diff options
Diffstat (limited to 'drivers/infiniband/hw/nes/nes_cm.c')
| -rw-r--r-- | drivers/infiniband/hw/nes/nes_cm.c | 3747 |
1 files changed, 2404 insertions, 1343 deletions
diff --git a/drivers/infiniband/hw/nes/nes_cm.c b/drivers/infiniband/hw/nes/nes_cm.c index 0bef878e0f6..6f09a72e78d 100644 --- a/drivers/infiniband/hw/nes/nes_cm.c +++ b/drivers/infiniband/hw/nes/nes_cm.c @@ -1,5 +1,5 @@ /* - * Copyright (c) 2006 - 2008 NetEffect, Inc. All rights reserved. + * Copyright (c) 2006 - 2014 Intel Corporation. All rights reserved. * * This software is available to you under a choice of one of two * licenses. You may choose to be licensed under the terms of the GNU @@ -34,7 +34,7 @@ #define TCPOPT_TIMESTAMP 8 -#include <asm/atomic.h> +#include <linux/atomic.h> #include <linux/skbuff.h> #include <linux/ip.h> #include <linux/tcp.h> @@ -52,10 +52,14 @@ #include <linux/random.h> #include <linux/list.h> #include <linux/threads.h> - +#include <linux/highmem.h> +#include <linux/slab.h> +#include <net/arp.h> #include <net/neighbour.h> #include <net/route.h> #include <net/ip_fib.h> +#include <net/tcp.h> +#include <linux/fcntl.h> #include "nes.h" @@ -65,8 +69,8 @@ u32 cm_packets_dropped; u32 cm_packets_retrans; u32 cm_packets_created; u32 cm_packets_received; -u32 cm_listens_created; -u32 cm_listens_destroyed; +atomic_t cm_listens_created; +atomic_t cm_listens_destroyed; u32 cm_backlog_drops; atomic_t cm_loopbacks; atomic_t cm_nodes_created; @@ -75,12 +79,57 @@ atomic_t cm_accel_dropped_pkts; atomic_t cm_resets_recvd; static inline int mini_cm_accelerated(struct nes_cm_core *, struct nes_cm_node *); -static struct nes_cm_listener *mini_cm_listen(struct nes_cm_core *, - struct nes_vnic *, struct nes_cm_info *); +static struct nes_cm_listener *mini_cm_listen(struct nes_cm_core *, struct nes_vnic *, struct nes_cm_info *); +static int mini_cm_del_listen(struct nes_cm_core *, struct nes_cm_listener *); +static struct nes_cm_node *mini_cm_connect(struct nes_cm_core *, struct nes_vnic *, u16, void *, struct nes_cm_info *); +static int mini_cm_close(struct nes_cm_core *, struct nes_cm_node *); +static int mini_cm_accept(struct nes_cm_core *, struct nes_cm_node *); +static int mini_cm_reject(struct nes_cm_core *, struct nes_cm_node *); +static int mini_cm_recv_pkt(struct nes_cm_core *, struct nes_vnic *, struct sk_buff *); +static int mini_cm_dealloc_core(struct nes_cm_core *); +static int mini_cm_get(struct nes_cm_core *); +static int mini_cm_set(struct nes_cm_core *, u32, u32); + +static void form_cm_frame(struct sk_buff *, struct nes_cm_node *, void *, u32, void *, u32, u8); static int add_ref_cm_node(struct nes_cm_node *); static int rem_ref_cm_node(struct nes_cm_core *, struct nes_cm_node *); -static int mini_cm_del_listen(struct nes_cm_core *, struct nes_cm_listener *); +static int nes_cm_disconn_true(struct nes_qp *); +static int nes_cm_post_event(struct nes_cm_event *event); +static int nes_disconnect(struct nes_qp *nesqp, int abrupt); +static void nes_disconnect_worker(struct work_struct *work); + +static int send_mpa_request(struct nes_cm_node *, struct sk_buff *); +static int send_mpa_reject(struct nes_cm_node *); +static int send_syn(struct nes_cm_node *, u32, struct sk_buff *); +static int send_reset(struct nes_cm_node *, struct sk_buff *); +static int send_ack(struct nes_cm_node *cm_node, struct sk_buff *skb); +static int send_fin(struct nes_cm_node *cm_node, struct sk_buff *skb); +static void process_packet(struct nes_cm_node *, struct sk_buff *, struct nes_cm_core *); + +static void active_open_err(struct nes_cm_node *, struct sk_buff *, int); +static void passive_open_err(struct nes_cm_node *, struct sk_buff *, int); +static void cleanup_retrans_entry(struct nes_cm_node *); +static void handle_rcv_mpa(struct nes_cm_node *, struct sk_buff *); +static void free_retrans_entry(struct nes_cm_node *cm_node); +static int handle_tcp_options(struct nes_cm_node *cm_node, struct tcphdr *tcph, struct sk_buff *skb, int optionsize, int passive); + +/* CM event handler functions */ +static void cm_event_connected(struct nes_cm_event *); +static void cm_event_connect_error(struct nes_cm_event *); +static void cm_event_reset(struct nes_cm_event *); +static void cm_event_mpa_req(struct nes_cm_event *); +static void cm_event_mpa_reject(struct nes_cm_event *); +static void handle_recv_entry(struct nes_cm_node *cm_node, u32 rem_node); + +/* MPA build functions */ +static int cm_build_mpa_frame(struct nes_cm_node *, u8 **, u16 *, u8 *, u8); +static void build_mpa_v2(struct nes_cm_node *, void *, u8); +static void build_mpa_v1(struct nes_cm_node *, void *, u8); +static void build_rdma0_msg(struct nes_cm_node *, struct nes_qp **); + +static void print_core(struct nes_cm_core *core); +static void record_ird_ord(struct nes_cm_node *, u16, u16); /* External CM API Interface */ /* instance of function pointers for client API */ @@ -99,7 +148,7 @@ static struct nes_cm_ops nes_cm_api = { mini_cm_set }; -struct nes_cm_core *g_cm_core; +static struct nes_cm_core *g_cm_core; atomic_t cm_connects; atomic_t cm_accepts; @@ -109,12 +158,20 @@ atomic_t cm_connecteds; atomic_t cm_connect_reqs; atomic_t cm_rejects; +int nes_add_ref_cm_node(struct nes_cm_node *cm_node) +{ + return add_ref_cm_node(cm_node); +} +int nes_rem_ref_cm_node(struct nes_cm_node *cm_node) +{ + return rem_ref_cm_node(cm_node->cm_core, cm_node); +} /** * create_event */ -static struct nes_cm_event *create_event(struct nes_cm_node *cm_node, - enum nes_cm_event_type type) +static struct nes_cm_event *create_event(struct nes_cm_node * cm_node, + enum nes_cm_event_type type) { struct nes_cm_event *event; @@ -135,11 +192,11 @@ static struct nes_cm_event *create_event(struct nes_cm_node *cm_node, event->cm_info.loc_port = cm_node->loc_port; event->cm_info.cm_id = cm_node->cm_id; - nes_debug(NES_DBG_CM, "Created event=%p, type=%u, dst_addr=%08x[%x]," - " src_addr=%08x[%x]\n", - event, type, - event->cm_info.loc_addr, event->cm_info.loc_port, - event->cm_info.rem_addr, event->cm_info.rem_port); + nes_debug(NES_DBG_CM, "cm_node=%p Created event=%p, type=%u, " + "dst_addr=%08x[%x], src_addr=%08x[%x]\n", + cm_node, event, type, event->cm_info.loc_addr, + event->cm_info.loc_port, event->cm_info.rem_addr, + event->cm_info.rem_port); nes_cm_post_event(event); return event; @@ -149,27 +206,50 @@ static struct nes_cm_event *create_event(struct nes_cm_node *cm_node, /** * send_mpa_request */ -int send_mpa_request(struct nes_cm_node *cm_node) +static int send_mpa_request(struct nes_cm_node *cm_node, struct sk_buff *skb) { - struct sk_buff *skb; - int ret; + u8 start_addr = 0; + u8 *start_ptr = &start_addr; + u8 **start_buff = &start_ptr; + u16 buff_len = 0; - skb = get_free_pkt(cm_node); if (!skb) { - nes_debug(NES_DBG_CM, "Failed to get a Free pkt\n"); + nes_debug(NES_DBG_CM, "skb set to NULL\n"); return -1; } /* send an MPA Request frame */ - form_cm_frame(skb, cm_node, NULL, 0, &cm_node->mpa_frame, - cm_node->mpa_frame_size, SET_ACK); + cm_build_mpa_frame(cm_node, start_buff, &buff_len, NULL, MPA_KEY_REQUEST); + form_cm_frame(skb, cm_node, NULL, 0, *start_buff, buff_len, SET_ACK); - ret = schedule_nes_timer(cm_node, skb, NES_TIMER_TYPE_SEND, 1, 0); - if (ret < 0) { - return ret; + return schedule_nes_timer(cm_node, skb, NES_TIMER_TYPE_SEND, 1, 0); +} + + + +static int send_mpa_reject(struct nes_cm_node *cm_node) +{ + struct sk_buff *skb = NULL; + u8 start_addr = 0; + u8 *start_ptr = &start_addr; + u8 **start_buff = &start_ptr; + u16 buff_len = 0; + struct ietf_mpa_v1 *mpa_frame; + + skb = dev_alloc_skb(MAX_CM_BUFFER); + if (!skb) { + nes_debug(NES_DBG_CM, "Failed to get a Free pkt\n"); + return -ENOMEM; } - return 0; + /* send an MPA reject frame */ + cm_build_mpa_frame(cm_node, start_buff, &buff_len, NULL, MPA_KEY_REPLY); + mpa_frame = (struct ietf_mpa_v1 *)*start_buff; + mpa_frame->flags |= IETF_MPA_FLAGS_REJECT; + form_cm_frame(skb, cm_node, NULL, 0, *start_buff, buff_len, SET_ACK | SET_FIN); + + cm_node->state = NES_CM_STATE_FIN_WAIT1; + return schedule_nes_timer(cm_node, skb, NES_TIMER_TYPE_SEND, 1, 0); } @@ -177,65 +257,174 @@ int send_mpa_request(struct nes_cm_node *cm_node) * recv_mpa - process a received TCP pkt, we are expecting an * IETF MPA frame */ -static int parse_mpa(struct nes_cm_node *cm_node, u8 *buffer, u32 len) +static int parse_mpa(struct nes_cm_node *cm_node, u8 *buffer, u32 *type, + u32 len) { - struct ietf_mpa_frame *mpa_frame; + struct ietf_mpa_v1 *mpa_frame; + struct ietf_mpa_v2 *mpa_v2_frame; + struct ietf_rtr_msg *rtr_msg; + int mpa_hdr_len; + int priv_data_len; + + *type = NES_MPA_REQUEST_ACCEPT; /* assume req frame is in tcp data payload */ - if (len < sizeof(struct ietf_mpa_frame)) { + if (len < sizeof(struct ietf_mpa_v1)) { nes_debug(NES_DBG_CM, "The received ietf buffer was too small (%x)\n", len); - return -1; + return -EINVAL; } - mpa_frame = (struct ietf_mpa_frame *)buffer; - cm_node->mpa_frame_size = ntohs(mpa_frame->priv_data_len); + /* points to the beginning of the frame, which could be MPA V1 or V2 */ + mpa_frame = (struct ietf_mpa_v1 *)buffer; + mpa_hdr_len = sizeof(struct ietf_mpa_v1); + priv_data_len = ntohs(mpa_frame->priv_data_len); - if (cm_node->mpa_frame_size + sizeof(struct ietf_mpa_frame) != len) { - nes_debug(NES_DBG_CM, "The received ietf buffer was not right" - " complete (%x + %x != %x)\n", - cm_node->mpa_frame_size, (u32)sizeof(struct ietf_mpa_frame), len); - return -1; + /* make sure mpa private data len is less than 512 bytes */ + if (priv_data_len > IETF_MAX_PRIV_DATA_LEN) { + nes_debug(NES_DBG_CM, "The received Length of Private" + " Data field exceeds 512 octets\n"); + return -EINVAL; + } + /* + * make sure MPA receiver interoperate with the + * received MPA version and MPA key information + * + */ + if (mpa_frame->rev != IETF_MPA_V1 && mpa_frame->rev != IETF_MPA_V2) { + nes_debug(NES_DBG_CM, "The received mpa version" + " is not supported\n"); + return -EINVAL; + } + /* + * backwards compatibility only + */ + if (mpa_frame->rev > cm_node->mpa_frame_rev) { + nes_debug(NES_DBG_CM, "The received mpa version" + " can not be interoperated\n"); + return -EINVAL; + } else { + cm_node->mpa_frame_rev = mpa_frame->rev; } - /* copy entire MPA frame to our cm_node's frame */ - memcpy(cm_node->mpa_frame_buf, buffer + sizeof(struct ietf_mpa_frame), - cm_node->mpa_frame_size); + if (cm_node->state != NES_CM_STATE_MPAREQ_SENT) { + if (memcmp(mpa_frame->key, IEFT_MPA_KEY_REQ, IETF_MPA_KEY_SIZE)) { + nes_debug(NES_DBG_CM, "Unexpected MPA Key received \n"); + return -EINVAL; + } + } else { + if (memcmp(mpa_frame->key, IEFT_MPA_KEY_REP, IETF_MPA_KEY_SIZE)) { + nes_debug(NES_DBG_CM, "Unexpected MPA Key received \n"); + return -EINVAL; + } + } - return 0; -} + if (priv_data_len + mpa_hdr_len != len) { + nes_debug(NES_DBG_CM, "The received ietf buffer was not right" + " complete (%x + %x != %x)\n", + priv_data_len, mpa_hdr_len, len); + return -EINVAL; + } + /* make sure it does not exceed the max size */ + if (len > MAX_CM_BUFFER) { + nes_debug(NES_DBG_CM, "The received ietf buffer was too large" + " (%x + %x != %x)\n", + priv_data_len, mpa_hdr_len, len); + return -EINVAL; + } + cm_node->mpa_frame_size = priv_data_len; + + switch (mpa_frame->rev) { + case IETF_MPA_V2: { + u16 ird_size; + u16 ord_size; + u16 rtr_ctrl_ird; + u16 rtr_ctrl_ord; + + mpa_v2_frame = (struct ietf_mpa_v2 *)buffer; + mpa_hdr_len += IETF_RTR_MSG_SIZE; + cm_node->mpa_frame_size -= IETF_RTR_MSG_SIZE; + rtr_msg = &mpa_v2_frame->rtr_msg; + + /* parse rtr message */ + rtr_ctrl_ird = ntohs(rtr_msg->ctrl_ird); + rtr_ctrl_ord = ntohs(rtr_msg->ctrl_ord); + ird_size = rtr_ctrl_ird & IETF_NO_IRD_ORD; + ord_size = rtr_ctrl_ord & IETF_NO_IRD_ORD; + + if (!(rtr_ctrl_ird & IETF_PEER_TO_PEER)) { + /* send reset */ + return -EINVAL; + } + if (ird_size == IETF_NO_IRD_ORD || ord_size == IETF_NO_IRD_ORD) + cm_node->mpav2_ird_ord = IETF_NO_IRD_ORD; + + if (cm_node->mpav2_ird_ord != IETF_NO_IRD_ORD) { + /* responder */ + if (cm_node->state != NES_CM_STATE_MPAREQ_SENT) { + /* we are still negotiating */ + if (ord_size > NES_MAX_IRD) { + cm_node->ird_size = NES_MAX_IRD; + } else { + cm_node->ird_size = ord_size; + if (ord_size == 0 && + (rtr_ctrl_ord & IETF_RDMA0_READ)) { + cm_node->ird_size = 1; + nes_debug(NES_DBG_CM, + "%s: Remote peer doesn't support RDMA0_READ (ord=%u)\n", + __func__, ord_size); + } + } + if (ird_size > NES_MAX_ORD) + cm_node->ord_size = NES_MAX_ORD; + else + cm_node->ord_size = ird_size; + } else { /* initiator */ + if (ord_size > NES_MAX_IRD) { + nes_debug(NES_DBG_CM, + "%s: Unable to support the requested (ord =%u)\n", + __func__, ord_size); + return -EINVAL; + } + cm_node->ird_size = ord_size; -/** - * handle_exception_pkt - process an exception packet. - * We have been in a TSA state, and we have now received SW - * TCP/IP traffic should be a FIN request or IP pkt with options - */ -static int handle_exception_pkt(struct nes_cm_node *cm_node, struct sk_buff *skb) -{ - int ret = 0; - struct tcphdr *tcph = tcp_hdr(skb); + if (ird_size > NES_MAX_ORD) { + cm_node->ord_size = NES_MAX_ORD; + } else { + if (ird_size == 0 && + (rtr_ctrl_ord & IETF_RDMA0_READ)) { + nes_debug(NES_DBG_CM, + "%s: Remote peer doesn't support RDMA0_READ (ird=%u)\n", + __func__, ird_size); + return -EINVAL; + } else { + cm_node->ord_size = ird_size; + } + } + } + } + + if (rtr_ctrl_ord & IETF_RDMA0_READ) { + cm_node->send_rdma0_op = SEND_RDMA_READ_ZERO; - /* first check to see if this a FIN pkt */ - if (tcph->fin) { - /* we need to ACK the FIN request */ - send_ack(cm_node); - - /* check which side we are (client/server) and set next state accordingly */ - if (cm_node->tcp_cntxt.client) - cm_node->state = NES_CM_STATE_CLOSING; - else { - /* we are the server side */ - cm_node->state = NES_CM_STATE_CLOSE_WAIT; - /* since this is a self contained CM we don't wait for */ - /* an APP to close us, just send final FIN immediately */ - ret = send_fin(cm_node, NULL); - cm_node->state = NES_CM_STATE_LAST_ACK; + } else if (rtr_ctrl_ord & IETF_RDMA0_WRITE) { + cm_node->send_rdma0_op = SEND_RDMA_WRITE_ZERO; + } else { /* Not supported RDMA0 operation */ + return -EINVAL; } - } else { - ret = -EINVAL; + break; + } + case IETF_MPA_V1: + default: + break; } - return ret; + /* copy entire MPA frame to our cm_node's frame */ + memcpy(cm_node->mpa_frame_buf, buffer + mpa_hdr_len, cm_node->mpa_frame_size); + + if (mpa_frame->flags & IETF_MPA_FLAGS_REJECT) + *type = NES_MPA_REQUEST_REJECT; + return 0; } @@ -243,8 +432,9 @@ static int handle_exception_pkt(struct nes_cm_node *cm_node, struct sk_buff *skb * form_cm_frame - get a free packet and build empty frame Use * node info to build. */ -struct sk_buff *form_cm_frame(struct sk_buff *skb, struct nes_cm_node *cm_node, - void *options, u32 optionsize, void *data, u32 datasize, u8 flags) +static void form_cm_frame(struct sk_buff *skb, + struct nes_cm_node *cm_node, void *options, u32 optionsize, + void *data, u32 datasize, u8 flags) { struct tcphdr *tcph; struct iphdr *iph; @@ -253,14 +443,14 @@ struct sk_buff *form_cm_frame(struct sk_buff *skb, struct nes_cm_node *cm_node, u16 packetsize = sizeof(*iph); packetsize += sizeof(*tcph); - packetsize += optionsize + datasize; + packetsize += optionsize + datasize; + skb_trim(skb, 0); memset(skb->data, 0x00, ETH_HLEN + sizeof(*iph) + sizeof(*tcph)); - skb->len = 0; buf = skb_put(skb, packetsize + ETH_HLEN); - ethh = (struct ethhdr *) buf; + ethh = (struct ethhdr *)buf; buf += ETH_HLEN; iph = (struct iphdr *)buf; @@ -268,10 +458,12 @@ struct sk_buff *form_cm_frame(struct sk_buff *skb, struct nes_cm_node *cm_node, tcph = (struct tcphdr *)buf; skb_reset_mac_header(skb); skb_set_network_header(skb, ETH_HLEN); - skb_set_transport_header(skb, ETH_HLEN+sizeof(*iph)); + skb_set_transport_header(skb, ETH_HLEN + sizeof(*iph)); buf += sizeof(*tcph); skb->ip_summed = CHECKSUM_PARTIAL; + if (!(cm_node->netdev->features & NETIF_F_IP_CSUM)) + skb->ip_summed = CHECKSUM_NONE; skb->protocol = htons(0x800); skb->data_len = 0; skb->mac_len = ETH_HLEN; @@ -281,37 +473,41 @@ struct sk_buff *form_cm_frame(struct sk_buff *skb, struct nes_cm_node *cm_node, ethh->h_proto = htons(0x0800); iph->version = IPVERSION; - iph->ihl = 5; /* 5 * 4Byte words, IP headr len */ + iph->ihl = 5; /* 5 * 4Byte words, IP headr len */ iph->tos = 0; iph->tot_len = htons(packetsize); iph->id = htons(++cm_node->tcp_cntxt.loc_id); iph->frag_off = htons(0x4000); iph->ttl = 0x40; - iph->protocol = 0x06; /* IPPROTO_TCP */ + iph->protocol = 0x06; /* IPPROTO_TCP */ - iph->saddr = htonl(cm_node->loc_addr); - iph->daddr = htonl(cm_node->rem_addr); + iph->saddr = htonl(cm_node->mapped_loc_addr); + iph->daddr = htonl(cm_node->mapped_rem_addr); - tcph->source = htons(cm_node->loc_port); - tcph->dest = htons(cm_node->rem_port); + tcph->source = htons(cm_node->mapped_loc_port); + tcph->dest = htons(cm_node->mapped_rem_port); tcph->seq = htonl(cm_node->tcp_cntxt.loc_seq_num); if (flags & SET_ACK) { cm_node->tcp_cntxt.loc_ack_num = cm_node->tcp_cntxt.rcv_nxt; tcph->ack_seq = htonl(cm_node->tcp_cntxt.loc_ack_num); tcph->ack = 1; - } else + } else { tcph->ack_seq = 0; + } if (flags & SET_SYN) { cm_node->tcp_cntxt.loc_seq_num++; tcph->syn = 1; - } else - cm_node->tcp_cntxt.loc_seq_num += datasize; /* data (no headers) */ + } else { + cm_node->tcp_cntxt.loc_seq_num += datasize; + } - if (flags & SET_FIN) + if (flags & SET_FIN) { + cm_node->tcp_cntxt.loc_seq_num++; tcph->fin = 1; + } if (flags & SET_RST) tcph->rst = 1; @@ -327,10 +523,101 @@ struct sk_buff *form_cm_frame(struct sk_buff *skb, struct nes_cm_node *cm_node, skb_shinfo(skb)->nr_frags = 0; cm_packets_created++; +} + +/* + * nes_create_sockaddr - Record ip addr and tcp port in a sockaddr struct + */ +static void nes_create_sockaddr(__be32 ip_addr, __be16 port, + struct sockaddr_storage *addr) +{ + struct sockaddr_in *nes_sockaddr = (struct sockaddr_in *)addr; + nes_sockaddr->sin_family = AF_INET; + memcpy(&nes_sockaddr->sin_addr.s_addr, &ip_addr, sizeof(__be32)); + nes_sockaddr->sin_port = port; +} + +/* + * nes_create_mapinfo - Create a mapinfo object in the port mapper data base + */ +static int nes_create_mapinfo(struct nes_cm_info *cm_info) +{ + struct sockaddr_storage local_sockaddr; + struct sockaddr_storage mapped_sockaddr; + + nes_create_sockaddr(htonl(cm_info->loc_addr), htons(cm_info->loc_port), + &local_sockaddr); + nes_create_sockaddr(htonl(cm_info->mapped_loc_addr), + htons(cm_info->mapped_loc_port), &mapped_sockaddr); + + return iwpm_create_mapinfo(&local_sockaddr, + &mapped_sockaddr, RDMA_NL_NES); +} + +/* + * nes_remove_mapinfo - Remove a mapinfo object from the port mapper data base + * and send a remove mapping op message to + * the userspace port mapper + */ +static int nes_remove_mapinfo(u32 loc_addr, u16 loc_port, + u32 mapped_loc_addr, u16 mapped_loc_port) +{ + struct sockaddr_storage local_sockaddr; + struct sockaddr_storage mapped_sockaddr; + + nes_create_sockaddr(htonl(loc_addr), htons(loc_port), &local_sockaddr); + nes_create_sockaddr(htonl(mapped_loc_addr), htons(mapped_loc_port), + &mapped_sockaddr); - return skb; + iwpm_remove_mapinfo(&local_sockaddr, &mapped_sockaddr); + return iwpm_remove_mapping(&local_sockaddr, RDMA_NL_NES); } +/* + * nes_form_pm_msg - Form a port mapper message with mapping info + */ +static void nes_form_pm_msg(struct nes_cm_info *cm_info, + struct iwpm_sa_data *pm_msg) +{ + nes_create_sockaddr(htonl(cm_info->loc_addr), htons(cm_info->loc_port), + &pm_msg->loc_addr); + nes_create_sockaddr(htonl(cm_info->rem_addr), htons(cm_info->rem_port), + &pm_msg->rem_addr); +} + +/* + * nes_form_reg_msg - Form a port mapper message with dev info + */ +static void nes_form_reg_msg(struct nes_vnic *nesvnic, + struct iwpm_dev_data *pm_msg) +{ + memcpy(pm_msg->dev_name, nesvnic->nesibdev->ibdev.name, + IWPM_DEVNAME_SIZE); + memcpy(pm_msg->if_name, nesvnic->netdev->name, IWPM_IFNAME_SIZE); +} + +/* + * nes_record_pm_msg - Save the received mapping info + */ +static void nes_record_pm_msg(struct nes_cm_info *cm_info, + struct iwpm_sa_data *pm_msg) +{ + struct sockaddr_in *mapped_loc_addr = + (struct sockaddr_in *)&pm_msg->mapped_loc_addr; + struct sockaddr_in *mapped_rem_addr = + (struct sockaddr_in *)&pm_msg->mapped_rem_addr; + + if (mapped_loc_addr->sin_family == AF_INET) { + cm_info->mapped_loc_addr = + ntohl(mapped_loc_addr->sin_addr.s_addr); + cm_info->mapped_loc_port = ntohs(mapped_loc_addr->sin_port); + } + if (mapped_rem_addr->sin_family == AF_INET) { + cm_info->mapped_rem_addr = + ntohl(mapped_rem_addr->sin_addr.s_addr); + cm_info->mapped_rem_port = ntohs(mapped_rem_addr->sin_port); + } +} /** * print_core - dump a cm core @@ -342,11 +629,9 @@ static void print_core(struct nes_cm_core *core) if (!core) return; nes_debug(NES_DBG_CM, "---------------------------------------------\n"); - nes_debug(NES_DBG_CM, "Session ID : %u \n", atomic_read(&core->session_id)); - nes_debug(NES_DBG_CM, "State : %u \n", core->state); + nes_debug(NES_DBG_CM, "State : %u \n", core->state); - nes_debug(NES_DBG_CM, "Tx Free cnt : %u \n", skb_queue_len(&core->tx_free_list)); nes_debug(NES_DBG_CM, "Listen Nodes : %u \n", atomic_read(&core->listen_node_cnt)); nes_debug(NES_DBG_CM, "Active Nodes : %u \n", atomic_read(&core->node_cnt)); @@ -355,6 +640,162 @@ static void print_core(struct nes_cm_core *core) nes_debug(NES_DBG_CM, "-------------- end core ---------------\n"); } +static void record_ird_ord(struct nes_cm_node *cm_node, + u16 conn_ird, u16 conn_ord) +{ + if (conn_ird > NES_MAX_IRD) + conn_ird = NES_MAX_IRD; + + if (conn_ord > NES_MAX_ORD) + conn_ord = NES_MAX_ORD; + + cm_node->ird_size = conn_ird; + cm_node->ord_size = conn_ord; +} + +/** + * cm_build_mpa_frame - build a MPA V1 frame or MPA V2 frame + */ +static int cm_build_mpa_frame(struct nes_cm_node *cm_node, u8 **start_buff, + u16 *buff_len, u8 *pci_mem, u8 mpa_key) +{ + int ret = 0; + + *start_buff = (pci_mem) ? pci_mem : &cm_node->mpa_frame_buf[0]; + + switch (cm_node->mpa_frame_rev) { + case IETF_MPA_V1: + *start_buff = (u8 *)*start_buff + sizeof(struct ietf_rtr_msg); + *buff_len = sizeof(struct ietf_mpa_v1) + cm_node->mpa_frame_size; + build_mpa_v1(cm_node, *start_buff, mpa_key); + break; + case IETF_MPA_V2: + *buff_len = sizeof(struct ietf_mpa_v2) + cm_node->mpa_frame_size; + build_mpa_v2(cm_node, *start_buff, mpa_key); + break; + default: + ret = -EINVAL; + } + return ret; +} + +/** + * build_mpa_v2 - build a MPA V2 frame + */ +static void build_mpa_v2(struct nes_cm_node *cm_node, + void *start_addr, u8 mpa_key) +{ + struct ietf_mpa_v2 *mpa_frame = (struct ietf_mpa_v2 *)start_addr; + struct ietf_rtr_msg *rtr_msg = &mpa_frame->rtr_msg; + u16 ctrl_ird; + u16 ctrl_ord; + + /* initialize the upper 5 bytes of the frame */ + build_mpa_v1(cm_node, start_addr, mpa_key); + mpa_frame->flags |= IETF_MPA_V2_FLAG; /* set a bit to indicate MPA V2 */ + mpa_frame->priv_data_len += htons(IETF_RTR_MSG_SIZE); + + /* initialize RTR msg */ + if (cm_node->mpav2_ird_ord == IETF_NO_IRD_ORD) { + ctrl_ird = IETF_NO_IRD_ORD; + ctrl_ord = IETF_NO_IRD_ORD; + } else { + ctrl_ird = cm_node->ird_size & IETF_NO_IRD_ORD; + ctrl_ord = cm_node->ord_size & IETF_NO_IRD_ORD; + } + ctrl_ird |= IETF_PEER_TO_PEER; + ctrl_ird |= IETF_FLPDU_ZERO_LEN; + + switch (mpa_key) { + case MPA_KEY_REQUEST: + ctrl_ord |= IETF_RDMA0_WRITE; + ctrl_ord |= IETF_RDMA0_READ; + break; + case MPA_KEY_REPLY: + switch (cm_node->send_rdma0_op) { + case SEND_RDMA_WRITE_ZERO: + ctrl_ord |= IETF_RDMA0_WRITE; + break; + case SEND_RDMA_READ_ZERO: + ctrl_ord |= IETF_RDMA0_READ; + break; + } + } + rtr_msg->ctrl_ird = htons(ctrl_ird); + rtr_msg->ctrl_ord = htons(ctrl_ord); +} + +/** + * build_mpa_v1 - build a MPA V1 frame + */ +static void build_mpa_v1(struct nes_cm_node *cm_node, void *start_addr, u8 mpa_key) +{ + struct ietf_mpa_v1 *mpa_frame = (struct ietf_mpa_v1 *)start_addr; + + switch (mpa_key) { + case MPA_KEY_REQUEST: + memcpy(mpa_frame->key, IEFT_MPA_KEY_REQ, IETF_MPA_KEY_SIZE); + break; + case MPA_KEY_REPLY: + memcpy(mpa_frame->key, IEFT_MPA_KEY_REP, IETF_MPA_KEY_SIZE); + break; + } + mpa_frame->flags = IETF_MPA_FLAGS_CRC; + mpa_frame->rev = cm_node->mpa_frame_rev; + mpa_frame->priv_data_len = htons(cm_node->mpa_frame_size); +} + +static void build_rdma0_msg(struct nes_cm_node *cm_node, struct nes_qp **nesqp_addr) +{ + u64 u64temp; + struct nes_qp *nesqp = *nesqp_addr; + struct nes_hw_qp_wqe *wqe = &nesqp->hwqp.sq_vbase[0]; + + u64temp = (unsigned long)nesqp->nesuqp_addr; + u64temp |= NES_SW_CONTEXT_ALIGN >> 1; + set_wqe_64bit_value(wqe->wqe_words, NES_IWARP_SQ_WQE_COMP_CTX_LOW_IDX, u64temp); + + wqe->wqe_words[NES_IWARP_SQ_WQE_FRAG0_LOW_IDX] = 0; + wqe->wqe_words[NES_IWARP_SQ_WQE_FRAG0_HIGH_IDX] = 0; + + switch (cm_node->send_rdma0_op) { + case SEND_RDMA_WRITE_ZERO: + nes_debug(NES_DBG_CM, "Sending first write.\n"); + wqe->wqe_words[NES_IWARP_SQ_WQE_MISC_IDX] = + cpu_to_le32(NES_IWARP_SQ_OP_RDMAW); + wqe->wqe_words[NES_IWARP_SQ_WQE_TOTAL_PAYLOAD_IDX] = 0; + wqe->wqe_words[NES_IWARP_SQ_WQE_LENGTH0_IDX] = 0; + wqe->wqe_words[NES_IWARP_SQ_WQE_STAG0_IDX] = 0; + break; + + case SEND_RDMA_READ_ZERO: + default: + if (cm_node->send_rdma0_op != SEND_RDMA_READ_ZERO) + WARN(1, "Unsupported RDMA0 len operation=%u\n", + cm_node->send_rdma0_op); + nes_debug(NES_DBG_CM, "Sending first rdma operation.\n"); + wqe->wqe_words[NES_IWARP_SQ_WQE_MISC_IDX] = + cpu_to_le32(NES_IWARP_SQ_OP_RDMAR); + wqe->wqe_words[NES_IWARP_SQ_WQE_RDMA_TO_LOW_IDX] = 1; + wqe->wqe_words[NES_IWARP_SQ_WQE_RDMA_TO_HIGH_IDX] = 0; + wqe->wqe_words[NES_IWARP_SQ_WQE_RDMA_LENGTH_IDX] = 0; + wqe->wqe_words[NES_IWARP_SQ_WQE_RDMA_STAG_IDX] = 1; + wqe->wqe_words[NES_IWARP_SQ_WQE_STAG0_IDX] = 1; + break; + } + + if (nesqp->sq_kmapped) { + nesqp->sq_kmapped = 0; + kunmap(nesqp->page); + } + + /*use the reserved spot on the WQ for the extra first WQE*/ + nesqp->nesqp_context->ird_ord_sizes &= cpu_to_le32(~(NES_QPCONTEXT_ORDIRD_LSMM_PRESENT | + NES_QPCONTEXT_ORDIRD_WRPDU | + NES_QPCONTEXT_ORDIRD_ALSMM)); + nesqp->skip_lsmm = 1; + nesqp->hwqp.sq_tail = 0; +} /** * schedule_nes_timer @@ -362,20 +803,17 @@ static void print_core(struct nes_cm_core *core) * rem_ref_cm_node(cm_core, cm_node);add_ref_cm_node(cm_node); */ int schedule_nes_timer(struct nes_cm_node *cm_node, struct sk_buff *skb, - enum nes_timer_type type, int send_retrans, - int close_when_complete) + enum nes_timer_type type, int send_retrans, + int close_when_complete) { - unsigned long flags; - struct nes_cm_core *cm_core; + unsigned long flags; + struct nes_cm_core *cm_core = cm_node->cm_core; struct nes_timer_entry *new_send; int ret = 0; - u32 was_timer_set; - if (!cm_node) - return -EINVAL; new_send = kzalloc(sizeof(*new_send), GFP_ATOMIC); if (!new_send) - return -1; + return -ENOMEM; /* new_send->timetosend = currenttime */ new_send->retrycount = NES_DEFAULT_RETRYS; @@ -388,255 +826,262 @@ int schedule_nes_timer(struct nes_cm_node *cm_node, struct sk_buff *skb, new_send->close_when_complete = close_when_complete; if (type == NES_TIMER_TYPE_CLOSE) { - new_send->timetosend += (HZ/2); /* TODO: decide on the correct value here */ - spin_lock_irqsave(&cm_node->recv_list_lock, flags); - list_add_tail(&new_send->list, &cm_node->recv_list); - spin_unlock_irqrestore(&cm_node->recv_list_lock, flags); + new_send->timetosend += (HZ / 10); + if (cm_node->recv_entry) { + kfree(new_send); + WARN_ON(1); + return -EINVAL; + } + cm_node->recv_entry = new_send; } if (type == NES_TIMER_TYPE_SEND) { - new_send->seq_num = htonl(tcp_hdr(skb)->seq); + new_send->seq_num = ntohl(tcp_hdr(skb)->seq); atomic_inc(&new_send->skb->users); + spin_lock_irqsave(&cm_node->retrans_list_lock, flags); + cm_node->send_entry = new_send; + add_ref_cm_node(cm_node); + spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); + new_send->timetosend = jiffies + NES_RETRY_TIMEOUT; ret = nes_nic_cm_xmit(new_send->skb, cm_node->netdev); if (ret != NETDEV_TX_OK) { - nes_debug(NES_DBG_CM, "Error sending packet %p (jiffies = %lu)\n", - new_send, jiffies); - atomic_dec(&new_send->skb->users); + nes_debug(NES_DBG_CM, "Error sending packet %p " + "(jiffies = %lu)\n", new_send, jiffies); new_send->timetosend = jiffies; + ret = NETDEV_TX_OK; } else { cm_packets_sent++; if (!send_retrans) { + cleanup_retrans_entry(cm_node); if (close_when_complete) - rem_ref_cm_node(cm_node->cm_core, cm_node); - dev_kfree_skb_any(new_send->skb); - kfree(new_send); + rem_ref_cm_node(cm_core, cm_node); return ret; } - new_send->timetosend = jiffies + NES_RETRY_TIMEOUT; } - spin_lock_irqsave(&cm_node->retrans_list_lock, flags); - list_add_tail(&new_send->list, &cm_node->retrans_list); - spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); - } - if (type == NES_TIMER_TYPE_RECV) { - new_send->seq_num = htonl(tcp_hdr(skb)->seq); - new_send->timetosend = jiffies; - spin_lock_irqsave(&cm_node->recv_list_lock, flags); - list_add_tail(&new_send->list, &cm_node->recv_list); - spin_unlock_irqrestore(&cm_node->recv_list_lock, flags); } - cm_core = cm_node->cm_core; - was_timer_set = timer_pending(&cm_core->tcp_timer); - - if (!was_timer_set) { - cm_core->tcp_timer.expires = new_send->timetosend; - add_timer(&cm_core->tcp_timer); - } + if (!timer_pending(&cm_core->tcp_timer)) + mod_timer(&cm_core->tcp_timer, new_send->timetosend); return ret; } +static void nes_retrans_expired(struct nes_cm_node *cm_node) +{ + struct iw_cm_id *cm_id = cm_node->cm_id; + enum nes_cm_node_state state = cm_node->state; + cm_node->state = NES_CM_STATE_CLOSED; + + switch (state) { + case NES_CM_STATE_SYN_RCVD: + case NES_CM_STATE_CLOSING: + rem_ref_cm_node(cm_node->cm_core, cm_node); + break; + case NES_CM_STATE_LAST_ACK: + case NES_CM_STATE_FIN_WAIT1: + if (cm_node->cm_id) + cm_id->rem_ref(cm_id); + send_reset(cm_node, NULL); + break; + default: + add_ref_cm_node(cm_node); + send_reset(cm_node, NULL); + create_event(cm_node, NES_CM_EVENT_ABORTED); + } +} + +static void handle_recv_entry(struct nes_cm_node *cm_node, u32 rem_node) +{ + struct nes_timer_entry *recv_entry = cm_node->recv_entry; + struct iw_cm_id *cm_id = cm_node->cm_id; + struct nes_qp *nesqp; + unsigned long qplockflags; + + if (!recv_entry) + return; + nesqp = (struct nes_qp *)recv_entry->skb; + if (nesqp) { + spin_lock_irqsave(&nesqp->lock, qplockflags); + if (nesqp->cm_id) { + nes_debug(NES_DBG_CM, "QP%u: cm_id = %p, " + "refcount = %d: HIT A " + "NES_TIMER_TYPE_CLOSE with something " + "to do!!!\n", nesqp->hwqp.qp_id, cm_id, + atomic_read(&nesqp->refcount)); + nesqp->hw_tcp_state = NES_AEQE_TCP_STATE_CLOSED; + nesqp->last_aeq = NES_AEQE_AEID_RESET_SENT; + nesqp->ibqp_state = IB_QPS_ERR; + spin_unlock_irqrestore(&nesqp->lock, qplockflags); + nes_cm_disconn(nesqp); + } else { + spin_unlock_irqrestore(&nesqp->lock, qplockflags); + nes_debug(NES_DBG_CM, "QP%u: cm_id = %p, " + "refcount = %d: HIT A " + "NES_TIMER_TYPE_CLOSE with nothing " + "to do!!!\n", nesqp->hwqp.qp_id, cm_id, + atomic_read(&nesqp->refcount)); + } + } else if (rem_node) { + /* TIME_WAIT state */ + rem_ref_cm_node(cm_node->cm_core, cm_node); + } + if (cm_node->cm_id) + cm_id->rem_ref(cm_id); + kfree(recv_entry); + cm_node->recv_entry = NULL; +} /** * nes_cm_timer_tick */ -void nes_cm_timer_tick(unsigned long pass) +static void nes_cm_timer_tick(unsigned long pass) { - unsigned long flags, qplockflags; + unsigned long flags; unsigned long nexttimeout = jiffies + NES_LONG_TIME; - struct iw_cm_id *cm_id; struct nes_cm_node *cm_node; struct nes_timer_entry *send_entry, *recv_entry; - struct list_head *list_core, *list_core_temp; - struct list_head *list_node, *list_node_temp; + struct list_head *list_core_temp; + struct list_head *list_node; struct nes_cm_core *cm_core = g_cm_core; - struct nes_qp *nesqp; - struct sk_buff *skb; u32 settimer = 0; + unsigned long timetosend; int ret = NETDEV_TX_OK; - int node_done; + struct list_head timer_list; + + INIT_LIST_HEAD(&timer_list); spin_lock_irqsave(&cm_core->ht_lock, flags); - list_for_each_safe(list_node, list_core_temp, &cm_core->connected_nodes) { + list_for_each_safe(list_node, list_core_temp, + &cm_core->connected_nodes) { cm_node = container_of(list_node, struct nes_cm_node, list); - add_ref_cm_node(cm_node); - spin_unlock_irqrestore(&cm_core->ht_lock, flags); - spin_lock_irqsave(&cm_node->recv_list_lock, flags); - list_for_each_safe(list_core, list_node_temp, &cm_node->recv_list) { - recv_entry = container_of(list_core, struct nes_timer_entry, list); - if ((time_after(recv_entry->timetosend, jiffies)) && - (recv_entry->type == NES_TIMER_TYPE_CLOSE)) { - if (nexttimeout > recv_entry->timetosend || !settimer) { + if ((cm_node->recv_entry) || (cm_node->send_entry)) { + add_ref_cm_node(cm_node); + list_add(&cm_node->timer_entry, &timer_list); + } + } + spin_unlock_irqrestore(&cm_core->ht_lock, flags); + + list_for_each_safe(list_node, list_core_temp, &timer_list) { + cm_node = container_of(list_node, struct nes_cm_node, + timer_entry); + recv_entry = cm_node->recv_entry; + + if (recv_entry) { + if (time_after(recv_entry->timetosend, jiffies)) { + if (nexttimeout > recv_entry->timetosend || + !settimer) { nexttimeout = recv_entry->timetosend; settimer = 1; } - continue; - } - list_del(&recv_entry->list); - cm_id = cm_node->cm_id; - spin_unlock_irqrestore(&cm_node->recv_list_lock, flags); - if (recv_entry->type == NES_TIMER_TYPE_CLOSE) { - nesqp = (struct nes_qp *)recv_entry->skb; - spin_lock_irqsave(&nesqp->lock, qplockflags); - if (nesqp->cm_id) { - nes_debug(NES_DBG_CM, "QP%u: cm_id = %p, refcount = %d: " - "****** HIT A NES_TIMER_TYPE_CLOSE" - " with something to do!!! ******\n", - nesqp->hwqp.qp_id, cm_id, - atomic_read(&nesqp->refcount)); - nesqp->hw_tcp_state = NES_AEQE_TCP_STATE_CLOSED; - nesqp->last_aeq = NES_AEQE_AEID_RESET_SENT; - nesqp->ibqp_state = IB_QPS_ERR; - spin_unlock_irqrestore(&nesqp->lock, qplockflags); - nes_cm_disconn(nesqp); - } else { - spin_unlock_irqrestore(&nesqp->lock, qplockflags); - nes_debug(NES_DBG_CM, "QP%u: cm_id = %p, refcount = %d:" - " ****** HIT A NES_TIMER_TYPE_CLOSE" - " with nothing to do!!! ******\n", - nesqp->hwqp.qp_id, cm_id, - atomic_read(&nesqp->refcount)); - nes_rem_ref(&nesqp->ibqp); - } - if (cm_id) - cm_id->rem_ref(cm_id); + } else { + handle_recv_entry(cm_node, 1); } - kfree(recv_entry); - spin_lock_irqsave(&cm_node->recv_list_lock, flags); } - spin_unlock_irqrestore(&cm_node->recv_list_lock, flags); spin_lock_irqsave(&cm_node->retrans_list_lock, flags); - node_done = 0; - list_for_each_safe(list_core, list_node_temp, &cm_node->retrans_list) { - if (node_done) { + do { + send_entry = cm_node->send_entry; + if (!send_entry) break; - } - send_entry = container_of(list_core, struct nes_timer_entry, list); if (time_after(send_entry->timetosend, jiffies)) { if (cm_node->state != NES_CM_STATE_TSA) { - if ((nexttimeout > send_entry->timetosend) || !settimer) { - nexttimeout = send_entry->timetosend; + if ((nexttimeout > + send_entry->timetosend) || + !settimer) { + nexttimeout = + send_entry->timetosend; settimer = 1; } - node_done = 1; - continue; } else { - list_del(&send_entry->list); - skb = send_entry->skb; - spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); - dev_kfree_skb_any(skb); - kfree(send_entry); - spin_lock_irqsave(&cm_node->retrans_list_lock, flags); - continue; + free_retrans_entry(cm_node); } + break; } - if (send_entry->type == NES_TIMER_NODE_CLEANUP) { - list_del(&send_entry->list); - spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); - kfree(send_entry); - spin_lock_irqsave(&cm_node->retrans_list_lock, flags); - continue; - } - if ((send_entry->seq_num < cm_node->tcp_cntxt.rem_ack_num) || - (cm_node->state == NES_CM_STATE_TSA) || - (cm_node->state == NES_CM_STATE_CLOSED)) { - skb = send_entry->skb; - list_del(&send_entry->list); - spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); - kfree(send_entry); - dev_kfree_skb_any(skb); - spin_lock_irqsave(&cm_node->retrans_list_lock, flags); - continue; + + if ((cm_node->state == NES_CM_STATE_TSA) || + (cm_node->state == NES_CM_STATE_CLOSED)) { + free_retrans_entry(cm_node); + break; } - if (!send_entry->retranscount || !send_entry->retrycount) { + if (!send_entry->retranscount || + !send_entry->retrycount) { cm_packets_dropped++; - skb = send_entry->skb; - list_del(&send_entry->list); - spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); - dev_kfree_skb_any(skb); - kfree(send_entry); - if (cm_node->state == NES_CM_STATE_SYN_RCVD) { - /* this node never even generated an indication up to the cm */ - rem_ref_cm_node(cm_core, cm_node); - } else { - cm_node->state = NES_CM_STATE_CLOSED; - create_event(cm_node, NES_CM_EVENT_ABORTED); - } - spin_lock_irqsave(&cm_node->retrans_list_lock, flags); - continue; + free_retrans_entry(cm_node); + + spin_unlock_irqrestore( + &cm_node->retrans_list_lock, flags); + nes_retrans_expired(cm_node); + cm_node->state = NES_CM_STATE_CLOSED; + spin_lock_irqsave(&cm_node->retrans_list_lock, + flags); + break; } - /* this seems like the correct place, but leave send entry unprotected */ - // spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); atomic_inc(&send_entry->skb->users); cm_packets_retrans++; - nes_debug(NES_DBG_CM, "Retransmitting send_entry %p for node %p," - " jiffies = %lu, time to send = %lu, retranscount = %u, " - "send_entry->seq_num = 0x%08X, cm_node->tcp_cntxt.rem_ack_num = 0x%08X\n", - send_entry, cm_node, jiffies, send_entry->timetosend, send_entry->retranscount, - send_entry->seq_num, cm_node->tcp_cntxt.rem_ack_num); - - spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); + nes_debug(NES_DBG_CM, "Retransmitting send_entry %p " + "for node %p, jiffies = %lu, time to send = " + "%lu, retranscount = %u, send_entry->seq_num = " + "0x%08X, cm_node->tcp_cntxt.rem_ack_num = " + "0x%08X\n", send_entry, cm_node, jiffies, + send_entry->timetosend, + send_entry->retranscount, + send_entry->seq_num, + cm_node->tcp_cntxt.rem_ack_num); + + spin_unlock_irqrestore(&cm_node->retrans_list_lock, + flags); ret = nes_nic_cm_xmit(send_entry->skb, cm_node->netdev); + spin_lock_irqsave(&cm_node->retrans_list_lock, flags); if (ret != NETDEV_TX_OK) { + nes_debug(NES_DBG_CM, "rexmit failed for " + "node=%p\n", cm_node); cm_packets_bounced++; - atomic_dec(&send_entry->skb->users); send_entry->retrycount--; nexttimeout = jiffies + NES_SHORT_TIME; settimer = 1; - node_done = 1; - spin_lock_irqsave(&cm_node->retrans_list_lock, flags); - continue; + break; } else { cm_packets_sent++; } - spin_lock_irqsave(&cm_node->retrans_list_lock, flags); - list_del(&send_entry->list); - nes_debug(NES_DBG_CM, "Packet Sent: retrans count = %u, retry count = %u.\n", - send_entry->retranscount, send_entry->retrycount); + nes_debug(NES_DBG_CM, "Packet Sent: retrans count = " + "%u, retry count = %u.\n", + send_entry->retranscount, + send_entry->retrycount); if (send_entry->send_retrans) { send_entry->retranscount--; - send_entry->timetosend = jiffies + NES_RETRY_TIMEOUT; - if (nexttimeout > send_entry->timetosend || !settimer) { + timetosend = (NES_RETRY_TIMEOUT << + (NES_DEFAULT_RETRANS - send_entry->retranscount)); + + send_entry->timetosend = jiffies + + min(timetosend, NES_MAX_TIMEOUT); + if (nexttimeout > send_entry->timetosend || + !settimer) { nexttimeout = send_entry->timetosend; settimer = 1; } - list_add(&send_entry->list, &cm_node->retrans_list); - continue; } else { int close_when_complete; - skb = send_entry->skb; - close_when_complete = send_entry->close_when_complete; - spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); - if (close_when_complete) { - BUG_ON(atomic_read(&cm_node->ref_count) == 1); - rem_ref_cm_node(cm_core, cm_node); - } - dev_kfree_skb_any(skb); - kfree(send_entry); - spin_lock_irqsave(&cm_node->retrans_list_lock, flags); - continue; + close_when_complete = + send_entry->close_when_complete; + nes_debug(NES_DBG_CM, "cm_node=%p state=%d\n", + cm_node, cm_node->state); + free_retrans_entry(cm_node); + if (close_when_complete) + rem_ref_cm_node(cm_node->cm_core, + cm_node); } - } - spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); + } while (0); - rem_ref_cm_node(cm_core, cm_node); - - spin_lock_irqsave(&cm_core->ht_lock, flags); - if (ret != NETDEV_TX_OK) - break; + spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); + rem_ref_cm_node(cm_node->cm_core, cm_node); } - spin_unlock_irqrestore(&cm_core->ht_lock, flags); if (settimer) { - if (!timer_pending(&cm_core->tcp_timer)) { - cm_core->tcp_timer.expires = nexttimeout; - add_timer(&cm_core->tcp_timer); - } + if (!timer_pending(&cm_core->tcp_timer)) + mod_timer(&cm_core->tcp_timer, nexttimeout); } } @@ -644,14 +1089,14 @@ void nes_cm_timer_tick(unsigned long pass) /** * send_syn */ -int send_syn(struct nes_cm_node *cm_node, u32 sendack) +static int send_syn(struct nes_cm_node *cm_node, u32 sendack, + struct sk_buff *skb) { int ret; int flags = SET_SYN; - struct sk_buff *skb; char optionsbuffer[sizeof(struct option_mss) + - sizeof(struct option_windowscale) + - sizeof(struct option_base) + 1]; + sizeof(struct option_windowscale) + sizeof(struct option_base) + + TCP_OPTIONS_PADDING]; int optionssize = 0; /* Sending MSS option */ @@ -672,8 +1117,7 @@ int send_syn(struct nes_cm_node *cm_node, u32 sendack) options->as_windowscale.shiftcount = cm_node->tcp_cntxt.rcv_wscale; optionssize += sizeof(struct option_windowscale); - if (sendack && !(NES_DRV_OPT_SUPRESS_OPTION_BC & nes_drv_opt) - ) { + if (sendack && !(NES_DRV_OPT_SUPRESS_OPTION_BC & nes_drv_opt)) { options = (union all_known_options *)&optionsbuffer[optionssize]; options->as_base.optionnum = OPTION_NUMBER_WRITE0; options->as_base.length = sizeof(struct option_base); @@ -691,7 +1135,8 @@ int send_syn(struct nes_cm_node *cm_node, u32 sendack) options->as_end = OPTION_NUMBER_END; optionssize += 1; - skb = get_free_pkt(cm_node); + if (!skb) + skb = dev_alloc_skb(MAX_CM_BUFFER); if (!skb) { nes_debug(NES_DBG_CM, "Failed to get a Free pkt\n"); return -1; @@ -710,18 +1155,18 @@ int send_syn(struct nes_cm_node *cm_node, u32 sendack) /** * send_reset */ -int send_reset(struct nes_cm_node *cm_node) +static int send_reset(struct nes_cm_node *cm_node, struct sk_buff *skb) { int ret; - struct sk_buff *skb = get_free_pkt(cm_node); int flags = SET_RST | SET_ACK; + if (!skb) + skb = dev_alloc_skb(MAX_CM_BUFFER); if (!skb) { nes_debug(NES_DBG_CM, "Failed to get a Free pkt\n"); - return -1; + return -ENOMEM; } - add_ref_cm_node(cm_node); form_cm_frame(skb, cm_node, NULL, 0, NULL, 0, flags); ret = schedule_nes_timer(cm_node, skb, NES_TIMER_TYPE_SEND, 0, 1); @@ -732,10 +1177,12 @@ int send_reset(struct nes_cm_node *cm_node) /** * send_ack */ -int send_ack(struct nes_cm_node *cm_node) +static int send_ack(struct nes_cm_node *cm_node, struct sk_buff *skb) { int ret; - struct sk_buff *skb = get_free_pkt(cm_node); + + if (!skb) + skb = dev_alloc_skb(MAX_CM_BUFFER); if (!skb) { nes_debug(NES_DBG_CM, "Failed to get a Free pkt\n"); @@ -752,13 +1199,13 @@ int send_ack(struct nes_cm_node *cm_node) /** * send_fin */ -int send_fin(struct nes_cm_node *cm_node, struct sk_buff *skb) +static int send_fin(struct nes_cm_node *cm_node, struct sk_buff *skb) { int ret; /* if we didn't get a frame get one */ if (!skb) - skb = get_free_pkt(cm_node); + skb = dev_alloc_skb(MAX_CM_BUFFER); if (!skb) { nes_debug(NES_DBG_CM, "Failed to get a Free pkt\n"); @@ -773,78 +1220,32 @@ int send_fin(struct nes_cm_node *cm_node, struct sk_buff *skb) /** - * get_free_pkt - */ -struct sk_buff *get_free_pkt(struct nes_cm_node *cm_node) -{ - struct sk_buff *skb, *new_skb; - - /* check to see if we need to repopulate the free tx pkt queue */ - if (skb_queue_len(&cm_node->cm_core->tx_free_list) < NES_CM_FREE_PKT_LO_WATERMARK) { - while (skb_queue_len(&cm_node->cm_core->tx_free_list) < - cm_node->cm_core->free_tx_pkt_max) { - /* replace the frame we took, we won't get it back */ - new_skb = dev_alloc_skb(cm_node->cm_core->mtu); - BUG_ON(!new_skb); - /* add a replacement frame to the free tx list head */ - skb_queue_head(&cm_node->cm_core->tx_free_list, new_skb); - } - } - - skb = skb_dequeue(&cm_node->cm_core->tx_free_list); - - return skb; -} - - -/** - * make_hashkey - generate hash key from node tuple - */ -static inline int make_hashkey(u16 loc_port, nes_addr_t loc_addr, u16 rem_port, - nes_addr_t rem_addr) -{ - u32 hashkey = 0; - - hashkey = loc_addr + rem_addr + loc_port + rem_port; - hashkey = (hashkey % NES_CM_HASHTABLE_SIZE); - - return hashkey; -} - - -/** * find_node - find a cm node that matches the reference cm node */ static struct nes_cm_node *find_node(struct nes_cm_core *cm_core, - u16 rem_port, nes_addr_t rem_addr, u16 loc_port, nes_addr_t loc_addr) + u16 rem_port, nes_addr_t rem_addr, u16 loc_port, nes_addr_t loc_addr) { unsigned long flags; - u32 hashkey; - struct list_head *list_pos; struct list_head *hte; struct nes_cm_node *cm_node; - /* make a hash index key for this packet */ - hashkey = make_hashkey(loc_port, loc_addr, rem_port, rem_addr); - /* get a handle on the hte */ hte = &cm_core->connected_nodes; - nes_debug(NES_DBG_CM, "Searching for an owner node:%x:%x from core %p->%p\n", - loc_addr, loc_port, cm_core, hte); - /* walk list and find cm_node associated with this session ID */ spin_lock_irqsave(&cm_core->ht_lock, flags); - list_for_each(list_pos, hte) { - cm_node = container_of(list_pos, struct nes_cm_node, list); + list_for_each_entry(cm_node, hte, list) { /* compare quad, return node handle if a match */ nes_debug(NES_DBG_CM, "finding node %x:%x =? %x:%x ^ %x:%x =? %x:%x\n", - cm_node->loc_addr, cm_node->loc_port, - loc_addr, loc_port, - cm_node->rem_addr, cm_node->rem_port, - rem_addr, rem_port); - if ((cm_node->loc_addr == loc_addr) && (cm_node->loc_port == loc_port) && - (cm_node->rem_addr == rem_addr) && (cm_node->rem_port == rem_port)) { + cm_node->loc_addr, cm_node->loc_port, + loc_addr, loc_port, + cm_node->rem_addr, cm_node->rem_port, + rem_addr, rem_port); + if ((cm_node->mapped_loc_addr == loc_addr) && + (cm_node->mapped_loc_port == loc_port) && + (cm_node->mapped_rem_addr == rem_addr) && + (cm_node->mapped_rem_port == rem_port)) { + add_ref_cm_node(cm_node); spin_unlock_irqrestore(&cm_core->ht_lock, flags); return cm_node; @@ -861,21 +1262,29 @@ static struct nes_cm_node *find_node(struct nes_cm_core *cm_core, * find_listener - find a cm node listening on this addr-port pair */ static struct nes_cm_listener *find_listener(struct nes_cm_core *cm_core, - nes_addr_t dst_addr, u16 dst_port, enum nes_cm_listener_state listener_state) + nes_addr_t dst_addr, u16 dst_port, + enum nes_cm_listener_state listener_state, int local) { unsigned long flags; - struct list_head *listen_list; struct nes_cm_listener *listen_node; + nes_addr_t listen_addr; + u16 listen_port; /* walk list and find cm_node associated with this session ID */ spin_lock_irqsave(&cm_core->listen_list_lock, flags); - list_for_each(listen_list, &cm_core->listen_list.list) { - listen_node = container_of(listen_list, struct nes_cm_listener, list); + list_for_each_entry(listen_node, &cm_core->listen_list.list, list) { + if (local) { + listen_addr = listen_node->loc_addr; + listen_port = listen_node->loc_port; + } else { + listen_addr = listen_node->mapped_loc_addr; + listen_port = listen_node->mapped_loc_port; + } /* compare node pair, return node handle if a match */ - if (((listen_node->loc_addr == dst_addr) || - listen_node->loc_addr == 0x00000000) && - (listen_node->loc_port == dst_port) && - (listener_state & listen_node->listener_state)) { + if (((listen_addr == dst_addr) || + listen_addr == 0x00000000) && + (listen_port == dst_port) && + (listener_state & listen_node->listener_state)) { atomic_inc(&listen_node->ref_count); spin_unlock_irqrestore(&cm_core->listen_list_lock, flags); return listen_node; @@ -883,32 +1292,23 @@ static struct nes_cm_listener *find_listener(struct nes_cm_core *cm_core, } spin_unlock_irqrestore(&cm_core->listen_list_lock, flags); - nes_debug(NES_DBG_CM, "Unable to find listener- %x:%x\n", - dst_addr, dst_port); - /* no listener */ return NULL; } - /** * add_hte_node - add a cm node to the hash table */ static int add_hte_node(struct nes_cm_core *cm_core, struct nes_cm_node *cm_node) { unsigned long flags; - u32 hashkey; struct list_head *hte; if (!cm_node || !cm_core) return -EINVAL; - nes_debug(NES_DBG_CM, "Adding Node to Active Connection HT\n"); - - /* first, make an index into our hash table */ - hashkey = make_hashkey(cm_node->loc_port, cm_node->loc_addr, - cm_node->rem_port, cm_node->rem_addr); - cm_node->hashkey = hashkey; + nes_debug(NES_DBG_CM, "Adding Node %p to Active Connection HT\n", + cm_node); spin_lock_irqsave(&cm_core->ht_lock, flags); @@ -927,10 +1327,86 @@ static int add_hte_node(struct nes_cm_core *cm_core, struct nes_cm_node *cm_node * mini_cm_dec_refcnt_listen */ static int mini_cm_dec_refcnt_listen(struct nes_cm_core *cm_core, - struct nes_cm_listener *listener, int free_hanging_nodes) + struct nes_cm_listener *listener, int free_hanging_nodes) { - int ret = 1; + int ret = -EINVAL; + int err = 0; unsigned long flags; + struct list_head *list_pos = NULL; + struct list_head *list_temp = NULL; + struct nes_cm_node *cm_node = NULL; + struct list_head reset_list; + + nes_debug(NES_DBG_CM, "attempting listener= %p free_nodes= %d, " + "refcnt=%d\n", listener, free_hanging_nodes, + atomic_read(&listener->ref_count)); + /* free non-accelerated child nodes for this listener */ + INIT_LIST_HEAD(&reset_list); + if (free_hanging_nodes) { + spin_lock_irqsave(&cm_core->ht_lock, flags); + list_for_each_safe(list_pos, list_temp, + &g_cm_core->connected_nodes) { + cm_node = container_of(list_pos, struct nes_cm_node, + list); + if ((cm_node->listener == listener) && + (!cm_node->accelerated)) { + add_ref_cm_node(cm_node); + list_add(&cm_node->reset_entry, &reset_list); + } + } + spin_unlock_irqrestore(&cm_core->ht_lock, flags); + } + + list_for_each_safe(list_pos, list_temp, &reset_list) { + cm_node = container_of(list_pos, struct nes_cm_node, + reset_entry); + { + struct nes_cm_node *loopback = cm_node->loopbackpartner; + enum nes_cm_node_state old_state; + if (NES_CM_STATE_FIN_WAIT1 <= cm_node->state) { + rem_ref_cm_node(cm_node->cm_core, cm_node); + } else { + if (!loopback) { + cleanup_retrans_entry(cm_node); + err = send_reset(cm_node, NULL); + if (err) { + cm_node->state = + NES_CM_STATE_CLOSED; + WARN_ON(1); + } else { + old_state = cm_node->state; + cm_node->state = NES_CM_STATE_LISTENER_DESTROYED; + if (old_state != NES_CM_STATE_MPAREQ_RCVD) + rem_ref_cm_node( + cm_node->cm_core, + cm_node); + } + } else { + struct nes_cm_event event; + + event.cm_node = loopback; + event.cm_info.rem_addr = + loopback->rem_addr; + event.cm_info.loc_addr = + loopback->loc_addr; + event.cm_info.rem_port = + loopback->rem_port; + event.cm_info.loc_port = + loopback->loc_port; + event.cm_info.cm_id = loopback->cm_id; + add_ref_cm_node(loopback); + loopback->state = NES_CM_STATE_CLOSED; + cm_event_connect_error(&event); + cm_node->state = NES_CM_STATE_LISTENER_DESTROYED; + + rem_ref_cm_node(cm_node->cm_core, + cm_node); + + } + } + } + } + spin_lock_irqsave(&cm_core->listen_list_lock, flags); if (!atomic_dec_return(&listener->ref_count)) { list_del(&listener->list); @@ -941,8 +1417,18 @@ static int mini_cm_dec_refcnt_listen(struct nes_cm_core *cm_core, spin_unlock_irqrestore(&cm_core->listen_list_lock, flags); if (listener->nesvnic) { - nes_manage_apbvt(listener->nesvnic, listener->loc_port, - PCI_FUNC(listener->nesvnic->nesdev->pcidev->devfn), NES_MANAGE_APBVT_DEL); + nes_manage_apbvt(listener->nesvnic, + listener->mapped_loc_port, + PCI_FUNC(listener->nesvnic->nesdev->pcidev->devfn), + NES_MANAGE_APBVT_DEL); + + nes_remove_mapinfo(listener->loc_addr, + listener->loc_port, + listener->mapped_loc_addr, + listener->mapped_loc_port); + nes_debug(NES_DBG_NLMSG, + "Delete APBVT mapped_loc_port = %04X\n", + listener->mapped_loc_port); } nes_debug(NES_DBG_CM, "destroying listener (%p)\n", listener); @@ -950,15 +1436,15 @@ static int mini_cm_dec_refcnt_listen(struct nes_cm_core *cm_core, kfree(listener); listener = NULL; ret = 0; - cm_listens_destroyed++; + atomic_inc(&cm_listens_destroyed); } else { spin_unlock_irqrestore(&cm_core->listen_list_lock, flags); } if (listener) { if (atomic_read(&listener->pend_accepts_cnt) > 0) nes_debug(NES_DBG_CM, "destroying listener (%p)" - " with non-zero pending accepts=%u\n", - listener, atomic_read(&listener->pend_accepts_cnt)); + " with non-zero pending accepts=%u\n", + listener, atomic_read(&listener->pend_accepts_cnt)); } return ret; @@ -969,7 +1455,7 @@ static int mini_cm_dec_refcnt_listen(struct nes_cm_core *cm_core, * mini_cm_del_listen */ static int mini_cm_del_listen(struct nes_cm_core *cm_core, - struct nes_cm_listener *listener) + struct nes_cm_listener *listener) { listener->listener_state = NES_CM_LISTENER_PASSIVE_STATE; listener->cm_id = NULL; /* going to be destroyed pretty soon */ @@ -981,57 +1467,95 @@ static int mini_cm_del_listen(struct nes_cm_core *cm_core, * mini_cm_accelerated */ static inline int mini_cm_accelerated(struct nes_cm_core *cm_core, - struct nes_cm_node *cm_node) + struct nes_cm_node *cm_node) { - u32 was_timer_set; cm_node->accelerated = 1; if (cm_node->accept_pend) { BUG_ON(!cm_node->listener); atomic_dec(&cm_node->listener->pend_accepts_cnt); + cm_node->accept_pend = 0; BUG_ON(atomic_read(&cm_node->listener->pend_accepts_cnt) < 0); } - was_timer_set = timer_pending(&cm_core->tcp_timer); - if (!was_timer_set) { - cm_core->tcp_timer.expires = jiffies + NES_SHORT_TIME; - add_timer(&cm_core->tcp_timer); - } + if (!timer_pending(&cm_core->tcp_timer)) + mod_timer(&cm_core->tcp_timer, (jiffies + NES_SHORT_TIME)); return 0; } /** - * nes_addr_send_arp + * nes_addr_resolve_neigh */ -static void nes_addr_send_arp(u32 dst_ip) +static int nes_addr_resolve_neigh(struct nes_vnic *nesvnic, u32 dst_ip, int arpindex) { struct rtable *rt; - struct flowi fl; + struct neighbour *neigh; + int rc = arpindex; + struct net_device *netdev; + struct nes_adapter *nesadapter = nesvnic->nesdev->nesadapter; + + rt = ip_route_output(&init_net, htonl(dst_ip), 0, 0, 0); + if (IS_ERR(rt)) { + printk(KERN_ERR "%s: ip_route_output_key failed for 0x%08X\n", + __func__, dst_ip); + return rc; + } - memset(&fl, 0, sizeof fl); - fl.nl_u.ip4_u.daddr = htonl(dst_ip); - if (ip_route_output_key(&init_net, &rt, &fl)) { - printk("%s: ip_route_output_key failed for 0x%08X\n", - __FUNCTION__, dst_ip); - return; + if (netif_is_bond_slave(nesvnic->netdev)) + netdev = netdev_master_upper_dev_get(nesvnic->netdev); + else + netdev = nesvnic->netdev; + + neigh = neigh_lookup(&arp_tbl, &rt->rt_gateway, netdev); + + rcu_read_lock(); + if (neigh) { + if (neigh->nud_state & NUD_VALID) { + nes_debug(NES_DBG_CM, "Neighbor MAC address for 0x%08X" + " is %pM, Gateway is 0x%08X \n", dst_ip, + neigh->ha, ntohl(rt->rt_gateway)); + + if (arpindex >= 0) { + if (ether_addr_equal(nesadapter->arp_table[arpindex].mac_addr, neigh->ha)) { + /* Mac address same as in nes_arp_table */ + goto out; + } + + nes_manage_arp_cache(nesvnic->netdev, + nesadapter->arp_table[arpindex].mac_addr, + dst_ip, NES_ARP_DELETE); + } + + nes_manage_arp_cache(nesvnic->netdev, neigh->ha, + dst_ip, NES_ARP_ADD); + rc = nes_arp_table(nesvnic->nesdev, dst_ip, NULL, + NES_ARP_RESOLVE); + } else { + neigh_event_send(neigh, NULL); + } } +out: + rcu_read_unlock(); + + if (neigh) + neigh_release(neigh); - neigh_event_send(rt->u.dst.neighbour, NULL); ip_rt_put(rt); + return rc; } - /** * make_cm_node - create a new instance of a cm node */ static struct nes_cm_node *make_cm_node(struct nes_cm_core *cm_core, - struct nes_vnic *nesvnic, struct nes_cm_info *cm_info, - struct nes_cm_listener *listener) + struct nes_vnic *nesvnic, struct nes_cm_info *cm_info, + struct nes_cm_listener *listener) { struct nes_cm_node *cm_node; struct timespec ts; + int oldarpindex = 0; int arpindex = 0; struct nes_device *nesdev; struct nes_adapter *nesadapter; @@ -1046,21 +1570,30 @@ static struct nes_cm_node *make_cm_node(struct nes_cm_core *cm_core, cm_node->rem_addr = cm_info->rem_addr; cm_node->loc_port = cm_info->loc_port; cm_node->rem_port = cm_info->rem_port; - cm_node->send_write0 = send_first; - nes_debug(NES_DBG_CM, "Make node addresses : loc = %x:%x, rem = %x:%x\n", - cm_node->loc_addr, cm_node->loc_port, cm_node->rem_addr, cm_node->rem_port); + + cm_node->mapped_loc_addr = cm_info->mapped_loc_addr; + cm_node->mapped_rem_addr = cm_info->mapped_rem_addr; + cm_node->mapped_loc_port = cm_info->mapped_loc_port; + cm_node->mapped_rem_port = cm_info->mapped_rem_port; + + cm_node->mpa_frame_rev = mpa_version; + cm_node->send_rdma0_op = SEND_RDMA_READ_ZERO; + cm_node->mpav2_ird_ord = 0; + cm_node->ird_size = 0; + cm_node->ord_size = 0; + + nes_debug(NES_DBG_CM, "Make node addresses : loc = %pI4:%x, rem = %pI4:%x\n", + &cm_node->loc_addr, cm_node->loc_port, + &cm_node->rem_addr, cm_node->rem_port); cm_node->listener = listener; cm_node->netdev = nesvnic->netdev; cm_node->cm_id = cm_info->cm_id; memcpy(cm_node->loc_mac, nesvnic->netdev->dev_addr, ETH_ALEN); - nes_debug(NES_DBG_CM, "listener=%p, cm_id=%p\n", - cm_node->listener, cm_node->cm_id); + nes_debug(NES_DBG_CM, "listener=%p, cm_id=%p\n", cm_node->listener, + cm_node->cm_id); - INIT_LIST_HEAD(&cm_node->retrans_list); spin_lock_init(&cm_node->retrans_list_lock); - INIT_LIST_HEAD(&cm_node->recv_list); - spin_lock_init(&cm_node->recv_list_lock); cm_node->loopbackpartner = NULL; atomic_set(&cm_node->ref_count, 1); @@ -1069,16 +1602,14 @@ static struct nes_cm_node *make_cm_node(struct nes_cm_core *cm_core, cm_node->tcp_cntxt.loc_id = NES_CM_DEF_LOCAL_ID; cm_node->tcp_cntxt.rcv_wscale = NES_CM_DEFAULT_RCV_WND_SCALE; cm_node->tcp_cntxt.rcv_wnd = NES_CM_DEFAULT_RCV_WND_SCALED >> - NES_CM_DEFAULT_RCV_WND_SCALE; + NES_CM_DEFAULT_RCV_WND_SCALE; ts = current_kernel_time(); cm_node->tcp_cntxt.loc_seq_num = htonl(ts.tv_nsec); cm_node->tcp_cntxt.mss = nesvnic->max_frame_size - sizeof(struct iphdr) - - sizeof(struct tcphdr) - ETH_HLEN - VLAN_HLEN; + sizeof(struct tcphdr) - ETH_HLEN - VLAN_HLEN; cm_node->tcp_cntxt.rcv_nxt = 0; /* get a unique session ID , add thread_id to an upcounter to handle race */ atomic_inc(&cm_core->node_cnt); - atomic_inc(&cm_core->session_id); - cm_node->session_id = (u32)(atomic_read(&cm_core->session_id) + current->tgid); cm_node->conn_type = cm_info->conn_type; cm_node->apbvt_set = 0; cm_node->accept_pend = 0; @@ -1089,21 +1620,21 @@ static struct nes_cm_node *make_cm_node(struct nes_cm_core *cm_core, nesadapter = nesdev->nesadapter; cm_node->loopbackpartner = NULL; + /* get the mac addr for the remote node */ - arpindex = nes_arp_table(nesdev, cm_node->rem_addr, NULL, NES_ARP_RESOLVE); + oldarpindex = nes_arp_table(nesdev, cm_node->mapped_rem_addr, + NULL, NES_ARP_RESOLVE); + arpindex = nes_addr_resolve_neigh(nesvnic, + cm_node->mapped_rem_addr, oldarpindex); if (arpindex < 0) { kfree(cm_node); - nes_addr_send_arp(cm_info->rem_addr); return NULL; } /* copy the mac addr to node context */ memcpy(cm_node->rem_mac, nesadapter->arp_table[arpindex].mac_addr, ETH_ALEN); - nes_debug(NES_DBG_CM, "Remote mac addr from arp table:%02x," - " %02x, %02x, %02x, %02x, %02x\n", - cm_node->rem_mac[0], cm_node->rem_mac[1], - cm_node->rem_mac[2], cm_node->rem_mac[3], - cm_node->rem_mac[4], cm_node->rem_mac[5]); + nes_debug(NES_DBG_CM, "Remote mac addr from arp table: %pM\n", + cm_node->rem_mac); add_hte_node(cm_core, cm_node); atomic_inc(&cm_nodes_created); @@ -1126,13 +1657,9 @@ static int add_ref_cm_node(struct nes_cm_node *cm_node) * rem_ref_cm_node - destroy an instance of a cm node */ static int rem_ref_cm_node(struct nes_cm_core *cm_core, - struct nes_cm_node *cm_node) + struct nes_cm_node *cm_node) { - unsigned long flags, qplockflags; - struct nes_timer_entry *send_entry; - struct nes_timer_entry *recv_entry; - struct iw_cm_id *cm_id; - struct list_head *list_core, *list_node_temp; + unsigned long flags; struct nes_qp *nesqp; if (!cm_node) @@ -1153,75 +1680,41 @@ static int rem_ref_cm_node(struct nes_cm_core *cm_core, atomic_dec(&cm_node->listener->pend_accepts_cnt); BUG_ON(atomic_read(&cm_node->listener->pend_accepts_cnt) < 0); } - - spin_lock_irqsave(&cm_node->retrans_list_lock, flags); - list_for_each_safe(list_core, list_node_temp, &cm_node->retrans_list) { - send_entry = container_of(list_core, struct nes_timer_entry, list); - list_del(&send_entry->list); - spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); - dev_kfree_skb_any(send_entry->skb); - kfree(send_entry); - spin_lock_irqsave(&cm_node->retrans_list_lock, flags); - continue; - } - spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); - - spin_lock_irqsave(&cm_node->recv_list_lock, flags); - list_for_each_safe(list_core, list_node_temp, &cm_node->recv_list) { - recv_entry = container_of(list_core, struct nes_timer_entry, list); - list_del(&recv_entry->list); - cm_id = cm_node->cm_id; - spin_unlock_irqrestore(&cm_node->recv_list_lock, flags); - if (recv_entry->type == NES_TIMER_TYPE_CLOSE) { - nesqp = (struct nes_qp *)recv_entry->skb; - spin_lock_irqsave(&nesqp->lock, qplockflags); - if (nesqp->cm_id) { - nes_debug(NES_DBG_CM, "QP%u: cm_id = %p: ****** HIT A NES_TIMER_TYPE_CLOSE" - " with something to do!!! ******\n", - nesqp->hwqp.qp_id, cm_id); - nesqp->hw_tcp_state = NES_AEQE_TCP_STATE_CLOSED; - nesqp->last_aeq = NES_AEQE_AEID_RESET_SENT; - nesqp->ibqp_state = IB_QPS_ERR; - spin_unlock_irqrestore(&nesqp->lock, qplockflags); - nes_cm_disconn(nesqp); - } else { - spin_unlock_irqrestore(&nesqp->lock, qplockflags); - nes_debug(NES_DBG_CM, "QP%u: cm_id = %p: ****** HIT A NES_TIMER_TYPE_CLOSE" - " with nothing to do!!! ******\n", - nesqp->hwqp.qp_id, cm_id); - nes_rem_ref(&nesqp->ibqp); - } - cm_id->rem_ref(cm_id); - } else if (recv_entry->type == NES_TIMER_TYPE_RECV) { - dev_kfree_skb_any(recv_entry->skb); - } - kfree(recv_entry); - spin_lock_irqsave(&cm_node->recv_list_lock, flags); - } - spin_unlock_irqrestore(&cm_node->recv_list_lock, flags); - + WARN_ON(cm_node->send_entry); + if (cm_node->recv_entry) + handle_recv_entry(cm_node, 0); if (cm_node->listener) { mini_cm_dec_refcnt_listen(cm_core, cm_node->listener, 0); } else { if (cm_node->apbvt_set && cm_node->nesvnic) { - nes_manage_apbvt(cm_node->nesvnic, cm_node->loc_port, - PCI_FUNC(cm_node->nesvnic->nesdev->pcidev->devfn), - NES_MANAGE_APBVT_DEL); + nes_manage_apbvt(cm_node->nesvnic, cm_node->mapped_loc_port, + PCI_FUNC(cm_node->nesvnic->nesdev->pcidev->devfn), + NES_MANAGE_APBVT_DEL); } + nes_debug(NES_DBG_NLMSG, "Delete APBVT mapped_loc_port = %04X\n", + cm_node->mapped_loc_port); + nes_remove_mapinfo(cm_node->loc_addr, cm_node->loc_port, + cm_node->mapped_loc_addr, cm_node->mapped_loc_port); } - kfree(cm_node); atomic_dec(&cm_core->node_cnt); atomic_inc(&cm_nodes_destroyed); + nesqp = cm_node->nesqp; + if (nesqp) { + nesqp->cm_node = NULL; + nes_rem_ref(&nesqp->ibqp); + cm_node->nesqp = NULL; + } + kfree(cm_node); return 0; } - /** * process_options */ -static int process_options(struct nes_cm_node *cm_node, u8 *optionsloc, u32 optionsize, u32 syn_packet) +static int process_options(struct nes_cm_node *cm_node, u8 *optionsloc, + u32 optionsize, u32 syn_packet) { u32 tmp; u32 offset = 0; @@ -1231,35 +1724,34 @@ static int process_options(struct nes_cm_node *cm_node, u8 *optionsloc, u32 opti while (offset < optionsize) { all_options = (union all_known_options *)(optionsloc + offset); switch (all_options->as_base.optionnum) { - case OPTION_NUMBER_END: - offset = optionsize; - break; - case OPTION_NUMBER_NONE: - offset += 1; - continue; - case OPTION_NUMBER_MSS: - nes_debug(NES_DBG_CM, "%s: MSS Length: %d Offset: %d Size: %d\n", - __FUNCTION__, - all_options->as_mss.length, offset, optionsize); - got_mss_option = 1; - if (all_options->as_mss.length != 4) { - return 1; - } else { - tmp = ntohs(all_options->as_mss.mss); - if (tmp > 0 && tmp < cm_node->tcp_cntxt.mss) - cm_node->tcp_cntxt.mss = tmp; - } - break; - case OPTION_NUMBER_WINDOW_SCALE: - cm_node->tcp_cntxt.snd_wscale = all_options->as_windowscale.shiftcount; - break; - case OPTION_NUMBER_WRITE0: - cm_node->send_write0 = 1; - break; - default: - nes_debug(NES_DBG_CM, "TCP Option not understood: %x\n", - all_options->as_base.optionnum); - break; + case OPTION_NUMBER_END: + offset = optionsize; + break; + case OPTION_NUMBER_NONE: + offset += 1; + continue; + case OPTION_NUMBER_MSS: + nes_debug(NES_DBG_CM, "%s: MSS Length: %d Offset: %d " + "Size: %d\n", __func__, + all_options->as_mss.length, offset, optionsize); + got_mss_option = 1; + if (all_options->as_mss.length != 4) { + return 1; + } else { + tmp = ntohs(all_options->as_mss.mss); + if (tmp > 0 && tmp < + cm_node->tcp_cntxt.mss) + cm_node->tcp_cntxt.mss = tmp; + } + break; + case OPTION_NUMBER_WINDOW_SCALE: + cm_node->tcp_cntxt.snd_wscale = + all_options->as_windowscale.shiftcount; + break; + default: + nes_debug(NES_DBG_CM, "TCP Option not understood: %x\n", + all_options->as_base.optionnum); + break; } offset += all_options->as_base.length; } @@ -1268,310 +1760,623 @@ static int process_options(struct nes_cm_node *cm_node, u8 *optionsloc, u32 opti return 0; } +static void drop_packet(struct sk_buff *skb) +{ + atomic_inc(&cm_accel_dropped_pkts); + dev_kfree_skb_any(skb); +} -/** - * process_packet - */ -int process_packet(struct nes_cm_node *cm_node, struct sk_buff *skb, - struct nes_cm_core *cm_core) +static void handle_fin_pkt(struct nes_cm_node *cm_node) { - int optionsize; - int datasize; - int ret = 0; - struct tcphdr *tcph = tcp_hdr(skb); - u32 inc_sequence; - if (cm_node->state == NES_CM_STATE_SYN_SENT && tcph->syn) { - inc_sequence = ntohl(tcph->seq); - cm_node->tcp_cntxt.rcv_nxt = inc_sequence; + nes_debug(NES_DBG_CM, "Received FIN, cm_node = %p, state = %u. " + "refcnt=%d\n", cm_node, cm_node->state, + atomic_read(&cm_node->ref_count)); + switch (cm_node->state) { + case NES_CM_STATE_SYN_RCVD: + case NES_CM_STATE_SYN_SENT: + case NES_CM_STATE_ESTABLISHED: + case NES_CM_STATE_MPAREJ_RCVD: + cm_node->tcp_cntxt.rcv_nxt++; + cleanup_retrans_entry(cm_node); + cm_node->state = NES_CM_STATE_LAST_ACK; + send_fin(cm_node, NULL); + break; + case NES_CM_STATE_MPAREQ_SENT: + create_event(cm_node, NES_CM_EVENT_ABORTED); + cm_node->tcp_cntxt.rcv_nxt++; + cleanup_retrans_entry(cm_node); + cm_node->state = NES_CM_STATE_CLOSED; + add_ref_cm_node(cm_node); + send_reset(cm_node, NULL); + break; + case NES_CM_STATE_FIN_WAIT1: + cm_node->tcp_cntxt.rcv_nxt++; + cleanup_retrans_entry(cm_node); + cm_node->state = NES_CM_STATE_CLOSING; + send_ack(cm_node, NULL); + /* Wait for ACK as this is simultaneous close.. + * After we receive ACK, do not send anything.. + * Just rm the node.. Done.. */ + break; + case NES_CM_STATE_FIN_WAIT2: + cm_node->tcp_cntxt.rcv_nxt++; + cleanup_retrans_entry(cm_node); + cm_node->state = NES_CM_STATE_TIME_WAIT; + send_ack(cm_node, NULL); + schedule_nes_timer(cm_node, NULL, NES_TIMER_TYPE_CLOSE, 1, 0); + break; + case NES_CM_STATE_TIME_WAIT: + cm_node->tcp_cntxt.rcv_nxt++; + cleanup_retrans_entry(cm_node); + cm_node->state = NES_CM_STATE_CLOSED; + rem_ref_cm_node(cm_node->cm_core, cm_node); + break; + case NES_CM_STATE_TSA: + default: + nes_debug(NES_DBG_CM, "Error Rcvd FIN for node-%p state = %d\n", + cm_node, cm_node->state); + break; } +} - if ((!tcph) || (cm_node->state == NES_CM_STATE_TSA)) { - BUG_ON(!tcph); - atomic_inc(&cm_accel_dropped_pkts); - return -1; + +static void handle_rst_pkt(struct nes_cm_node *cm_node, struct sk_buff *skb, + struct tcphdr *tcph) +{ + + int reset = 0; /* whether to send reset in case of err.. */ + atomic_inc(&cm_resets_recvd); + nes_debug(NES_DBG_CM, "Received Reset, cm_node = %p, state = %u." + " refcnt=%d\n", cm_node, cm_node->state, + atomic_read(&cm_node->ref_count)); + cleanup_retrans_entry(cm_node); + switch (cm_node->state) { + case NES_CM_STATE_SYN_SENT: + case NES_CM_STATE_MPAREQ_SENT: + nes_debug(NES_DBG_CM, "%s[%u] create abort for cm_node=%p " + "listener=%p state=%d\n", __func__, __LINE__, cm_node, + cm_node->listener, cm_node->state); + switch (cm_node->mpa_frame_rev) { + case IETF_MPA_V2: + cm_node->mpa_frame_rev = IETF_MPA_V1; + /* send a syn and goto syn sent state */ + cm_node->state = NES_CM_STATE_SYN_SENT; + if (send_syn(cm_node, 0, NULL)) { + active_open_err(cm_node, skb, reset); + } + break; + case IETF_MPA_V1: + default: + active_open_err(cm_node, skb, reset); + break; + } + break; + case NES_CM_STATE_MPAREQ_RCVD: + atomic_inc(&cm_node->passive_state); + dev_kfree_skb_any(skb); + break; + case NES_CM_STATE_ESTABLISHED: + case NES_CM_STATE_SYN_RCVD: + case NES_CM_STATE_LISTENING: + nes_debug(NES_DBG_CM, "Bad state %s[%u]\n", __func__, __LINE__); + passive_open_err(cm_node, skb, reset); + break; + case NES_CM_STATE_TSA: + active_open_err(cm_node, skb, reset); + break; + case NES_CM_STATE_CLOSED: + drop_packet(skb); + break; + case NES_CM_STATE_FIN_WAIT2: + case NES_CM_STATE_FIN_WAIT1: + case NES_CM_STATE_LAST_ACK: + cm_node->cm_id->rem_ref(cm_node->cm_id); + case NES_CM_STATE_TIME_WAIT: + cm_node->state = NES_CM_STATE_CLOSED; + rem_ref_cm_node(cm_node->cm_core, cm_node); + drop_packet(skb); + break; + default: + drop_packet(skb); + break; } +} - if (tcph->rst) { - atomic_inc(&cm_resets_recvd); - nes_debug(NES_DBG_CM, "Received Reset, cm_node = %p, state = %u. refcnt=%d\n", - cm_node, cm_node->state, atomic_read(&cm_node->ref_count)); - switch (cm_node->state) { - case NES_CM_STATE_LISTENING: - rem_ref_cm_node(cm_core, cm_node); - break; - case NES_CM_STATE_TSA: - case NES_CM_STATE_CLOSED: - break; - case NES_CM_STATE_SYN_RCVD: - nes_debug(NES_DBG_CM, "Received a reset for local 0x%08X:%04X," - " remote 0x%08X:%04X, node state = %u\n", - cm_node->loc_addr, cm_node->loc_port, - cm_node->rem_addr, cm_node->rem_port, - cm_node->state); - rem_ref_cm_node(cm_core, cm_node); - break; - case NES_CM_STATE_ONE_SIDE_ESTABLISHED: - case NES_CM_STATE_ESTABLISHED: - case NES_CM_STATE_MPAREQ_SENT: - default: - nes_debug(NES_DBG_CM, "Received a reset for local 0x%08X:%04X," - " remote 0x%08X:%04X, node state = %u refcnt=%d\n", - cm_node->loc_addr, cm_node->loc_port, - cm_node->rem_addr, cm_node->rem_port, - cm_node->state, atomic_read(&cm_node->ref_count)); - // create event - cm_node->state = NES_CM_STATE_CLOSED; - create_event(cm_node, NES_CM_EVENT_ABORTED); - break; +static void handle_rcv_mpa(struct nes_cm_node *cm_node, struct sk_buff *skb) +{ + int ret = 0; + int datasize = skb->len; + u8 *dataloc = skb->data; + + enum nes_cm_event_type type = NES_CM_EVENT_UNKNOWN; + u32 res_type; + + ret = parse_mpa(cm_node, dataloc, &res_type, datasize); + if (ret) { + nes_debug(NES_DBG_CM, "didn't like MPA Request\n"); + if (cm_node->state == NES_CM_STATE_MPAREQ_SENT) { + nes_debug(NES_DBG_CM, "%s[%u] create abort for " + "cm_node=%p listener=%p state=%d\n", __func__, + __LINE__, cm_node, cm_node->listener, + cm_node->state); + active_open_err(cm_node, skb, 1); + } else { + passive_open_err(cm_node, skb, 1); + } + return; + } + switch (cm_node->state) { + case NES_CM_STATE_ESTABLISHED: + if (res_type == NES_MPA_REQUEST_REJECT) + /*BIG problem as we are receiving the MPA.. So should + * not be REJECT.. This is Passive Open.. We can + * only receive it Reject for Active Open...*/ + WARN_ON(1); + cm_node->state = NES_CM_STATE_MPAREQ_RCVD; + type = NES_CM_EVENT_MPA_REQ; + atomic_set(&cm_node->passive_state, + NES_PASSIVE_STATE_INDICATED); + break; + case NES_CM_STATE_MPAREQ_SENT: + cleanup_retrans_entry(cm_node); + if (res_type == NES_MPA_REQUEST_REJECT) { + type = NES_CM_EVENT_MPA_REJECT; + cm_node->state = NES_CM_STATE_MPAREJ_RCVD; + } else { + type = NES_CM_EVENT_CONNECTED; + cm_node->state = NES_CM_STATE_TSA; } - return -1; + + break; + default: + WARN_ON(1); + break; } + dev_kfree_skb_any(skb); + create_event(cm_node, type); +} - optionsize = (tcph->doff << 2) - sizeof(struct tcphdr); +static void indicate_pkt_err(struct nes_cm_node *cm_node, struct sk_buff *skb) +{ + switch (cm_node->state) { + case NES_CM_STATE_SYN_SENT: + case NES_CM_STATE_MPAREQ_SENT: + nes_debug(NES_DBG_CM, "%s[%u] create abort for cm_node=%p " + "listener=%p state=%d\n", __func__, __LINE__, cm_node, + cm_node->listener, cm_node->state); + active_open_err(cm_node, skb, 1); + break; + case NES_CM_STATE_ESTABLISHED: + case NES_CM_STATE_SYN_RCVD: + passive_open_err(cm_node, skb, 1); + break; + case NES_CM_STATE_TSA: + default: + drop_packet(skb); + } +} - skb_pull(skb, ip_hdr(skb)->ihl << 2); - skb_pull(skb, tcph->doff << 2); +static int check_syn(struct nes_cm_node *cm_node, struct tcphdr *tcph, + struct sk_buff *skb) +{ + int err; - datasize = skb->len; - inc_sequence = ntohl(tcph->seq); - nes_debug(NES_DBG_CM, "datasize = %u, sequence = 0x%08X, ack_seq = 0x%08X," - " rcv_nxt = 0x%08X Flags: %s %s.\n", - datasize, inc_sequence, ntohl(tcph->ack_seq), - cm_node->tcp_cntxt.rcv_nxt, (tcph->syn ? "SYN":""), - (tcph->ack ? "ACK":"")); - - if (!tcph->syn && (inc_sequence != cm_node->tcp_cntxt.rcv_nxt) - ) { - nes_debug(NES_DBG_CM, "dropping packet, datasize = %u, sequence = 0x%08X," - " ack_seq = 0x%08X, rcv_nxt = 0x%08X Flags: %s.\n", - datasize, inc_sequence, ntohl(tcph->ack_seq), - cm_node->tcp_cntxt.rcv_nxt, (tcph->ack ? "ACK":"")); - if (cm_node->state == NES_CM_STATE_LISTENING) { - rem_ref_cm_node(cm_core, cm_node); - } - return -1; + err = ((ntohl(tcph->ack_seq) == cm_node->tcp_cntxt.loc_seq_num)) ? 0 : 1; + if (err) + active_open_err(cm_node, skb, 1); + + return err; +} + +static int check_seq(struct nes_cm_node *cm_node, struct tcphdr *tcph, + struct sk_buff *skb) +{ + int err = 0; + u32 seq; + u32 ack_seq; + u32 loc_seq_num = cm_node->tcp_cntxt.loc_seq_num; + u32 rcv_nxt = cm_node->tcp_cntxt.rcv_nxt; + u32 rcv_wnd; + + seq = ntohl(tcph->seq); + ack_seq = ntohl(tcph->ack_seq); + rcv_wnd = cm_node->tcp_cntxt.rcv_wnd; + if (ack_seq != loc_seq_num) + err = 1; + else if (!between(seq, rcv_nxt, (rcv_nxt + rcv_wnd))) + err = 1; + if (err) { + nes_debug(NES_DBG_CM, "%s[%u] create abort for cm_node=%p " + "listener=%p state=%d\n", __func__, __LINE__, cm_node, + cm_node->listener, cm_node->state); + indicate_pkt_err(cm_node, skb); + nes_debug(NES_DBG_CM, "seq ERROR cm_node =%p seq=0x%08X " + "rcv_nxt=0x%08X rcv_wnd=0x%x\n", cm_node, seq, rcv_nxt, + rcv_wnd); } + return err; +} - cm_node->tcp_cntxt.rcv_nxt = inc_sequence + datasize; +/* + * handle_syn_pkt() is for Passive node. The syn packet is received when a node + * is created with a listener or it may comein as rexmitted packet which in + * that case will be just dropped. + */ +static void handle_syn_pkt(struct nes_cm_node *cm_node, struct sk_buff *skb, + struct tcphdr *tcph) +{ + int ret; + u32 inc_sequence; + int optionsize; + optionsize = (tcph->doff << 2) - sizeof(struct tcphdr); + skb_trim(skb, 0); + inc_sequence = ntohl(tcph->seq); - if (optionsize) { - u8 *optionsloc = (u8 *)&tcph[1]; - if (process_options(cm_node, optionsloc, optionsize, (u32)tcph->syn)) { - nes_debug(NES_DBG_CM, "%s: Node %p, Sending RESET\n", __FUNCTION__, cm_node); - send_reset(cm_node); - if (cm_node->state != NES_CM_STATE_SYN_SENT) - rem_ref_cm_node(cm_core, cm_node); - return 0; + switch (cm_node->state) { + case NES_CM_STATE_SYN_SENT: + case NES_CM_STATE_MPAREQ_SENT: + /* Rcvd syn on active open connection*/ + active_open_err(cm_node, skb, 1); + break; + case NES_CM_STATE_LISTENING: + /* Passive OPEN */ + if (atomic_read(&cm_node->listener->pend_accepts_cnt) > + cm_node->listener->backlog) { + nes_debug(NES_DBG_CM, "drop syn due to backlog " + "pressure \n"); + cm_backlog_drops++; + passive_open_err(cm_node, skb, 0); + break; } - } else if (tcph->syn) - cm_node->tcp_cntxt.mss = NES_CM_DEFAULT_MSS; + ret = handle_tcp_options(cm_node, tcph, skb, optionsize, + 1); + if (ret) { + passive_open_err(cm_node, skb, 0); + /* drop pkt */ + break; + } + cm_node->tcp_cntxt.rcv_nxt = inc_sequence + 1; + BUG_ON(cm_node->send_entry); + cm_node->accept_pend = 1; + atomic_inc(&cm_node->listener->pend_accepts_cnt); + + cm_node->state = NES_CM_STATE_SYN_RCVD; + send_syn(cm_node, 1, skb); + break; + case NES_CM_STATE_CLOSED: + cleanup_retrans_entry(cm_node); + add_ref_cm_node(cm_node); + send_reset(cm_node, skb); + break; + case NES_CM_STATE_TSA: + case NES_CM_STATE_ESTABLISHED: + case NES_CM_STATE_FIN_WAIT1: + case NES_CM_STATE_FIN_WAIT2: + case NES_CM_STATE_MPAREQ_RCVD: + case NES_CM_STATE_LAST_ACK: + case NES_CM_STATE_CLOSING: + case NES_CM_STATE_UNKNOWN: + default: + drop_packet(skb); + break; + } +} - cm_node->tcp_cntxt.snd_wnd = ntohs(tcph->window) << - cm_node->tcp_cntxt.snd_wscale; +static void handle_synack_pkt(struct nes_cm_node *cm_node, struct sk_buff *skb, + struct tcphdr *tcph) +{ + int ret; + u32 inc_sequence; + int optionsize; - if (cm_node->tcp_cntxt.snd_wnd > cm_node->tcp_cntxt.max_snd_wnd) { - cm_node->tcp_cntxt.max_snd_wnd = cm_node->tcp_cntxt.snd_wnd; + optionsize = (tcph->doff << 2) - sizeof(struct tcphdr); + skb_trim(skb, 0); + inc_sequence = ntohl(tcph->seq); + switch (cm_node->state) { + case NES_CM_STATE_SYN_SENT: + cleanup_retrans_entry(cm_node); + /* active open */ + if (check_syn(cm_node, tcph, skb)) + return; + cm_node->tcp_cntxt.rem_ack_num = ntohl(tcph->ack_seq); + /* setup options */ + ret = handle_tcp_options(cm_node, tcph, skb, optionsize, 0); + if (ret) { + nes_debug(NES_DBG_CM, "cm_node=%p tcp_options failed\n", + cm_node); + break; + } + cleanup_retrans_entry(cm_node); + cm_node->tcp_cntxt.rcv_nxt = inc_sequence + 1; + send_mpa_request(cm_node, skb); + cm_node->state = NES_CM_STATE_MPAREQ_SENT; + break; + case NES_CM_STATE_MPAREQ_RCVD: + /* passive open, so should not be here */ + passive_open_err(cm_node, skb, 1); + break; + case NES_CM_STATE_LISTENING: + cm_node->tcp_cntxt.loc_seq_num = ntohl(tcph->ack_seq); + cleanup_retrans_entry(cm_node); + cm_node->state = NES_CM_STATE_CLOSED; + send_reset(cm_node, skb); + break; + case NES_CM_STATE_CLOSED: + cm_node->tcp_cntxt.loc_seq_num = ntohl(tcph->ack_seq); + cleanup_retrans_entry(cm_node); + add_ref_cm_node(cm_node); + send_reset(cm_node, skb); + break; + case NES_CM_STATE_ESTABLISHED: + case NES_CM_STATE_FIN_WAIT1: + case NES_CM_STATE_FIN_WAIT2: + case NES_CM_STATE_LAST_ACK: + case NES_CM_STATE_TSA: + case NES_CM_STATE_CLOSING: + case NES_CM_STATE_UNKNOWN: + case NES_CM_STATE_MPAREQ_SENT: + default: + drop_packet(skb); + break; } +} + +static int handle_ack_pkt(struct nes_cm_node *cm_node, struct sk_buff *skb, + struct tcphdr *tcph) +{ + int datasize = 0; + u32 inc_sequence; + int ret = 0; + int optionsize; + + optionsize = (tcph->doff << 2) - sizeof(struct tcphdr); - if (tcph->ack) { + if (check_seq(cm_node, tcph, skb)) + return -EINVAL; + + skb_pull(skb, tcph->doff << 2); + inc_sequence = ntohl(tcph->seq); + datasize = skb->len; + switch (cm_node->state) { + case NES_CM_STATE_SYN_RCVD: + /* Passive OPEN */ + cleanup_retrans_entry(cm_node); + ret = handle_tcp_options(cm_node, tcph, skb, optionsize, 1); + if (ret) + break; cm_node->tcp_cntxt.rem_ack_num = ntohl(tcph->ack_seq); - switch (cm_node->state) { - case NES_CM_STATE_SYN_RCVD: - case NES_CM_STATE_SYN_SENT: - /* read and stash current sequence number */ - if (cm_node->tcp_cntxt.rem_ack_num != cm_node->tcp_cntxt.loc_seq_num) { - nes_debug(NES_DBG_CM, "ERROR - cm_node->tcp_cntxt.rem_ack_num !=" - " cm_node->tcp_cntxt.loc_seq_num\n"); - send_reset(cm_node); - return 0; - } - if (cm_node->state == NES_CM_STATE_SYN_SENT) - cm_node->state = NES_CM_STATE_ONE_SIDE_ESTABLISHED; - else { - cm_node->state = NES_CM_STATE_ESTABLISHED; - } - break; - case NES_CM_STATE_LAST_ACK: - cm_node->state = NES_CM_STATE_CLOSED; - break; - case NES_CM_STATE_FIN_WAIT1: - cm_node->state = NES_CM_STATE_FIN_WAIT2; - break; - case NES_CM_STATE_CLOSING: - cm_node->state = NES_CM_STATE_TIME_WAIT; - /* need to schedule this to happen in 2MSL timeouts */ - cm_node->state = NES_CM_STATE_CLOSED; - break; - case NES_CM_STATE_ONE_SIDE_ESTABLISHED: - case NES_CM_STATE_ESTABLISHED: - case NES_CM_STATE_MPAREQ_SENT: - case NES_CM_STATE_CLOSE_WAIT: - case NES_CM_STATE_TIME_WAIT: - case NES_CM_STATE_CLOSED: - break; - case NES_CM_STATE_LISTENING: - nes_debug(NES_DBG_CM, "Received an ACK on a listening port (SYN %d)\n", tcph->syn); - cm_node->tcp_cntxt.loc_seq_num = ntohl(tcph->ack_seq); - send_reset(cm_node); - /* send_reset bumps refcount, this should have been a new node */ - rem_ref_cm_node(cm_core, cm_node); - return -1; - break; - case NES_CM_STATE_TSA: - nes_debug(NES_DBG_CM, "Received a packet with the ack bit set while in TSA state\n"); - break; - case NES_CM_STATE_UNKNOWN: - case NES_CM_STATE_INITED: - case NES_CM_STATE_ACCEPTING: - case NES_CM_STATE_FIN_WAIT2: - default: - nes_debug(NES_DBG_CM, "Received ack from unknown state: %x\n", - cm_node->state); - send_reset(cm_node); - break; + cm_node->state = NES_CM_STATE_ESTABLISHED; + if (datasize) { + cm_node->tcp_cntxt.rcv_nxt = inc_sequence + datasize; + handle_rcv_mpa(cm_node, skb); + } else { /* rcvd ACK only */ + dev_kfree_skb_any(skb); + } + break; + case NES_CM_STATE_ESTABLISHED: + /* Passive OPEN */ + cleanup_retrans_entry(cm_node); + if (datasize) { + cm_node->tcp_cntxt.rcv_nxt = inc_sequence + datasize; + handle_rcv_mpa(cm_node, skb); + } else { + drop_packet(skb); + } + break; + case NES_CM_STATE_MPAREQ_SENT: + cm_node->tcp_cntxt.rem_ack_num = ntohl(tcph->ack_seq); + if (datasize) { + cm_node->tcp_cntxt.rcv_nxt = inc_sequence + datasize; + handle_rcv_mpa(cm_node, skb); + } else { /* Could be just an ack pkt.. */ + dev_kfree_skb_any(skb); } + break; + case NES_CM_STATE_LISTENING: + cleanup_retrans_entry(cm_node); + cm_node->state = NES_CM_STATE_CLOSED; + send_reset(cm_node, skb); + break; + case NES_CM_STATE_CLOSED: + cleanup_retrans_entry(cm_node); + add_ref_cm_node(cm_node); + send_reset(cm_node, skb); + break; + case NES_CM_STATE_LAST_ACK: + case NES_CM_STATE_CLOSING: + cleanup_retrans_entry(cm_node); + cm_node->state = NES_CM_STATE_CLOSED; + cm_node->cm_id->rem_ref(cm_node->cm_id); + rem_ref_cm_node(cm_node->cm_core, cm_node); + drop_packet(skb); + break; + case NES_CM_STATE_FIN_WAIT1: + cleanup_retrans_entry(cm_node); + drop_packet(skb); + cm_node->state = NES_CM_STATE_FIN_WAIT2; + break; + case NES_CM_STATE_SYN_SENT: + case NES_CM_STATE_FIN_WAIT2: + case NES_CM_STATE_TSA: + case NES_CM_STATE_MPAREQ_RCVD: + case NES_CM_STATE_UNKNOWN: + default: + cleanup_retrans_entry(cm_node); + drop_packet(skb); + break; } + return ret; +} - if (tcph->syn) { - if (cm_node->state == NES_CM_STATE_LISTENING) { - /* do not exceed backlog */ - atomic_inc(&cm_node->listener->pend_accepts_cnt); - if (atomic_read(&cm_node->listener->pend_accepts_cnt) > - cm_node->listener->backlog) { - nes_debug(NES_DBG_CM, "drop syn due to backlog pressure \n"); - cm_backlog_drops++; - atomic_dec(&cm_node->listener->pend_accepts_cnt); - rem_ref_cm_node(cm_core, cm_node); - return 0; - } - cm_node->accept_pend = 1; - } - if (datasize == 0) - cm_node->tcp_cntxt.rcv_nxt ++; - if (cm_node->state == NES_CM_STATE_LISTENING) { - cm_node->state = NES_CM_STATE_SYN_RCVD; - send_syn(cm_node, 1); - } - if (cm_node->state == NES_CM_STATE_ONE_SIDE_ESTABLISHED) { - cm_node->state = NES_CM_STATE_ESTABLISHED; - /* send final handshake ACK */ - ret = send_ack(cm_node); - if (ret < 0) - return ret; +static int handle_tcp_options(struct nes_cm_node *cm_node, struct tcphdr *tcph, + struct sk_buff *skb, int optionsize, int passive) +{ + u8 *optionsloc = (u8 *)&tcph[1]; - cm_node->state = NES_CM_STATE_MPAREQ_SENT; - ret = send_mpa_request(cm_node); - if (ret < 0) - return ret; + if (optionsize) { + if (process_options(cm_node, optionsloc, optionsize, + (u32)tcph->syn)) { + nes_debug(NES_DBG_CM, "%s: Node %p, Sending RESET\n", + __func__, cm_node); + if (passive) + passive_open_err(cm_node, skb, 1); + else + active_open_err(cm_node, skb, 1); + return 1; } } - if (tcph->fin) { - cm_node->tcp_cntxt.rcv_nxt++; - switch (cm_node->state) { - case NES_CM_STATE_SYN_RCVD: - case NES_CM_STATE_SYN_SENT: - case NES_CM_STATE_ONE_SIDE_ESTABLISHED: - case NES_CM_STATE_ESTABLISHED: - case NES_CM_STATE_ACCEPTING: - case NES_CM_STATE_MPAREQ_SENT: - cm_node->state = NES_CM_STATE_CLOSE_WAIT; - cm_node->state = NES_CM_STATE_LAST_ACK; - ret = send_fin(cm_node, NULL); - break; - case NES_CM_STATE_FIN_WAIT1: - cm_node->state = NES_CM_STATE_CLOSING; - ret = send_ack(cm_node); - break; - case NES_CM_STATE_FIN_WAIT2: - cm_node->state = NES_CM_STATE_TIME_WAIT; - cm_node->tcp_cntxt.loc_seq_num ++; - ret = send_ack(cm_node); - /* need to schedule this to happen in 2MSL timeouts */ - cm_node->state = NES_CM_STATE_CLOSED; - break; - case NES_CM_STATE_CLOSE_WAIT: - case NES_CM_STATE_LAST_ACK: - case NES_CM_STATE_CLOSING: - case NES_CM_STATE_TSA: - default: - nes_debug(NES_DBG_CM, "Received a fin while in %x state\n", - cm_node->state); - ret = -EINVAL; - break; - } + cm_node->tcp_cntxt.snd_wnd = ntohs(tcph->window) << + cm_node->tcp_cntxt.snd_wscale; + + if (cm_node->tcp_cntxt.snd_wnd > cm_node->tcp_cntxt.max_snd_wnd) + cm_node->tcp_cntxt.max_snd_wnd = cm_node->tcp_cntxt.snd_wnd; + return 0; +} + +/* + * active_open_err() will send reset() if flag set.. + * It will also send ABORT event. + */ +static void active_open_err(struct nes_cm_node *cm_node, struct sk_buff *skb, + int reset) +{ + cleanup_retrans_entry(cm_node); + if (reset) { + nes_debug(NES_DBG_CM, "ERROR active err called for cm_node=%p, " + "state=%d\n", cm_node, cm_node->state); + add_ref_cm_node(cm_node); + send_reset(cm_node, skb); + } else { + dev_kfree_skb_any(skb); } - if (datasize) { - u8 *dataloc = skb->data; - /* figure out what state we are in and handle transition to next state */ - switch (cm_node->state) { - case NES_CM_STATE_LISTENING: - case NES_CM_STATE_SYN_RCVD: - case NES_CM_STATE_SYN_SENT: - case NES_CM_STATE_FIN_WAIT1: - case NES_CM_STATE_FIN_WAIT2: - case NES_CM_STATE_CLOSE_WAIT: - case NES_CM_STATE_LAST_ACK: - case NES_CM_STATE_CLOSING: - break; - case NES_CM_STATE_MPAREQ_SENT: - /* recv the mpa res frame, ret=frame len (incl priv data) */ - ret = parse_mpa(cm_node, dataloc, datasize); - if (ret < 0) - break; - /* set the req frame payload len in skb */ - /* we are done handling this state, set node to a TSA state */ - cm_node->state = NES_CM_STATE_TSA; - send_ack(cm_node); - create_event(cm_node, NES_CM_EVENT_CONNECTED); - break; + cm_node->state = NES_CM_STATE_CLOSED; + create_event(cm_node, NES_CM_EVENT_ABORTED); +} - case NES_CM_STATE_ESTABLISHED: - /* we are expecting an MPA req frame */ - ret = parse_mpa(cm_node, dataloc, datasize); - if (ret < 0) { - break; - } - cm_node->state = NES_CM_STATE_TSA; - send_ack(cm_node); - /* we got a valid MPA request, create an event */ - create_event(cm_node, NES_CM_EVENT_MPA_REQ); - break; - case NES_CM_STATE_TSA: - handle_exception_pkt(cm_node, skb); - break; - case NES_CM_STATE_UNKNOWN: - case NES_CM_STATE_INITED: - default: - ret = -1; - } +/* + * passive_open_err() will either do a reset() or will free up the skb and + * remove the cm_node. + */ +static void passive_open_err(struct nes_cm_node *cm_node, struct sk_buff *skb, + int reset) +{ + cleanup_retrans_entry(cm_node); + cm_node->state = NES_CM_STATE_CLOSED; + if (reset) { + nes_debug(NES_DBG_CM, "passive_open_err sending RST for " + "cm_node=%p state =%d\n", cm_node, cm_node->state); + send_reset(cm_node, skb); + } else { + dev_kfree_skb_any(skb); + rem_ref_cm_node(cm_node->cm_core, cm_node); } +} - return ret; +/* + * free_retrans_entry() routines assumes that the retrans_list_lock has + * been acquired before calling. + */ +static void free_retrans_entry(struct nes_cm_node *cm_node) +{ + struct nes_timer_entry *send_entry; + + send_entry = cm_node->send_entry; + if (send_entry) { + cm_node->send_entry = NULL; + dev_kfree_skb_any(send_entry->skb); + kfree(send_entry); + rem_ref_cm_node(cm_node->cm_core, cm_node); + } } +static void cleanup_retrans_entry(struct nes_cm_node *cm_node) +{ + unsigned long flags; + + spin_lock_irqsave(&cm_node->retrans_list_lock, flags); + free_retrans_entry(cm_node); + spin_unlock_irqrestore(&cm_node->retrans_list_lock, flags); +} + +/** + * process_packet + * Returns skb if to be freed, else it will return NULL if already used.. + */ +static void process_packet(struct nes_cm_node *cm_node, struct sk_buff *skb, + struct nes_cm_core *cm_core) +{ + enum nes_tcpip_pkt_type pkt_type = NES_PKT_TYPE_UNKNOWN; + struct tcphdr *tcph = tcp_hdr(skb); + u32 fin_set = 0; + int ret = 0; + + skb_pull(skb, ip_hdr(skb)->ihl << 2); + + nes_debug(NES_DBG_CM, "process_packet: cm_node=%p state =%d syn=%d " + "ack=%d rst=%d fin=%d\n", cm_node, cm_node->state, tcph->syn, + tcph->ack, tcph->rst, tcph->fin); + + if (tcph->rst) { + pkt_type = NES_PKT_TYPE_RST; + } else if (tcph->syn) { + pkt_type = NES_PKT_TYPE_SYN; + if (tcph->ack) + pkt_type = NES_PKT_TYPE_SYNACK; + } else if (tcph->ack) { + pkt_type = NES_PKT_TYPE_ACK; + } + if (tcph->fin) + fin_set = 1; + + switch (pkt_type) { + case NES_PKT_TYPE_SYN: + handle_syn_pkt(cm_node, skb, tcph); + break; + case NES_PKT_TYPE_SYNACK: + handle_synack_pkt(cm_node, skb, tcph); + break; + case NES_PKT_TYPE_ACK: + ret = handle_ack_pkt(cm_node, skb, tcph); + if (fin_set && !ret) + handle_fin_pkt(cm_node); + break; + case NES_PKT_TYPE_RST: + handle_rst_pkt(cm_node, skb, tcph); + break; + default: + if ((fin_set) && (!check_seq(cm_node, tcph, skb))) + handle_fin_pkt(cm_node); + drop_packet(skb); + break; + } +} /** * mini_cm_listen - create a listen node with params */ static struct nes_cm_listener *mini_cm_listen(struct nes_cm_core *cm_core, - struct nes_vnic *nesvnic, struct nes_cm_info *cm_info) + struct nes_vnic *nesvnic, struct nes_cm_info *cm_info) { struct nes_cm_listener *listener; + struct iwpm_dev_data pm_reg_msg; + struct iwpm_sa_data pm_msg; unsigned long flags; + int iwpm_err = 0; nes_debug(NES_DBG_CM, "Search for 0x%08x : 0x%04x\n", - cm_info->loc_addr, cm_info->loc_port); + cm_info->loc_addr, cm_info->loc_port); /* cannot have multiple matching listeners */ - listener = find_listener(cm_core, htonl(cm_info->loc_addr), - htons(cm_info->loc_port), NES_CM_LISTENER_EITHER_STATE); + listener = find_listener(cm_core, cm_info->loc_addr, cm_info->loc_port, + NES_CM_LISTENER_EITHER_STATE, 1); + if (listener && listener->listener_state == NES_CM_LISTENER_ACTIVE_STATE) { /* find automatically incs ref count ??? */ atomic_dec(&listener->ref_count); @@ -1580,6 +2385,22 @@ static struct nes_cm_listener *mini_cm_listen(struct nes_cm_core *cm_core, } if (!listener) { + nes_form_reg_msg(nesvnic, &pm_reg_msg); + iwpm_err = iwpm_register_pid(&pm_reg_msg, RDMA_NL_NES); + if (iwpm_err) { + nes_debug(NES_DBG_NLMSG, + "Port Mapper reg pid fail (err = %d).\n", iwpm_err); + } + if (iwpm_valid_pid() && !iwpm_err) { + nes_form_pm_msg(cm_info, &pm_msg); + iwpm_err = iwpm_add_mapping(&pm_msg, RDMA_NL_NES); + if (iwpm_err) + nes_debug(NES_DBG_NLMSG, + "Port Mapper query fail (err = %d).\n", iwpm_err); + else + nes_record_pm_msg(cm_info, &pm_msg); + } + /* create a CM listen node (1/2 node to compare incoming traffic to) */ listener = kzalloc(sizeof(*listener), GFP_ATOMIC); if (!listener) { @@ -1587,9 +2408,10 @@ static struct nes_cm_listener *mini_cm_listen(struct nes_cm_core *cm_core, return NULL; } - memset(listener, 0, sizeof(struct nes_cm_listener)); - listener->loc_addr = htonl(cm_info->loc_addr); - listener->loc_port = htons(cm_info->loc_port); + listener->loc_addr = cm_info->loc_addr; + listener->loc_port = cm_info->loc_port; + listener->mapped_loc_addr = cm_info->mapped_loc_addr; + listener->mapped_loc_port = cm_info->mapped_loc_port; listener->reused_node = 0; atomic_set(&listener->ref_count, 1); @@ -1605,9 +2427,7 @@ static struct nes_cm_listener *mini_cm_listen(struct nes_cm_core *cm_core, listener->cm_core = cm_core; listener->nesvnic = nesvnic; atomic_inc(&cm_core->node_cnt); - atomic_inc(&cm_core->session_id); - listener->session_id = (u32)(atomic_read(&cm_core->session_id) + current->tgid); listener->conn_type = cm_info->conn_type; listener->backlog = cm_info->backlog; listener->listener_state = NES_CM_LISTENER_ACTIVE_STATE; @@ -1620,9 +2440,9 @@ static struct nes_cm_listener *mini_cm_listen(struct nes_cm_core *cm_core, } nes_debug(NES_DBG_CM, "Api - listen(): addr=0x%08X, port=0x%04x," - " listener = %p, backlog = %d, cm_id = %p.\n", - cm_info->loc_addr, cm_info->loc_port, - listener, listener->backlog, listener->cm_id); + " listener = %p, backlog = %d, cm_id = %p.\n", + cm_info->loc_addr, cm_info->loc_port, + listener, listener->backlog, listener->cm_id); return listener; } @@ -1631,83 +2451,106 @@ static struct nes_cm_listener *mini_cm_listen(struct nes_cm_core *cm_core, /** * mini_cm_connect - make a connection node with params */ -struct nes_cm_node *mini_cm_connect(struct nes_cm_core *cm_core, - struct nes_vnic *nesvnic, struct ietf_mpa_frame *mpa_frame, - struct nes_cm_info *cm_info) +static struct nes_cm_node *mini_cm_connect(struct nes_cm_core *cm_core, + struct nes_vnic *nesvnic, u16 private_data_len, + void *private_data, struct nes_cm_info *cm_info) { int ret = 0; struct nes_cm_node *cm_node; struct nes_cm_listener *loopbackremotelistener; struct nes_cm_node *loopbackremotenode; struct nes_cm_info loopback_cm_info; - - u16 mpa_frame_size = sizeof(struct ietf_mpa_frame) + - ntohs(mpa_frame->priv_data_len); - - cm_info->loc_addr = htonl(cm_info->loc_addr); - cm_info->rem_addr = htonl(cm_info->rem_addr); - cm_info->loc_port = htons(cm_info->loc_port); - cm_info->rem_port = htons(cm_info->rem_port); + u8 *start_buff; /* create a CM connection node */ cm_node = make_cm_node(cm_core, nesvnic, cm_info, NULL); if (!cm_node) return NULL; - // set our node side to client (active) side + /* set our node side to client (active) side */ cm_node->tcp_cntxt.client = 1; cm_node->tcp_cntxt.rcv_wscale = NES_CM_DEFAULT_RCV_WND_SCALE; if (cm_info->loc_addr == cm_info->rem_addr) { - loopbackremotelistener = find_listener(cm_core, cm_node->rem_addr, - cm_node->rem_port, NES_CM_LISTENER_ACTIVE_STATE); + loopbackremotelistener = find_listener(cm_core, + cm_node->mapped_loc_addr, cm_node->mapped_rem_port, + NES_CM_LISTENER_ACTIVE_STATE, 0); if (loopbackremotelistener == NULL) { create_event(cm_node, NES_CM_EVENT_ABORTED); } else { - atomic_inc(&cm_loopbacks); loopback_cm_info = *cm_info; loopback_cm_info.loc_port = cm_info->rem_port; loopback_cm_info.rem_port = cm_info->loc_port; + loopback_cm_info.mapped_loc_port = + cm_info->mapped_rem_port; + loopback_cm_info.mapped_rem_port = + cm_info->mapped_loc_port; loopback_cm_info.cm_id = loopbackremotelistener->cm_id; - loopbackremotenode = make_cm_node(cm_core, nesvnic, &loopback_cm_info, - loopbackremotelistener); + loopbackremotenode = make_cm_node(cm_core, nesvnic, + &loopback_cm_info, loopbackremotelistener); + if (!loopbackremotenode) { + rem_ref_cm_node(cm_node->cm_core, cm_node); + return NULL; + } + atomic_inc(&cm_loopbacks); loopbackremotenode->loopbackpartner = cm_node; - loopbackremotenode->tcp_cntxt.rcv_wscale = NES_CM_DEFAULT_RCV_WND_SCALE; + loopbackremotenode->tcp_cntxt.rcv_wscale = + NES_CM_DEFAULT_RCV_WND_SCALE; cm_node->loopbackpartner = loopbackremotenode; - memcpy(loopbackremotenode->mpa_frame_buf, &mpa_frame->priv_data, - mpa_frame_size); - loopbackremotenode->mpa_frame_size = mpa_frame_size - - sizeof(struct ietf_mpa_frame); + memcpy(loopbackremotenode->mpa_frame_buf, private_data, + private_data_len); + loopbackremotenode->mpa_frame_size = private_data_len; - // we are done handling this state, set node to a TSA state + /* we are done handling this state. */ + /* set node to a TSA state */ cm_node->state = NES_CM_STATE_TSA; - cm_node->tcp_cntxt.rcv_nxt = loopbackremotenode->tcp_cntxt.loc_seq_num; - loopbackremotenode->tcp_cntxt.rcv_nxt = cm_node->tcp_cntxt.loc_seq_num; - cm_node->tcp_cntxt.max_snd_wnd = loopbackremotenode->tcp_cntxt.rcv_wnd; - loopbackremotenode->tcp_cntxt.max_snd_wnd = cm_node->tcp_cntxt.rcv_wnd; - cm_node->tcp_cntxt.snd_wnd = loopbackremotenode->tcp_cntxt.rcv_wnd; - loopbackremotenode->tcp_cntxt.snd_wnd = cm_node->tcp_cntxt.rcv_wnd; - cm_node->tcp_cntxt.snd_wscale = loopbackremotenode->tcp_cntxt.rcv_wscale; - loopbackremotenode->tcp_cntxt.snd_wscale = cm_node->tcp_cntxt.rcv_wscale; - + cm_node->tcp_cntxt.rcv_nxt = + loopbackremotenode->tcp_cntxt.loc_seq_num; + loopbackremotenode->tcp_cntxt.rcv_nxt = + cm_node->tcp_cntxt.loc_seq_num; + cm_node->tcp_cntxt.max_snd_wnd = + loopbackremotenode->tcp_cntxt.rcv_wnd; + loopbackremotenode->tcp_cntxt.max_snd_wnd = + cm_node->tcp_cntxt.rcv_wnd; + cm_node->tcp_cntxt.snd_wnd = + loopbackremotenode->tcp_cntxt.rcv_wnd; + loopbackremotenode->tcp_cntxt.snd_wnd = + cm_node->tcp_cntxt.rcv_wnd; + cm_node->tcp_cntxt.snd_wscale = + loopbackremotenode->tcp_cntxt.rcv_wscale; + loopbackremotenode->tcp_cntxt.snd_wscale = + cm_node->tcp_cntxt.rcv_wscale; + loopbackremotenode->state = NES_CM_STATE_MPAREQ_RCVD; create_event(loopbackremotenode, NES_CM_EVENT_MPA_REQ); } return cm_node; } - /* set our node side to client (active) side */ - cm_node->tcp_cntxt.client = 1; - /* init our MPA frame ptr */ - memcpy(&cm_node->mpa_frame, mpa_frame, mpa_frame_size); - cm_node->mpa_frame_size = mpa_frame_size; + start_buff = &cm_node->mpa_frame_buf[0] + sizeof(struct ietf_mpa_v2); + cm_node->mpa_frame_size = private_data_len; + + memcpy(start_buff, private_data, private_data_len); /* send a syn and goto syn sent state */ cm_node->state = NES_CM_STATE_SYN_SENT; - ret = send_syn(cm_node, 0); + ret = send_syn(cm_node, 0, NULL); + + if (ret) { + /* error in sending the syn free up the cm_node struct */ + nes_debug(NES_DBG_CM, "Api - connect() FAILED: dest " + "addr=0x%08X, port=0x%04x, cm_node=%p, cm_id = %p.\n", + cm_node->rem_addr, cm_node->rem_port, cm_node, + cm_node->cm_id); + rem_ref_cm_node(cm_node->cm_core, cm_node); + cm_node = NULL; + } - nes_debug(NES_DBG_CM, "Api - connect(): dest addr=0x%08X, port=0x%04x," - " cm_node=%p, cm_id = %p.\n", - cm_node->rem_addr, cm_node->rem_port, cm_node, cm_node->cm_id); + if (cm_node) { + nes_debug(NES_DBG_CM, "Api - connect(): dest addr=0x%08X," + "port=0x%04x, cm_node=%p, cm_id = %p.\n", + cm_node->rem_addr, cm_node->rem_port, cm_node, + cm_node->cm_id); + } return cm_node; } @@ -1717,8 +2560,7 @@ struct nes_cm_node *mini_cm_connect(struct nes_cm_core *cm_core, * mini_cm_accept - accept a connection * This function is never called */ -int mini_cm_accept(struct nes_cm_core *cm_core, struct ietf_mpa_frame *mpa_frame, - struct nes_cm_node *cm_node) +static int mini_cm_accept(struct nes_cm_core *cm_core, struct nes_cm_node *cm_node) { return 0; } @@ -1727,31 +2569,62 @@ int mini_cm_accept(struct nes_cm_core *cm_core, struct ietf_mpa_frame *mpa_frame /** * mini_cm_reject - reject and teardown a connection */ -int mini_cm_reject(struct nes_cm_core *cm_core, - struct ietf_mpa_frame *mpa_frame, - struct nes_cm_node *cm_node) +static int mini_cm_reject(struct nes_cm_core *cm_core, struct nes_cm_node *cm_node) { int ret = 0; - struct sk_buff *skb; - u16 mpa_frame_size = sizeof(struct ietf_mpa_frame) + - ntohs(mpa_frame->priv_data_len); + int err = 0; + int passive_state; + struct nes_cm_event event; + struct iw_cm_id *cm_id = cm_node->cm_id; + struct nes_cm_node *loopback = cm_node->loopbackpartner; - skb = get_free_pkt(cm_node); - if (!skb) { - nes_debug(NES_DBG_CM, "Failed to get a Free pkt\n"); - return -1; - } + nes_debug(NES_DBG_CM, "%s cm_node=%p type=%d state=%d\n", + __func__, cm_node, cm_node->tcp_cntxt.client, cm_node->state); - /* send an MPA Request frame */ - form_cm_frame(skb, cm_node, NULL, 0, mpa_frame, mpa_frame_size, SET_ACK | SET_FIN); - ret = schedule_nes_timer(cm_node, skb, NES_TIMER_TYPE_SEND, 1, 0); + if (cm_node->tcp_cntxt.client) + return ret; + cleanup_retrans_entry(cm_node); - cm_node->state = NES_CM_STATE_CLOSED; - ret = send_fin(cm_node, NULL); + if (!loopback) { + passive_state = atomic_add_return(1, &cm_node->passive_state); + if (passive_state == NES_SEND_RESET_EVENT) { + cm_node->state = NES_CM_STATE_CLOSED; + rem_ref_cm_node(cm_core, cm_node); + } else { + if (cm_node->state == NES_CM_STATE_LISTENER_DESTROYED) { + rem_ref_cm_node(cm_core, cm_node); + } else { + ret = send_mpa_reject(cm_node); + if (ret) { + cm_node->state = NES_CM_STATE_CLOSED; + err = send_reset(cm_node, NULL); + if (err) + WARN_ON(1); + } else { + cm_id->add_ref(cm_id); + } + } + } + } else { + cm_node->cm_id = NULL; + if (cm_node->state == NES_CM_STATE_LISTENER_DESTROYED) { + rem_ref_cm_node(cm_core, cm_node); + rem_ref_cm_node(cm_core, loopback); + } else { + event.cm_node = loopback; + event.cm_info.rem_addr = loopback->rem_addr; + event.cm_info.loc_addr = loopback->loc_addr; + event.cm_info.rem_port = loopback->rem_port; + event.cm_info.loc_port = loopback->loc_port; + event.cm_info.cm_id = loopback->cm_id; + cm_event_mpa_reject(&event); + rem_ref_cm_node(cm_core, cm_node); + loopback->state = NES_CM_STATE_CLOSING; - if (ret < 0) { - printk(KERN_INFO PFX "failed to send MPA Reply (reject)\n"); - return ret; + cm_id = loopback->cm_id; + rem_ref_cm_node(cm_core, loopback); + cm_id->rem_ref(cm_id); + } } return ret; @@ -1761,7 +2634,7 @@ int mini_cm_reject(struct nes_cm_core *cm_core, /** * mini_cm_close */ -int mini_cm_close(struct nes_cm_core *cm_core, struct nes_cm_node *cm_node) +static int mini_cm_close(struct nes_cm_core *cm_core, struct nes_cm_node *cm_node) { int ret = 0; @@ -1769,37 +2642,45 @@ int mini_cm_close(struct nes_cm_core *cm_core, struct nes_cm_node *cm_node) return -EINVAL; switch (cm_node->state) { - /* if passed in node is null, create a reference key node for node search */ - /* check if we found an owner node for this pkt */ - case NES_CM_STATE_SYN_RCVD: - case NES_CM_STATE_SYN_SENT: - case NES_CM_STATE_ONE_SIDE_ESTABLISHED: - case NES_CM_STATE_ESTABLISHED: - case NES_CM_STATE_ACCEPTING: - case NES_CM_STATE_MPAREQ_SENT: - cm_node->state = NES_CM_STATE_FIN_WAIT1; - send_fin(cm_node, NULL); - break; - case NES_CM_STATE_CLOSE_WAIT: - cm_node->state = NES_CM_STATE_LAST_ACK; - send_fin(cm_node, NULL); - break; - case NES_CM_STATE_FIN_WAIT1: - case NES_CM_STATE_FIN_WAIT2: - case NES_CM_STATE_LAST_ACK: - case NES_CM_STATE_TIME_WAIT: - case NES_CM_STATE_CLOSING: - ret = -1; - break; - case NES_CM_STATE_LISTENING: - case NES_CM_STATE_UNKNOWN: - case NES_CM_STATE_INITED: - case NES_CM_STATE_CLOSED: - case NES_CM_STATE_TSA: - ret = rem_ref_cm_node(cm_core, cm_node); - break; + case NES_CM_STATE_SYN_RCVD: + case NES_CM_STATE_SYN_SENT: + case NES_CM_STATE_ONE_SIDE_ESTABLISHED: + case NES_CM_STATE_ESTABLISHED: + case NES_CM_STATE_ACCEPTING: + case NES_CM_STATE_MPAREQ_SENT: + case NES_CM_STATE_MPAREQ_RCVD: + cleanup_retrans_entry(cm_node); + send_reset(cm_node, NULL); + break; + case NES_CM_STATE_CLOSE_WAIT: + cm_node->state = NES_CM_STATE_LAST_ACK; + send_fin(cm_node, NULL); + break; + case NES_CM_STATE_FIN_WAIT1: + case NES_CM_STATE_FIN_WAIT2: + case NES_CM_STATE_LAST_ACK: + case NES_CM_STATE_TIME_WAIT: + case NES_CM_STATE_CLOSING: + ret = -1; + break; + case NES_CM_STATE_LISTENING: + cleanup_retrans_entry(cm_node); + send_reset(cm_node, NULL); + break; + case NES_CM_STATE_MPAREJ_RCVD: + case NES_CM_STATE_UNKNOWN: + case NES_CM_STATE_INITED: + case NES_CM_STATE_CLOSED: + case NES_CM_STATE_LISTENER_DESTROYED: + ret = rem_ref_cm_node(cm_core, cm_node); + break; + case NES_CM_STATE_TSA: + if (cm_node->send_entry) + printk(KERN_ERR "ERROR Close got called from STATE_TSA " + "send_entry=%p\n", cm_node->send_entry); + ret = rem_ref_cm_node(cm_core, cm_node); + break; } - cm_node->cm_id = NULL; return ret; } @@ -1808,102 +2689,111 @@ int mini_cm_close(struct nes_cm_core *cm_core, struct nes_cm_node *cm_node) * recv_pkt - recv an ETHERNET packet, and process it through CM * node state machine */ -int mini_cm_recv_pkt(struct nes_cm_core *cm_core, struct nes_vnic *nesvnic, - struct sk_buff *skb) +static int mini_cm_recv_pkt(struct nes_cm_core *cm_core, + struct nes_vnic *nesvnic, struct sk_buff *skb) { struct nes_cm_node *cm_node = NULL; struct nes_cm_listener *listener = NULL; struct iphdr *iph; struct tcphdr *tcph; struct nes_cm_info nfo; - int ret = 0; + int skb_handled = 1; + __be32 tmp_daddr, tmp_saddr; - if (!skb || skb->len < sizeof(struct iphdr) + sizeof(struct tcphdr)) { - ret = -EINVAL; - goto out; - } + if (!skb) + return 0; + if (skb->len < sizeof(struct iphdr) + sizeof(struct tcphdr)) + return 0; iph = (struct iphdr *)skb->data; tcph = (struct tcphdr *)(skb->data + sizeof(struct iphdr)); - skb_reset_network_header(skb); - skb_set_transport_header(skb, sizeof(*tcph)); - skb->len = ntohs(iph->tot_len); nfo.loc_addr = ntohl(iph->daddr); nfo.loc_port = ntohs(tcph->dest); nfo.rem_addr = ntohl(iph->saddr); nfo.rem_port = ntohs(tcph->source); - nes_debug(NES_DBG_CM, "Received packet: dest=0x%08X:0x%04X src=0x%08X:0x%04X\n", - iph->daddr, tcph->dest, iph->saddr, tcph->source); + /* If port mapper is available these should be mapped address info */ + nfo.mapped_loc_addr = ntohl(iph->daddr); + nfo.mapped_loc_port = ntohs(tcph->dest); + nfo.mapped_rem_addr = ntohl(iph->saddr); + nfo.mapped_rem_port = ntohs(tcph->source); - /* note: this call is going to increment cm_node ref count */ - cm_node = find_node(cm_core, - nfo.rem_port, nfo.rem_addr, - nfo.loc_port, nfo.loc_addr); + tmp_daddr = cpu_to_be32(iph->daddr); + tmp_saddr = cpu_to_be32(iph->saddr); - if (!cm_node) { - listener = find_listener(cm_core, nfo.loc_addr, nfo.loc_port, - NES_CM_LISTENER_ACTIVE_STATE); - if (listener) { - nfo.cm_id = listener->cm_id; - nfo.conn_type = listener->conn_type; - } else { - nfo.cm_id = NULL; - nfo.conn_type = 0; - } + nes_debug(NES_DBG_CM, "Received packet: dest=%pI4:0x%04X src=%pI4:0x%04X\n", + &tmp_daddr, tcph->dest, &tmp_saddr, tcph->source); + + do { + cm_node = find_node(cm_core, + nfo.mapped_rem_port, nfo.mapped_rem_addr, + nfo.mapped_loc_port, nfo.mapped_loc_addr); - cm_node = make_cm_node(cm_core, nesvnic, &nfo, listener); if (!cm_node) { - nes_debug(NES_DBG_CM, "Unable to allocate node\n"); - if (listener) { - nes_debug(NES_DBG_CM, "unable to allocate node and decrementing listener refcount\n"); + /* Only type of packet accepted are for */ + /* the PASSIVE open (syn only) */ + if ((!tcph->syn) || (tcph->ack)) { + skb_handled = 0; + break; + } + listener = find_listener(cm_core, nfo.mapped_loc_addr, + nfo.mapped_loc_port, + NES_CM_LISTENER_ACTIVE_STATE, 0); + if (!listener) { + nfo.cm_id = NULL; + nfo.conn_type = 0; + nes_debug(NES_DBG_CM, "Unable to find listener for the pkt\n"); + skb_handled = 0; + break; + } + nfo.cm_id = listener->cm_id; + nfo.conn_type = listener->conn_type; + cm_node = make_cm_node(cm_core, nesvnic, &nfo, + listener); + if (!cm_node) { + nes_debug(NES_DBG_CM, "Unable to allocate " + "node\n"); + cm_packets_dropped++; atomic_dec(&listener->ref_count); + dev_kfree_skb_any(skb); + break; } - ret = -1; - goto out; - } - if (!listener) { - nes_debug(NES_DBG_CM, "Packet found for unknown port %x refcnt=%d\n", - nfo.loc_port, atomic_read(&cm_node->ref_count)); - if (!tcph->rst) { - nes_debug(NES_DBG_CM, "Packet found for unknown port=%d" - " rem_port=%d refcnt=%d\n", - nfo.loc_port, nfo.rem_port, atomic_read(&cm_node->ref_count)); - - cm_node->tcp_cntxt.rcv_nxt = ntohl(tcph->seq); - cm_node->tcp_cntxt.loc_seq_num = ntohl(tcph->ack_seq); - send_reset(cm_node); + if (!tcph->rst && !tcph->fin) { + cm_node->state = NES_CM_STATE_LISTENING; + } else { + cm_packets_dropped++; + rem_ref_cm_node(cm_core, cm_node); + dev_kfree_skb_any(skb); + break; } - rem_ref_cm_node(cm_core, cm_node); - ret = -1; - goto out; + add_ref_cm_node(cm_node); + } else if (cm_node->state == NES_CM_STATE_TSA) { + if (cm_node->nesqp->pau_mode) + nes_queue_mgt_skbs(skb, nesvnic, cm_node->nesqp); + else { + rem_ref_cm_node(cm_core, cm_node); + atomic_inc(&cm_accel_dropped_pkts); + dev_kfree_skb_any(skb); + } + break; } - add_ref_cm_node(cm_node); - cm_node->state = NES_CM_STATE_LISTENING; - } - - nes_debug(NES_DBG_CM, "Processing Packet for node %p, data = (%p):\n", - cm_node, skb->data); - process_packet(cm_node, skb, cm_core); - - rem_ref_cm_node(cm_core, cm_node); - out: - if (skb) - dev_kfree_skb_any(skb); - return ret; + skb_reset_network_header(skb); + skb_set_transport_header(skb, sizeof(*tcph)); + skb->len = ntohs(iph->tot_len); + process_packet(cm_node, skb, cm_core); + rem_ref_cm_node(cm_core, cm_node); + } while (0); + return skb_handled; } /** * nes_cm_alloc_core - allocate a top level instance of a cm core */ -struct nes_cm_core *nes_cm_alloc_core(void) +static struct nes_cm_core *nes_cm_alloc_core(void) { - int i; - struct nes_cm_core *cm_core; - struct sk_buff *skb = NULL; /* setup the CM core */ /* alloc top level core control structure */ @@ -1915,26 +2805,12 @@ struct nes_cm_core *nes_cm_alloc_core(void) init_timer(&cm_core->tcp_timer); cm_core->tcp_timer.function = nes_cm_timer_tick; - cm_core->mtu = NES_CM_DEFAULT_MTU; + cm_core->mtu = NES_CM_DEFAULT_MTU; cm_core->state = NES_CM_STATE_INITED; cm_core->free_tx_pkt_max = NES_CM_DEFAULT_FREE_PKTS; - atomic_set(&cm_core->session_id, 0); atomic_set(&cm_core->events_posted, 0); - /* init the packet lists */ - skb_queue_head_init(&cm_core->tx_free_list); - - for (i = 0; i < NES_CM_DEFAULT_FRAME_CNT; i++) { - skb = dev_alloc_skb(cm_core->mtu); - if (!skb) { - kfree(cm_core); - return NULL; - } - /* add 'raw' skb to free frame list */ - skb_queue_head(&cm_core->tx_free_list, skb); - } - cm_core->api = &nes_cm_api; spin_lock_init(&cm_core->ht_lock); @@ -1958,7 +2834,7 @@ struct nes_cm_core *nes_cm_alloc_core(void) /** * mini_cm_dealloc_core - deallocate a top level instance of a cm core */ -int mini_cm_dealloc_core(struct nes_cm_core *cm_core) +static int mini_cm_dealloc_core(struct nes_cm_core *cm_core) { nes_debug(NES_DBG_CM, "De-Alloc CM Core (%p)\n", cm_core); @@ -1967,9 +2843,8 @@ int mini_cm_dealloc_core(struct nes_cm_core *cm_core) barrier(); - if (timer_pending(&cm_core->tcp_timer)) { + if (timer_pending(&cm_core->tcp_timer)) del_timer(&cm_core->tcp_timer); - } destroy_workqueue(cm_core->event_wq); destroy_workqueue(cm_core->disconn_wq); @@ -1983,7 +2858,7 @@ int mini_cm_dealloc_core(struct nes_cm_core *cm_core) /** * mini_cm_get */ -int mini_cm_get(struct nes_cm_core *cm_core) +static int mini_cm_get(struct nes_cm_core *cm_core) { return cm_core->state; } @@ -1992,20 +2867,20 @@ int mini_cm_get(struct nes_cm_core *cm_core) /** * mini_cm_set */ -int mini_cm_set(struct nes_cm_core *cm_core, u32 type, u32 value) +static int mini_cm_set(struct nes_cm_core *cm_core, u32 type, u32 value) { int ret = 0; switch (type) { - case NES_CM_SET_PKT_SIZE: - cm_core->mtu = value; - break; - case NES_CM_SET_FREE_PKT_Q_SIZE: - cm_core->free_tx_pkt_max = value; - break; - default: - /* unknown set option */ - ret = -EINVAL; + case NES_CM_SET_PKT_SIZE: + cm_core->mtu = value; + break; + case NES_CM_SET_FREE_PKT_Q_SIZE: + cm_core->free_tx_pkt_max = value; + break; + default: + /* unknown set option */ + ret = -EINVAL; } return ret; @@ -2024,8 +2899,8 @@ static int nes_cm_init_tsa_conn(struct nes_qp *nesqp, struct nes_cm_node *cm_nod return -EINVAL; nesqp->nesqp_context->misc |= cpu_to_le32(NES_QPCONTEXT_MISC_IPV4 | - NES_QPCONTEXT_MISC_NO_NAGLE | NES_QPCONTEXT_MISC_DO_NOT_FRAG | - NES_QPCONTEXT_MISC_DROS); + NES_QPCONTEXT_MISC_NO_NAGLE | NES_QPCONTEXT_MISC_DO_NOT_FRAG | + NES_QPCONTEXT_MISC_DROS); if (cm_node->tcp_cntxt.snd_wscale || cm_node->tcp_cntxt.rcv_wscale) nesqp->nesqp_context->misc |= cpu_to_le32(NES_QPCONTEXT_MISC_WSCALE); @@ -2035,15 +2910,15 @@ static int nes_cm_init_tsa_conn(struct nes_qp *nesqp, struct nes_cm_node *cm_nod nesqp->nesqp_context->mss |= cpu_to_le32(((u32)cm_node->tcp_cntxt.mss) << 16); nesqp->nesqp_context->tcp_state_flow_label |= cpu_to_le32( - (u32)NES_QPCONTEXT_TCPSTATE_EST << NES_QPCONTEXT_TCPFLOW_TCP_STATE_SHIFT); + (u32)NES_QPCONTEXT_TCPSTATE_EST << NES_QPCONTEXT_TCPFLOW_TCP_STATE_SHIFT); nesqp->nesqp_context->pd_index_wscale |= cpu_to_le32( - (cm_node->tcp_cntxt.snd_wscale << NES_QPCONTEXT_PDWSCALE_SND_WSCALE_SHIFT) & - NES_QPCONTEXT_PDWSCALE_SND_WSCALE_MASK); + (cm_node->tcp_cntxt.snd_wscale << NES_QPCONTEXT_PDWSCALE_SND_WSCALE_SHIFT) & + NES_QPCONTEXT_PDWSCALE_SND_WSCALE_MASK); nesqp->nesqp_context->pd_index_wscale |= cpu_to_le32( - (cm_node->tcp_cntxt.rcv_wscale << NES_QPCONTEXT_PDWSCALE_RCV_WSCALE_SHIFT) & - NES_QPCONTEXT_PDWSCALE_RCV_WSCALE_MASK); + (cm_node->tcp_cntxt.rcv_wscale << NES_QPCONTEXT_PDWSCALE_RCV_WSCALE_SHIFT) & + NES_QPCONTEXT_PDWSCALE_RCV_WSCALE_MASK); nesqp->nesqp_context->keepalive = cpu_to_le32(0x80); nesqp->nesqp_context->ts_recent = 0; @@ -2052,24 +2927,24 @@ static int nes_cm_init_tsa_conn(struct nes_qp *nesqp, struct nes_cm_node *cm_nod nesqp->nesqp_context->snd_wnd = cpu_to_le32(cm_node->tcp_cntxt.snd_wnd); nesqp->nesqp_context->rcv_nxt = cpu_to_le32(cm_node->tcp_cntxt.rcv_nxt); nesqp->nesqp_context->rcv_wnd = cpu_to_le32(cm_node->tcp_cntxt.rcv_wnd << - cm_node->tcp_cntxt.rcv_wscale); + cm_node->tcp_cntxt.rcv_wscale); nesqp->nesqp_context->snd_max = cpu_to_le32(cm_node->tcp_cntxt.loc_seq_num); nesqp->nesqp_context->snd_una = cpu_to_le32(cm_node->tcp_cntxt.loc_seq_num); nesqp->nesqp_context->srtt = 0; nesqp->nesqp_context->rttvar = cpu_to_le32(0x6); nesqp->nesqp_context->ssthresh = cpu_to_le32(0x3FFFC000); - nesqp->nesqp_context->cwnd = cpu_to_le32(2*cm_node->tcp_cntxt.mss); + nesqp->nesqp_context->cwnd = cpu_to_le32(2 * cm_node->tcp_cntxt.mss); nesqp->nesqp_context->snd_wl1 = cpu_to_le32(cm_node->tcp_cntxt.rcv_nxt); nesqp->nesqp_context->snd_wl2 = cpu_to_le32(cm_node->tcp_cntxt.loc_seq_num); nesqp->nesqp_context->max_snd_wnd = cpu_to_le32(cm_node->tcp_cntxt.max_snd_wnd); nes_debug(NES_DBG_CM, "QP%u: rcv_nxt = 0x%08X, snd_nxt = 0x%08X," - " Setting MSS to %u, PDWscale = 0x%08X, rcv_wnd = %u, context misc = 0x%08X.\n", - nesqp->hwqp.qp_id, le32_to_cpu(nesqp->nesqp_context->rcv_nxt), - le32_to_cpu(nesqp->nesqp_context->snd_nxt), - cm_node->tcp_cntxt.mss, le32_to_cpu(nesqp->nesqp_context->pd_index_wscale), - le32_to_cpu(nesqp->nesqp_context->rcv_wnd), - le32_to_cpu(nesqp->nesqp_context->misc)); + " Setting MSS to %u, PDWscale = 0x%08X, rcv_wnd = %u, context misc = 0x%08X.\n", + nesqp->hwqp.qp_id, le32_to_cpu(nesqp->nesqp_context->rcv_nxt), + le32_to_cpu(nesqp->nesqp_context->snd_nxt), + cm_node->tcp_cntxt.mss, le32_to_cpu(nesqp->nesqp_context->pd_index_wscale), + le32_to_cpu(nesqp->nesqp_context->rcv_wnd), + le32_to_cpu(nesqp->nesqp_context->misc)); nes_debug(NES_DBG_CM, " snd_wnd = 0x%08X.\n", le32_to_cpu(nesqp->nesqp_context->snd_wnd)); nes_debug(NES_DBG_CM, " snd_cwnd = 0x%08X.\n", le32_to_cpu(nesqp->nesqp_context->cwnd)); nes_debug(NES_DBG_CM, " max_swnd = 0x%08X.\n", le32_to_cpu(nesqp->nesqp_context->max_snd_wnd)); @@ -2086,22 +2961,16 @@ static int nes_cm_init_tsa_conn(struct nes_qp *nesqp, struct nes_cm_node *cm_nod */ int nes_cm_disconn(struct nes_qp *nesqp) { - unsigned long flags; + struct disconn_work *work; - spin_lock_irqsave(&nesqp->lock, flags); - if (nesqp->disconn_pending == 0) { - nesqp->disconn_pending++; - spin_unlock_irqrestore(&nesqp->lock, flags); - /* nes_add_ref(&nesqp->ibqp); */ - /* init our disconnect work element, to */ - INIT_WORK(&nesqp->disconn_work, nes_disconnect_worker); - - queue_work(g_cm_core->disconn_wq, &nesqp->disconn_work); - } else { - spin_unlock_irqrestore(&nesqp->lock, flags); - nes_rem_ref(&nesqp->ibqp); - } + work = kzalloc(sizeof *work, GFP_ATOMIC); + if (!work) + return -ENOMEM; /* Timer will clean up */ + nes_add_ref(&nesqp->ibqp); + work->nesqp = nesqp; + INIT_WORK(&work->work, nes_disconnect_worker); + queue_work(g_cm_core->disconn_wq, &work->work); return 0; } @@ -2109,20 +2978,23 @@ int nes_cm_disconn(struct nes_qp *nesqp) /** * nes_disconnect_worker */ -void nes_disconnect_worker(struct work_struct *work) +static void nes_disconnect_worker(struct work_struct *work) { - struct nes_qp *nesqp = container_of(work, struct nes_qp, disconn_work); + struct disconn_work *dwork = container_of(work, struct disconn_work, work); + struct nes_qp *nesqp = dwork->nesqp; + kfree(dwork); nes_debug(NES_DBG_CM, "processing AEQE id 0x%04X for QP%u.\n", - nesqp->last_aeq, nesqp->hwqp.qp_id); + nesqp->last_aeq, nesqp->hwqp.qp_id); nes_cm_disconn_true(nesqp); + nes_rem_ref(&nesqp->ibqp); } /** * nes_cm_disconn_true */ -int nes_cm_disconn_true(struct nes_qp *nesqp) +static int nes_cm_disconn_true(struct nes_qp *nesqp) { unsigned long flags; int ret = 0; @@ -2132,7 +3004,12 @@ int nes_cm_disconn_true(struct nes_qp *nesqp) u16 last_ae; u8 original_hw_tcp_state; u8 original_ibqp_state; - u8 issued_disconnect_reset = 0; + int disconn_status = 0; + int issue_disconn = 0; + int issue_close = 0; + int issue_flush = 0; + u32 flush_q = NES_CQP_FLUSH_RQ; + struct ib_event ibevent; if (!nesqp) { nes_debug(NES_DBG_CM, "disconnect_worker nesqp is NULL\n"); @@ -2144,9 +3021,8 @@ int nes_cm_disconn_true(struct nes_qp *nesqp) /* make sure we havent already closed this connection */ if (!cm_id) { nes_debug(NES_DBG_CM, "QP%u disconnect_worker cmid is NULL\n", - nesqp->hwqp.qp_id); + nesqp->hwqp.qp_id); spin_unlock_irqrestore(&nesqp->lock, flags); - nes_rem_ref(&nesqp->ibqp); return -1; } @@ -2154,67 +3030,86 @@ int nes_cm_disconn_true(struct nes_qp *nesqp) nes_debug(NES_DBG_CM, "Disconnecting QP%u\n", nesqp->hwqp.qp_id); original_hw_tcp_state = nesqp->hw_tcp_state; - original_ibqp_state = nesqp->ibqp_state; + original_ibqp_state = nesqp->ibqp_state; last_ae = nesqp->last_aeq; + if (nesqp->term_flags) { + issue_disconn = 1; + issue_close = 1; + nesqp->cm_id = NULL; + del_timer(&nesqp->terminate_timer); + if (nesqp->flush_issued == 0) { + nesqp->flush_issued = 1; + issue_flush = 1; + } + } else if ((original_hw_tcp_state == NES_AEQE_TCP_STATE_CLOSE_WAIT) || + ((original_ibqp_state == IB_QPS_RTS) && + (last_ae == NES_AEQE_AEID_LLP_CONNECTION_RESET))) { + issue_disconn = 1; + if (last_ae == NES_AEQE_AEID_LLP_CONNECTION_RESET) + disconn_status = -ECONNRESET; + } + + if (((original_hw_tcp_state == NES_AEQE_TCP_STATE_CLOSED) || + (original_hw_tcp_state == NES_AEQE_TCP_STATE_TIME_WAIT) || + (last_ae == NES_AEQE_AEID_RDMAP_ROE_BAD_LLP_CLOSE) || + (last_ae == NES_AEQE_AEID_LLP_CONNECTION_RESET))) { + issue_close = 1; + nesqp->cm_id = NULL; + if (nesqp->flush_issued == 0) { + nesqp->flush_issued = 1; + issue_flush = 1; + } + } - nes_debug(NES_DBG_CM, "set ibqp_state=%u\n", nesqp->ibqp_state); + spin_unlock_irqrestore(&nesqp->lock, flags); - if ((nesqp->cm_id) && (cm_id->event_handler)) { - if ((original_hw_tcp_state == NES_AEQE_TCP_STATE_CLOSE_WAIT) || - ((original_ibqp_state == IB_QPS_RTS) && - (last_ae == NES_AEQE_AEID_LLP_CONNECTION_RESET))) { + if ((issue_flush) && (nesqp->destroyed == 0)) { + /* Flush the queue(s) */ + if (nesqp->hw_iwarp_state >= NES_AEQE_IWARP_STATE_TERMINATE) + flush_q |= NES_CQP_FLUSH_SQ; + flush_wqes(nesvnic->nesdev, nesqp, flush_q, 1); + + if (nesqp->term_flags) { + ibevent.device = nesqp->ibqp.device; + ibevent.event = nesqp->terminate_eventtype; + ibevent.element.qp = &nesqp->ibqp; + if (nesqp->ibqp.event_handler) + nesqp->ibqp.event_handler(&ibevent, nesqp->ibqp.qp_context); + } + } + + if ((cm_id) && (cm_id->event_handler)) { + if (issue_disconn) { atomic_inc(&cm_disconnects); cm_event.event = IW_CM_EVENT_DISCONNECT; - if (last_ae == NES_AEQE_AEID_LLP_CONNECTION_RESET) { - issued_disconnect_reset = 1; - cm_event.status = IW_CM_EVENT_STATUS_RESET; - nes_debug(NES_DBG_CM, "Generating a CM Disconnect Event (status reset) for " - " QP%u, cm_id = %p. \n", - nesqp->hwqp.qp_id, cm_id); - } else { - cm_event.status = IW_CM_EVENT_STATUS_OK; - } - + cm_event.status = disconn_status; cm_event.local_addr = cm_id->local_addr; cm_event.remote_addr = cm_id->remote_addr; cm_event.private_data = NULL; cm_event.private_data_len = 0; - nes_debug(NES_DBG_CM, "Generating a CM Disconnect Event for " - " QP%u, SQ Head = %u, SQ Tail = %u. cm_id = %p, refcount = %u.\n", - nesqp->hwqp.qp_id, - nesqp->hwqp.sq_head, nesqp->hwqp.sq_tail, cm_id, - atomic_read(&nesqp->refcount)); + nes_debug(NES_DBG_CM, "Generating a CM Disconnect Event" + " for QP%u, SQ Head = %u, SQ Tail = %u. " + "cm_id = %p, refcount = %u.\n", + nesqp->hwqp.qp_id, nesqp->hwqp.sq_head, + nesqp->hwqp.sq_tail, cm_id, + atomic_read(&nesqp->refcount)); - spin_unlock_irqrestore(&nesqp->lock, flags); ret = cm_id->event_handler(cm_id, &cm_event); if (ret) - nes_debug(NES_DBG_CM, "OFA CM event_handler returned, ret=%d\n", ret); - spin_lock_irqsave(&nesqp->lock, flags); + nes_debug(NES_DBG_CM, "OFA CM event_handler " + "returned, ret=%d\n", ret); } - nesqp->disconn_pending = 0; - /* There might have been another AE while the lock was released */ - original_hw_tcp_state = nesqp->hw_tcp_state; - original_ibqp_state = nesqp->ibqp_state; - last_ae = nesqp->last_aeq; - - if ((issued_disconnect_reset == 0) && (nesqp->cm_id) && - ((original_hw_tcp_state == NES_AEQE_TCP_STATE_CLOSED) || - (original_hw_tcp_state == NES_AEQE_TCP_STATE_TIME_WAIT) || - (last_ae == NES_AEQE_AEID_RDMAP_ROE_BAD_LLP_CLOSE) || - (last_ae == NES_AEQE_AEID_LLP_CONNECTION_RESET))) { + if (issue_close) { atomic_inc(&cm_closes); - nesqp->cm_id = NULL; - nesqp->in_disconnect = 0; - spin_unlock_irqrestore(&nesqp->lock, flags); nes_disconnect(nesqp, 1); cm_id->provider_data = nesqp; /* Send up the close complete event */ cm_event.event = IW_CM_EVENT_CLOSE; - cm_event.status = IW_CM_EVENT_STATUS_OK; + cm_event.status = 0; cm_event.provider_data = cm_id->provider_data; cm_event.local_addr = cm_id->local_addr; cm_event.remote_addr = cm_id->remote_addr; @@ -2222,41 +3117,12 @@ int nes_cm_disconn_true(struct nes_qp *nesqp) cm_event.private_data_len = 0; ret = cm_id->event_handler(cm_id, &cm_event); - if (ret) { + if (ret) nes_debug(NES_DBG_CM, "OFA CM event_handler returned, ret=%d\n", ret); - } cm_id->rem_ref(cm_id); - - spin_lock_irqsave(&nesqp->lock, flags); - if (nesqp->flush_issued == 0) { - nesqp->flush_issued = 1; - spin_unlock_irqrestore(&nesqp->lock, flags); - flush_wqes(nesvnic->nesdev, nesqp, NES_CQP_FLUSH_RQ, 1); - } else { - spin_unlock_irqrestore(&nesqp->lock, flags); - } - - /* This reference is from either ModifyQP or the AE processing, - there is still a race here with modifyqp */ - nes_rem_ref(&nesqp->ibqp); - - } else { - cm_id = nesqp->cm_id; - spin_unlock_irqrestore(&nesqp->lock, flags); - /* check to see if the inbound reset beat the outbound reset */ - if ((!cm_id) && (last_ae==NES_AEQE_AEID_RESET_SENT)) { - nes_debug(NES_DBG_CM, "QP%u: Decing refcount due to inbound reset" - " beating the outbound reset.\n", - nesqp->hwqp.qp_id); - nes_rem_ref(&nesqp->ibqp); - } } - } else { - nesqp->disconn_pending = 0; - spin_unlock_irqrestore(&nesqp->lock, flags); } - nes_rem_ref(&nesqp->ibqp); return 0; } @@ -2265,20 +3131,22 @@ int nes_cm_disconn_true(struct nes_qp *nesqp) /** * nes_disconnect */ -int nes_disconnect(struct nes_qp *nesqp, int abrupt) +static int nes_disconnect(struct nes_qp *nesqp, int abrupt) { int ret = 0; struct nes_vnic *nesvnic; struct nes_device *nesdev; + struct nes_ib_device *nesibdev; nesvnic = to_nesvnic(nesqp->ibqp.device); if (!nesvnic) return -EINVAL; nesdev = nesvnic->nesdev; + nesibdev = nesvnic->nesibdev; nes_debug(NES_DBG_CM, "netdev refcnt = %u.\n", - atomic_read(&nesvnic->netdev->refcnt)); + netdev_refcnt_read(nesvnic->netdev)); if (nesqp->active_conn) { @@ -2287,9 +3155,11 @@ int nes_disconnect(struct nes_qp *nesqp, int abrupt) } else { /* Need to free the Last Streaming Mode Message */ if (nesqp->ietf_frame) { + if (nesqp->lsmm_mr) + nesibdev->ibdev.dereg_mr(nesqp->lsmm_mr); pci_free_consistent(nesdev->pcidev, - nesqp->private_data_len+sizeof(struct ietf_mpa_frame), - nesqp->ietf_frame, nesqp->ietf_frame_pbase); + nesqp->private_data_len + nesqp->ietf_frame_size, + nesqp->ietf_frame, nesqp->ietf_frame_pbase); } } @@ -2298,7 +3168,6 @@ int nes_disconnect(struct nes_qp *nesqp, int abrupt) nes_debug(NES_DBG_CM, "Call close API\n"); g_cm_core->api->close(g_cm_core, nesqp->cm_node); - nesqp->cm_node = NULL; } return ret; @@ -2323,6 +3192,20 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) struct nes_v4_quad nes_quad; u32 crc_value; int ret; + int passive_state; + struct nes_ib_device *nesibdev; + struct ib_mr *ibmr = NULL; + struct ib_phys_buf ibphysbuf; + struct nes_pd *nespd; + u64 tagged_offset; + u8 mpa_frame_offset = 0; + struct ietf_mpa_v2 *mpa_v2_frame; + u8 start_addr = 0; + u8 *start_ptr = &start_addr; + u8 **start_buff = &start_ptr; + u16 buff_len = 0; + struct sockaddr_in *laddr = (struct sockaddr_in *)&cm_id->local_addr; + struct sockaddr_in *raddr = (struct sockaddr_in *)&cm_id->remote_addr; ibqp = nes_get_qp(cm_id->device, conn_param->qpn); if (!ibqp) @@ -2334,72 +3217,117 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) nesdev = nesvnic->nesdev; adapter = nesdev->nesadapter; - nes_debug(NES_DBG_CM, "nesvnic=%p, netdev=%p, %s\n", - nesvnic, nesvnic->netdev, nesvnic->netdev->name); - - /* since this is from a listen, we were able to put node handle into cm_id */ cm_node = (struct nes_cm_node *)cm_id->provider_data; + nes_debug(NES_DBG_CM, "nes_accept: cm_node= %p nesvnic=%p, netdev=%p," + "%s\n", cm_node, nesvnic, nesvnic->netdev, + nesvnic->netdev->name); + + if (NES_CM_STATE_LISTENER_DESTROYED == cm_node->state) { + if (cm_node->loopbackpartner) + rem_ref_cm_node(cm_node->cm_core, cm_node->loopbackpartner); + rem_ref_cm_node(cm_node->cm_core, cm_node); + return -EINVAL; + } + passive_state = atomic_add_return(1, &cm_node->passive_state); + if (passive_state == NES_SEND_RESET_EVENT) { + rem_ref_cm_node(cm_node->cm_core, cm_node); + return -ECONNRESET; + } /* associate the node with the QP */ nesqp->cm_node = (void *)cm_node; + cm_node->nesqp = nesqp; + - nes_debug(NES_DBG_CM, "QP%u, cm_node=%p, jiffies = %lu\n", - nesqp->hwqp.qp_id, cm_node, jiffies); + nes_debug(NES_DBG_CM, "QP%u, cm_node=%p, jiffies = %lu listener = %p\n", + nesqp->hwqp.qp_id, cm_node, jiffies, cm_node->listener); atomic_inc(&cm_accepts); nes_debug(NES_DBG_CM, "netdev refcnt = %u.\n", - atomic_read(&nesvnic->netdev->refcnt)); + netdev_refcnt_read(nesvnic->netdev)); + + nesqp->ietf_frame_size = sizeof(struct ietf_mpa_v2); + /* allocate the ietf frame and space for private data */ + nesqp->ietf_frame = pci_alloc_consistent(nesdev->pcidev, + nesqp->ietf_frame_size + conn_param->private_data_len, + &nesqp->ietf_frame_pbase); + + if (!nesqp->ietf_frame) { + nes_debug(NES_DBG_CM, "Unable to allocate memory for private data\n"); + return -ENOMEM; + } + mpa_v2_frame = (struct ietf_mpa_v2 *)nesqp->ietf_frame; - /* allocate the ietf frame and space for private data */ - nesqp->ietf_frame = pci_alloc_consistent(nesdev->pcidev, - sizeof(struct ietf_mpa_frame) + conn_param->private_data_len, - &nesqp->ietf_frame_pbase); + if (cm_node->mpa_frame_rev == IETF_MPA_V1) + mpa_frame_offset = 4; - if (!nesqp->ietf_frame) { - nes_debug(NES_DBG_CM, "Unable to allocate memory for private data\n"); + if (cm_node->mpa_frame_rev == IETF_MPA_V1 || + cm_node->mpav2_ird_ord == IETF_NO_IRD_ORD) { + record_ird_ord(cm_node, (u16)conn_param->ird, (u16)conn_param->ord); + } + + memcpy(mpa_v2_frame->priv_data, conn_param->private_data, + conn_param->private_data_len); + + cm_build_mpa_frame(cm_node, start_buff, &buff_len, nesqp->ietf_frame, MPA_KEY_REPLY); + nesqp->private_data_len = conn_param->private_data_len; + + /* setup our first outgoing iWarp send WQE (the IETF frame response) */ + wqe = &nesqp->hwqp.sq_vbase[0]; + + if (raddr->sin_addr.s_addr != laddr->sin_addr.s_addr) { + u64temp = (unsigned long)nesqp; + nesibdev = nesvnic->nesibdev; + nespd = nesqp->nespd; + ibphysbuf.addr = nesqp->ietf_frame_pbase + mpa_frame_offset; + ibphysbuf.size = buff_len; + tagged_offset = (u64)(unsigned long)*start_buff; + ibmr = nesibdev->ibdev.reg_phys_mr((struct ib_pd *)nespd, + &ibphysbuf, 1, + IB_ACCESS_LOCAL_WRITE, + &tagged_offset); + if (!ibmr) { + nes_debug(NES_DBG_CM, "Unable to register memory region" + "for lSMM for cm_node = %p \n", + cm_node); + pci_free_consistent(nesdev->pcidev, + nesqp->private_data_len + nesqp->ietf_frame_size, + nesqp->ietf_frame, nesqp->ietf_frame_pbase); return -ENOMEM; } - - /* setup the MPA frame */ - nesqp->private_data_len = conn_param->private_data_len; - memcpy(nesqp->ietf_frame->key, IEFT_MPA_KEY_REP, IETF_MPA_KEY_SIZE); - - memcpy(nesqp->ietf_frame->priv_data, conn_param->private_data, - conn_param->private_data_len); - - nesqp->ietf_frame->priv_data_len = cpu_to_be16(conn_param->private_data_len); - nesqp->ietf_frame->rev = mpa_version; - nesqp->ietf_frame->flags = IETF_MPA_FLAGS_CRC; - - /* setup our first outgoing iWarp send WQE (the IETF frame response) */ - wqe = &nesqp->hwqp.sq_vbase[0]; - - if (cm_id->remote_addr.sin_addr.s_addr != cm_id->local_addr.sin_addr.s_addr) { - u64temp = (unsigned long)nesqp; - u64temp |= NES_SW_CONTEXT_ALIGN>>1; - set_wqe_64bit_value(wqe->wqe_words, NES_IWARP_SQ_WQE_COMP_CTX_LOW_IDX, - u64temp); - wqe->wqe_words[NES_IWARP_SQ_WQE_MISC_IDX] = - cpu_to_le32(NES_IWARP_SQ_WQE_STREAMING | NES_IWARP_SQ_WQE_WRPDU); - wqe->wqe_words[NES_IWARP_SQ_WQE_TOTAL_PAYLOAD_IDX] = - cpu_to_le32(conn_param->private_data_len + sizeof(struct ietf_mpa_frame)); - wqe->wqe_words[NES_IWARP_SQ_WQE_FRAG0_LOW_IDX] = - cpu_to_le32((u32)nesqp->ietf_frame_pbase); - wqe->wqe_words[NES_IWARP_SQ_WQE_FRAG0_HIGH_IDX] = - cpu_to_le32((u32)((u64)nesqp->ietf_frame_pbase >> 32)); - wqe->wqe_words[NES_IWARP_SQ_WQE_LENGTH0_IDX] = - cpu_to_le32(conn_param->private_data_len + sizeof(struct ietf_mpa_frame)); - wqe->wqe_words[NES_IWARP_SQ_WQE_STAG0_IDX] = 0; - - nesqp->nesqp_context->ird_ord_sizes |= cpu_to_le32( - NES_QPCONTEXT_ORDIRD_LSMM_PRESENT | NES_QPCONTEXT_ORDIRD_WRPDU); - } else { - nesqp->nesqp_context->ird_ord_sizes |= cpu_to_le32((NES_QPCONTEXT_ORDIRD_LSMM_PRESENT | - NES_QPCONTEXT_ORDIRD_WRPDU | NES_QPCONTEXT_ORDIRD_ALSMM)); + ibmr->pd = &nespd->ibpd; + ibmr->device = nespd->ibpd.device; + nesqp->lsmm_mr = ibmr; + + u64temp |= NES_SW_CONTEXT_ALIGN >> 1; + set_wqe_64bit_value(wqe->wqe_words, + NES_IWARP_SQ_WQE_COMP_CTX_LOW_IDX, + u64temp); + wqe->wqe_words[NES_IWARP_SQ_WQE_MISC_IDX] = + cpu_to_le32(NES_IWARP_SQ_WQE_STREAMING | + NES_IWARP_SQ_WQE_WRPDU); + wqe->wqe_words[NES_IWARP_SQ_WQE_TOTAL_PAYLOAD_IDX] = + cpu_to_le32(buff_len); + set_wqe_64bit_value(wqe->wqe_words, + NES_IWARP_SQ_WQE_FRAG0_LOW_IDX, + (u64)(unsigned long)(*start_buff)); + wqe->wqe_words[NES_IWARP_SQ_WQE_LENGTH0_IDX] = + cpu_to_le32(buff_len); + wqe->wqe_words[NES_IWARP_SQ_WQE_STAG0_IDX] = ibmr->lkey; + if (nesqp->sq_kmapped) { + nesqp->sq_kmapped = 0; + kunmap(nesqp->page); } - nesqp->skip_lsmm = 1; + nesqp->nesqp_context->ird_ord_sizes |= + cpu_to_le32(NES_QPCONTEXT_ORDIRD_LSMM_PRESENT | + NES_QPCONTEXT_ORDIRD_WRPDU); + } else { + nesqp->nesqp_context->ird_ord_sizes |= + cpu_to_le32(NES_QPCONTEXT_ORDIRD_WRPDU); + } + nesqp->skip_lsmm = 1; /* Cache the cm_id in the qp */ nesqp->cm_id = cm_id; @@ -2407,82 +3335,95 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) /* nesqp->cm_node = (void *)cm_id->provider_data; */ cm_id->provider_data = nesqp; - nesqp->active_conn = 0; + nesqp->active_conn = 0; + + if (cm_node->state == NES_CM_STATE_TSA) + nes_debug(NES_DBG_CM, "Already state = TSA for cm_node=%p\n", + cm_node); nes_cm_init_tsa_conn(nesqp, cm_node); - nesqp->nesqp_context->tcpPorts[0] = cpu_to_le16(ntohs(cm_id->local_addr.sin_port)); - nesqp->nesqp_context->tcpPorts[1] = cpu_to_le16(ntohs(cm_id->remote_addr.sin_port)); - nesqp->nesqp_context->ip0 = cpu_to_le32(ntohl(cm_id->remote_addr.sin_addr.s_addr)); + nesqp->nesqp_context->tcpPorts[0] = + cpu_to_le16(cm_node->mapped_loc_port); + nesqp->nesqp_context->tcpPorts[1] = + cpu_to_le16(cm_node->mapped_rem_port); + + nesqp->nesqp_context->ip0 = cpu_to_le32(cm_node->mapped_rem_addr); nesqp->nesqp_context->misc2 |= cpu_to_le32( - (u32)PCI_FUNC(nesdev->pcidev->devfn) << NES_QPCONTEXT_MISC2_SRC_IP_SHIFT); + (u32)PCI_FUNC(nesdev->pcidev->devfn) << + NES_QPCONTEXT_MISC2_SRC_IP_SHIFT); - nesqp->nesqp_context->arp_index_vlan |= cpu_to_le32( - nes_arp_table(nesdev, le32_to_cpu(nesqp->nesqp_context->ip0), NULL, - NES_ARP_RESOLVE) << 16); + nesqp->nesqp_context->arp_index_vlan |= + cpu_to_le32(nes_arp_table(nesdev, + le32_to_cpu(nesqp->nesqp_context->ip0), NULL, + NES_ARP_RESOLVE) << 16); nesqp->nesqp_context->ts_val_delta = cpu_to_le32( - jiffies - nes_read_indexed(nesdev, NES_IDX_TCP_NOW)); + jiffies - nes_read_indexed(nesdev, NES_IDX_TCP_NOW)); nesqp->nesqp_context->ird_index = cpu_to_le32(nesqp->hwqp.qp_id); nesqp->nesqp_context->ird_ord_sizes |= cpu_to_le32( - ((u32)1 << NES_QPCONTEXT_ORDIRD_IWARP_MODE_SHIFT)); - nesqp->nesqp_context->ird_ord_sizes |= cpu_to_le32((u32)conn_param->ord); + ((u32)1 << NES_QPCONTEXT_ORDIRD_IWARP_MODE_SHIFT)); + nesqp->nesqp_context->ird_ord_sizes |= + cpu_to_le32((u32)cm_node->ord_size); memset(&nes_quad, 0, sizeof(nes_quad)); - nes_quad.DstIpAdrIndex = cpu_to_le32((u32)PCI_FUNC(nesdev->pcidev->devfn) << 24); - nes_quad.SrcIpadr = cm_id->remote_addr.sin_addr.s_addr; - nes_quad.TcpPorts[0] = cm_id->remote_addr.sin_port; - nes_quad.TcpPorts[1] = cm_id->local_addr.sin_port; + nes_quad.DstIpAdrIndex = + cpu_to_le32((u32)PCI_FUNC(nesdev->pcidev->devfn) << 24); + nes_quad.SrcIpadr = htonl(cm_node->mapped_rem_addr); + nes_quad.TcpPorts[0] = htons(cm_node->mapped_rem_port); + nes_quad.TcpPorts[1] = htons(cm_node->mapped_loc_port); /* Produce hash key */ crc_value = get_crc_value(&nes_quad); nesqp->hte_index = cpu_to_be32(crc_value ^ 0xffffffff); nes_debug(NES_DBG_CM, "HTE Index = 0x%08X, CRC = 0x%08X\n", - nesqp->hte_index, nesqp->hte_index & adapter->hte_index_mask); + nesqp->hte_index, nesqp->hte_index & adapter->hte_index_mask); nesqp->hte_index &= adapter->hte_index_mask; nesqp->nesqp_context->hte_index = cpu_to_le32(nesqp->hte_index); cm_node->cm_core->api->accelerated(cm_node->cm_core, cm_node); - nes_debug(NES_DBG_CM, "QP%u, Destination IP = 0x%08X:0x%04X, local = 0x%08X:0x%04X," - " rcv_nxt=0x%08X, snd_nxt=0x%08X, mpa + private data length=%zu.\n", - nesqp->hwqp.qp_id, - ntohl(cm_id->remote_addr.sin_addr.s_addr), - ntohs(cm_id->remote_addr.sin_port), - ntohl(cm_id->local_addr.sin_addr.s_addr), - ntohs(cm_id->local_addr.sin_port), - le32_to_cpu(nesqp->nesqp_context->rcv_nxt), - le32_to_cpu(nesqp->nesqp_context->snd_nxt), - conn_param->private_data_len+sizeof(struct ietf_mpa_frame)); + nes_debug(NES_DBG_CM, "QP%u, Destination IP = 0x%08X:0x%04X, local = " + "0x%08X:0x%04X, rcv_nxt=0x%08X, snd_nxt=0x%08X, mpa + " + "private data length=%u.\n", nesqp->hwqp.qp_id, + ntohl(raddr->sin_addr.s_addr), ntohs(raddr->sin_port), + ntohl(laddr->sin_addr.s_addr), ntohs(laddr->sin_port), + le32_to_cpu(nesqp->nesqp_context->rcv_nxt), + le32_to_cpu(nesqp->nesqp_context->snd_nxt), + buff_len); - attr.qp_state = IB_QPS_RTS; - nes_modify_qp(&nesqp->ibqp, &attr, IB_QP_STATE, NULL); - - /* notify OF layer that accept event was successfull */ + /* notify OF layer that accept event was successful */ cm_id->add_ref(cm_id); + nes_add_ref(&nesqp->ibqp); cm_event.event = IW_CM_EVENT_ESTABLISHED; - cm_event.status = IW_CM_EVENT_STATUS_ACCEPTED; + cm_event.status = 0; cm_event.provider_data = (void *)nesqp; cm_event.local_addr = cm_id->local_addr; cm_event.remote_addr = cm_id->remote_addr; cm_event.private_data = NULL; cm_event.private_data_len = 0; + cm_event.ird = cm_node->ird_size; + cm_event.ord = cm_node->ord_size; + ret = cm_id->event_handler(cm_id, &cm_event); + attr.qp_state = IB_QPS_RTS; + nes_modify_qp(&nesqp->ibqp, &attr, IB_QP_STATE, NULL); if (cm_node->loopbackpartner) { - cm_node->loopbackpartner->mpa_frame_size = nesqp->private_data_len; + cm_node->loopbackpartner->mpa_frame_size = + nesqp->private_data_len; /* copy entire MPA frame to our cm_node's frame */ - memcpy(cm_node->loopbackpartner->mpa_frame_buf, nesqp->ietf_frame->priv_data, - nesqp->private_data_len); + memcpy(cm_node->loopbackpartner->mpa_frame_buf, + conn_param->private_data, conn_param->private_data_len); create_event(cm_node->loopbackpartner, NES_CM_EVENT_CONNECTED); } if (ret) - printk("%s[%u] OFA CM event_handler returned, ret=%d\n", - __FUNCTION__, __LINE__, ret); + printk(KERN_ERR "%s[%u] OFA CM event_handler returned, " + "ret=%d\n", __func__, __LINE__, ret); return 0; } @@ -2494,23 +3435,29 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) int nes_reject(struct iw_cm_id *cm_id, const void *pdata, u8 pdata_len) { struct nes_cm_node *cm_node; + struct nes_cm_node *loopback; struct nes_cm_core *cm_core; + u8 *start_buff; atomic_inc(&cm_rejects); - cm_node = (struct nes_cm_node *) cm_id->provider_data; + cm_node = (struct nes_cm_node *)cm_id->provider_data; + loopback = cm_node->loopbackpartner; cm_core = cm_node->cm_core; - cm_node->mpa_frame_size = sizeof(struct ietf_mpa_frame) + pdata_len; - - strcpy(&cm_node->mpa_frame.key[0], IEFT_MPA_KEY_REP); - memcpy(&cm_node->mpa_frame.priv_data, pdata, pdata_len); - - cm_node->mpa_frame.priv_data_len = cpu_to_be16(pdata_len); - cm_node->mpa_frame.rev = mpa_version; - cm_node->mpa_frame.flags = IETF_MPA_FLAGS_CRC | IETF_MPA_FLAGS_REJECT; + cm_node->cm_id = cm_id; - cm_core->api->reject(cm_core, &cm_node->mpa_frame, cm_node); + if (pdata_len + sizeof(struct ietf_mpa_v2) > MAX_CM_BUFFER) + return -EINVAL; - return 0; + if (loopback) { + memcpy(&loopback->mpa_frame.priv_data, pdata, pdata_len); + loopback->mpa_frame.priv_data_len = pdata_len; + loopback->mpa_frame_size = pdata_len; + } else { + start_buff = &cm_node->mpa_frame_buf[0] + sizeof(struct ietf_mpa_v2); + cm_node->mpa_frame_size = pdata_len; + memcpy(start_buff, pdata, pdata_len); + } + return cm_core->api->reject(cm_core, cm_node); } @@ -2526,7 +3473,15 @@ int nes_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) struct nes_device *nesdev; struct nes_cm_node *cm_node; struct nes_cm_info cm_info; - + int apbvt_set = 0; + struct sockaddr_in *laddr = (struct sockaddr_in *)&cm_id->local_addr; + struct sockaddr_in *raddr = (struct sockaddr_in *)&cm_id->remote_addr; + struct iwpm_dev_data pm_reg_msg; + struct iwpm_sa_data pm_msg; + int iwpm_err = 0; + + if (cm_id->remote_addr.ss_family != AF_INET) + return -ENOSYS; ibqp = nes_get_qp(cm_id->device, conn_param->qpn); if (!ibqp) return -EINVAL; @@ -2536,78 +3491,99 @@ int nes_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) nesvnic = to_nesvnic(nesqp->ibqp.device); if (!nesvnic) return -EINVAL; - nesdev = nesvnic->nesdev; + nesdev = nesvnic->nesdev; if (!nesdev) return -EINVAL; - atomic_inc(&cm_connects); + if (!laddr->sin_port || !raddr->sin_port) + return -EINVAL; - nesqp->ietf_frame = kzalloc(sizeof(struct ietf_mpa_frame) + - conn_param->private_data_len, GFP_KERNEL); - if (!nesqp->ietf_frame) - return -ENOMEM; + nes_debug(NES_DBG_CM, "QP%u, current IP = 0x%08X, Destination IP = " + "0x%08X:0x%04X, local = 0x%08X:0x%04X.\n", nesqp->hwqp.qp_id, + ntohl(nesvnic->local_ipaddr), ntohl(raddr->sin_addr.s_addr), + ntohs(raddr->sin_port), ntohl(laddr->sin_addr.s_addr), + ntohs(laddr->sin_port)); - /* set qp as having an active connection */ + atomic_inc(&cm_connects); nesqp->active_conn = 1; - nes_debug(NES_DBG_CM, "QP%u, Destination IP = 0x%08X:0x%04X, local = 0x%08X:0x%04X.\n", - nesqp->hwqp.qp_id, - ntohl(cm_id->remote_addr.sin_addr.s_addr), - ntohs(cm_id->remote_addr.sin_port), - ntohl(cm_id->local_addr.sin_addr.s_addr), - ntohs(cm_id->local_addr.sin_port)); - /* cache the cm_id in the qp */ nesqp->cm_id = cm_id; - cm_id->provider_data = nesqp; - - /* copy the private data */ - if (conn_param->private_data_len) { - memcpy(nesqp->ietf_frame->priv_data, conn_param->private_data, - conn_param->private_data_len); - } - nesqp->private_data_len = conn_param->private_data_len; - nesqp->nesqp_context->ird_ord_sizes |= cpu_to_le32((u32)conn_param->ord); - nes_debug(NES_DBG_CM, "requested ord = 0x%08X.\n", (u32)conn_param->ord); - nes_debug(NES_DBG_CM, "mpa private data len =%u\n", conn_param->private_data_len); - strcpy(&nesqp->ietf_frame->key[0], IEFT_MPA_KEY_REQ); - nesqp->ietf_frame->flags = IETF_MPA_FLAGS_CRC; - nesqp->ietf_frame->rev = IETF_MPA_VERSION; - nesqp->ietf_frame->priv_data_len = htons(conn_param->private_data_len); - - if (cm_id->local_addr.sin_addr.s_addr != cm_id->remote_addr.sin_addr.s_addr) - nes_manage_apbvt(nesvnic, ntohs(cm_id->local_addr.sin_port), - PCI_FUNC(nesdev->pcidev->devfn), NES_MANAGE_APBVT_ADD); + nes_debug(NES_DBG_CM, "requested ord = 0x%08X.\n", (u32)conn_param->ord); + nes_debug(NES_DBG_CM, "mpa private data len =%u\n", + conn_param->private_data_len); /* set up the connection params for the node */ - cm_info.loc_addr = (cm_id->local_addr.sin_addr.s_addr); - cm_info.loc_port = (cm_id->local_addr.sin_port); - cm_info.rem_addr = (cm_id->remote_addr.sin_addr.s_addr); - cm_info.rem_port = (cm_id->remote_addr.sin_port); + cm_info.loc_addr = ntohl(laddr->sin_addr.s_addr); + cm_info.loc_port = ntohs(laddr->sin_port); + cm_info.rem_addr = ntohl(raddr->sin_addr.s_addr); + cm_info.rem_port = ntohs(raddr->sin_port); cm_info.cm_id = cm_id; cm_info.conn_type = NES_CM_IWARP_CONN_TYPE; + /* No port mapper available, go with the specified peer information */ + cm_info.mapped_loc_addr = cm_info.loc_addr; + cm_info.mapped_loc_port = cm_info.loc_port; + cm_info.mapped_rem_addr = cm_info.rem_addr; + cm_info.mapped_rem_port = cm_info.rem_port; + + nes_form_reg_msg(nesvnic, &pm_reg_msg); + iwpm_err = iwpm_register_pid(&pm_reg_msg, RDMA_NL_NES); + if (iwpm_err) { + nes_debug(NES_DBG_NLMSG, + "Port Mapper reg pid fail (err = %d).\n", iwpm_err); + } + if (iwpm_valid_pid() && !iwpm_err) { + nes_form_pm_msg(&cm_info, &pm_msg); + iwpm_err = iwpm_add_and_query_mapping(&pm_msg, RDMA_NL_NES); + if (iwpm_err) + nes_debug(NES_DBG_NLMSG, + "Port Mapper query fail (err = %d).\n", iwpm_err); + else + nes_record_pm_msg(&cm_info, &pm_msg); + } + + if (laddr->sin_addr.s_addr != raddr->sin_addr.s_addr) { + nes_manage_apbvt(nesvnic, cm_info.mapped_loc_port, + PCI_FUNC(nesdev->pcidev->devfn), NES_MANAGE_APBVT_ADD); + apbvt_set = 1; + } + + if (nes_create_mapinfo(&cm_info)) + return -ENOMEM; + cm_id->add_ref(cm_id); - nes_add_ref(&nesqp->ibqp); /* create a connect CM node connection */ - cm_node = g_cm_core->api->connect(g_cm_core, nesvnic, nesqp->ietf_frame, &cm_info); + cm_node = g_cm_core->api->connect(g_cm_core, nesvnic, + conn_param->private_data_len, (void *)conn_param->private_data, + &cm_info); if (!cm_node) { - if (cm_id->local_addr.sin_addr.s_addr != cm_id->remote_addr.sin_addr.s_addr) - nes_manage_apbvt(nesvnic, ntohs(cm_id->local_addr.sin_port), - PCI_FUNC(nesdev->pcidev->devfn), NES_MANAGE_APBVT_DEL); - nes_rem_ref(&nesqp->ibqp); - kfree(nesqp->ietf_frame); - nesqp->ietf_frame = NULL; + if (apbvt_set) + nes_manage_apbvt(nesvnic, cm_info.mapped_loc_port, + PCI_FUNC(nesdev->pcidev->devfn), + NES_MANAGE_APBVT_DEL); + + nes_debug(NES_DBG_NLMSG, "Delete mapped_loc_port = %04X\n", + cm_info.mapped_loc_port); + nes_remove_mapinfo(cm_info.loc_addr, cm_info.loc_port, + cm_info.mapped_loc_addr, cm_info.mapped_loc_port); cm_id->rem_ref(cm_id); return -ENOMEM; } - cm_node->apbvt_set = 1; + record_ird_ord(cm_node, (u16)conn_param->ird, (u16)conn_param->ord); + if (cm_node->send_rdma0_op == SEND_RDMA_READ_ZERO && + cm_node->ord_size == 0) + cm_node->ord_size = 1; + + cm_node->apbvt_set = apbvt_set; nesqp->cm_node = cm_node; + cm_node->nesqp = nesqp; + nes_add_ref(&nesqp->ibqp); return 0; } @@ -2621,50 +3597,59 @@ int nes_create_listen(struct iw_cm_id *cm_id, int backlog) struct nes_vnic *nesvnic; struct nes_cm_listener *cm_node; struct nes_cm_info cm_info; - struct nes_adapter *adapter; int err; - + struct sockaddr_in *laddr = (struct sockaddr_in *)&cm_id->local_addr; nes_debug(NES_DBG_CM, "cm_id = %p, local port = 0x%04X.\n", - cm_id, ntohs(cm_id->local_addr.sin_port)); + cm_id, ntohs(laddr->sin_port)); + if (cm_id->local_addr.ss_family != AF_INET) + return -ENOSYS; nesvnic = to_nesvnic(cm_id->device); if (!nesvnic) return -EINVAL; - adapter = nesvnic->nesdev->nesadapter; + nes_debug(NES_DBG_CM, "nesvnic=%p, netdev=%p, %s\n", nesvnic, nesvnic->netdev, nesvnic->netdev->name); nes_debug(NES_DBG_CM, "nesvnic->local_ipaddr=0x%08x, sin_addr.s_addr=0x%08x\n", - nesvnic->local_ipaddr, cm_id->local_addr.sin_addr.s_addr); + nesvnic->local_ipaddr, laddr->sin_addr.s_addr); /* setup listen params in our api call struct */ - cm_info.loc_addr = nesvnic->local_ipaddr; - cm_info.loc_port = cm_id->local_addr.sin_port; + cm_info.loc_addr = ntohl(nesvnic->local_ipaddr); + cm_info.loc_port = ntohs(laddr->sin_port); cm_info.backlog = backlog; cm_info.cm_id = cm_id; cm_info.conn_type = NES_CM_IWARP_CONN_TYPE; + /* No port mapper available, go with the specified info */ + cm_info.mapped_loc_addr = cm_info.loc_addr; + cm_info.mapped_loc_port = cm_info.loc_port; cm_node = g_cm_core->api->listen(g_cm_core, nesvnic, &cm_info); if (!cm_node) { - printk("%s[%u] Error returned from listen API call\n", - __FUNCTION__, __LINE__); + printk(KERN_ERR "%s[%u] Error returned from listen API call\n", + __func__, __LINE__); return -ENOMEM; } cm_id->provider_data = cm_node; if (!cm_node->reused_node) { - err = nes_manage_apbvt(nesvnic, ntohs(cm_id->local_addr.sin_port), - PCI_FUNC(nesvnic->nesdev->pcidev->devfn), NES_MANAGE_APBVT_ADD); + if (nes_create_mapinfo(&cm_info)) + return -ENOMEM; + + err = nes_manage_apbvt(nesvnic, cm_node->mapped_loc_port, + PCI_FUNC(nesvnic->nesdev->pcidev->devfn), + NES_MANAGE_APBVT_ADD); if (err) { - printk("nes_manage_apbvt call returned %d.\n", err); + printk(KERN_ERR "nes_manage_apbvt call returned %d.\n", + err); g_cm_core->api->stop_listener(g_cm_core, (void *)cm_node); return err; } - cm_listens_created++; + atomic_inc(&cm_listens_created); } cm_id->add_ref(cm_id); @@ -2696,15 +3681,16 @@ int nes_destroy_listen(struct iw_cm_id *cm_id) */ int nes_cm_recv(struct sk_buff *skb, struct net_device *netdevice) { + int rc = 0; + cm_packets_received++; - if ((g_cm_core) && (g_cm_core->api)) { - g_cm_core->api->recv_pkt(g_cm_core, netdev_priv(netdevice), skb); - } else { + if ((g_cm_core) && (g_cm_core->api)) + rc = g_cm_core->api->recv_pkt(g_cm_core, netdev_priv(netdevice), skb); + else nes_debug(NES_DBG_CM, "Unable to process packet for CM," - " cm is not setup properly.\n"); - } + " cm is not setup properly.\n"); - return 0; + return rc; } @@ -2717,11 +3703,10 @@ int nes_cm_start(void) nes_debug(NES_DBG_CM, "\n"); /* create the primary CM core, pass this handle to subsequent core inits */ g_cm_core = nes_cm_alloc_core(); - if (g_cm_core) { + if (g_cm_core) return 0; - } else { + else return -ENOMEM; - } } @@ -2740,9 +3725,8 @@ int nes_cm_stop(void) * cm_event_connected * handle a connected event, setup QPs and HW */ -void cm_event_connected(struct nes_cm_event *event) +static void cm_event_connected(struct nes_cm_event *event) { - u64 u64temp; struct nes_qp *nesqp; struct nes_vnic *nesvnic; struct nes_device *nesdev; @@ -2751,10 +3735,12 @@ void cm_event_connected(struct nes_cm_event *event) struct ib_qp_attr attr; struct iw_cm_id *cm_id; struct iw_cm_event cm_event; - struct nes_hw_qp_wqe *wqe; struct nes_v4_quad nes_quad; u32 crc_value; int ret; + struct sockaddr_in *laddr; + struct sockaddr_in *raddr; + struct sockaddr_in *cm_event_laddr; /* get all our handles */ cm_node = event->cm_node; @@ -2764,111 +3750,97 @@ void cm_event_connected(struct nes_cm_event *event) nesvnic = to_nesvnic(nesqp->ibqp.device); nesdev = nesvnic->nesdev; nesadapter = nesdev->nesadapter; + laddr = (struct sockaddr_in *)&cm_id->local_addr; + raddr = (struct sockaddr_in *)&cm_id->remote_addr; + cm_event_laddr = (struct sockaddr_in *)&cm_event.local_addr; - if (nesqp->destroyed) { + if (nesqp->destroyed) return; - } atomic_inc(&cm_connecteds); nes_debug(NES_DBG_CM, "QP%u attempting to connect to 0x%08X:0x%04X on" - " local port 0x%04X. jiffies = %lu.\n", - nesqp->hwqp.qp_id, - ntohl(cm_id->remote_addr.sin_addr.s_addr), - ntohs(cm_id->remote_addr.sin_port), - ntohs(cm_id->local_addr.sin_port), - jiffies); + " local port 0x%04X. jiffies = %lu.\n", + nesqp->hwqp.qp_id, ntohl(raddr->sin_addr.s_addr), + ntohs(raddr->sin_port), ntohs(laddr->sin_port), jiffies); nes_cm_init_tsa_conn(nesqp, cm_node); /* set the QP tsa context */ - nesqp->nesqp_context->tcpPorts[0] = cpu_to_le16(ntohs(cm_id->local_addr.sin_port)); - nesqp->nesqp_context->tcpPorts[1] = cpu_to_le16(ntohs(cm_id->remote_addr.sin_port)); - nesqp->nesqp_context->ip0 = cpu_to_le32(ntohl(cm_id->remote_addr.sin_addr.s_addr)); + nesqp->nesqp_context->tcpPorts[0] = + cpu_to_le16(cm_node->mapped_loc_port); + nesqp->nesqp_context->tcpPorts[1] = + cpu_to_le16(cm_node->mapped_rem_port); + nesqp->nesqp_context->ip0 = cpu_to_le32(cm_node->mapped_rem_addr); nesqp->nesqp_context->misc2 |= cpu_to_le32( - (u32)PCI_FUNC(nesdev->pcidev->devfn) << NES_QPCONTEXT_MISC2_SRC_IP_SHIFT); + (u32)PCI_FUNC(nesdev->pcidev->devfn) << + NES_QPCONTEXT_MISC2_SRC_IP_SHIFT); nesqp->nesqp_context->arp_index_vlan |= cpu_to_le32( - nes_arp_table(nesdev, le32_to_cpu(nesqp->nesqp_context->ip0), + nes_arp_table(nesdev, + le32_to_cpu(nesqp->nesqp_context->ip0), NULL, NES_ARP_RESOLVE) << 16); nesqp->nesqp_context->ts_val_delta = cpu_to_le32( jiffies - nes_read_indexed(nesdev, NES_IDX_TCP_NOW)); nesqp->nesqp_context->ird_index = cpu_to_le32(nesqp->hwqp.qp_id); nesqp->nesqp_context->ird_ord_sizes |= - cpu_to_le32((u32)1 << NES_QPCONTEXT_ORDIRD_IWARP_MODE_SHIFT); + cpu_to_le32((u32)1 << + NES_QPCONTEXT_ORDIRD_IWARP_MODE_SHIFT); + nesqp->nesqp_context->ird_ord_sizes |= + cpu_to_le32((u32)cm_node->ord_size); /* Adjust tail for not having a LSMM */ - nesqp->hwqp.sq_tail = 1; - -#if defined(NES_SEND_FIRST_WRITE) - if (cm_node->send_write0) { - nes_debug(NES_DBG_CM, "Sending first write.\n"); - wqe = &nesqp->hwqp.sq_vbase[0]; - u64temp = (unsigned long)nesqp; - u64temp |= NES_SW_CONTEXT_ALIGN>>1; - set_wqe_64bit_value(wqe->wqe_words, NES_IWARP_SQ_WQE_COMP_CTX_LOW_IDX, - u64temp); - wqe->wqe_words[NES_IWARP_SQ_WQE_MISC_IDX] = cpu_to_le32(NES_IWARP_SQ_OP_RDMAW); - wqe->wqe_words[NES_IWARP_SQ_WQE_TOTAL_PAYLOAD_IDX] = 0; - wqe->wqe_words[NES_IWARP_SQ_WQE_FRAG0_LOW_IDX] = 0; - wqe->wqe_words[NES_IWARP_SQ_WQE_FRAG0_HIGH_IDX] = 0; - wqe->wqe_words[NES_IWARP_SQ_WQE_LENGTH0_IDX] = 0; - wqe->wqe_words[NES_IWARP_SQ_WQE_STAG0_IDX] = 0; - - /* use the reserved spot on the WQ for the extra first WQE */ - nesqp->nesqp_context->ird_ord_sizes &= cpu_to_le32(~(NES_QPCONTEXT_ORDIRD_LSMM_PRESENT | - NES_QPCONTEXT_ORDIRD_WRPDU | NES_QPCONTEXT_ORDIRD_ALSMM)); - nesqp->skip_lsmm = 1; - nesqp->hwqp.sq_tail = 0; - nes_write32(nesdev->regs + NES_WQE_ALLOC, - (1 << 24) | 0x00800000 | nesqp->hwqp.qp_id); - } -#endif + /*nesqp->hwqp.sq_tail = 1;*/ + + build_rdma0_msg(cm_node, &nesqp); + + nes_write32(nesdev->regs + NES_WQE_ALLOC, + (1 << 24) | 0x00800000 | nesqp->hwqp.qp_id); memset(&nes_quad, 0, sizeof(nes_quad)); - nes_quad.DstIpAdrIndex = cpu_to_le32((u32)PCI_FUNC(nesdev->pcidev->devfn) << 24); - nes_quad.SrcIpadr = cm_id->remote_addr.sin_addr.s_addr; - nes_quad.TcpPorts[0] = cm_id->remote_addr.sin_port; - nes_quad.TcpPorts[1] = cm_id->local_addr.sin_port; + nes_quad.DstIpAdrIndex = + cpu_to_le32((u32)PCI_FUNC(nesdev->pcidev->devfn) << 24); + nes_quad.SrcIpadr = htonl(cm_node->mapped_rem_addr); + nes_quad.TcpPorts[0] = htons(cm_node->mapped_rem_port); + nes_quad.TcpPorts[1] = htons(cm_node->mapped_loc_port); /* Produce hash key */ crc_value = get_crc_value(&nes_quad); nesqp->hte_index = cpu_to_be32(crc_value ^ 0xffffffff); nes_debug(NES_DBG_CM, "HTE Index = 0x%08X, After CRC = 0x%08X\n", - nesqp->hte_index, nesqp->hte_index & nesadapter->hte_index_mask); + nesqp->hte_index, nesqp->hte_index & nesadapter->hte_index_mask); nesqp->hte_index &= nesadapter->hte_index_mask; nesqp->nesqp_context->hte_index = cpu_to_le32(nesqp->hte_index); nesqp->ietf_frame = &cm_node->mpa_frame; - nesqp->private_data_len = (u8) cm_node->mpa_frame_size; + nesqp->private_data_len = (u8)cm_node->mpa_frame_size; cm_node->cm_core->api->accelerated(cm_node->cm_core, cm_node); - /* modify QP state to rts */ - attr.qp_state = IB_QPS_RTS; - nes_modify_qp(&nesqp->ibqp, &attr, IB_QP_STATE, NULL); - /* notify OF layer we successfully created the requested connection */ cm_event.event = IW_CM_EVENT_CONNECT_REPLY; - cm_event.status = IW_CM_EVENT_STATUS_ACCEPTED; + cm_event.status = 0; cm_event.provider_data = cm_id->provider_data; - cm_event.local_addr.sin_family = AF_INET; - cm_event.local_addr.sin_port = cm_id->local_addr.sin_port; + cm_event_laddr->sin_family = AF_INET; + cm_event_laddr->sin_port = laddr->sin_port; cm_event.remote_addr = cm_id->remote_addr; - cm_event.private_data = (void *)event->cm_node->mpa_frame_buf; - cm_event.private_data_len = (u8) event->cm_node->mpa_frame_size; + cm_event.private_data = (void *)event->cm_node->mpa_frame_buf; + cm_event.private_data_len = (u8)event->cm_node->mpa_frame_size; + cm_event.ird = cm_node->ird_size; + cm_event.ord = cm_node->ord_size; - cm_event.local_addr.sin_addr.s_addr = event->cm_info.rem_addr; + cm_event_laddr->sin_addr.s_addr = htonl(event->cm_info.rem_addr); ret = cm_id->event_handler(cm_id, &cm_event); nes_debug(NES_DBG_CM, "OFA CM event_handler returned, ret=%d\n", ret); if (ret) - printk("%s[%u] OFA CM event_handler returned, ret=%d\n", - __FUNCTION__, __LINE__, ret); - nes_debug(NES_DBG_CM, "Exiting connect thread for QP%u. jiffies = %lu\n", - nesqp->hwqp.qp_id, jiffies ); + printk(KERN_ERR "%s[%u] OFA CM event_handler returned, " + "ret=%d\n", __func__, __LINE__, ret); + attr.qp_state = IB_QPS_RTS; + nes_modify_qp(&nesqp->ibqp, &attr, IB_QP_STATE, NULL); - nes_rem_ref(&nesqp->ibqp); + nes_debug(NES_DBG_CM, "Exiting connect thread for QP%u. jiffies = " + "%lu\n", nesqp->hwqp.qp_id, jiffies); return; } @@ -2877,7 +3849,7 @@ void cm_event_connected(struct nes_cm_event *event) /** * cm_event_connect_error */ -void cm_event_connect_error(struct nes_cm_event *event) +static void cm_event_connect_error(struct nes_cm_event *event) { struct nes_qp *nesqp; struct iw_cm_id *cm_id; @@ -2889,40 +3861,46 @@ void cm_event_connect_error(struct nes_cm_event *event) return; cm_id = event->cm_node->cm_id; - if (!cm_id) { + if (!cm_id) return; - } nes_debug(NES_DBG_CM, "cm_node=%p, cm_id=%p\n", event->cm_node, cm_id); nesqp = cm_id->provider_data; - if (!nesqp) { + if (!nesqp) return; - } /* notify OF layer about this connection error event */ /* cm_id->rem_ref(cm_id); */ nesqp->cm_id = NULL; cm_id->provider_data = NULL; cm_event.event = IW_CM_EVENT_CONNECT_REPLY; - cm_event.status = IW_CM_EVENT_STATUS_REJECTED; + cm_event.status = -ECONNRESET; cm_event.provider_data = cm_id->provider_data; cm_event.local_addr = cm_id->local_addr; cm_event.remote_addr = cm_id->remote_addr; cm_event.private_data = NULL; cm_event.private_data_len = 0; - nes_debug(NES_DBG_CM, "call CM_EVENT REJECTED, local_addr=%08x, remove_addr=%08x\n", - cm_event.local_addr.sin_addr.s_addr, cm_event.remote_addr.sin_addr.s_addr); +#ifdef CONFIG_INFINIBAND_NES_DEBUG + { + struct sockaddr_in *cm_event_laddr = (struct sockaddr_in *) + &cm_event.local_addr; + struct sockaddr_in *cm_event_raddr = (struct sockaddr_in *) + &cm_event.remote_addr; + nes_debug(NES_DBG_CM, "call CM_EVENT REJECTED, local_addr=%08x, remote_addr=%08x\n", + cm_event_laddr->sin_addr.s_addr, cm_event_raddr->sin_addr.s_addr); + } +#endif ret = cm_id->event_handler(cm_id, &cm_event); nes_debug(NES_DBG_CM, "OFA CM event_handler returned, ret=%d\n", ret); if (ret) - printk("%s[%u] OFA CM event_handler returned, ret=%d\n", - __FUNCTION__, __LINE__, ret); - nes_rem_ref(&nesqp->ibqp); - cm_id->rem_ref(cm_id); + printk(KERN_ERR "%s[%u] OFA CM event_handler returned, " + "ret=%d\n", __func__, __LINE__, ret); + cm_id->rem_ref(cm_id); + rem_ref_cm_node(event->cm_node->cm_core, event->cm_node); return; } @@ -2930,7 +3908,7 @@ void cm_event_connect_error(struct nes_cm_event *event) /** * cm_event_reset */ -void cm_event_reset(struct nes_cm_event *event) +static void cm_event_reset(struct nes_cm_event *event) { struct nes_qp *nesqp; struct iw_cm_id *cm_id; @@ -2948,18 +3926,32 @@ void cm_event_reset(struct nes_cm_event *event) nes_debug(NES_DBG_CM, "%p - cm_id = %p\n", event->cm_node, cm_id); nesqp = cm_id->provider_data; + if (!nesqp) + return; nesqp->cm_id = NULL; /* cm_id->provider_data = NULL; */ cm_event.event = IW_CM_EVENT_DISCONNECT; - cm_event.status = IW_CM_EVENT_STATUS_RESET; + cm_event.status = -ECONNRESET; cm_event.provider_data = cm_id->provider_data; cm_event.local_addr = cm_id->local_addr; cm_event.remote_addr = cm_id->remote_addr; cm_event.private_data = NULL; cm_event.private_data_len = 0; + cm_id->add_ref(cm_id); ret = cm_id->event_handler(cm_id, &cm_event); + atomic_inc(&cm_closes); + cm_event.event = IW_CM_EVENT_CLOSE; + cm_event.status = 0; + cm_event.provider_data = cm_id->provider_data; + cm_event.local_addr = cm_id->local_addr; + cm_event.remote_addr = cm_id->remote_addr; + cm_event.private_data = NULL; + cm_event.private_data_len = 0; + nes_debug(NES_DBG_CM, "NODE %p Generating CLOSE\n", event->cm_node); + ret = cm_id->event_handler(cm_id, &cm_event); + nes_debug(NES_DBG_CM, "OFA CM event_handler returned, ret=%d\n", ret); @@ -2973,12 +3965,16 @@ void cm_event_reset(struct nes_cm_event *event) /** * cm_event_mpa_req */ -void cm_event_mpa_req(struct nes_cm_event *event) +static void cm_event_mpa_req(struct nes_cm_event *event) { - struct iw_cm_id *cm_id; + struct iw_cm_id *cm_id; struct iw_cm_event cm_event; int ret; struct nes_cm_node *cm_node; + struct sockaddr_in *cm_event_laddr = (struct sockaddr_in *) + &cm_event.local_addr; + struct sockaddr_in *cm_event_raddr = (struct sockaddr_in *) + &cm_event.remote_addr; cm_node = event->cm_node; if (!cm_node) @@ -2987,27 +3983,81 @@ void cm_event_mpa_req(struct nes_cm_event *event) atomic_inc(&cm_connect_reqs); nes_debug(NES_DBG_CM, "cm_node = %p - cm_id = %p, jiffies = %lu\n", - cm_node, cm_id, jiffies); + cm_node, cm_id, jiffies); cm_event.event = IW_CM_EVENT_CONNECT_REQUEST; - cm_event.status = IW_CM_EVENT_STATUS_OK; + cm_event.status = 0; cm_event.provider_data = (void *)cm_node; - cm_event.local_addr.sin_family = AF_INET; - cm_event.local_addr.sin_port = htons(event->cm_info.loc_port); - cm_event.local_addr.sin_addr.s_addr = htonl(event->cm_info.loc_addr); + cm_event_laddr->sin_family = AF_INET; + cm_event_laddr->sin_port = htons(event->cm_info.loc_port); + cm_event_laddr->sin_addr.s_addr = htonl(event->cm_info.loc_addr); + + cm_event_raddr->sin_family = AF_INET; + cm_event_raddr->sin_port = htons(event->cm_info.rem_port); + cm_event_raddr->sin_addr.s_addr = htonl(event->cm_info.rem_addr); + cm_event.private_data = cm_node->mpa_frame_buf; + cm_event.private_data_len = (u8)cm_node->mpa_frame_size; + if (cm_node->mpa_frame_rev == IETF_MPA_V1) { + cm_event.ird = NES_MAX_IRD; + cm_event.ord = NES_MAX_ORD; + } else { + cm_event.ird = cm_node->ird_size; + cm_event.ord = cm_node->ord_size; + } - cm_event.remote_addr.sin_family = AF_INET; - cm_event.remote_addr.sin_port = htons(event->cm_info.rem_port); - cm_event.remote_addr.sin_addr.s_addr = htonl(event->cm_info.rem_addr); + ret = cm_id->event_handler(cm_id, &cm_event); + if (ret) + printk(KERN_ERR "%s[%u] OFA CM event_handler returned, ret=%d\n", + __func__, __LINE__, ret); + return; +} - cm_event.private_data = cm_node->mpa_frame_buf; - cm_event.private_data_len = (u8) cm_node->mpa_frame_size; + +static void cm_event_mpa_reject(struct nes_cm_event *event) +{ + struct iw_cm_id *cm_id; + struct iw_cm_event cm_event; + struct nes_cm_node *cm_node; + int ret; + struct sockaddr_in *cm_event_laddr = (struct sockaddr_in *) + &cm_event.local_addr; + struct sockaddr_in *cm_event_raddr = (struct sockaddr_in *) + &cm_event.remote_addr; + + cm_node = event->cm_node; + if (!cm_node) + return; + cm_id = cm_node->cm_id; + + atomic_inc(&cm_connect_reqs); + nes_debug(NES_DBG_CM, "cm_node = %p - cm_id = %p, jiffies = %lu\n", + cm_node, cm_id, jiffies); + + cm_event.event = IW_CM_EVENT_CONNECT_REPLY; + cm_event.status = -ECONNREFUSED; + cm_event.provider_data = cm_id->provider_data; + + cm_event_laddr->sin_family = AF_INET; + cm_event_laddr->sin_port = htons(event->cm_info.loc_port); + cm_event_laddr->sin_addr.s_addr = htonl(event->cm_info.loc_addr); + + cm_event_raddr->sin_family = AF_INET; + cm_event_raddr->sin_port = htons(event->cm_info.rem_port); + cm_event_raddr->sin_addr.s_addr = htonl(event->cm_info.rem_addr); + + cm_event.private_data = cm_node->mpa_frame_buf; + cm_event.private_data_len = (u8)cm_node->mpa_frame_size; + + nes_debug(NES_DBG_CM, "call CM_EVENT_MPA_REJECTED, local_addr=%08x, " + "remove_addr=%08x\n", + cm_event_laddr->sin_addr.s_addr, + cm_event_raddr->sin_addr.s_addr); ret = cm_id->event_handler(cm_id, &cm_event); if (ret) - printk("%s[%u] OFA CM event_handler returned, ret=%d\n", - __FUNCTION__, __LINE__, ret); + printk(KERN_ERR "%s[%u] OFA CM event_handler returned, ret=%d\n", + __func__, __LINE__, ret); return; } @@ -3019,13 +4069,14 @@ static void nes_cm_event_handler(struct work_struct *); * nes_cm_post_event * post an event to the cm event handler */ -int nes_cm_post_event(struct nes_cm_event *event) +static int nes_cm_post_event(struct nes_cm_event *event) { atomic_inc(&event->cm_node->cm_core->events_posted); add_ref_cm_node(event->cm_node); event->cm_info.cm_id->add_ref(event->cm_info.cm_id); INIT_WORK(&event->event_work, nes_cm_event_handler); - nes_debug(NES_DBG_CM, "queue_work, event=%p\n", event); + nes_debug(NES_DBG_CM, "cm_node=%p queue_work, event=%p\n", + event->cm_node, event); queue_work(event->cm_node->cm_core->event_wq, &event->event_work); @@ -3041,46 +4092,56 @@ int nes_cm_post_event(struct nes_cm_event *event) */ static void nes_cm_event_handler(struct work_struct *work) { - struct nes_cm_event *event = container_of(work, struct nes_cm_event, event_work); + struct nes_cm_event *event = container_of(work, struct nes_cm_event, + event_work); struct nes_cm_core *cm_core; - if ((!event) || (!event->cm_node) || (!event->cm_node->cm_core)) { + if ((!event) || (!event->cm_node) || (!event->cm_node->cm_core)) return; - } + cm_core = event->cm_node->cm_core; nes_debug(NES_DBG_CM, "event=%p, event->type=%u, events posted=%u\n", - event, event->type, atomic_read(&cm_core->events_posted)); + event, event->type, atomic_read(&cm_core->events_posted)); switch (event->type) { - case NES_CM_EVENT_MPA_REQ: - cm_event_mpa_req(event); - nes_debug(NES_DBG_CM, "CM Event: MPA REQUEST\n"); - break; - case NES_CM_EVENT_RESET: - nes_debug(NES_DBG_CM, "CM Event: RESET\n"); - cm_event_reset(event); - break; - case NES_CM_EVENT_CONNECTED: - if ((!event->cm_node->cm_id) || - (event->cm_node->state != NES_CM_STATE_TSA)) { - break; - } - cm_event_connected(event); - nes_debug(NES_DBG_CM, "CM Event: CONNECTED\n"); - break; - case NES_CM_EVENT_ABORTED: - if ((!event->cm_node->cm_id) || (event->cm_node->state == NES_CM_STATE_TSA)) { - break; - } - cm_event_connect_error(event); - nes_debug(NES_DBG_CM, "CM Event: ABORTED\n"); + case NES_CM_EVENT_MPA_REQ: + cm_event_mpa_req(event); + nes_debug(NES_DBG_CM, "cm_node=%p CM Event: MPA REQUEST\n", + event->cm_node); + break; + case NES_CM_EVENT_RESET: + nes_debug(NES_DBG_CM, "cm_node = %p CM Event: RESET\n", + event->cm_node); + cm_event_reset(event); + break; + case NES_CM_EVENT_CONNECTED: + if ((!event->cm_node->cm_id) || + (event->cm_node->state != NES_CM_STATE_TSA)) break; - case NES_CM_EVENT_DROPPED_PKT: - nes_debug(NES_DBG_CM, "CM Event: DROPPED PKT\n"); + cm_event_connected(event); + nes_debug(NES_DBG_CM, "CM Event: CONNECTED\n"); + break; + case NES_CM_EVENT_MPA_REJECT: + if ((!event->cm_node->cm_id) || + (event->cm_node->state == NES_CM_STATE_TSA)) break; - default: - nes_debug(NES_DBG_CM, "CM Event: UNKNOWN EVENT TYPE\n"); + cm_event_mpa_reject(event); + nes_debug(NES_DBG_CM, "CM Event: REJECT\n"); + break; + + case NES_CM_EVENT_ABORTED: + if ((!event->cm_node->cm_id) || + (event->cm_node->state == NES_CM_STATE_TSA)) break; + cm_event_connect_error(event); + nes_debug(NES_DBG_CM, "CM Event: ABORTED\n"); + break; + case NES_CM_EVENT_DROPPED_PKT: + nes_debug(NES_DBG_CM, "CM Event: DROPPED PKT\n"); + break; + default: + nes_debug(NES_DBG_CM, "CM Event: UNKNOWN EVENT TYPE\n"); + break; } atomic_dec(&cm_core->events_posted); |
