RDS/IB: add refcount tracking to struct rds_ib_device

The RDS IB client .remove callback used to free the rds_ibdev for the given
device unconditionally.  This could race other users of the struct.  This patch
adds refcounting so that we only free the rds_ibdev once all of its users are
done.

Many rds_ibdev users are tied to connections.  We give the connection a
reference and change these users to reference the device in the connection
instead of looking it up in the IB client data.  The only user of the IB client
data remaining is the first lookup of the device as connections are built up.

Incrementing the reference count of a device found in the IB client data could
race with final freeing so we use an RCU grace period to make sure that freeing
won't happen until those lookups are done.

MRs need the rds_ibdev to get at the pool that they're freed in to.  They exist
outside a connection and many MRs can reference different devices from one
socket, so it was natural to have each MR hold a reference.  MR refs can be
dropped from interrupt handlers and final device teardown can block so we push
it off to a work struct.  Pool teardown had to be fixed to cancel its pending
work instead of deadlocking waiting for all queued work, including itself, to
finish.

MRs get their reference from the global device list, which gets a reference.
It is left unprotected by locks and remains racy.  A simple global lock would
be a significant bottleneck.  More scalable (complicated) locking should be
done carefully in a later patch.

Signed-off-by: Zach Brown <zach.brown@oracle.com>

1 files changed, 14 insertions, 19 deletions

diff --git a/net/rds/ib_cm.c b/net/rds/ib_cm.c
index 73253f7..a9fb917 100644
--- a/net/rds/ib_cm.c
+++ b/net/rds/ib_cm.c
@@ -95,7 +95,6 @@ void rds_ib_cm_connect_complete(struct rds_connection *conn, struct rdma_cm_even
 {
 	const struct rds_ib_connect_private *dp = NULL;
 	struct rds_ib_connection *ic = conn->c_transport_data;
-	struct rds_ib_device *rds_ibdev;
 	struct ib_qp_attr qp_attr;
 	int err;
 
@@ -145,12 +144,11 @@ void rds_ib_cm_connect_complete(struct rds_connection *conn, struct rdma_cm_even
 	if (err)
 		printk(KERN_NOTICE "ib_modify_qp(IB_QP_STATE, RTS): err=%d\n", err);
 
-	/* update ib_device with this local ipaddr & conn */
-	rds_ibdev = ib_get_client_data(ic->i_cm_id->device, &rds_ib_client);
-	err = rds_ib_update_ipaddr(rds_ibdev, conn->c_laddr);
+	/* update ib_device with this local ipaddr */
+	err = rds_ib_update_ipaddr(ic->rds_ibdev, conn->c_laddr);
 	if (err)
-		printk(KERN_ERR "rds_ib_update_ipaddr failed (%d)\n", err);
-	rds_ib_add_conn(rds_ibdev, conn);
+		printk(KERN_ERR "rds_ib_update_ipaddr failed (%d)\n",
+			err);
 
 	/* If the peer gave us the last packet it saw, process this as if
 	 * we had received a regular ACK. */
@@ -168,12 +166,10 @@ static void rds_ib_cm_fill_conn_param(struct rds_connection *conn,
 			u32 max_initiator_depth)
 {
 	struct rds_ib_connection *ic = conn->c_transport_data;
-	struct rds_ib_device *rds_ibdev;
+	struct rds_ib_device *rds_ibdev = ic->rds_ibdev;
 
 	memset(conn_param, 0, sizeof(struct rdma_conn_param));
 
-	rds_ibdev = ib_get_client_data(ic->i_cm_id->device, &rds_ib_client);
-
 	conn_param->responder_resources =
 		min_t(u32, rds_ibdev->max_responder_resources, max_responder_resources);
 	conn_param->initiator_depth =
@@ -241,18 +237,16 @@ static int rds_ib_setup_qp(struct rds_connection *conn)
 	struct rds_ib_device *rds_ibdev;
 	int ret;
 
-	/* rds_ib_add_one creates a rds_ib_device object per IB device,
-	 * and allocates a protection domain, memory range and FMR pool
-	 * for each.  If that fails for any reason, it will not register
-	 * the rds_ibdev at all.
+	/*
+	 * It's normal to see a null device if an incoming connection races
+	 * with device removal, so we don't print a warning.
 	 */
-	rds_ibdev = ib_get_client_data(dev, &rds_ib_client);
-	if (!rds_ibdev) {
-		if (printk_ratelimit())
-			printk(KERN_NOTICE "RDS/IB: No client_data for device %s\n",
-					dev->name);
+	rds_ibdev = rds_ib_get_client_data(dev);
+	if (!rds_ibdev)
 		return -EOPNOTSUPP;
-	}
+
+	/* add the conn now so that connection establishment has the dev */
+	rds_ib_add_conn(rds_ibdev, conn);
 
 	if (rds_ibdev->max_wrs < ic->i_send_ring.w_nr + 1)
 		rds_ib_ring_resize(&ic->i_send_ring, rds_ibdev->max_wrs - 1);
@@ -371,6 +365,7 @@ static int rds_ib_setup_qp(struct rds_connection *conn)
 		 ic->i_send_cq, ic->i_recv_cq);
 
 out:
+	rds_ib_dev_put(rds_ibdev);
 	return ret;
 }