[PATCH 0/4] Misc. fixes for iw_cxgb4

[PATCH 0/4] Misc. fixes for iw_cxgb4

[Hariprasad Shenai]

This patch series adds support to determine ingress padding boundary at runtime.
Advertise a larger max read queue depth for qps, and gather the resource limits
from fw and use them to avoid exhausting all the resources and display TPTE on
errors and add support for work request logging feature.

The patches series is created against 'net-next' tree.
And includes patches on cxgb4 and iw_cxgb4 driver.

Since this patch-series contains changes which are dependent on commit id
fc5ab02 ("cxgb4: Replaced the backdoor mechanism to access the HW memory with
PCIe Window method") of net-next tree, we would like to request this patch
series to get merged via David Miller's 'net-next' tree.

We have included all the maintainers of respective drivers. Kindly review the
change and let us know in case of any review comments.

Hariprasad Shenai (4):
  iw_cxgb4: Detect Ing. Padding Boundary at run-time
  cxgb4/iw_cxgb4: use firmware ord/ird resource limits
  cxgb4/iw_cxgb4: display TPTE on errors
  cxgb4/iw_cxgb4: work request logging feature

 drivers/infiniband/hw/cxgb4/cm.c                |   80 +++++++---
 drivers/infiniband/hw/cxgb4/cq.c                |    8 +-
 drivers/infiniband/hw/cxgb4/device.c            |  187 ++++++++++++++++++++++-
 drivers/infiniband/hw/cxgb4/ev.c                |   53 ++++++-
 drivers/infiniband/hw/cxgb4/iw_cxgb4.h          |   38 +++++-
 drivers/infiniband/hw/cxgb4/provider.c          |   10 +-
 drivers/infiniband/hw/cxgb4/qp.c                |   75 ++++++++--
 drivers/infiniband/hw/cxgb4/t4.h                |   16 +--
 drivers/net/ethernet/chelsio/cxgb4/cxgb4.h      |    3 +
 drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c |  100 ++++++++++++
 drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h  |    7 +
 drivers/net/ethernet/chelsio/cxgb4/t4_regs.h    |    6 +
 drivers/net/ethernet/chelsio/cxgb4/t4fw_api.h   |    2 +
 13 files changed, 524 insertions(+), 61 deletions(-)

[PATCH 2/4] cxgb4/iw_cxgb4: use firmware ord/ird resource limits

[Hariprasad Shenai]

Advertise a larger max read queue depth for qps, and gather the resource limits
from fw and use them to avoid exhaustinq all the resources.

Design:

cxgb4:

Obtain the max_ordird_qp and max_ird_adapter device params from FW
at init time and pass them up to the ULDs when they attach.  If these
parameters are not available, due to older firmware, then hard-code
the values based on the known values for older firmware.
iw_cxgb4:

Fix the c4iw_query_device() to report these correct values based on
adapter parameters.  ibv_query_device() will always return:

max_qp_rd_atom = max_qp_init_rd_atom = min(module_max, max_ordird_qp)
max_res_rd_atom = max_ird_adapter

Bump up the per qp max module option to 32, allowing it to be increased
by the user up to the device max of max_ordird_qp.  32 seems to be
sufficient to maximize throughput for streaming read benchmarks.

Fail connection setup if the negotiated IRD exhausts the available
adapter ird resources.  So the driver will track the amount of ird
resource in use and not send an RI_WR/INIT to FW that would reduce the
available ird resources below zero.

Signed-off-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
---
 drivers/infiniband/hw/cxgb4/cm.c                |   80 ++++++++++++++++-------
 drivers/infiniband/hw/cxgb4/device.c            |    2 +
 drivers/infiniband/hw/cxgb4/iw_cxgb4.h          |    9 ++-
 drivers/infiniband/hw/cxgb4/provider.c          |    6 +-
 drivers/infiniband/hw/cxgb4/qp.c                |   53 +++++++++++++--
 drivers/net/ethernet/chelsio/cxgb4/cxgb4.h      |    3 +
 drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c |   18 +++++
 drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h  |    2 +
 drivers/net/ethernet/chelsio/cxgb4/t4fw_api.h   |    2 +
 9 files changed, 141 insertions(+), 34 deletions(-)

diff --git a/drivers/infiniband/hw/cxgb4/cm.c b/drivers/infiniband/hw/cxgb4/cm.c
index d62a0f9..df5bd3d 100644
--- a/drivers/infiniband/hw/cxgb4/cm.c
+++ b/drivers/infiniband/hw/cxgb4/cm.c
@@ -79,9 +79,10 @@ static int dack_mode = 1;
 module_param(dack_mode, int, 0644);
 MODULE_PARM_DESC(dack_mode, "Delayed ack mode (default=1)");
 
-int c4iw_max_read_depth = 8;
+uint c4iw_max_read_depth = 32;
 module_param(c4iw_max_read_depth, int, 0644);
-MODULE_PARM_DESC(c4iw_max_read_depth, "Per-connection max ORD/IRD (default=8)");
+MODULE_PARM_DESC(c4iw_max_read_depth,
+		 "Per-connection max ORD/IRD (default=32)");
 
 static int enable_tcp_timestamps;
 module_param(enable_tcp_timestamps, int, 0644);
@@ -813,6 +814,8 @@ static void 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));
+		PDBG("%s initiator ird %u ord %u\n", __func__, ep->ird,
+		     ep->ord);
 		mpa_v2_params.ird = htons((u16)ep->ird);
 		mpa_v2_params.ord = htons((u16)ep->ord);
 
@@ -1182,8 +1185,8 @@ static int connect_request_upcall(struct c4iw_ep *ep)
 			sizeof(struct mpa_v2_conn_params);
 	} else {
 		/* this means MPA_v1 is used. Send max supported */
-		event.ord = c4iw_max_read_depth;
-		event.ird = c4iw_max_read_depth;
+		event.ord = cur_max_read_depth(ep->com.dev);
+		event.ird = cur_max_read_depth(ep->com.dev);
 		event.private_data_len = ep->plen;
 		event.private_data = ep->mpa_pkt + sizeof(struct mpa_message);
 	}
@@ -1247,6 +1250,8 @@ static int update_rx_credits(struct c4iw_ep *ep, u32 credits)
 	return credits;
 }
 
+#define RELAXED_IRD_NEGOTIATION 1
+
 static int process_mpa_reply(struct c4iw_ep *ep, struct sk_buff *skb)
 {
 	struct mpa_message *mpa;
@@ -1358,17 +1363,33 @@ 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;
+			PDBG("%s responder ird %u ord %u ep ird %u ord %u\n",
+			     __func__, resp_ird, resp_ord, ep->ird, ep->ord);
 
 			/*
 			 * This is a double-check. Ideally, below checks are
 			 * not required since ird/ord stuff has been taken
 			 * care of in c4iw_accept_cr
 			 */
-			if ((ep->ird < resp_ord) || (ep->ord > resp_ird)) {
+			if (ep->ird < resp_ord) {
+				if (RELAXED_IRD_NEGOTIATION && resp_ord <=
+				    ep->com.dev->rdev.lldi.max_ordird_qp)
+					ep->ird = resp_ord;
+				else
+					insuff_ird = 1;
+			} else if (ep->ird > resp_ord) {
+				ep->ird = resp_ord;
+			}
+			if (ep->ord > resp_ird) {
+				if (RELAXED_IRD_NEGOTIATION)
+					ep->ord = resp_ird;
+				else
+					insuff_ird = 1;
+			}
+			if (insuff_ird) {
 				err = -ENOMEM;
 				ep->ird = resp_ord;
 				ep->ord = resp_ird;
-				insuff_ird = 1;
 			}
 
 			if (ntohs(mpa_v2_params->ird) &
@@ -1571,6 +1592,8 @@ static void process_mpa_request(struct c4iw_ep *ep, struct sk_buff *skb)
 				MPA_V2_IRD_ORD_MASK;
 			ep->ord = ntohs(mpa_v2_params->ord) &
 				MPA_V2_IRD_ORD_MASK;
+			PDBG("%s initiator ird %u ord %u\n", __func__, ep->ird,
+			     ep->ord);
 			if (ntohs(mpa_v2_params->ird) & MPA_V2_PEER2PEER_MODEL)
 				if (peer2peer) {
 					if (ntohs(mpa_v2_params->ord) &
@@ -2724,8 +2747,8 @@ int c4iw_accept_cr(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
 	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)) {
+	if ((conn_param->ord > cur_max_read_depth(ep->com.dev)) ||
+	    (conn_param->ird > cur_max_read_depth(ep->com.dev))) {
 		abort_connection(ep, NULL, GFP_KERNEL);
 		err = -EINVAL;
 		goto err;
@@ -2733,31 +2756,41 @@ int c4iw_accept_cr(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
 
 	if (ep->mpa_attr.version == 2 && ep->mpa_attr.enhanced_rdma_conn) {
 		if (conn_param->ord > ep->ird) {
-			ep->ird = conn_param->ird;
-			ep->ord = conn_param->ord;
-			send_mpa_reject(ep, conn_param->private_data,
-					conn_param->private_data_len);
-			abort_connection(ep, NULL, GFP_KERNEL);
-			err = -ENOMEM;
-			goto err;
+			if (RELAXED_IRD_NEGOTIATION) {
+				ep->ord = ep->ird;
+			} else {
+				ep->ird = conn_param->ird;
+				ep->ord = conn_param->ord;
+				send_mpa_reject(ep, conn_param->private_data,
+						conn_param->private_data_len);
+				abort_connection(ep, NULL, GFP_KERNEL);
+				err = -ENOMEM;
+				goto err;
+			}
 		}
-		if (conn_param->ird > ep->ord) {
-			if (!ep->ord)
-				conn_param->ird = 1;
-			else {
+		if (conn_param->ird < ep->ord) {
+			if (RELAXED_IRD_NEGOTIATION &&
+			    ep->ord <= h->rdev.lldi.max_ordird_qp) {
+				conn_param->ird = ep->ord;
+			} else {
 				abort_connection(ep, NULL, GFP_KERNEL);
 				err = -ENOMEM;
 				goto err;
 			}
 		}
-
 	}
 	ep->ird = conn_param->ird;
 	ep->ord = conn_param->ord;
 
-	if (ep->mpa_attr.version != 2)
+	if (ep->mpa_attr.version == 1) {
 		if (peer2peer && ep->ird == 0)
 			ep->ird = 1;
+	} else {
+		if (peer2peer &&
+		    (ep->mpa_attr.p2p_type != FW_RI_INIT_P2PTYPE_DISABLED) &&
+		    (p2p_type == FW_RI_INIT_P2PTYPE_READ_REQ) && ep->ord == 0)
+			ep->ird = 1;
+	}
 
 	PDBG("%s %d ird %d ord %d\n", __func__, __LINE__, ep->ird, ep->ord);
 
@@ -2796,6 +2829,7 @@ int c4iw_accept_cr(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
 	return 0;
 err1:
 	ep->com.cm_id = NULL;
+	abort_connection(ep, NULL, GFP_KERNEL);
 	cm_id->rem_ref(cm_id);
 err:
 	mutex_unlock(&ep->com.mutex);
@@ -2879,8 +2913,8 @@ int c4iw_connect(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param)
 	int iptype;
 	int iwpm_err = 0;
 
-	if ((conn_param->ord > c4iw_max_read_depth) ||
-	    (conn_param->ird > c4iw_max_read_depth)) {
+	if ((conn_param->ord > cur_max_read_depth(dev)) ||
+	    (conn_param->ird > cur_max_read_depth(dev))) {
 		err = -EINVAL;
 		goto out;
 	}
diff --git a/drivers/infiniband/hw/cxgb4/device.c b/drivers/infiniband/hw/cxgb4/device.c
index 95e6c6c..d2d3dba 100644
--- a/drivers/infiniband/hw/cxgb4/device.c
+++ b/drivers/infiniband/hw/cxgb4/device.c
@@ -348,6 +348,7 @@ static int stats_show(struct seq_file *seq, void *v)
 		   dev->rdev.stats.act_ofld_conn_fails);
 	seq_printf(seq, "PAS_OFLD_CONN_FAILS: %10llu\n",
 		   dev->rdev.stats.pas_ofld_conn_fails);
+	seq_printf(seq, "AVAILABLE IRD: %10u\n", dev->avail_ird);
 	return 0;
 }
 
@@ -839,6 +840,7 @@ static struct c4iw_dev *c4iw_alloc(const struct cxgb4_lld_info *infop)
 	mutex_init(&devp->rdev.stats.lock);
 	mutex_init(&devp->db_mutex);
 	INIT_LIST_HEAD(&devp->db_fc_list);
+	devp->avail_ird = devp->rdev.lldi.max_ird_adapter;
 
 	if (c4iw_debugfs_root) {
 		devp->debugfs_root = debugfs_create_dir(
diff --git a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
index 9b9754c..75541cb 100644
--- a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
+++ b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
@@ -249,6 +249,7 @@ struct c4iw_dev {
 	struct idr atid_idr;
 	struct idr stid_idr;
 	struct list_head db_fc_list;
+	u32 avail_ird;
 };
 
 static inline struct c4iw_dev *to_c4iw_dev(struct ib_device *ibdev)
@@ -330,6 +331,13 @@ static inline void remove_handle_nolock(struct c4iw_dev *rhp,
 	_remove_handle(rhp, idr, id, 0);
 }
 
+extern uint c4iw_max_read_depth;
+
+static inline int cur_max_read_depth(struct c4iw_dev *dev)
+{
+	return min(dev->rdev.lldi.max_ordird_qp, c4iw_max_read_depth);
+}
+
 struct c4iw_pd {
 	struct ib_pd ibpd;
 	u32 pdid;
@@ -1003,7 +1011,6 @@ void c4iw_ev_dispatch(struct c4iw_dev *dev, struct t4_cqe *err_cqe);
 
 extern struct cxgb4_client t4c_client;
 extern c4iw_handler_func c4iw_handlers[NUM_CPL_CMDS];
-extern int c4iw_max_read_depth;
 extern int db_fc_threshold;
 extern int db_coalescing_threshold;
 extern int use_dsgl;
diff --git a/drivers/infiniband/hw/cxgb4/provider.c b/drivers/infiniband/hw/cxgb4/provider.c
index 1d41b92..67c4a69 100644
--- a/drivers/infiniband/hw/cxgb4/provider.c
+++ b/drivers/infiniband/hw/cxgb4/provider.c
@@ -322,8 +322,10 @@ static int c4iw_query_device(struct ib_device *ibdev,
 	props->max_qp_wr = dev->rdev.hw_queue.t4_max_qp_depth;
 	props->max_sge = T4_MAX_RECV_SGE;
 	props->max_sge_rd = 1;
-	props->max_qp_rd_atom = c4iw_max_read_depth;
-	props->max_qp_init_rd_atom = c4iw_max_read_depth;
+	props->max_res_rd_atom = dev->rdev.lldi.max_ird_adapter;
+	props->max_qp_rd_atom = min(dev->rdev.lldi.max_ordird_qp,
+				    c4iw_max_read_depth);
+	props->max_qp_init_rd_atom = props->max_qp_rd_atom;
 	props->max_cq = T4_MAX_NUM_CQ;
 	props->max_cqe = dev->rdev.hw_queue.t4_max_cq_depth;
 	props->max_mr = c4iw_num_stags(&dev->rdev);
diff --git a/drivers/infiniband/hw/cxgb4/qp.c b/drivers/infiniband/hw/cxgb4/qp.c
index 6f74e0e..c911e96 100644
--- a/drivers/infiniband/hw/cxgb4/qp.c
+++ b/drivers/infiniband/hw/cxgb4/qp.c
@@ -58,6 +58,30 @@ static int max_fr_immd = T4_MAX_FR_IMMD;
 module_param(max_fr_immd, int, 0644);
 MODULE_PARM_DESC(max_fr_immd, "fastreg threshold for using DSGL instead of immedate");
 
+static int alloc_ird(struct c4iw_dev *dev, u32 ird)
+{
+	int ret;
+
+	spin_lock_irq(&dev->lock);
+	if (ird <= dev->avail_ird) {
+		dev->avail_ird -= ird;
+		ret = 0;
+	} else {
+		ret = -ENOMEM;
+		pr_info("%s: device IRD resources exhausted\n",
+			pci_name(dev->rdev.lldi.pdev));
+	}
+	spin_unlock_irq(&dev->lock);
+	return ret;
+}
+
+static void free_ird(struct c4iw_dev *dev, int ird)
+{
+	spin_lock_irq(&dev->lock);
+	dev->avail_ird += ird;
+	spin_unlock_irq(&dev->lock);
+}
+
 static void set_state(struct c4iw_qp *qhp, enum c4iw_qp_state state)
 {
 	unsigned long flag;
@@ -1204,12 +1228,20 @@ static int rdma_init(struct c4iw_dev *rhp, struct c4iw_qp *qhp)
 	int ret;
 	struct sk_buff *skb;
 
-	PDBG("%s qhp %p qid 0x%x tid %u\n", __func__, qhp, qhp->wq.sq.qid,
-	     qhp->ep->hwtid);
+	PDBG("%s qhp %p qid 0x%x tid %u ird %u ord %u\n", __func__, qhp,
+	     qhp->wq.sq.qid, qhp->ep->hwtid, qhp->ep->ird, qhp->ep->ord);
 
 	skb = alloc_skb(sizeof *wqe, GFP_KERNEL);
-	if (!skb)
-		return -ENOMEM;
+	if (!skb) {
+		ret = -ENOMEM;
+		goto out;
+	}
+	ret = alloc_ird(rhp, qhp->attr.max_ird);
+	if (ret) {
+		qhp->attr.max_ird = 0;
+		kfree_skb(skb);
+		goto out;
+	}
 	set_wr_txq(skb, CPL_PRIORITY_DATA, qhp->ep->txq_idx);
 
 	wqe = (struct fw_ri_wr *)__skb_put(skb, sizeof(*wqe));
@@ -1260,10 +1292,14 @@ static int rdma_init(struct c4iw_dev *rhp, struct c4iw_qp *qhp)
 
 	ret = c4iw_ofld_send(&rhp->rdev, skb);
 	if (ret)
-		goto out;
+		goto err1;
 
 	ret = c4iw_wait_for_reply(&rhp->rdev, &qhp->ep->com.wr_wait,
 				  qhp->ep->hwtid, qhp->wq.sq.qid, __func__);
+	if (!ret)
+		goto out;
+err1:
+	free_ird(rhp, qhp->attr.max_ird);
 out:
 	PDBG("%s ret %d\n", __func__, ret);
 	return ret;
@@ -1308,7 +1344,7 @@ int c4iw_modify_qp(struct c4iw_dev *rhp, struct c4iw_qp *qhp,
 			newattr.max_ord = attrs->max_ord;
 		}
 		if (mask & C4IW_QP_ATTR_MAX_IRD) {
-			if (attrs->max_ird > c4iw_max_read_depth) {
+			if (attrs->max_ird > cur_max_read_depth(rhp)) {
 				ret = -EINVAL;
 				goto out;
 			}
@@ -1531,6 +1567,7 @@ int c4iw_destroy_qp(struct ib_qp *ib_qp)
 	if (!list_empty(&qhp->db_fc_entry))
 		list_del_init(&qhp->db_fc_entry);
 	spin_unlock_irq(&rhp->lock);
+	free_ird(rhp, qhp->attr.max_ird);
 
 	ucontext = ib_qp->uobject ?
 		   to_c4iw_ucontext(ib_qp->uobject->context) : NULL;
@@ -1621,8 +1658,8 @@ struct ib_qp *c4iw_create_qp(struct ib_pd *pd, struct ib_qp_init_attr *attrs,
 	qhp->attr.enable_rdma_read = 1;
 	qhp->attr.enable_rdma_write = 1;
 	qhp->attr.enable_bind = 1;
-	qhp->attr.max_ord = 1;
-	qhp->attr.max_ird = 1;
+	qhp->attr.max_ord = 0;
+	qhp->attr.max_ird = 0;
 	qhp->sq_sig_all = attrs->sq_sig_type == IB_SIGNAL_ALL_WR;
 	spin_lock_init(&qhp->lock);
 	mutex_init(&qhp->mutex);
diff --git a/drivers/net/ethernet/chelsio/cxgb4/cxgb4.h b/drivers/net/ethernet/chelsio/cxgb4/cxgb4.h
index f338a7f..4615621 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/cxgb4.h
+++ b/drivers/net/ethernet/chelsio/cxgb4/cxgb4.h
@@ -310,6 +310,9 @@ struct adapter_params {
 
 	unsigned int ofldq_wr_cred;
 	bool ulptx_memwrite_dsgl;          /* use of T5 DSGL allowed */
+
+	unsigned int max_ordird_qp;       /* Max read depth per RDMA QP */
+	unsigned int max_ird_adapter;     /* Max read depth per adapter */
 };
 
 #include "t4fw_api.h"
diff --git a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
index a7ce996..767cbba 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
+++ b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
@@ -4113,6 +4113,8 @@ static void uld_attach(struct adapter *adap, unsigned int uld)
 	lli.sge_egrstatuspagesize = adap->sge.stat_len;
 	lli.sge_pktshift = adap->sge.pktshift;
 	lli.enable_fw_ofld_conn = adap->flags & FW_OFLD_CONN;
+	lli.max_ordird_qp = adap->params.max_ordird_qp;
+	lli.max_ird_adapter = adap->params.max_ird_adapter;
 	lli.ulptx_memwrite_dsgl = adap->params.ulptx_memwrite_dsgl;
 
 	handle = ulds[uld].add(&lli);
@@ -5877,6 +5879,22 @@ static int adap_init0(struct adapter *adap)
 		adap->vres.cq.size = val[3] - val[2] + 1;
 		adap->vres.ocq.start = val[4];
 		adap->vres.ocq.size = val[5] - val[4] + 1;
+
+		params[0] = FW_PARAM_DEV(MAXORDIRD_QP);
+		params[1] = FW_PARAM_DEV(MAXIRD_ADAPTER);
+		ret = t4_query_params(adap, 0, 0, 0, 2, params, val);
+		if (ret < 0) {
+			adap->params.max_ordird_qp = 8;
+			adap->params.max_ird_adapter = 32 * adap->tids.ntids;
+			ret = 0;
+		} else {
+			adap->params.max_ordird_qp = val[0];
+			adap->params.max_ird_adapter = val[1];
+		}
+		dev_info(adap->pdev_dev,
+			 "max_ordird_qp %d max_ird_adapter %d\n",
+			 adap->params.max_ordird_qp,
+			 adap->params.max_ird_adapter);
 	}
 	if (caps_cmd.iscsicaps) {
 		params[0] = FW_PARAM_PFVF(ISCSI_START);
diff --git a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h
index 962458f..df1d944 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h
+++ b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h
@@ -258,6 +258,8 @@ struct cxgb4_lld_info {
 	unsigned int pf;		     /* Physical Function we're using */
 	bool enable_fw_ofld_conn;            /* Enable connection through fw */
 					     /* WR */
+	unsigned int max_ordird_qp;          /* Max ORD/IRD depth per RDMA QP */
+	unsigned int max_ird_adapter;        /* Max IRD memory per adapter */
 	bool ulptx_memwrite_dsgl;            /* use of T5 DSGL allowed */
 };
 
diff --git a/drivers/net/ethernet/chelsio/cxgb4/t4fw_api.h b/drivers/net/ethernet/chelsio/cxgb4/t4fw_api.h
index 4a6ae4d..ff709e3 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/t4fw_api.h
+++ b/drivers/net/ethernet/chelsio/cxgb4/t4fw_api.h
@@ -934,6 +934,8 @@ enum fw_params_param_dev {
 	FW_PARAMS_PARAM_DEV_FWREV = 0x0B,
 	FW_PARAMS_PARAM_DEV_TPREV = 0x0C,
 	FW_PARAMS_PARAM_DEV_CF = 0x0D,
+	FW_PARAMS_PARAM_DEV_MAXORDIRD_QP = 0x13, /* max supported QP IRD/ORD */
+	FW_PARAMS_PARAM_DEV_MAXIRD_ADAPTER = 0x14, /* max supported adap IRD */
 	FW_PARAMS_PARAM_DEV_ULPTX_MEMWRITE_DSGL = 0x17,
 };
 
-- 
1.7.1

Re: [PATCH 2/4] cxgb4/iw_cxgb4: use firmware ord/ird resource limits

[Yann Droneaud]

Hi,

Le vendredi 11 juillet 2014 à 20:44 +0530, Hariprasad Shenai a écrit :
> diff --git a/drivers/infiniband/hw/cxgb4/qp.c
> b/drivers/infiniband/hw/cxgb4/qp.c
> index 6f74e0e..c911e96 100644
> --- a/drivers/infiniband/hw/cxgb4/qp.c
> +++ b/drivers/infiniband/hw/cxgb4/qp.c
> @@ -58,6 +58,30 @@ static int max_fr_immd = T4_MAX_FR_IMMD;
>  module_param(max_fr_immd, int, 0644);
>  MODULE_PARM_DESC(max_fr_immd, "fastreg threshold for using DSGL
> instead of immedate");
>  
> +static int alloc_ird(struct c4iw_dev *dev, u32 ird)
> +{
> +       int ret;
> +
> +       spin_lock_irq(&dev->lock);
> +       if (ird <= dev->avail_ird) {
> +               dev->avail_ird -= ird;
> +               ret = 0;
> +       } else {
> +               ret = -ENOMEM;
> +               pr_info("%s: device IRD resources exhausted\n",
> +                       pci_name(dev->rdev.lldi.pdev));
> +       }
> +       spin_unlock_irq(&dev->lock);
> +       return ret;
> +}
> +

I would have written this function has:

static int alloc_ird(struct c4iw_dev *dev, u32 ird)
{
	int ret = 0

	spin_lock_irq(&dev->lock);
	if (ird <= dev->avail_ird)
		dev->avail_ird -= ird;
	else
		ret = -ENOMEM;
	spin_unlock_irq(&dev->lock);

	if (ret)
		dev_warn(dev->rdev.lldi.pdev,
			 "device IRD resources exhausted\n");

	return ret;
}

Regards.

-- 
Yann Droneaud
OPTEYA


--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH 3/4] cxgb4/iw_cxgb4: display TPTE on errors

[Hariprasad Shenai]

With ingress WRITE or READ RESPONSE errors, HW provides the offending
stag from the packet.  This patch adds logic to log the parsed TPTE
in this case. cxgb4 now exports a function to read a TPTE entry
from adapter memory.

Signed-off-by: Steve Wise <swise@opengridcomputing.com>
Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
---
 drivers/infiniband/hw/cxgb4/device.c            |   27 ++++++++--
 drivers/infiniband/hw/cxgb4/ev.c                |   53 ++++++++++++++++--
 drivers/infiniband/hw/cxgb4/t4.h                |    4 +-
 drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c |   66 +++++++++++++++++++++++
 drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h  |    1 +
 5 files changed, 140 insertions(+), 11 deletions(-)

diff --git a/drivers/infiniband/hw/cxgb4/device.c b/drivers/infiniband/hw/cxgb4/device.c
index d2d3dba..e8b79a3 100644
--- a/drivers/infiniband/hw/cxgb4/device.c
+++ b/drivers/infiniband/hw/cxgb4/device.c
@@ -241,12 +241,31 @@ static int dump_stag(int id, void *p, void *data)
 	struct c4iw_debugfs_data *stagd = data;
 	int space;
 	int cc;
+	struct fw_ri_tpte tpte;
+	int ret;
 
 	space = stagd->bufsize - stagd->pos - 1;
 	if (space == 0)
 		return 1;
 
-	cc = snprintf(stagd->buf + stagd->pos, space, "0x%x\n", id<<8);
+	ret = cxgb4_read_tpte(stagd->devp->rdev.lldi.ports[0], (u32)id<<8,
+			      (__be32 *)&tpte);
+	if (ret) {
+		pr_info("%s cxgb4_read_tpte err %d\n", __func__, ret);
+		return ret;
+	}
+	cc = snprintf(stagd->buf + stagd->pos, space,
+		      "stag: idx 0x%x valid %d key 0x%x state %d pdid %d "
+		      "perm 0x%x ps %d len 0x%llx va 0x%llx\n",
+		      (u32)id<<8,
+		      G_FW_RI_TPTE_VALID(ntohl(tpte.valid_to_pdid)),
+		      G_FW_RI_TPTE_STAGKEY(ntohl(tpte.valid_to_pdid)),
+		      G_FW_RI_TPTE_STAGSTATE(ntohl(tpte.valid_to_pdid)),
+		      G_FW_RI_TPTE_PDID(ntohl(tpte.valid_to_pdid)),
+		      G_FW_RI_TPTE_PERM(ntohl(tpte.locread_to_qpid)),
+		      G_FW_RI_TPTE_PS(ntohl(tpte.locread_to_qpid)),
+		      ((u64)ntohl(tpte.len_hi) << 32) | ntohl(tpte.len_lo),
+		      ((u64)ntohl(tpte.va_hi) << 32) | ntohl(tpte.va_lo_fbo));
 	if (cc < space)
 		stagd->pos += cc;
 	return 0;
@@ -259,7 +278,7 @@ static int stag_release(struct inode *inode, struct file *file)
 		printk(KERN_INFO "%s null stagd?\n", __func__);
 		return 0;
 	}
-	kfree(stagd->buf);
+	vfree(stagd->buf);
 	kfree(stagd);
 	return 0;
 }
@@ -282,8 +301,8 @@ static int stag_open(struct inode *inode, struct file *file)
 	idr_for_each(&stagd->devp->mmidr, count_idrs, &count);
 	spin_unlock_irq(&stagd->devp->lock);
 
-	stagd->bufsize = count * sizeof("0x12345678\n");
-	stagd->buf = kmalloc(stagd->bufsize, GFP_KERNEL);
+	stagd->bufsize = count * 256;
+	stagd->buf = vmalloc(stagd->bufsize);
 	if (!stagd->buf) {
 		ret = -ENOMEM;
 		goto err1;
diff --git a/drivers/infiniband/hw/cxgb4/ev.c b/drivers/infiniband/hw/cxgb4/ev.c
index d61d0a1..97379f4 100644
--- a/drivers/infiniband/hw/cxgb4/ev.c
+++ b/drivers/infiniband/hw/cxgb4/ev.c
@@ -35,6 +35,53 @@
 
 #include "iw_cxgb4.h"
 
+static void print_tpte(struct c4iw_dev *dev, u32 stag)
+{
+	int ret;
+	struct fw_ri_tpte tpte;
+
+	ret = cxgb4_read_tpte(dev->rdev.lldi.ports[0], stag,
+			      (__be32 *)&tpte);
+	if (ret) {
+		pr_err("%s cxgb4_read_tpte err %d\n", __func__, ret);
+		return;
+	}
+	pr_err("stag idx 0x%x valid %d key 0x%x state %d pdid %d "
+	       "perm 0x%x ps %d len 0x%llx va 0x%llx\n",
+	       stag & 0xffffff00,
+	       G_FW_RI_TPTE_VALID(ntohl(tpte.valid_to_pdid)),
+	       G_FW_RI_TPTE_STAGKEY(ntohl(tpte.valid_to_pdid)),
+	       G_FW_RI_TPTE_STAGSTATE(ntohl(tpte.valid_to_pdid)),
+	       G_FW_RI_TPTE_PDID(ntohl(tpte.valid_to_pdid)),
+	       G_FW_RI_TPTE_PERM(ntohl(tpte.locread_to_qpid)),
+	       G_FW_RI_TPTE_PS(ntohl(tpte.locread_to_qpid)),
+	       ((u64)ntohl(tpte.len_hi) << 32) | ntohl(tpte.len_lo),
+	       ((u64)ntohl(tpte.va_hi) << 32) | ntohl(tpte.va_lo_fbo));
+}
+
+static void dump_err_cqe(struct c4iw_dev *dev, struct t4_cqe *err_cqe)
+{
+	__be64 *p = (void *)err_cqe;
+
+	pr_err("AE qpid %d opcode %d status 0x%x "
+	       "type %d len 0x%x wrid.hi 0x%x wrid.lo 0x%x\n",
+	       CQE_QPID(err_cqe), CQE_OPCODE(err_cqe),
+	       CQE_STATUS(err_cqe), CQE_TYPE(err_cqe), ntohl(err_cqe->len),
+	       CQE_WRID_HI(err_cqe), CQE_WRID_LOW(err_cqe));
+
+	pr_err("%016llx %016llx %016llx %016llx\n",
+	       be64_to_cpu(p[0]), be64_to_cpu(p[1]), be64_to_cpu(p[2]),
+	       be64_to_cpu(p[3]));
+
+	/*
+	 * Ingress WRITE and READ_RESP errors provide
+	 * the offending stag, so parse and log it.
+	 */
+	if (RQ_TYPE(err_cqe) && (CQE_OPCODE(err_cqe) == FW_RI_RDMA_WRITE ||
+				 CQE_OPCODE(err_cqe) == FW_RI_READ_RESP))
+		print_tpte(dev, CQE_WRID_STAG(err_cqe));
+}
+
 static void post_qp_event(struct c4iw_dev *dev, struct c4iw_cq *chp,
 			  struct c4iw_qp *qhp,
 			  struct t4_cqe *err_cqe,
@@ -44,11 +91,7 @@ static void post_qp_event(struct c4iw_dev *dev, struct c4iw_cq *chp,
 	struct c4iw_qp_attributes attrs;
 	unsigned long flag;
 
-	printk(KERN_ERR MOD "AE qpid 0x%x opcode %d status 0x%x "
-	       "type %d wrid.hi 0x%x wrid.lo 0x%x\n",
-	       CQE_QPID(err_cqe), CQE_OPCODE(err_cqe),
-	       CQE_STATUS(err_cqe), CQE_TYPE(err_cqe),
-	       CQE_WRID_HI(err_cqe), CQE_WRID_LOW(err_cqe));
+	dump_err_cqe(dev, err_cqe);
 
 	if (qhp->attr.state == C4IW_QP_STATE_RTS) {
 		attrs.next_state = C4IW_QP_STATE_TERMINATE;
diff --git a/drivers/infiniband/hw/cxgb4/t4.h b/drivers/infiniband/hw/cxgb4/t4.h
index e64fa8b..dd45186 100644
--- a/drivers/infiniband/hw/cxgb4/t4.h
+++ b/drivers/infiniband/hw/cxgb4/t4.h
@@ -236,8 +236,8 @@ struct t4_cqe {
 #define CQE_WRID_SQ_IDX(x)	((x)->u.scqe.cidx)
 
 /* generic accessor macros */
-#define CQE_WRID_HI(x)		((x)->u.gen.wrid_hi)
-#define CQE_WRID_LOW(x)		((x)->u.gen.wrid_low)
+#define CQE_WRID_HI(x)		(be32_to_cpu((x)->u.gen.wrid_hi))
+#define CQE_WRID_LOW(x)		(be32_to_cpu((x)->u.gen.wrid_low))
 
 /* macros for flit 3 of the cqe */
 #define S_CQE_GENBIT	63
diff --git a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
index 767cbba..ba7d13d 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
+++ b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
@@ -3832,6 +3832,72 @@ void cxgb4_enable_db_coalescing(struct net_device *dev)
 }
 EXPORT_SYMBOL(cxgb4_enable_db_coalescing);
 
+int cxgb4_read_tpte(struct net_device *dev, u32 stag, __be32 *tpte)
+{
+	struct adapter *adap;
+	u32 offset, memtype, memaddr;
+	u32 edc0_size, edc1_size, mc0_size, mc1_size;
+	u32 edc0_end, edc1_end, mc0_end, mc1_end;
+	int ret;
+
+	adap = netdev2adap(dev);
+
+	offset = ((stag >> 8) * 32) + adap->vres.stag.start;
+
+	/* Figure out where the offset lands in the Memory Type/Address scheme.
+	 * This code assumes that the memory is laid out starting at offset 0
+	 * with no breaks as: EDC0, EDC1, MC0, MC1. All cards have both EDC0
+	 * and EDC1.  Some cards will have neither MC0 nor MC1, most cards have
+	 * MC0, and some have both MC0 and MC1.
+	 */
+	edc0_size = EDRAM_SIZE_GET(t4_read_reg(adap, MA_EDRAM0_BAR)) << 20;
+	edc1_size = EDRAM_SIZE_GET(t4_read_reg(adap, MA_EDRAM1_BAR)) << 20;
+	mc0_size = EXT_MEM_SIZE_GET(t4_read_reg(adap, MA_EXT_MEMORY_BAR)) << 20;
+
+	edc0_end = edc0_size;
+	edc1_end = edc0_end + edc1_size;
+	mc0_end = edc1_end + mc0_size;
+
+	if (offset < edc0_end) {
+		memtype = MEM_EDC0;
+		memaddr = offset;
+	} else if (offset < edc1_end) {
+		memtype = MEM_EDC1;
+		memaddr = offset - edc0_end;
+	} else {
+		if (offset < mc0_end) {
+			memtype = MEM_MC0;
+			memaddr = offset - edc1_end;
+		} else if (is_t4(adap->params.chip)) {
+			/* T4 only has a single memory channel */
+			goto err;
+		} else {
+			mc1_size = EXT_MEM_SIZE_GET(
+					t4_read_reg(adap,
+						    MA_EXT_MEMORY1_BAR)) << 20;
+			mc1_end = mc0_end + mc1_size;
+			if (offset < mc1_end) {
+				memtype = MEM_MC1;
+				memaddr = offset - mc0_end;
+			} else {
+				/* offset beyond the end of any memory */
+				goto err;
+			}
+		}
+	}
+
+	spin_lock(&adap->win0_lock);
+	ret = t4_memory_rw(adap, 0, memtype, memaddr, 32, tpte, T4_MEMORY_READ);
+	spin_unlock(&adap->win0_lock);
+	return ret;
+
+err:
+	dev_err(adap->pdev_dev, "stag %#x, offset %#x out of range\n",
+		stag, offset);
+	return -EINVAL;
+}
+EXPORT_SYMBOL(cxgb4_read_tpte);
+
 static struct pci_driver cxgb4_driver;
 
 static void check_neigh_update(struct neighbour *neigh)
diff --git a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h
index df1d944..c7170d6 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h
+++ b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h
@@ -296,5 +296,6 @@ int cxgb4_sync_txq_pidx(struct net_device *dev, u16 qid, u16 pidx, u16 size);
 int cxgb4_flush_eq_cache(struct net_device *dev);
 void cxgb4_disable_db_coalescing(struct net_device *dev);
 void cxgb4_enable_db_coalescing(struct net_device *dev);
+int cxgb4_read_tpte(struct net_device *dev, u32 stag, __be32 *tpte);
 
 #endif  /* !__CXGB4_OFLD_H */
-- 
1.7.1

Re: [PATCH 3/4] cxgb4/iw_cxgb4: display TPTE on errors

[Yann Droneaud]

Hi,

Le vendredi 11 juillet 2014 à 20:44 +0530, Hariprasad Shenai a écrit :
> With ingress WRITE or READ RESPONSE errors, HW provides the offending
> stag from the packet.  This patch adds logic to log the parsed TPTE
> in this case. cxgb4 now exports a function to read a TPTE entry
> from adapter memory.
> 
> Signed-off-by: Steve Wise <swise-7bPotxP6k4+P2YhJcF5u+vpXobYPEAuW@public.gmane.org>
> Signed-off-by: Hariprasad Shenai <hariprasad-ut6Up61K2wZBDgjK7y7TUQ@public.gmane.org>
> ---
>  drivers/infiniband/hw/cxgb4/device.c            |   27 ++++++++--
>  drivers/infiniband/hw/cxgb4/ev.c                |   53 ++++++++++++++++--
>  drivers/infiniband/hw/cxgb4/t4.h                |    4 +-
>  drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c |   66 +++++++++++++++++++++++
>  drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h  |    1 +
>  5 files changed, 140 insertions(+), 11 deletions(-)
> 
> diff --git a/drivers/infiniband/hw/cxgb4/device.c b/drivers/infiniband/hw/cxgb4/device.c
> index d2d3dba..e8b79a3 100644
> --- a/drivers/infiniband/hw/cxgb4/device.c
> +++ b/drivers/infiniband/hw/cxgb4/device.c
> @@ -241,12 +241,31 @@ static int dump_stag(int id, void *p, void *data)
>  	struct c4iw_debugfs_data *stagd = data;
>  	int space;
>  	int cc;
> +	struct fw_ri_tpte tpte;
> +	int ret;
>  
>  	space = stagd->bufsize - stagd->pos - 1;
>  	if (space == 0)
>  		return 1;
>  
> -	cc = snprintf(stagd->buf + stagd->pos, space, "0x%x\n", id<<8);
> +	ret = cxgb4_read_tpte(stagd->devp->rdev.lldi.ports[0], (u32)id<<8,
> +			      (__be32 *)&tpte);
> +	if (ret) {
> +		pr_info("%s cxgb4_read_tpte err %d\n", __func__, ret);
> +		return ret;

If it's an error, use pr_err(), and perhaps better, dev_err().

> +	}
> +	cc = snprintf(stagd->buf + stagd->pos, space,
> +		      "stag: idx 0x%x valid %d key 0x%x state %d pdid %d "
> +		      "perm 0x%x ps %d len 0x%llx va 0x%llx\n",
> +		      (u32)id<<8,
> +		      G_FW_RI_TPTE_VALID(ntohl(tpte.valid_to_pdid)),
> +		      G_FW_RI_TPTE_STAGKEY(ntohl(tpte.valid_to_pdid)),
> +		      G_FW_RI_TPTE_STAGSTATE(ntohl(tpte.valid_to_pdid)),
> +		      G_FW_RI_TPTE_PDID(ntohl(tpte.valid_to_pdid)),
> +		      G_FW_RI_TPTE_PERM(ntohl(tpte.locread_to_qpid)),
> +		      G_FW_RI_TPTE_PS(ntohl(tpte.locread_to_qpid)),
> +		      ((u64)ntohl(tpte.len_hi) << 32) | ntohl(tpte.len_lo),
> +		      ((u64)ntohl(tpte.va_hi) << 32) | ntohl(tpte.va_lo_fbo));
>  	if (cc < space)
>  		stagd->pos += cc;
>  	return 0;
> @@ -259,7 +278,7 @@ static int stag_release(struct inode *inode, struct file *file)
>  		printk(KERN_INFO "%s null stagd?\n", __func__);
>  		return 0;
>  	}
> -	kfree(stagd->buf);
> +	vfree(stagd->buf);
>  	kfree(stagd);
>  	return 0;
>  }
> @@ -282,8 +301,8 @@ static int stag_open(struct inode *inode, struct file *file)
>  	idr_for_each(&stagd->devp->mmidr, count_idrs, &count);
>  	spin_unlock_irq(&stagd->devp->lock);
>  
> -	stagd->bufsize = count * sizeof("0x12345678\n");
> -	stagd->buf = kmalloc(stagd->bufsize, GFP_KERNEL);
> +	stagd->bufsize = count * 256;
> +	stagd->buf = vmalloc(stagd->bufsize);
>  	if (!stagd->buf) {
>  		ret = -ENOMEM;
>  		goto err1;
> diff --git a/drivers/infiniband/hw/cxgb4/ev.c b/drivers/infiniband/hw/cxgb4/ev.c
> index d61d0a1..97379f4 100644
> --- a/drivers/infiniband/hw/cxgb4/ev.c
> +++ b/drivers/infiniband/hw/cxgb4/ev.c
> @@ -35,6 +35,53 @@
>  
>  #include "iw_cxgb4.h"
>  
> +static void print_tpte(struct c4iw_dev *dev, u32 stag)
> +{
> +	int ret;
> +	struct fw_ri_tpte tpte;
> +
> +	ret = cxgb4_read_tpte(dev->rdev.lldi.ports[0], stag,
> +			      (__be32 *)&tpte);
> +	if (ret) {
> +		pr_err("%s cxgb4_read_tpte err %d\n", __func__, ret);

pr_err() is used here. Perhaps dev_err() can be used here too.

> +		return;
> +	}
> +	pr_err("stag idx 0x%x valid %d key 0x%x state %d pdid %d "
> +	       "perm 0x%x ps %d len 0x%llx va 0x%llx\n",
> +	       stag & 0xffffff00,
> +	       G_FW_RI_TPTE_VALID(ntohl(tpte.valid_to_pdid)),
> +	       G_FW_RI_TPTE_STAGKEY(ntohl(tpte.valid_to_pdid)),
> +	       G_FW_RI_TPTE_STAGSTATE(ntohl(tpte.valid_to_pdid)),
> +	       G_FW_RI_TPTE_PDID(ntohl(tpte.valid_to_pdid)),
> +	       G_FW_RI_TPTE_PERM(ntohl(tpte.locread_to_qpid)),
> +	       G_FW_RI_TPTE_PS(ntohl(tpte.locread_to_qpid)),
> +	       ((u64)ntohl(tpte.len_hi) << 32) | ntohl(tpte.len_lo),
> +	       ((u64)ntohl(tpte.va_hi) << 32) | ntohl(tpte.va_lo_fbo));

That's not an error.

Perhaps it's a debug message, then use dev_dbg().

> +}
> +
> +static void dump_err_cqe(struct c4iw_dev *dev, struct t4_cqe *err_cqe)
> +{
> +	__be64 *p = (void *)err_cqe;
> +
> +	pr_err("AE qpid %d opcode %d status 0x%x "
> +	       "type %d len 0x%x wrid.hi 0x%x wrid.lo 0x%x\n",
> +	       CQE_QPID(err_cqe), CQE_OPCODE(err_cqe),
> +	       CQE_STATUS(err_cqe), CQE_TYPE(err_cqe), ntohl(err_cqe->len),
> +	       CQE_WRID_HI(err_cqe), CQE_WRID_LOW(err_cqe));
> +

You could use dev_err().

> +	pr_err("%016llx %016llx %016llx %016llx\n",
> +	       be64_to_cpu(p[0]), be64_to_cpu(p[1]), be64_to_cpu(p[2]),
> +	       be64_to_cpu(p[3]));
> +

Is it really required to do a "raw dump" of the err_cqe content ?
It looks like a debug message, so use dev_dbg().

> +	/*
> +	 * Ingress WRITE and READ_RESP errors provide
> +	 * the offending stag, so parse and log it.
> +	 */
> +	if (RQ_TYPE(err_cqe) && (CQE_OPCODE(err_cqe) == FW_RI_RDMA_WRITE ||
> +				 CQE_OPCODE(err_cqe) == FW_RI_READ_RESP))
> +		print_tpte(dev, CQE_WRID_STAG(err_cqe));
> +}
> +
>  static void post_qp_event(struct c4iw_dev *dev, struct c4iw_cq *chp,
>  			  struct c4iw_qp *qhp,
>  			  struct t4_cqe *err_cqe,
> @@ -44,11 +91,7 @@ static void post_qp_event(struct c4iw_dev *dev, struct c4iw_cq *chp,
>  	struct c4iw_qp_attributes attrs;
>  	unsigned long flag;
>  
> -	printk(KERN_ERR MOD "AE qpid 0x%x opcode %d status 0x%x "
> -	       "type %d wrid.hi 0x%x wrid.lo 0x%x\n",
> -	       CQE_QPID(err_cqe), CQE_OPCODE(err_cqe),
> -	       CQE_STATUS(err_cqe), CQE_TYPE(err_cqe),
> -	       CQE_WRID_HI(err_cqe), CQE_WRID_LOW(err_cqe));
> +	dump_err_cqe(dev, err_cqe);
>  
>  	if (qhp->attr.state == C4IW_QP_STATE_RTS) {
>  		attrs.next_state = C4IW_QP_STATE_TERMINATE;
> diff --git a/drivers/infiniband/hw/cxgb4/t4.h b/drivers/infiniband/hw/cxgb4/t4.h
> index e64fa8b..dd45186 100644
> --- a/drivers/infiniband/hw/cxgb4/t4.h
> +++ b/drivers/infiniband/hw/cxgb4/t4.h
> @@ -236,8 +236,8 @@ struct t4_cqe {
>  #define CQE_WRID_SQ_IDX(x)	((x)->u.scqe.cidx)
>  
>  /* generic accessor macros */
> -#define CQE_WRID_HI(x)		((x)->u.gen.wrid_hi)
> -#define CQE_WRID_LOW(x)		((x)->u.gen.wrid_low)
> +#define CQE_WRID_HI(x)		(be32_to_cpu((x)->u.gen.wrid_hi))
> +#define CQE_WRID_LOW(x)		(be32_to_cpu((x)->u.gen.wrid_low))

Is these accessor macros used elsewhere ?

If yes, changing the endianness should be taken in account.

Perhaps this kink of change should be put in its own commit.

Regards.

-- 
Yann Droneaud
OPTEYA



--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH 1/4] iw_cxgb4: Detect Ing. Padding Boundary at run-time

[Hariprasad Shenai]

Updates iw_cxgb4 to determine the Ingress Padding Boundary from
cxgb4_lld_info, and take subsequent actions.

Signed-off-by: Steve Wise <swise-7bPotxP6k4+P2YhJcF5u+vpXobYPEAuW@public.gmane.org>
Signed-off-by: Hariprasad Shenai <hariprasad-ut6Up61K2wZBDgjK7y7TUQ@public.gmane.org>
---
 drivers/infiniband/hw/cxgb4/cq.c                |    4 ++--
 drivers/infiniband/hw/cxgb4/device.c            |   21 +++++++++++++++++++++
 drivers/infiniband/hw/cxgb4/iw_cxgb4.h          |   12 ++++++++++++
 drivers/infiniband/hw/cxgb4/provider.c          |    4 ++--
 drivers/infiniband/hw/cxgb4/qp.c                |   10 ++++++----
 drivers/infiniband/hw/cxgb4/t4.h                |    8 --------
 drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c |    2 ++
 drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h  |    2 ++
 8 files changed, 47 insertions(+), 16 deletions(-)

diff --git a/drivers/infiniband/hw/cxgb4/cq.c b/drivers/infiniband/hw/cxgb4/cq.c
index c04292c..f04a838 100644
--- a/drivers/infiniband/hw/cxgb4/cq.c
+++ b/drivers/infiniband/hw/cxgb4/cq.c
@@ -895,7 +895,7 @@ struct ib_cq *c4iw_create_cq(struct ib_device *ibdev, int entries,
 	/*
 	 * Make actual HW queue 2x to avoid cdix_inc overflows.
 	 */
-	hwentries = min(entries * 2, T4_MAX_IQ_SIZE);
+	hwentries = min(entries * 2, rhp->rdev.hw_queue.t4_max_iq_size);
 
 	/*
 	 * Make HW queue at least 64 entries so GTS updates aren't too
@@ -912,7 +912,7 @@ struct ib_cq *c4iw_create_cq(struct ib_device *ibdev, int entries,
 	if (ucontext) {
 		memsize = roundup(memsize, PAGE_SIZE);
 		hwentries = memsize / sizeof *chp->cq.queue;
-		while (hwentries > T4_MAX_IQ_SIZE) {
+		while (hwentries > rhp->rdev.hw_queue.t4_max_iq_size) {
 			memsize -= PAGE_SIZE;
 			hwentries = memsize / sizeof *chp->cq.queue;
 		}
diff --git a/drivers/infiniband/hw/cxgb4/device.c b/drivers/infiniband/hw/cxgb4/device.c
index dd93aad..95e6c6c 100644
--- a/drivers/infiniband/hw/cxgb4/device.c
+++ b/drivers/infiniband/hw/cxgb4/device.c
@@ -768,6 +768,27 @@ static struct c4iw_dev *c4iw_alloc(const struct cxgb4_lld_info *infop)
 	}
 	devp->rdev.lldi = *infop;
 
+	/* init various hw-queue params based on lld info */
+	pr_info("%s: ing. padding boundary is %d, egrsstatuspagesize = %d\n",
+		__func__, devp->rdev.lldi.sge_ingpadboundary,
+		devp->rdev.lldi.sge_egrstatuspagesize);
+
+	devp->rdev.hw_queue.t4_eq_status_entries =
+		devp->rdev.lldi.sge_ingpadboundary > 64 ? 2 : 1;
+	devp->rdev.hw_queue.t4_max_eq_size =
+		65520 - devp->rdev.hw_queue.t4_eq_status_entries;
+	devp->rdev.hw_queue.t4_max_iq_size = 65520 - 1;
+	devp->rdev.hw_queue.t4_max_rq_size =
+		8192 - devp->rdev.hw_queue.t4_eq_status_entries;
+	devp->rdev.hw_queue.t4_max_sq_size =
+		devp->rdev.hw_queue.t4_max_eq_size - 1;
+	devp->rdev.hw_queue.t4_max_qp_depth =
+		devp->rdev.hw_queue.t4_max_rq_size - 1;
+	devp->rdev.hw_queue.t4_max_cq_depth =
+		devp->rdev.hw_queue.t4_max_iq_size - 1;
+	devp->rdev.hw_queue.t4_stat_len =
+		devp->rdev.lldi.sge_egrstatuspagesize;
+
 	/*
 	 * For T5 devices, we map all of BAR2 with WC.
 	 * For T4 devices with onchip qp mem, we map only that part
diff --git a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
index 125bc5d..9b9754c 100644
--- a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
+++ b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
@@ -139,6 +139,17 @@ struct c4iw_stats {
 	u64  pas_ofld_conn_fails;
 };
 
+struct c4iw_hw_queue {
+	int t4_eq_status_entries;
+	int t4_max_eq_size;
+	int t4_max_iq_size;
+	int t4_max_rq_size;
+	int t4_max_sq_size;
+	int t4_max_qp_depth;
+	int t4_max_cq_depth;
+	int t4_stat_len;
+};
+
 struct c4iw_rdev {
 	struct c4iw_resource resource;
 	unsigned long qpshift;
@@ -156,6 +167,7 @@ struct c4iw_rdev {
 	unsigned long oc_mw_pa;
 	void __iomem *oc_mw_kva;
 	struct c4iw_stats stats;
+	struct c4iw_hw_queue hw_queue;
 	struct t4_dev_status_page *status_page;
 };
 
diff --git a/drivers/infiniband/hw/cxgb4/provider.c b/drivers/infiniband/hw/cxgb4/provider.c
index b1d3053..1d41b92 100644
--- a/drivers/infiniband/hw/cxgb4/provider.c
+++ b/drivers/infiniband/hw/cxgb4/provider.c
@@ -319,13 +319,13 @@ static int c4iw_query_device(struct ib_device *ibdev,
 	props->vendor_part_id = (u32)dev->rdev.lldi.pdev->device;
 	props->max_mr_size = T4_MAX_MR_SIZE;
 	props->max_qp = T4_MAX_NUM_QP;
-	props->max_qp_wr = T4_MAX_QP_DEPTH;
+	props->max_qp_wr = dev->rdev.hw_queue.t4_max_qp_depth;
 	props->max_sge = T4_MAX_RECV_SGE;
 	props->max_sge_rd = 1;
 	props->max_qp_rd_atom = c4iw_max_read_depth;
 	props->max_qp_init_rd_atom = c4iw_max_read_depth;
 	props->max_cq = T4_MAX_NUM_CQ;
-	props->max_cqe = T4_MAX_CQ_DEPTH;
+	props->max_cqe = dev->rdev.hw_queue.t4_max_cq_depth;
 	props->max_mr = c4iw_num_stags(&dev->rdev);
 	props->max_pd = T4_MAX_NUM_PD;
 	props->local_ca_ack_delay = 0;
diff --git a/drivers/infiniband/hw/cxgb4/qp.c b/drivers/infiniband/hw/cxgb4/qp.c
index 086f62f..6f74e0e 100644
--- a/drivers/infiniband/hw/cxgb4/qp.c
+++ b/drivers/infiniband/hw/cxgb4/qp.c
@@ -258,7 +258,8 @@ static int create_qp(struct c4iw_rdev *rdev, struct t4_wq *wq,
 	/*
 	 * eqsize is the number of 64B entries plus the status page size.
 	 */
-	eqsize = wq->sq.size * T4_SQ_NUM_SLOTS + T4_EQ_STATUS_ENTRIES;
+	eqsize = wq->sq.size * T4_SQ_NUM_SLOTS +
+		rdev->hw_queue.t4_eq_status_entries;
 
 	res->u.sqrq.fetchszm_to_iqid = cpu_to_be32(
 		V_FW_RI_RES_WR_HOSTFCMODE(0) |	/* no host cidx updates */
@@ -283,7 +284,8 @@ static int create_qp(struct c4iw_rdev *rdev, struct t4_wq *wq,
 	/*
 	 * eqsize is the number of 64B entries plus the status page size.
 	 */
-	eqsize = wq->rq.size * T4_RQ_NUM_SLOTS + T4_EQ_STATUS_ENTRIES;
+	eqsize = wq->rq.size * T4_RQ_NUM_SLOTS +
+		rdev->hw_queue.t4_eq_status_entries;
 	res->u.sqrq.fetchszm_to_iqid = cpu_to_be32(
 		V_FW_RI_RES_WR_HOSTFCMODE(0) |	/* no host cidx updates */
 		V_FW_RI_RES_WR_CPRIO(0) |	/* don't keep in chip cache */
@@ -1570,11 +1572,11 @@ struct ib_qp *c4iw_create_qp(struct ib_pd *pd, struct ib_qp_init_attr *attrs,
 		return ERR_PTR(-EINVAL);
 
 	rqsize = roundup(attrs->cap.max_recv_wr + 1, 16);
-	if (rqsize > T4_MAX_RQ_SIZE)
+	if (rqsize > rhp->rdev.hw_queue.t4_max_rq_size)
 		return ERR_PTR(-E2BIG);
 
 	sqsize = roundup(attrs->cap.max_send_wr + 1, 16);
-	if (sqsize > T4_MAX_SQ_SIZE)
+	if (sqsize > rhp->rdev.hw_queue.t4_max_sq_size)
 		return ERR_PTR(-E2BIG);
 
 	ucontext = pd->uobject ? to_c4iw_ucontext(pd->uobject->context) : NULL;
diff --git a/drivers/infiniband/hw/cxgb4/t4.h b/drivers/infiniband/hw/cxgb4/t4.h
index 68b0a6b..e64fa8b 100644
--- a/drivers/infiniband/hw/cxgb4/t4.h
+++ b/drivers/infiniband/hw/cxgb4/t4.h
@@ -39,19 +39,11 @@
 #define T4_MAX_NUM_QP 65536
 #define T4_MAX_NUM_CQ 65536
 #define T4_MAX_NUM_PD 65536
-#define T4_EQ_STATUS_ENTRIES (L1_CACHE_BYTES > 64 ? 2 : 1)
-#define T4_MAX_EQ_SIZE (65520 - T4_EQ_STATUS_ENTRIES)
-#define T4_MAX_IQ_SIZE (65520 - 1)
-#define T4_MAX_RQ_SIZE (8192 - T4_EQ_STATUS_ENTRIES)
-#define T4_MAX_SQ_SIZE (T4_MAX_EQ_SIZE - 1)
-#define T4_MAX_QP_DEPTH (T4_MAX_RQ_SIZE - 1)
-#define T4_MAX_CQ_DEPTH (T4_MAX_IQ_SIZE - 1)
 #define T4_MAX_NUM_STAG (1<<15)
 #define T4_MAX_MR_SIZE (~0ULL)
 #define T4_PAGESIZE_MASK 0xffff000  /* 4KB-128MB */
 #define T4_STAG_UNSET 0xffffffff
 #define T4_FW_MAJ 0
-#define T4_EQ_STATUS_ENTRIES (L1_CACHE_BYTES > 64 ? 2 : 1)
 #define A_PCIE_MA_SYNC 0x30b4
 
 struct t4_status_page {
diff --git a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
index 2b438bd..a7ce996 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
+++ b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
@@ -4109,6 +4109,8 @@ static void uld_attach(struct adapter *adap, unsigned int uld)
 	lli.db_reg = adap->regs + MYPF_REG(SGE_PF_KDOORBELL);
 	lli.fw_vers = adap->params.fw_vers;
 	lli.dbfifo_int_thresh = dbfifo_int_thresh;
+	lli.sge_ingpadboundary = adap->sge.fl_align;
+	lli.sge_egrstatuspagesize = adap->sge.stat_len;
 	lli.sge_pktshift = adap->sge.pktshift;
 	lli.enable_fw_ofld_conn = adap->flags & FW_OFLD_CONN;
 	lli.ulptx_memwrite_dsgl = adap->params.ulptx_memwrite_dsgl;
diff --git a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h
index 8f60851..962458f 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h
+++ b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h
@@ -251,6 +251,8 @@ struct cxgb4_lld_info {
 	void __iomem *gts_reg;               /* address of GTS register */
 	void __iomem *db_reg;                /* address of kernel doorbell */
 	int dbfifo_int_thresh;		     /* doorbell fifo int threshold */
+	unsigned int sge_ingpadboundary;     /* SGE ingress padding boundary */
+	unsigned int sge_egrstatuspagesize;  /* SGE egress status page size */
 	unsigned int sge_pktshift;           /* Padding between CPL and */
 					     /*	packet data */
 	unsigned int pf;		     /* Physical Function we're using */
-- 
1.7.1

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH 1/4] iw_cxgb4: Detect Ing. Padding Boundary at run-time

[Yann Droneaud]

Hi,

Le vendredi 11 juillet 2014 à 20:44 +0530, Hariprasad Shenai a écrit :
> Updates iw_cxgb4 to determine the Ingress Padding Boundary from
> cxgb4_lld_info, and take subsequent actions.
> 
> Signed-off-by: Steve Wise <swise@opengridcomputing.com>
> Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
> ---
>  drivers/infiniband/hw/cxgb4/cq.c                |    4 ++--
>  drivers/infiniband/hw/cxgb4/device.c            |   21 +++++++++++++++++++++
>  drivers/infiniband/hw/cxgb4/iw_cxgb4.h          |   12 ++++++++++++
>  drivers/infiniband/hw/cxgb4/provider.c          |    4 ++--
>  drivers/infiniband/hw/cxgb4/qp.c                |   10 ++++++----
>  drivers/infiniband/hw/cxgb4/t4.h                |    8 --------
>  drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c |    2 ++
>  drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h  |    2 ++
>  8 files changed, 47 insertions(+), 16 deletions(-)
> 

[...]

> diff --git a/drivers/infiniband/hw/cxgb4/device.c b/drivers/infiniband/hw/cxgb4/device.c
> index dd93aad..95e6c6c 100644
> --- a/drivers/infiniband/hw/cxgb4/device.c
> +++ b/drivers/infiniband/hw/cxgb4/device.c
> @@ -768,6 +768,27 @@ static struct c4iw_dev *c4iw_alloc(const struct cxgb4_lld_info *infop)
>  	}
>  	devp->rdev.lldi = *infop;
>  
> +	/* init various hw-queue params based on lld info */
> +	pr_info("%s: ing. padding boundary is %d, egrsstatuspagesize = %d\n",
> +		__func__, devp->rdev.lldi.sge_ingpadboundary,
> +		devp->rdev.lldi.sge_egrstatuspagesize);

Is this really needed ?

Use dev_info() or perhaps dev_dbg() instead of pr_info().

Regards.

-- 
Yann Droneaud
OPTEYA

[PATCH 4/4] cxgb4/iw_cxgb4: work request logging feature

[Hariprasad Shenai]

This commit enhances the iwarp driver to optionally keep a log of rdma
work request timining data for kernel mode QPs.  If iw_cxgb4 module option
c4iw_wr_log is set to non-zero, each work request is tracked and timing
data maintained in a rolling log that is 4096 entries deep by default.
Module option c4iw_wr_log_size_order allows specifing a log2 size to use
instead of the default order of 12 (4096 entries). Both module options
are read-only and must be passed in at module load time to set them. IE:

modprobe iw_cxgb4 c4iw_wr_log=1 c4iw_wr_log_size_order=10

The timing data is viewable via the iw_cxgb4 debugfs file "wr_log".
Writing anything to this file will clear all the timing data.
Data tracked includes:

- The host time when the work request was posted, just before ringing
the doorbell.  The host time when the completion was polled by the
application.  This is also the time the log entry is created.  The delta
of these two times is the amount of time took processing the work request.

- The qid of the EQ used to post the work request.

- The work request opcode.

- The cqe wr_id field.  For sq completions requests this is the swsqe
index.  For recv completions this is the MSN of the ingress SEND.
This value can be used to match log entries from this log with firmware
flowc event entries.

- The sge timestamp value just before ringing the doorbell when
posting,  the sge timestamp value just after polling the completion,
and CQE.timestamp field from the completion itself.  With these three
timestamps we can track the latency from post to poll, and the amount
of time the completion resided in the CQ before being reaped by the
application.  With debug firmware, the sge timestamp is also logged by
firmware in its flowc history so that we can compute the latency from
posting the work request until the firmware sees it.

Signed-off-by: Steve Wise <swise-7bPotxP6k4+P2YhJcF5u+vpXobYPEAuW@public.gmane.org>
Signed-off-by: Hariprasad Shenai <hariprasad-ut6Up61K2wZBDgjK7y7TUQ@public.gmane.org>
---
 drivers/infiniband/hw/cxgb4/cq.c                |    4 +
 drivers/infiniband/hw/cxgb4/device.c            |  137 +++++++++++++++++++++++
 drivers/infiniband/hw/cxgb4/iw_cxgb4.h          |   17 +++
 drivers/infiniband/hw/cxgb4/qp.c                |   12 ++
 drivers/infiniband/hw/cxgb4/t4.h                |    4 +
 drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c |   14 +++
 drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h  |    2 +
 drivers/net/ethernet/chelsio/cxgb4/t4_regs.h    |    6 +
 8 files changed, 196 insertions(+), 0 deletions(-)

diff --git a/drivers/infiniband/hw/cxgb4/cq.c b/drivers/infiniband/hw/cxgb4/cq.c
index f04a838..de9bcf2 100644
--- a/drivers/infiniband/hw/cxgb4/cq.c
+++ b/drivers/infiniband/hw/cxgb4/cq.c
@@ -633,11 +633,15 @@ proc_cqe:
 		wq->sq.cidx = (uint16_t)idx;
 		PDBG("%s completing sq idx %u\n", __func__, wq->sq.cidx);
 		*cookie = wq->sq.sw_sq[wq->sq.cidx].wr_id;
+		if (c4iw_wr_log)
+			c4iw_log_wr_stats(wq, hw_cqe);
 		t4_sq_consume(wq);
 	} else {
 		PDBG("%s completing rq idx %u\n", __func__, wq->rq.cidx);
 		*cookie = wq->rq.sw_rq[wq->rq.cidx].wr_id;
 		BUG_ON(t4_rq_empty(wq));
+		if (c4iw_wr_log)
+			c4iw_log_wr_stats(wq, hw_cqe);
 		t4_rq_consume(wq);
 		goto skip_cqe;
 	}
diff --git a/drivers/infiniband/hw/cxgb4/device.c b/drivers/infiniband/hw/cxgb4/device.c
index e8b79a3..f070b08 100644
--- a/drivers/infiniband/hw/cxgb4/device.c
+++ b/drivers/infiniband/hw/cxgb4/device.c
@@ -55,6 +55,15 @@ module_param(allow_db_coalescing_on_t5, int, 0644);
 MODULE_PARM_DESC(allow_db_coalescing_on_t5,
 		 "Allow DB Coalescing on T5 (default = 0)");
 
+int c4iw_wr_log = 0;
+module_param(c4iw_wr_log, int, 0444);
+MODULE_PARM_DESC(c4iw_wr_log, "Enables logging of work request timing data.");
+
+int c4iw_wr_log_size_order = 12;
+module_param(c4iw_wr_log_size_order, int, 0444);
+MODULE_PARM_DESC(c4iw_wr_log_size_order,
+		 "Number of entries (log2) in the work request timing log.");
+
 struct uld_ctx {
 	struct list_head entry;
 	struct cxgb4_lld_info lldi;
@@ -103,6 +112,117 @@ static ssize_t debugfs_read(struct file *file, char __user *buf, size_t count,
 	return simple_read_from_buffer(buf, count, ppos, d->buf, d->pos);
 }
 
+void c4iw_log_wr_stats(struct t4_wq *wq, struct t4_cqe *cqe)
+{
+	struct wr_log_entry le;
+	int idx;
+
+	if (!wq->rdev->wr_log)
+		return;
+
+	idx = (atomic_inc_return(&wq->rdev->wr_log_idx) - 1) &
+		(wq->rdev->wr_log_size - 1);
+	le.poll_sge_ts = cxgb4_read_sge_timestamp(wq->rdev->lldi.ports[0]);
+	getnstimeofday(&le.poll_host_ts);
+	le.valid = 1;
+	le.cqe_sge_ts = CQE_TS(cqe);
+	if (SQ_TYPE(cqe)) {
+		le.qid = wq->sq.qid;
+		le.opcode = CQE_OPCODE(cqe);
+		le.post_host_ts = wq->sq.sw_sq[wq->sq.cidx].host_ts;
+		le.post_sge_ts = wq->sq.sw_sq[wq->sq.cidx].sge_ts;
+		le.wr_id = CQE_WRID_SQ_IDX(cqe);
+	} else {
+		le.qid = wq->rq.qid;
+		le.opcode = FW_RI_RECEIVE;
+		le.post_host_ts = wq->rq.sw_rq[wq->rq.cidx].host_ts;
+		le.post_sge_ts = wq->rq.sw_rq[wq->rq.cidx].sge_ts;
+		le.wr_id = CQE_WRID_MSN(cqe);
+	}
+	wq->rdev->wr_log[idx] = le;
+}
+
+static int wr_log_show(struct seq_file *seq, void *v)
+{
+	struct c4iw_dev *dev = seq->private;
+	struct timespec prev_ts = {0, 0};
+	struct wr_log_entry *lep;
+	int prev_ts_set = 0;
+	int idx, end;
+
+#define ts2ns(ts) ((ts) * dev->rdev.lldi.cclk_ps / 1000)
+
+	idx = atomic_read(&dev->rdev.wr_log_idx) &
+		(dev->rdev.wr_log_size - 1);
+	end = idx - 1;
+	if (end < 0)
+		end = dev->rdev.wr_log_size - 1;
+	lep = &dev->rdev.wr_log[idx];
+	while (idx != end) {
+		if (lep->valid) {
+			if (!prev_ts_set) {
+				prev_ts_set = 1;
+				prev_ts = lep->poll_host_ts;
+			}
+			seq_printf(seq, "%04u: sec %lu nsec %lu qid %u opcode "
+				   "%u %s 0x%x host_wr_delta sec %lu nsec %lu "
+				   "post_sge_ts 0x%llx cqe_sge_ts 0x%llx "
+				   "poll_sge_ts 0x%llx post_poll_delta_ns %llu "
+				   "cqe_poll_delta_ns %llu\n",
+				   idx,
+				   timespec_sub(lep->poll_host_ts,
+						prev_ts).tv_sec,
+				   timespec_sub(lep->poll_host_ts,
+						prev_ts).tv_nsec,
+				   lep->qid, lep->opcode,
+				   lep->opcode == FW_RI_RECEIVE ?
+							"msn" : "wrid",
+				   lep->wr_id,
+				   timespec_sub(lep->poll_host_ts,
+						lep->post_host_ts).tv_sec,
+				   timespec_sub(lep->poll_host_ts,
+						lep->post_host_ts).tv_nsec,
+				   lep->post_sge_ts, lep->cqe_sge_ts,
+				   lep->poll_sge_ts,
+				   ts2ns(lep->poll_sge_ts - lep->post_sge_ts),
+				   ts2ns(lep->poll_sge_ts - lep->cqe_sge_ts));
+			prev_ts = lep->poll_host_ts;
+		}
+		idx++;
+		if (idx > (dev->rdev.wr_log_size - 1))
+			idx = 0;
+		lep = &dev->rdev.wr_log[idx];
+	}
+#undef ts2ns
+	return 0;
+}
+
+static int wr_log_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, wr_log_show, inode->i_private);
+}
+
+static ssize_t wr_log_clear(struct file *file, const char __user *buf,
+			    size_t count, loff_t *pos)
+{
+	struct c4iw_dev *dev = ((struct seq_file *)file->private_data)->private;
+	int i;
+
+	if (dev->rdev.wr_log)
+		for (i = 0; i < dev->rdev.wr_log_size; i++)
+			dev->rdev.wr_log[i].valid = 0;
+	return count;
+}
+
+static const struct file_operations wr_log_debugfs_fops = {
+	.owner   = THIS_MODULE,
+	.open    = wr_log_open,
+	.release = single_release,
+	.read    = seq_read,
+	.llseek  = seq_lseek,
+	.write   = wr_log_clear,
+};
+
 static int dump_qp(int id, void *p, void *data)
 {
 	struct c4iw_qp *qp = p;
@@ -603,6 +723,12 @@ static int setup_debugfs(struct c4iw_dev *devp)
 	if (de && de->d_inode)
 		de->d_inode->i_size = 4096;
 
+	if (c4iw_wr_log) {
+		de = debugfs_create_file("wr_log", S_IWUSR, devp->debugfs_root,
+					 (void *)devp, &wr_log_debugfs_fops);
+		if (de && de->d_inode)
+			de->d_inode->i_size = 4096;
+	}
 	return 0;
 }
 
@@ -716,6 +842,16 @@ static int c4iw_rdev_open(struct c4iw_rdev *rdev)
 		pr_err(MOD "error allocating status page\n");
 		goto err4;
 	}
+	if (c4iw_wr_log) {
+		rdev->wr_log = kzalloc((1 << c4iw_wr_log_size_order) *
+				       sizeof(*rdev->wr_log), GFP_KERNEL);
+		if (rdev->wr_log) {
+			rdev->wr_log_size = 1 << c4iw_wr_log_size_order;
+			atomic_set(&rdev->wr_log_idx, 0);
+		} else {
+			pr_err(MOD "error allocating wr_log. Logging disabled\n");
+		}
+	}
 	return 0;
 err4:
 	c4iw_rqtpool_destroy(rdev);
@@ -729,6 +865,7 @@ err1:
 
 static void c4iw_rdev_close(struct c4iw_rdev *rdev)
 {
+	kfree(rdev->wr_log);
 	free_page((unsigned long)rdev->status_page);
 	c4iw_pblpool_destroy(rdev);
 	c4iw_rqtpool_destroy(rdev);
diff --git a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
index 75541cb..69f047c 100644
--- a/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
+++ b/drivers/infiniband/hw/cxgb4/iw_cxgb4.h
@@ -150,6 +150,18 @@ struct c4iw_hw_queue {
 	int t4_stat_len;
 };
 
+struct wr_log_entry {
+	struct timespec post_host_ts;
+	struct timespec poll_host_ts;
+	u64 post_sge_ts;
+	u64 cqe_sge_ts;
+	u64 poll_sge_ts;
+	u16 qid;
+	u16 wr_id;
+	u8 opcode;
+	u8 valid;
+};
+
 struct c4iw_rdev {
 	struct c4iw_resource resource;
 	unsigned long qpshift;
@@ -169,6 +181,9 @@ struct c4iw_rdev {
 	struct c4iw_stats stats;
 	struct c4iw_hw_queue hw_queue;
 	struct t4_dev_status_page *status_page;
+	atomic_t wr_log_idx;
+	struct wr_log_entry *wr_log;
+	int wr_log_size;
 };
 
 static inline int c4iw_fatal_error(struct c4iw_rdev *rdev)
@@ -1011,6 +1026,8 @@ void c4iw_ev_dispatch(struct c4iw_dev *dev, struct t4_cqe *err_cqe);
 
 extern struct cxgb4_client t4c_client;
 extern c4iw_handler_func c4iw_handlers[NUM_CPL_CMDS];
+extern void c4iw_log_wr_stats(struct t4_wq *wq, struct t4_cqe *cqe);
+extern int c4iw_wr_log;
 extern int db_fc_threshold;
 extern int db_coalescing_threshold;
 extern int use_dsgl;
diff --git a/drivers/infiniband/hw/cxgb4/qp.c b/drivers/infiniband/hw/cxgb4/qp.c
index c911e96..0495b63 100644
--- a/drivers/infiniband/hw/cxgb4/qp.c
+++ b/drivers/infiniband/hw/cxgb4/qp.c
@@ -822,6 +822,11 @@ int c4iw_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 				  qhp->sq_sig_all;
 		swsqe->flushed = 0;
 		swsqe->wr_id = wr->wr_id;
+		if (c4iw_wr_log) {
+			swsqe->sge_ts = cxgb4_read_sge_timestamp(
+					qhp->rhp->rdev.lldi.ports[0]);
+			getnstimeofday(&swsqe->host_ts);
+		}
 
 		init_wr_hdr(wqe, qhp->wq.sq.pidx, fw_opcode, fw_flags, len16);
 
@@ -885,6 +890,13 @@ int c4iw_post_receive(struct ib_qp *ibqp, struct ib_recv_wr *wr,
 		}
 
 		qhp->wq.rq.sw_rq[qhp->wq.rq.pidx].wr_id = wr->wr_id;
+		if (c4iw_wr_log) {
+			qhp->wq.rq.sw_rq[qhp->wq.rq.pidx].sge_ts =
+				cxgb4_read_sge_timestamp(
+						qhp->rhp->rdev.lldi.ports[0]);
+			getnstimeofday(
+				&qhp->wq.rq.sw_rq[qhp->wq.rq.pidx].host_ts);
+		}
 
 		wqe->recv.opcode = FW_RI_RECV_WR;
 		wqe->recv.r1 = 0;
diff --git a/drivers/infiniband/hw/cxgb4/t4.h b/drivers/infiniband/hw/cxgb4/t4.h
index dd45186..c9f7034 100644
--- a/drivers/infiniband/hw/cxgb4/t4.h
+++ b/drivers/infiniband/hw/cxgb4/t4.h
@@ -269,6 +269,8 @@ struct t4_swsqe {
 	int			signaled;
 	u16			idx;
 	int                     flushed;
+	struct timespec         host_ts;
+	u64                     sge_ts;
 };
 
 static inline pgprot_t t4_pgprot_wc(pgprot_t prot)
@@ -306,6 +308,8 @@ struct t4_sq {
 
 struct t4_swrqe {
 	u64 wr_id;
+	struct timespec host_ts;
+	u64 sge_ts;
 };
 
 struct t4_rq {
diff --git a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
index ba7d13d..9c7e4f0 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
+++ b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
@@ -3898,6 +3898,19 @@ err:
 }
 EXPORT_SYMBOL(cxgb4_read_tpte);
 
+u64 cxgb4_read_sge_timestamp(struct net_device *dev)
+{
+	u32 hi, lo;
+	struct adapter *adap;
+
+	adap = netdev2adap(dev);
+	lo = t4_read_reg(adap, SGE_TIMESTAMP_LO);
+	hi = GET_TSVAL(t4_read_reg(adap, SGE_TIMESTAMP_HI));
+
+	return ((u64)hi << 32) | (u64)lo;
+}
+EXPORT_SYMBOL(cxgb4_read_sge_timestamp);
+
 static struct pci_driver cxgb4_driver;
 
 static void check_neigh_update(struct neighbour *neigh)
@@ -4161,6 +4174,7 @@ static void uld_attach(struct adapter *adap, unsigned int uld)
 	lli.wr_cred = adap->params.ofldq_wr_cred;
 	lli.adapter_type = adap->params.chip;
 	lli.iscsi_iolen = MAXRXDATA_GET(t4_read_reg(adap, TP_PARA_REG2));
+	lli.cclk_ps = 1000000000 / adap->params.vpd.cclk;
 	lli.udb_density = 1 << QUEUESPERPAGEPF0_GET(
 			t4_read_reg(adap, SGE_EGRESS_QUEUES_PER_PAGE_PF) >>
 			(adap->fn * 4));
diff --git a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h
index c7170d6..79a84de 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h
+++ b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_uld.h
@@ -243,6 +243,7 @@ struct cxgb4_lld_info {
 	unsigned char fw_api_ver;            /* FW API version */
 	unsigned int fw_vers;                /* FW version */
 	unsigned int iscsi_iolen;            /* iSCSI max I/O length */
+	unsigned int cclk_ps;                /* Core clock period in psec */
 	unsigned short udb_density;          /* # of user DB/page */
 	unsigned short ucq_density;          /* # of user CQs/page */
 	unsigned short filt_mode;            /* filter optional components */
@@ -297,5 +298,6 @@ int cxgb4_flush_eq_cache(struct net_device *dev);
 void cxgb4_disable_db_coalescing(struct net_device *dev);
 void cxgb4_enable_db_coalescing(struct net_device *dev);
 int cxgb4_read_tpte(struct net_device *dev, u32 stag, __be32 *tpte);
+u64 cxgb4_read_sge_timestamp(struct net_device *dev);
 
 #endif  /* !__CXGB4_OFLD_H */
diff --git a/drivers/net/ethernet/chelsio/cxgb4/t4_regs.h b/drivers/net/ethernet/chelsio/cxgb4/t4_regs.h
index ae77764..3b244ab 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/t4_regs.h
+++ b/drivers/net/ethernet/chelsio/cxgb4/t4_regs.h
@@ -251,6 +251,12 @@
 #define V_NOCOALESCE(x) ((x) << S_NOCOALESCE)
 #define F_NOCOALESCE    V_NOCOALESCE(1U)
 
+#define SGE_TIMESTAMP_LO 0x1098
+#define SGE_TIMESTAMP_HI 0x109c
+#define S_TSVAL    0
+#define M_TSVAL    0xfffffffU
+#define GET_TSVAL(x) (((x) >> S_TSVAL) & M_TSVAL)
+
 #define SGE_TIMER_VALUE_0_AND_1 0x10b8
 #define  TIMERVALUE0_MASK   0xffff0000U
 #define  TIMERVALUE0_SHIFT  16
-- 
1.7.1

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html