[PATCH] IB: Add automatic retries to MAD layer

Add automatic retries to MAD layer.

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Hal Rosenstock <halr@voltaire.com>
Cc: Roland Dreier <rolandd@cisco.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

5 files changed, 31 insertions, 2 deletions

diff --git a/drivers/infiniband/core/mad.c b/drivers/infiniband/core/mad.c
index 430a6ee..04f88d3 100644
--- a/drivers/infiniband/core/mad.c
+++ b/drivers/infiniband/core/mad.c
@@ -954,7 +954,7 @@ int ib_post_send_mad(struct ib_mad_agent *mad_agent,
 		/* Timeout will be updated after send completes */
 		mad_send_wr->timeout = msecs_to_jiffies(send_wr->wr.
 							ud.timeout_ms);
-		mad_send_wr->retry = 0;
+		mad_send_wr->retries = mad_send_wr->send_wr.wr.ud.retries;
 		/* One reference for each work request to QP + response */
 		mad_send_wr->refcount = 1 + (mad_send_wr->timeout > 0);
 		mad_send_wr->status = IB_WC_SUCCESS;
@@ -2174,6 +2174,27 @@ local_send_completion:
 	spin_unlock_irqrestore(&mad_agent_priv->lock, flags);
 }
 
+static int retry_send(struct ib_mad_send_wr_private *mad_send_wr)
+{
+	int ret;
+
+	if (!mad_send_wr->retries--)
+		return -ETIMEDOUT;
+
+	mad_send_wr->timeout = msecs_to_jiffies(mad_send_wr->send_wr.
+						wr.ud.timeout_ms);
+
+	ret = ib_send_mad(mad_send_wr);
+
+	if (!ret) {
+		mad_send_wr->refcount++;
+		list_del(&mad_send_wr->agent_list);
+		list_add_tail(&mad_send_wr->agent_list,
+			      &mad_send_wr->mad_agent_priv->send_list);
+	}
+	return ret;
+}
+
 static void timeout_sends(void *data)
 {
 	struct ib_mad_agent_private *mad_agent_priv;
@@ -2202,6 +2223,9 @@ static void timeout_sends(void *data)
 			break;
 		}
 
+		if (!retry_send(mad_send_wr))
+			continue;
+
 		list_del(&mad_send_wr->agent_list);
 		spin_unlock_irqrestore(&mad_agent_priv->lock, flags);
 
diff --git a/drivers/infiniband/core/mad_priv.h b/drivers/infiniband/core/mad_priv.h
index 6fcab00..8a61dd9 100644
--- a/drivers/infiniband/core/mad_priv.h
+++ b/drivers/infiniband/core/mad_priv.h
@@ -123,6 +123,7 @@ struct ib_mad_send_wr_private {
 	u64 wr_id;			/* client WR ID */
 	u64 tid;
 	unsigned long timeout;
+	int retries;
 	int retry;
 	int refcount;
 	enum ib_wc_status status;
@@ -136,6 +137,7 @@ struct ib_mad_local_private {
 	struct ib_sge sg_list[IB_MAD_SEND_REQ_MAX_SG];
 	u64 wr_id;			/* client WR ID */
 	u64 tid;
+	int retries;
 };
 
 struct ib_mad_mgmt_method_table {
diff --git a/drivers/infiniband/core/sa_query.c b/drivers/infiniband/core/sa_query.c
index 649824e..9ef8fe0 100644
--- a/drivers/infiniband/core/sa_query.c
+++ b/drivers/infiniband/core/sa_query.c
@@ -462,7 +462,8 @@ static int send_mad(struct ib_sa_query *query, int timeout_ms)
 				 .mad_hdr     = &query->mad->mad_hdr,
 				 .remote_qpn  = 1,
 				 .remote_qkey = IB_QP1_QKEY,
-				 .timeout_ms  = timeout_ms
+				 .timeout_ms  = timeout_ms,
+				 .retries     = 0
 			 }
 		 }
 	};
diff --git a/drivers/infiniband/core/user_mad.c b/drivers/infiniband/core/user_mad.c
index 9d912d6..088bb1f 100644
--- a/drivers/infiniband/core/user_mad.c
+++ b/drivers/infiniband/core/user_mad.c
@@ -322,6 +322,7 @@ static ssize_t ib_umad_write(struct file *filp, const char __user *buf,
 	wr.wr.ud.remote_qpn  = be32_to_cpu(packet->mad.qpn);
 	wr.wr.ud.remote_qkey = be32_to_cpu(packet->mad.qkey);
 	wr.wr.ud.timeout_ms  = packet->mad.timeout_ms;
+	wr.wr.ud.retries     = 0;
 
 	wr.wr_id            = (unsigned long) packet;
 
diff --git a/drivers/infiniband/include/ib_verbs.h b/drivers/infiniband/include/ib_verbs.h
index e5bd9a1..b6107c4 100644
--- a/drivers/infiniband/include/ib_verbs.h
+++ b/drivers/infiniband/include/ib_verbs.h
@@ -566,6 +566,7 @@ struct ib_send_wr {
 			u32	remote_qpn;
 			u32	remote_qkey;
 			int	timeout_ms; /* valid for MADs only */
+			int	retries;    /* valid for MADs only */
 			u16	pkey_index; /* valid for GSI only */
 			u8	port_num;   /* valid for DR SMPs on switch only */
 		} ud;