } + + return 0; } -void +int bond_ethdev_close(struct rte_eth_dev *dev) { struct bond_dev_private *internals = dev->data->dev_private; - uint8_t bond_port_id = internals->port_id; + uint16_t bond_port_id = internals->port_id; int skipped = 0; struct rte_flow_error ferror; + if (rte_eal_process_type() != RTE_PROC_PRIMARY) + return 0; + RTE_BOND_LOG(INFO, "Closing bonded device %s", dev->device->name); while (internals->slave_count != skipped) { uint16_t port_id = internals->slaves[skipped].port_id; - rte_eth_dev_stop(port_id); + if (rte_eth_dev_stop(port_id) != 0) { + RTE_BOND_LOG(ERR, "Failed to stop device on port %u", + port_id); + skipped++; + } if (rte_eth_bond_slave_remove(bond_port_id, port_id) != 0) { RTE_BOND_LOG(ERR, @@ -2192,24 +2137,37 @@ bond_ethdev_close(struct rte_eth_dev *dev) bond_flow_ops.flush(dev, &ferror); bond_ethdev_free_queues(dev); rte_bitmap_reset(internals->vlan_filter_bmp); + rte_bitmap_free(internals->vlan_filter_bmp); + rte_free(internals->vlan_filter_bmpmem); + + /* Try to release mempool used in mode6. If the bond + * device is not mode6, free the NULL is not problem. + */ + rte_mempool_free(internals->mode6.mempool); + + return 0; } /* forward declaration */ static int bond_ethdev_configure(struct rte_eth_dev *dev); -static void +static int bond_ethdev_info(struct rte_eth_dev *dev, struct rte_eth_dev_info *dev_info) { struct bond_dev_private *internals = dev->data->dev_private; + struct bond_slave_details slave; + int ret; uint16_t max_nb_rx_queues = UINT16_MAX; uint16_t max_nb_tx_queues = UINT16_MAX; + uint16_t max_rx_desc_lim = UINT16_MAX; + uint16_t max_tx_desc_lim = UINT16_MAX; dev_info->max_mac_addrs = BOND_MAX_MAC_ADDRS; dev_info->max_rx_pktlen = internals->candidate_max_rx_pktlen ? internals->candidate_max_rx_pktlen : - ETHER_MAX_JUMBO_FRAME_LEN; + RTE_ETHER_MAX_JUMBO_FRAME_LEN; /* Max number of tx/rx queues that the bonded device can support is the * minimum values of the bonded slaves, as all slaves must be capable @@ -2217,23 +2175,46 @@ bond_ethdev_info(struct rte_eth_dev *dev, struct rte_eth_dev_info *dev_info) */ if (internals->slave_count > 0) { struct rte_eth_dev_info slave_info; - uint8_t idx; + uint16_t idx; for (idx = 0; idx < internals->slave_count; idx++) { - rte_eth_dev_info_get(internals->slaves[idx].port_id, - &slave_info); + slave = internals->slaves[idx]; + ret = rte_eth_dev_info_get(slave.port_id, &slave_info); + if (ret != 0) { + RTE_BOND_LOG(ERR, + "%s: Error during getting device (port %u) info: %s\n", + __func__, + slave.port_id, + strerror(-ret)); + + return ret; + } if (slave_info.max_rx_queues < max_nb_rx_queues) max_nb_rx_queues = slave_info.max_rx_queues; if (slave_info.max_tx_queues < max_nb_tx_queues) max_nb_tx_queues = slave_info.max_tx_queues; + + if (slave_info.rx_desc_lim.nb_max < max_rx_desc_lim) + max_rx_desc_lim = slave_info.rx_desc_lim.nb_max; + + if (slave_info.tx_desc_lim.nb_max < max_tx_desc_lim) + max_tx_desc_lim = slave_info.tx_desc_lim.nb_max; } } dev_info->max_rx_queues = max_nb_rx_queues; dev_info->max_tx_queues = max_nb_tx_queues; + memcpy(&dev_info->default_rxconf, &internals->default_rxconf, + sizeof(dev_info->default_rxconf)); + memcpy(&dev_info->default_txconf, &internals->default_txconf, + sizeof(dev_info->default_txconf)); + + dev_info->rx_desc_lim.nb_max = max_rx_desc_lim; + dev_info->tx_desc_lim.nb_max = max_tx_desc_lim; + /** * If dedicated hw queues enabled for link bonding device in LACP mode * then we need to reduce the maximum number of data path queues by 1. @@ -2253,6 +2234,8 @@ bond_ethdev_info(struct rte_eth_dev *dev, struct rte_eth_dev_info *dev_info) dev_info->flow_type_rss_offloads = internals->flow_type_rss_offloads; dev_info->reta_size = internals->reta_size; + + return 0; } static int @@ -2362,8 +2345,8 @@ bond_ethdev_slave_link_status_change_monitor(void *cb_arg) if (cb_arg == NULL) return; - bonded_ethdev = (struct rte_eth_dev *)cb_arg; - internals = (struct bond_dev_private *)bonded_ethdev->data->dev_private; + bonded_ethdev = cb_arg; + internals = bonded_ethdev->data->dev_private; if (!bonded_ethdev->data->dev_started || !internals->link_status_polling_enabled) @@ -2411,12 +2394,14 @@ bond_ethdev_slave_link_status_change_monitor(void *cb_arg) static int bond_ethdev_link_update(struct rte_eth_dev *ethdev, int wait_to_complete) { - void (*link_update)(uint16_t port_id, struct rte_eth_link *eth_link); + int (*link_update)(uint16_t port_id, struct rte_eth_link *eth_link); struct bond_dev_private *bond_ctx; struct rte_eth_link slave_link; + bool one_link_update_succeeded; uint32_t idx; + int ret; bond_ctx = ethdev->data->dev_private; @@ -2448,8 +2433,18 @@ bond_ethdev_link_update(struct rte_eth_dev *ethdev, int wait_to_complete) * packet loss will occur on this slave if transmission at rates * greater than this are attempted */ - for (idx = 1; idx < bond_ctx->active_slave_count; idx++) { - link_update(bond_ctx->active_slaves[0], &slave_link); + for (idx = 0; idx < bond_ctx->active_slave_count; idx++) { + ret = link_update(bond_ctx->active_slaves[idx], + &slave_link); + if (ret < 0) { + ethdev->data->dev_link.link_speed = + ETH_SPEED_NUM_NONE; + RTE_BOND_LOG(ERR, + "Slave (port %u) link get failed: %s", + bond_ctx->active_slaves[idx], + rte_strerror(-ret)); + return 0; + } if (slave_link.link_speed < ethdev->data->dev_link.link_speed) @@ -2459,7 +2454,13 @@ bond_ethdev_link_update(struct rte_eth_dev *ethdev, int wait_to_complete) break; case BONDING_MODE_ACTIVE_BACKUP: /* Current primary slave */ - link_update(bond_ctx->current_primary_port, &slave_link); + ret = link_update(bond_ctx->current_primary_port, &slave_link); + if (ret < 0) { + RTE_BOND_LOG(ERR, "Slave (port %u) link get failed: %s", + bond_ctx->current_primary_port, + rte_strerror(-ret)); + return 0; + } ethdev->data->dev_link.link_speed = slave_link.link_speed; break; @@ -2468,7 +2469,8 @@ bond_ethdev_link_update(struct rte_eth_dev *ethdev, int wait_to_complete) bond_ctx->mode4.slave_link.link_autoneg; ethdev->data->dev_link.link_duplex = bond_ctx->mode4.slave_link.link_duplex; - /* fall through to update link speed */ + /* fall through */ + /* to update link speed */ case BONDING_MODE_ROUND_ROBIN: case BONDING_MODE_BALANCE: case BONDING_MODE_TLB: @@ -2479,13 +2481,28 @@ bond_ethdev_link_update(struct rte_eth_dev *ethdev, int wait_to_complete) * of all the slaves */ ethdev->data->dev_link.link_speed = ETH_SPEED_NUM_NONE; + one_link_update_succeeded = false; for (idx = 0; idx < bond_ctx->active_slave_count; idx++) { - link_update(bond_ctx->active_slaves[idx], &slave_link); + ret = link_update(bond_ctx->active_slaves[idx], + &slave_link); + if (ret < 0) { + RTE_BOND_LOG(ERR, + "Slave (port %u) link get failed: %s", + bond_ctx->active_slaves[idx], + rte_strerror(-ret)); + continue; + } + one_link_update_succeeded = true; ethdev->data->dev_link.link_speed += slave_link.link_speed; } + + if (!one_link_update_succeeded) { + RTE_BOND_LOG(ERR, "All slaves link get failed"); + return 0; + } } @@ -2525,70 +2542,248 @@ bond_ethdev_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats) return 0; } -static void +static int bond_ethdev_stats_reset(struct rte_eth_dev *dev) { struct bond_dev_private *internals = dev->data->dev_private; int i; + int err; + int ret; - for (i = 0; i < internals->slave_count; i++) - rte_eth_stats_reset(internals->slaves[i].port_id); + for (i = 0, err = 0; i < internals->slave_count; i++) { + ret = rte_eth_stats_reset(internals->slaves[i].port_id); + if (ret != 0) + err = ret; + } + + return err; } -static void +static int bond_ethdev_promiscuous_enable(struct rte_eth_dev *eth_dev) { struct bond_dev_private *internals = eth_dev->data->dev_private; int i; - - internals->promiscuous_en = 1; + int ret = 0; + uint16_t port_id; switch (internals->mode) { /* Promiscuous mode is propagated to all slaves */ case BONDING_MODE_ROUND_ROBIN: case BONDING_MODE_BALANCE: case BONDING_MODE_BROADCAST: - for (i = 0; i < internals->slave_count; i++) - rte_eth_promiscuous_enable(internals->slaves[i].port_id); - break; - /* In mode4 promiscus mode is managed when slave is added/removed */ - case BONDING_MODE_8023AD: + case BONDING_MODE_8023AD: { + unsigned int slave_ok = 0; + + for (i = 0; i < internals->slave_count; i++) { + port_id = internals->slaves[i].port_id; + + ret = rte_eth_promiscuous_enable(port_id); + if (ret != 0) + RTE_BOND_LOG(ERR, + "Failed to enable promiscuous mode for port %u: %s", + port_id, rte_strerror(-ret)); + else + slave_ok++; + } + /* + * Report success if operation is successful on at least + * on one slave. Otherwise return last error code. + */ + if (slave_ok > 0) + ret = 0; break; + } /* Promiscuous mode is propagated only to primary slave */ case BONDING_MODE_ACTIVE_BACKUP: case BONDING_MODE_TLB: case BONDING_MODE_ALB: default: - rte_eth_promiscuous_enable(internals->current_primary_port); + /* Do not touch promisc when there cannot be primary ports */ + if (internals->slave_count == 0) + break; + port_id = internals->current_primary_port; + ret = rte_eth_promiscuous_enable(port_id); + if (ret != 0) + RTE_BOND_LOG(ERR, + "Failed to enable promiscuous mode for port %u: %s", + port_id, rte_strerror(-ret)); } + + return ret; } -static void +static int bond_ethdev_promiscuous_disable(struct rte_eth_dev *dev) { struct bond_dev_private *internals = dev->data->dev_private; int i; - - internals->promiscuous_en = 0; + int ret = 0; + uint16_t port_id; switch (internals->mode) { /* Promiscuous mode is propagated to all slaves */ case BONDING_MODE_ROUND_ROBIN: case BONDING_MODE_BALANCE: case BONDING_MODE_BROADCAST: - for (i = 0; i < internals->slave_count; i++) - rte_eth_promiscuous_disable(internals->slaves[i].port_id); - break; - /* In mode4 promiscus mode is set managed when slave is added/removed */ - case BONDING_MODE_8023AD: + case BONDING_MODE_8023AD: { + unsigned int slave_ok = 0; + + for (i = 0; i < internals->slave_count; i++) { + port_id = internals->slaves[i].port_id; + + if (internals->mode == BONDING_MODE_8023AD && + bond_mode_8023ad_ports[port_id].forced_rx_flags == + BOND_8023AD_FORCED_PROMISC) { + slave_ok++; + continue; + } + ret = rte_eth_promiscuous_disable(port_id); + if (ret != 0) + RTE_BOND_LOG(ERR, + "Failed to disable promiscuous mode for port %u: %s", + port_id, rte_strerror(-ret)); + else + slave_ok++; + } + /* + * Report success if operation is successful on at least + * on one slave. Otherwise return last error code. + */ + if (slave_ok > 0) + ret = 0; break; + } /* Promiscuous mode is propagated only to primary slave */ case BONDING_MODE_ACTIVE_BACKUP: case BONDING_MODE_TLB: case BONDING_MODE_ALB: default: - rte_eth_promiscuous_disable(internals->current_primary_port); + /* Do not touch promisc when there cannot be primary ports */ + if (internals->slave_count == 0) + break; + port_id = internals->current_primary_port; + ret = rte_eth_promiscuous_disable(port_id); + if (ret != 0) + RTE_BOND_LOG(ERR, + "Failed to disable promiscuous mode for port %u: %s", + port_id, rte_strerror(-ret)); + } + + return ret; +} + +static int +bond_ethdev_allmulticast_enable(struct rte_eth_dev *eth_dev) +{ + struct bond_dev_private *internals = eth_dev->data->dev_private; + int i; + int ret = 0; + uint16_t port_id; + + switch (internals->mode) { + /* allmulti mode is propagated to all slaves */ + case BONDING_MODE_ROUND_ROBIN: + case BONDING_MODE_BALANCE: + case BONDING_MODE_BROADCAST: + case BONDING_MODE_8023AD: { + unsigned int slave_ok = 0; + + for (i = 0; i < internals->slave_count; i++) { + port_id = internals->slaves[i].port_id; + + ret = rte_eth_allmulticast_enable(port_id); + if (ret != 0) + RTE_BOND_LOG(ERR, + "Failed to enable allmulti mode for port %u: %s", + port_id, rte_strerror(-ret)); + else + slave_ok++; + } + /* + * Report success if operation is successful on at least + * on one slave. Otherwise return last error code. + */ + if (slave_ok > 0) + ret = 0; + break; + } + /* allmulti mode is propagated only to primary slave */ + case BONDING_MODE_ACTIVE_BACKUP: + case BONDING_MODE_TLB: + case BONDING_MODE_ALB: + default: + /* Do not touch allmulti when there cannot be primary ports */ + if (internals->slave_count == 0) + break; + port_id = internals->current_primary_port; + ret = rte_eth_allmulticast_enable(port_id); + if (ret != 0) + RTE_BOND_LOG(ERR, + "Failed to enable allmulti mode for port %u: %s", + port_id, rte_strerror(-ret)); } + + return ret; +} + +static int +bond_ethdev_allmulticast_disable(struct rte_eth_dev *eth_dev) +{ + struct bond_dev_private *internals = eth_dev->data->dev_private; + int i; + int ret = 0; + uint16_t port_id; + + switch (internals->mode) { + /* allmulti mode is propagated to all slaves */ + case BONDING_MODE_ROUND_ROBIN: + case BONDING_MODE_BALANCE: + case BONDING_MODE_BROADCAST: + case BONDING_MODE_8023AD: { + unsigned int slave_ok = 0; + + for (i = 0; i < internals->slave_count; i++) { + uint16_t port_id = internals->slaves[i].port_id; + + if (internals->mode == BONDING_MODE_8023AD && + bond_mode_8023ad_ports[port_id].forced_rx_flags == + BOND_8023AD_FORCED_ALLMULTI) + continue; + + ret = rte_eth_allmulticast_disable(port_id); + if (ret != 0) + RTE_BOND_LOG(ERR, + "Failed to disable allmulti mode for port %u: %s", + port_id, rte_strerror(-ret)); + else + slave_ok++; + } + /* + * Report success if operation is successful on at least + * on one slave. Otherwise return last error code. + */ + if (slave_ok > 0) + ret = 0; + break; + } + /* allmulti mode is propagated only to primary slave */ + case BONDING_MODE_ACTIVE_BACKUP: + case BONDING_MODE_TLB: + case BONDING_MODE_ALB: + default: + /* Do not touch allmulti when there cannot be primary ports */ + if (internals->slave_count == 0) + break; + port_id = internals->current_primary_port; + ret = rte_eth_allmulticast_disable(port_id); + if (ret != 0) + RTE_BOND_LOG(ERR, + "Failed to disable allmulti mode for port %u: %s", + port_id, rte_strerror(-ret)); + } + + return ret; } static void @@ -2597,7 +2792,7 @@ bond_ethdev_delayed_lsc_propagation(void *arg) if (arg == NULL) return; - _rte_eth_dev_callback_process((struct rte_eth_dev *)arg, + rte_eth_dev_callback_process((struct rte_eth_dev *)arg, RTE_ETH_EVENT_INTR_LSC, NULL); } @@ -2609,15 +2804,17 @@ bond_ethdev_lsc_event_callback(uint16_t port_id, enum rte_eth_event_type type, struct bond_dev_private *internals; struct rte_eth_link link; int rc = -1; + int ret; - int i, valid_slave = 0; - uint8_t active_pos; uint8_t lsc_flag = 0; + int valid_slave = 0; + uint16_t active_pos; + uint16_t i; if (type != RTE_ETH_EVENT_INTR_LSC || param == NULL) return rc; - bonded_eth_dev = &rte_eth_devices[*(uint8_t *)param]; + bonded_eth_dev = &rte_eth_devices[*(uint16_t *)param]; if (check_for_bonded_ethdev(bonded_eth_dev)) return rc; @@ -2648,21 +2845,14 @@ bond_ethdev_lsc_event_callback(uint16_t port_id, enum rte_eth_event_type type, active_pos = find_slave_by_id(internals->active_slaves, internals->active_slave_count, port_id); - rte_eth_link_get_nowait(port_id, &link); - if (link.link_status) { + ret = rte_eth_link_get_nowait(port_id, &link); + if (ret < 0) + RTE_BOND_LOG(ERR, "Slave (port %u) link get failed", port_id); + + if (ret == 0 && link.link_status) { if (active_pos < internals->active_slave_count) goto link_update; - /* if no active slave ports then set this port to be primary port */ - if (internals->active_slave_count < 1) { - /* If first active slave, then change link status */ - bonded_eth_dev->data->dev_link.link_status = ETH_LINK_UP; - internals->current_primary_port = port_id; - lsc_flag = 1; - - mac_address_slaves_update(bonded_eth_dev); - } - /* check link state properties if bonded link is up*/ if (bonded_eth_dev->data->dev_link.link_status == ETH_LINK_UP) { if (link_properties_valid(bonded_eth_dev, &link) != 0) @@ -2674,9 +2864,24 @@ bond_ethdev_lsc_event_callback(uint16_t port_id, enum rte_eth_event_type type, link_properties_set(bonded_eth_dev, &link); } + /* If no active slave ports then set this port to be + * the primary port. + */ + if (internals->active_slave_count < 1) { + /* If first active slave, then change link status */ + bonded_eth_dev->data->dev_link.link_status = + ETH_LINK_UP; + internals->current_primary_port = port_id; + lsc_flag = 1; + + mac_address_slaves_update(bonded_eth_dev); + } + activate_slave(bonded_eth_dev, port_id); - /* If user has defined the primary port then default to using it */ + /* If the user has defined the primary port then default to + * using it. + */ if (internals->user_defined_primary_port && internals->primary_port == port_id) bond_ethdev_primary_set(internals, port_id); @@ -2698,6 +2903,7 @@ bond_ethdev_lsc_event_callback(uint16_t port_id, enum rte_eth_event_type type, internals->active_slaves[0]); else internals->current_primary_port = internals->primary_port; + mac_address_slaves_update(bonded_eth_dev); } } @@ -2721,7 +2927,7 @@ link_update: bond_ethdev_delayed_lsc_propagation, (void *)bonded_eth_dev); else - _rte_eth_dev_callback_process(bonded_eth_dev, + rte_eth_dev_callback_process(bonded_eth_dev, RTE_ETH_EVENT_INTR_LSC, NULL); @@ -2731,7 +2937,7 @@ link_update: bond_ethdev_delayed_lsc_propagation, (void *)bonded_eth_dev); else - _rte_eth_dev_callback_process(bonded_eth_dev, + rte_eth_dev_callback_process(bonded_eth_dev, RTE_ETH_EVENT_INTR_LSC, NULL); } @@ -2756,7 +2962,8 @@ bond_ethdev_rss_reta_update(struct rte_eth_dev *dev, return -EINVAL; /* Copy RETA table */ - reta_count = reta_size / RTE_RETA_GROUP_SIZE; + reta_count = (reta_size + RTE_RETA_GROUP_SIZE - 1) / + RTE_RETA_GROUP_SIZE; for (i = 0; i < reta_count; i++) { internals->reta_conf[i].mask = reta_conf[i].mask; @@ -2878,7 +3085,8 @@ bond_ethdev_mtu_set(struct rte_eth_dev *dev, uint16_t mtu) } static int -bond_ethdev_mac_address_set(struct rte_eth_dev *dev, struct ether_addr *addr) +bond_ethdev_mac_address_set(struct rte_eth_dev *dev, + struct rte_ether_addr *addr) { if (mac_address_set(dev, addr)) { RTE_BOND_LOG(ERR, "Failed to update MAC address"); @@ -2900,8 +3108,9 @@ bond_filter_ctrl(struct rte_eth_dev *dev __rte_unused, } static int -bond_ethdev_mac_addr_add(struct rte_eth_dev *dev, struct ether_addr *mac_addr, - __rte_unused uint32_t index, uint32_t vmdq) +bond_ethdev_mac_addr_add(struct rte_eth_dev *dev, + struct rte_ether_addr *mac_addr, + __rte_unused uint32_t index, uint32_t vmdq) { struct rte_eth_dev *slave_eth_dev; struct bond_dev_private *internals = dev->data->dev_private; @@ -2951,7 +3160,7 @@ bond_ethdev_mac_addr_remove(struct rte_eth_dev *dev, uint32_t index) goto end; } - struct ether_addr *mac_addr = &dev->data->mac_addrs[index]; + struct rte_ether_addr *mac_addr = &dev->data->mac_addrs[index]; for (i = 0; i < internals->slave_count; i++) rte_eth_dev_mac_addr_remove(internals->slaves[i].port_id, @@ -2977,6 +3186,8 @@ const struct eth_dev_ops default_dev_ops = { .stats_reset = bond_ethdev_stats_reset, .promiscuous_enable = bond_ethdev_promiscuous_enable, .promiscuous_disable = bond_ethdev_promiscuous_disable, + .allmulticast_enable = bond_ethdev_allmulticast_enable, + .allmulticast_disable = bond_ethdev_allmulticast_disable, .reta_update = bond_ethdev_rss_reta_update, .reta_query = bond_ethdev_rss_reta_query, .rss_hash_update = bond_ethdev_rss_hash_update, @@ -3013,17 +3224,18 @@ bond_alloc(struct rte_vdev_device *dev, uint8_t mode) eth_dev->data->nb_tx_queues = (uint16_t)1; /* Allocate memory for storing MAC addresses */ - eth_dev->data->mac_addrs = rte_zmalloc_socket(name, ETHER_ADDR_LEN * + eth_dev->data->mac_addrs = rte_zmalloc_socket(name, RTE_ETHER_ADDR_LEN * BOND_MAX_MAC_ADDRS, 0, socket_id); if (eth_dev->data->mac_addrs == NULL) { RTE_BOND_LOG(ERR, "Failed to allocate %u bytes needed to store MAC addresses", - ETHER_ADDR_LEN * BOND_MAX_MAC_ADDRS); + RTE_ETHER_ADDR_LEN * BOND_MAX_MAC_ADDRS); goto err; } eth_dev->dev_ops = &default_dev_ops; - eth_dev->data->dev_flags = RTE_ETH_DEV_INTR_LSC; + eth_dev->data->dev_flags = RTE_ETH_DEV_INTR_LSC | + RTE_ETH_DEV_AUTOFILL_QUEUE_XSTATS; rte_spinlock_init(&internals->lock); rte_spinlock_init(&internals->lsc_lock); @@ -3054,6 +3266,14 @@ bond_alloc(struct rte_vdev_device *dev, uint8_t mode) /* Initially allow to choose any offload type */ internals->flow_type_rss_offloads = ETH_RSS_PROTO_MASK; + memset(&internals->default_rxconf, 0, + sizeof(internals->default_rxconf)); + memset(&internals->default_txconf, 0, + sizeof(internals->default_txconf)); + + memset(&internals->rx_desc_lim, 0, sizeof(internals->rx_desc_lim)); + memset(&internals->tx_desc_lim, 0, sizeof(internals->tx_desc_lim)); + memset(internals->active_slaves, 0, sizeof(internals->active_slaves)); memset(internals->slaves, 0, sizeof(internals->slaves)); @@ -3069,7 +3289,7 @@ bond_alloc(struct rte_vdev_device *dev, uint8_t mode) } vlan_filter_bmp_size = - rte_bitmap_get_memory_footprint(ETHER_MAX_VLAN_ID + 1); + rte_bitmap_get_memory_footprint(RTE_ETHER_MAX_VLAN_ID + 1); internals->vlan_filter_bmpmem = rte_malloc(name, vlan_filter_bmp_size, RTE_CACHE_LINE_SIZE); if (internals->vlan_filter_bmpmem == NULL) { @@ -3079,7 +3299,7 @@ bond_alloc(struct rte_vdev_device *dev, uint8_t mode) goto err; } - internals->vlan_filter_bmp = rte_bitmap_init(ETHER_MAX_VLAN_ID + 1, + internals->vlan_filter_bmp = rte_bitmap_init(RTE_ETHER_MAX_VLAN_ID + 1, internals->vlan_filter_bmpmem, vlan_filter_bmp_size); if (internals->vlan_filter_bmp == NULL) { RTE_BOND_LOG(ERR, @@ -3093,10 +3313,9 @@ bond_alloc(struct rte_vdev_device *dev, uint8_t mode) err: rte_free(internals); - if (eth_dev != NULL) { - rte_free(eth_dev->data->mac_addrs); - rte_eth_dev_release_port(eth_dev); - } + if (eth_dev != NULL) + eth_dev->data->dev_private = NULL; + rte_eth_dev_release_port(eth_dev); return -1; } @@ -3117,8 +3336,7 @@ bond_probe(struct rte_vdev_device *dev) name = rte_vdev_device_name(dev); RTE_BOND_LOG(INFO, "Initializing pmd_bond for %s", name); - if (rte_eal_process_type() == RTE_PROC_SECONDARY && - strlen(rte_vdev_device_args(dev)) == 0) { + if (rte_eal_process_type() == RTE_PROC_SECONDARY) { eth_dev = rte_eth_dev_attach_secondary(name); if (!eth_dev) { RTE_BOND_LOG(ERR, "Failed to probe %s", name); @@ -3181,8 +3399,6 @@ bond_probe(struct rte_vdev_device *dev) internals = rte_eth_devices[port_id].data->dev_private; internals->kvlist = kvlist; - rte_eth_dev_probing_finish(&rte_eth_devices[port_id]); - if (rte_kvargs_count(kvlist, PMD_BOND_AGG_MODE_KVARG) == 1) { if (rte_kvargs_process(kvlist, PMD_BOND_AGG_MODE_KVARG, @@ -3195,12 +3411,12 @@ bond_probe(struct rte_vdev_device *dev) } if (internals->mode == BONDING_MODE_8023AD) - rte_eth_bond_8023ad_agg_selection_set(port_id, - agg_mode); + internals->mode4.agg_selection = agg_mode; } else { - rte_eth_bond_8023ad_agg_selection_set(port_id, AGG_STABLE); + internals->mode4.agg_selection = AGG_STABLE; } + rte_eth_dev_probing_finish(&rte_eth_devices[port_id]); RTE_BOND_LOG(INFO, "Create bonded device %s on port %d in mode %u on " "socket %u.", name, port_id, bonding_mode, socket_id); return 0; @@ -3217,6 +3433,7 @@ bond_remove(struct rte_vdev_device *dev) struct rte_eth_dev *eth_dev; struct bond_dev_private *internals; const char *name; + int ret = 0; if (!dev) return -EINVAL; @@ -3224,14 +3441,13 @@ bond_remove(struct rte_vdev_device *dev) name = rte_vdev_device_name(dev); RTE_BOND_LOG(INFO, "Uninitializing pmd_bond for %s", name); - /* now free all data allocation - for eth_dev structure, - * dummy pci driver and internal (private) data - */ - /* find an ethdev entry */ eth_dev = rte_eth_dev_allocated(name); if (eth_dev == NULL) - return -ENODEV; + return 0; /* port already released */ + + if (rte_eal_process_type() != RTE_PROC_PRIMARY) + return rte_eth_dev_release_port(eth_dev); RTE_ASSERT(eth_dev->device == &dev->device); @@ -3240,27 +3456,12 @@ bond_remove(struct rte_vdev_device *dev) return -EBUSY; if (eth_dev->data->dev_started == 1) { - bond_ethdev_stop(eth_dev); + ret = bond_ethdev_stop(eth_dev); bond_ethdev_close(eth_dev); } - - eth_dev->dev_ops = NULL; - eth_dev->rx_pkt_burst = NULL; - eth_dev->tx_pkt_burst = NULL; - - internals = eth_dev->data->dev_private; - /* Try to release mempool used in mode6. If the bond - * device is not mode6, free the NULL is not problem. - */ - rte_mempool_free(internals->mode6.mempool); - rte_bitmap_free(internals->vlan_filter_bmp); - rte_free(internals->vlan_filter_bmpmem); - rte_free(eth_dev->data->dev_private); - rte_free(eth_dev->data->mac_addrs); - rte_eth_dev_release_port(eth_dev); - return 0; + return ret; } /* this part will resolve the slave portids after all the other pdev and vdev @@ -3284,11 +3485,23 @@ bond_ethdev_configure(struct rte_eth_dev *dev) unsigned i, j; - /* If RSS is enabled, fill table and key with default values */ + /* + * If RSS is enabled, fill table with default values and + * set key to the the value specified in port RSS configuration. + * Fall back to default RSS key if the key is not specified + */ if (dev->data->dev_conf.rxmode.mq_mode & ETH_MQ_RX_RSS) { - dev->data->dev_conf.rx_adv_conf.rss_conf.rss_key = internals->rss_key; - dev->data->dev_conf.rx_adv_conf.rss_conf.rss_key_len = 0; - memcpy(internals->rss_key, default_rss_key, 40); + if (dev->data->dev_conf.rx_adv_conf.rss_conf.rss_key != NULL) { + internals->rss_key_len = + dev->data->dev_conf.rx_adv_conf.rss_conf.rss_key_len; + memcpy(internals->rss_key, + dev->data->dev_conf.rx_adv_conf.rss_conf.rss_key, + internals->rss_key_len); + } else { + internals->rss_key_len = sizeof(default_rss_key); + memcpy(internals->rss_key, default_rss_key, + internals->rss_key_len); + } for (i = 0; i < RTE_DIM(internals->reta_conf); i++) { internals->reta_conf[i].mask = ~0LL; @@ -3312,7 +3525,7 @@ bond_ethdev_configure(struct rte_eth_dev *dev) /* Parse MAC address for bonded device */ arg_count = rte_kvargs_count(kvlist, PMD_BOND_MAC_ADDR_KVARG); if (arg_count == 1) { - struct ether_addr bond_mac; + struct rte_ether_addr bond_mac; if (rte_kvargs_process(kvlist, PMD_BOND_MAC_ADDR_KVARG, &bond_ethdev_parse_bond_mac_addr_kvarg, &bond_mac) < 0) { @@ -3372,9 +3585,16 @@ bond_ethdev_configure(struct rte_eth_dev *dev) "Failed to parse agg selection mode for bonded device %s", name); } - if (internals->mode == BONDING_MODE_8023AD) - rte_eth_bond_8023ad_agg_selection_set(port_id, - agg_mode); + if (internals->mode == BONDING_MODE_8023AD) { + int ret = rte_eth_bond_8023ad_agg_selection_set(port_id, + agg_mode); + if (ret < 0) { + RTE_BOND_LOG(ERR, + "Invalid args for agg selection set for bonded device %s", + name); + return -1; + } + } } /* Parse/add slave ports to bonded device */ @@ -3547,11 +3767,4 @@ RTE_PMD_REGISTER_PARAM_STRING(net_bonding, "up_delay= " "down_delay="); -int bond_logtype; - -RTE_INIT(bond_init_log) -{ - bond_logtype = rte_log_register("pmd.net.bon"); - if (bond_logtype >= 0) - rte_log_set_level(bond_logtype, RTE_LOG_NOTICE); -} +RTE_LOG_REGISTER(bond_logtype, pmd.net.bond, NOTICE);