Patchwork [5/5] RDMA/cxgb4: Fix bug for active and passive LE hash collision path

login
register
mail settings
Submitter Vipul Pandya
Date Nov. 29, 2012, 2:52 p.m.
Message ID <1354200745-23598-6-git-send-email-vipul@chelsio.com>
Download mbox | patch
Permalink /patch/202765/
State Superseded
Delegated to: David Miller
Headers show

Comments

Vipul Pandya - Nov. 29, 2012, 2:52 p.m.
Retries active opens for INUSE errors.

Logs any active ofld_connect_wr error replies.

Sends ofld_connect_wr on same ctrlq. It needs to go  on the same control txq as
regular CPL active/passive messages.

Retries on active open replies with EADDRINUSE.

Uses active open fw wr only if active filter region is set.

Adds stat for ofld_connect_wr failures.

This patch also adds debugfs file to show endpoints.

Signed-off-by: Vipul Pandya <vipul@chelsio.com>
---
 drivers/infiniband/hw/cxgb4/cm.c                |  277 ++++++++++++++++-------
 drivers/infiniband/hw/cxgb4/device.c            |  125 ++++++++++
 drivers/infiniband/hw/cxgb4/iw_cxgb4.h          |   32 +++
 drivers/net/ethernet/chelsio/cxgb4/cxgb4.h      |    1 +
 drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c |   19 ++-
 drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h  |    8 +-
 drivers/net/ethernet/chelsio/cxgb4/t4_regs.h    |    2 +
 7 files changed, 376 insertions(+), 88 deletions(-)

Patch

diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c
index b5840b4..255bd3f 100644
--- a/drivers/infiniband/hw/cxgb4/cm.c
+++ b/drivers/infiniband/hw/cxgb4/cm.c
@@ -276,6 +276,7 @@  void _c4iw_free_ep(struct kref *kref)
 		cxgb4_remove_tid(ep->com.dev->rdev.lldi.tids, 0, ep->hwtid);
 		dst_release(ep->dst);
 		cxgb4_l2t_release(ep->l2t);
+		remove_handle(ep->com.dev, &ep->com.dev->hwtid_idr, ep->hwtid);
 	}
 	kfree(ep);
 }
@@ -551,6 +552,7 @@  static int send_connect(struct c4iw_ep *ep)
 	req->opt0 = cpu_to_be64(opt0);
 	req->params = cpu_to_be32(select_ntuple(ep->com.dev, ep->dst, ep->l2t));
 	req->opt2 = cpu_to_be32(opt2);
+	set_bit(ACT_OPEN_REQ, &ep->com.history);
 	return c4iw_l2t_send(&ep->com.dev->rdev, skb, ep->l2t);
 }
 
@@ -827,6 +829,7 @@  static int act_establish(struct c4iw_dev *dev, struct sk_buff *skb)
 	/* setup the hwtid for this connection */
 	ep->hwtid = tid;
 	cxgb4_insert_tid(t, ep, tid);
+	insert_handle(dev, &dev->hwtid_idr, ep, ep->hwtid);
 
 	ep->snd_seq = be32_to_cpu(req->snd_isn);
 	ep->rcv_seq = be32_to_cpu(req->rcv_isn);
@@ -834,7 +837,9 @@  static int act_establish(struct c4iw_dev *dev, struct sk_buff *skb)
 	set_emss(ep, ntohs(req->tcp_opt));
 
 	/* dealloc the atid */
+	remove_handle(ep->com.dev, &ep->com.dev->atid_idr, atid);
 	cxgb4_free_atid(t, atid);
+	set_bit(ACT_ESTAB, &ep->com.history);
 
 	/* start MPA negotiation */
 	send_flowc(ep, NULL);
@@ -860,6 +865,7 @@  static void close_complete_upcall(struct c4iw_ep *ep)
 		ep->com.cm_id->rem_ref(ep->com.cm_id);
 		ep->com.cm_id = NULL;
 		ep->com.qp = NULL;
+		set_bit(CLOSE_UPCALL, &ep->com.history);
 	}
 }
 
@@ -868,6 +874,7 @@  static int abort_connection(struct c4iw_ep *ep, struct sk_buff *skb, gfp_t gfp)
 	PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
 	close_complete_upcall(ep);
 	state_set(&ep->com, ABORTING);
+	set_bit(ABORT_CONN, &ep->com.history);
 	return send_abort(ep, skb, gfp);
 }
 
@@ -882,6 +889,7 @@  static void peer_close_upcall(struct c4iw_ep *ep)
 		PDBG("peer close delivered ep %p cm_id %p tid %u\n",
 		     ep, ep->com.cm_id, ep->hwtid);
 		ep->com.cm_id->event_handler(ep->com.cm_id, &event);
+		set_bit(DISCONN_UPCALL, &ep->com.history);
 	}
 }
 
@@ -900,6 +908,7 @@  static void peer_abort_upcall(struct c4iw_ep *ep)
 		ep->com.cm_id->rem_ref(ep->com.cm_id);
 		ep->com.cm_id = NULL;
 		ep->com.qp = NULL;
+		set_bit(ABORT_UPCALL, &ep->com.history);
 	}
 }
 
@@ -932,6 +941,7 @@  static void connect_reply_upcall(struct c4iw_ep *ep, int status)
 
 	PDBG("%s ep %p tid %u status %d\n", __func__, ep,
 	     ep->hwtid, status);
+	set_bit(CONN_RPL_UPCALL, &ep->com.history);
 	ep->com.cm_id->event_handler(ep->com.cm_id, &event);
 
 	if (status < 0) {
@@ -972,6 +982,7 @@  static void connect_request_upcall(struct c4iw_ep *ep)
 						ep->parent_ep->com.cm_id,
 						&event);
 	}
+	set_bit(CONNREQ_UPCALL, &ep->com.history);
 	c4iw_put_ep(&ep->parent_ep->com);
 	ep->parent_ep = NULL;
 }
@@ -988,6 +999,7 @@  static void established_upcall(struct c4iw_ep *ep)
 	if (ep->com.cm_id) {
 		PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
 		ep->com.cm_id->event_handler(ep->com.cm_id, &event);
+		set_bit(ESTAB_UPCALL, &ep->com.history);
 	}
 }
 
@@ -1373,6 +1385,7 @@  static int rx_data(struct c4iw_dev *dev, struct sk_buff *skb)
 	unsigned int dlen = ntohs(hdr->len);
 	unsigned int tid = GET_TID(hdr);
 	struct tid_info *t = dev->rdev.lldi.tids;
+	__u8 status = hdr->status;
 
 	ep = lookup_tid(t, tid);
 	PDBG("%s ep %p tid %u dlen %u\n", __func__, ep, ep->hwtid, dlen);
@@ -1395,9 +1408,9 @@  static int rx_data(struct c4iw_dev *dev, struct sk_buff *skb)
 	case MPA_REP_SENT:
 		break;
 	default:
-		printk(KERN_ERR MOD "%s Unexpected streaming data."
-		       " ep %p state %d tid %u\n",
-		       __func__, ep, state_read(&ep->com), ep->hwtid);
+		pr_err("%s Unexpected streaming data." \
+		       " ep %p state %d tid %u status %d\n",
+		       __func__, ep, state_read(&ep->com), ep->hwtid, status);
 
 		/*
 		 * The ep will timeout and inform the ULP of the failure.
@@ -1464,6 +1477,7 @@  static void send_fw_act_open_req(struct c4iw_ep *ep, unsigned int atid)
 	req->tcb.cplrxdataack_cplpassacceptrpl =
 			htons(F_FW_OFLD_CONNECTION_WR_CPLRXDATAACK);
 	req->tcb.tx_max = jiffies;
+	req->tcb.rcv_adv = htons(1);
 	cxgb4_best_mtu(ep->com.dev->rdev.lldi.mtus, ep->mtu, &mtu_idx);
 	wscale = compute_wscale(rcv_win);
 	req->tcb.opt0 = TCAM_BYPASS(1) |
@@ -1491,7 +1505,8 @@  static void send_fw_act_open_req(struct c4iw_ep *ep, unsigned int atid)
 		req->tcb.opt2 |= WND_SCALE_EN(1);
 	req->tcb.opt0 = cpu_to_be64(req->tcb.opt0);
 	req->tcb.opt2 = cpu_to_be32(req->tcb.opt2);
-	set_wr_txq(skb, CPL_PRIORITY_CONTROL, 0);
+	set_wr_txq(skb, CPL_PRIORITY_CONTROL, ep->ctrlq_idx);
+	set_bit(ACT_OFLD_CONN, &ep->com.history);
 	c4iw_l2t_send(&ep->com.dev->rdev, skb, ep->l2t);
 }
 
@@ -1504,6 +1519,111 @@  static inline int act_open_has_tid(int status)
 	       status != CPL_ERR_ARP_MISS;
 }
 
+#define ACT_OPEN_RETRY_COUNT 2
+
+static int c4iw_reconnect(struct c4iw_ep *ep)
+{
+	int err = 0;
+	struct rtable *rt;
+	struct port_info *pi;
+	struct net_device *pdev;
+	int step;
+	struct neighbour *neigh;
+
+	PDBG("%s qp %p cm_id %p\n", __func__, ep->com.qp, ep->com.cm_id);
+	init_timer(&ep->timer);
+
+	/*
+	 * Allocate an active TID to initiate a TCP connection.
+	 */
+	ep->atid = cxgb4_alloc_atid(ep->com.dev->rdev.lldi.tids, ep);
+	if (ep->atid == -1) {
+		pr_err("%s - cannot alloc atid.\n", __func__);
+		err = -ENOMEM;
+		goto fail2;
+	}
+	insert_handle(ep->com.dev, &ep->com.dev->atid_idr, ep, ep->atid);
+
+	/* find a route */
+	rt = find_route(ep->com.dev,
+			ep->com.cm_id->local_addr.sin_addr.s_addr,
+			ep->com.cm_id->remote_addr.sin_addr.s_addr,
+			ep->com.cm_id->local_addr.sin_port,
+			ep->com.cm_id->remote_addr.sin_port, 0);
+	if (!rt) {
+		pr_err("%s - cannot find route.\n", __func__);
+		err = -EHOSTUNREACH;
+		goto fail3;
+	}
+	ep->dst = &rt->dst;
+
+	neigh = dst_neigh_lookup(ep->dst,
+			&ep->com.cm_id->remote_addr.sin_addr.s_addr);
+	/* get a l2t entry */
+	if (neigh->dev->flags & IFF_LOOPBACK) {
+		PDBG("%s LOOPBACK\n", __func__);
+		pdev = ip_dev_find(&init_net,
+				ep->com.cm_id->remote_addr.sin_addr.s_addr);
+		ep->l2t = cxgb4_l2t_get(ep->com.dev->rdev.lldi.l2t,
+				neigh, pdev, 0);
+		pi = (struct port_info *)netdev_priv(pdev);
+		ep->mtu = pdev->mtu;
+		ep->tx_chan = cxgb4_port_chan(pdev);
+		ep->smac_idx = (cxgb4_port_viid(pdev) & 0x7F) << 1;
+		dev_put(pdev);
+	} else {
+		ep->l2t = cxgb4_l2t_get(ep->com.dev->rdev.lldi.l2t,
+				neigh, neigh->dev, 0);
+		pi = (struct port_info *)netdev_priv(neigh->dev);
+		ep->mtu = dst_mtu(ep->dst);
+		ep->tx_chan = cxgb4_port_chan(neigh->dev);
+		ep->smac_idx = (cxgb4_port_viid(neigh->dev) &
+				0x7F) << 1;
+	}
+
+	step = ep->com.dev->rdev.lldi.ntxq / ep->com.dev->rdev.lldi.nchan;
+	ep->txq_idx = pi->port_id * step;
+	ep->ctrlq_idx = pi->port_id;
+	step = ep->com.dev->rdev.lldi.nrxq / ep->com.dev->rdev.lldi.nchan;
+	ep->rss_qid = ep->com.dev->rdev.lldi.rxq_ids[pi->port_id * step];
+
+	if (!ep->l2t) {
+		pr_err("%s - cannot alloc l2e.\n", __func__);
+		err = -ENOMEM;
+		goto fail4;
+	}
+
+	PDBG("%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,
+	     ep->l2t->idx);
+
+	state_set(&ep->com, CONNECTING);
+	ep->tos = 0;
+
+	/* send connect request to rnic */
+	err = send_connect(ep);
+	if (!err)
+		goto out;
+
+	cxgb4_l2t_release(ep->l2t);
+fail4:
+	dst_release(ep->dst);
+fail3:
+	remove_handle(ep->com.dev, &ep->com.dev->atid_idr, ep->atid);
+	cxgb4_free_atid(ep->com.dev->rdev.lldi.tids, ep->atid);
+fail2:
+	/*
+	 * remember to send notification to upper layer.
+	 * We are in here so the upper layer is not aware that this is
+	 * re-connect attempt and so, upper layer is still waiting for
+	 * response of 1st connect request.
+	 */
+	connect_reply_upcall(ep, -ECONNRESET);
+	c4iw_put_ep(&ep->com);
+out:
+	return err;
+}
+
 static int act_open_rpl(struct c4iw_dev *dev, struct sk_buff *skb)
 {
 	struct c4iw_ep *ep;
@@ -1524,6 +1644,8 @@  static int act_open_rpl(struct c4iw_dev *dev, struct sk_buff *skb)
 		return 0;
 	}
 
+	set_bit(ACT_OPEN_RPL, &ep->com.history);
+
 	/*
 	 * Log interesting failures.
 	 */
@@ -1532,12 +1654,27 @@  static int act_open_rpl(struct c4iw_dev *dev, struct sk_buff *skb)
 	case CPL_ERR_CONN_TIMEDOUT:
 		break;
 	case CPL_ERR_TCAM_FULL:
-		mutex_lock(&dev->rdev.stats.lock);
-		dev->rdev.stats.tcam_full++;
-		mutex_unlock(&dev->rdev.stats.lock);
-		send_fw_act_open_req(ep,
-			GET_TID_TID(GET_AOPEN_ATID(ntohl(rpl->atid_status))));
-		return 0;
+		if (dev->rdev.lldi.enable_fw_ofld_conn) {
+			mutex_lock(&dev->rdev.stats.lock);
+			dev->rdev.stats.tcam_full++;
+			mutex_unlock(&dev->rdev.stats.lock);
+			send_fw_act_open_req(ep,
+					     GET_TID_TID(GET_AOPEN_ATID(
+					     ntohl(rpl->atid_status))));
+			return 0;
+		}
+		break;
+	case CPL_ERR_CONN_EXIST:
+		if (ep->retry_count++ < ACT_OPEN_RETRY_COUNT) {
+			set_bit(ACT_RETRY_INUSE, &ep->com.history);
+			remove_handle(ep->com.dev, &ep->com.dev->atid_idr,
+					atid);
+			cxgb4_free_atid(t, atid);
+			dst_release(ep->dst);
+			cxgb4_l2t_release(ep->l2t);
+			c4iw_reconnect(ep);
+			return 0;
+		}
 		break;
 	default:
 		printk(KERN_INFO MOD "Active open failure - "
@@ -1556,6 +1693,7 @@  static int act_open_rpl(struct c4iw_dev *dev, struct sk_buff *skb)
 	if (status && act_open_has_tid(status))
 		cxgb4_remove_tid(ep->com.dev->rdev.lldi.tids, 0, GET_TID(rpl));
 
+	remove_handle(ep->com.dev, &ep->com.dev->atid_idr, atid);
 	cxgb4_free_atid(t, atid);
 	dst_release(ep->dst);
 	cxgb4_l2t_release(ep->l2t);
@@ -1776,7 +1914,7 @@  out:
 
 static int pass_accept_req(struct c4iw_dev *dev, struct sk_buff *skb)
 {
-	struct c4iw_ep *child_ep, *parent_ep;
+	struct c4iw_ep *child_ep = NULL, *parent_ep;
 	struct cpl_pass_accept_req *req = cplhdr(skb);
 	unsigned int stid = GET_POPEN_TID(ntohl(req->tos_stid));
 	struct tid_info *t = dev->rdev.lldi.tids;
@@ -1857,6 +1995,7 @@  static int pass_accept_req(struct c4iw_dev *dev, struct sk_buff *skb)
 	init_timer(&child_ep->timer);
 	cxgb4_insert_tid(t, child_ep, hwtid);
 	accept_cr(child_ep, peer_ip, skb, req);
+	set_bit(PASS_ACCEPT_REQ, &child_ep->com.history);
 	goto out;
 reject:
 	reject_cr(dev, hwtid, peer_ip, skb);
@@ -1880,11 +2019,13 @@  static int pass_establish(struct c4iw_dev *dev, struct sk_buff *skb)
 	     ntohs(req->tcp_opt));
 
 	set_emss(ep, ntohs(req->tcp_opt));
+	insert_handle(dev, &dev->hwtid_idr, ep, ep->hwtid);
 
 	dst_confirm(ep->dst);
 	state_set(&ep->com, MPA_REQ_WAIT);
 	start_ep_timer(ep);
 	send_flowc(ep, skb);
+	set_bit(PASS_ESTAB, &ep->com.history);
 
 	return 0;
 }
@@ -1904,6 +2045,7 @@  static int peer_close(struct c4iw_dev *dev, struct sk_buff *skb)
 	PDBG("%s ep %p tid %u\n", __func__, ep, ep->hwtid);
 	dst_confirm(ep->dst);
 
+	set_bit(PEER_CLOSE, &ep->com.history);
 	mutex_lock(&ep->com.mutex);
 	switch (ep->com.state) {
 	case MPA_REQ_WAIT:
@@ -1983,74 +2125,6 @@  static int is_neg_adv_abort(unsigned int status)
 	       status == CPL_ERR_PERSIST_NEG_ADVICE;
 }
 
-static int c4iw_reconnect(struct c4iw_ep *ep)
-{
-	struct rtable *rt;
-	int err = 0;
-
-	PDBG("%s qp %p cm_id %p\n", __func__, ep->com.qp, ep->com.cm_id);
-	init_timer(&ep->timer);
-
-	/*
-	 * Allocate an active TID to initiate a TCP connection.
-	 */
-	ep->atid = cxgb4_alloc_atid(ep->com.dev->rdev.lldi.tids, ep);
-	if (ep->atid == -1) {
-		printk(KERN_ERR MOD "%s - cannot alloc atid.\n", __func__);
-		err = -ENOMEM;
-		goto fail2;
-	}
-
-	/* find a route */
-	rt = find_route(ep->com.dev,
-			ep->com.cm_id->local_addr.sin_addr.s_addr,
-			ep->com.cm_id->remote_addr.sin_addr.s_addr,
-			ep->com.cm_id->local_addr.sin_port,
-			ep->com.cm_id->remote_addr.sin_port, 0);
-	if (!rt) {
-		printk(KERN_ERR MOD "%s - cannot find route.\n", __func__);
-		err = -EHOSTUNREACH;
-		goto fail3;
-	}
-	ep->dst = &rt->dst;
-
-	err = import_ep(ep, ep->com.cm_id->remote_addr.sin_addr.s_addr,
-			ep->dst, ep->com.dev, false);
-	if (err) {
-		printk(KERN_ERR MOD "%s - cannot alloc l2e.\n", __func__);
-		goto fail4;
-	}
-
-	PDBG("%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,
-	     ep->l2t->idx);
-
-	state_set(&ep->com, CONNECTING);
-	ep->tos = 0;
-
-	/* send connect request to rnic */
-	err = send_connect(ep);
-	if (!err)
-		goto out;
-
-	cxgb4_l2t_release(ep->l2t);
-fail4:
-	dst_release(ep->dst);
-fail3:
-	cxgb4_free_atid(ep->com.dev->rdev.lldi.tids, ep->atid);
-fail2:
-	/*
-	 * remember to send notification to upper layer.
-	 * We are in here so the upper layer is not aware that this is
-	 * re-connect attempt and so, upper layer is still waiting for
-	 * response of 1st connect request.
-	 */
-	connect_reply_upcall(ep, -ECONNRESET);
-	c4iw_put_ep(&ep->com);
-out:
-	return err;
-}
-
 static int peer_abort(struct c4iw_dev *dev, struct sk_buff *skb)
 {
 	struct cpl_abort_req_rss *req = cplhdr(skb);
@@ -2071,6 +2145,7 @@  static int peer_abort(struct c4iw_dev *dev, struct sk_buff *skb)
 	}
 	PDBG("%s ep %p tid %u state %u\n", __func__, ep, ep->hwtid,
 	     ep->com.state);
+	set_bit(PEER_ABORT, &ep->com.history);
 
 	/*
 	 * Wake up any threads in rdma_init() or rdma_fini().
@@ -2285,6 +2360,7 @@  int c4iw_reject_cr(struct iw_cm_id *cm_id, const void *pdata, u8 pdata_len)
 		c4iw_put_ep(&ep->com);
 		return -ECONNRESET;
 	}
+	set_bit(ULP_REJECT, &ep->com.history);
 	BUG_ON(state_read(&ep->com) != MPA_REQ_RCVD);
 	if (mpa_rev == 0)
 		abort_connection(ep, NULL, GFP_KERNEL);
@@ -2314,6 +2390,7 @@  int c4iw_accept_cr(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
 	BUG_ON(state_read(&ep->com) != MPA_REQ_RCVD);
 	BUG_ON(!qp);
 
+	set_bit(ULP_ACCEPT, &ep->com.history);
 	if ((conn_param->ord > c4iw_max_read_depth) ||
 	    (conn_param->ird > c4iw_max_read_depth)) {
 		abort_connection(ep, NULL, GFP_KERNEL);
@@ -2437,6 +2514,7 @@  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);
 
 	PDBG("%s saddr 0x%x sport 0x%x raddr 0x%x rport 0x%x\n", __func__,
 	     ntohl(cm_id->local_addr.sin_addr.s_addr),
@@ -2482,6 +2560,7 @@  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);
 	cxgb4_free_atid(ep->com.dev->rdev.lldi.tids, ep->atid);
 fail2:
 	cm_id->rem_ref(cm_id);
@@ -2524,6 +2603,7 @@  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);
 	state_set(&ep->com, LISTEN);
 	if (dev->rdev.lldi.enable_fw_ofld_conn) {
 		do {
@@ -2531,7 +2611,10 @@  int c4iw_create_listen(struct iw_cm_id *cm_id, int backlog)
 				ep->com.dev->rdev.lldi.ports[0], ep->stid,
 				ep->com.local_addr.sin_addr.s_addr,
 				ep->com.local_addr.sin_port,
-				ep->com.dev->rdev.lldi.rxq_ids[0]);
+				0,
+				ep->com.dev->rdev.lldi.rxq_ids[0],
+				0,
+				0);
 			if (err == -EBUSY) {
 				set_current_state(TASK_UNINTERRUPTIBLE);
 				schedule_timeout(usecs_to_jiffies(100));
@@ -2542,6 +2625,7 @@  int c4iw_create_listen(struct iw_cm_id *cm_id, int backlog)
 		err = cxgb4_create_server(ep->com.dev->rdev.lldi.ports[0],
 				ep->stid, ep->com.local_addr.sin_addr.s_addr,
 				ep->com.local_addr.sin_port,
+				0,
 				ep->com.dev->rdev.lldi.rxq_ids[0]);
 		if (!err)
 			err = c4iw_wait_for_reply(&ep->com.dev->rdev,
@@ -2587,6 +2671,7 @@  int c4iw_destroy_listen(struct iw_cm_id *cm_id)
 		err = c4iw_wait_for_reply(&ep->com.dev->rdev, &ep->com.wr_wait,
 					  0, 0, __func__);
 	}
+	remove_handle(ep->com.dev, &ep->com.dev->stid_idr, ep->stid);
 	cxgb4_free_stid(ep->com.dev->rdev.lldi.tids, ep->stid, PF_INET);
 done:
 	cm_id->rem_ref(cm_id);
@@ -2650,10 +2735,13 @@  int c4iw_ep_disconnect(struct c4iw_ep *ep, int abrupt, gfp_t gfp)
 
 	if (close) {
 		if (abrupt) {
+			set_bit(EP_DISC_ABORT, &ep->com.history);
 			close_complete_upcall(ep);
 			ret = send_abort(ep, NULL, gfp);
-		} else
+		} else {
+			set_bit(EP_DISC_CLOSE, &ep->com.history);
 			ret = send_halfclose(ep, gfp);
+		}
 		if (ret)
 			fatal = 1;
 	}
@@ -2667,6 +2755,7 @@  static void active_ofld_conn_reply(struct c4iw_dev *dev, struct sk_buff *skb,
 			struct cpl_fw6_msg_ofld_connection_wr_rpl *req)
 {
 	struct c4iw_ep *ep;
+	int atid = be32_to_cpu(req->tid);
 
 	ep = (struct c4iw_ep *)lookup_atid(dev->rdev.lldi.tids, req->tid);
 	if (!ep)
@@ -2674,15 +2763,35 @@  static void active_ofld_conn_reply(struct c4iw_dev *dev, struct sk_buff *skb,
 
 	switch (req->retval) {
 	case FW_ENOMEM:
+		set_bit(ACT_RETRY_NOMEM, &ep->com.history);
+		if (ep->retry_count++ < ACT_OPEN_RETRY_COUNT) {
+			send_fw_act_open_req(ep, atid);
+			return;
+		}
 	case FW_EADDRINUSE:
-		PDBG("%s ofld conn wr ret %d\n", __func__, req->retval);
+		set_bit(ACT_RETRY_INUSE, &ep->com.history);
+		if (ep->retry_count++ < ACT_OPEN_RETRY_COUNT) {
+			send_fw_act_open_req(ep, atid);
+			return;
+		}
 		break;
 	default:
 		pr_info("%s unexpected ofld conn wr retval %d\n",
 		       __func__, req->retval);
 		break;
 	}
+	pr_err("active ofld_connect_wr failure %d atid %d\n",
+	       req->retval, atid);
+	mutex_lock(&dev->rdev.stats.lock);
+	dev->rdev.stats.act_ofld_conn_fails++;
+	mutex_unlock(&dev->rdev.stats.lock);
 	connect_reply_upcall(ep, status2errno(req->retval));
+	state_set(&ep->com, DEAD);
+	remove_handle(dev, &dev->atid_idr, atid);
+	cxgb4_free_atid(dev->rdev.lldi.tids, atid);
+	dst_release(ep->dst);
+	cxgb4_l2t_release(ep->l2t);
+	c4iw_put_ep(&ep->com);
 }
 
 static void passive_ofld_conn_reply(struct c4iw_dev *dev, struct sk_buff *skb,
@@ -2696,6 +2805,9 @@  static void passive_ofld_conn_reply(struct c4iw_dev *dev, struct sk_buff *skb,
 	BUG_ON(!rpl_skb);
 	if (req->retval) {
 		PDBG("%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);
 		kfree_skb(rpl_skb);
 	} else {
 		cpl = (struct cpl_pass_accept_req *)cplhdr(rpl_skb);
@@ -2989,6 +3101,7 @@  static void process_timeout(struct c4iw_ep *ep)
 	mutex_lock(&ep->com.mutex);
 	PDBG("%s ep %p tid %u state %d\n", __func__, ep, ep->hwtid,
 	     ep->com.state);
+	set_bit(TIMEDOUT, &ep->com.history);
 	switch (ep->com.state) {
 	case MPA_REQ_SENT:
 		__state_set(&ep->com, ABORTING);
diff --git a/drivers/infiniband/hw/cxgb4/device.c b/drivers/infiniband/hw/cxgb4/device.c
index 6b5b3d1..ba11c76 100644
--- a/drivers/infiniband/hw/cxgb4/device.c
+++ b/drivers/infiniband/hw/cxgb4/device.c
@@ -280,6 +280,10 @@  static int stats_show(struct seq_file *seq, void *v)
 		   db_state_str[dev->db_state],
 		   dev->rdev.stats.db_state_transitions);
 	seq_printf(seq, "TCAM_FULL: %10llu\n", dev->rdev.stats.tcam_full);
+	seq_printf(seq, "ACT_OFLD_CONN_FAILS: %10llu\n",
+		   dev->rdev.stats.act_ofld_conn_fails);
+	seq_printf(seq, "PAS_OFLD_CONN_FAILS: %10llu\n",
+		   dev->rdev.stats.pas_ofld_conn_fails);
 	return 0;
 }
 
@@ -310,6 +314,9 @@  static ssize_t stats_clear(struct file *file, const char __user *buf,
 	dev->rdev.stats.db_empty = 0;
 	dev->rdev.stats.db_drop = 0;
 	dev->rdev.stats.db_state_transitions = 0;
+	dev->rdev.stats.tcam_full = 0;
+	dev->rdev.stats.act_ofld_conn_fails = 0;
+	dev->rdev.stats.pas_ofld_conn_fails = 0;
 	mutex_unlock(&dev->rdev.stats.lock);
 	return count;
 }
@@ -323,6 +330,113 @@  static const struct file_operations stats_debugfs_fops = {
 	.write   = stats_clear,
 };
 
+static int dump_ep(int id, void *p, void *data)
+{
+	struct c4iw_ep *ep = p;
+	struct c4iw_debugfs_data *epd = data;
+	int space;
+	int cc;
+
+	space = epd->bufsize - epd->pos - 1;
+	if (space == 0)
+		return 1;
+
+	cc = snprintf(epd->buf + epd->pos, space,
+			"ep %p cm_id %p qp %p state %d flags 0x%lx history 0x%lx "
+			"hwtid %d atid %d %pI4:%d <-> %pI4:%d\n",
+			ep, ep->com.cm_id, ep->com.qp, (int)ep->com.state,
+			ep->com.flags, ep->com.history, ep->hwtid, ep->atid,
+			&ep->com.local_addr.sin_addr.s_addr,
+			ntohs(ep->com.local_addr.sin_port),
+			&ep->com.remote_addr.sin_addr.s_addr,
+			ntohs(ep->com.remote_addr.sin_port));
+	if (cc < space)
+		epd->pos += cc;
+	return 0;
+}
+
+static int dump_listen_ep(int id, void *p, void *data)
+{
+	struct c4iw_listen_ep *ep = p;
+	struct c4iw_debugfs_data *epd = data;
+	int space;
+	int cc;
+
+	space = epd->bufsize - epd->pos - 1;
+	if (space == 0)
+		return 1;
+
+	cc = snprintf(epd->buf + epd->pos, space,
+			"ep %p cm_id %p state %d flags 0x%lx stid %d backlog %d "
+			"%pI4:%d\n", ep, ep->com.cm_id, (int)ep->com.state,
+			ep->com.flags, ep->stid, ep->backlog,
+			&ep->com.local_addr.sin_addr.s_addr,
+			ntohs(ep->com.local_addr.sin_port));
+	if (cc < space)
+		epd->pos += cc;
+	return 0;
+}
+
+static int ep_release(struct inode *inode, struct file *file)
+{
+	struct c4iw_debugfs_data *epd = file->private_data;
+	if (!epd) {
+		pr_info("%s null qpd?\n", __func__);
+		return 0;
+	}
+	vfree(epd->buf);
+	kfree(epd);
+	return 0;
+}
+
+static int ep_open(struct inode *inode, struct file *file)
+{
+	struct c4iw_debugfs_data *epd;
+	int ret = 0;
+	int count = 1;
+
+	epd = kmalloc(sizeof(*epd), GFP_KERNEL);
+	if (!epd) {
+		ret = -ENOMEM;
+		goto out;
+	}
+	epd->devp = inode->i_private;
+	epd->pos = 0;
+
+	spin_lock_irq(&epd->devp->lock);
+	idr_for_each(&epd->devp->hwtid_idr, count_idrs, &count);
+	idr_for_each(&epd->devp->atid_idr, count_idrs, &count);
+	idr_for_each(&epd->devp->stid_idr, count_idrs, &count);
+	spin_unlock_irq(&epd->devp->lock);
+
+	epd->bufsize = count * 160;
+	epd->buf = vmalloc(epd->bufsize);
+	if (!epd->buf) {
+		ret = -ENOMEM;
+		goto err1;
+	}
+
+	spin_lock_irq(&epd->devp->lock);
+	idr_for_each(&epd->devp->hwtid_idr, dump_ep, epd);
+	idr_for_each(&epd->devp->atid_idr, dump_ep, epd);
+	idr_for_each(&epd->devp->stid_idr, dump_listen_ep, epd);
+	spin_unlock_irq(&epd->devp->lock);
+
+	file->private_data = epd;
+	goto out;
+err1:
+	kfree(epd);
+out:
+	return ret;
+}
+
+static const struct file_operations ep_debugfs_fops = {
+	.owner   = THIS_MODULE,
+	.open    = ep_open,
+	.release = ep_release,
+	.read    = debugfs_read,
+};
+
 static int setup_debugfs(struct c4iw_dev *devp)
 {
 	struct dentry *de;
@@ -345,6 +459,11 @@  static int setup_debugfs(struct c4iw_dev *devp)
 	if (de && de->d_inode)
 		de->d_inode->i_size = 4096;
 
+	de = debugfs_create_file("eps", S_IWUSR, devp->debugfs_root,
+			(void *)devp, &ep_debugfs_fops);
+	if (de && de->d_inode)
+		de->d_inode->i_size = 4096;
+
 	return 0;
 }
 
@@ -476,6 +595,9 @@  static void c4iw_dealloc(struct uld_ctx *ctx)
 	idr_destroy(&ctx->dev->cqidr);
 	idr_destroy(&ctx->dev->qpidr);
 	idr_destroy(&ctx->dev->mmidr);
+	idr_destroy(&ctx->dev->hwtid_idr);
+	idr_destroy(&ctx->dev->stid_idr);
+	idr_destroy(&ctx->dev->atid_idr);
 	iounmap(ctx->dev->rdev.oc_mw_kva);
 	ib_dealloc_device(&ctx->dev->ibdev);
 	ctx->dev = NULL;
@@ -533,6 +655,9 @@  static struct c4iw_dev *c4iw_alloc(const struct cxgb4_lld_info *infop)
 	idr_init(&devp->cqidr);
 	idr_init(&devp->qpidr);
 	idr_init(&devp->mmidr);
+	idr_init(&devp->hwtid_idr);
+	idr_init(&devp->stid_idr);
+	idr_init(&devp->atid_idr);
 	spin_lock_init(&devp->lock);
 	mutex_init(&devp->rdev.stats.lock);
 	mutex_init(&devp->db_mutex);
diff --git a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
index 6a17fde..9c1644f 100644
--- a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
+++ b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
@@ -131,6 +131,8 @@  struct c4iw_stats {
 	u64  db_drop;
 	u64  db_state_transitions;
 	u64  tcam_full;
+	u64  act_ofld_conn_fails;
+	u64  pas_ofld_conn_fails;
 };
 
 struct c4iw_rdev {
@@ -224,6 +226,9 @@  struct c4iw_dev {
 	struct dentry *debugfs_root;
 	enum db_state db_state;
 	int qpcnt;
+	struct idr hwtid_idr;
+	struct idr atid_idr;
+	struct idr stid_idr;
 };
 
 static inline struct c4iw_dev *to_c4iw_dev(struct ib_device *ibdev)
@@ -713,6 +718,31 @@  enum c4iw_ep_flags {
 	CLOSE_SENT		= 3,
 };
 
+enum c4iw_ep_history {
+	ACT_OPEN_REQ            = 0,
+	ACT_OFLD_CONN           = 1,
+	ACT_OPEN_RPL            = 2,
+	ACT_ESTAB               = 3,
+	PASS_ACCEPT_REQ         = 4,
+	PASS_ESTAB              = 5,
+	ABORT_UPCALL            = 6,
+	ESTAB_UPCALL            = 7,
+	CLOSE_UPCALL            = 8,
+	ULP_ACCEPT              = 9,
+	ULP_REJECT              = 10,
+	TIMEDOUT                = 11,
+	PEER_ABORT              = 12,
+	PEER_CLOSE              = 13,
+	CONNREQ_UPCALL          = 14,
+	ABORT_CONN              = 15,
+	DISCONN_UPCALL          = 16,
+	EP_DISC_CLOSE           = 17,
+	EP_DISC_ABORT           = 18,
+	CONN_RPL_UPCALL         = 19,
+	ACT_RETRY_NOMEM         = 20,
+	ACT_RETRY_INUSE         = 21
+};
+
 struct c4iw_ep_common {
 	struct iw_cm_id *cm_id;
 	struct c4iw_qp *qp;
@@ -724,6 +754,7 @@  struct c4iw_ep_common {
 	struct sockaddr_in remote_addr;
 	struct c4iw_wr_wait wr_wait;
 	unsigned long flags;
+	unsigned long history;
 };
 
 struct c4iw_listen_ep {
@@ -761,6 +792,7 @@  struct c4iw_ep {
 	u8 tos;
 	u8 retry_with_mpa_v1;
 	u8 tried_with_mpa_v1;
+	unsigned int retry_count;
 };
 
 static inline struct c4iw_ep *to_ep(struct iw_cm_id *cm_id)
diff --git a/drivers/net/ethernet/chelsio/cxgb4/cxgb4.h b/drivers/net/ethernet/chelsio/cxgb4/cxgb4.h
index b5d299d..682d490 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/cxgb4.h
+++ b/drivers/net/ethernet/chelsio/cxgb4/cxgb4.h
@@ -530,6 +530,7 @@  struct adapter {
 	struct net_device *port[MAX_NPORTS];
 	u8 chan_map[NCHAN];                   /* channel -> port map */
 
+	u32 filter_mode;
 	unsigned int l2t_start;
 	unsigned int l2t_end;
 	struct l2t_data *l2t;
diff --git a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
index b9ce750..97ab36f 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
+++ b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
@@ -2688,7 +2688,8 @@  static int tid_init(struct tid_info *t)
  *	Returns <0 on error and one of the %NET_XMIT_* values on success.
  */
 int cxgb4_create_server(const struct net_device *dev, unsigned int stid,
-			__be32 sip, __be16 sport, unsigned int queue)
+			__be32 sip, __be16 sport, __be16 vlan,
+			unsigned int queue)
 {
 	unsigned int chan;
 	struct sk_buff *skb;
@@ -3061,7 +3062,7 @@  static void uld_attach(struct adapter *adap, unsigned int uld)
 	lli.ucq_density = 1 << QUEUESPERPAGEPF0_GET(
 			t4_read_reg(adap, SGE_INGRESS_QUEUES_PER_PAGE_PF) >>
 			(adap->fn * 4));
-	lli.filt_mode = tp_vlan_pri_map;
+	lli.filt_mode = adap->filter_mode;
 	/* MODQ_REQ_MAP sets queues 0-3 to chan 0-3 */
 	for (i = 0; i < NCHAN; i++)
 		lli.tx_modq[i] = i;
@@ -3327,7 +3328,8 @@  static int delete_filter(struct adapter *adapter, unsigned int fidx)
 }
 
 int cxgb4_create_server_filter(const struct net_device *dev, unsigned int stid,
-		__be32 sip, __be16 sport, unsigned int queue)
+		__be32 sip, __be16 sport, __be16 vlan,
+		unsigned int queue, unsigned char port, unsigned char mask)
 {
 	int ret;
 	struct filter_entry *f;
@@ -3361,11 +3363,16 @@  int cxgb4_create_server_filter(const struct net_device *dev, unsigned int stid,
 	f->fs.val.lport = cpu_to_be16(sport);
 	f->fs.mask.lport  = ~0;
 	val = (u8 *)&sip;
-	if ((val[0] | val[1] | val[2] | val[3]) != 0)
+	if ((val[0] | val[1] | val[2] | val[3]) != 0) {
 		for (i = 0; i < 4; i++) {
 			f->fs.val.lip[i] = val[i];
 			f->fs.mask.lip[i] = ~0;
 		}
+		if (adap->filter_mode & F_PORT) {
+			f->fs.val.iport = port;
+			f->fs.mask.iport = mask;
+		}
+	}
 
 	f->fs.dirsteer = 1;
 	f->fs.iq = queue;
@@ -4472,6 +4479,10 @@  static int adap_init0(struct adapter *adap)
 	for (j = 0; j < NCHAN; j++)
 		adap->params.tp.tx_modq[j] = j;
 
+	t4_read_indirect(adap, TP_PIO_ADDR, TP_PIO_DATA,
+			 &adap->filter_mode, 1,
+			 TP_VLAN_PRI_MAP);
+
 	adap->flags |= FW_OK;
 	return 0;
 
diff --git a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h
index 065bbd5..e2bbc7f 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h
+++ b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h
@@ -38,6 +38,7 @@ 
 #include <linux/cache.h>
 #include <linux/spinlock.h>
 #include <linux/skbuff.h>
+#include <linux/inetdevice.h>
 #include <linux/atomic.h>
 
 /* CPL message priority levels */
@@ -151,9 +152,12 @@  void cxgb4_remove_tid(struct tid_info *t, unsigned int qid, unsigned int tid);
 struct in6_addr;
 
 int cxgb4_create_server(const struct net_device *dev, unsigned int stid,
-			__be32 sip, __be16 sport, unsigned int queue);
+			__be32 sip, __be16 sport, __be16 vlan,
+			unsigned int queue);
 int cxgb4_create_server_filter(const struct net_device *dev, unsigned int stid,
-			       __be32 sip, __be16 sport, unsigned int queue);
+			       __be32 sip, __be16 sport, __be16 vlan,
+			       unsigned int queue,
+			       unsigned char port, unsigned char mask);
 int cxgb4_remove_server_filter(const struct net_device *dev, unsigned int stid,
 			       unsigned int queue, bool ipv6);
 static inline void set_wr_txq(struct sk_buff *skb, int prio, int queue)
diff --git a/drivers/net/ethernet/chelsio/cxgb4/t4_regs.h b/drivers/net/ethernet/chelsio/cxgb4/t4_regs.h
index bdc1bbc..3139221 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/t4_regs.h
+++ b/drivers/net/ethernet/chelsio/cxgb4/t4_regs.h
@@ -1066,5 +1066,7 @@ 
 #define A_TP_TX_SCHED_PCMD 0x25
 
 #define S_PORT    1
+#define V_PORT(x) ((x) << S_PORT)
+#define F_PORT    V_PORT(1U)
 
 #endif /* __T4_REGS_H */