sunrpc/xprtrdma: clean up workqueue usage

* Create and use svc_rdma_wq instead of using the system workqueue and
  flush_scheduled_work().  This workqueue is necessary to serve as
  flushing domain for rdma->sc_work which is used to destroy itself
  and thus can't be flushed explicitly.

* Replace cancel_delayed_work() + flush_scheduled_work() with
  cancel_delayed_work_sync().

* Implement synchronous connect in xprt_rdma_connect() using
  flush_delayed_work() on the rdma_connect work instead of using
  flush_scheduled_work().

This is to prepare for the deprecation and removal of
flush_scheduled_work().

Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: J. Bruce Fields <bfields@redhat.com>

3 files changed, 17 insertions, 5 deletions

diff --git a/net/sunrpc/xprtrdma/svc_rdma.c b/net/sunrpc/xprtrdma/svc_rdma.c
index d718b8fa952..09af4fab1a4 100644
--- a/net/sunrpc/xprtrdma/svc_rdma.c
+++ b/net/sunrpc/xprtrdma/svc_rdma.c
@@ -43,6 +43,7 @@
 #include <linux/slab.h>
 #include <linux/fs.h>
 #include <linux/sysctl.h>
+#include <linux/workqueue.h>
 #include <linux/sunrpc/clnt.h>
 #include <linux/sunrpc/sched.h>
 #include <linux/sunrpc/svc_rdma.h>
@@ -74,6 +75,8 @@ atomic_t rdma_stat_sq_prod;
 struct kmem_cache *svc_rdma_map_cachep;
 struct kmem_cache *svc_rdma_ctxt_cachep;
 
+struct workqueue_struct *svc_rdma_wq;
+
 /*
  * This function implements reading and resetting an atomic_t stat
  * variable through read/write to a proc file. Any write to the file
@@ -231,7 +234,7 @@ static ctl_table svcrdma_root_table[] = {
 void svc_rdma_cleanup(void)
 {
 	dprintk("SVCRDMA Module Removed, deregister RPC RDMA transport\n");
-	flush_scheduled_work();
+	destroy_workqueue(svc_rdma_wq);
 	if (svcrdma_table_header) {
 		unregister_sysctl_table(svcrdma_table_header);
 		svcrdma_table_header = NULL;
@@ -249,6 +252,11 @@ int svc_rdma_init(void)
 	dprintk("\tsq_depth         : %d\n",
 		svcrdma_max_requests * RPCRDMA_SQ_DEPTH_MULT);
 	dprintk("\tmax_inline       : %d\n", svcrdma_max_req_size);
+
+	svc_rdma_wq = alloc_workqueue("svc_rdma", 0, 0);
+	if (!svc_rdma_wq)
+		return -ENOMEM;
+
 	if (!svcrdma_table_header)
 		svcrdma_table_header =
 			register_sysctl_table(svcrdma_root_table);
@@ -283,6 +291,7 @@ int svc_rdma_init(void)
 	kmem_cache_destroy(svc_rdma_map_cachep);
  err0:
 	unregister_sysctl_table(svcrdma_table_header);
+	destroy_workqueue(svc_rdma_wq);
 	return -ENOMEM;
 }
 MODULE_AUTHOR("Tom Tucker <tom@opengridcomputing.com>");
diff --git a/net/sunrpc/xprtrdma/svc_rdma_transport.c b/net/sunrpc/xprtrdma/svc_rdma_transport.c
index 22f65cc46fe..9df1eadc912 100644
--- a/net/sunrpc/xprtrdma/svc_rdma_transport.c
+++ b/net/sunrpc/xprtrdma/svc_rdma_transport.c
@@ -45,6 +45,7 @@
 #include <linux/sched.h>
 #include <linux/slab.h>
 #include <linux/spinlock.h>
+#include <linux/workqueue.h>
 #include <rdma/ib_verbs.h>
 #include <rdma/rdma_cm.h>
 #include <linux/sunrpc/svc_rdma.h>
@@ -90,6 +91,9 @@ struct svc_xprt_class svc_rdma_class = {
 /* WR context cache. Created in svc_rdma.c  */
 extern struct kmem_cache *svc_rdma_ctxt_cachep;
 
+/* Workqueue created in svc_rdma.c */
+extern struct workqueue_struct *svc_rdma_wq;
+
 struct svc_rdma_op_ctxt *svc_rdma_get_context(struct svcxprt_rdma *xprt)
 {
 	struct svc_rdma_op_ctxt *ctxt;
@@ -1187,7 +1191,7 @@ static void svc_rdma_free(struct svc_xprt *xprt)
 	struct svcxprt_rdma *rdma =
 		container_of(xprt, struct svcxprt_rdma, sc_xprt);
 	INIT_WORK(&rdma->sc_work, __svc_rdma_free);
-	schedule_work(&rdma->sc_work);
+	queue_work(svc_rdma_wq, &rdma->sc_work);
 }
 
 static int svc_rdma_has_wspace(struct svc_xprt *xprt)
diff --git a/net/sunrpc/xprtrdma/transport.c b/net/sunrpc/xprtrdma/transport.c
index 2da32b40bfc..0867070bb5c 100644
--- a/net/sunrpc/xprtrdma/transport.c
+++ b/net/sunrpc/xprtrdma/transport.c
@@ -237,8 +237,7 @@ xprt_rdma_destroy(struct rpc_xprt *xprt)
 
 	dprintk("RPC:       %s: called\n", __func__);
 
-	cancel_delayed_work(&r_xprt->rdma_connect);
-	flush_scheduled_work();
+	cancel_delayed_work_sync(&r_xprt->rdma_connect);
 
 	xprt_clear_connected(xprt);
 
@@ -448,7 +447,7 @@ xprt_rdma_connect(struct rpc_task *task)
 	} else {
 		schedule_delayed_work(&r_xprt->rdma_connect, 0);
 		if (!RPC_IS_ASYNC(task))
-			flush_scheduled_work();
+			flush_delayed_work(&r_xprt->rdma_connect);
 	}
 }