Merge tag 'nfsd-5.14' of git://linux-nfs.org/~bfields/linux

Pull nfsd updates from Bruce Fields:

 - add tracepoints for callbacks and for client creation and destruction

 - cache the mounts used for server-to-server copies

 - expose callback information in /proc/fs/nfsd/clients/*/info

 - don't hold locks unnecessarily while waiting for commits

 - update NLM to use xdr_stream, as we have for NFSv2/v3/v4

* tag 'nfsd-5.14' of git://linux-nfs.org/~bfields/linux: (69 commits)
  nfsd: fix NULL dereference in nfs3svc_encode_getaclres
  NFSD: Prevent a possible oops in the nfs_dirent() tracepoint
  nfsd: remove redundant assignment to pointer 'this'
  nfsd: Reduce contention for the nfsd_file nf_rwsem
  lockd: Update the NLMv4 SHARE results encoder to use struct xdr_stream
  lockd: Update the NLMv4 nlm_res results encoder to use struct xdr_stream
  lockd: Update the NLMv4 TEST results encoder to use struct xdr_stream
  lockd: Update the NLMv4 void results encoder to use struct xdr_stream
  lockd: Update the NLMv4 FREE_ALL arguments decoder to use struct xdr_stream
  lockd: Update the NLMv4 SHARE arguments decoder to use struct xdr_stream
  lockd: Update the NLMv4 SM_NOTIFY arguments decoder to use struct xdr_stream
  lockd: Update the NLMv4 nlm_res arguments decoder to use struct xdr_stream
  lockd: Update the NLMv4 UNLOCK arguments decoder to use struct xdr_stream
  lockd: Update the NLMv4 CANCEL arguments decoder to use struct xdr_stream
  lockd: Update the NLMv4 LOCK arguments decoder to use struct xdr_stream
  lockd: Update the NLMv4 TEST arguments decoder to use struct xdr_stream
  lockd: Update the NLMv4 void arguments decoder to use struct xdr_stream
  lockd: Update the NLMv1 SHARE results encoder to use struct xdr_stream
  lockd: Update the NLMv1 nlm_res results encoder to use struct xdr_stream
  lockd: Update the NLMv1 TEST results encoder to use struct xdr_stream
  ...

fs/lockd/svc.c

@@ -766,6 +766,46 @@ static void __exit exit_nlm(void)
 module_init(init_nlm);
 module_exit(exit_nlm);
 
+/**
+ * nlmsvc_dispatch - Process an NLM Request
+ * @rqstp: incoming request
+ * @statp: pointer to location of accept_stat field in RPC Reply buffer
+ *
+ * Return values:
+ *  %0: Processing complete; do not send a Reply
+ *  %1: Processing complete; send Reply in rqstp->rq_res
+ */
+static int nlmsvc_dispatch(struct svc_rqst *rqstp, __be32 *statp)
+{
+	const struct svc_procedure *procp = rqstp->rq_procinfo;
+	struct kvec *argv = rqstp->rq_arg.head;
+	struct kvec *resv = rqstp->rq_res.head;
+
+	svcxdr_init_decode(rqstp);
+	if (!procp->pc_decode(rqstp, argv->iov_base))
+		goto out_decode_err;
+
+	*statp = procp->pc_func(rqstp);
+	if (*statp == rpc_drop_reply)
+		return 0;
+	if (*statp != rpc_success)
+		return 1;
+
+	svcxdr_init_encode(rqstp);
+	if (!procp->pc_encode(rqstp, resv->iov_base + resv->iov_len))
+		goto out_encode_err;
+
+	return 1;
+
+out_decode_err:
+	*statp = rpc_garbage_args;
+	return 1;
+
+out_encode_err:
+	*statp = rpc_system_err;
+	return 1;
+}
+
 /*
  * Define NLM program and procedures
  */
@@ -775,6 +815,7 @@ static const struct svc_version	nlmsvc_version1 = {
 	.vs_nproc	= 17,
 	.vs_proc	= nlmsvc_procedures,
 	.vs_count	= nlmsvc_version1_count,
+	.vs_dispatch	= nlmsvc_dispatch,
 	.vs_xdrsize	= NLMSVC_XDRSIZE,
 };
 static unsigned int nlmsvc_version3_count[24];
@@ -783,6 +824,7 @@ static const struct svc_version	nlmsvc_version3 = {
 	.vs_nproc	= 24,
 	.vs_proc	= nlmsvc_procedures,
 	.vs_count	= nlmsvc_version3_count,
+	.vs_dispatch	= nlmsvc_dispatch,
 	.vs_xdrsize	= NLMSVC_XDRSIZE,
 };
 #ifdef CONFIG_LOCKD_V4
@@ -792,6 +834,7 @@ static const struct svc_version	nlmsvc_version4 = {
 	.vs_nproc	= 24,
 	.vs_proc	= nlmsvc_procedures4,
 	.vs_count	= nlmsvc_version4_count,
+	.vs_dispatch	= nlmsvc_dispatch,
 	.vs_xdrsize	= NLMSVC_XDRSIZE,
 };
 #endif

fs/lockd/svcxdr.h

+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Encode/decode NLM basic data types
+ *
+ * Basic NLMv3 XDR data types are not defined in an IETF standards
+ * document.  X/Open has a description of these data types that
+ * is useful.  See Chapter 10 of "Protocols for Interworking:
+ * XNFS, Version 3W".
+ *
+ * Basic NLMv4 XDR data types are defined in Appendix II.1.4 of
+ * RFC 1813: "NFS Version 3 Protocol Specification".
+ *
+ * Author: Chuck Lever <chuck.lever@oracle.com>
+ *
+ * Copyright (c) 2020, Oracle and/or its affiliates.
+ */
+
+#ifndef _LOCKD_SVCXDR_H_
+#define _LOCKD_SVCXDR_H_
+
+static inline bool
+svcxdr_decode_stats(struct xdr_stream *xdr, __be32 *status)
+{
+	__be32 *p;
+
+	p = xdr_inline_decode(xdr, XDR_UNIT);
+	if (!p)
+		return false;
+	*status = *p;
+
+	return true;
+}
+
+static inline bool
+svcxdr_encode_stats(struct xdr_stream *xdr, __be32 status)
+{
+	__be32 *p;
+
+	p = xdr_reserve_space(xdr, XDR_UNIT);
+	if (!p)
+		return false;
+	*p = status;
+
+	return true;
+}
+
+static inline bool
+svcxdr_decode_string(struct xdr_stream *xdr, char **data, unsigned int *data_len)
+{
+	__be32 *p;
+	u32 len;
+
+	if (xdr_stream_decode_u32(xdr, &len) < 0)
+		return false;
+	if (len > NLM_MAXSTRLEN)
+		return false;
+	p = xdr_inline_decode(xdr, len);
+	if (!p)
+		return false;
+	*data_len = len;
+	*data = (char *)p;
+
+	return true;
+}
+
+/*
+ * NLM cookies are defined by specification to be a variable-length
+ * XDR opaque no longer than 1024 bytes. However, this implementation
+ * limits their length to 32 bytes, and treats zero-length cookies
+ * specially.
+ */
+static inline bool
+svcxdr_decode_cookie(struct xdr_stream *xdr, struct nlm_cookie *cookie)
+{
+	__be32 *p;
+	u32 len;
+
+	if (xdr_stream_decode_u32(xdr, &len) < 0)
+		return false;
+	if (len > NLM_MAXCOOKIELEN)
+		return false;
+	if (!len)
+		goto out_hpux;
+
+	p = xdr_inline_decode(xdr, len);
+	if (!p)
+		return false;
+	cookie->len = len;
+	memcpy(cookie->data, p, len);
+
+	return true;
+
+	/* apparently HPUX can return empty cookies */
+out_hpux:
+	cookie->len = 4;
+	memset(cookie->data, 0, 4);
+	return true;
+}
+
+static inline bool
+svcxdr_encode_cookie(struct xdr_stream *xdr, const struct nlm_cookie *cookie)
+{
+	__be32 *p;
+
+	if (xdr_stream_encode_u32(xdr, cookie->len) < 0)
+		return false;
+	p = xdr_reserve_space(xdr, cookie->len);
+	if (!p)
+		return false;
+	memcpy(p, cookie->data, cookie->len);
+
+	return true;
+}
+
+static inline bool
+svcxdr_decode_owner(struct xdr_stream *xdr, struct xdr_netobj *obj)
+{
+	__be32 *p;
+	u32 len;
+
+	if (xdr_stream_decode_u32(xdr, &len) < 0)
+		return false;
+	if (len > XDR_MAX_NETOBJ)
+		return false;
+	p = xdr_inline_decode(xdr, len);
+	if (!p)
+		return false;
+	obj->len = len;
+	obj->data = (u8 *)p;
+
+	return true;
+}
+
+static inline bool
+svcxdr_encode_owner(struct xdr_stream *xdr, const struct xdr_netobj *obj)
+{
+	unsigned int quadlen = XDR_QUADLEN(obj->len);
+	__be32 *p;
+
+	if (xdr_stream_encode_u32(xdr, obj->len) < 0)
+		return false;
+	p = xdr_reserve_space(xdr, obj->len);
+	if (!p)
+		return false;
+	p[quadlen - 1] = 0;	/* XDR pad */
+	memcpy(p, obj->data, obj->len);
+
+	return true;
+}
+
+#endif /* _LOCKD_SVCXDR_H_ */

fs/nfs_common/grace.c

@@ -82,6 +82,7 @@ __state_in_grace(struct net *net, bool open)
 
 /**
  * locks_in_grace
+ * @net: network namespace
  *
  * Lock managers call this function to determine when it is OK for them
  * to answer ordinary lock requests, and when they should accept only

fs/nfsd/netns.h

@@ -176,6 +176,12 @@ struct nfsd_net {
 	unsigned int             longest_chain_cachesize;
 
 	struct shrinker		nfsd_reply_cache_shrinker;
+
+	/* tracking server-to-server copy mounts */
+	spinlock_t              nfsd_ssc_lock;
+	struct list_head        nfsd_ssc_mount_list;
+	wait_queue_head_t       nfsd_ssc_waitq;
+
 	/* utsname taken from the process that starts the server */
 	char			nfsd_name[UNX_MAXNODENAME+1];
 };

fs/nfsd/nfs3acl.c

@@ -172,7 +172,7 @@ static int nfs3svc_encode_getaclres(struct svc_rqst *rqstp, __be32 *p)
 	struct nfsd3_getaclres *resp = rqstp->rq_resp;
 	struct dentry *dentry = resp->fh.fh_dentry;
 	struct kvec *head = rqstp->rq_res.head;
-	struct inode *inode = d_inode(dentry);
+	struct inode *inode;
 	unsigned int base;
 	int n;
 	int w;
@@ -181,6 +181,7 @@ static int nfs3svc_encode_getaclres(struct svc_rqst *rqstp, __be32 *p)
 		return 0;
 	switch (resp->status) {
 	case nfs_ok:
+		inode = d_inode(dentry);
 		if (!svcxdr_encode_post_op_attr(rqstp, xdr, &resp->fh))
 			return 0;
 		if (xdr_stream_encode_u32(xdr, resp->mask) < 0)

fs/nfsd/nfs4callback.c

@@ -915,10 +915,8 @@ static int setup_callback_client(struct nfs4_client *clp, struct nfs4_cb_conn *c
 		args.authflavor = clp->cl_cred.cr_flavor;
 		clp->cl_cb_ident = conn->cb_ident;
 	} else {
-		if (!conn->cb_xprt) {
-			trace_nfsd_cb_setup_err(clp, -EINVAL);
+		if (!conn->cb_xprt)
 			return -EINVAL;
-		}
 		clp->cl_cb_conn.cb_xprt = conn->cb_xprt;
 		clp->cl_cb_session = ses;
 		args.bc_xprt = conn->cb_xprt;
@@ -941,37 +939,43 @@ static int setup_callback_client(struct nfs4_client *clp, struct nfs4_cb_conn *c
 	}
 	clp->cl_cb_client = client;
 	clp->cl_cb_cred = cred;
-	trace_nfsd_cb_setup(clp);
+	rcu_read_lock();
+	trace_nfsd_cb_setup(clp, rpc_peeraddr2str(client, RPC_DISPLAY_NETID),
+			    args.authflavor);
+	rcu_read_unlock();
 	return 0;
 }
 
+static void nfsd4_mark_cb_state(struct nfs4_client *clp, int newstate)
+{
+	if (clp->cl_cb_state != newstate) {
+		clp->cl_cb_state = newstate;
+		trace_nfsd_cb_state(clp);
+	}
+}
+
 static void nfsd4_mark_cb_down(struct nfs4_client *clp, int reason)
 {
 	if (test_bit(NFSD4_CLIENT_CB_UPDATE, &clp->cl_flags))
 		return;
-	clp->cl_cb_state = NFSD4_CB_DOWN;
-	trace_nfsd_cb_state(clp);
+	nfsd4_mark_cb_state(clp, NFSD4_CB_DOWN);
 }
 
 static void nfsd4_mark_cb_fault(struct nfs4_client *clp, int reason)
 {
 	if (test_bit(NFSD4_CLIENT_CB_UPDATE, &clp->cl_flags))
 		return;
-	clp->cl_cb_state = NFSD4_CB_FAULT;
-	trace_nfsd_cb_state(clp);
+	nfsd4_mark_cb_state(clp, NFSD4_CB_FAULT);
 }
 
 static void nfsd4_cb_probe_done(struct rpc_task *task, void *calldata)
 {
 	struct nfs4_client *clp = container_of(calldata, struct nfs4_client, cl_cb_null);
 
-	trace_nfsd_cb_done(clp, task->tk_status);
 	if (task->tk_status)
 		nfsd4_mark_cb_down(clp, task->tk_status);
-	else {
-		clp->cl_cb_state = NFSD4_CB_UP;
-		trace_nfsd_cb_state(clp);
-	}
+	else
+		nfsd4_mark_cb_state(clp, NFSD4_CB_UP);
 }
 
 static void nfsd4_cb_probe_release(void *calldata)
@@ -995,8 +999,8 @@ static const struct rpc_call_ops nfsd4_cb_probe_ops = {
  */
 void nfsd4_probe_callback(struct nfs4_client *clp)
 {
-	clp->cl_cb_state = NFSD4_CB_UNKNOWN;
-	trace_nfsd_cb_state(clp);
+	trace_nfsd_cb_probe(clp);
+	nfsd4_mark_cb_state(clp, NFSD4_CB_UNKNOWN);
 	set_bit(NFSD4_CLIENT_CB_UPDATE, &clp->cl_flags);
 	nfsd4_run_cb(&clp->cl_cb_null);
 }
@@ -1009,11 +1013,10 @@ void nfsd4_probe_callback_sync(struct nfs4_client *clp)
 
 void nfsd4_change_callback(struct nfs4_client *clp, struct nfs4_cb_conn *conn)
 {
-	clp->cl_cb_state = NFSD4_CB_UNKNOWN;
+	nfsd4_mark_cb_state(clp, NFSD4_CB_UNKNOWN);
 	spin_lock(&clp->cl_lock);
 	memcpy(&clp->cl_cb_conn, conn, sizeof(struct nfs4_cb_conn));
 	spin_unlock(&clp->cl_lock);
-	trace_nfsd_cb_state(clp);
 }
 
 /*
@@ -1170,8 +1173,6 @@ static void nfsd4_cb_done(struct rpc_task *task, void *calldata)
 	struct nfsd4_callback *cb = calldata;
 	struct nfs4_client *clp = cb->cb_clp;
 
-	trace_nfsd_cb_done(clp, task->tk_status);
-
 	if (!nfsd4_cb_sequence_done(task, cb))
 		return;
 
@@ -1231,6 +1232,9 @@ void nfsd4_destroy_callback_queue(void)
 /* must be called under the state lock */
 void nfsd4_shutdown_callback(struct nfs4_client *clp)
 {
+	if (clp->cl_cb_state != NFSD4_CB_UNKNOWN)
+		trace_nfsd_cb_shutdown(clp);
+
 	set_bit(NFSD4_CLIENT_CB_KILL, &clp->cl_flags);
 	/*
 	 * Note this won't actually result in a null callback;
@@ -1276,7 +1280,6 @@ static void nfsd4_process_cb_update(struct nfsd4_callback *cb)
 	 * kill the old client:
 	 */
 	if (clp->cl_cb_client) {
-		trace_nfsd_cb_shutdown(clp);
 		rpc_shutdown_client(clp->cl_cb_client);
 		clp->cl_cb_client = NULL;
 		put_cred(clp->cl_cb_cred);
@@ -1322,8 +1325,6 @@ nfsd4_run_cb_work(struct work_struct *work)
 	struct rpc_clnt *clnt;
 	int flags;
 
-	trace_nfsd_cb_work(clp, cb->cb_msg.rpc_proc->p_name);
-
 	if (cb->cb_need_restart) {
 		cb->cb_need_restart = false;
 	} else {
@@ -1345,7 +1346,7 @@ nfsd4_run_cb_work(struct work_struct *work)
 	 * Don't send probe messages for 4.1 or later.
 	 */
 	if (!cb->cb_ops && clp->cl_minorversion) {
-		clp->cl_cb_state = NFSD4_CB_UP;
+		nfsd4_mark_cb_state(clp, NFSD4_CB_UP);
 		nfsd41_destroy_cb(cb);
 		return;
 	}

fs/nfsd/nfs4proc.c

@@ -55,6 +55,13 @@ module_param(inter_copy_offload_enable, bool, 0644);
 MODULE_PARM_DESC(inter_copy_offload_enable,
 		 "Enable inter server to server copy offload. Default: false");
 
+#ifdef CONFIG_NFSD_V4_2_INTER_SSC
+static int nfsd4_ssc_umount_timeout = 900000;		/* default to 15 mins */
+module_param(nfsd4_ssc_umount_timeout, int, 0644);
+MODULE_PARM_DESC(nfsd4_ssc_umount_timeout,
+		"idle msecs before unmount export from source server");
+#endif
+
 #ifdef CONFIG_NFSD_V4_SECURITY_LABEL
 #include <linux/security.h>
 
@@ -1165,6 +1172,81 @@ extern void nfs_sb_deactive(struct super_block *sb);
 
 #define NFSD42_INTERSSC_MOUNTOPS "vers=4.2,addr=%s,sec=sys"
 
+/*
+ * setup a work entry in the ssc delayed unmount list.
+ */
+static __be32 nfsd4_ssc_setup_dul(struct nfsd_net *nn, char *ipaddr,
+		struct nfsd4_ssc_umount_item **retwork, struct vfsmount **ss_mnt)
+{
+	struct nfsd4_ssc_umount_item *ni = 0;
+	struct nfsd4_ssc_umount_item *work = NULL;
+	struct nfsd4_ssc_umount_item *tmp;
+	DEFINE_WAIT(wait);
+
+	*ss_mnt = NULL;
+	*retwork = NULL;
+	work = kzalloc(sizeof(*work), GFP_KERNEL);
+try_again:
+	spin_lock(&nn->nfsd_ssc_lock);
+	list_for_each_entry_safe(ni, tmp, &nn->nfsd_ssc_mount_list, nsui_list) {
+		if (strncmp(ni->nsui_ipaddr, ipaddr, sizeof(ni->nsui_ipaddr)))
+			continue;
+		/* found a match */
+		if (ni->nsui_busy) {
+			/*  wait - and try again */
+			prepare_to_wait(&nn->nfsd_ssc_waitq, &wait,
+				TASK_INTERRUPTIBLE);
+			spin_unlock(&nn->nfsd_ssc_lock);
+
+			/* allow 20secs for mount/unmount for now - revisit */
+			if (signal_pending(current) ||
+					(schedule_timeout(20*HZ) == 0)) {
+				kfree(work);
+				return nfserr_eagain;
+			}
+			finish_wait(&nn->nfsd_ssc_waitq, &wait);
+			goto try_again;
+		}
+		*ss_mnt = ni->nsui_vfsmount;
+		refcount_inc(&ni->nsui_refcnt);
+		spin_unlock(&nn->nfsd_ssc_lock);
+		kfree(work);
+
+		/* return vfsmount in ss_mnt */
+		return 0;
+	}
+	if (work) {
+		strncpy(work->nsui_ipaddr, ipaddr, sizeof(work->nsui_ipaddr));
+		refcount_set(&work->nsui_refcnt, 2);
+		work->nsui_busy = true;
+		list_add_tail(&work->nsui_list, &nn->nfsd_ssc_mount_list);
+		*retwork = work;
+	}
+	spin_unlock(&nn->nfsd_ssc_lock);
+	return 0;
+}
+
+static void nfsd4_ssc_update_dul_work(struct nfsd_net *nn,
+		struct nfsd4_ssc_umount_item *work, struct vfsmount *ss_mnt)
+{
+	/* set nsui_vfsmount, clear busy flag and wakeup waiters */
+	spin_lock(&nn->nfsd_ssc_lock);
+	work->nsui_vfsmount = ss_mnt;
+	work->nsui_busy = false;
+	wake_up_all(&nn->nfsd_ssc_waitq);
+	spin_unlock(&nn->nfsd_ssc_lock);
+}
+
+static void nfsd4_ssc_cancel_dul_work(struct nfsd_net *nn,
+		struct nfsd4_ssc_umount_item *work)
+{
+	spin_lock(&nn->nfsd_ssc_lock);
+	list_del(&work->nsui_list);
+	wake_up_all(&nn->nfsd_ssc_waitq);
+	spin_unlock(&nn->nfsd_ssc_lock);
+	kfree(work);
+}
+
 /*
  * Support one copy source server for now.
  */
@@ -1181,6 +1263,8 @@ nfsd4_interssc_connect(struct nl4_server *nss, struct svc_rqst *rqstp,
 	char *ipaddr, *dev_name, *raw_data;
 	int len, raw_len;
 	__be32 status = nfserr_inval;
+	struct nfsd4_ssc_umount_item *work = NULL;
+	struct nfsd_net *nn = net_generic(SVC_NET(rqstp), nfsd_net_id);
 
 	naddr = &nss->u.nl4_addr;
 	tmp_addrlen = rpc_uaddr2sockaddr(SVC_NET(rqstp), naddr->addr,
@@ -1229,12 +1313,24 @@ nfsd4_interssc_connect(struct nl4_server *nss, struct svc_rqst *rqstp,
 		goto out_free_rawdata;
 	snprintf(dev_name, len + 5, "%s%s%s:/", startsep, ipaddr, endsep);
 
+	status = nfsd4_ssc_setup_dul(nn, ipaddr, &work, &ss_mnt);
+	if (status)
+		goto out_free_devname;
+	if (ss_mnt)
+		goto out_done;
+
 	/* Use an 'internal' mount: SB_KERNMOUNT -> MNT_INTERNAL */
 	ss_mnt = vfs_kern_mount(type, SB_KERNMOUNT, dev_name, raw_data);
 	module_put(type->owner);
-	if (IS_ERR(ss_mnt))
+	if (IS_ERR(ss_mnt)) {
+		status = nfserr_nodev;
+		if (work)
+			nfsd4_ssc_cancel_dul_work(nn, work);
 		goto out_free_devname;
-
+	}
+	if (work)
+		nfsd4_ssc_update_dul_work(nn, work, ss_mnt);
+out_done:
 	status = 0;
 	*mount = ss_mnt;
 
@@ -1301,10 +1397,42 @@ static void
 nfsd4_cleanup_inter_ssc(struct vfsmount *ss_mnt, struct nfsd_file *src,
 			struct nfsd_file *dst)
 {
+	bool found = false;
+	long timeout;
+	struct nfsd4_ssc_umount_item *tmp;
+	struct nfsd4_ssc_umount_item *ni = NULL;
+	struct nfsd_net *nn = net_generic(dst->nf_net, nfsd_net_id);
+
 	nfs42_ssc_close(src->nf_file);
-	fput(src->nf_file);
 	nfsd_file_put(dst);
+	fput(src->nf_file);
+
+	if (!nn) {
 		mntput(ss_mnt);
+		return;
+	}
+	spin_lock(&nn->nfsd_ssc_lock);
+	timeout = msecs_to_jiffies(nfsd4_ssc_umount_timeout);
+	list_for_each_entry_safe(ni, tmp, &nn->nfsd_ssc_mount_list, nsui_list) {
+		if (ni->nsui_vfsmount->mnt_sb == ss_mnt->mnt_sb) {
+			list_del(&ni->nsui_list);
+			/*
+			 * vfsmount can be shared by multiple exports,
+			 * decrement refcnt. If the count drops to 1 it
+			 * will be unmounted when nsui_expire expires.
+			 */
+			refcount_dec(&ni->nsui_refcnt);
+			ni->nsui_expire = jiffies + timeout;
+			list_add_tail(&ni->nsui_list, &nn->nfsd_ssc_mount_list);
+			found = true;
+			break;
+		}
+	}
+	spin_unlock(&nn->nfsd_ssc_lock);
+	if (!found) {
+		mntput(ss_mnt);
+		return;
+	}
 }
 
 #else /* CONFIG_NFSD_V4_2_INTER_SSC */
@@ -1375,7 +1503,8 @@ static const struct nfsd4_callback_ops nfsd4_cb_offload_ops = {
 
 static void nfsd4_init_copy_res(struct nfsd4_copy *copy, bool sync)
 {
-	copy->cp_res.wr_stable_how = NFS_UNSTABLE;
+	copy->cp_res.wr_stable_how =
+		copy->committed ? NFS_FILE_SYNC : NFS_UNSTABLE;
 	copy->cp_synchronous = sync;
 	gen_boot_verifier(&copy->cp_res.wr_verifier, copy->cp_clp->net);
 }
@@ -1386,6 +1515,7 @@ static ssize_t _nfsd_copy_file_range(struct nfsd4_copy *copy)
 	u64 bytes_total = copy->cp_count;
 	u64 src_pos = copy->cp_src_pos;
 	u64 dst_pos = copy->cp_dst_pos;
+	__be32 status;
 
 	/* See RFC 7862 p.67: */
 	if (bytes_total == 0)
@@ -1403,6 +1533,16 @@ static ssize_t _nfsd_copy_file_range(struct nfsd4_copy *copy)
 		src_pos += bytes_copied;
 		dst_pos += bytes_copied;
 	} while (bytes_total > 0 && !copy->cp_synchronous);
+	/* for a non-zero asynchronous copy do a commit of data */
+	if (!copy->cp_synchronous && copy->cp_res.wr_bytes_written > 0) {
+		down_write(&copy->nf_dst->nf_rwsem);
+		status = vfs_fsync_range(copy->nf_dst->nf_file,
+					 copy->cp_dst_pos,
+					 copy->cp_res.wr_bytes_written, 0);
+		up_write(&copy->nf_dst->nf_rwsem);
+		if (!status)
+			copy->committed = true;
+	}
 	return bytes_copied;
 }
 
@@ -1497,6 +1637,8 @@ static int nfsd4_do_async_copy(void *data)
 	memcpy(&cb_copy->fh, &copy->fh, sizeof(copy->fh));
 	nfsd4_init_cb(&cb_copy->cp_cb, cb_copy->cp_clp,
 			&nfsd4_cb_offload_ops, NFSPROC4_CLNT_CB_OFFLOAD);
+	trace_nfsd_cb_offload(copy->cp_clp, &copy->cp_res.cb_stateid,
+			      &copy->fh, copy->cp_count, copy->nfserr);
 	nfsd4_run_cb(&cb_copy->cp_cb);
 out:
 	if (!copy->cp_intra)
@@ -3232,7 +3374,7 @@ bool nfsd4_spo_must_allow(struct svc_rqst *rqstp)
 {
 	struct nfsd4_compoundres *resp = rqstp->rq_resp;
 	struct nfsd4_compoundargs *argp = rqstp->rq_argp;
-	struct nfsd4_op *this = &argp->ops[resp->opcnt - 1];
+	struct nfsd4_op *this;
 	struct nfsd4_compound_state *cstate = &resp->cstate;
 	struct nfs4_op_map *allow = &cstate->clp->cl_spo_must_allow;
 	u32 opiter;

fs/nfsd/nfsd.h

@@ -484,6 +484,10 @@ static inline bool nfsd_attrs_supported(u32 minorversion, const u32 *bmval)
 extern int nfsd4_is_junction(struct dentry *dentry);
 extern int register_cld_notifier(void);
 extern void unregister_cld_notifier(void);
+#ifdef CONFIG_NFSD_V4_2_INTER_SSC
+extern void nfsd4_ssc_init_umount_work(struct nfsd_net *nn);
+#endif
+
 #else /* CONFIG_NFSD_V4 */
 static inline int nfsd4_is_junction(struct dentry *dentry)
 {

fs/nfsd/nfsfh.h

@@ -225,15 +225,12 @@ static inline bool fh_fsid_match(struct knfsd_fh *fh1, struct knfsd_fh *fh2)
  * returns a crc32 hash for the filehandle that is compatible with
  * the one displayed by "wireshark".
  */
-
-static inline u32
-knfsd_fh_hash(struct knfsd_fh *fh)
+static inline u32 knfsd_fh_hash(const struct knfsd_fh *fh)
 {
 	return ~crc32_le(0xFFFFFFFF, (unsigned char *)&fh->fh_base, fh->fh_size);
 }
 #else
-static inline u32
-knfsd_fh_hash(struct knfsd_fh *fh)
+static inline u32 knfsd_fh_hash(const struct knfsd_fh *fh)
 {
 	return 0;
 }

fs/nfsd/nfssvc.c

@@ -403,6 +403,9 @@ static int nfsd_startup_net(struct net *net, const struct cred *cred)
 	if (ret)
 		goto out_filecache;
 
+#ifdef CONFIG_NFSD_V4_2_INTER_SSC
+	nfsd4_ssc_init_umount_work(nn);
+#endif
 	nn->nfsd_net_up = true;
 	return 0;
 

fs/nfsd/vfs.c

@@ -1123,6 +1123,19 @@ nfsd_write(struct svc_rqst *rqstp, struct svc_fh *fhp, loff_t offset,
 }
 
 #ifdef CONFIG_NFSD_V3
+static int
+nfsd_filemap_write_and_wait_range(struct nfsd_file *nf, loff_t offset,
+				  loff_t end)
+{
+	struct address_space *mapping = nf->nf_file->f_mapping;
+	int ret = filemap_fdatawrite_range(mapping, offset, end);
+
+	if (ret)
+		return ret;
+	filemap_fdatawait_range_keep_errors(mapping, offset, end);
+	return 0;
+}
+
 /*
  * Commit all pending writes to stable storage.
  *
@@ -1153,9 +1166,10 @@ nfsd_commit(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	if (err)
 		goto out;
 	if (EX_ISSYNC(fhp->fh_export)) {
-		int err2;
+		int err2 = nfsd_filemap_write_and_wait_range(nf, offset, end);
 
 		down_write(&nf->nf_rwsem);
+		if (!err2)
 			err2 = vfs_fsync_range(nf->nf_file, offset, end, 0);
 		switch (err2) {
 		case 0:
@@ -1613,9 +1627,9 @@ nfsd_symlink(struct svc_rqst *rqstp, struct svc_fh *fhp,
 
 	host_err = vfs_symlink(&init_user_ns, d_inode(dentry), dnew, path);
 	err = nfserrno(host_err);
+	fh_unlock(fhp);
 	if (!err)
 		err = nfserrno(commit_metadata(fhp));
-	fh_unlock(fhp);
 
 	fh_drop_write(fhp);
 
@@ -1680,6 +1694,7 @@ nfsd_link(struct svc_rqst *rqstp, struct svc_fh *ffhp,
 	if (d_really_is_negative(dold))
 		goto out_dput;
 	host_err = vfs_link(dold, &init_user_ns, dirp, dnew, NULL);
+	fh_unlock(ffhp);
 	if (!host_err) {
 		err = nfserrno(commit_metadata(ffhp));
 		if (!err)
@@ -1859,6 +1874,7 @@ nfsd_unlink(struct svc_rqst *rqstp, struct svc_fh *fhp, int type,
 {
 	struct dentry	*dentry, *rdentry;
 	struct inode	*dirp;
+	struct inode	*rinode;
 	__be32		err;
 	int		host_err;
 
@@ -1887,6 +1903,8 @@ nfsd_unlink(struct svc_rqst *rqstp, struct svc_fh *fhp, int type,
 		host_err = -ENOENT;
 		goto out_drop_write;
 	}
+	rinode = d_inode(rdentry);
+	ihold(rinode);
 
 	if (!type)
 		type = d_inode(rdentry)->i_mode & S_IFMT;
@@ -1899,9 +1917,11 @@ nfsd_unlink(struct svc_rqst *rqstp, struct svc_fh *fhp, int type,
 		host_err = vfs_rmdir(&init_user_ns, dirp, rdentry);
 	}
 
+	fh_unlock(fhp);
 	if (!host_err)
 		host_err = commit_metadata(fhp);
 	dput(rdentry);
+	iput(rinode);    /* truncate the inode here */
 
 out_drop_write:
 	fh_drop_write(fhp);

fs/nfsd/xdr4.h

@@ -567,6 +567,7 @@ struct nfsd4_copy {
 	struct vfsmount		*ss_mnt;
 	struct nfs_fh		c_fh;
 	nfs4_stateid		stateid;
+	bool			committed;
 };
 
 struct nfsd4_seek {

include/linux/lockd/xdr.h

@@ -109,11 +109,5 @@ int	nlmsvc_decode_shareargs(struct svc_rqst *, __be32 *);
 int	nlmsvc_encode_shareres(struct svc_rqst *, __be32 *);
 int	nlmsvc_decode_notify(struct svc_rqst *, __be32 *);
 int	nlmsvc_decode_reboot(struct svc_rqst *, __be32 *);
-/*
-int	nlmclt_encode_testargs(struct rpc_rqst *, u32 *, struct nlm_args *);
-int	nlmclt_encode_lockargs(struct rpc_rqst *, u32 *, struct nlm_args *);
-int	nlmclt_encode_cancargs(struct rpc_rqst *, u32 *, struct nlm_args *);
-int	nlmclt_encode_unlockargs(struct rpc_rqst *, u32 *, struct nlm_args *);
- */
 
 #endif /* LOCKD_XDR_H */

include/linux/lockd/xdr4.h

@@ -37,12 +37,7 @@ int	nlm4svc_decode_shareargs(struct svc_rqst *, __be32 *);
 int	nlm4svc_encode_shareres(struct svc_rqst *, __be32 *);
 int	nlm4svc_decode_notify(struct svc_rqst *, __be32 *);
 int	nlm4svc_decode_reboot(struct svc_rqst *, __be32 *);
-/*
-int	nlmclt_encode_testargs(struct rpc_rqst *, u32 *, struct nlm_args *);
-int	nlmclt_encode_lockargs(struct rpc_rqst *, u32 *, struct nlm_args *);
-int	nlmclt_encode_cancargs(struct rpc_rqst *, u32 *, struct nlm_args *);
-int	nlmclt_encode_unlockargs(struct rpc_rqst *, u32 *, struct nlm_args *);
- */
+
 extern const struct rpc_version nlm_version4;
 
 #endif /* LOCKD_XDR4_H */

include/linux/nfs_ssc.h

@@ -8,6 +8,7 @@
  */
 
 #include <linux/nfs_fs.h>
+#include <linux/sunrpc/svc.h>
 
 extern struct nfs_ssc_client_ops_tbl nfs_ssc_client_tbl;
 
@@ -52,6 +53,19 @@ static inline void nfs42_ssc_close(struct file *filep)
 	if (nfs_ssc_client_tbl.ssc_nfs4_ops)
 		(*nfs_ssc_client_tbl.ssc_nfs4_ops->sco_close)(filep);
 }
+
+struct nfsd4_ssc_umount_item {
+	struct list_head nsui_list;
+	bool nsui_busy;
+	/*
+	 * nsui_refcnt inited to 2, 1 on list and 1 for consumer. Entry
+	 * is removed when refcnt drops to 1 and nsui_expire expires.
+	 */
+	refcount_t nsui_refcnt;
+	unsigned long nsui_expire;
+	struct vfsmount *nsui_vfsmount;
+	char nsui_ipaddr[RPC_MAX_ADDRBUFLEN];
+};
 #endif
 
 /*

net/sunrpc/auth_gss/svcauth_gss.c

@@ -1275,7 +1275,7 @@ static int gss_proxy_save_rsc(struct cache_detail *cd,
 	long long ctxh;
 	struct gss_api_mech *gm = NULL;
 	time64_t expiry;
-	int status = -EINVAL;
+	int status;
 
 	memset(&rsci, 0, sizeof(rsci));
 	/* context handle */

net/sunrpc/xprtrdma/svc_rdma_rw.c

@@ -483,7 +483,7 @@ svc_rdma_build_writes(struct svc_rdma_write_info *info,
  * @iov: kvec to write
  *
  * Returns:
- *   On succes, returns zero
+ *   On success, returns zero
  *   %-E2BIG if the client-provided Write chunk is too small
  *   %-ENOMEM if a resource has been exhausted
  *   %-EIO if an rdma-rw error occurred
@@ -504,7 +504,7 @@ static int svc_rdma_iov_write(struct svc_rdma_write_info *info,
  * @length: number of bytes to write
  *
  * Returns:
- *   On succes, returns zero
+ *   On success, returns zero
  *   %-E2BIG if the client-provided Write chunk is too small
  *   %-ENOMEM if a resource has been exhausted
  *   %-EIO if an rdma-rw error occurred
@@ -526,7 +526,7 @@ static int svc_rdma_pages_write(struct svc_rdma_write_info *info,
  * @data: pointer to write arguments
  *
  * Returns:
- *   On succes, returns zero
+ *   On success, returns zero
  *   %-E2BIG if the client-provided Write chunk is too small
  *   %-ENOMEM if a resource has been exhausted
  *   %-EIO if an rdma-rw error occurred