Merge tag 'nfs-for-6.1-1' of git://git.linux-nfs.org/projects/anna/linux-nfs

Pull NFS client updates from Anna Schumaker:
 "New Features:
   - Add NFSv4.2 xattr tracepoints
   - Replace xprtiod WQ in rpcrdma
   - Flexfiles cancels I/O on layout recall or revoke

  Bugfixes and Cleanups:
   - Directly use ida_alloc() / ida_free()
   - Don't open-code max_t()
   - Prefer using strscpy over strlcpy
   - Remove unused forward declarations
   - Always return layout states on flexfiles layout return
   - Have LISTXATTR treat NFS4ERR_NOXATTR as an empty reply instead of
     error
   - Allow more xprtrdma memory allocations to fail without triggering a
     reclaim
   - Various other xprtrdma clean ups
   - Fix rpc_killall_tasks() races"

* tag 'nfs-for-6.1-1' of git://git.linux-nfs.org/projects/anna/linux-nfs: (27 commits)
  NFSv4/flexfiles: Cancel I/O if the layout is recalled or revoked
  SUNRPC: Add API to force the client to disconnect
  SUNRPC: Add a helper to allow pNFS drivers to selectively cancel RPC calls
  SUNRPC: Fix races with rpc_killall_tasks()
  xprtrdma: Fix uninitialized variable
  xprtrdma: Prevent memory allocations from driving a reclaim
  xprtrdma: Memory allocation should be allowed to fail during connect
  xprtrdma: MR-related memory allocation should be allowed to fail
  xprtrdma: Clean up synopsis of rpcrdma_regbuf_alloc()
  xprtrdma: Clean up synopsis of rpcrdma_req_create()
  svcrdma: Clean up RPCRDMA_DEF_GFP
  SUNRPC: Replace the use of the xprtiod WQ in rpcrdma
  NFSv4.2: Add a tracepoint for listxattr
  NFSv4.2: Add tracepoints for getxattr, setxattr, and removexattr
  NFSv4.2: Move TRACE_DEFINE_ENUM(NFS4_CONTENT_*) under CONFIG_NFS_V4_2
  NFSv4.2: Add special handling for LISTXATTR receiving NFS4ERR_NOXATTR
  nfs: remove nfs_wait_atomic_killable() and nfs_write_prepare() declaration
  NFSv4: remove nfs4_renewd_prepare_shutdown() declaration
  fs/nfs/pnfs_nfs.c: fix spelling typo and syntax error in comment
  NFSv4/pNFS: Always return layout stats on layout return for flexfiles
  ...

30 files changed, 342 insertions, 116 deletions

diff --git a/fs/nfs/file.c b/fs/nfs/file.c
index e0b205b63986..d8ec889a4b3f 100644
--- a/fs/nfs/file.c
+++ b/fs/nfs/file.c
@@ -656,9 +656,9 @@ ssize_t nfs_file_write(struct kiocb *iocb, struct iov_iter *from)
 			goto out;
 	}
 	if (mntflags & NFS_MOUNT_WRITE_WAIT) {
-		result = filemap_fdatawait_range(file->f_mapping,
-						 iocb->ki_pos - written,
-						 iocb->ki_pos - 1);
+		filemap_fdatawait_range(file->f_mapping,
+					iocb->ki_pos - written,
+					iocb->ki_pos - 1);
 	}
 	result = generic_write_sync(iocb, written);
 	if (result < 0)
diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c
index 7d285561e59f..1ec79ccf89ad 100644
--- a/fs/nfs/flexfilelayout/flexfilelayout.c
+++ b/fs/nfs/flexfilelayout/flexfilelayout.c
@@ -30,14 +30,20 @@
 #define FF_LAYOUT_POLL_RETRY_MAX     (15*HZ)
 #define FF_LAYOUTRETURN_MAXERR 20
 
+enum nfs4_ff_op_type {
+	NFS4_FF_OP_LAYOUTSTATS,
+	NFS4_FF_OP_LAYOUTRETURN,
+};
+
 static unsigned short io_maxretrans;
 
 static const struct pnfs_commit_ops ff_layout_commit_ops;
 static void ff_layout_read_record_layoutstats_done(struct rpc_task *task,
 		struct nfs_pgio_header *hdr);
-static int ff_layout_mirror_prepare_stats(struct pnfs_layout_hdr *lo,
+static int
+ff_layout_mirror_prepare_stats(struct pnfs_layout_hdr *lo,
 			       struct nfs42_layoutstat_devinfo *devinfo,
-			       int dev_limit);
+			       int dev_limit, enum nfs4_ff_op_type type);
 static void ff_layout_encode_ff_layoutupdate(struct xdr_stream *xdr,
 			      const struct nfs42_layoutstat_devinfo *devinfo,
 			      struct nfs4_ff_layout_mirror *mirror);
@@ -1373,6 +1379,11 @@ static int ff_layout_read_prepare_common(struct rpc_task *task,
 		return -EIO;
 	}
 
+	if (!pnfs_is_valid_lseg(hdr->lseg)) {
+		rpc_exit(task, -EAGAIN);
+		return -EAGAIN;
+	}
+
 	ff_layout_read_record_layoutstats_start(task, hdr);
 	return 0;
 }
@@ -1553,6 +1564,11 @@ static int ff_layout_write_prepare_common(struct rpc_task *task,
 		return -EIO;
 	}
 
+	if (!pnfs_is_valid_lseg(hdr->lseg)) {
+		rpc_exit(task, -EAGAIN);
+		return -EAGAIN;
+	}
+
 	ff_layout_write_record_layoutstats_start(task, hdr);
 	return 0;
 }
@@ -1645,15 +1661,23 @@ static void ff_layout_commit_record_layoutstats_done(struct rpc_task *task,
 	set_bit(NFS_LSEG_LAYOUTRETURN, &cdata->lseg->pls_flags);
 }
 
-static void ff_layout_commit_prepare_common(struct rpc_task *task,
-		struct nfs_commit_data *cdata)
+static int ff_layout_commit_prepare_common(struct rpc_task *task,
+					   struct nfs_commit_data *cdata)
 {
+	if (!pnfs_is_valid_lseg(cdata->lseg)) {
+		rpc_exit(task, -EAGAIN);
+		return -EAGAIN;
+	}
+
 	ff_layout_commit_record_layoutstats_start(task, cdata);
+	return 0;
 }
 
 static void ff_layout_commit_prepare_v3(struct rpc_task *task, void *data)
 {
-	ff_layout_commit_prepare_common(task, data);
+	if (ff_layout_commit_prepare_common(task, data))
+		return;
+
 	rpc_call_start(task);
 }
 
@@ -1949,6 +1973,65 @@ ff_layout_commit_pagelist(struct inode *inode, struct list_head *mds_pages,
 					    ff_layout_initiate_commit);
 }
 
+static bool ff_layout_match_rw(const struct rpc_task *task,
+			       const struct nfs_pgio_header *hdr,
+			       const struct pnfs_layout_segment *lseg)
+{
+	return hdr->lseg == lseg;
+}
+
+static bool ff_layout_match_commit(const struct rpc_task *task,
+				   const struct nfs_commit_data *cdata,
+				   const struct pnfs_layout_segment *lseg)
+{
+	return cdata->lseg == lseg;
+}
+
+static bool ff_layout_match_io(const struct rpc_task *task, const void *data)
+{
+	const struct rpc_call_ops *ops = task->tk_ops;
+
+	if (ops == &ff_layout_read_call_ops_v3 ||
+	    ops == &ff_layout_read_call_ops_v4 ||
+	    ops == &ff_layout_write_call_ops_v3 ||
+	    ops == &ff_layout_write_call_ops_v4)
+		return ff_layout_match_rw(task, task->tk_calldata, data);
+	if (ops == &ff_layout_commit_call_ops_v3 ||
+	    ops == &ff_layout_commit_call_ops_v4)
+		return ff_layout_match_commit(task, task->tk_calldata, data);
+	return false;
+}
+
+static void ff_layout_cancel_io(struct pnfs_layout_segment *lseg)
+{
+	struct nfs4_ff_layout_segment *flseg = FF_LAYOUT_LSEG(lseg);
+	struct nfs4_ff_layout_mirror *mirror;
+	struct nfs4_ff_layout_ds *mirror_ds;
+	struct nfs4_pnfs_ds *ds;
+	struct nfs_client *ds_clp;
+	struct rpc_clnt *clnt;
+	u32 idx;
+
+	for (idx = 0; idx < flseg->mirror_array_cnt; idx++) {
+		mirror = flseg->mirror_array[idx];
+		mirror_ds = mirror->mirror_ds;
+		if (!mirror_ds)
+			continue;
+		ds = mirror->mirror_ds->ds;
+		if (!ds)
+			continue;
+		ds_clp = ds->ds_clp;
+		if (!ds_clp)
+			continue;
+		clnt = ds_clp->cl_rpcclient;
+		if (!clnt)
+			continue;
+		if (!rpc_cancel_tasks(clnt, -EAGAIN, ff_layout_match_io, lseg))
+			continue;
+		rpc_clnt_disconnect(clnt);
+	}
+}
+
 static struct pnfs_ds_commit_info *
 ff_layout_get_ds_info(struct inode *inode)
 {
@@ -2161,8 +2244,9 @@ ff_layout_prepare_layoutreturn(struct nfs4_layoutreturn_args *args)
 			FF_LAYOUTRETURN_MAXERR);
 
 	spin_lock(&args->inode->i_lock);
-	ff_args->num_dev = ff_layout_mirror_prepare_stats(&ff_layout->generic_hdr,
-			&ff_args->devinfo[0], ARRAY_SIZE(ff_args->devinfo));
+	ff_args->num_dev = ff_layout_mirror_prepare_stats(
+		&ff_layout->generic_hdr, &ff_args->devinfo[0],
+		ARRAY_SIZE(ff_args->devinfo), NFS4_FF_OP_LAYOUTRETURN);
 	spin_unlock(&args->inode->i_lock);
 
 	args->ld_private->ops = &layoutreturn_ops;
@@ -2396,7 +2480,7 @@ static const struct nfs4_xdr_opaque_ops layoutstat_ops = {
 static int
 ff_layout_mirror_prepare_stats(struct pnfs_layout_hdr *lo,
 			       struct nfs42_layoutstat_devinfo *devinfo,
-			       int dev_limit)
+			       int dev_limit, enum nfs4_ff_op_type type)
 {
 	struct nfs4_flexfile_layout *ff_layout = FF_LAYOUT_FROM_HDR(lo);
 	struct nfs4_ff_layout_mirror *mirror;
@@ -2408,7 +2492,9 @@ ff_layout_mirror_prepare_stats(struct pnfs_layout_hdr *lo,
 			break;
 		if (IS_ERR_OR_NULL(mirror->mirror_ds))
 			continue;
-		if (!test_and_clear_bit(NFS4_FF_MIRROR_STAT_AVAIL, &mirror->flags))
+		if (!test_and_clear_bit(NFS4_FF_MIRROR_STAT_AVAIL,
+					&mirror->flags) &&
+		    type != NFS4_FF_OP_LAYOUTRETURN)
 			continue;
 		/* mirror refcount put in cleanup_layoutstats */
 		if (!refcount_inc_not_zero(&mirror->ref))
@@ -2448,7 +2534,9 @@ ff_layout_prepare_layoutstats(struct nfs42_layoutstat_args *args)
 	spin_lock(&args->inode->i_lock);
 	ff_layout = FF_LAYOUT_FROM_HDR(NFS_I(args->inode)->layout);
 	args->num_dev = ff_layout_mirror_prepare_stats(&ff_layout->generic_hdr,
-			&args->devinfo[0], dev_count);
+						       &args->devinfo[0],
+						       dev_count,
+						       NFS4_FF_OP_LAYOUTSTATS);
 	spin_unlock(&args->inode->i_lock);
 	if (!args->num_dev) {
 		kfree(args->devinfo);
@@ -2501,6 +2589,7 @@ static struct pnfs_layoutdriver_type flexfilelayout_type = {
 	.prepare_layoutreturn   = ff_layout_prepare_layoutreturn,
 	.sync			= pnfs_nfs_generic_sync,
 	.prepare_layoutstats	= ff_layout_prepare_layoutstats,
+	.cancel_io		= ff_layout_cancel_io,
 };
 
 static int __init nfs4flexfilelayout_init(void)
diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
index 8f3773dc38dd..6b2cfa59a1a2 100644
--- a/fs/nfs/inode.c
+++ b/fs/nfs/inode.c
@@ -313,7 +313,7 @@ struct nfs_find_desc {
 static int
 nfs_find_actor(struct inode *inode, void *opaque)
 {
-	struct nfs_find_desc	*desc = (struct nfs_find_desc *)opaque;
+	struct nfs_find_desc	*desc = opaque;
 	struct nfs_fh		*fh = desc->fh;
 	struct nfs_fattr	*fattr = desc->fattr;
 
@@ -331,7 +331,7 @@ nfs_find_actor(struct inode *inode, void *opaque)
 static int
 nfs_init_locked(struct inode *inode, void *opaque)
 {
-	struct nfs_find_desc	*desc = (struct nfs_find_desc *)opaque;
+	struct nfs_find_desc	*desc = opaque;
 	struct nfs_fattr	*fattr = desc->fattr;
 
 	set_nfs_fileid(inode, fattr->fileid);
@@ -2267,7 +2267,7 @@ static inline void nfs4_init_once(struct nfs_inode *nfsi)
 
 static void init_once(void *foo)
 {
-	struct nfs_inode *nfsi = (struct nfs_inode *) foo;
+	struct nfs_inode *nfsi = foo;
 
 	inode_init_once(&nfsi->vfs_inode);
 	INIT_LIST_HEAD(&nfsi->open_files);
diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h
index 898dd95bc7a7..d914d609b85b 100644
--- a/fs/nfs/internal.h
+++ b/fs/nfs/internal.h
@@ -435,7 +435,6 @@ extern void nfs_zap_acl_cache(struct inode *inode);
 extern void nfs_set_cache_invalid(struct inode *inode, unsigned long flags);
 extern bool nfs_check_cache_invalid(struct inode *, unsigned long);
 extern int nfs_wait_bit_killable(struct wait_bit_key *key, int mode);
-extern int nfs_wait_atomic_killable(atomic_t *p, unsigned int mode);
 
 /* super.c */
 extern const struct super_operations nfs_sops;
@@ -503,7 +502,6 @@ extern void nfs_pageio_init_write(struct nfs_pageio_descriptor *pgio,
 			const struct nfs_pgio_completion_ops *compl_ops);
 extern void nfs_pageio_reset_write_mds(struct nfs_pageio_descriptor *pgio);
 extern void nfs_commit_free(struct nfs_commit_data *p);
-extern void nfs_write_prepare(struct rpc_task *task, void *calldata);
 extern void nfs_commit_prepare(struct rpc_task *task, void *calldata);
 extern int nfs_initiate_commit(struct rpc_clnt *clnt,
 			       struct nfs_commit_data *data,
diff --git a/fs/nfs/nfs42proc.c b/fs/nfs/nfs42proc.c
index d37e4a5401b1..13424f0d793b 100644
--- a/fs/nfs/nfs42proc.c
+++ b/fs/nfs/nfs42proc.c
@@ -1175,6 +1175,7 @@ static int _nfs42_proc_removexattr(struct inode *inode, const char *name)
 
 	ret = nfs4_call_sync(server->client, server, &msg, &args.seq_args,
 	    &res.seq_res, 1);
+	trace_nfs4_removexattr(inode, name, ret);
 	if (!ret)
 		nfs4_update_changeattr(inode, &res.cinfo, timestamp, 0);
 
@@ -1214,6 +1215,7 @@ static int _nfs42_proc_setxattr(struct inode *inode, const char *name,
 
 	ret = nfs4_call_sync(server->client, server, &msg, &arg.seq_args,
 	    &res.seq_res, 1);
+	trace_nfs4_setxattr(inode, name, ret);
 
 	for (; np > 0; np--)
 		put_page(pages[np - 1]);
@@ -1246,6 +1248,7 @@ static ssize_t _nfs42_proc_getxattr(struct inode *inode, const char *name,
 
 	ret = nfs4_call_sync(server->client, server, &msg, &arg.seq_args,
 	    &res.seq_res, 0);
+	trace_nfs4_getxattr(inode, name, ret);
 	if (ret < 0)
 		return ret;
 
@@ -1317,6 +1320,7 @@ static ssize_t _nfs42_proc_listxattrs(struct inode *inode, void *buf,
 
 	ret = nfs4_call_sync(server->client, server, &msg, &arg.seq_args,
 	    &res.seq_res, 0);
+	trace_nfs4_listxattr(inode, ret);
 
 	if (ret >= 0) {
 		ret = res.copied;
diff --git a/fs/nfs/nfs42xattr.c b/fs/nfs/nfs42xattr.c
index a9bf09fdf2c3..76ae11834206 100644
--- a/fs/nfs/nfs42xattr.c
+++ b/fs/nfs/nfs42xattr.c
@@ -981,7 +981,7 @@ nfs4_xattr_entry_count(struct shrinker *shrink, struct shrink_control *sc)
 
 static void nfs4_xattr_cache_init_once(void *p)
 {
-	struct nfs4_xattr_cache *cache = (struct nfs4_xattr_cache *)p;
+	struct nfs4_xattr_cache *cache = p;
 
 	spin_lock_init(&cache->listxattr_lock);
 	atomic_long_set(&cache->nent, 0);
diff --git a/fs/nfs/nfs42xdr.c b/fs/nfs/nfs42xdr.c
index b56f05113d36..fe1aeb0f048f 100644
--- a/fs/nfs/nfs42xdr.c
+++ b/fs/nfs/nfs42xdr.c
@@ -569,6 +569,14 @@ static int decode_listxattrs(struct xdr_stream *xdr,
 		 */
 		if (status == -ETOOSMALL)
 			status = -ERANGE;
+		/*
+		 * Special case: for LISTXATTRS, NFS4ERR_NOXATTR
+		 * should be translated to success with zero-length reply.
+		 */
+		if (status == -ENODATA) {
+			res->eof = true;
+			status = 0;
+		}
 		goto out;
 	}
 
diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h
index 79df6e83881b..400a71e75238 100644
--- a/fs/nfs/nfs4_fs.h
+++ b/fs/nfs/nfs4_fs.h
@@ -459,7 +459,6 @@ struct nfs_client *nfs4_alloc_client(const struct nfs_client_initdata *);
 
 /* nfs4renewd.c */
 extern void nfs4_schedule_state_renewal(struct nfs_client *);
-extern void nfs4_renewd_prepare_shutdown(struct nfs_server *);
 extern void nfs4_kill_renewd(struct nfs_client *);
 extern void nfs4_renew_state(struct work_struct *);
 extern void nfs4_set_lease_period(struct nfs_client *clp, unsigned long lease);
diff --git a/fs/nfs/nfs4client.c b/fs/nfs/nfs4client.c
index 3c5678aec006..7a5162afa5c0 100644
--- a/fs/nfs/nfs4client.c
+++ b/fs/nfs/nfs4client.c
@@ -254,7 +254,7 @@ struct nfs_client *nfs4_alloc_client(const struct nfs_client_initdata *cl_init)
 			goto error;
 		ip_addr = (const char *)buf;
 	}
-	strlcpy(clp->cl_ipaddr, ip_addr, sizeof(clp->cl_ipaddr));
+	strscpy(clp->cl_ipaddr, ip_addr, sizeof(clp->cl_ipaddr));
 
 	err = nfs_idmap_new(clp);
 	if (err < 0) {
diff --git a/fs/nfs/nfs4idmap.c b/fs/nfs/nfs4idmap.c
index ec6afd3c4bca..e3fdd2f45b01 100644
--- a/fs/nfs/nfs4idmap.c
+++ b/fs/nfs/nfs4idmap.c
@@ -583,7 +583,7 @@ static int nfs_idmap_legacy_upcall(struct key *authkey, void *aux)
 	struct request_key_auth *rka = get_request_key_auth(authkey);
 	struct rpc_pipe_msg *msg;
 	struct idmap_msg *im;
-	struct idmap *idmap = (struct idmap *)aux;
+	struct idmap *idmap = aux;
 	struct key *key = rka->target_key;
 	int ret = -ENOKEY;
 
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index 4553803538e5..e2efcd26336c 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -6608,7 +6608,7 @@ static void nfs4_delegreturn_prepare(struct rpc_task *task, void *data)
 	struct nfs4_delegreturndata *d_data;
 	struct pnfs_layout_hdr *lo;
 
-	d_data = (struct nfs4_delegreturndata *)data;
+	d_data = data;
 
 	if (!d_data->lr.roc && nfs4_wait_on_layoutreturn(d_data->inode, task)) {
 		nfs4_sequence_done(task, &d_data->res.seq_res);
@@ -8900,7 +8900,7 @@ int nfs4_proc_exchange_id(struct nfs_client *clp, const struct cred *cred)
 void nfs4_test_session_trunk(struct rpc_clnt *clnt, struct rpc_xprt *xprt,
 			    void *data)
 {
-	struct nfs4_add_xprt_data *adata = (struct nfs4_add_xprt_data *)data;
+	struct nfs4_add_xprt_data *adata = data;
 	struct rpc_task *task;
 	int status;
 
diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c
index 7e185f7eb260..c3503fb26fa2 100644
--- a/fs/nfs/nfs4state.c
+++ b/fs/nfs/nfs4state.c
@@ -497,8 +497,7 @@ nfs4_alloc_state_owner(struct nfs_server *server,
 	sp = kzalloc(sizeof(*sp), gfp_flags);
 	if (!sp)
 		return NULL;
-	sp->so_seqid.owner_id = ida_simple_get(&server->openowner_id, 0, 0,
-						gfp_flags);
+	sp->so_seqid.owner_id = ida_alloc(&server->openowner_id, gfp_flags);
 	if (sp->so_seqid.owner_id < 0) {
 		kfree(sp);
 		return NULL;
@@ -534,7 +533,7 @@ static void nfs4_free_state_owner(struct nfs4_state_owner *sp)
 {
 	nfs4_destroy_seqid_counter(&sp->so_seqid);
 	put_cred(sp->so_cred);
-	ida_simple_remove(&sp->so_server->openowner_id, sp->so_seqid.owner_id);
+	ida_free(&sp->so_server->openowner_id, sp->so_seqid.owner_id);
 	kfree(sp);
 }
 
@@ -877,8 +876,7 @@ static struct nfs4_lock_state *nfs4_alloc_lock_state(struct nfs4_state *state, f
 	refcount_set(&lsp->ls_count, 1);
 	lsp->ls_state = state;
 	lsp->ls_owner = fl_owner;
-	lsp->ls_seqid.owner_id = ida_simple_get(&server->lockowner_id,
-						0, 0, GFP_KERNEL_ACCOUNT);
+	lsp->ls_seqid.owner_id = ida_alloc(&server->lockowner_id, GFP_KERNEL_ACCOUNT);
 	if (lsp->ls_seqid.owner_id < 0)
 		goto out_free;
 	INIT_LIST_HEAD(&lsp->ls_locks);
@@ -890,7 +888,7 @@ out_free:
 
 void nfs4_free_lock_state(struct nfs_server *server, struct nfs4_lock_state *lsp)
 {
-	ida_simple_remove(&server->lockowner_id, lsp->ls_seqid.owner_id);
+	ida_free(&server->lockowner_id, lsp->ls_seqid.owner_id);
 	nfs4_destroy_seqid_counter(&lsp->ls_seqid);
 	kfree(lsp);
 }
diff --git a/fs/nfs/nfs4trace.h b/fs/nfs/nfs4trace.h
index 6ee6ad3674a2..2cff5901c689 100644
--- a/fs/nfs/nfs4trace.h
+++ b/fs/nfs/nfs4trace.h
@@ -2097,6 +2097,7 @@ TRACE_EVENT(ff_layout_commit_error,
 		)
 );
 
+#ifdef CONFIG_NFS_V4_2
 TRACE_DEFINE_ENUM(NFS4_CONTENT_DATA);
 TRACE_DEFINE_ENUM(NFS4_CONTENT_HOLE);
 
@@ -2105,7 +2106,6 @@ TRACE_DEFINE_ENUM(NFS4_CONTENT_HOLE);
 		{ NFS4_CONTENT_DATA, "DATA" },		\
 		{ NFS4_CONTENT_HOLE, "HOLE" })
 
-#ifdef CONFIG_NFS_V4_2
 TRACE_EVENT(nfs4_llseek,
 		TP_PROTO(
 			const struct inode *inode,
@@ -2496,6 +2496,54 @@ TRACE_EVENT(nfs4_offload_cancel,
 			__entry->stateid_seq, __entry->stateid_hash
 		)
 );
+
+DECLARE_EVENT_CLASS(nfs4_xattr_event,
+		TP_PROTO(
+			const struct inode *inode,
+			const char *name,
+			int error
+		),
+
+		TP_ARGS(inode, name, error),
+
+		TP_STRUCT__entry(
+			__field(unsigned long, error)
+			__field(dev_t, dev)
+			__field(u32, fhandle)
+			__field(u64, fileid)
+			__string(name, name)
+		),
+
+		TP_fast_assign(
+			__entry->error = error < 0 ? -error : 0;
+			__entry->dev = inode->i_sb->s_dev;
+			__entry->fileid = NFS_FILEID(inode);
+			__entry->fhandle = nfs_fhandle_hash(NFS_FH(inode));
+			__assign_str(name, name);
+		),
+
+		TP_printk(
+			"error=%ld (%s) fileid=%02x:%02x:%llu fhandle=0x%08x "
+			"name=%s",
+			-__entry->error, show_nfs4_status(__entry->error),
+			MAJOR(__entry->dev), MINOR(__entry->dev),
+			(unsigned long long)__entry->fileid,
+			__entry->fhandle, __get_str(name)
+		)
+);
+#define DEFINE_NFS4_XATTR_EVENT(name) \
+	DEFINE_EVENT(nfs4_xattr_event, name,  \
+			TP_PROTO( \
+				const struct inode *inode, \
+				const char *name, \
+				int error \
+			), \
+			TP_ARGS(inode, name, error))
+DEFINE_NFS4_XATTR_EVENT(nfs4_getxattr);
+DEFINE_NFS4_XATTR_EVENT(nfs4_setxattr);
+DEFINE_NFS4_XATTR_EVENT(nfs4_removexattr);
+
+DEFINE_NFS4_INODE_EVENT(nfs4_listxattr);
 #endif /* CONFIG_NFS_V4_2 */
 
 #endif /* CONFIG_NFS_V4_1 */
diff --git a/fs/nfs/nfsroot.c b/fs/nfs/nfsroot.c
index fa148308822c..620329b7e6ae 100644
--- a/fs/nfs/nfsroot.c
+++ b/fs/nfs/nfsroot.c
@@ -139,7 +139,7 @@ static int __init nfs_root_setup(char *line)
 	ROOT_DEV = Root_NFS;
 
 	if (line[0] == '/' || line[0] == ',' || (line[0] >= '0' && line[0] <= '9')) {
-		strlcpy(nfs_root_parms, line, sizeof(nfs_root_parms));
+		strscpy(nfs_root_parms, line, sizeof(nfs_root_parms));
 	} else {
 		size_t n = strlen(line) + sizeof(NFS_ROOT) - 1;
 		if (n >= sizeof(nfs_root_parms))
diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c
index 1f2801bfecd1..a5db5158c634 100644
--- a/fs/nfs/pnfs.c
+++ b/fs/nfs/pnfs.c
@@ -710,6 +710,7 @@ pnfs_mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo,
 			    u32 seq)
 {
 	struct pnfs_layout_segment *lseg, *next;
+	struct nfs_server *server = NFS_SERVER(lo->plh_inode);
 	int remaining = 0;
 
 	dprintk("%s:Begin lo %p\n", __func__, lo);
@@ -722,8 +723,10 @@ pnfs_mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo,
 				"offset %llu length %llu\n", __func__,
 				lseg, lseg->pls_range.iomode, lseg->pls_seq,
 				lseg->pls_range.offset, lseg->pls_range.length);
-			if (!mark_lseg_invalid(lseg, tmp_list))
-				remaining++;
+			if (mark_lseg_invalid(lseg, tmp_list))
+				continue;
+			remaining++;
+			pnfs_lseg_cancel_io(server, lseg);
 		}
 	dprintk("%s:Return %i\n", __func__, remaining);
 	return remaining;
@@ -2485,6 +2488,7 @@ pnfs_mark_matching_lsegs_return(struct pnfs_layout_hdr *lo,
 				u32 seq)
 {
 	struct pnfs_layout_segment *lseg, *next;
+	struct nfs_server *server = NFS_SERVER(lo->plh_inode);
 	int remaining = 0;
 
 	dprintk("%s:Begin lo %p\n", __func__, lo);
@@ -2507,6 +2511,7 @@ pnfs_mark_matching_lsegs_return(struct pnfs_layout_hdr *lo,
 				continue;
 			remaining++;
 			set_bit(NFS_LSEG_LAYOUTRETURN, &lseg->pls_flags);
+			pnfs_lseg_cancel_io(server, lseg);
 		}
 
 	if (remaining) {
diff --git a/fs/nfs/pnfs.h b/fs/nfs/pnfs.h
index f331f067691b..e3e6a41f19de 100644
--- a/fs/nfs/pnfs.h
+++ b/fs/nfs/pnfs.h
@@ -169,6 +169,8 @@ struct pnfs_layoutdriver_type {
 	void (*cleanup_layoutcommit) (struct nfs4_layoutcommit_data *data);
 	int (*prepare_layoutcommit) (struct nfs4_layoutcommit_args *args);
 	int (*prepare_layoutstats) (struct nfs42_layoutstat_args *args);
+
+	void (*cancel_io)(struct pnfs_layout_segment *lseg);
 };
 
 struct pnfs_commit_ops {
@@ -685,6 +687,13 @@ pnfs_lseg_request_intersecting(struct pnfs_layout_segment *lseg, struct nfs_page
 				req_offset(req), req_last);
 }
 
+static inline void pnfs_lseg_cancel_io(struct nfs_server *server,
+				       struct pnfs_layout_segment *lseg)
+{
+	if (server->pnfs_curr_ld->cancel_io)
+		server->pnfs_curr_ld->cancel_io(lseg);
+}
+
 extern unsigned int layoutstats_timer;
 
 #ifdef NFS_DEBUG
diff --git a/fs/nfs/pnfs_nfs.c b/fs/nfs/pnfs_nfs.c
index 657c242a18ff..987c88ddeaf0 100644
--- a/fs/nfs/pnfs_nfs.c
+++ b/fs/nfs/pnfs_nfs.c
@@ -374,12 +374,12 @@ pnfs_bucket_search_commit_reqs(struct pnfs_commit_bucket *buckets,
 	return NULL;
 }
 
-/* pnfs_generic_search_commit_reqs - Search lists in @cinfo for the head reqest
+/* pnfs_generic_search_commit_reqs - Search lists in @cinfo for the head request
  *				   for @page
  * @cinfo - commit info for current inode
  * @page - page to search for matching head request
  *
- * Returns a the head request if one is found, otherwise returns NULL.
+ * Return: the head request if one is found, otherwise %NULL.
  */
 struct nfs_page *
 pnfs_generic_search_commit_reqs(struct nfs_commit_info *cinfo, struct page *page)
diff --git a/include/linux/sunrpc/clnt.h b/include/linux/sunrpc/clnt.h
index 75eea5ebb179..770ef2cb5775 100644
--- a/include/linux/sunrpc/clnt.h
+++ b/include/linux/sunrpc/clnt.h
@@ -246,6 +246,7 @@ void rpc_clnt_xprt_switch_remove_xprt(struct rpc_clnt *, struct rpc_xprt *);
 bool rpc_clnt_xprt_switch_has_addr(struct rpc_clnt *clnt,
 			const struct sockaddr *sap);
 void rpc_clnt_xprt_set_online(struct rpc_clnt *clnt, struct rpc_xprt *xprt);
+void rpc_clnt_disconnect(struct rpc_clnt *clnt);
 void rpc_cleanup_clids(void);
 
 static inline int rpc_reply_expected(struct rpc_task *task)
diff --git a/include/linux/sunrpc/sched.h b/include/linux/sunrpc/sched.h
index baeca2f564dc..b8ca3ecaf8d7 100644
--- a/include/linux/sunrpc/sched.h
+++ b/include/linux/sunrpc/sched.h
@@ -209,11 +209,17 @@ struct rpc_task *rpc_run_task(const struct rpc_task_setup *);
 struct rpc_task *rpc_run_bc_task(struct rpc_rqst *req);
 void		rpc_put_task(struct rpc_task *);
 void		rpc_put_task_async(struct rpc_task *);
+bool		rpc_task_set_rpc_status(struct rpc_task *task, int rpc_status);
+void		rpc_task_try_cancel(struct rpc_task *task, int error);
 void		rpc_signal_task(struct rpc_task *);
 void		rpc_exit_task(struct rpc_task *);
 void		rpc_exit(struct rpc_task *, int);
 void		rpc_release_calldata(const struct rpc_call_ops *, void *);
 void		rpc_killall_tasks(struct rpc_clnt *);
+unsigned long	rpc_cancel_tasks(struct rpc_clnt *clnt, int error,
+				 bool (*fnmatch)(const struct rpc_task *,
+						 const void *),
+				 const void *data);
 void		rpc_execute(struct rpc_task *);
 void		rpc_init_priority_wait_queue(struct rpc_wait_queue *, const char *);
 void		rpc_init_wait_queue(struct rpc_wait_queue *, const char *);
diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c
index c284efa3d1ef..993acf38af87 100644
--- a/net/sunrpc/clnt.c
+++ b/net/sunrpc/clnt.c
@@ -345,7 +345,7 @@ static int rpc_alloc_clid(struct rpc_clnt *clnt)
 {
 	int clid;
 
-	clid = ida_simple_get(&rpc_clids, 0, 0, GFP_KERNEL);
+	clid = ida_alloc(&rpc_clids, GFP_KERNEL);
 	if (clid < 0)
 		return clid;
 	clnt->cl_clid = clid;
@@ -354,7 +354,7 @@ static int rpc_alloc_clid(struct rpc_clnt *clnt)
 
 static void rpc_free_clid(struct rpc_clnt *clnt)
 {
-	ida_simple_remove(&rpc_clids, clnt->cl_clid);
+	ida_free(&rpc_clids, clnt->cl_clid);
 }
 
 static struct rpc_clnt * rpc_new_client(const struct rpc_create_args *args,
@@ -873,6 +873,57 @@ void rpc_killall_tasks(struct rpc_clnt *clnt)
 }
 EXPORT_SYMBOL_GPL(rpc_killall_tasks);
 
+/**
+ * rpc_cancel_tasks - try to cancel a set of RPC tasks
+ * @clnt: Pointer to RPC client
+ * @error: RPC task error value to set
+ * @fnmatch: Pointer to selector function
+ * @data: User data
+ *
+ * Uses @fnmatch to define a set of RPC tasks that are to be cancelled.
+ * The argument @error must be a negative error value.
+ */
+unsigned long rpc_cancel_tasks(struct rpc_clnt *clnt, int error,
+			       bool (*fnmatch)(const struct rpc_task *,
+					       const void *),
+			       const void *data)
+{
+	struct rpc_task *task;
+	unsigned long count = 0;
+
+	if (list_empty(&clnt->cl_tasks))
+		return 0;
+	/*
+	 * Spin lock all_tasks to prevent changes...
+	 */
+	spin_lock(&clnt->cl_lock);
+	list_for_each_entry(task, &clnt->cl_tasks, tk_task) {
+		if (!RPC_IS_ACTIVATED(task))
+			continue;
+		if (!fnmatch(task, data))
+			continue;
+		rpc_task_try_cancel(task, error);
+		count++;
+	}
+	spin_unlock(&clnt->cl_lock);
+	return count;
+}
+EXPORT_SYMBOL_GPL(rpc_cancel_tasks);
+
+static int rpc_clnt_disconnect_xprt(struct rpc_clnt *clnt,
+				    struct rpc_xprt *xprt, void *dummy)
+{
+	if (xprt_connected(xprt))
+		xprt_force_disconnect(xprt);
+	return 0;
+}
+
+void rpc_clnt_disconnect(struct rpc_clnt *clnt)
+{
+	rpc_clnt_iterate_for_each_xprt(clnt, rpc_clnt_disconnect_xprt, NULL);
+}
+EXPORT_SYMBOL_GPL(rpc_clnt_disconnect);
+
 /*
  * Properly shut down an RPC client, terminating all outstanding
  * requests.
@@ -1642,7 +1693,7 @@ static void
 __rpc_call_rpcerror(struct rpc_task *task, int tk_status, int rpc_status)
 {
 	trace_rpc_call_rpcerror(task, tk_status, rpc_status);
-	task->tk_rpc_status = rpc_status;
+	rpc_task_set_rpc_status(task, rpc_status);
 	rpc_exit(task, tk_status);
 }
 
@@ -2435,10 +2486,8 @@ rpc_check_timeout(struct rpc_task *task)
 {
 	struct rpc_clnt	*clnt = task->tk_client;
 
-	if (RPC_SIGNALLED(task)) {
-		rpc_call_rpcerror(task, -ERESTARTSYS);
+	if (RPC_SIGNALLED(task))
 		return;
-	}
 
 	if (xprt_adjust_timeout(task->tk_rqstp) == 0)
 		return;
diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c
index 46cbf151a50b..be587a308e05 100644
--- a/net/sunrpc/sched.c
+++ b/net/sunrpc/sched.c
@@ -65,6 +65,13 @@ gfp_t rpc_task_gfp_mask(void)
 }
 EXPORT_SYMBOL_GPL(rpc_task_gfp_mask);
 
+bool rpc_task_set_rpc_status(struct rpc_task *task, int rpc_status)
+{
+	if (cmpxchg(&task->tk_rpc_status, 0, rpc_status) == 0)
+		return true;
+	return false;
+}
+
 unsigned long
 rpc_task_timeout(const struct rpc_task *task)
 {
@@ -853,12 +860,25 @@ void rpc_signal_task(struct rpc_task *task)
 	if (!RPC_IS_ACTIVATED(task))
 		return;
 
+	if (!rpc_task_set_rpc_status(task, -ERESTARTSYS))
+		return;
 	trace_rpc_task_signalled(task, task->tk_action);
 	set_bit(RPC_TASK_SIGNALLED, &task->tk_runstate);
 	smp_mb__after_atomic();
 	queue = READ_ONCE(task->tk_waitqueue);
 	if (queue)
-		rpc_wake_up_queued_task_set_status(queue, task, -ERESTARTSYS);
+		rpc_wake_up_queued_task(queue, task);
+}
+
+void rpc_task_try_cancel(struct rpc_task *task, int error)
+{
+	struct rpc_wait_queue *queue;
+
+	if (!rpc_task_set_rpc_status(task, error))
+		return;
+	queue = READ_ONCE(task->tk_waitqueue);
+	if (queue)
+		rpc_wake_up_queued_task(queue, task);
 }
 
 void rpc_exit(struct rpc_task *task, int status)
@@ -905,10 +925,16 @@ static void __rpc_execute(struct rpc_task *task)
 		 * Perform the next FSM step or a pending callback.
 		 *
 		 * tk_action may be NULL if the task has been killed.
-		 * In particular, note that rpc_killall_tasks may
-		 * do this at any time, so beware when dereferencing.
 		 */
 		do_action = task->tk_action;
+		/* Tasks with an RPC error status should exit */
+		if (do_action != rpc_exit_task &&
+		    (status = READ_ONCE(task->tk_rpc_status)) != 0) {
+			task->tk_status = status;
+			if (do_action != NULL)
+				do_action = rpc_exit_task;
+		}
+		/* Callbacks override all actions */
 		if (task->tk_callback) {
 			do_action = task->tk_callback;
 			task->tk_callback = NULL;
@@ -931,14 +957,6 @@ static void __rpc_execute(struct rpc_task *task)
 		}
 
 		/*
-		 * Signalled tasks should exit rather than sleep.
-		 */
-		if (RPC_SIGNALLED(task)) {
-			task->tk_rpc_status = -ERESTARTSYS;
-			rpc_exit(task, -ERESTARTSYS);
-		}
-
-		/*
 		 * The queue->lock protects against races with
 		 * rpc_make_runnable().
 		 *
@@ -953,6 +971,12 @@ static void __rpc_execute(struct rpc_task *task)
 			spin_unlock(&queue->lock);
 			continue;
 		}
+		/* Wake up any task that has an exit status */
+		if (READ_ONCE(task->tk_rpc_status) != 0) {
+			rpc_wake_up_task_queue_locked(queue, task);
+			spin_unlock(&queue->lock);
+			continue;
+		}
 		rpc_clear_running(task);
 		spin_unlock(&queue->lock);
 		if (task_is_async)
@@ -970,10 +994,7 @@ static void __rpc_execute(struct rpc_task *task)
 			 * clean up after sleeping on some queue, we don't
 			 * break the loop here, but go around once more.
 			 */
-			trace_rpc_task_signalled(task, task->tk_action);
-			set_bit(RPC_TASK_SIGNALLED, &task->tk_runstate);
-			task->tk_rpc_status = -ERESTARTSYS;
-			rpc_exit(task, -ERESTARTSYS);
+			rpc_signal_task(task);
 		}
 		trace_rpc_task_sync_wake(task, task->tk_action);
 	}
diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index f8fae7815649..71dc26373444 100644
--- a/net/sunrpc/xprt.c
+++ b/net/sunrpc/xprt.c
@@ -1788,7 +1788,7 @@ static int xprt_alloc_id(struct rpc_xprt *xprt)
 {
 	int id;
 
-	id = ida_simple_get(&rpc_xprt_ids, 0, 0, GFP_KERNEL);
+	id = ida_alloc(&rpc_xprt_ids, GFP_KERNEL);
 	if (id < 0)
 		return id;
 
@@ -1798,7 +1798,7 @@ static int xprt_alloc_id(struct rpc_xprt *xprt)
 
 static void xprt_free_id(struct rpc_xprt *xprt)
 {
-	ida_simple_remove(&rpc_xprt_ids, xprt->id);
+	ida_free(&rpc_xprt_ids, xprt->id);
 }
 
 struct rpc_xprt *xprt_alloc(struct net *net, size_t size,
@@ -1822,10 +1822,7 @@ struct rpc_xprt *xprt_alloc(struct net *net, size_t size,
 			goto out_free;
 		list_add(&req->rq_list, &xprt->free);
 	}
-	if (max_alloc > num_prealloc)
-		xprt->max_reqs = max_alloc;
-	else
-		xprt->max_reqs = num_prealloc;
+	xprt->max_reqs = max_t(unsigned int, max_alloc, num_prealloc);
 	xprt->min_reqs = num_prealloc;
 	xprt->num_reqs = num_prealloc;
 
diff --git a/net/sunrpc/xprtmultipath.c b/net/sunrpc/xprtmultipath.c
index 685db598acbe..701250b305db 100644
--- a/net/sunrpc/xprtmultipath.c
+++ b/net/sunrpc/xprtmultipath.c
@@ -103,7 +103,7 @@ static int xprt_switch_alloc_id(struct rpc_xprt_switch *xps, gfp_t gfp_flags)
 {
 	int id;
 
-	id = ida_simple_get(&rpc_xprtswitch_ids, 0, 0, gfp_flags);
+	id = ida_alloc(&rpc_xprtswitch_ids, gfp_flags);
 	if (id < 0)
 		return id;
 
@@ -113,7 +113,7 @@ static int xprt_switch_alloc_id(struct rpc_xprt_switch *xps, gfp_t gfp_flags)
 
 static void xprt_switch_free_id(struct rpc_xprt_switch *xps)
 {
-	ida_simple_remove(&rpc_xprtswitch_ids, xps->xps_id);
+	ida_free(&rpc_xprtswitch_ids, xps->xps_id);
 }
 
 /**
diff --git a/net/sunrpc/xprtrdma/backchannel.c b/net/sunrpc/xprtrdma/backchannel.c
index faba7136dd9a..e4d84a13c566 100644
--- a/net/sunrpc/xprtrdma/backchannel.c
+++ b/net/sunrpc/xprtrdma/backchannel.c
@@ -189,7 +189,7 @@ create_req:
 		return NULL;
 
 	size = min_t(size_t, r_xprt->rx_ep->re_inline_recv, PAGE_SIZE);
-	req = rpcrdma_req_create(r_xprt, size, GFP_KERNEL);
+	req = rpcrdma_req_create(r_xprt, size);
 	if (!req)
 		return NULL;
 	if (rpcrdma_req_setup(r_xprt, req)) {
diff --git a/net/sunrpc/xprtrdma/frwr_ops.c b/net/sunrpc/xprtrdma/frwr_ops.c
index de0bdb6b729f..ffbf99894970 100644
--- a/net/sunrpc/xprtrdma/frwr_ops.c
+++ b/net/sunrpc/xprtrdma/frwr_ops.c
@@ -124,16 +124,16 @@ int frwr_mr_init(struct rpcrdma_xprt *r_xprt, struct rpcrdma_mr *mr)
 	unsigned int depth = ep->re_max_fr_depth;
 	struct scatterlist *sg;
 	struct ib_mr *frmr;
-	int rc;
+
+	sg = kcalloc_node(depth, sizeof(*sg), XPRTRDMA_GFP_FLAGS,
+			  ibdev_to_node(ep->re_id->device));
+	if (!sg)
+		return -ENOMEM;
 
 	frmr = ib_alloc_mr(ep->re_pd, ep->re_mrtype, depth);
 	if (IS_ERR(frmr))
 		goto out_mr_err;
 
-	sg = kmalloc_array(depth, sizeof(*sg), GFP_KERNEL);
-	if (!sg)
-		goto out_list_err;
-
 	mr->mr_xprt = r_xprt;
 	mr->mr_ibmr = frmr;
 	mr->mr_device = NULL;
@@ -146,13 +146,9 @@ int frwr_mr_init(struct rpcrdma_xprt *r_xprt, struct rpcrdma_mr *mr)
 	return 0;
 
 out_mr_err:
-	rc = PTR_ERR(frmr);
-	trace_xprtrdma_frwr_alloc(mr, rc);
-	return rc;
-
-out_list_err:
-	ib_dereg_mr(frmr);
-	return -ENOMEM;
+	kfree(sg);
+	trace_xprtrdma_frwr_alloc(mr, PTR_ERR(frmr));
+	return PTR_ERR(frmr);
 }
 
 /**
diff --git a/net/sunrpc/xprtrdma/svc_rdma_backchannel.c b/net/sunrpc/xprtrdma/svc_rdma_backchannel.c
index 85c8cdda98b1..aa2227a7e552 100644
--- a/net/sunrpc/xprtrdma/svc_rdma_backchannel.c
+++ b/net/sunrpc/xprtrdma/svc_rdma_backchannel.c
@@ -119,12 +119,12 @@ xprt_rdma_bc_allocate(struct rpc_task *task)
 		return -EINVAL;
 	}
 
-	page = alloc_page(RPCRDMA_DEF_GFP);
+	page = alloc_page(GFP_NOIO | __GFP_NOWARN);
 	if (!page)
 		return -ENOMEM;
 	rqst->rq_buffer = page_address(page);
 
-	rqst->rq_rbuffer = kmalloc(rqst->rq_rcvsize, RPCRDMA_DEF_GFP);
+	rqst->rq_rbuffer = kmalloc(rqst->rq_rcvsize, GFP_NOIO | __GFP_NOWARN);
 	if (!rqst->rq_rbuffer) {
 		put_page(page);
 		return -ENOMEM;
diff --git a/net/sunrpc/xprtrdma/transport.c b/net/sunrpc/xprtrdma/transport.c
index bcb37b51adf6..10bb2b929c6d 100644
--- a/net/sunrpc/xprtrdma/transport.c
+++ b/net/sunrpc/xprtrdma/transport.c
@@ -494,8 +494,7 @@ xprt_rdma_connect(struct rpc_xprt *xprt, struct rpc_task *task)
 		xprt_reconnect_backoff(xprt, RPCRDMA_INIT_REEST_TO);
 	}
 	trace_xprtrdma_op_connect(r_xprt, delay);
-	queue_delayed_work(xprtiod_workqueue, &r_xprt->rx_connect_worker,
-			   delay);
+	queue_delayed_work(system_long_wq, &r_xprt->rx_connect_worker, delay);
 }
 
 /**
diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c
index 2fbe9aaeec34..44b87e4274b4 100644
--- a/net/sunrpc/xprtrdma/verbs.c
+++ b/net/sunrpc/xprtrdma/verbs.c
@@ -76,8 +76,7 @@ static void rpcrdma_mrs_destroy(struct rpcrdma_xprt *r_xprt);
 static void rpcrdma_ep_get(struct rpcrdma_ep *ep);
 static int rpcrdma_ep_put(struct rpcrdma_ep *ep);
 static struct rpcrdma_regbuf *
-rpcrdma_regbuf_alloc(size_t size, enum dma_data_direction direction,
-		     gfp_t flags);
+rpcrdma_regbuf_alloc(size_t size, enum dma_data_direction direction);
 static void rpcrdma_regbuf_dma_unmap(struct rpcrdma_regbuf *rb);
 static void rpcrdma_regbuf_free(struct rpcrdma_regbuf *rb);
 
@@ -373,7 +372,7 @@ static int rpcrdma_ep_create(struct rpcrdma_xprt *r_xprt)
 	struct rpcrdma_ep *ep;
 	int rc;
 
-	ep = kzalloc(sizeof(*ep), GFP_KERNEL);
+	ep = kzalloc(sizeof(*ep), XPRTRDMA_GFP_FLAGS);
 	if (!ep)
 		return -ENOTCONN;
 	ep->re_xprt = &r_xprt->rx_xprt;
@@ -606,7 +605,7 @@ static struct rpcrdma_sendctx *rpcrdma_sendctx_create(struct rpcrdma_ep *ep)
 	struct rpcrdma_sendctx *sc;
 
 	sc = kzalloc(struct_size(sc, sc_sges, ep->re_attr.cap.max_send_sge),
-		     GFP_KERNEL);
+		     XPRTRDMA_GFP_FLAGS);
 	if (!sc)
 		return NULL;
 
@@ -629,7 +628,7 @@ static int rpcrdma_sendctxs_create(struct rpcrdma_xprt *r_xprt)
 	 * Sends are posted.
 	 */
 	i = r_xprt->rx_ep->re_max_requests + RPCRDMA_MAX_BC_REQUESTS;
-	buf->rb_sc_ctxs = kcalloc(i, sizeof(sc), GFP_KERNEL);
+	buf->rb_sc_ctxs = kcalloc(i, sizeof(sc), XPRTRDMA_GFP_FLAGS);
 	if (!buf->rb_sc_ctxs)
 		return -ENOMEM;
 
@@ -740,13 +739,16 @@ rpcrdma_mrs_create(struct rpcrdma_xprt *r_xprt)
 {
 	struct rpcrdma_buffer *buf = &r_xprt->rx_buf;
 	struct rpcrdma_ep *ep = r_xprt->rx_ep;
+	struct ib_device *device = ep->re_id->device;
 	unsigned int count;
 
+	/* Try to allocate enough to perform one full-sized I/O */
 	for (count = 0; count < ep->re_max_rdma_segs; count++) {
 		struct rpcrdma_mr *mr;
 		int rc;
 
-		mr = kzalloc(sizeof(*mr), GFP_KERNEL);
+		mr = kzalloc_node(sizeof(*mr), XPRTRDMA_GFP_FLAGS,
+				  ibdev_to_node(device));
 		if (!mr)
 			break;
 
@@ -791,38 +793,33 @@ void rpcrdma_mrs_refresh(struct rpcrdma_xprt *r_xprt)
 	/* If there is no underlying connection, it's no use
 	 * to wake the refresh worker.
 	 */
-	if (ep->re_connect_status == 1) {
-		/* The work is scheduled on a WQ_MEM_RECLAIM
-		 * workqueue in order to prevent MR allocation
-		 * from recursing into NFS during direct reclaim.
-		 */
-		queue_work(xprtiod_workqueue, &buf->rb_refresh_worker);
-	}
+	if (ep->re_connect_status != 1)
+		return;
+	queue_work(system_highpri_wq, &buf->rb_refresh_worker);
 }
 
 /**
  * rpcrdma_req_create - Allocate an rpcrdma_req object
  * @r_xprt: controlling r_xprt
  * @size: initial size, in bytes, of send and receive buffers
- * @flags: GFP flags passed to memory allocators
  *
  * Returns an allocated and fully initialized rpcrdma_req or NULL.
  */
-struct rpcrdma_req *rpcrdma_req_create(struct rpcrdma_xprt *r_xprt, size_t size,
-				       gfp_t flags)
+struct rpcrdma_req *rpcrdma_req_create(struct rpcrdma_xprt *r_xprt,
+				       size_t size)
 {
 	struct rpcrdma_buffer *buffer = &r_xprt->rx_buf;
 	struct rpcrdma_req *req;
 
-	req = kzalloc(sizeof(*req), flags);
+	req = kzalloc(sizeof(*req), XPRTRDMA_GFP_FLAGS);
 	if (req == NULL)
 		goto out1;
 
-	req->rl_sendbuf = rpcrdma_regbuf_alloc(size, DMA_TO_DEVICE, flags);
+	req->rl_sendbuf = rpcrdma_regbuf_alloc(size, DMA_TO_DEVICE);
 	if (!req->rl_sendbuf)
 		goto out2;
 
-	req->rl_recvbuf = rpcrdma_regbuf_alloc(size, DMA_NONE, flags);
+	req->rl_recvbuf = rpcrdma_regbuf_alloc(size, DMA_NONE);
 	if (!req->rl_recvbuf)
 		goto out3;
 
@@ -858,7 +855,7 @@ int rpcrdma_req_setup(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req)
 		     r_xprt->rx_ep->re_max_rdma_segs * rpcrdma_readchunk_maxsz;
 	maxhdrsize *= sizeof(__be32);
 	rb = rpcrdma_regbuf_alloc(__roundup_pow_of_two(maxhdrsize),
-				  DMA_TO_DEVICE, GFP_KERNEL);
+				  DMA_TO_DEVICE);
 	if (!rb)
 		goto out;
 
@@ -929,12 +926,12 @@ struct rpcrdma_rep *rpcrdma_rep_create(struct rpcrdma_xprt *r_xprt,
 	struct rpcrdma_buffer *buf = &r_xprt->rx_buf;
 	struct rpcrdma_rep *rep;
 
-	rep = kzalloc(sizeof(*rep), GFP_KERNEL);
+	rep = kzalloc(sizeof(*rep), XPRTRDMA_GFP_FLAGS);
 	if (rep == NULL)
 		goto out;
 
 	rep->rr_rdmabuf = rpcrdma_regbuf_alloc(r_xprt->rx_ep->re_inline_recv,
-					       DMA_FROM_DEVICE, GFP_KERNEL);
+					       DMA_FROM_DEVICE);
 	if (!rep->rr_rdmabuf)
 		goto out_free;
 
@@ -1064,8 +1061,8 @@ int rpcrdma_buffer_create(struct rpcrdma_xprt *r_xprt)
 	for (i = 0; i < r_xprt->rx_xprt.max_reqs; i++) {
 		struct rpcrdma_req *req;
 
-		req = rpcrdma_req_create(r_xprt, RPCRDMA_V1_DEF_INLINE_SIZE * 2,
-					 GFP_KERNEL);
+		req = rpcrdma_req_create(r_xprt,
+					 RPCRDMA_V1_DEF_INLINE_SIZE * 2);
 		if (!req)
 			goto out;
 		list_add(&req->rl_list, &buf->rb_send_bufs);
@@ -1235,15 +1232,14 @@ void rpcrdma_buffer_put(struct rpcrdma_buffer *buffers, struct rpcrdma_req *req)
  * or Replies they may be registered externally via frwr_map.
  */
 static struct rpcrdma_regbuf *
-rpcrdma_regbuf_alloc(size_t size, enum dma_data_direction direction,
-		     gfp_t flags)
+rpcrdma_regbuf_alloc(size_t size, enum dma_data_direction direction)
 {
 	struct rpcrdma_regbuf *rb;
 
-	rb = kmalloc(sizeof(*rb), flags);
+	rb = kmalloc(sizeof(*rb), XPRTRDMA_GFP_FLAGS);
 	if (!rb)
 		return NULL;
-	rb->rg_data = kmalloc(size, flags);
+	rb->rg_data = kmalloc(size, XPRTRDMA_GFP_FLAGS);
 	if (!rb->rg_data) {
 		kfree(rb);
 		return NULL;
diff --git a/net/sunrpc/xprtrdma/xprt_rdma.h b/net/sunrpc/xprtrdma/xprt_rdma.h
index c79f92eeda76..5e5ff6784ef5 100644
--- a/net/sunrpc/xprtrdma/xprt_rdma.h
+++ b/net/sunrpc/xprtrdma/xprt_rdma.h
@@ -149,7 +149,11 @@ static inline void *rdmab_data(const struct rpcrdma_regbuf *rb)
 	return rb->rg_data;
 }
 
-#define RPCRDMA_DEF_GFP		(GFP_NOIO | __GFP_NOWARN)
+/* Do not use emergency memory reserves, and fail quickly if memory
+ * cannot be allocated easily. These flags may be used wherever there
+ * is robust logic to handle a failure to allocate.
+ */
+#define XPRTRDMA_GFP_FLAGS  (__GFP_NOMEMALLOC | __GFP_NORETRY | __GFP_NOWARN)
 
 /* To ensure a transport can always make forward progress,
  * the number of RDMA segments allowed in header chunk lists
@@ -467,8 +471,8 @@ void rpcrdma_post_recvs(struct rpcrdma_xprt *r_xprt, int needed, bool temp);
 /*
  * Buffer calls - xprtrdma/verbs.c
  */
-struct rpcrdma_req *rpcrdma_req_create(struct rpcrdma_xprt *r_xprt, size_t size,
-				       gfp_t flags);
+struct rpcrdma_req *rpcrdma_req_create(struct rpcrdma_xprt *r_xprt,
+				       size_t size);
 int rpcrdma_req_setup(struct rpcrdma_xprt *r_xprt, struct rpcrdma_req *req);
 void rpcrdma_req_destroy(struct rpcrdma_req *req);
 int rpcrdma_buffer_create(struct rpcrdma_xprt *);
diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c
index e976007f4fd0..f34d5427b66c 100644
--- a/net/sunrpc/xprtsock.c
+++ b/net/sunrpc/xprtsock.c
@@ -261,7 +261,7 @@ static void xs_format_common_peer_addresses(struct rpc_xprt *xprt)
 	switch (sap->sa_family) {
 	case AF_LOCAL:
 		sun = xs_addr_un(xprt);
-		strlcpy(buf, sun->sun_path, sizeof(buf));
+		strscpy(buf, sun->sun_path, sizeof(buf));
 		xprt->address_strings[RPC_DISPLAY_ADDR] =
 						kstrdup(buf, GFP_KERNEL);
 		break;
@@ -1978,8 +1978,7 @@ static void xs_local_connect(struct rpc_xprt *xprt, struct rpc_task *task)
 		 * we'll need to figure out how to pass a namespace to
 		 * connect.
 		 */
-		task->tk_rpc_status = -ENOTCONN;
-		rpc_exit(task, -ENOTCONN);
+		rpc_task_set_rpc_status(task, -ENOTCONN);
 		goto out_wake;
 	}
 	ret = xs_local_setup_socket(transport);