diff options
Diffstat (limited to 'drivers/infiniband/hw/cxgb4/cm.c')
| -rw-r--r-- | drivers/infiniband/hw/cxgb4/cm.c | 810 |
1 files changed, 509 insertions, 301 deletions
diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c index e49b34c3b136..b3b45c49077d 100644 --- a/drivers/infiniband/hw/cxgb4/cm.c +++ b/drivers/infiniband/hw/cxgb4/cm.c @@ -77,9 +77,9 @@ static int enable_ecn; module_param(enable_ecn, int, 0644); MODULE_PARM_DESC(enable_ecn, "Enable ECN (default=0/disabled)"); -static int dack_mode = 1; +static int dack_mode; module_param(dack_mode, int, 0644); -MODULE_PARM_DESC(dack_mode, "Delayed ack mode (default=1)"); +MODULE_PARM_DESC(dack_mode, "Delayed ack mode (default=0)"); uint c4iw_max_read_depth = 32; module_param(c4iw_max_read_depth, int, 0644); @@ -99,10 +99,6 @@ module_param(enable_tcp_window_scaling, int, 0644); MODULE_PARM_DESC(enable_tcp_window_scaling, "Enable tcp window scaling (default=1)"); -int c4iw_debug; -module_param(c4iw_debug, int, 0644); -MODULE_PARM_DESC(c4iw_debug, "obsolete"); - static int peer2peer = 1; module_param(peer2peer, int, 0644); MODULE_PARM_DESC(peer2peer, "Support peer2peer ULPs (default=1)"); @@ -144,12 +140,12 @@ static struct workqueue_struct *workq; static struct sk_buff_head rxq; static struct sk_buff *get_skb(struct sk_buff *skb, int len, gfp_t gfp); -static void ep_timeout(unsigned long arg); +static void ep_timeout(struct timer_list *t); static void connect_reply_upcall(struct c4iw_ep *ep, int status); static int sched(struct c4iw_dev *dev, struct sk_buff *skb); static LIST_HEAD(timeout_list); -static spinlock_t timeout_lock; +static DEFINE_SPINLOCK(timeout_lock); static void deref_cm_id(struct c4iw_ep_common *epc) { @@ -180,7 +176,7 @@ static void ref_qp(struct c4iw_ep *ep) static void start_ep_timer(struct c4iw_ep *ep) { - pr_debug("%s ep %p\n", __func__, ep); + pr_debug("ep %p\n", ep); if (timer_pending(&ep->timer)) { pr_err("%s timer already started! ep %p\n", __func__, ep); @@ -189,15 +185,13 @@ static void start_ep_timer(struct c4iw_ep *ep) clear_bit(TIMEOUT, &ep->com.flags); c4iw_get_ep(&ep->com); ep->timer.expires = jiffies + ep_timeout_secs * HZ; - ep->timer.data = (unsigned long)ep; - ep->timer.function = ep_timeout; add_timer(&ep->timer); } static int stop_ep_timer(struct c4iw_ep *ep) { - pr_debug("%s ep %p stopping\n", __func__, ep); - del_timer_sync(&ep->timer); + pr_debug("ep %p stopping\n", ep); + timer_delete_sync(&ep->timer); if (!test_and_set_bit(TIMEOUT, &ep->com.flags)) { c4iw_put_ep(&ep->com); return 0; @@ -212,7 +206,7 @@ static int c4iw_l2t_send(struct c4iw_rdev *rdev, struct sk_buff *skb, if (c4iw_fatal_error(rdev)) { kfree_skb(skb); - pr_debug("%s - device in error state - dropping\n", __func__); + pr_err("%s - device in error state - dropping\n", __func__); return -EIO; } error = cxgb4_l2t_send(rdev->lldi.ports[0], skb, l2e); @@ -229,7 +223,7 @@ int c4iw_ofld_send(struct c4iw_rdev *rdev, struct sk_buff *skb) if (c4iw_fatal_error(rdev)) { kfree_skb(skb); - pr_debug("%s - device in error state - dropping\n", __func__); + pr_err("%s - device in error state - dropping\n", __func__); return -EIO; } error = cxgb4_ofld_send(rdev->lldi.ports[0], skb); @@ -265,8 +259,8 @@ static void set_emss(struct c4iw_ep *ep, u16 opt) if (ep->emss & 7) pr_debug("Warning: misaligned mtu idx %u mss %u emss=%u\n", TCPOPT_MSS_G(opt), ep->mss, ep->emss); - pr_debug("%s mss_idx %u mss %u emss=%u\n", __func__, TCPOPT_MSS_G(opt), - ep->mss, ep->emss); + pr_debug("mss_idx %u mss %u emss=%u\n", TCPOPT_MSS_G(opt), ep->mss, + ep->emss); } static enum c4iw_ep_state state_read(struct c4iw_ep_common *epc) @@ -287,7 +281,7 @@ static void __state_set(struct c4iw_ep_common *epc, enum c4iw_ep_state new) static void state_set(struct c4iw_ep_common *epc, enum c4iw_ep_state new) { mutex_lock(&epc->mutex); - pr_debug("%s - %s -> %s\n", __func__, states[epc->state], states[new]); + pr_debug("%s -> %s\n", states[epc->state], states[new]); __state_set(epc, new); mutex_unlock(&epc->mutex); return; @@ -318,11 +312,18 @@ static void *alloc_ep(int size, gfp_t gfp) epc = kzalloc(size, gfp); if (epc) { + epc->wr_waitp = c4iw_alloc_wr_wait(gfp); + if (!epc->wr_waitp) { + kfree(epc); + epc = NULL; + goto out; + } kref_init(&epc->kref); mutex_init(&epc->mutex); - c4iw_init_wr_wait(&epc->wr_wait); + c4iw_init_wr_wait(epc->wr_waitp); } - pr_debug("%s alloc ep %p\n", __func__, epc); + pr_debug("alloc ep %p\n", epc); +out: return epc; } @@ -330,20 +331,23 @@ static void remove_ep_tid(struct c4iw_ep *ep) { unsigned long flags; - spin_lock_irqsave(&ep->com.dev->lock, flags); - _remove_handle(ep->com.dev, &ep->com.dev->hwtid_idr, ep->hwtid, 0); - if (idr_is_empty(&ep->com.dev->hwtid_idr)) + xa_lock_irqsave(&ep->com.dev->hwtids, flags); + __xa_erase(&ep->com.dev->hwtids, ep->hwtid); + if (xa_empty(&ep->com.dev->hwtids)) wake_up(&ep->com.dev->wait); - spin_unlock_irqrestore(&ep->com.dev->lock, flags); + xa_unlock_irqrestore(&ep->com.dev->hwtids, flags); } -static void insert_ep_tid(struct c4iw_ep *ep) +static int insert_ep_tid(struct c4iw_ep *ep) { unsigned long flags; + int err; + + xa_lock_irqsave(&ep->com.dev->hwtids, flags); + err = __xa_insert(&ep->com.dev->hwtids, ep->hwtid, ep, GFP_KERNEL); + xa_unlock_irqrestore(&ep->com.dev->hwtids, flags); - spin_lock_irqsave(&ep->com.dev->lock, flags); - _insert_handle(ep->com.dev, &ep->com.dev->hwtid_idr, ep, ep->hwtid, 0); - spin_unlock_irqrestore(&ep->com.dev->lock, flags); + return err; } /* @@ -354,11 +358,11 @@ static struct c4iw_ep *get_ep_from_tid(struct c4iw_dev *dev, unsigned int tid) struct c4iw_ep *ep; unsigned long flags; - spin_lock_irqsave(&dev->lock, flags); - ep = idr_find(&dev->hwtid_idr, tid); + xa_lock_irqsave(&dev->hwtids, flags); + ep = xa_load(&dev->hwtids, tid); if (ep) c4iw_get_ep(&ep->com); - spin_unlock_irqrestore(&dev->lock, flags); + xa_unlock_irqrestore(&dev->hwtids, flags); return ep; } @@ -371,11 +375,11 @@ static struct c4iw_listen_ep *get_ep_from_stid(struct c4iw_dev *dev, struct c4iw_listen_ep *ep; unsigned long flags; - spin_lock_irqsave(&dev->lock, flags); - ep = idr_find(&dev->stid_idr, stid); + xa_lock_irqsave(&dev->stids, flags); + ep = xa_load(&dev->stids, stid); if (ep) c4iw_get_ep(&ep->com); - spin_unlock_irqrestore(&dev->lock, flags); + xa_unlock_irqrestore(&dev->stids, flags); return ep; } @@ -384,7 +388,7 @@ void _c4iw_free_ep(struct kref *kref) struct c4iw_ep *ep; ep = container_of(kref, struct c4iw_ep, com.kref); - pr_debug("%s ep %p state %s\n", __func__, ep, states[ep->com.state]); + pr_debug("ep %p state %s\n", ep, states[ep->com.state]); if (test_bit(QP_REFERENCED, &ep->com.flags)) deref_qp(ep); if (test_bit(RELEASE_RESOURCES, &ep->com.flags)) { @@ -402,11 +406,11 @@ void _c4iw_free_ep(struct kref *kref) ep->com.local_addr.ss_family); dst_release(ep->dst); cxgb4_l2t_release(ep->l2t); - if (ep->mpa_skb) - kfree_skb(ep->mpa_skb); + kfree_skb(ep->mpa_skb); } if (!skb_queue_empty(&ep->com.ep_skb_list)) skb_queue_purge(&ep->com.ep_skb_list); + c4iw_put_wr_wait(ep->com.wr_waitp); kfree(ep); } @@ -456,6 +460,8 @@ static struct sk_buff *get_skb(struct sk_buff *skb, int len, gfp_t gfp) skb_reset_transport_header(skb); } else { skb = alloc_skb(len, gfp); + if (!skb) + return NULL; } t4_set_arp_err_handler(skb, NULL, NULL); return skb; @@ -489,7 +495,6 @@ static int _put_ep_safe(struct c4iw_dev *dev, struct sk_buff *skb) ep = *((struct c4iw_ep **)(skb->cb + 2 * sizeof(void *))); release_ep_resources(ep); - kfree_skb(skb); return 0; } @@ -500,7 +505,6 @@ static int _put_pass_ep_safe(struct c4iw_dev *dev, struct sk_buff *skb) ep = *((struct c4iw_ep **)(skb->cb + 2 * sizeof(void *))); c4iw_put_ep(&ep->parent_ep->com); release_ep_resources(ep); - kfree_skb(skb); return 0; } @@ -554,7 +558,7 @@ static void act_open_req_arp_failure(void *handle, struct sk_buff *skb) cxgb4_clip_release(ep->com.dev->rdev.lldi.ports[0], (const u32 *)&sin6->sin6_addr.s6_addr, 1); } - remove_handle(ep->com.dev, &ep->com.dev->atid_idr, ep->atid); + xa_erase_irq(&ep->com.dev->atids, ep->atid); cxgb4_free_atid(ep->com.dev->rdev.lldi.tids, ep->atid); queue_arp_failure_cpl(ep, skb, FAKE_CPL_PUT_EP_SAFE); } @@ -570,7 +574,7 @@ static void abort_arp_failure(void *handle, struct sk_buff *skb) struct c4iw_rdev *rdev = &ep->com.dev->rdev; struct cpl_abort_req *req = cplhdr(skb); - pr_debug("%s rdev %p\n", __func__, rdev); + pr_debug("rdev %p\n", rdev); req->cmd = CPL_ABORT_NO_RST; skb_get(skb); ret = c4iw_ofld_send(rdev, skb); @@ -585,24 +589,29 @@ static int send_flowc(struct c4iw_ep *ep) { struct fw_flowc_wr *flowc; struct sk_buff *skb = skb_dequeue(&ep->com.ep_skb_list); - int i; u16 vlan = ep->l2t->vlan; int nparams; + int flowclen, flowclen16; if (WARN_ON(!skb)) return -ENOMEM; if (vlan == CPL_L2T_VLAN_NONE) - nparams = 8; - else nparams = 9; + else + nparams = 10; + + flowclen = offsetof(struct fw_flowc_wr, mnemval[nparams]); + flowclen16 = DIV_ROUND_UP(flowclen, 16); + flowclen = flowclen16 * 16; - flowc = __skb_put(skb, FLOWC_LEN); + flowc = __skb_put(skb, flowclen); + memset(flowc, 0, flowclen); flowc->op_to_nparams = cpu_to_be32(FW_WR_OP_V(FW_FLOWC_WR) | FW_FLOWC_WR_NPARAMS_V(nparams)); - flowc->flowid_len16 = cpu_to_be32(FW_WR_LEN16_V(DIV_ROUND_UP(FLOWC_LEN, - 16)) | FW_WR_FLOWID_V(ep->hwtid)); + flowc->flowid_len16 = cpu_to_be32(FW_WR_LEN16_V(flowclen16) | + FW_WR_FLOWID_V(ep->hwtid)); flowc->mnemval[0].mnemonic = FW_FLOWC_MNEM_PFNVFN; flowc->mnemval[0].val = cpu_to_be32(FW_PFVF_CMD_PFN_V @@ -621,21 +630,13 @@ static int send_flowc(struct c4iw_ep *ep) flowc->mnemval[6].val = cpu_to_be32(ep->snd_win); flowc->mnemval[7].mnemonic = FW_FLOWC_MNEM_MSS; flowc->mnemval[7].val = cpu_to_be32(ep->emss); - if (nparams == 9) { + flowc->mnemval[8].mnemonic = FW_FLOWC_MNEM_RCV_SCALE; + flowc->mnemval[8].val = cpu_to_be32(ep->snd_wscale); + if (nparams == 10) { u16 pri; - pri = (vlan & VLAN_PRIO_MASK) >> VLAN_PRIO_SHIFT; - flowc->mnemval[8].mnemonic = FW_FLOWC_MNEM_SCHEDCLASS; - flowc->mnemval[8].val = cpu_to_be32(pri); - } else { - /* Pad WR to 16 byte boundary */ - flowc->mnemval[8].mnemonic = 0; - flowc->mnemval[8].val = 0; - } - for (i = 0; i < 9; i++) { - flowc->mnemval[i].r4[0] = 0; - flowc->mnemval[i].r4[1] = 0; - flowc->mnemval[i].r4[2] = 0; + flowc->mnemval[9].mnemonic = FW_FLOWC_MNEM_SCHEDCLASS; + flowc->mnemval[9].val = cpu_to_be32(pri); } set_wr_txq(skb, CPL_PRIORITY_DATA, ep->txq_idx); @@ -647,7 +648,7 @@ static int send_halfclose(struct c4iw_ep *ep) struct sk_buff *skb = skb_dequeue(&ep->com.ep_skb_list); u32 wrlen = roundup(sizeof(struct cpl_close_con_req), 16); - pr_debug("%s ep %p tid %u\n", __func__, ep, ep->hwtid); + pr_debug("ep %p tid %u\n", ep, ep->hwtid); if (WARN_ON(!skb)) return -ENOMEM; @@ -657,12 +658,38 @@ static int send_halfclose(struct c4iw_ep *ep) return c4iw_l2t_send(&ep->com.dev->rdev, skb, ep->l2t); } -static int send_abort(struct c4iw_ep *ep) +static void read_tcb(struct c4iw_ep *ep) +{ + struct sk_buff *skb; + struct cpl_get_tcb *req; + int wrlen = roundup(sizeof(*req), 16); + + skb = get_skb(NULL, sizeof(*req), GFP_KERNEL); + if (WARN_ON(!skb)) + return; + + set_wr_txq(skb, CPL_PRIORITY_CONTROL, ep->ctrlq_idx); + req = (struct cpl_get_tcb *) skb_put(skb, wrlen); + memset(req, 0, wrlen); + INIT_TP_WR(req, ep->hwtid); + OPCODE_TID(req) = cpu_to_be32(MK_OPCODE_TID(CPL_GET_TCB, ep->hwtid)); + req->reply_ctrl = htons(REPLY_CHAN_V(0) | QUEUENO_V(ep->rss_qid)); + + /* + * keep a ref on the ep so the tcb is not unlocked before this + * cpl completes. The ref is released in read_tcb_rpl(). + */ + c4iw_get_ep(&ep->com); + if (WARN_ON(c4iw_ofld_send(&ep->com.dev->rdev, skb))) + c4iw_put_ep(&ep->com); +} + +static int send_abort_req(struct c4iw_ep *ep) { u32 wrlen = roundup(sizeof(struct cpl_abort_req), 16); struct sk_buff *req_skb = skb_dequeue(&ep->com.ep_skb_list); - pr_debug("%s ep %p tid %u\n", __func__, ep, ep->hwtid); + pr_debug("ep %p tid %u\n", ep, ep->hwtid); if (WARN_ON(!req_skb)) return -ENOMEM; @@ -672,6 +699,17 @@ static int send_abort(struct c4iw_ep *ep) return c4iw_l2t_send(&ep->com.dev->rdev, req_skb, ep->l2t); } +static int send_abort(struct c4iw_ep *ep) +{ + if (!ep->com.qp || !ep->com.qp->srq) { + send_abort_req(ep); + return 0; + } + set_bit(ABORT_REQ_IN_PROGRESS, &ep->com.flags); + read_tcb(ep); + return 0; +} + static int send_connect(struct c4iw_ep *ep) { struct cpl_act_open_req *req = NULL; @@ -696,7 +734,7 @@ static int send_connect(struct c4iw_ep *ep) &ep->com.remote_addr; int ret; enum chip_type adapter_type = ep->com.dev->rdev.lldi.adapter_type; - u32 isn = (prandom_u32() & ~7UL) - 1; + u32 isn = (get_random_u32() & ~7UL) - 1; struct net_device *netdev; u64 params; @@ -725,7 +763,7 @@ static int send_connect(struct c4iw_ep *ep) roundup(sizev4, 16) : roundup(sizev6, 16); - pr_debug("%s ep %p atid %u\n", __func__, ep, ep->atid); + pr_debug("ep %p atid %u\n", ep, ep->atid); skb = get_skb(NULL, wrlen, GFP_KERNEL); if (!skb) { @@ -824,13 +862,13 @@ static int send_connect(struct c4iw_ep *ep) t5req->params = cpu_to_be64(FILTER_TUPLE_V(params)); t5req->rsvd = cpu_to_be32(isn); - pr_debug("%s snd_isn %u\n", __func__, t5req->rsvd); + pr_debug("snd_isn %u\n", t5req->rsvd); t5req->opt2 = cpu_to_be32(opt2); } else { t6req->params = cpu_to_be64(FILTER_TUPLE_V(params)); t6req->rsvd = cpu_to_be32(isn); - pr_debug("%s snd_isn %u\n", __func__, t6req->rsvd); + pr_debug("snd_isn %u\n", t6req->rsvd); t6req->opt2 = cpu_to_be32(opt2); } } @@ -877,13 +915,13 @@ static int send_connect(struct c4iw_ep *ep) t5req6->params = cpu_to_be64(FILTER_TUPLE_V(params)); t5req6->rsvd = cpu_to_be32(isn); - pr_debug("%s snd_isn %u\n", __func__, t5req6->rsvd); + pr_debug("snd_isn %u\n", t5req6->rsvd); t5req6->opt2 = cpu_to_be32(opt2); } else { t6req6->params = cpu_to_be64(FILTER_TUPLE_V(params)); t6req6->rsvd = cpu_to_be32(isn); - pr_debug("%s snd_isn %u\n", __func__, t6req6->rsvd); + pr_debug("snd_isn %u\n", t6req6->rsvd); t6req6->opt2 = cpu_to_be32(opt2); } @@ -907,15 +945,13 @@ static int send_mpa_req(struct c4iw_ep *ep, struct sk_buff *skb, struct mpa_message *mpa; struct mpa_v2_conn_params mpa_v2_params; - pr_debug("%s ep %p tid %u pd_len %d\n", - __func__, ep, ep->hwtid, ep->plen); - - BUG_ON(skb_cloned(skb)); + pr_debug("ep %p tid %u pd_len %d\n", + ep, ep->hwtid, ep->plen); mpalen = sizeof(*mpa) + ep->plen; if (mpa_rev_to_use == 2) mpalen += sizeof(struct mpa_v2_conn_params); - wrlen = roundup(mpalen + sizeof *req, 16); + wrlen = roundup(mpalen + sizeof(*req), 16); skb = get_skb(skb, wrlen, GFP_KERNEL); if (!skb) { connect_reply_upcall(ep, -ENOMEM); @@ -959,9 +995,10 @@ static int send_mpa_req(struct c4iw_ep *ep, struct sk_buff *skb, } if (mpa_rev_to_use == 2) { - mpa->private_data_size = htons(ntohs(mpa->private_data_size) + - sizeof (struct mpa_v2_conn_params)); - pr_debug("%s initiator ird %u ord %u\n", __func__, ep->ird, + mpa->private_data_size = + htons(ntohs(mpa->private_data_size) + + sizeof(struct mpa_v2_conn_params)); + pr_debug("initiator ird %u ord %u\n", ep->ird, ep->ord); mpa_v2_params.ird = htons((u16)ep->ird); mpa_v2_params.ord = htons((u16)ep->ord); @@ -994,7 +1031,6 @@ static int send_mpa_req(struct c4iw_ep *ep, struct sk_buff *skb, */ skb_get(skb); t4_set_arp_err_handler(skb, NULL, arp_failure_discard); - BUG_ON(ep->mpa_skb); ep->mpa_skb = skb; ret = c4iw_l2t_send(&ep->com.dev->rdev, skb, ep->l2t); if (ret) @@ -1014,13 +1050,13 @@ static int send_mpa_reject(struct c4iw_ep *ep, const void *pdata, u8 plen) struct sk_buff *skb; struct mpa_v2_conn_params mpa_v2_params; - pr_debug("%s ep %p tid %u pd_len %d\n", - __func__, ep, ep->hwtid, ep->plen); + pr_debug("ep %p tid %u pd_len %d\n", + ep, ep->hwtid, ep->plen); mpalen = sizeof(*mpa) + plen; if (ep->mpa_attr.version == 2 && ep->mpa_attr.enhanced_rdma_conn) mpalen += sizeof(struct mpa_v2_conn_params); - wrlen = roundup(mpalen + sizeof *req, 16); + wrlen = roundup(mpalen + sizeof(*req), 16); skb = get_skb(NULL, wrlen, GFP_KERNEL); if (!skb) { @@ -1051,8 +1087,9 @@ static int send_mpa_reject(struct c4iw_ep *ep, const void *pdata, u8 plen) if (ep->mpa_attr.version == 2 && ep->mpa_attr.enhanced_rdma_conn) { mpa->flags |= MPA_ENHANCED_RDMA_CONN; - mpa->private_data_size = htons(ntohs(mpa->private_data_size) + - sizeof (struct mpa_v2_conn_params)); + mpa->private_data_size = + htons(ntohs(mpa->private_data_size) + + sizeof(struct mpa_v2_conn_params)); mpa_v2_params.ird = htons(((u16)ep->ird) | (peer2peer ? MPA_V2_PEER2PEER_MODEL : 0)); @@ -1080,7 +1117,6 @@ static int send_mpa_reject(struct c4iw_ep *ep, const void *pdata, u8 plen) skb_get(skb); set_wr_txq(skb, CPL_PRIORITY_DATA, ep->txq_idx); t4_set_arp_err_handler(skb, NULL, mpa_start_arp_failure); - BUG_ON(ep->mpa_skb); ep->mpa_skb = skb; ep->snd_seq += mpalen; return c4iw_l2t_send(&ep->com.dev->rdev, skb, ep->l2t); @@ -1094,13 +1130,13 @@ static int send_mpa_reply(struct c4iw_ep *ep, const void *pdata, u8 plen) struct sk_buff *skb; struct mpa_v2_conn_params mpa_v2_params; - pr_debug("%s ep %p tid %u pd_len %d\n", - __func__, ep, ep->hwtid, ep->plen); + pr_debug("ep %p tid %u pd_len %d\n", + ep, ep->hwtid, ep->plen); mpalen = sizeof(*mpa) + plen; if (ep->mpa_attr.version == 2 && ep->mpa_attr.enhanced_rdma_conn) mpalen += sizeof(struct mpa_v2_conn_params); - wrlen = roundup(mpalen + sizeof *req, 16); + wrlen = roundup(mpalen + sizeof(*req), 16); skb = get_skb(NULL, wrlen, GFP_KERNEL); if (!skb) { @@ -1135,8 +1171,9 @@ static int send_mpa_reply(struct c4iw_ep *ep, const void *pdata, u8 plen) if (ep->mpa_attr.version == 2 && ep->mpa_attr.enhanced_rdma_conn) { mpa->flags |= MPA_ENHANCED_RDMA_CONN; - mpa->private_data_size = htons(ntohs(mpa->private_data_size) + - sizeof (struct mpa_v2_conn_params)); + mpa->private_data_size = + htons(ntohs(mpa->private_data_size) + + sizeof(struct mpa_v2_conn_params)); mpa_v2_params.ird = htons((u16)ep->ird); mpa_v2_params.ord = htons((u16)ep->ord); if (peer2peer && (ep->mpa_attr.p2p_type != @@ -1178,14 +1215,17 @@ static int act_establish(struct c4iw_dev *dev, struct sk_buff *skb) { struct c4iw_ep *ep; struct cpl_act_establish *req = cplhdr(skb); + unsigned short tcp_opt = ntohs(req->tcp_opt); unsigned int tid = GET_TID(req); unsigned int atid = TID_TID_G(ntohl(req->tos_atid)); struct tid_info *t = dev->rdev.lldi.tids; int ret; ep = lookup_atid(t, atid); + if (!ep) + return -EINVAL; - pr_debug("%s ep %p tid %u snd_isn %u rcv_isn %u\n", __func__, ep, tid, + pr_debug("ep %p tid %u snd_isn %u rcv_isn %u\n", ep, tid, be32_to_cpu(req->snd_isn), be32_to_cpu(req->rcv_isn)); mutex_lock(&ep->com.mutex); @@ -1198,11 +1238,12 @@ static int act_establish(struct c4iw_dev *dev, struct sk_buff *skb) ep->snd_seq = be32_to_cpu(req->snd_isn); ep->rcv_seq = be32_to_cpu(req->rcv_isn); + ep->snd_wscale = TCPOPT_SND_WSCALE_G(tcp_opt); - set_emss(ep, ntohs(req->tcp_opt)); + set_emss(ep, tcp_opt); /* dealloc the atid */ - remove_handle(ep->com.dev, &ep->com.dev->atid_idr, atid); + xa_erase_irq(&ep->com.dev->atids, atid); cxgb4_free_atid(t, atid); set_bit(ACT_ESTAB, &ep->com.history); @@ -1229,7 +1270,7 @@ static void close_complete_upcall(struct c4iw_ep *ep, int status) { struct iw_cm_event event; - pr_debug("%s ep %p tid %u\n", __func__, ep, ep->hwtid); + pr_debug("ep %p tid %u\n", ep, ep->hwtid); memset(&event, 0, sizeof(event)); event.event = IW_CM_EVENT_CLOSE; event.status = status; @@ -1246,7 +1287,7 @@ static void peer_close_upcall(struct c4iw_ep *ep) { struct iw_cm_event event; - pr_debug("%s ep %p tid %u\n", __func__, ep, ep->hwtid); + pr_debug("ep %p tid %u\n", ep, ep->hwtid); memset(&event, 0, sizeof(event)); event.event = IW_CM_EVENT_DISCONNECT; if (ep->com.cm_id) { @@ -1261,7 +1302,7 @@ static void peer_abort_upcall(struct c4iw_ep *ep) { struct iw_cm_event event; - pr_debug("%s ep %p tid %u\n", __func__, ep, ep->hwtid); + pr_debug("ep %p tid %u\n", ep, ep->hwtid); memset(&event, 0, sizeof(event)); event.event = IW_CM_EVENT_CLOSE; event.status = -ECONNRESET; @@ -1278,8 +1319,8 @@ static void connect_reply_upcall(struct c4iw_ep *ep, int status) { struct iw_cm_event event; - pr_debug("%s ep %p tid %u status %d\n", - __func__, ep, ep->hwtid, status); + pr_debug("ep %p tid %u status %d\n", + ep, ep->hwtid, status); memset(&event, 0, sizeof(event)); event.event = IW_CM_EVENT_CONNECT_REPLY; event.status = status; @@ -1308,7 +1349,7 @@ static void connect_reply_upcall(struct c4iw_ep *ep, int status) } } - pr_debug("%s ep %p tid %u status %d\n", __func__, ep, + pr_debug("ep %p tid %u status %d\n", ep, ep->hwtid, status); set_bit(CONN_RPL_UPCALL, &ep->com.history); ep->com.cm_id->event_handler(ep->com.cm_id, &event); @@ -1322,7 +1363,7 @@ static int connect_request_upcall(struct c4iw_ep *ep) struct iw_cm_event event; int ret; - pr_debug("%s ep %p tid %u\n", __func__, ep, ep->hwtid); + pr_debug("ep %p tid %u\n", ep, ep->hwtid); memset(&event, 0, sizeof(event)); event.event = IW_CM_EVENT_CONNECT_REQUEST; memcpy(&event.local_addr, &ep->com.local_addr, @@ -1359,13 +1400,13 @@ static void established_upcall(struct c4iw_ep *ep) { struct iw_cm_event event; - pr_debug("%s ep %p tid %u\n", __func__, ep, ep->hwtid); + pr_debug("ep %p tid %u\n", ep, ep->hwtid); memset(&event, 0, sizeof(event)); event.event = IW_CM_EVENT_ESTABLISHED; event.ird = ep->ord; event.ord = ep->ird; if (ep->com.cm_id) { - pr_debug("%s ep %p tid %u\n", __func__, ep, ep->hwtid); + pr_debug("ep %p tid %u\n", ep, ep->hwtid); ep->com.cm_id->event_handler(ep->com.cm_id, &event); set_bit(ESTAB_UPCALL, &ep->com.history); } @@ -1377,8 +1418,8 @@ static int update_rx_credits(struct c4iw_ep *ep, u32 credits) u32 wrlen = roundup(sizeof(struct cpl_rx_data_ack), 16); u32 credit_dack; - pr_debug("%s ep %p tid %u credits %u\n", - __func__, ep, ep->hwtid, credits); + pr_debug("ep %p tid %u credits %u\n", + ep, ep->hwtid, credits); skb = get_skb(NULL, wrlen, GFP_KERNEL); if (!skb) { pr_err("update_rx_credits - cannot alloc skb!\n"); @@ -1429,7 +1470,7 @@ static int process_mpa_reply(struct c4iw_ep *ep, struct sk_buff *skb) int err; int disconnect = 0; - pr_debug("%s ep %p tid %u\n", __func__, ep, ep->hwtid); + pr_debug("ep %p tid %u\n", ep, ep->hwtid); /* * If we get more than the supported amount of private data @@ -1527,8 +1568,7 @@ static int process_mpa_reply(struct c4iw_ep *ep, struct sk_buff *skb) MPA_V2_IRD_ORD_MASK; resp_ord = ntohs(mpa_v2_params->ord) & MPA_V2_IRD_ORD_MASK; - pr_debug("%s responder ird %u ord %u ep ird %u ord %u\n", - __func__, + pr_debug("responder ird %u ord %u ep ird %u ord %u\n", resp_ird, resp_ord, ep->ird, ep->ord); /* @@ -1573,8 +1613,8 @@ static int process_mpa_reply(struct c4iw_ep *ep, struct sk_buff *skb) if (peer2peer) ep->mpa_attr.p2p_type = p2p_type; - pr_debug("%s - crc_enabled=%d, recv_marker_enabled=%d, xmit_marker_enabled=%d, version=%d p2p_type=%d local-p2p_type = %d\n", - __func__, ep->mpa_attr.crc_enabled, + pr_debug("crc_enabled=%d, recv_marker_enabled=%d, xmit_marker_enabled=%d, version=%d p2p_type=%d local-p2p_type = %d\n", + ep->mpa_attr.crc_enabled, ep->mpa_attr.recv_marker_enabled, ep->mpa_attr.xmit_marker_enabled, ep->mpa_attr.version, ep->mpa_attr.p2p_type, p2p_type); @@ -1670,7 +1710,7 @@ static int process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb) struct mpa_v2_conn_params *mpa_v2_params; u16 plen; - pr_debug("%s ep %p tid %u\n", __func__, ep, ep->hwtid); + pr_debug("ep %p tid %u\n", ep, ep->hwtid); /* * If we get more than the supported amount of private data @@ -1679,7 +1719,7 @@ static int process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb) if (ep->mpa_pkt_len + skb->len > sizeof(ep->mpa_pkt)) goto err_stop_timer; - pr_debug("%s enter (%s line %u)\n", __func__, __FILE__, __LINE__); + pr_debug("enter (%s line %u)\n", __FILE__, __LINE__); /* * Copy the new data into our accumulation buffer. @@ -1695,7 +1735,7 @@ static int process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb) if (ep->mpa_pkt_len < sizeof(*mpa)) return 0; - pr_debug("%s enter (%s line %u)\n", __func__, __FILE__, __LINE__); + pr_debug("enter (%s line %u)\n", __FILE__, __LINE__); mpa = (struct mpa_message *) ep->mpa_pkt; /* @@ -1758,8 +1798,8 @@ static int process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb) MPA_V2_IRD_ORD_MASK; ep->ord = min_t(u32, ep->ord, cur_max_read_depth(ep->com.dev)); - pr_debug("%s initiator ird %u ord %u\n", - __func__, ep->ird, ep->ord); + pr_debug("initiator ird %u ord %u\n", + ep->ird, ep->ord); if (ntohs(mpa_v2_params->ird) & MPA_V2_PEER2PEER_MODEL) if (peer2peer) { if (ntohs(mpa_v2_params->ord) & @@ -1776,8 +1816,7 @@ static int process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb) if (peer2peer) ep->mpa_attr.p2p_type = p2p_type; - pr_debug("%s - crc_enabled=%d, recv_marker_enabled=%d, xmit_marker_enabled=%d, version=%d p2p_type=%d\n", - __func__, + pr_debug("crc_enabled=%d, recv_marker_enabled=%d, xmit_marker_enabled=%d, version=%d p2p_type=%d\n", ep->mpa_attr.crc_enabled, ep->mpa_attr.recv_marker_enabled, ep->mpa_attr.xmit_marker_enabled, ep->mpa_attr.version, ep->mpa_attr.p2p_type); @@ -1816,7 +1855,7 @@ static int rx_data(struct c4iw_dev *dev, struct sk_buff *skb) ep = get_ep_from_tid(dev, tid); if (!ep) return 0; - pr_debug("%s ep %p tid %u dlen %u\n", __func__, ep, ep->hwtid, dlen); + pr_debug("ep %p tid %u dlen %u\n", ep, ep->hwtid, dlen); skb_pull(skb, sizeof(*hdr)); skb_trim(skb, dlen); mutex_lock(&ep->com.mutex); @@ -1836,7 +1875,6 @@ static int rx_data(struct c4iw_dev *dev, struct sk_buff *skb) struct c4iw_qp_attributes attrs; update_rx_credits(ep, dlen); - BUG_ON(!ep->com.qp); if (status) pr_err("%s Unexpected streaming data." \ " qpid %u ep %p state %d tid %u status %d\n", @@ -1858,10 +1896,31 @@ static int rx_data(struct c4iw_dev *dev, struct sk_buff *skb) return 0; } +static void complete_cached_srq_buffers(struct c4iw_ep *ep, u32 srqidx) +{ + enum chip_type adapter_type; + + adapter_type = ep->com.dev->rdev.lldi.adapter_type; + + /* + * If this TCB had a srq buffer cached, then we must complete + * it. For user mode, that means saving the srqidx in the + * user/kernel status page for this qp. For kernel mode, just + * synthesize the CQE now. + */ + if (CHELSIO_CHIP_VERSION(adapter_type) > CHELSIO_T5 && srqidx) { + if (ep->com.qp->ibqp.uobject) + t4_set_wq_in_error(&ep->com.qp->wq, srqidx); + else + c4iw_flush_srqidx(ep->com.qp, srqidx); + } +} + static int abort_rpl(struct c4iw_dev *dev, struct sk_buff *skb) { + u32 srqidx; struct c4iw_ep *ep; - struct cpl_abort_rpl_rss *rpl = cplhdr(skb); + struct cpl_abort_rpl_rss6 *rpl = cplhdr(skb); int release = 0; unsigned int tid = GET_TID(rpl); @@ -1870,11 +1929,17 @@ static int abort_rpl(struct c4iw_dev *dev, struct sk_buff *skb) pr_warn("Abort rpl to freed endpoint\n"); return 0; } - pr_debug("%s ep %p tid %u\n", __func__, ep, ep->hwtid); + + if (ep->com.qp && ep->com.qp->srq) { + srqidx = ABORT_RSS_SRQIDX_G(be32_to_cpu(rpl->srqidx_status)); + complete_cached_srq_buffers(ep, srqidx ? srqidx : ep->srqe_idx); + } + + pr_debug("ep %p tid %u\n", ep, ep->hwtid); mutex_lock(&ep->com.mutex); switch (ep->com.state) { case ABORTING: - c4iw_wake_up(&ep->com.wr_wait, -ECONNRESET); + c4iw_wake_up_noref(ep->com.wr_waitp, -ECONNRESET); __state_set(&ep->com, DEAD); release = 1; break; @@ -1884,8 +1949,10 @@ static int abort_rpl(struct c4iw_dev *dev, struct sk_buff *skb) } mutex_unlock(&ep->com.mutex); - if (release) + if (release) { + close_complete_upcall(ep, -ECONNRESET); release_ep_resources(ep); + } c4iw_put_ep(&ep->com); return 0; } @@ -1900,6 +1967,9 @@ static int send_fw_act_open_req(struct c4iw_ep *ep, unsigned int atid) int win; skb = get_skb(NULL, sizeof(*req), GFP_KERNEL); + if (!skb) + return -ENOMEM; + req = __skb_put_zero(skb, sizeof(*req)); req->op_compl = htonl(WR_OP_V(FW_OFLD_CONNECTION_WR)); req->len16_pkd = htonl(FW_WR_LEN16_V(DIV_ROUND_UP(sizeof(*req), 16))); @@ -1994,8 +2064,8 @@ static void set_tcp_window(struct c4iw_ep *ep, struct port_info *pi) { ep->snd_win = snd_win; ep->rcv_win = rcv_win; - pr_debug("%s snd_win %d rcv_win %d\n", - __func__, ep->snd_win, ep->rcv_win); + pr_debug("snd_win %d rcv_win %d\n", + ep->snd_win, ep->rcv_win); } #define ACT_OPEN_RETRY_COUNT 2 @@ -2016,7 +2086,7 @@ static int import_ep(struct c4iw_ep *ep, int iptype, __u8 *peer_ip, err = -ENOMEM; if (n->dev->flags & IFF_LOOPBACK) { if (iptype == 4) - pdev = ip_dev_find(&init_net, *(__be32 *)peer_ip); + pdev = __ip_dev_find(&init_net, *(__be32 *)peer_ip, false); else if (IS_ENABLED(CONFIG_IPV6)) for_each_netdev(&init_net, pdev) { if (ipv6_chk_addr(&init_net, @@ -2031,16 +2101,15 @@ static int import_ep(struct c4iw_ep *ep, int iptype, __u8 *peer_ip, err = -ENODEV; goto out; } + if (is_vlan_dev(pdev)) + pdev = vlan_dev_real_dev(pdev); ep->l2t = cxgb4_l2t_get(cdev->rdev.lldi.l2t, n, pdev, rt_tos2priority(tos)); - if (!ep->l2t) { - dev_put(pdev); + if (!ep->l2t) goto out; - } ep->mtu = pdev->mtu; ep->tx_chan = cxgb4_port_chan(pdev); - ep->smac_idx = cxgb4_tp_smt_idx(adapter_type, - cxgb4_port_viid(pdev)); + ep->smac_idx = ((struct port_info *)netdev_priv(pdev))->smt_idx; step = cdev->rdev.lldi.ntxq / cdev->rdev.lldi.nchan; ep->txq_idx = cxgb4_port_idx(pdev) * step; @@ -2050,17 +2119,15 @@ static int import_ep(struct c4iw_ep *ep, int iptype, __u8 *peer_ip, ep->rss_qid = cdev->rdev.lldi.rxq_ids[ cxgb4_port_idx(pdev) * step]; set_tcp_window(ep, (struct port_info *)netdev_priv(pdev)); - dev_put(pdev); } else { pdev = get_real_dev(n->dev); ep->l2t = cxgb4_l2t_get(cdev->rdev.lldi.l2t, - n, pdev, 0); + n, pdev, rt_tos2priority(tos)); if (!ep->l2t) goto out; ep->mtu = dst_mtu(dst); ep->tx_chan = cxgb4_port_chan(pdev); - ep->smac_idx = cxgb4_tp_smt_idx(adapter_type, - cxgb4_port_viid(pdev)); + ep->smac_idx = ((struct port_info *)netdev_priv(pdev))->smt_idx; step = cdev->rdev.lldi.ntxq / cdev->rdev.lldi.nchan; ep->txq_idx = cxgb4_port_idx(pdev) * step; @@ -2100,9 +2167,8 @@ static int c4iw_reconnect(struct c4iw_ep *ep) int iptype; __u8 *ra; - pr_debug("%s qp %p cm_id %p\n", __func__, ep->com.qp, ep->com.cm_id); - init_timer(&ep->timer); - c4iw_init_wr_wait(&ep->com.wr_wait); + pr_debug("qp %p cm_id %p\n", ep->com.qp, ep->com.cm_id); + c4iw_init_wr_wait(ep->com.wr_waitp); /* When MPA revision is different on nodes, the node with MPA_rev=2 * tries to reconnect with MPA_rev 1 for the same EP through @@ -2110,7 +2176,7 @@ static int c4iw_reconnect(struct c4iw_ep *ep) * further connection establishment. As we are using the same EP pointer * for reconnect, few skbs are used during the previous c4iw_connect(), * which leaves the EP with inadequate skbs for further - * c4iw_reconnect(), Further causing an assert BUG_ON() due to empty + * c4iw_reconnect(), Further causing a crash due to an empty * skb_list() during peer_abort(). Allocate skbs which is already used. */ size = (CN_MAX_CON_BUF - skb_queue_len(&ep->com.ep_skb_list)); @@ -2128,7 +2194,9 @@ static int c4iw_reconnect(struct c4iw_ep *ep) err = -ENOMEM; goto fail2; } - insert_handle(ep->com.dev, &ep->com.dev->atid_idr, ep, ep->atid); + err = xa_insert_irq(&ep->com.dev->atids, ep->atid, ep, GFP_KERNEL); + if (err) + goto fail2a; /* find a route */ if (ep->com.cm_id->m_local_addr.ss_family == AF_INET) { @@ -2145,7 +2213,8 @@ static int c4iw_reconnect(struct c4iw_ep *ep) laddr6->sin6_addr.s6_addr, raddr6->sin6_addr.s6_addr, laddr6->sin6_port, - raddr6->sin6_port, 0, + raddr6->sin6_port, + ep->com.cm_id->tos, raddr6->sin6_scope_id); iptype = 6; ra = (__u8 *)&raddr6->sin6_addr; @@ -2163,8 +2232,8 @@ static int c4iw_reconnect(struct c4iw_ep *ep) goto fail4; } - pr_debug("%s txq_idx %u tx_chan %u smac_idx %u rss_qid %u l2t_idx %u\n", - __func__, ep->txq_idx, ep->tx_chan, ep->smac_idx, ep->rss_qid, + pr_debug("txq_idx %u tx_chan %u smac_idx %u rss_qid %u l2t_idx %u\n", + ep->txq_idx, ep->tx_chan, ep->smac_idx, ep->rss_qid, ep->l2t->idx); state_set(&ep->com, CONNECTING); @@ -2179,7 +2248,8 @@ static int c4iw_reconnect(struct c4iw_ep *ep) fail4: dst_release(ep->dst); fail3: - remove_handle(ep->com.dev, &ep->com.dev->atid_idr, ep->atid); + xa_erase_irq(&ep->com.dev->atids, ep->atid); +fail2a: cxgb4_free_atid(ep->com.dev->rdev.lldi.tids, ep->atid); fail2: /* @@ -2210,17 +2280,20 @@ static int act_open_rpl(struct c4iw_dev *dev, struct sk_buff *skb) int ret = 0; ep = lookup_atid(t, atid); + if (!ep) + return -EINVAL; + la = (struct sockaddr_in *)&ep->com.local_addr; ra = (struct sockaddr_in *)&ep->com.remote_addr; la6 = (struct sockaddr_in6 *)&ep->com.local_addr; ra6 = (struct sockaddr_in6 *)&ep->com.remote_addr; - pr_debug("%s ep %p atid %u status %u errno %d\n", __func__, ep, atid, + pr_debug("ep %p atid %u status %u errno %d\n", ep, atid, status, status2errno(status)); if (cxgb_is_neg_adv(status)) { - pr_debug("%s Connection problems for atid %u status %u (%s)\n", - __func__, atid, status, neg_adv_str(status)); + pr_debug("Connection problems for atid %u status %u (%s)\n", + atid, status, neg_adv_str(status)); ep->stats.connect_neg_adv++; mutex_lock(&dev->rdev.stats.lock); dev->rdev.stats.neg_adv++; @@ -2262,8 +2335,7 @@ static int act_open_rpl(struct c4iw_dev *dev, struct sk_buff *skb) (const u32 *) &sin6->sin6_addr.s6_addr, 1); } - remove_handle(ep->com.dev, &ep->com.dev->atid_idr, - atid); + xa_erase_irq(&ep->com.dev->atids, atid); cxgb4_free_atid(t, atid); dst_release(ep->dst); cxgb4_l2t_release(ep->l2t); @@ -2300,7 +2372,7 @@ fail: cxgb4_remove_tid(ep->com.dev->rdev.lldi.tids, 0, GET_TID(rpl), ep->com.local_addr.ss_family); - remove_handle(ep->com.dev, &ep->com.dev->atid_idr, atid); + xa_erase_irq(&ep->com.dev->atids, atid); cxgb4_free_atid(t, atid); dst_release(ep->dst); cxgb4_l2t_release(ep->l2t); @@ -2316,12 +2388,12 @@ static int pass_open_rpl(struct c4iw_dev *dev, struct sk_buff *skb) struct c4iw_listen_ep *ep = get_ep_from_stid(dev, stid); if (!ep) { - pr_debug("%s stid %d lookup failure!\n", __func__, stid); + pr_warn("%s stid %d lookup failure!\n", __func__, stid); goto out; } - pr_debug("%s ep %p status %d error %d\n", __func__, ep, + pr_debug("ep %p status %d error %d\n", ep, rpl->status, status2errno(rpl->status)); - c4iw_wake_up(&ep->com.wr_wait, status2errno(rpl->status)); + c4iw_wake_up_noref(ep->com.wr_waitp, status2errno(rpl->status)); c4iw_put_ep(&ep->com); out: return 0; @@ -2333,9 +2405,14 @@ static int close_listsrv_rpl(struct c4iw_dev *dev, struct sk_buff *skb) unsigned int stid = GET_TID(rpl); struct c4iw_listen_ep *ep = get_ep_from_stid(dev, stid); - pr_debug("%s ep %p\n", __func__, ep); - c4iw_wake_up(&ep->com.wr_wait, status2errno(rpl->status)); + if (!ep) { + pr_warn("%s stid %d lookup failure!\n", __func__, stid); + goto out; + } + pr_debug("ep %p\n", ep); + c4iw_wake_up_noref(ep->com.wr_waitp, status2errno(rpl->status)); c4iw_put_ep(&ep->com); +out: return 0; } @@ -2351,22 +2428,7 @@ static int accept_cr(struct c4iw_ep *ep, struct sk_buff *skb, int win; enum chip_type adapter_type = ep->com.dev->rdev.lldi.adapter_type; - pr_debug("%s ep %p tid %u\n", __func__, ep, ep->hwtid); - BUG_ON(skb_cloned(skb)); - - skb_get(skb); - rpl = cplhdr(skb); - if (!is_t4(adapter_type)) { - skb_trim(skb, roundup(sizeof(*rpl5), 16)); - rpl5 = (void *)rpl; - INIT_TP_WR(rpl5, ep->hwtid); - } else { - skb_trim(skb, sizeof(*rpl)); - INIT_TP_WR(rpl, ep->hwtid); - } - OPCODE_TID(rpl) = cpu_to_be32(MK_OPCODE_TID(CPL_PASS_ACCEPT_RPL, - ep->hwtid)); - + pr_debug("ep %p tid %u\n", ep, ep->hwtid); cxgb_best_mtu(ep->com.dev->rdev.lldi.mtus, ep->mtu, &mtu_idx, enable_tcp_timestamps && req->tcpopt.tstamp, (ep->com.remote_addr.ss_family == AF_INET) ? 0 : 1); @@ -2412,17 +2474,25 @@ static int accept_cr(struct c4iw_ep *ep, struct sk_buff *skb, if (tcph->ece && tcph->cwr) opt2 |= CCTRL_ECN_V(1); } - if (CHELSIO_CHIP_VERSION(adapter_type) > CHELSIO_T4) { - u32 isn = (prandom_u32() & ~7UL) - 1; + + if (!is_t4(adapter_type)) { + u32 isn = (get_random_u32() & ~7UL) - 1; + + skb = get_skb(skb, roundup(sizeof(*rpl5), 16), GFP_KERNEL); + rpl5 = __skb_put_zero(skb, roundup(sizeof(*rpl5), 16)); + rpl = (void *)rpl5; + INIT_TP_WR_CPL(rpl5, CPL_PASS_ACCEPT_RPL, ep->hwtid); opt2 |= T5_OPT_2_VALID_F; opt2 |= CONG_CNTRL_V(CONG_ALG_TAHOE); opt2 |= T5_ISS_F; - rpl5 = (void *)rpl; - memset(&rpl5->iss, 0, roundup(sizeof(*rpl5)-sizeof(*rpl), 16)); if (peer2peer) isn += 4; rpl5->iss = cpu_to_be32(isn); - pr_debug("%s iss %u\n", __func__, be32_to_cpu(rpl5->iss)); + pr_debug("iss %u\n", be32_to_cpu(rpl5->iss)); + } else { + skb = get_skb(skb, sizeof(*rpl), GFP_KERNEL); + rpl = __skb_put_zero(skb, sizeof(*rpl)); + INIT_TP_WR_CPL(rpl, CPL_PASS_ACCEPT_RPL, ep->hwtid); } rpl->opt0 = cpu_to_be64(opt0); @@ -2435,8 +2505,7 @@ static int accept_cr(struct c4iw_ep *ep, struct sk_buff *skb, static void reject_cr(struct c4iw_dev *dev, u32 hwtid, struct sk_buff *skb) { - pr_debug("%s c4iw_dev %p tid %u\n", __func__, dev, hwtid); - BUG_ON(skb_cloned(skb)); + pr_debug("c4iw_dev %p tid %u\n", dev, hwtid); skb_trim(skb, sizeof(struct cpl_tid_release)); release_tid(&dev->rdev, hwtid, skb); return; @@ -2457,40 +2526,45 @@ static int pass_accept_req(struct c4iw_dev *dev, struct sk_buff *skb) u16 peer_mss = ntohs(req->tcpopt.mss); int iptype; unsigned short hdrs; - u8 tos = PASS_OPEN_TOS_G(ntohl(req->tos_stid)); + u8 tos; parent_ep = (struct c4iw_ep *)get_ep_from_stid(dev, stid); if (!parent_ep) { - pr_debug("%s connect request on invalid stid %d\n", - __func__, stid); + pr_err("%s connect request on invalid stid %d\n", + __func__, stid); goto reject; } if (state_read(&parent_ep->com) != LISTEN) { - pr_debug("%s - listening ep not in LISTEN\n", __func__); + pr_err("%s - listening ep not in LISTEN\n", __func__); goto reject; } + if (parent_ep->com.cm_id->tos_set) + tos = parent_ep->com.cm_id->tos; + else + tos = PASS_OPEN_TOS_G(ntohl(req->tos_stid)); + cxgb_get_4tuple(req, parent_ep->com.dev->rdev.lldi.adapter_type, &iptype, local_ip, peer_ip, &local_port, &peer_port); /* Find output route */ if (iptype == 4) { - pr_debug("%s parent ep %p hwtid %u laddr %pI4 raddr %pI4 lport %d rport %d peer_mss %d\n" - , __func__, parent_ep, hwtid, + pr_debug("parent ep %p hwtid %u laddr %pI4 raddr %pI4 lport %d rport %d peer_mss %d\n" + , parent_ep, hwtid, local_ip, peer_ip, ntohs(local_port), ntohs(peer_port), peer_mss); dst = cxgb_find_route(&dev->rdev.lldi, get_real_dev, *(__be32 *)local_ip, *(__be32 *)peer_ip, local_port, peer_port, tos); } else { - pr_debug("%s parent ep %p hwtid %u laddr %pI6 raddr %pI6 lport %d rport %d peer_mss %d\n" - , __func__, parent_ep, hwtid, + pr_debug("parent ep %p hwtid %u laddr %pI6 raddr %pI6 lport %d rport %d peer_mss %d\n" + , parent_ep, hwtid, local_ip, peer_ip, ntohs(local_port), ntohs(peer_port), peer_mss); dst = cxgb_find_route6(&dev->rdev.lldi, get_real_dev, local_ip, peer_ip, local_port, peer_port, - PASS_OPEN_TOS_G(ntohl(req->tos_stid)), + tos, ((struct sockaddr_in6 *) &parent_ep->com.local_addr)->sin6_scope_id); } @@ -2571,10 +2645,10 @@ static int pass_accept_req(struct c4iw_dev *dev, struct sk_buff *skb) child_ep->dst = dst; child_ep->hwtid = hwtid; - pr_debug("%s tx_chan %u smac_idx %u rss_qid %u\n", __func__, + pr_debug("tx_chan %u smac_idx %u rss_qid %u\n", child_ep->tx_chan, child_ep->smac_idx, child_ep->rss_qid); - init_timer(&child_ep->timer); + timer_setup(&child_ep->timer, ep_timeout, 0); cxgb4_insert_tid(t, child_ep, hwtid, child_ep->com.local_addr.ss_family); insert_ep_tid(child_ep); @@ -2594,9 +2668,9 @@ fail: c4iw_put_ep(&child_ep->com); reject: reject_cr(dev, hwtid, skb); +out: if (parent_ep) c4iw_put_ep(&parent_ep->com); -out: return 0; } @@ -2606,16 +2680,20 @@ static int pass_establish(struct c4iw_dev *dev, struct sk_buff *skb) struct cpl_pass_establish *req = cplhdr(skb); unsigned int tid = GET_TID(req); int ret; + u16 tcp_opt = ntohs(req->tcp_opt); ep = get_ep_from_tid(dev, tid); - pr_debug("%s ep %p tid %u\n", __func__, ep, ep->hwtid); + if (!ep) + return 0; + + pr_debug("ep %p tid %u\n", ep, ep->hwtid); ep->snd_seq = be32_to_cpu(req->snd_isn); ep->rcv_seq = be32_to_cpu(req->rcv_isn); + ep->snd_wscale = TCPOPT_SND_WSCALE_G(tcp_opt); - pr_debug("%s ep %p hwtid %u tcp_opt 0x%02x\n", __func__, ep, tid, - ntohs(req->tcp_opt)); + pr_debug("ep %p hwtid %u tcp_opt 0x%02x\n", ep, tid, tcp_opt); - set_emss(ep, ntohs(req->tcp_opt)); + set_emss(ep, tcp_opt); dst_confirm(ep->dst); mutex_lock(&ep->com.mutex); @@ -2645,7 +2723,7 @@ static int peer_close(struct c4iw_dev *dev, struct sk_buff *skb) if (!ep) return 0; - pr_debug("%s ep %p tid %u\n", __func__, ep, ep->hwtid); + pr_debug("ep %p tid %u\n", ep, ep->hwtid); dst_confirm(ep->dst); set_bit(PEER_CLOSE, &ep->com.history); @@ -2668,12 +2746,12 @@ static int peer_close(struct c4iw_dev *dev, struct sk_buff *skb) */ __state_set(&ep->com, CLOSING); pr_debug("waking up ep %p tid %u\n", ep, ep->hwtid); - c4iw_wake_up(&ep->com.wr_wait, -ECONNRESET); + c4iw_wake_up_noref(ep->com.wr_waitp, -ECONNRESET); break; case MPA_REP_SENT: __state_set(&ep->com, CLOSING); pr_debug("waking up ep %p tid %u\n", ep, ep->hwtid); - c4iw_wake_up(&ep->com.wr_wait, -ECONNRESET); + c4iw_wake_up_noref(ep->com.wr_waitp, -ECONNRESET); break; case FPDU_MODE: start_ep_timer(ep); @@ -2709,7 +2787,7 @@ static int peer_close(struct c4iw_dev *dev, struct sk_buff *skb) disconnect = 0; break; default: - BUG_ON(1); + WARN_ONCE(1, "Bad endpoint state %u\n", ep->com.state); } mutex_unlock(&ep->com.mutex); if (disconnect) @@ -2720,32 +2798,52 @@ static int peer_close(struct c4iw_dev *dev, struct sk_buff *skb) return 0; } +static void finish_peer_abort(struct c4iw_dev *dev, struct c4iw_ep *ep) +{ + complete_cached_srq_buffers(ep, ep->srqe_idx); + if (ep->com.cm_id && ep->com.qp) { + struct c4iw_qp_attributes attrs; + + attrs.next_state = C4IW_QP_STATE_ERROR; + c4iw_modify_qp(ep->com.qp->rhp, ep->com.qp, + C4IW_QP_ATTR_NEXT_STATE, &attrs, 1); + } + peer_abort_upcall(ep); + release_ep_resources(ep); + c4iw_put_ep(&ep->com); +} + static int peer_abort(struct c4iw_dev *dev, struct sk_buff *skb) { - struct cpl_abort_req_rss *req = cplhdr(skb); + struct cpl_abort_req_rss6 *req = cplhdr(skb); struct c4iw_ep *ep; struct sk_buff *rpl_skb; struct c4iw_qp_attributes attrs; int ret; int release = 0; unsigned int tid = GET_TID(req); + u8 status; + u32 srqidx; + u32 len = roundup(sizeof(struct cpl_abort_rpl), 16); ep = get_ep_from_tid(dev, tid); if (!ep) return 0; - if (cxgb_is_neg_adv(req->status)) { - pr_debug("%s Negative advice on abort- tid %u status %d (%s)\n", - __func__, ep->hwtid, req->status, - neg_adv_str(req->status)); + status = ABORT_RSS_STATUS_G(be32_to_cpu(req->srqidx_status)); + + if (cxgb_is_neg_adv(status)) { + pr_debug("Negative advice on abort- tid %u status %d (%s)\n", + ep->hwtid, status, neg_adv_str(status)); ep->stats.abort_neg_adv++; mutex_lock(&dev->rdev.stats.lock); dev->rdev.stats.neg_adv++; mutex_unlock(&dev->rdev.stats.lock); goto deref_ep; } - pr_debug("%s ep %p tid %u state %u\n", __func__, ep, ep->hwtid, + + pr_debug("ep %p tid %u state %u\n", ep, ep->hwtid, ep->com.state); set_bit(PEER_ABORT, &ep->com.history); @@ -2755,7 +2853,7 @@ static int peer_abort(struct c4iw_dev *dev, struct sk_buff *skb) * MPA_REQ_SENT */ if (ep->com.state != MPA_REQ_SENT) - c4iw_wake_up(&ep->com.wr_wait, -ECONNRESET); + c4iw_wake_up_noref(ep->com.wr_waitp, -ECONNRESET); mutex_lock(&ep->com.mutex); switch (ep->com.state) { @@ -2767,7 +2865,8 @@ static int peer_abort(struct c4iw_dev *dev, struct sk_buff *skb) break; case MPA_REQ_SENT: (void)stop_ep_timer(ep); - if (mpa_rev == 1 || (mpa_rev == 2 && ep->tried_with_mpa_v1)) + if (status != CPL_ERR_CONN_RESET || mpa_rev == 1 || + (mpa_rev == 2 && ep->tried_with_mpa_v1)) connect_reply_upcall(ep, -ECONNRESET); else { /* @@ -2778,8 +2877,8 @@ static int peer_abort(struct c4iw_dev *dev, struct sk_buff *skb) * do some housekeeping so as to re-initiate the * connection */ - pr_debug("%s: mpa_rev=%d. Retrying with mpav1\n", - __func__, mpa_rev); + pr_info("%s: mpa_rev=%d. Retrying with mpav1\n", + __func__, mpa_rev); ep->retry_with_mpa_v1 = 1; } break; @@ -2790,8 +2889,24 @@ static int peer_abort(struct c4iw_dev *dev, struct sk_buff *skb) case MORIBUND: case CLOSING: stop_ep_timer(ep); - /*FALLTHROUGH*/ + fallthrough; case FPDU_MODE: + if (ep->com.qp && ep->com.qp->srq) { + srqidx = ABORT_RSS_SRQIDX_G( + be32_to_cpu(req->srqidx_status)); + if (srqidx) { + complete_cached_srq_buffers(ep, srqidx); + } else { + /* Hold ep ref until finish_peer_abort() */ + c4iw_get_ep(&ep->com); + __state_set(&ep->com, ABORTING); + set_bit(PEER_ABORT_IN_PROGRESS, &ep->com.flags); + read_tcb(ep); + break; + + } + } + if (ep->com.cm_id && ep->com.qp) { attrs.next_state = C4IW_QP_STATE_ERROR; ret = c4iw_modify_qp(ep->com.qp->rhp, @@ -2805,11 +2920,11 @@ static int peer_abort(struct c4iw_dev *dev, struct sk_buff *skb) case ABORTING: break; case DEAD: - pr_debug("%s PEER_ABORT IN DEAD STATE!!!!\n", __func__); + pr_warn("%s PEER_ABORT IN DEAD STATE!!!!\n", __func__); mutex_unlock(&ep->com.mutex); goto deref_ep; default: - BUG_ON(1); + WARN_ONCE(1, "Bad endpoint state %u\n", ep->com.state); break; } dst_confirm(ep->dst); @@ -2843,7 +2958,7 @@ out: (const u32 *)&sin6->sin6_addr.s6_addr, 1); } - remove_handle(ep->com.dev, &ep->com.dev->hwtid_idr, ep->hwtid); + xa_erase_irq(&ep->com.dev->hwtids, ep->hwtid); cxgb4_remove_tid(ep->com.dev->rdev.lldi.tids, 0, ep->hwtid, ep->com.local_addr.ss_family); dst_release(ep->dst); @@ -2870,8 +2985,7 @@ static int close_con_rpl(struct c4iw_dev *dev, struct sk_buff *skb) if (!ep) return 0; - pr_debug("%s ep %p tid %u\n", __func__, ep, ep->hwtid); - BUG_ON(!ep); + pr_debug("ep %p tid %u\n", ep, ep->hwtid); /* The cm_id may be null if we failed to connect */ mutex_lock(&ep->com.mutex); @@ -2897,7 +3011,7 @@ static int close_con_rpl(struct c4iw_dev *dev, struct sk_buff *skb) case DEAD: break; default: - BUG_ON(1); + WARN_ONCE(1, "Bad endpoint state %u\n", ep->com.state); break; } mutex_unlock(&ep->com.mutex); @@ -2915,17 +3029,23 @@ static int terminate(struct c4iw_dev *dev, struct sk_buff *skb) struct c4iw_qp_attributes attrs; ep = get_ep_from_tid(dev, tid); - BUG_ON(!ep); - if (ep && ep->com.qp) { - pr_warn("TERM received tid %u qpid %u\n", - tid, ep->com.qp->wq.sq.qid); - attrs.next_state = C4IW_QP_STATE_TERMINATE; - c4iw_modify_qp(ep->com.qp->rhp, ep->com.qp, - C4IW_QP_ATTR_NEXT_STATE, &attrs, 1); + if (ep) { + if (ep->com.qp) { + pr_warn("TERM received tid %u qpid %u\n", tid, + ep->com.qp->wq.sq.qid); + attrs.next_state = C4IW_QP_STATE_TERMINATE; + c4iw_modify_qp(ep->com.qp->rhp, ep->com.qp, + C4IW_QP_ATTR_NEXT_STATE, &attrs, 1); + } + + /* As per draft-hilland-iwarp-verbs-v1.0, sec 6.2.3, + * when entering the TERM state the RNIC MUST initiate a CLOSE. + */ + c4iw_ep_disconnect(ep, 1, GFP_KERNEL); + c4iw_put_ep(&ep->com); } else pr_warn("TERM received tid %u no ep/qp\n", tid); - c4iw_put_ep(&ep->com); return 0; } @@ -2946,19 +3066,19 @@ static int fw4_ack(struct c4iw_dev *dev, struct sk_buff *skb) ep = get_ep_from_tid(dev, tid); if (!ep) return 0; - pr_debug("%s ep %p tid %u credits %u\n", - __func__, ep, ep->hwtid, credits); + pr_debug("ep %p tid %u credits %u\n", + ep, ep->hwtid, credits); if (credits == 0) { - pr_debug("%s 0 credit ack ep %p tid %u state %u\n", - __func__, ep, ep->hwtid, state_read(&ep->com)); + pr_debug("0 credit ack ep %p tid %u state %u\n", + ep, ep->hwtid, state_read(&ep->com)); goto out; } dst_confirm(ep->dst); if (ep->mpa_skb) { - pr_debug("%s last streaming msg ack ep %p tid %u state %u initiator %u freeing skb\n", - __func__, ep, ep->hwtid, - state_read(&ep->com), ep->mpa_attr.initiator ? 1 : 0); + pr_debug("last streaming msg ack ep %p tid %u state %u initiator %u freeing skb\n", + ep, ep->hwtid, state_read(&ep->com), + ep->mpa_attr.initiator ? 1 : 0); mutex_lock(&ep->com.mutex); kfree_skb(ep->mpa_skb); ep->mpa_skb = NULL; @@ -2976,7 +3096,7 @@ int c4iw_reject_cr(struct iw_cm_id *cm_id, const void *pdata, u8 pdata_len) int abort; struct c4iw_ep *ep = to_ep(cm_id); - pr_debug("%s ep %p tid %u\n", __func__, ep, ep->hwtid); + pr_debug("ep %p tid %u\n", ep, ep->hwtid); mutex_lock(&ep->com.mutex); if (ep->com.state != MPA_REQ_RCVD) { @@ -3007,7 +3127,7 @@ int c4iw_accept_cr(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) struct c4iw_qp *qp = get_qhp(h, conn_param->qpn); int abort = 0; - pr_debug("%s ep %p tid %u\n", __func__, ep, ep->hwtid); + pr_debug("ep %p tid %u\n", ep, ep->hwtid); mutex_lock(&ep->com.mutex); if (ep->com.state != MPA_REQ_RCVD) { @@ -3015,7 +3135,10 @@ int c4iw_accept_cr(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) goto err_out; } - BUG_ON(!qp); + if (!qp) { + err = -EINVAL; + goto err_out; + } set_bit(ULP_ACCEPT, &ep->com.history); if ((conn_param->ord > cur_max_read_depth(ep->com.dev)) || @@ -3060,7 +3183,7 @@ int c4iw_accept_cr(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) ep->ird = 1; } - pr_debug("%s %d ird %d ord %d\n", __func__, __LINE__, ep->ird, ep->ord); + pr_debug("ird %d ord %d\n", ep->ird, ep->ord); ep->com.cm_id = cm_id; ref_cm_id(&ep->com); @@ -3115,17 +3238,22 @@ static int pick_local_ipaddrs(struct c4iw_dev *dev, struct iw_cm_id *cm_id) int found = 0; struct sockaddr_in *laddr = (struct sockaddr_in *)&cm_id->m_local_addr; struct sockaddr_in *raddr = (struct sockaddr_in *)&cm_id->m_remote_addr; + const struct in_ifaddr *ifa; ind = in_dev_get(dev->rdev.lldi.ports[0]); if (!ind) return -EADDRNOTAVAIL; - for_primary_ifa(ind) { + rcu_read_lock(); + in_dev_for_each_ifa_rcu(ifa, ind) { + if (ifa->ifa_flags & IFA_F_SECONDARY) + continue; laddr->sin_addr.s_addr = ifa->ifa_address; raddr->sin_addr.s_addr = ifa->ifa_address; found = 1; break; } - endfor_ifa(ind); + rcu_read_unlock(); + in_dev_put(ind); return found ? 0 : -EADDRNOTAVAIL; } @@ -3158,7 +3286,7 @@ static int get_lladdr(struct net_device *dev, struct in6_addr *addr, static int pick_local_ip6addrs(struct c4iw_dev *dev, struct iw_cm_id *cm_id) { - struct in6_addr uninitialized_var(addr); + struct in6_addr addr; struct sockaddr_in6 *la6 = (struct sockaddr_in6 *)&cm_id->m_local_addr; struct sockaddr_in6 *ra6 = (struct sockaddr_in6 *)&cm_id->m_remote_addr; @@ -3200,7 +3328,7 @@ int c4iw_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) goto fail1; } - init_timer(&ep->timer); + timer_setup(&ep->timer, ep_timeout, 0); ep->plen = conn_param->private_data_len; if (ep->plen) memcpy(ep->mpa_pkt + sizeof(struct mpa_message), @@ -3213,15 +3341,16 @@ int c4iw_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) ep->com.cm_id = cm_id; ref_cm_id(&ep->com); + cm_id->provider_data = ep; ep->com.dev = dev; ep->com.qp = get_qhp(dev, conn_param->qpn); if (!ep->com.qp) { - pr_debug("%s qpn 0x%x not found!\n", __func__, conn_param->qpn); + pr_warn("%s qpn 0x%x not found!\n", __func__, conn_param->qpn); err = -EINVAL; goto fail2; } ref_qp(ep); - pr_debug("%s qpn 0x%x qp %p cm_id %p\n", __func__, conn_param->qpn, + pr_debug("qpn 0x%x qp %p cm_id %p\n", conn_param->qpn, ep->com.qp, cm_id); /* @@ -3233,7 +3362,9 @@ int c4iw_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) err = -ENOMEM; goto fail2; } - insert_handle(dev, &dev->atid_idr, ep, ep->atid); + err = xa_insert_irq(&dev->atids, ep->atid, ep, GFP_KERNEL); + if (err) + goto fail5; memcpy(&ep->com.local_addr, &cm_id->m_local_addr, sizeof(ep->com.local_addr)); @@ -3255,12 +3386,12 @@ int c4iw_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) if (raddr->sin_addr.s_addr == htonl(INADDR_ANY)) { err = pick_local_ipaddrs(dev, cm_id); if (err) - goto fail2; + goto fail3; } /* find a route */ - pr_debug("%s saddr %pI4 sport 0x%x raddr %pI4 rport 0x%x\n", - __func__, &laddr->sin_addr, ntohs(laddr->sin_port), + pr_debug("saddr %pI4 sport 0x%x raddr %pI4 rport 0x%x\n", + &laddr->sin_addr, ntohs(laddr->sin_port), ra, ntohs(raddr->sin_port)); ep->dst = cxgb_find_route(&dev->rdev.lldi, get_real_dev, laddr->sin_addr.s_addr, @@ -3277,19 +3408,19 @@ int c4iw_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) if (ipv6_addr_type(&raddr6->sin6_addr) == IPV6_ADDR_ANY) { err = pick_local_ip6addrs(dev, cm_id); if (err) - goto fail2; + goto fail3; } /* find a route */ - pr_debug("%s saddr %pI6 sport 0x%x raddr %pI6 rport 0x%x\n", - __func__, laddr6->sin6_addr.s6_addr, + pr_debug("saddr %pI6 sport 0x%x raddr %pI6 rport 0x%x\n", + laddr6->sin6_addr.s6_addr, ntohs(laddr6->sin6_port), raddr6->sin6_addr.s6_addr, ntohs(raddr6->sin6_port)); ep->dst = cxgb_find_route6(&dev->rdev.lldi, get_real_dev, laddr6->sin6_addr.s6_addr, raddr6->sin6_addr.s6_addr, laddr6->sin6_port, - raddr6->sin6_port, 0, + raddr6->sin6_port, cm_id->tos, raddr6->sin6_scope_id); } if (!ep->dst) { @@ -3305,8 +3436,8 @@ int c4iw_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) goto fail4; } - pr_debug("%s txq_idx %u tx_chan %u smac_idx %u rss_qid %u l2t_idx %u\n", - __func__, ep->txq_idx, ep->tx_chan, ep->smac_idx, ep->rss_qid, + pr_debug("txq_idx %u tx_chan %u smac_idx %u rss_qid %u l2t_idx %u\n", + ep->txq_idx, ep->tx_chan, ep->smac_idx, ep->rss_qid, ep->l2t->idx); state_set(&ep->com, CONNECTING); @@ -3321,7 +3452,8 @@ int c4iw_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) fail4: dst_release(ep->dst); fail3: - remove_handle(ep->com.dev, &ep->com.dev->atid_idr, ep->atid); + xa_erase_irq(&ep->com.dev->atids, ep->atid); +fail5: cxgb4_free_atid(ep->com.dev->rdev.lldi.tids, ep->atid); fail2: skb_queue_purge(&ep->com.ep_skb_list); @@ -3344,14 +3476,14 @@ static int create_server6(struct c4iw_dev *dev, struct c4iw_listen_ep *ep) if (err) return err; } - c4iw_init_wr_wait(&ep->com.wr_wait); + c4iw_init_wr_wait(ep->com.wr_waitp); err = cxgb4_create_server6(ep->com.dev->rdev.lldi.ports[0], ep->stid, &sin6->sin6_addr, sin6->sin6_port, ep->com.dev->rdev.lldi.rxq_ids[0]); if (!err) err = c4iw_wait_for_reply(&ep->com.dev->rdev, - &ep->com.wr_wait, + ep->com.wr_waitp, 0, 0, __func__); else if (err > 0) err = net_xmit_errno(err); @@ -3387,13 +3519,13 @@ static int create_server4(struct c4iw_dev *dev, struct c4iw_listen_ep *ep) } } while (err == -EBUSY); } else { - c4iw_init_wr_wait(&ep->com.wr_wait); + c4iw_init_wr_wait(ep->com.wr_waitp); err = cxgb4_create_server(ep->com.dev->rdev.lldi.ports[0], ep->stid, sin->sin_addr.s_addr, sin->sin_port, 0, ep->com.dev->rdev.lldi.rxq_ids[0]); if (!err) err = c4iw_wait_for_reply(&ep->com.dev->rdev, - &ep->com.wr_wait, + ep->com.wr_waitp, 0, 0, __func__); else if (err > 0) err = net_xmit_errno(err); @@ -3420,7 +3552,7 @@ int c4iw_create_listen(struct iw_cm_id *cm_id, int backlog) goto fail1; } skb_queue_head_init(&ep->com.ep_skb_list); - pr_debug("%s ep %p\n", __func__, ep); + pr_debug("ep %p\n", ep); ep->com.cm_id = cm_id; ref_cm_id(&ep->com); ep->com.dev = dev; @@ -3444,10 +3576,9 @@ int c4iw_create_listen(struct iw_cm_id *cm_id, int backlog) err = -ENOMEM; goto fail2; } - insert_handle(dev, &dev->stid_idr, ep, ep->stid); - - memcpy(&ep->com.local_addr, &cm_id->m_local_addr, - sizeof(ep->com.local_addr)); + err = xa_insert_irq(&dev->stids, ep->stid, ep, GFP_KERNEL); + if (err) + goto fail3; state_set(&ep->com, LISTEN); if (ep->com.local_addr.ss_family == AF_INET) @@ -3458,7 +3589,8 @@ int c4iw_create_listen(struct iw_cm_id *cm_id, int backlog) cm_id->provider_data = ep; goto out; } - + xa_erase_irq(&ep->com.dev->stids, ep->stid); +fail3: cxgb4_free_stid(ep->com.dev->rdev.lldi.tids, ep->stid, ep->com.local_addr.ss_family); fail2: @@ -3474,7 +3606,7 @@ int c4iw_destroy_listen(struct iw_cm_id *cm_id) int err; struct c4iw_listen_ep *ep = to_listen_ep(cm_id); - pr_debug("%s ep %p\n", __func__, ep); + pr_debug("ep %p\n", ep); might_sleep(); state_set(&ep->com, DEAD); @@ -3482,22 +3614,23 @@ int c4iw_destroy_listen(struct iw_cm_id *cm_id) ep->com.local_addr.ss_family == AF_INET) { err = cxgb4_remove_server_filter( ep->com.dev->rdev.lldi.ports[0], ep->stid, - ep->com.dev->rdev.lldi.rxq_ids[0], 0); + ep->com.dev->rdev.lldi.rxq_ids[0], false); } else { struct sockaddr_in6 *sin6; - c4iw_init_wr_wait(&ep->com.wr_wait); + c4iw_init_wr_wait(ep->com.wr_waitp); err = cxgb4_remove_server( ep->com.dev->rdev.lldi.ports[0], ep->stid, - ep->com.dev->rdev.lldi.rxq_ids[0], 0); + ep->com.dev->rdev.lldi.rxq_ids[0], + ep->com.local_addr.ss_family == AF_INET6); if (err) goto done; - err = c4iw_wait_for_reply(&ep->com.dev->rdev, &ep->com.wr_wait, + err = c4iw_wait_for_reply(&ep->com.dev->rdev, ep->com.wr_waitp, 0, 0, __func__); sin6 = (struct sockaddr_in6 *)&ep->com.local_addr; cxgb4_clip_release(ep->com.dev->rdev.lldi.ports[0], (const u32 *)&sin6->sin6_addr.s6_addr, 1); } - remove_handle(ep->com.dev, &ep->com.dev->stid_idr, ep->stid); + xa_erase_irq(&ep->com.dev->stids, ep->stid); cxgb4_free_stid(ep->com.dev->rdev.lldi.tids, ep->stid, ep->com.local_addr.ss_family); done: @@ -3515,7 +3648,7 @@ int c4iw_ep_disconnect(struct c4iw_ep *ep, int abrupt, gfp_t gfp) mutex_lock(&ep->com.mutex); - pr_debug("%s ep %p state %s, abrupt %d\n", __func__, ep, + pr_debug("ep %p state %s, abrupt %d\n", ep, states[ep->com.state], abrupt); /* @@ -3569,18 +3702,17 @@ int c4iw_ep_disconnect(struct c4iw_ep *ep, int abrupt, gfp_t gfp) case MORIBUND: case ABORTING: case DEAD: - pr_debug("%s ignoring disconnect ep %p state %u\n", - __func__, ep, ep->com.state); + pr_debug("ignoring disconnect ep %p state %u\n", + ep, ep->com.state); break; default: - BUG(); + WARN_ONCE(1, "Bad endpoint state %u\n", ep->com.state); break; } if (close) { if (abrupt) { set_bit(EP_DISC_ABORT, &ep->com.history); - close_complete_upcall(ep, -ECONNRESET); ret = send_abort(ep); } else { set_bit(EP_DISC_CLOSE, &ep->com.history); @@ -3632,6 +3764,7 @@ static void active_ofld_conn_reply(struct c4iw_dev *dev, struct sk_buff *skb, send_fw_act_open_req(ep, atid); return; } + fallthrough; case FW_EADDRINUSE: set_bit(ACT_RETRY_INUSE, &ep->com.history); if (ep->retry_count++ < ACT_OPEN_RETRY_COUNT) { @@ -3657,7 +3790,7 @@ static void active_ofld_conn_reply(struct c4iw_dev *dev, struct sk_buff *skb, cxgb4_clip_release(ep->com.dev->rdev.lldi.ports[0], (const u32 *)&sin6->sin6_addr.s6_addr, 1); } - remove_handle(dev, &dev->atid_idr, atid); + xa_erase_irq(&dev->atids, atid); cxgb4_free_atid(dev->rdev.lldi.tids, atid); dst_release(ep->dst); cxgb4_l2t_release(ep->l2t); @@ -3672,9 +3805,8 @@ static void passive_ofld_conn_reply(struct c4iw_dev *dev, struct sk_buff *skb, int ret; rpl_skb = (struct sk_buff *)(unsigned long)req->cookie; - BUG_ON(!rpl_skb); if (req->retval) { - pr_debug("%s passive open failure %d\n", __func__, req->retval); + pr_err("%s passive open failure %d\n", __func__, req->retval); mutex_lock(&dev->rdev.stats.lock); dev->rdev.stats.pas_ofld_conn_fails++; mutex_unlock(&dev->rdev.stats.lock); @@ -3691,6 +3823,80 @@ static void passive_ofld_conn_reply(struct c4iw_dev *dev, struct sk_buff *skb, return; } +static inline u64 t4_tcb_get_field64(__be64 *tcb, u16 word) +{ + u64 tlo = be64_to_cpu(tcb[((31 - word) / 2)]); + u64 thi = be64_to_cpu(tcb[((31 - word) / 2) - 1]); + u64 t; + u32 shift = 32; + + t = (thi << shift) | (tlo >> shift); + + return t; +} + +static inline u32 t4_tcb_get_field32(__be64 *tcb, u16 word, u32 mask, u32 shift) +{ + u32 v; + u64 t = be64_to_cpu(tcb[(31 - word) / 2]); + + if (word & 0x1) + shift += 32; + v = (t >> shift) & mask; + return v; +} + +static int read_tcb_rpl(struct c4iw_dev *dev, struct sk_buff *skb) +{ + struct cpl_get_tcb_rpl *rpl = cplhdr(skb); + __be64 *tcb = (__be64 *)(rpl + 1); + unsigned int tid = GET_TID(rpl); + struct c4iw_ep *ep; + u64 t_flags_64; + u32 rx_pdu_out; + + ep = get_ep_from_tid(dev, tid); + if (!ep) + return 0; + /* Examine the TF_RX_PDU_OUT (bit 49 of the t_flags) in order to + * determine if there's a rx PDU feedback event pending. + * + * If that bit is set, it means we'll need to re-read the TCB's + * rq_start value. The final value is the one present in a TCB + * with the TF_RX_PDU_OUT bit cleared. + */ + + t_flags_64 = t4_tcb_get_field64(tcb, TCB_T_FLAGS_W); + rx_pdu_out = (t_flags_64 & TF_RX_PDU_OUT_V(1)) >> TF_RX_PDU_OUT_S; + + c4iw_put_ep(&ep->com); /* from get_ep_from_tid() */ + c4iw_put_ep(&ep->com); /* from read_tcb() */ + + /* If TF_RX_PDU_OUT bit is set, re-read the TCB */ + if (rx_pdu_out) { + if (++ep->rx_pdu_out_cnt >= 2) { + WARN_ONCE(1, "tcb re-read() reached the guard limit, finishing the cleanup\n"); + goto cleanup; + } + read_tcb(ep); + return 0; + } + + ep->srqe_idx = t4_tcb_get_field32(tcb, TCB_RQ_START_W, TCB_RQ_START_M, + TCB_RQ_START_S); +cleanup: + pr_debug("ep %p tid %u %016x\n", ep, ep->hwtid, ep->srqe_idx); + + if (test_bit(PEER_ABORT_IN_PROGRESS, &ep->com.flags)) + finish_peer_abort(dev, ep); + else if (test_bit(ABORT_REQ_IN_PROGRESS, &ep->com.flags)) + send_abort_req(ep); + else + WARN_ONCE(1, "unexpected state!"); + + return 0; +} + static int deferred_fw6_msg(struct c4iw_dev *dev, struct sk_buff *skb) { struct cpl_fw6_msg *rpl = cplhdr(skb); @@ -3870,7 +4076,6 @@ static int rx_pkt(struct c4iw_dev *dev, struct sk_buff *skb) struct net_device *pdev; u16 rss_qid, eth_hdr_len; int step; - u32 tx_chan; struct neighbour *neigh; /* Drop all non-SYN packets */ @@ -3891,8 +4096,8 @@ static int rx_pkt(struct c4iw_dev *dev, struct sk_buff *skb) lep = (struct c4iw_ep *)get_ep_from_stid(dev, stid); if (!lep) { - pr_debug("%s connect request on invalid stid %d\n", - __func__, stid); + pr_warn("%s connect request on invalid stid %d\n", + __func__, stid); goto reject; } @@ -3918,7 +4123,7 @@ static int rx_pkt(struct c4iw_dev *dev, struct sk_buff *skb) } else { vlan_eh = (struct vlan_ethhdr *)(req + 1); iph = (struct iphdr *)(vlan_eh + 1); - skb->vlan_tci = ntohs(cpl->vlan); + __vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q), ntohs(cpl->vlan)); } if (iph->version != 0x4) @@ -3929,7 +4134,7 @@ static int rx_pkt(struct c4iw_dev *dev, struct sk_buff *skb) skb_set_transport_header(skb, (void *)tcph - (void *)rss); skb_get(skb); - pr_debug("%s lip 0x%x lport %u pip 0x%x pport %u tos %d\n", __func__, + pr_debug("lip 0x%x lport %u pip 0x%x pport %u tos %d\n", ntohl(iph->daddr), ntohs(tcph->dest), ntohl(iph->saddr), ntohs(tcph->source), iph->tos); @@ -3937,31 +4142,31 @@ static int rx_pkt(struct c4iw_dev *dev, struct sk_buff *skb) iph->daddr, iph->saddr, tcph->dest, tcph->source, iph->tos); if (!dst) { - pr_err("%s - failed to find dst entry!\n", - __func__); + pr_err("%s - failed to find dst entry!\n", __func__); goto reject; } neigh = dst_neigh_lookup_skb(dst, skb); if (!neigh) { - pr_err("%s - failed to allocate neigh!\n", - __func__); + pr_err("%s - failed to allocate neigh!\n", __func__); goto free_dst; } if (neigh->dev->flags & IFF_LOOPBACK) { pdev = ip_dev_find(&init_net, iph->daddr); + if (!pdev) { + pr_err("%s - failed to find device!\n", __func__); + goto free_dst; + } e = cxgb4_l2t_get(dev->rdev.lldi.l2t, neigh, pdev, 0); pi = (struct port_info *)netdev_priv(pdev); - tx_chan = cxgb4_port_chan(pdev); dev_put(pdev); } else { pdev = get_real_dev(neigh->dev); e = cxgb4_l2t_get(dev->rdev.lldi.l2t, neigh, pdev, 0); pi = (struct port_info *)netdev_priv(pdev); - tx_chan = cxgb4_port_chan(pdev); } neigh_release(neigh); if (!e) { @@ -4016,6 +4221,7 @@ static c4iw_handler_func work_handlers[NUM_CPL_CMDS + NUM_FAKE_CPLS] = { [CPL_CLOSE_CON_RPL] = close_con_rpl, [CPL_RDMA_TERMINATE] = terminate, [CPL_FW4_ACK] = fw4_ack, + [CPL_GET_TCB_RPL] = read_tcb_rpl, [CPL_FW6_MSG] = deferred_fw6_msg, [CPL_RX_PKT] = rx_pkt, [FAKE_CPL_PUT_EP_SAFE] = _put_ep_safe, @@ -4028,8 +4234,7 @@ static void process_timeout(struct c4iw_ep *ep) int abort = 1; mutex_lock(&ep->com.mutex); - pr_debug("%s ep %p tid %u state %d\n", __func__, ep, ep->hwtid, - ep->com.state); + pr_debug("ep %p tid %u state %d\n", ep, ep->hwtid, ep->com.state); set_bit(TIMEDOUT, &ep->com.history); switch (ep->com.state) { case MPA_REQ_SENT: @@ -4105,19 +4310,24 @@ static void process_work(struct work_struct *work) dev = *((struct c4iw_dev **) (skb->cb + sizeof(void *))); opcode = rpl->ot.opcode; - BUG_ON(!work_handlers[opcode]); - ret = work_handlers[opcode](dev, skb); - if (!ret) + if (opcode >= ARRAY_SIZE(work_handlers) || + !work_handlers[opcode]) { + pr_err("No handler for opcode 0x%x.\n", opcode); kfree_skb(skb); + } else { + ret = work_handlers[opcode](dev, skb); + if (!ret) + kfree_skb(skb); + } process_timedout_eps(); } } static DECLARE_WORK(skb_work, process_work); -static void ep_timeout(unsigned long arg) +static void ep_timeout(struct timer_list *t) { - struct c4iw_ep *ep = (struct c4iw_ep *)arg; + struct c4iw_ep *ep = timer_container_of(ep, t, timer); int kickit = 0; spin_lock(&timeout_lock); @@ -4172,15 +4382,15 @@ static int fw6_msg(struct c4iw_dev *dev, struct sk_buff *skb) struct c4iw_wr_wait *wr_waitp; int ret; - pr_debug("%s type %u\n", __func__, rpl->type); + pr_debug("type %u\n", rpl->type); switch (rpl->type) { case FW6_TYPE_WR_RPL: ret = (int)((be64_to_cpu(rpl->data[0]) >> 8) & 0xff); wr_waitp = (struct c4iw_wr_wait *)(__force unsigned long) rpl->data[1]; - pr_debug("%s wr_waitp %p ret %u\n", __func__, wr_waitp, ret); + pr_debug("wr_waitp %p ret %u\n", wr_waitp, ret); if (wr_waitp) - c4iw_wake_up(wr_waitp, ret ? -ret : 0); + c4iw_wake_up_deref(wr_waitp, ret ? -ret : 0); kfree_skb(skb); break; case FW6_TYPE_CQE: @@ -4210,15 +4420,14 @@ static int peer_abort_intr(struct c4iw_dev *dev, struct sk_buff *skb) return 0; } if (cxgb_is_neg_adv(req->status)) { - pr_debug("%s Negative advice on abort- tid %u status %d (%s)\n", - __func__, ep->hwtid, req->status, + pr_debug("Negative advice on abort- tid %u status %d (%s)\n", + ep->hwtid, req->status, neg_adv_str(req->status)); goto out; } - pr_debug("%s ep %p tid %u state %u\n", __func__, ep, ep->hwtid, - ep->com.state); + pr_debug("ep %p tid %u state %u\n", ep, ep->hwtid, ep->com.state); - c4iw_wake_up(&ep->com.wr_wait, -ECONNRESET); + c4iw_wake_up_noref(ep->com.wr_waitp, -ECONNRESET); out: sched(dev, skb); return 0; @@ -4244,13 +4453,13 @@ c4iw_handler_func c4iw_handlers[NUM_CPL_CMDS] = { [CPL_RDMA_TERMINATE] = sched, [CPL_FW4_ACK] = sched, [CPL_SET_TCB_RPL] = set_tcb_rpl, + [CPL_GET_TCB_RPL] = sched, [CPL_FW6_MSG] = fw6_msg, [CPL_RX_PKT] = sched }; int __init c4iw_cm_init(void) { - spin_lock_init(&timeout_lock); skb_queue_head_init(&rxq); workq = alloc_ordered_workqueue("iw_cxgb4", WQ_MEM_RECLAIM); @@ -4263,6 +4472,5 @@ int __init c4iw_cm_init(void) void c4iw_cm_term(void) { WARN_ON(!list_empty(&timeout_list)); - flush_workqueue(workq); destroy_workqueue(workq); } |
