From fb14ae8853e4f0347950f98e604fa2f4f3b3abe1 Mon Sep 17 00:00:00 2001
From: Chuck Lever <chuck.lever@oracle.com>
Date: Wed, 28 Feb 2018 15:30:54 -0500
Subject: xprtrdma: "Support" call-only RPCs

RPC-over-RDMA version 1 credit accounting relies on there being a
response message for every RPC Call. This means that RPC procedures
that have no reply will disrupt credit accounting, just in the same
way as a retransmit would (since it is sent because no reply has
arrived). Deal with the "no reply" case the same way.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 include/linux/sunrpc/clnt.h | 7 +++++++
 1 file changed, 7 insertions(+)

(limited to 'include')

diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h
index ed761f751ecb..9b11b6a0978c 100644
--- a/include/linux/sunrpc/clnt.h
+++ b/include/linux/sunrpc/clnt.h
@@ -217,5 +217,12 @@ void rpc_clnt_xprt_switch_add_xprt(struct rpc_clnt *, struct rpc_xprt *);
 bool rpc_clnt_xprt_switch_has_addr(struct rpc_clnt *clnt,
 			const struct sockaddr *sap);
 void rpc_cleanup_clids(void);
+
+static inline int rpc_reply_expected(struct rpc_task *task)
+{
+	return (task->tk_msg.rpc_proc != NULL) &&
+		(task->tk_msg.rpc_proc->p_decode != NULL);
+}
+
 #endif /* __KERNEL__ */
 #endif /* _LINUX_SUNRPC_CLNT_H */
-- 
cgit v1.2.3-58-ga151


From ecd465ee88bb6648c06c82b1abae6ec28cf5fccb Mon Sep 17 00:00:00 2001
From: Chuck Lever <chuck.lever@oracle.com>
Date: Mon, 5 Mar 2018 15:12:57 -0500
Subject: SUNRPC: Move xprt_update_rtt callsite

Since commit 33849792cbcd ("xprtrdma: Detect unreachable NFS/RDMA
servers more reliably"), the xprtrdma transport now has a ->timer
callout. But xprtrdma does not need to compute RTT data, only UDP
needs that. Move the xprt_update_rtt call into the UDP transport
implementation.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 include/linux/sunrpc/xprt.h |  1 +
 net/sunrpc/xprt.c           | 11 ++++++++---
 net/sunrpc/xprtsock.c       |  1 +
 3 files changed, 10 insertions(+), 3 deletions(-)

(limited to 'include')

diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h
index 7fad83881ce1..ad322cec049a 100644
--- a/include/linux/sunrpc/xprt.h
+++ b/include/linux/sunrpc/xprt.h
@@ -373,6 +373,7 @@ void			xprt_wait_for_buffer_space(struct rpc_task *task, rpc_action action);
 void			xprt_write_space(struct rpc_xprt *xprt);
 void			xprt_adjust_cwnd(struct rpc_xprt *xprt, struct rpc_task *task, int result);
 struct rpc_rqst *	xprt_lookup_rqst(struct rpc_xprt *xprt, __be32 xid);
+void			xprt_update_rtt(struct rpc_task *task);
 void			xprt_complete_rqst(struct rpc_task *task, int copied);
 void			xprt_pin_rqst(struct rpc_rqst *req);
 void			xprt_unpin_rqst(struct rpc_rqst *req);
diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index 8f0ad4f268da..13fbb4849188 100644
--- a/net/sunrpc/xprt.c
+++ b/net/sunrpc/xprt.c
@@ -889,7 +889,13 @@ __must_hold(&req->rq_xprt->recv_lock)
 	}
 }
 
-static void xprt_update_rtt(struct rpc_task *task)
+/**
+ * xprt_update_rtt - Update RPC RTT statistics
+ * @task: RPC request that recently completed
+ *
+ * Caller holds xprt->recv_lock.
+ */
+void xprt_update_rtt(struct rpc_task *task)
 {
 	struct rpc_rqst *req = task->tk_rqstp;
 	struct rpc_rtt *rtt = task->tk_client->cl_rtt;
@@ -902,6 +908,7 @@ static void xprt_update_rtt(struct rpc_task *task)
 		rpc_set_timeo(rtt, timer, req->rq_ntrans - 1);
 	}
 }
+EXPORT_SYMBOL_GPL(xprt_update_rtt);
 
 /**
  * xprt_complete_rqst - called when reply processing is complete
@@ -921,8 +928,6 @@ void xprt_complete_rqst(struct rpc_task *task, int copied)
 
 	xprt->stat.recvs++;
 	req->rq_rtt = ktime_sub(ktime_get(), req->rq_xtime);
-	if (xprt->ops->timer != NULL)
-		xprt_update_rtt(task);
 
 	list_del_init(&req->rq_list);
 	req->rq_private_buf.len = copied;
diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c
index a6b8c1f8f92a..61df77f41304 100644
--- a/net/sunrpc/xprtsock.c
+++ b/net/sunrpc/xprtsock.c
@@ -1060,6 +1060,7 @@ static void xs_udp_data_read_skb(struct rpc_xprt *xprt,
 	if (!rovr)
 		goto out_unlock;
 	xprt_pin_rqst(rovr);
+	xprt_update_rtt(rovr->rq_task);
 	spin_unlock(&xprt->recv_lock);
 	task = rovr->rq_task;
 
-- 
cgit v1.2.3-58-ga151


From ff699ea8269a02d977c6ee42d58f76efe83a34f9 Mon Sep 17 00:00:00 2001
From: Chuck Lever <chuck.lever@oracle.com>
Date: Mon, 5 Mar 2018 15:13:13 -0500
Subject: SUNRPC: Make num_reqs a non-atomic integer

If recording xprt->stat.max_slots is moved into xprt_alloc_slot,
then xprt->num_reqs is never manipulated outside
xprt->reserve_lock. There's no longer a need for xprt->num_reqs to
be atomic.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 include/linux/sunrpc/xprt.h |  2 +-
 net/sunrpc/xprt.c           | 17 +++++++++--------
 2 files changed, 10 insertions(+), 9 deletions(-)

(limited to 'include')

diff --git a/include/linux/sunrpc/xprt.h b/include/linux/sunrpc/xprt.h
index ad322cec049a..5fea0fb420df 100644
--- a/include/linux/sunrpc/xprt.h
+++ b/include/linux/sunrpc/xprt.h
@@ -197,7 +197,7 @@ struct rpc_xprt {
 	struct list_head	free;		/* free slots */
 	unsigned int		max_reqs;	/* max number of slots */
 	unsigned int		min_reqs;	/* min number of slots */
-	atomic_t		num_reqs;	/* total slots */
+	unsigned int		num_reqs;	/* total slots */
 	unsigned long		state;		/* transport state */
 	unsigned char		resvport   : 1; /* use a reserved port */
 	atomic_t		swapper;	/* we're swapping over this
diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index 73f05a1fbc19..70f005044f06 100644
--- a/net/sunrpc/xprt.c
+++ b/net/sunrpc/xprt.c
@@ -1009,7 +1009,7 @@ void xprt_transmit(struct rpc_task *task)
 	struct rpc_rqst	*req = task->tk_rqstp;
 	struct rpc_xprt	*xprt = req->rq_xprt;
 	unsigned int connect_cookie;
-	int status, numreqs;
+	int status;
 
 	dprintk("RPC: %5u xprt_transmit(%u)\n", task->tk_pid, req->rq_slen);
 
@@ -1047,9 +1047,6 @@ void xprt_transmit(struct rpc_task *task)
 
 	xprt->ops->set_retrans_timeout(task);
 
-	numreqs = atomic_read(&xprt->num_reqs);
-	if (numreqs > xprt->stat.max_slots)
-		xprt->stat.max_slots = numreqs;
 	xprt->stat.sends++;
 	xprt->stat.req_u += xprt->stat.sends - xprt->stat.recvs;
 	xprt->stat.bklog_u += xprt->backlog.qlen;
@@ -1111,14 +1108,15 @@ static struct rpc_rqst *xprt_dynamic_alloc_slot(struct rpc_xprt *xprt)
 {
 	struct rpc_rqst *req = ERR_PTR(-EAGAIN);
 
-	if (!atomic_add_unless(&xprt->num_reqs, 1, xprt->max_reqs))
+	if (xprt->num_reqs >= xprt->max_reqs)
 		goto out;
+	++xprt->num_reqs;
 	spin_unlock(&xprt->reserve_lock);
 	req = kzalloc(sizeof(struct rpc_rqst), GFP_NOFS);
 	spin_lock(&xprt->reserve_lock);
 	if (req != NULL)
 		goto out;
-	atomic_dec(&xprt->num_reqs);
+	--xprt->num_reqs;
 	req = ERR_PTR(-ENOMEM);
 out:
 	return req;
@@ -1126,7 +1124,8 @@ out:
 
 static bool xprt_dynamic_free_slot(struct rpc_xprt *xprt, struct rpc_rqst *req)
 {
-	if (atomic_add_unless(&xprt->num_reqs, -1, xprt->min_reqs)) {
+	if (xprt->num_reqs > xprt->min_reqs) {
+		--xprt->num_reqs;
 		kfree(req);
 		return true;
 	}
@@ -1162,6 +1161,8 @@ void xprt_alloc_slot(struct rpc_xprt *xprt, struct rpc_task *task)
 	spin_unlock(&xprt->reserve_lock);
 	return;
 out_init_req:
+	xprt->stat.max_slots = max_t(unsigned int, xprt->stat.max_slots,
+				     xprt->num_reqs);
 	task->tk_status = 0;
 	task->tk_rqstp = req;
 	xprt_request_init(task, xprt);
@@ -1229,7 +1230,7 @@ struct rpc_xprt *xprt_alloc(struct net *net, size_t size,
 	else
 		xprt->max_reqs = num_prealloc;
 	xprt->min_reqs = num_prealloc;
-	atomic_set(&xprt->num_reqs, num_prealloc);
+	xprt->num_reqs = num_prealloc;
 
 	return xprt;
 
-- 
cgit v1.2.3-58-ga151


From e671edb9428c8a61662aaf8c39f5edced7cc45c7 Mon Sep 17 00:00:00 2001
From: Chuck Lever <chuck.lever@oracle.com>
Date: Fri, 16 Mar 2018 10:33:44 -0400
Subject: sunrpc: Simplify synopsis of some trace points

Clean up: struct rpc_task carries a pointer to a struct rpc_clnt,
and in fact task->tk_client is always what is passed into trace
points that are already passing @task.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 include/trace/events/sunrpc.h | 40 +++++++++++++++++++++-------------------
 net/sunrpc/clnt.c             |  2 +-
 net/sunrpc/sched.c            | 10 +++++-----
 3 files changed, 27 insertions(+), 25 deletions(-)

(limited to 'include')

diff --git a/include/trace/events/sunrpc.h b/include/trace/events/sunrpc.h
index 970c91a83173..53bbaeac08dd 100644
--- a/include/trace/events/sunrpc.h
+++ b/include/trace/events/sunrpc.h
@@ -50,9 +50,9 @@ DEFINE_EVENT(rpc_task_status, rpc_bind_status,
 );
 
 TRACE_EVENT(rpc_connect_status,
-	TP_PROTO(struct rpc_task *task, int status),
+	TP_PROTO(const struct rpc_task *task),
 
-	TP_ARGS(task, status),
+	TP_ARGS(task),
 
 	TP_STRUCT__entry(
 		__field(unsigned int, task_id)
@@ -63,7 +63,7 @@ TRACE_EVENT(rpc_connect_status,
 	TP_fast_assign(
 		__entry->task_id = task->tk_pid;
 		__entry->client_id = task->tk_client->cl_clid;
-		__entry->status = status;
+		__entry->status = task->tk_status;
 	),
 
 	TP_printk("task:%u@%u status=%d",
@@ -103,9 +103,9 @@ TRACE_EVENT(rpc_request,
 
 DECLARE_EVENT_CLASS(rpc_task_running,
 
-	TP_PROTO(const struct rpc_clnt *clnt, const struct rpc_task *task, const void *action),
+	TP_PROTO(const struct rpc_task *task, const void *action),
 
-	TP_ARGS(clnt, task, action),
+	TP_ARGS(task, action),
 
 	TP_STRUCT__entry(
 		__field(unsigned int, task_id)
@@ -117,7 +117,8 @@ DECLARE_EVENT_CLASS(rpc_task_running,
 		),
 
 	TP_fast_assign(
-		__entry->client_id = clnt ? clnt->cl_clid : -1;
+		__entry->client_id = task->tk_client ?
+				     task->tk_client->cl_clid : -1;
 		__entry->task_id = task->tk_pid;
 		__entry->action = action;
 		__entry->runstate = task->tk_runstate;
@@ -136,33 +137,33 @@ DECLARE_EVENT_CLASS(rpc_task_running,
 
 DEFINE_EVENT(rpc_task_running, rpc_task_begin,
 
-	TP_PROTO(const struct rpc_clnt *clnt, const struct rpc_task *task, const void *action),
+	TP_PROTO(const struct rpc_task *task, const void *action),
 
-	TP_ARGS(clnt, task, action)
+	TP_ARGS(task, action)
 
 );
 
 DEFINE_EVENT(rpc_task_running, rpc_task_run_action,
 
-	TP_PROTO(const struct rpc_clnt *clnt, const struct rpc_task *task, const void *action),
+	TP_PROTO(const struct rpc_task *task, const void *action),
 
-	TP_ARGS(clnt, task, action)
+	TP_ARGS(task, action)
 
 );
 
 DEFINE_EVENT(rpc_task_running, rpc_task_complete,
 
-	TP_PROTO(const struct rpc_clnt *clnt, const struct rpc_task *task, const void *action),
+	TP_PROTO(const struct rpc_task *task, const void *action),
 
-	TP_ARGS(clnt, task, action)
+	TP_ARGS(task, action)
 
 );
 
 DECLARE_EVENT_CLASS(rpc_task_queued,
 
-	TP_PROTO(const struct rpc_clnt *clnt, const struct rpc_task *task, const struct rpc_wait_queue *q),
+	TP_PROTO(const struct rpc_task *task, const struct rpc_wait_queue *q),
 
-	TP_ARGS(clnt, task, q),
+	TP_ARGS(task, q),
 
 	TP_STRUCT__entry(
 		__field(unsigned int, task_id)
@@ -175,7 +176,8 @@ DECLARE_EVENT_CLASS(rpc_task_queued,
 		),
 
 	TP_fast_assign(
-		__entry->client_id = clnt ? clnt->cl_clid : -1;
+		__entry->client_id = task->tk_client ?
+				     task->tk_client->cl_clid : -1;
 		__entry->task_id = task->tk_pid;
 		__entry->timeout = task->tk_timeout;
 		__entry->runstate = task->tk_runstate;
@@ -196,17 +198,17 @@ DECLARE_EVENT_CLASS(rpc_task_queued,
 
 DEFINE_EVENT(rpc_task_queued, rpc_task_sleep,
 
-	TP_PROTO(const struct rpc_clnt *clnt, const struct rpc_task *task, const struct rpc_wait_queue *q),
+	TP_PROTO(const struct rpc_task *task, const struct rpc_wait_queue *q),
 
-	TP_ARGS(clnt, task, q)
+	TP_ARGS(task, q)
 
 );
 
 DEFINE_EVENT(rpc_task_queued, rpc_task_wakeup,
 
-	TP_PROTO(const struct rpc_clnt *clnt, const struct rpc_task *task, const struct rpc_wait_queue *q),
+	TP_PROTO(const struct rpc_task *task, const struct rpc_wait_queue *q),
 
-	TP_ARGS(clnt, task, q)
+	TP_ARGS(task, q)
 
 );
 
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index 6e432ecd7f99..079f06fe2c5d 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -1887,7 +1887,7 @@ call_connect_status(struct rpc_task *task)
 
 	dprint_status(task);
 
-	trace_rpc_connect_status(task, status);
+	trace_rpc_connect_status(task);
 	task->tk_status = 0;
 	switch (status) {
 	case -ECONNREFUSED:
diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c
index d9db2eab3a8d..3fe5d60ab0e2 100644
--- a/net/sunrpc/sched.c
+++ b/net/sunrpc/sched.c
@@ -276,7 +276,7 @@ static void rpc_set_active(struct rpc_task *task)
 {
 	rpc_task_set_debuginfo(task);
 	set_bit(RPC_TASK_ACTIVE, &task->tk_runstate);
-	trace_rpc_task_begin(task->tk_client, task, NULL);
+	trace_rpc_task_begin(task, NULL);
 }
 
 /*
@@ -291,7 +291,7 @@ static int rpc_complete_task(struct rpc_task *task)
 	unsigned long flags;
 	int ret;
 
-	trace_rpc_task_complete(task->tk_client, task, NULL);
+	trace_rpc_task_complete(task, NULL);
 
 	spin_lock_irqsave(&wq->lock, flags);
 	clear_bit(RPC_TASK_ACTIVE, &task->tk_runstate);
@@ -358,7 +358,7 @@ static void __rpc_sleep_on_priority(struct rpc_wait_queue *q,
 	dprintk("RPC: %5u sleep_on(queue \"%s\" time %lu)\n",
 			task->tk_pid, rpc_qname(q), jiffies);
 
-	trace_rpc_task_sleep(task->tk_client, task, q);
+	trace_rpc_task_sleep(task, q);
 
 	__rpc_add_wait_queue(q, task, queue_priority);
 
@@ -428,7 +428,7 @@ static void __rpc_do_wake_up_task_on_wq(struct workqueue_struct *wq,
 		return;
 	}
 
-	trace_rpc_task_wakeup(task->tk_client, task, queue);
+	trace_rpc_task_wakeup(task, queue);
 
 	__rpc_remove_wait_queue(queue, task);
 
@@ -780,7 +780,7 @@ static void __rpc_execute(struct rpc_task *task)
 		}
 		if (!do_action)
 			break;
-		trace_rpc_task_run_action(task->tk_client, task, do_action);
+		trace_rpc_task_run_action(task, do_action);
 		do_action(task);
 
 		/*
-- 
cgit v1.2.3-58-ga151


From 40bf7eb304b5659991ed932c0cd5bee6a7c88f4f Mon Sep 17 00:00:00 2001
From: Chuck Lever <chuck.lever@oracle.com>
Date: Fri, 16 Mar 2018 10:33:49 -0400
Subject: sunrpc: Add static trace point to report RPC latency stats

Introduce a low-overhead mechanism to report information about
latencies of individual RPCs. The goal is to enable user space to
filter the trace record for latency outliers, or build histograms,
etc.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 include/trace/events/sunrpc.h | 45 +++++++++++++++++++++++++++++++++++++++++++
 net/sunrpc/stats.c            | 16 ++++++++++-----
 2 files changed, 56 insertions(+), 5 deletions(-)

(limited to 'include')

diff --git a/include/trace/events/sunrpc.h b/include/trace/events/sunrpc.h
index 53bbaeac08dd..fda9e726f1cc 100644
--- a/include/trace/events/sunrpc.h
+++ b/include/trace/events/sunrpc.h
@@ -212,6 +212,51 @@ DEFINE_EVENT(rpc_task_queued, rpc_task_wakeup,
 
 );
 
+TRACE_EVENT(rpc_stats_latency,
+
+	TP_PROTO(
+		const struct rpc_task *task,
+		ktime_t backlog,
+		ktime_t rtt,
+		ktime_t execute
+	),
+
+	TP_ARGS(task, backlog, rtt, execute),
+
+	TP_STRUCT__entry(
+		__field(u32, xid)
+		__field(int, version)
+		__string(progname, task->tk_client->cl_program->name)
+		__string(procname, rpc_proc_name(task))
+		__field(unsigned long, backlog)
+		__field(unsigned long, rtt)
+		__field(unsigned long, execute)
+		__string(addr,
+			 task->tk_xprt->address_strings[RPC_DISPLAY_ADDR])
+		__string(port,
+			 task->tk_xprt->address_strings[RPC_DISPLAY_PORT])
+	),
+
+	TP_fast_assign(
+		__entry->xid = be32_to_cpu(task->tk_rqstp->rq_xid);
+		__entry->version = task->tk_client->cl_vers;
+		__assign_str(progname, task->tk_client->cl_program->name)
+		__assign_str(procname, rpc_proc_name(task))
+		__entry->backlog = ktime_to_us(backlog);
+		__entry->rtt = ktime_to_us(rtt);
+		__entry->execute = ktime_to_us(execute);
+		__assign_str(addr,
+			     task->tk_xprt->address_strings[RPC_DISPLAY_ADDR]);
+		__assign_str(port,
+			     task->tk_xprt->address_strings[RPC_DISPLAY_PORT]);
+	),
+
+	TP_printk("peer=[%s]:%s xid=0x%08x %sv%d %s backlog=%lu rtt=%lu execute=%lu",
+		__get_str(addr), __get_str(port), __entry->xid,
+		__get_str(progname), __entry->version, __get_str(procname),
+		__entry->backlog, __entry->rtt, __entry->execute)
+);
+
 /*
  * First define the enums in the below macros to be exported to userspace
  * via TRACE_DEFINE_ENUM().
diff --git a/net/sunrpc/stats.c b/net/sunrpc/stats.c
index 1e671333c3d5..f68aa46c9dd7 100644
--- a/net/sunrpc/stats.c
+++ b/net/sunrpc/stats.c
@@ -24,6 +24,8 @@
 #include <linux/sunrpc/metrics.h>
 #include <linux/rcupdate.h>
 
+#include <trace/events/sunrpc.h>
+
 #include "netns.h"
 
 #define RPCDBG_FACILITY	RPCDBG_MISC
@@ -148,7 +150,7 @@ void rpc_count_iostats_metrics(const struct rpc_task *task,
 			       struct rpc_iostats *op_metrics)
 {
 	struct rpc_rqst *req = task->tk_rqstp;
-	ktime_t delta, now;
+	ktime_t backlog, execute, now;
 
 	if (!op_metrics || !req)
 		return;
@@ -164,16 +166,20 @@ void rpc_count_iostats_metrics(const struct rpc_task *task,
 	op_metrics->om_bytes_sent += req->rq_xmit_bytes_sent;
 	op_metrics->om_bytes_recv += req->rq_reply_bytes_recvd;
 
+	backlog = 0;
 	if (ktime_to_ns(req->rq_xtime)) {
-		delta = ktime_sub(req->rq_xtime, task->tk_start);
-		op_metrics->om_queue = ktime_add(op_metrics->om_queue, delta);
+		backlog = ktime_sub(req->rq_xtime, task->tk_start);
+		op_metrics->om_queue = ktime_add(op_metrics->om_queue, backlog);
 	}
+
 	op_metrics->om_rtt = ktime_add(op_metrics->om_rtt, req->rq_rtt);
 
-	delta = ktime_sub(now, task->tk_start);
-	op_metrics->om_execute = ktime_add(op_metrics->om_execute, delta);
+	execute = ktime_sub(now, task->tk_start);
+	op_metrics->om_execute = ktime_add(op_metrics->om_execute, execute);
 
 	spin_unlock(&op_metrics->om_lock);
+
+	trace_rpc_stats_latency(req->rq_task, backlog, req->rq_rtt, execute);
 }
 EXPORT_SYMBOL_GPL(rpc_count_iostats_metrics);
 
-- 
cgit v1.2.3-58-ga151


From a25a4cb3af177a4cf5621ffbf4fa89ae60c6d4d7 Mon Sep 17 00:00:00 2001
From: Chuck Lever <chuck.lever@oracle.com>
Date: Fri, 16 Mar 2018 10:33:55 -0400
Subject: sunrpc: Add static trace point to report result of RPC ping

This information can help track down local misconfiguration issues
as well as network partitions and unresponsive servers.

There are several ways to send a ping, and with transport multi-
plexing, the exact rpc_xprt that is used is sometimes not known by
the upper layer. The rpc_xprt pointer passed to the trace point
call also has to be RCU-safe.

I found a spot inside the client FSM where an rpc_xprt pointer is
always available and safe to use.

Suggested-by: Bill Baker <Bill.Baker@oracle.com>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 include/trace/events/sunrpc.h | 21 +++++++++++++++++++++
 net/sunrpc/clnt.c             |  6 ++++++
 2 files changed, 27 insertions(+)

(limited to 'include')

diff --git a/include/trace/events/sunrpc.h b/include/trace/events/sunrpc.h
index fda9e726f1cc..76887d60f0c0 100644
--- a/include/trace/events/sunrpc.h
+++ b/include/trace/events/sunrpc.h
@@ -453,6 +453,27 @@ DEFINE_EVENT(rpc_xprt_event, xprt_complete_rqst,
 	TP_PROTO(struct rpc_xprt *xprt, __be32 xid, int status),
 	TP_ARGS(xprt, xid, status));
 
+TRACE_EVENT(xprt_ping,
+	TP_PROTO(const struct rpc_xprt *xprt, int status),
+
+	TP_ARGS(xprt, status),
+
+	TP_STRUCT__entry(
+		__field(int, status)
+		__string(addr, xprt->address_strings[RPC_DISPLAY_ADDR])
+		__string(port, xprt->address_strings[RPC_DISPLAY_PORT])
+	),
+
+	TP_fast_assign(
+		__entry->status = status;
+		__assign_str(addr, xprt->address_strings[RPC_DISPLAY_ADDR]);
+		__assign_str(port, xprt->address_strings[RPC_DISPLAY_PORT]);
+	),
+
+	TP_printk("peer=[%s]:%s status=%d",
+			__get_str(addr), __get_str(port), __entry->status)
+);
+
 TRACE_EVENT(xs_tcp_data_ready,
 	TP_PROTO(struct rpc_xprt *xprt, int err, unsigned int total),
 
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index 079f06fe2c5d..166f8c1680d1 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -2014,6 +2014,9 @@ call_transmit_status(struct rpc_task *task)
 	case -EPERM:
 		if (RPC_IS_SOFTCONN(task)) {
 			xprt_end_transmit(task);
+			if (!task->tk_msg.rpc_proc->p_proc)
+				trace_xprt_ping(task->tk_xprt,
+						task->tk_status);
 			rpc_exit(task, task->tk_status);
 			break;
 		}
@@ -2112,6 +2115,9 @@ call_status(struct rpc_task *task)
 	struct rpc_rqst	*req = task->tk_rqstp;
 	int		status;
 
+	if (!task->tk_msg.rpc_proc->p_proc)
+		trace_xprt_ping(task->tk_xprt, task->tk_status);
+
 	if (req->rq_reply_bytes_recvd > 0 && !req->rq_bytes_sent)
 		task->tk_status = req->rq_reply_bytes_recvd;
 
-- 
cgit v1.2.3-58-ga151


From 912678dbc592db7ad618f383866ad23e43cd51f3 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <trond.myklebust@primarydata.com>
Date: Tue, 20 Mar 2018 16:43:15 -0400
Subject: NFS: Move the delegation return down into nfs4_proc_remove()

Move the delegation return out of generic code and down into the
NFSv4 specific unlink code.

Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 fs/nfs/dir.c            |  5 ++---
 fs/nfs/nfs3proc.c       |  6 +++---
 fs/nfs/nfs4proc.c       | 22 ++++++++++++++++++++--
 fs/nfs/proc.c           |  6 +++---
 include/linux/nfs_xdr.h |  2 +-
 5 files changed, 29 insertions(+), 12 deletions(-)

(limited to 'include')

diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index b4549e54007d..eb9d782ed674 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -1798,12 +1798,11 @@ static int nfs_safe_remove(struct dentry *dentry)
 
 	trace_nfs_remove_enter(dir, dentry);
 	if (inode != NULL) {
-		NFS_PROTO(inode)->return_delegation(inode);
-		error = NFS_PROTO(dir)->remove(dir, &dentry->d_name);
+		error = NFS_PROTO(dir)->remove(dir, dentry);
 		if (error == 0)
 			nfs_drop_nlink(inode);
 	} else
-		error = NFS_PROTO(dir)->remove(dir, &dentry->d_name);
+		error = NFS_PROTO(dir)->remove(dir, dentry);
 	if (error == -ENOENT)
 		nfs_dentry_handle_enoent(dentry);
 	trace_nfs_remove_exit(dir, dentry, error);
diff --git a/fs/nfs/nfs3proc.c b/fs/nfs/nfs3proc.c
index 7327930ad970..f4ead71e4350 100644
--- a/fs/nfs/nfs3proc.c
+++ b/fs/nfs/nfs3proc.c
@@ -383,11 +383,11 @@ out:
 }
 
 static int
-nfs3_proc_remove(struct inode *dir, const struct qstr *name)
+nfs3_proc_remove(struct inode *dir, struct dentry *dentry)
 {
 	struct nfs_removeargs arg = {
 		.fh = NFS_FH(dir),
-		.name = *name,
+		.name = dentry->d_name,
 	};
 	struct nfs_removeres res;
 	struct rpc_message msg = {
@@ -397,7 +397,7 @@ nfs3_proc_remove(struct inode *dir, const struct qstr *name)
 	};
 	int status = -ENOMEM;
 
-	dprintk("NFS call  remove %s\n", name->name);
+	dprintk("NFS call  remove %pd2\n", dentry);
 	res.dir_attr = nfs_alloc_fattr();
 	if (res.dir_attr == NULL)
 		goto out;
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index f4216b6b01c9..810ebd2f5174 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -4200,10 +4200,28 @@ static int _nfs4_proc_remove(struct inode *dir, const struct qstr *name)
 	return status;
 }
 
-static int nfs4_proc_remove(struct inode *dir, const struct qstr *name)
+static int nfs4_proc_remove(struct inode *dir, struct dentry *dentry)
 {
 	struct nfs4_exception exception = { };
+	struct inode *inode = d_inode(dentry);
 	int err;
+
+	if (inode)
+		nfs4_inode_return_delegation(inode);
+	do {
+		err = _nfs4_proc_remove(dir, &dentry->d_name);
+		trace_nfs4_remove(dir, &dentry->d_name, err);
+		err = nfs4_handle_exception(NFS_SERVER(dir), err,
+				&exception);
+	} while (exception.retry);
+	return err;
+}
+
+static int nfs4_proc_rmdir(struct inode *dir, const struct qstr *name)
+{
+	struct nfs4_exception exception = { };
+	int err;
+
 	do {
 		err = _nfs4_proc_remove(dir, name);
 		trace_nfs4_remove(dir, name, err);
@@ -9601,7 +9619,7 @@ const struct nfs_rpc_ops nfs_v4_clientops = {
 	.link		= nfs4_proc_link,
 	.symlink	= nfs4_proc_symlink,
 	.mkdir		= nfs4_proc_mkdir,
-	.rmdir		= nfs4_proc_remove,
+	.rmdir		= nfs4_proc_rmdir,
 	.readdir	= nfs4_proc_readdir,
 	.mknod		= nfs4_proc_mknod,
 	.statfs		= nfs4_proc_statfs,
diff --git a/fs/nfs/proc.c b/fs/nfs/proc.c
index f7fd9192d4bc..b2e81a110133 100644
--- a/fs/nfs/proc.c
+++ b/fs/nfs/proc.c
@@ -300,11 +300,11 @@ out:
 }
   
 static int
-nfs_proc_remove(struct inode *dir, const struct qstr *name)
+nfs_proc_remove(struct inode *dir, struct dentry *dentry)
 {
 	struct nfs_removeargs arg = {
 		.fh = NFS_FH(dir),
-		.name = *name,
+		.name = dentry->d_name,
 	};
 	struct rpc_message msg = { 
 		.rpc_proc = &nfs_procedures[NFSPROC_REMOVE],
@@ -312,7 +312,7 @@ nfs_proc_remove(struct inode *dir, const struct qstr *name)
 	};
 	int			status;
 
-	dprintk("NFS call  remove %s\n", name->name);
+	dprintk("NFS call  remove %pd2\n",dentry);
 	status = rpc_call_sync(NFS_CLIENT(dir), &msg, 0);
 	nfs_mark_for_revalidate(dir);
 
diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h
index 6959968dc36a..3ebf14b3bf0b 100644
--- a/include/linux/nfs_xdr.h
+++ b/include/linux/nfs_xdr.h
@@ -1590,7 +1590,7 @@ struct nfs_rpc_ops {
 			    unsigned int);
 	int	(*create)  (struct inode *, struct dentry *,
 			    struct iattr *, int);
-	int	(*remove)  (struct inode *, const struct qstr *);
+	int	(*remove)  (struct inode *, struct dentry *);
 	void	(*unlink_setup)  (struct rpc_message *, struct inode *dir);
 	void	(*unlink_rpc_prepare) (struct rpc_task *, struct nfs_unlinkdata *);
 	int	(*unlink_done) (struct rpc_task *, struct inode *);
-- 
cgit v1.2.3-58-ga151


From f2c2c552f119db84d85a53a8bd76479f34df02b1 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <trond.myklebust@primarydata.com>
Date: Tue, 20 Mar 2018 16:43:16 -0400
Subject: NFS: Move delegation recall into the NFSv4 callback for
 rename_setup()

Move the delegation recall out of the generic code, and into the NFSv4
specific callback.

Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 fs/nfs/dir.c            |  4 ----
 fs/nfs/nfs3proc.c       |  4 +++-
 fs/nfs/nfs4proc.c       | 13 ++++++++++---
 fs/nfs/proc.c           |  4 +++-
 fs/nfs/unlink.c         |  5 +----
 include/linux/nfs_xdr.h |  4 +++-
 6 files changed, 20 insertions(+), 14 deletions(-)

(limited to 'include')

diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index eb9d782ed674..8001f8c7ad0e 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -2020,10 +2020,6 @@ int nfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 		}
 	}
 
-	NFS_PROTO(old_inode)->return_delegation(old_inode);
-	if (new_inode != NULL)
-		NFS_PROTO(new_inode)->return_delegation(new_inode);
-
 	task = nfs_async_rename(old_dir, new_dir, old_dentry, new_dentry, NULL);
 	if (IS_ERR(task)) {
 		error = PTR_ERR(task);
diff --git a/fs/nfs/nfs3proc.c b/fs/nfs/nfs3proc.c
index f4ead71e4350..08875dc17b52 100644
--- a/fs/nfs/nfs3proc.c
+++ b/fs/nfs/nfs3proc.c
@@ -433,7 +433,9 @@ nfs3_proc_unlink_done(struct rpc_task *task, struct inode *dir)
 }
 
 static void
-nfs3_proc_rename_setup(struct rpc_message *msg, struct inode *dir)
+nfs3_proc_rename_setup(struct rpc_message *msg,
+		struct dentry *old_dentry,
+		struct dentry *new_dentry)
 {
 	msg->rpc_proc = &nfs3_procedures[NFS3PROC_RENAME];
 }
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index 810ebd2f5174..b06a820b570d 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -4267,14 +4267,21 @@ static int nfs4_proc_unlink_done(struct rpc_task *task, struct inode *dir)
 	return 1;
 }
 
-static void nfs4_proc_rename_setup(struct rpc_message *msg, struct inode *dir)
+static void nfs4_proc_rename_setup(struct rpc_message *msg,
+		struct dentry *old_dentry,
+		struct dentry *new_dentry)
 {
-	struct nfs_server *server = NFS_SERVER(dir);
 	struct nfs_renameargs *arg = msg->rpc_argp;
 	struct nfs_renameres *res = msg->rpc_resp;
+	struct inode *old_inode = d_inode(old_dentry);
+	struct inode *new_inode = d_inode(new_dentry);
 
+	if (old_inode)
+		nfs4_inode_return_delegation(old_inode);
+	if (new_inode)
+		nfs4_inode_return_delegation(new_inode);
 	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_RENAME];
-	res->server = server;
+	res->server = NFS_SB(old_dentry->d_sb);
 	nfs4_init_sequence(&arg->seq_args, &res->seq_res, 1);
 }
 
diff --git a/fs/nfs/proc.c b/fs/nfs/proc.c
index b2e81a110133..39dc9276b1f6 100644
--- a/fs/nfs/proc.c
+++ b/fs/nfs/proc.c
@@ -338,7 +338,9 @@ static int nfs_proc_unlink_done(struct rpc_task *task, struct inode *dir)
 }
 
 static void
-nfs_proc_rename_setup(struct rpc_message *msg, struct inode *dir)
+nfs_proc_rename_setup(struct rpc_message *msg,
+		struct dentry *old_dentry,
+		struct dentry *new_dentry)
 {
 	msg->rpc_proc = &nfs_procedures[NFSPROC_RENAME];
 }
diff --git a/fs/nfs/unlink.c b/fs/nfs/unlink.c
index 630b4a3c1a93..44f101411422 100644
--- a/fs/nfs/unlink.c
+++ b/fs/nfs/unlink.c
@@ -386,7 +386,7 @@ nfs_async_rename(struct inode *old_dir, struct inode *new_dir,
 
 	nfs_sb_active(old_dir->i_sb);
 
-	NFS_PROTO(data->old_dir)->rename_setup(&msg, old_dir);
+	NFS_PROTO(data->old_dir)->rename_setup(&msg, old_dentry, new_dentry);
 
 	return rpc_run_task(&task_setup_data);
 }
@@ -463,9 +463,6 @@ nfs_sillyrename(struct inode *dir, struct dentry *dentry)
 
 	fileid = NFS_FILEID(d_inode(dentry));
 
-	/* Return delegation in anticipation of the rename */
-	NFS_PROTO(d_inode(dentry))->return_delegation(d_inode(dentry));
-
 	sdentry = NULL;
 	do {
 		int slen;
diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h
index 3ebf14b3bf0b..4b87e2d726b1 100644
--- a/include/linux/nfs_xdr.h
+++ b/include/linux/nfs_xdr.h
@@ -1594,7 +1594,9 @@ struct nfs_rpc_ops {
 	void	(*unlink_setup)  (struct rpc_message *, struct inode *dir);
 	void	(*unlink_rpc_prepare) (struct rpc_task *, struct nfs_unlinkdata *);
 	int	(*unlink_done) (struct rpc_task *, struct inode *);
-	void	(*rename_setup)  (struct rpc_message *msg, struct inode *dir);
+	void	(*rename_setup)  (struct rpc_message *msg,
+			struct dentry *old_dentry,
+			struct dentry *new_dentry);
 	void	(*rename_rpc_prepare)(struct rpc_task *task, struct nfs_renamedata *);
 	int	(*rename_done) (struct rpc_task *task, struct inode *old_dir, struct inode *new_dir);
 	int	(*link)    (struct inode *, struct inode *, const struct qstr *);
-- 
cgit v1.2.3-58-ga151


From 977fcc2b0b41c1fc82e8349995695e207ccb6684 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <trond.myklebust@primarydata.com>
Date: Tue, 20 Mar 2018 16:43:17 -0400
Subject: NFS: Add a delegation return into nfs4_proc_unlink_setup()

Ensure that when we do finally delete the file, then we return the
delegation.

Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 fs/nfs/nfs3proc.c       | 2 +-
 fs/nfs/nfs4proc.c       | 9 ++++++---
 fs/nfs/proc.c           | 2 +-
 fs/nfs/unlink.c         | 2 +-
 include/linux/nfs_xdr.h | 2 +-
 5 files changed, 10 insertions(+), 7 deletions(-)

(limited to 'include')

diff --git a/fs/nfs/nfs3proc.c b/fs/nfs/nfs3proc.c
index 08875dc17b52..ae03307bd8cd 100644
--- a/fs/nfs/nfs3proc.c
+++ b/fs/nfs/nfs3proc.c
@@ -411,7 +411,7 @@ out:
 }
 
 static void
-nfs3_proc_unlink_setup(struct rpc_message *msg, struct inode *dir)
+nfs3_proc_unlink_setup(struct rpc_message *msg, struct dentry *dentry)
 {
 	msg->rpc_proc = &nfs3_procedures[NFS3PROC_REMOVE];
 }
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index b06a820b570d..21daeac114fe 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -4231,17 +4231,20 @@ static int nfs4_proc_rmdir(struct inode *dir, const struct qstr *name)
 	return err;
 }
 
-static void nfs4_proc_unlink_setup(struct rpc_message *msg, struct inode *dir)
+static void nfs4_proc_unlink_setup(struct rpc_message *msg, struct dentry *dentry)
 {
-	struct nfs_server *server = NFS_SERVER(dir);
 	struct nfs_removeargs *args = msg->rpc_argp;
 	struct nfs_removeres *res = msg->rpc_resp;
+	struct inode *inode = d_inode(dentry);
 
-	res->server = server;
+	res->server = NFS_SB(dentry->d_sb);
 	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_REMOVE];
 	nfs4_init_sequence(&args->seq_args, &res->seq_res, 1);
 
 	nfs_fattr_init(res->dir_attr);
+
+	if (inode)
+		nfs4_inode_return_delegation(inode);
 }
 
 static void nfs4_proc_unlink_rpc_prepare(struct rpc_task *task, struct nfs_unlinkdata *data)
diff --git a/fs/nfs/proc.c b/fs/nfs/proc.c
index 39dc9276b1f6..318b3f34a6d0 100644
--- a/fs/nfs/proc.c
+++ b/fs/nfs/proc.c
@@ -321,7 +321,7 @@ nfs_proc_remove(struct inode *dir, struct dentry *dentry)
 }
 
 static void
-nfs_proc_unlink_setup(struct rpc_message *msg, struct inode *dir)
+nfs_proc_unlink_setup(struct rpc_message *msg, struct dentry *dentry)
 {
 	msg->rpc_proc = &nfs_procedures[NFSPROC_REMOVE];
 }
diff --git a/fs/nfs/unlink.c b/fs/nfs/unlink.c
index 44f101411422..bf54fc9ae135 100644
--- a/fs/nfs/unlink.c
+++ b/fs/nfs/unlink.c
@@ -105,7 +105,7 @@ static void nfs_do_call_unlink(struct nfs_unlinkdata *data)
 	data->args.fh = NFS_FH(dir);
 	nfs_fattr_init(data->res.dir_attr);
 
-	NFS_PROTO(dir)->unlink_setup(&msg, dir);
+	NFS_PROTO(dir)->unlink_setup(&msg, data->dentry);
 
 	task_setup_data.rpc_client = NFS_CLIENT(dir);
 	task = rpc_run_task(&task_setup_data);
diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h
index 4b87e2d726b1..c4ba58b3c0f8 100644
--- a/include/linux/nfs_xdr.h
+++ b/include/linux/nfs_xdr.h
@@ -1591,7 +1591,7 @@ struct nfs_rpc_ops {
 	int	(*create)  (struct inode *, struct dentry *,
 			    struct iattr *, int);
 	int	(*remove)  (struct inode *, struct dentry *);
-	void	(*unlink_setup)  (struct rpc_message *, struct inode *dir);
+	void	(*unlink_setup)  (struct rpc_message *, struct dentry *);
 	void	(*unlink_rpc_prepare) (struct rpc_task *, struct nfs_unlinkdata *);
 	int	(*unlink_done) (struct rpc_task *, struct inode *);
 	void	(*rename_setup)  (struct rpc_message *msg,
-- 
cgit v1.2.3-58-ga151


From c135cb39a907b85aef5389c191b6f02cffbadb8a Mon Sep 17 00:00:00 2001
From: Trond Myklebust <trond.myklebust@primarydata.com>
Date: Tue, 20 Mar 2018 16:43:19 -0400
Subject: NFS: Remove the unused return_delegation() callback

Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 fs/nfs/nfs3proc.c       | 7 -------
 fs/nfs/nfs4proc.c       | 1 -
 fs/nfs/proc.c           | 7 -------
 include/linux/nfs_xdr.h | 1 -
 4 files changed, 16 deletions(-)

(limited to 'include')

diff --git a/fs/nfs/nfs3proc.c b/fs/nfs/nfs3proc.c
index ae03307bd8cd..cc04d988db7c 100644
--- a/fs/nfs/nfs3proc.c
+++ b/fs/nfs/nfs3proc.c
@@ -910,12 +910,6 @@ static int nfs3_have_delegation(struct inode *inode, fmode_t flags)
 	return 0;
 }
 
-static int nfs3_return_delegation(struct inode *inode)
-{
-	nfs_wb_all(inode);
-	return 0;
-}
-
 static const struct inode_operations nfs3_dir_inode_operations = {
 	.create		= nfs_create,
 	.lookup		= nfs_lookup,
@@ -992,7 +986,6 @@ const struct nfs_rpc_ops nfs_v3_clientops = {
 	.clear_acl_cache = forget_all_cached_acls,
 	.close_context	= nfs_close_context,
 	.have_delegation = nfs3_have_delegation,
-	.return_delegation = nfs3_return_delegation,
 	.alloc_client	= nfs_alloc_client,
 	.init_client	= nfs_init_client,
 	.free_client	= nfs_free_client,
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index 173089f5bf7e..04978bfd6beb 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -9654,7 +9654,6 @@ const struct nfs_rpc_ops nfs_v4_clientops = {
 	.close_context  = nfs4_close_context,
 	.open_context	= nfs4_atomic_open,
 	.have_delegation = nfs4_have_delegation,
-	.return_delegation = nfs4_inode_return_delegation,
 	.alloc_client	= nfs4_alloc_client,
 	.init_client	= nfs4_init_client,
 	.free_client	= nfs4_free_client,
diff --git a/fs/nfs/proc.c b/fs/nfs/proc.c
index 318b3f34a6d0..4e93d6308733 100644
--- a/fs/nfs/proc.c
+++ b/fs/nfs/proc.c
@@ -673,12 +673,6 @@ static int nfs_have_delegation(struct inode *inode, fmode_t flags)
 	return 0;
 }
 
-static int nfs_return_delegation(struct inode *inode)
-{
-	nfs_wb_all(inode);
-	return 0;
-}
-
 static const struct inode_operations nfs_dir_inode_operations = {
 	.create		= nfs_create,
 	.lookup		= nfs_lookup,
@@ -743,7 +737,6 @@ const struct nfs_rpc_ops nfs_v2_clientops = {
 	.lock_check_bounds = nfs_lock_check_bounds,
 	.close_context	= nfs_close_context,
 	.have_delegation = nfs_have_delegation,
-	.return_delegation = nfs_return_delegation,
 	.alloc_client	= nfs_alloc_client,
 	.init_client	= nfs_init_client,
 	.free_client	= nfs_free_client,
diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h
index c4ba58b3c0f8..34d28564ecf3 100644
--- a/include/linux/nfs_xdr.h
+++ b/include/linux/nfs_xdr.h
@@ -1635,7 +1635,6 @@ struct nfs_rpc_ops {
 				struct iattr *iattr,
 				int *);
 	int (*have_delegation)(struct inode *, fmode_t);
-	int (*return_delegation)(struct inode *);
 	struct nfs_client *(*alloc_client) (const struct nfs_client_initdata *);
 	struct nfs_client *(*init_client) (struct nfs_client *,
 				const struct nfs_client_initdata *);
-- 
cgit v1.2.3-58-ga151


From 909728821e366599b899251e0a0f023c0ccf4fb0 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <trond.myklebust@primarydata.com>
Date: Tue, 20 Mar 2018 16:53:27 -0400
Subject: NFS: Convert NFS_INO_INVALID flags to unsigned long

The cache validity attribute is unsigned long, so make sure that
the flags are too.

Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 include/linux/nfs_fs.h | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

(limited to 'include')

diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h
index 38187c68063d..4afb11be73f4 100644
--- a/include/linux/nfs_fs.h
+++ b/include/linux/nfs_fs.h
@@ -198,14 +198,14 @@ struct nfs_inode {
 /*
  * Cache validity bit flags
  */
-#define NFS_INO_INVALID_ATTR	0x0001		/* cached attrs are invalid */
-#define NFS_INO_INVALID_DATA	0x0002		/* cached data is invalid */
-#define NFS_INO_INVALID_ATIME	0x0004		/* cached atime is invalid */
-#define NFS_INO_INVALID_ACCESS	0x0008		/* cached access cred invalid */
-#define NFS_INO_INVALID_ACL	0x0010		/* cached acls are invalid */
-#define NFS_INO_REVAL_PAGECACHE	0x0020		/* must revalidate pagecache */
-#define NFS_INO_REVAL_FORCED	0x0040		/* force revalidation ignoring a delegation */
-#define NFS_INO_INVALID_LABEL	0x0080		/* cached label is invalid */
+#define NFS_INO_INVALID_ATTR	BIT(0)		/* cached attrs are invalid */
+#define NFS_INO_INVALID_DATA	BIT(1)		/* cached data is invalid */
+#define NFS_INO_INVALID_ATIME	BIT(2)		/* cached atime is invalid */
+#define NFS_INO_INVALID_ACCESS	BIT(3)		/* cached access cred invalid */
+#define NFS_INO_INVALID_ACL	BIT(4)		/* cached acls are invalid */
+#define NFS_INO_REVAL_PAGECACHE	BIT(5)		/* must revalidate pagecache */
+#define NFS_INO_REVAL_FORCED	BIT(6)		/* force revalidation ignoring a delegation */
+#define NFS_INO_INVALID_LABEL	BIT(7)		/* cached label is invalid */
 
 /*
  * Bit offsets in flags field
-- 
cgit v1.2.3-58-ga151


From 16e143751727471f9a565515344196693bbc8762 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <trond.myklebust@primarydata.com>
Date: Tue, 20 Mar 2018 16:53:31 -0400
Subject: NFS: More fine grained attribute tracking

Currently, if the NFS_INO_INVALID_ATTR flag is set, for instance by
a call to nfs_post_op_update_inode_locked(), then it will not be cleared
until all the attributes have been revalidated. This means, for instance,
that NFSv4 writes will always force a full attribute revalidation.

Track the ctime, mtime, size and change attribute separately from the
other attributes so that we can have nfs_post_op_update_inode_locked()
set them correctly, and later have the cache consistency bitmask be
able to clear them.

Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 fs/nfs/dir.c           |   4 +-
 fs/nfs/inode.c         | 109 +++++++++++++++++++++++++++++--------------------
 fs/nfs/nfs4proc.c      |   7 +++-
 fs/nfs/write.c         |   7 +++-
 include/linux/nfs_fs.h |  21 +++++++---
 5 files changed, 94 insertions(+), 54 deletions(-)

(limited to 'include')

diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index 8001f8c7ad0e..73f8b43d988c 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -1272,7 +1272,9 @@ static void nfs_drop_nlink(struct inode *inode)
 	/* drop the inode if we're reasonably sure this is the last link */
 	if (inode->i_nlink == 1)
 		clear_nlink(inode);
-	NFS_I(inode)->cache_validity |= NFS_INO_INVALID_ATTR;
+	NFS_I(inode)->cache_validity |= NFS_INO_INVALID_CHANGE
+		| NFS_INO_INVALID_CTIME
+		| NFS_INO_INVALID_OTHER;
 	spin_unlock(&inode->i_lock);
 }
 
diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
index 6e5a96e2e9a0..23880f45c1e4 100644
--- a/fs/nfs/inode.c
+++ b/fs/nfs/inode.c
@@ -452,7 +452,7 @@ nfs_fhget(struct super_block *sb, struct nfs_fh *fh, struct nfs_fattr *fattr, st
 		inode->i_mode = fattr->mode;
 		if ((fattr->valid & NFS_ATTR_FATTR_MODE) == 0
 				&& nfs_server_capable(inode, NFS_CAP_MODE))
-			nfs_set_cache_invalid(inode, NFS_INO_INVALID_ATTR);
+			nfs_set_cache_invalid(inode, NFS_INO_INVALID_OTHER);
 		/* Why so? Because we want revalidate for devices/FIFOs, and
 		 * that's precisely what we have in nfs_file_inode_operations.
 		 */
@@ -498,37 +498,35 @@ nfs_fhget(struct super_block *sb, struct nfs_fh *fh, struct nfs_fattr *fattr, st
 		if (fattr->valid & NFS_ATTR_FATTR_ATIME)
 			inode->i_atime = fattr->atime;
 		else if (nfs_server_capable(inode, NFS_CAP_ATIME))
-			nfs_set_cache_invalid(inode, NFS_INO_INVALID_ATTR);
+			nfs_set_cache_invalid(inode, NFS_INO_INVALID_ATIME);
 		if (fattr->valid & NFS_ATTR_FATTR_MTIME)
 			inode->i_mtime = fattr->mtime;
 		else if (nfs_server_capable(inode, NFS_CAP_MTIME))
-			nfs_set_cache_invalid(inode, NFS_INO_INVALID_ATTR);
+			nfs_set_cache_invalid(inode, NFS_INO_INVALID_MTIME);
 		if (fattr->valid & NFS_ATTR_FATTR_CTIME)
 			inode->i_ctime = fattr->ctime;
 		else if (nfs_server_capable(inode, NFS_CAP_CTIME))
-			nfs_set_cache_invalid(inode, NFS_INO_INVALID_ATTR);
+			nfs_set_cache_invalid(inode, NFS_INO_INVALID_CTIME);
 		if (fattr->valid & NFS_ATTR_FATTR_CHANGE)
 			inode_set_iversion_raw(inode, fattr->change_attr);
 		else
-			nfs_set_cache_invalid(inode, NFS_INO_INVALID_ATTR
-				| NFS_INO_REVAL_PAGECACHE);
+			nfs_set_cache_invalid(inode, NFS_INO_INVALID_CHANGE);
 		if (fattr->valid & NFS_ATTR_FATTR_SIZE)
 			inode->i_size = nfs_size_to_loff_t(fattr->size);
 		else
-			nfs_set_cache_invalid(inode, NFS_INO_INVALID_ATTR
-				| NFS_INO_REVAL_PAGECACHE);
+			nfs_set_cache_invalid(inode, NFS_INO_INVALID_SIZE);
 		if (fattr->valid & NFS_ATTR_FATTR_NLINK)
 			set_nlink(inode, fattr->nlink);
 		else if (nfs_server_capable(inode, NFS_CAP_NLINK))
-			nfs_set_cache_invalid(inode, NFS_INO_INVALID_ATTR);
+			nfs_set_cache_invalid(inode, NFS_INO_INVALID_OTHER);
 		if (fattr->valid & NFS_ATTR_FATTR_OWNER)
 			inode->i_uid = fattr->uid;
 		else if (nfs_server_capable(inode, NFS_CAP_OWNER))
-			nfs_set_cache_invalid(inode, NFS_INO_INVALID_ATTR);
+			nfs_set_cache_invalid(inode, NFS_INO_INVALID_OTHER);
 		if (fattr->valid & NFS_ATTR_FATTR_GROUP)
 			inode->i_gid = fattr->gid;
 		else if (nfs_server_capable(inode, NFS_CAP_OWNER_GROUP))
-			nfs_set_cache_invalid(inode, NFS_INO_INVALID_ATTR);
+			nfs_set_cache_invalid(inode, NFS_INO_INVALID_OTHER);
 		if (fattr->valid & NFS_ATTR_FATTR_BLOCKS_USED)
 			inode->i_blocks = fattr->du.nfs2.blocks;
 		if (fattr->valid & NFS_ATTR_FATTR_SPACE_USED) {
@@ -657,6 +655,7 @@ out:
  * nfs_setattr_update_inode - Update inode metadata after a setattr call.
  * @inode: pointer to struct inode
  * @attr: pointer to struct iattr
+ * @fattr: pointer to struct nfs_fattr
  *
  * Note: we do this in the *proc.c in order to ensure that
  *       it works for things like exclusive creates too.
@@ -669,6 +668,8 @@ void nfs_setattr_update_inode(struct inode *inode, struct iattr *attr,
 
 	spin_lock(&inode->i_lock);
 	NFS_I(inode)->attr_gencount = fattr->gencount;
+	nfs_set_cache_invalid(inode, NFS_INO_INVALID_CHANGE
+			| NFS_INO_INVALID_CTIME);
 	if ((attr->ia_valid & (ATTR_MODE|ATTR_UID|ATTR_GID)) != 0) {
 		if ((attr->ia_valid & ATTR_MODE) != 0) {
 			int mode = attr->ia_mode & S_IALLUGO;
@@ -683,13 +684,12 @@ void nfs_setattr_update_inode(struct inode *inode, struct iattr *attr,
 				| NFS_INO_INVALID_ACL);
 	}
 	if ((attr->ia_valid & ATTR_SIZE) != 0) {
+		nfs_set_cache_invalid(inode, NFS_INO_INVALID_MTIME);
 		nfs_inc_stats(inode, NFSIOS_SETATTRTRUNC);
 		nfs_vmtruncate(inode, attr->ia_size);
 	}
 	if (fattr->valid)
 		nfs_update_inode(inode, fattr);
-	else
-		NFS_I(inode)->cache_validity |= NFS_INO_INVALID_ATTR;
 	spin_unlock(&inode->i_lock);
 }
 EXPORT_SYMBOL_GPL(nfs_setattr_update_inode);
@@ -1361,33 +1361,41 @@ static int nfs_check_inode_attributes(struct inode *inode, struct nfs_fattr *fat
 	if (!nfs_file_has_buffered_writers(nfsi)) {
 		/* Verify a few of the more important attributes */
 		if ((fattr->valid & NFS_ATTR_FATTR_CHANGE) != 0 && !inode_eq_iversion_raw(inode, fattr->change_attr))
-			invalid |= NFS_INO_INVALID_ATTR | NFS_INO_REVAL_PAGECACHE;
+			invalid |= NFS_INO_INVALID_CHANGE
+				| NFS_INO_REVAL_PAGECACHE;
 
 		if ((fattr->valid & NFS_ATTR_FATTR_MTIME) && !timespec_equal(&inode->i_mtime, &fattr->mtime))
-			invalid |= NFS_INO_INVALID_ATTR;
+			invalid |= NFS_INO_INVALID_MTIME;
 
 		if ((fattr->valid & NFS_ATTR_FATTR_CTIME) && !timespec_equal(&inode->i_ctime, &fattr->ctime))
-			invalid |= NFS_INO_INVALID_ATTR;
+			invalid |= NFS_INO_INVALID_CTIME;
 
 		if (fattr->valid & NFS_ATTR_FATTR_SIZE) {
 			cur_size = i_size_read(inode);
 			new_isize = nfs_size_to_loff_t(fattr->size);
 			if (cur_size != new_isize)
-				invalid |= NFS_INO_INVALID_ATTR|NFS_INO_REVAL_PAGECACHE;
+				invalid |= NFS_INO_INVALID_SIZE
+					| NFS_INO_REVAL_PAGECACHE;
 		}
 	}
 
 	/* Have any file permissions changed? */
 	if ((fattr->valid & NFS_ATTR_FATTR_MODE) && (inode->i_mode & S_IALLUGO) != (fattr->mode & S_IALLUGO))
-		invalid |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ACCESS | NFS_INO_INVALID_ACL;
+		invalid |= NFS_INO_INVALID_ACCESS
+			| NFS_INO_INVALID_ACL
+			| NFS_INO_INVALID_OTHER;
 	if ((fattr->valid & NFS_ATTR_FATTR_OWNER) && !uid_eq(inode->i_uid, fattr->uid))
-		invalid |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ACCESS | NFS_INO_INVALID_ACL;
+		invalid |= NFS_INO_INVALID_ACCESS
+			| NFS_INO_INVALID_ACL
+			| NFS_INO_INVALID_OTHER;
 	if ((fattr->valid & NFS_ATTR_FATTR_GROUP) && !gid_eq(inode->i_gid, fattr->gid))
-		invalid |= NFS_INO_INVALID_ATTR | NFS_INO_INVALID_ACCESS | NFS_INO_INVALID_ACL;
+		invalid |= NFS_INO_INVALID_ACCESS
+			| NFS_INO_INVALID_ACL
+			| NFS_INO_INVALID_OTHER;
 
 	/* Has the link count changed? */
 	if ((fattr->valid & NFS_ATTR_FATTR_NLINK) && inode->i_nlink != fattr->nlink)
-		invalid |= NFS_INO_INVALID_ATTR;
+		invalid |= NFS_INO_INVALID_OTHER;
 
 	if ((fattr->valid & NFS_ATTR_FATTR_ATIME) && !timespec_equal(&inode->i_atime, &fattr->atime))
 		invalid |= NFS_INO_INVALID_ATIME;
@@ -1589,10 +1597,9 @@ int nfs_refresh_inode(struct inode *inode, struct nfs_fattr *fattr)
 }
 EXPORT_SYMBOL_GPL(nfs_refresh_inode);
 
-static int nfs_post_op_update_inode_locked(struct inode *inode, struct nfs_fattr *fattr)
+static int nfs_post_op_update_inode_locked(struct inode *inode,
+		struct nfs_fattr *fattr, unsigned int invalid)
 {
-	unsigned long invalid = NFS_INO_INVALID_ATTR;
-
 	if (S_ISDIR(inode->i_mode))
 		invalid |= NFS_INO_INVALID_DATA;
 	nfs_set_cache_invalid(inode, invalid);
@@ -1621,7 +1628,9 @@ int nfs_post_op_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 
 	spin_lock(&inode->i_lock);
 	nfs_fattr_set_barrier(fattr);
-	status = nfs_post_op_update_inode_locked(inode, fattr);
+	status = nfs_post_op_update_inode_locked(inode, fattr,
+			NFS_INO_INVALID_CHANGE
+			| NFS_INO_INVALID_CTIME);
 	spin_unlock(&inode->i_lock);
 
 	return status;
@@ -1673,7 +1682,10 @@ int nfs_post_op_update_inode_force_wcc_locked(struct inode *inode, struct nfs_fa
 		fattr->valid |= NFS_ATTR_FATTR_PRESIZE;
 	}
 out_noforce:
-	status = nfs_post_op_update_inode_locked(inode, fattr);
+	status = nfs_post_op_update_inode_locked(inode, fattr,
+			NFS_INO_INVALID_CHANGE
+			| NFS_INO_INVALID_CTIME
+			| NFS_INO_INVALID_MTIME);
 	return status;
 }
 
@@ -1795,12 +1807,15 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 					inode->i_sb->s_id, inode->i_ino);
 			/* Could it be a race with writeback? */
 			if (!have_writers) {
-				invalid |= NFS_INO_INVALID_ATTR
+				invalid |= NFS_INO_INVALID_CHANGE
 					| NFS_INO_INVALID_DATA
 					| NFS_INO_INVALID_ACCESS
 					| NFS_INO_INVALID_ACL;
 				/* Force revalidate of all attributes */
-				save_cache_validity |= NFS_INO_INVALID_ATTR;
+				save_cache_validity |= NFS_INO_INVALID_CTIME
+					| NFS_INO_INVALID_MTIME
+					| NFS_INO_INVALID_SIZE
+					| NFS_INO_INVALID_OTHER;
 				if (S_ISDIR(inode->i_mode))
 					nfs_force_lookup_revalidate(inode);
 			}
@@ -1808,7 +1823,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 		}
 	} else {
 		nfsi->cache_validity |= save_cache_validity &
-				(NFS_INO_INVALID_ATTR
+				(NFS_INO_INVALID_CHANGE
 				| NFS_INO_REVAL_PAGECACHE
 				| NFS_INO_REVAL_FORCED);
 		cache_revalidated = false;
@@ -1818,7 +1833,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 		memcpy(&inode->i_mtime, &fattr->mtime, sizeof(inode->i_mtime));
 	} else if (server->caps & NFS_CAP_MTIME) {
 		nfsi->cache_validity |= save_cache_validity &
-				(NFS_INO_INVALID_ATTR
+				(NFS_INO_INVALID_MTIME
 				| NFS_INO_REVAL_FORCED);
 		cache_revalidated = false;
 	}
@@ -1827,7 +1842,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 		memcpy(&inode->i_ctime, &fattr->ctime, sizeof(inode->i_ctime));
 	} else if (server->caps & NFS_CAP_CTIME) {
 		nfsi->cache_validity |= save_cache_validity &
-				(NFS_INO_INVALID_ATTR
+				(NFS_INO_INVALID_CTIME
 				| NFS_INO_REVAL_FORCED);
 		cache_revalidated = false;
 	}
@@ -1842,7 +1857,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 			if (!nfs_have_writebacks(inode) || new_isize > cur_isize) {
 				i_size_write(inode, new_isize);
 				if (!have_writers)
-					invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA;
+					invalid |= NFS_INO_INVALID_DATA;
 			}
 			dprintk("NFS: isize change on server for file %s/%ld "
 					"(%Ld to %Ld)\n",
@@ -1853,7 +1868,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 		}
 	} else {
 		nfsi->cache_validity |= save_cache_validity &
-				(NFS_INO_INVALID_ATTR
+				(NFS_INO_INVALID_SIZE
 				| NFS_INO_REVAL_PAGECACHE
 				| NFS_INO_REVAL_FORCED);
 		cache_revalidated = false;
@@ -1874,55 +1889,61 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 			umode_t newmode = inode->i_mode & S_IFMT;
 			newmode |= fattr->mode & S_IALLUGO;
 			inode->i_mode = newmode;
-			invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
+			invalid |= NFS_INO_INVALID_ACCESS
+				| NFS_INO_INVALID_ACL
+				| NFS_INO_INVALID_OTHER;
 		}
 	} else if (server->caps & NFS_CAP_MODE) {
 		nfsi->cache_validity |= save_cache_validity &
-				(NFS_INO_INVALID_ATTR
-				| NFS_INO_INVALID_ACCESS
+				(NFS_INO_INVALID_ACCESS
 				| NFS_INO_INVALID_ACL
+				| NFS_INO_INVALID_OTHER
 				| NFS_INO_REVAL_FORCED);
 		cache_revalidated = false;
 	}
 
 	if (fattr->valid & NFS_ATTR_FATTR_OWNER) {
 		if (!uid_eq(inode->i_uid, fattr->uid)) {
-			invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
+			invalid |= NFS_INO_INVALID_ACCESS
+				| NFS_INO_INVALID_ACL
+				| NFS_INO_INVALID_OTHER;
 			inode->i_uid = fattr->uid;
 		}
 	} else if (server->caps & NFS_CAP_OWNER) {
 		nfsi->cache_validity |= save_cache_validity &
-				(NFS_INO_INVALID_ATTR
-				| NFS_INO_INVALID_ACCESS
+				(NFS_INO_INVALID_ACCESS
 				| NFS_INO_INVALID_ACL
+				| NFS_INO_INVALID_OTHER
 				| NFS_INO_REVAL_FORCED);
 		cache_revalidated = false;
 	}
 
 	if (fattr->valid & NFS_ATTR_FATTR_GROUP) {
 		if (!gid_eq(inode->i_gid, fattr->gid)) {
-			invalid |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ACCESS|NFS_INO_INVALID_ACL;
+			invalid |= NFS_INO_INVALID_ACCESS
+				| NFS_INO_INVALID_ACL
+				| NFS_INO_INVALID_OTHER;
 			inode->i_gid = fattr->gid;
 		}
 	} else if (server->caps & NFS_CAP_OWNER_GROUP) {
 		nfsi->cache_validity |= save_cache_validity &
-				(NFS_INO_INVALID_ATTR
-				| NFS_INO_INVALID_ACCESS
+				(NFS_INO_INVALID_ACCESS
 				| NFS_INO_INVALID_ACL
+				| NFS_INO_INVALID_OTHER
 				| NFS_INO_REVAL_FORCED);
 		cache_revalidated = false;
 	}
 
 	if (fattr->valid & NFS_ATTR_FATTR_NLINK) {
 		if (inode->i_nlink != fattr->nlink) {
-			invalid |= NFS_INO_INVALID_ATTR;
+			invalid |= NFS_INO_INVALID_OTHER;
 			if (S_ISDIR(inode->i_mode))
 				invalid |= NFS_INO_INVALID_DATA;
 			set_nlink(inode, fattr->nlink);
 		}
 	} else if (server->caps & NFS_CAP_NLINK) {
 		nfsi->cache_validity |= save_cache_validity &
-				(NFS_INO_INVALID_ATTR
+				(NFS_INO_INVALID_OTHER
 				| NFS_INO_REVAL_FORCED);
 		cache_revalidated = false;
 	}
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index 24b5c80b8128..e5ad2c186927 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -1045,7 +1045,9 @@ static void update_changeattr(struct inode *dir, struct nfs4_change_info *cinfo,
 	struct nfs_inode *nfsi = NFS_I(dir);
 
 	spin_lock(&dir->i_lock);
-	nfsi->cache_validity |= NFS_INO_INVALID_ATTR|NFS_INO_INVALID_DATA;
+	nfsi->cache_validity |= NFS_INO_INVALID_CTIME
+		| NFS_INO_INVALID_MTIME
+		| NFS_INO_INVALID_DATA;
 	if (cinfo->atomic && cinfo->before == inode_peek_iversion_raw(dir)) {
 		nfsi->cache_validity &= ~NFS_INO_REVAL_PAGECACHE;
 		nfsi->attrtimeo_timestamp = jiffies;
@@ -5364,7 +5366,8 @@ static int __nfs4_proc_set_acl(struct inode *inode, const void *buf, size_t bufl
 	 * so mark the attribute cache invalid.
 	 */
 	spin_lock(&inode->i_lock);
-	NFS_I(inode)->cache_validity |= NFS_INO_INVALID_ATTR;
+	NFS_I(inode)->cache_validity |= NFS_INO_INVALID_CHANGE
+		| NFS_INO_INVALID_CTIME;
 	spin_unlock(&inode->i_lock);
 	nfs_access_zap_cache(inode);
 	nfs_zap_acl_cache(inode);
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index 7428a669d7a7..3efce54ef1cd 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -1562,8 +1562,11 @@ static int nfs_writeback_done(struct rpc_task *task,
 	}
 
 	/* Deal with the suid/sgid bit corner case */
-	if (nfs_should_remove_suid(inode))
-		nfs_mark_for_revalidate(inode);
+	if (nfs_should_remove_suid(inode)) {
+		spin_lock(&inode->i_lock);
+		NFS_I(inode)->cache_validity |= NFS_INO_INVALID_OTHER;
+		spin_unlock(&inode->i_lock);
+	}
 	return 0;
 }
 
diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h
index 4afb11be73f4..2f129bbfaae8 100644
--- a/include/linux/nfs_fs.h
+++ b/include/linux/nfs_fs.h
@@ -198,7 +198,6 @@ struct nfs_inode {
 /*
  * Cache validity bit flags
  */
-#define NFS_INO_INVALID_ATTR	BIT(0)		/* cached attrs are invalid */
 #define NFS_INO_INVALID_DATA	BIT(1)		/* cached data is invalid */
 #define NFS_INO_INVALID_ATIME	BIT(2)		/* cached atime is invalid */
 #define NFS_INO_INVALID_ACCESS	BIT(3)		/* cached access cred invalid */
@@ -206,6 +205,17 @@ struct nfs_inode {
 #define NFS_INO_REVAL_PAGECACHE	BIT(5)		/* must revalidate pagecache */
 #define NFS_INO_REVAL_FORCED	BIT(6)		/* force revalidation ignoring a delegation */
 #define NFS_INO_INVALID_LABEL	BIT(7)		/* cached label is invalid */
+#define NFS_INO_INVALID_CHANGE	BIT(8)		/* cached change is invalid */
+#define NFS_INO_INVALID_CTIME	BIT(9)		/* cached ctime is invalid */
+#define NFS_INO_INVALID_MTIME	BIT(10)		/* cached mtime is invalid */
+#define NFS_INO_INVALID_SIZE	BIT(11)		/* cached size is invalid */
+#define NFS_INO_INVALID_OTHER	BIT(12)		/* other attrs are invalid */
+
+#define NFS_INO_INVALID_ATTR	(NFS_INO_INVALID_CHANGE \
+		| NFS_INO_INVALID_CTIME \
+		| NFS_INO_INVALID_MTIME \
+		| NFS_INO_INVALID_SIZE \
+		| NFS_INO_INVALID_OTHER)	/* inode metadata is invalid */
 
 /*
  * Bit offsets in flags field
@@ -292,10 +302,11 @@ static inline void nfs_mark_for_revalidate(struct inode *inode)
 	struct nfs_inode *nfsi = NFS_I(inode);
 
 	spin_lock(&inode->i_lock);
-	nfsi->cache_validity |= NFS_INO_INVALID_ATTR |
-				NFS_INO_REVAL_PAGECACHE |
-				NFS_INO_INVALID_ACCESS |
-				NFS_INO_INVALID_ACL;
+	nfsi->cache_validity |= NFS_INO_REVAL_PAGECACHE
+		| NFS_INO_INVALID_ACCESS
+		| NFS_INO_INVALID_ACL
+		| NFS_INO_INVALID_CHANGE
+		| NFS_INO_INVALID_CTIME;
 	if (S_ISDIR(inode->i_mode))
 		nfsi->cache_validity |= NFS_INO_INVALID_DATA;
 	spin_unlock(&inode->i_lock);
-- 
cgit v1.2.3-58-ga151


From 0e779aa70308462e45f7cd1a54de418dfe101694 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <trond.myklebust@primarydata.com>
Date: Tue, 20 Mar 2018 17:03:05 -0400
Subject: SUNRPC: Add helpers for decoding opaque and string types

Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 include/linux/sunrpc/xdr.h |  6 ++++
 net/sunrpc/xdr.c           | 82 ++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 88 insertions(+)

(limited to 'include')

diff --git a/include/linux/sunrpc/xdr.h b/include/linux/sunrpc/xdr.h
index d950223c64b1..7e609de34d85 100644
--- a/include/linux/sunrpc/xdr.h
+++ b/include/linux/sunrpc/xdr.h
@@ -253,6 +253,12 @@ xdr_stream_remaining(const struct xdr_stream *xdr)
 	return xdr->nwords << 2;
 }
 
+ssize_t xdr_stream_decode_opaque(struct xdr_stream *xdr, void *ptr,
+		size_t size);
+ssize_t xdr_stream_decode_opaque_dup(struct xdr_stream *xdr, void **ptr,
+		size_t maxlen, gfp_t gfp_flags);
+ssize_t xdr_stream_decode_string(struct xdr_stream *xdr, char *str,
+		size_t size);
 ssize_t xdr_stream_decode_string_dup(struct xdr_stream *xdr, char **str,
 		size_t maxlen, gfp_t gfp_flags);
 /**
diff --git a/net/sunrpc/xdr.c b/net/sunrpc/xdr.c
index e34f4ee7f2b6..30afbd236656 100644
--- a/net/sunrpc/xdr.c
+++ b/net/sunrpc/xdr.c
@@ -1518,6 +1518,88 @@ out:
 }
 EXPORT_SYMBOL_GPL(xdr_process_buf);
 
+/**
+ * xdr_stream_decode_opaque - Decode variable length opaque
+ * @xdr: pointer to xdr_stream
+ * @ptr: location to store opaque data
+ * @size: size of storage buffer @ptr
+ *
+ * Return values:
+ *   On success, returns size of object stored in *@ptr
+ *   %-EBADMSG on XDR buffer overflow
+ *   %-EMSGSIZE on overflow of storage buffer @ptr
+ */
+ssize_t xdr_stream_decode_opaque(struct xdr_stream *xdr, void *ptr, size_t size)
+{
+	ssize_t ret;
+	void *p;
+
+	ret = xdr_stream_decode_opaque_inline(xdr, &p, size);
+	if (ret <= 0)
+		return ret;
+	memcpy(ptr, p, ret);
+	return ret;
+}
+EXPORT_SYMBOL_GPL(xdr_stream_decode_opaque);
+
+/**
+ * xdr_stream_decode_opaque_dup - Decode and duplicate variable length opaque
+ * @xdr: pointer to xdr_stream
+ * @ptr: location to store pointer to opaque data
+ * @maxlen: maximum acceptable object size
+ * @gfp_flags: GFP mask to use
+ *
+ * Return values:
+ *   On success, returns size of object stored in *@ptr
+ *   %-EBADMSG on XDR buffer overflow
+ *   %-EMSGSIZE if the size of the object would exceed @maxlen
+ *   %-ENOMEM on memory allocation failure
+ */
+ssize_t xdr_stream_decode_opaque_dup(struct xdr_stream *xdr, void **ptr,
+		size_t maxlen, gfp_t gfp_flags)
+{
+	ssize_t ret;
+	void *p;
+
+	ret = xdr_stream_decode_opaque_inline(xdr, &p, maxlen);
+	if (ret > 0) {
+		*ptr = kmemdup(p, ret, gfp_flags);
+		if (*ptr != NULL)
+			return ret;
+		ret = -ENOMEM;
+	}
+	*ptr = NULL;
+	return ret;
+}
+EXPORT_SYMBOL_GPL(xdr_stream_decode_opaque_dup);
+
+/**
+ * xdr_stream_decode_string - Decode variable length string
+ * @xdr: pointer to xdr_stream
+ * @str: location to store string
+ * @size: size of storage buffer @str
+ *
+ * Return values:
+ *   On success, returns length of NUL-terminated string stored in *@str
+ *   %-EBADMSG on XDR buffer overflow
+ *   %-EMSGSIZE on overflow of storage buffer @str
+ */
+ssize_t xdr_stream_decode_string(struct xdr_stream *xdr, char *str, size_t size)
+{
+	ssize_t ret;
+	void *p;
+
+	ret = xdr_stream_decode_opaque_inline(xdr, &p, size);
+	if (ret > 0) {
+		memcpy(str, p, ret);
+		str[ret] = '\0';
+		return strlen(str);
+	}
+	*str = '\0';
+	return ret;
+}
+EXPORT_SYMBOL_GPL(xdr_stream_decode_string);
+
 /**
  * xdr_stream_decode_string_dup - Decode and duplicate variable length string
  * @xdr: pointer to xdr_stream
-- 
cgit v1.2.3-58-ga151


From 85e3dd44c514a8bed6c713df4af657be83d00f68 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <trond.myklebust@primarydata.com>
Date: Tue, 20 Mar 2018 17:03:06 -0400
Subject: SUNRPC: Add a helper for encoding opaque data inline

Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 include/linux/sunrpc/xdr.h | 25 +++++++++++++++++++++++++
 1 file changed, 25 insertions(+)

(limited to 'include')

diff --git a/include/linux/sunrpc/xdr.h b/include/linux/sunrpc/xdr.h
index 7e609de34d85..a43c3b6455b6 100644
--- a/include/linux/sunrpc/xdr.h
+++ b/include/linux/sunrpc/xdr.h
@@ -318,6 +318,31 @@ xdr_stream_encode_u64(struct xdr_stream *xdr, __u64 n)
 	return len;
 }
 
+/**
+ * xdr_stream_encode_opaque_inline - Encode opaque xdr data
+ * @xdr: pointer to xdr_stream
+ * @ptr: pointer to void pointer
+ * @len: size of object
+ *
+ * Return values:
+ *   On success, returns length in bytes of XDR buffer consumed
+ *   %-EMSGSIZE on XDR buffer overflow
+ */
+static inline ssize_t
+xdr_stream_encode_opaque_inline(struct xdr_stream *xdr, void **ptr, size_t len)
+{
+	size_t count = sizeof(__u32) + xdr_align_size(len);
+	__be32 *p = xdr_reserve_space(xdr, count);
+
+	if (unlikely(!p)) {
+		*ptr = NULL;
+		return -EMSGSIZE;
+	}
+	xdr_encode_opaque(p, NULL, len);
+	*ptr = ++p;
+	return count;
+}
+
 /**
  * xdr_stream_encode_opaque_fixed - Encode fixed length opaque xdr data
  * @xdr: pointer to xdr_stream
-- 
cgit v1.2.3-58-ga151


From 37c88763def8474bc0972fbd1adb0d21670104b7 Mon Sep 17 00:00:00 2001
From: Trond Myklebust <trond.myklebust@primarydata.com>
Date: Tue, 20 Mar 2018 17:03:08 -0400
Subject: NFSv4; Clean up XDR encoding of type bitmap4

Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
---
 fs/nfs/nfs4xdr.c           | 166 ++++++++++++++++++++-------------------------
 include/linux/sunrpc/xdr.h |  63 +++++++++++++++++
 2 files changed, 135 insertions(+), 94 deletions(-)

(limited to 'include')

diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c
index 6b08f6b1addf..80c5b519fd6a 100644
--- a/fs/nfs/nfs4xdr.c
+++ b/fs/nfs/nfs4xdr.c
@@ -958,6 +958,35 @@ static void encode_uint64(struct xdr_stream *xdr, u64 n)
 	WARN_ON_ONCE(xdr_stream_encode_u64(xdr, n) < 0);
 }
 
+static ssize_t xdr_encode_bitmap4(struct xdr_stream *xdr,
+		const __u32 *bitmap, size_t len)
+{
+	ssize_t ret;
+
+	/* Trim empty words */
+	while (len > 0 && bitmap[len-1] == 0)
+		len--;
+	ret = xdr_stream_encode_uint32_array(xdr, bitmap, len);
+	if (WARN_ON_ONCE(ret < 0))
+		return ret;
+	return len;
+}
+
+static size_t mask_bitmap4(const __u32 *bitmap, const __u32 *mask,
+		__u32 *res, size_t len)
+{
+	size_t i;
+	__u32 tmp;
+
+	while (len > 0 && (bitmap[len-1] == 0 || mask[len-1] == 0))
+		len--;
+	for (i = len; i-- > 0;) {
+		tmp = bitmap[i] & mask[i];
+		res[i] = tmp;
+	}
+	return len;
+}
+
 static void encode_nfs4_seqid(struct xdr_stream *xdr,
 		const struct nfs_seqid *seqid)
 {
@@ -1200,85 +1229,45 @@ static void encode_create(struct xdr_stream *xdr, const struct nfs4_create_arg *
 			create->server, create->server->attr_bitmask);
 }
 
-static void encode_getattr_one(struct xdr_stream *xdr, uint32_t bitmap, struct compound_hdr *hdr)
-{
-	__be32 *p;
-
-	encode_op_hdr(xdr, OP_GETATTR, decode_getattr_maxsz, hdr);
-	p = reserve_space(xdr, 8);
-	*p++ = cpu_to_be32(1);
-	*p = cpu_to_be32(bitmap);
-}
-
-static void encode_getattr_two(struct xdr_stream *xdr, uint32_t bm0, uint32_t bm1, struct compound_hdr *hdr)
-{
-	__be32 *p;
-
-	encode_op_hdr(xdr, OP_GETATTR, decode_getattr_maxsz, hdr);
-	p = reserve_space(xdr, 12);
-	*p++ = cpu_to_be32(2);
-	*p++ = cpu_to_be32(bm0);
-	*p = cpu_to_be32(bm1);
-}
-
-static void
-encode_getattr_three(struct xdr_stream *xdr,
-		     uint32_t bm0, uint32_t bm1, uint32_t bm2,
-		     struct compound_hdr *hdr)
+static void encode_getattr(struct xdr_stream *xdr,
+		const __u32 *bitmap, const __u32 *mask, size_t len,
+		struct compound_hdr *hdr)
 {
-	__be32 *p;
+	__u32 masked_bitmap[nfs4_fattr_bitmap_maxsz];
 
 	encode_op_hdr(xdr, OP_GETATTR, decode_getattr_maxsz, hdr);
-	if (bm2) {
-		p = reserve_space(xdr, 16);
-		*p++ = cpu_to_be32(3);
-		*p++ = cpu_to_be32(bm0);
-		*p++ = cpu_to_be32(bm1);
-		*p = cpu_to_be32(bm2);
-	} else if (bm1) {
-		p = reserve_space(xdr, 12);
-		*p++ = cpu_to_be32(2);
-		*p++ = cpu_to_be32(bm0);
-		*p = cpu_to_be32(bm1);
-	} else {
-		p = reserve_space(xdr, 8);
-		*p++ = cpu_to_be32(1);
-		*p = cpu_to_be32(bm0);
+	if (mask) {
+		if (WARN_ON_ONCE(len > ARRAY_SIZE(masked_bitmap)))
+			len = ARRAY_SIZE(masked_bitmap);
+		len = mask_bitmap4(bitmap, mask, masked_bitmap, len);
+		bitmap = masked_bitmap;
 	}
+	xdr_encode_bitmap4(xdr, bitmap, len);
 }
 
 static void encode_getfattr(struct xdr_stream *xdr, const u32* bitmask, struct compound_hdr *hdr)
 {
-	encode_getattr_three(xdr, bitmask[0] & nfs4_fattr_bitmap[0],
-			   bitmask[1] & nfs4_fattr_bitmap[1],
-			   bitmask[2] & nfs4_fattr_bitmap[2],
-			   hdr);
+	encode_getattr(xdr, nfs4_fattr_bitmap, bitmask,
+			ARRAY_SIZE(nfs4_fattr_bitmap), hdr);
 }
 
 static void encode_getfattr_open(struct xdr_stream *xdr, const u32 *bitmask,
 				 const u32 *open_bitmap,
 				 struct compound_hdr *hdr)
 {
-	encode_getattr_three(xdr,
-			     bitmask[0] & open_bitmap[0],
-			     bitmask[1] & open_bitmap[1],
-			     bitmask[2] & open_bitmap[2],
-			     hdr);
+	encode_getattr(xdr, open_bitmap, bitmask, 3, hdr);
 }
 
 static void encode_fsinfo(struct xdr_stream *xdr, const u32* bitmask, struct compound_hdr *hdr)
 {
-	encode_getattr_three(xdr,
-			     bitmask[0] & nfs4_fsinfo_bitmap[0],
-			     bitmask[1] & nfs4_fsinfo_bitmap[1],
-			     bitmask[2] & nfs4_fsinfo_bitmap[2],
-			     hdr);
+	encode_getattr(xdr, nfs4_fsinfo_bitmap, bitmask,
+			ARRAY_SIZE(nfs4_fsinfo_bitmap), hdr);
 }
 
 static void encode_fs_locations(struct xdr_stream *xdr, const u32* bitmask, struct compound_hdr *hdr)
 {
-	encode_getattr_two(xdr, bitmask[0] & nfs4_fs_locations_bitmap[0],
-			   bitmask[1] & nfs4_fs_locations_bitmap[1], hdr);
+	encode_getattr(xdr, nfs4_fs_locations_bitmap, bitmask,
+			ARRAY_SIZE(nfs4_fs_locations_bitmap), hdr);
 }
 
 static void encode_getfh(struct xdr_stream *xdr, struct compound_hdr *hdr)
@@ -2559,13 +2548,17 @@ static void nfs4_xdr_enc_getacl(struct rpc_rqst *req, struct xdr_stream *xdr,
 	struct compound_hdr hdr = {
 		.minorversion = nfs4_xdr_minorversion(&args->seq_args),
 	};
+	const __u32 nfs4_acl_bitmap[1] = {
+		[0] = FATTR4_WORD0_ACL,
+	};
 	uint32_t replen;
 
 	encode_compound_hdr(xdr, req, &hdr);
 	encode_sequence(xdr, &args->seq_args, &hdr);
 	encode_putfh(xdr, args->fh, &hdr);
 	replen = hdr.replen + op_decode_hdr_maxsz;
-	encode_getattr_two(xdr, FATTR4_WORD0_ACL, 0, &hdr);
+	encode_getattr(xdr, nfs4_acl_bitmap, NULL,
+			ARRAY_SIZE(nfs4_acl_bitmap), &hdr);
 
 	xdr_inline_pages(&req->rq_rcv_buf, replen << 2,
 		args->acl_pages, 0, args->acl_len);
@@ -2644,8 +2637,8 @@ static void nfs4_xdr_enc_pathconf(struct rpc_rqst *req, struct xdr_stream *xdr,
 	encode_compound_hdr(xdr, req, &hdr);
 	encode_sequence(xdr, &args->seq_args, &hdr);
 	encode_putfh(xdr, args->fh, &hdr);
-	encode_getattr_one(xdr, args->bitmask[0] & nfs4_pathconf_bitmap[0],
-			   &hdr);
+	encode_getattr(xdr, nfs4_pathconf_bitmap, args->bitmask,
+			ARRAY_SIZE(nfs4_pathconf_bitmap), &hdr);
 	encode_nops(&hdr);
 }
 
@@ -2663,8 +2656,8 @@ static void nfs4_xdr_enc_statfs(struct rpc_rqst *req, struct xdr_stream *xdr,
 	encode_compound_hdr(xdr, req, &hdr);
 	encode_sequence(xdr, &args->seq_args, &hdr);
 	encode_putfh(xdr, args->fh, &hdr);
-	encode_getattr_two(xdr, args->bitmask[0] & nfs4_statfs_bitmap[0],
-			   args->bitmask[1] & nfs4_statfs_bitmap[1], &hdr);
+	encode_getattr(xdr, nfs4_statfs_bitmap, args->bitmask,
+			ARRAY_SIZE(nfs4_statfs_bitmap), &hdr);
 	encode_nops(&hdr);
 }
 
@@ -2684,7 +2677,7 @@ static void nfs4_xdr_enc_server_caps(struct rpc_rqst *req,
 	encode_compound_hdr(xdr, req, &hdr);
 	encode_sequence(xdr, &args->seq_args, &hdr);
 	encode_putfh(xdr, args->fhandle, &hdr);
-	encode_getattr_three(xdr, bitmask[0], bitmask[1], bitmask[2], &hdr);
+	encode_getattr(xdr, bitmask, NULL, 3, &hdr);
 	encode_nops(&hdr);
 }
 
@@ -3218,34 +3211,27 @@ static int decode_ace(struct xdr_stream *xdr, void *ace)
 	return -EIO;
 }
 
-static int decode_attr_bitmap(struct xdr_stream *xdr, uint32_t *bitmap)
+static ssize_t
+decode_bitmap4(struct xdr_stream *xdr, uint32_t *bitmap, size_t sz)
 {
-	uint32_t bmlen;
-	__be32 *p;
-
-	p = xdr_inline_decode(xdr, 4);
-	if (unlikely(!p))
-		goto out_overflow;
-	bmlen = be32_to_cpup(p);
+	ssize_t ret;
 
-	bitmap[0] = bitmap[1] = bitmap[2] = 0;
-	p = xdr_inline_decode(xdr, (bmlen << 2));
-	if (unlikely(!p))
-		goto out_overflow;
-	if (bmlen > 0) {
-		bitmap[0] = be32_to_cpup(p++);
-		if (bmlen > 1) {
-			bitmap[1] = be32_to_cpup(p++);
-			if (bmlen > 2)
-				bitmap[2] = be32_to_cpup(p);
-		}
-	}
-	return 0;
-out_overflow:
+	ret = xdr_stream_decode_uint32_array(xdr, bitmap, sz);
+	if (likely(ret >= 0))
+		return ret;
+	if (ret == -EMSGSIZE)
+		return sz;
 	print_overflow_msg(__func__, xdr);
 	return -EIO;
 }
 
+static int decode_attr_bitmap(struct xdr_stream *xdr, uint32_t *bitmap)
+{
+	ssize_t ret;
+	ret = decode_bitmap4(xdr, bitmap, 3);
+	return ret < 0 ? ret : 0;
+}
+
 static int decode_attr_length(struct xdr_stream *xdr, uint32_t *attrlen, unsigned int *savep)
 {
 	__be32 *p;
@@ -5471,21 +5457,13 @@ decode_savefh(struct xdr_stream *xdr)
 
 static int decode_setattr(struct xdr_stream *xdr)
 {
-	__be32 *p;
-	uint32_t bmlen;
 	int status;
 
 	status = decode_op_hdr(xdr, OP_SETATTR);
 	if (status)
 		return status;
-	p = xdr_inline_decode(xdr, 4);
-	if (unlikely(!p))
-		goto out_overflow;
-	bmlen = be32_to_cpup(p);
-	p = xdr_inline_decode(xdr, bmlen << 2);
-	if (likely(p))
+	if (decode_bitmap4(xdr, NULL, 0) >= 0)
 		return 0;
-out_overflow:
 	print_overflow_msg(__func__, xdr);
 	return -EIO;
 }
diff --git a/include/linux/sunrpc/xdr.h b/include/linux/sunrpc/xdr.h
index a43c3b6455b6..2bd68177a442 100644
--- a/include/linux/sunrpc/xdr.h
+++ b/include/linux/sunrpc/xdr.h
@@ -386,6 +386,31 @@ xdr_stream_encode_opaque(struct xdr_stream *xdr, const void *ptr, size_t len)
 	return count;
 }
 
+/**
+ * xdr_stream_encode_uint32_array - Encode variable length array of integers
+ * @xdr: pointer to xdr_stream
+ * @array: array of integers
+ * @array_size: number of elements in @array
+ *
+ * Return values:
+ *   On success, returns length in bytes of XDR buffer consumed
+ *   %-EMSGSIZE on XDR buffer overflow
+ */
+static inline ssize_t
+xdr_stream_encode_uint32_array(struct xdr_stream *xdr,
+		const __u32 *array, size_t array_size)
+{
+	ssize_t ret = (array_size+1) * sizeof(__u32);
+	__be32 *p = xdr_reserve_space(xdr, ret);
+
+	if (unlikely(!p))
+		return -EMSGSIZE;
+	*p++ = cpu_to_be32(array_size);
+	for (; array_size > 0; p++, array++, array_size--)
+		*p = cpu_to_be32p(array);
+	return ret;
+}
+
 /**
  * xdr_stream_decode_u32 - Decode a 32-bit integer
  * @xdr: pointer to xdr_stream
@@ -463,6 +488,44 @@ xdr_stream_decode_opaque_inline(struct xdr_stream *xdr, void **ptr, size_t maxle
 	}
 	return len;
 }
+
+/**
+ * xdr_stream_decode_uint32_array - Decode variable length array of integers
+ * @xdr: pointer to xdr_stream
+ * @array: location to store the integer array or NULL
+ * @array_size: number of elements to store
+ *
+ * Return values:
+ *   On success, returns number of elements stored in @array
+ *   %-EBADMSG on XDR buffer overflow
+ *   %-EMSGSIZE if the size of the array exceeds @array_size
+ */
+static inline ssize_t
+xdr_stream_decode_uint32_array(struct xdr_stream *xdr,
+		__u32 *array, size_t array_size)
+{
+	__be32 *p;
+	__u32 len;
+	ssize_t retval;
+
+	if (unlikely(xdr_stream_decode_u32(xdr, &len) < 0))
+		return -EBADMSG;
+	p = xdr_inline_decode(xdr, len * sizeof(*p));
+	if (unlikely(!p))
+		return -EBADMSG;
+	if (array == NULL)
+		return len;
+	if (len <= array_size) {
+		if (len < array_size)
+			memset(array+len, 0, (array_size-len)*sizeof(*array));
+		array_size = len;
+		retval = len;
+	} else
+		retval = -EMSGSIZE;
+	for (; array_size > 0; p++, array++, array_size--)
+		*array = be32_to_cpup(p);
+	return retval;
+}
 #endif /* __KERNEL__ */
 
 #endif /* _SUNRPC_XDR_H_ */
-- 
cgit v1.2.3-58-ga151