3 files changed, 306 insertions, 237 deletions

diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c
index bb9148b83166..1ec79ccf89ad 100644
--- a/fs/nfs/flexfilelayout/flexfilelayout.c
+++ b/fs/nfs/flexfilelayout/flexfilelayout.c
@@ -30,13 +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);
@@ -48,9 +55,11 @@ ff_layout_alloc_layout_hdr(struct inode *inode, gfp_t gfp_flags)
 
 	ffl = kzalloc(sizeof(*ffl), gfp_flags);
 	if (ffl) {
+		pnfs_init_ds_commit_info(&ffl->commit_info);
 		INIT_LIST_HEAD(&ffl->error_list);
 		INIT_LIST_HEAD(&ffl->mirrors);
 		ffl->last_report_time = ktime_get();
+		ffl->commit_info.ops = &ff_layout_commit_ops;
 		return &ffl->generic_hdr;
 	} else
 		return NULL;
@@ -59,14 +68,14 @@ ff_layout_alloc_layout_hdr(struct inode *inode, gfp_t gfp_flags)
 static void
 ff_layout_free_layout_hdr(struct pnfs_layout_hdr *lo)
 {
+	struct nfs4_flexfile_layout *ffl = FF_LAYOUT_FROM_HDR(lo);
 	struct nfs4_ff_layout_ds_err *err, *n;
 
-	list_for_each_entry_safe(err, n, &FF_LAYOUT_FROM_HDR(lo)->error_list,
-				 list) {
+	list_for_each_entry_safe(err, n, &ffl->error_list, list) {
 		list_del(&err->list);
 		kfree(err);
 	}
-	kfree(FF_LAYOUT_FROM_HDR(lo));
+	kfree_rcu(ffl, generic_hdr.plh_rcu);
 }
 
 static int decode_pnfs_stateid(struct xdr_stream *xdr, nfs4_stateid *stateid)
@@ -103,7 +112,7 @@ static int decode_nfs_fh(struct xdr_stream *xdr, struct nfs_fh *fh)
 	if (unlikely(!p))
 		return -ENOBUFS;
 	fh->size = be32_to_cpup(p++);
-	if (fh->size > sizeof(struct nfs_fh)) {
+	if (fh->size > NFS_MAXFHSIZE) {
 		printk(KERN_ERR "NFS flexfiles: Too big fh received %d\n",
 		       fh->size);
 		return -EOVERFLOW;
@@ -248,36 +257,10 @@ static void ff_layout_put_mirror(struct nfs4_ff_layout_mirror *mirror)
 
 static void ff_layout_free_mirror_array(struct nfs4_ff_layout_segment *fls)
 {
-	int i;
-
-	if (fls->mirror_array) {
-		for (i = 0; i < fls->mirror_array_cnt; i++) {
-			/* normally mirror_ds is freed in
-			 * .free_deviceid_node but we still do it here
-			 * for .alloc_lseg error path */
-			ff_layout_put_mirror(fls->mirror_array[i]);
-		}
-		kfree(fls->mirror_array);
-		fls->mirror_array = NULL;
-	}
-}
-
-static int ff_layout_check_layout(struct nfs4_layoutget_res *lgr)
-{
-	int ret = 0;
-
-	dprintk("--> %s\n", __func__);
-
-	/* FIXME: remove this check when layout segment support is added */
-	if (lgr->range.offset != 0 ||
-	    lgr->range.length != NFS4_MAX_UINT64) {
-		dprintk("%s Only whole file layouts supported. Use MDS i/o\n",
-			__func__);
-		ret = -EINVAL;
-	}
+	u32 i;
 
-	dprintk("--> %s returns %d\n", __func__, ret);
-	return ret;
+	for (i = 0; i < fls->mirror_array_cnt; i++)
+		ff_layout_put_mirror(fls->mirror_array[i]);
 }
 
 static void _ff_layout_free_lseg(struct nfs4_ff_layout_segment *fls)
@@ -289,6 +272,23 @@ static void _ff_layout_free_lseg(struct nfs4_ff_layout_segment *fls)
 }
 
 static bool
+ff_lseg_match_mirrors(struct pnfs_layout_segment *l1,
+		struct pnfs_layout_segment *l2)
+{
+	const struct nfs4_ff_layout_segment *fl1 = FF_LAYOUT_LSEG(l1);
+	const struct nfs4_ff_layout_segment *fl2 = FF_LAYOUT_LSEG(l1);
+	u32 i;
+
+	if (fl1->mirror_array_cnt != fl2->mirror_array_cnt)
+		return false;
+	for (i = 0; i < fl1->mirror_array_cnt; i++) {
+		if (fl1->mirror_array[i] != fl2->mirror_array[i])
+			return false;
+	}
+	return true;
+}
+
+static bool
 ff_lseg_range_is_after(const struct pnfs_layout_range *l1,
 		const struct pnfs_layout_range *l2)
 {
@@ -323,6 +323,8 @@ ff_lseg_merge(struct pnfs_layout_segment *new,
 			new->pls_range.length);
 	if (new_end < old->pls_range.offset)
 		return false;
+	if (!ff_lseg_match_mirrors(new, old))
+		return false;
 
 	/* Mergeable: copy info from 'old' to 'new' */
 	if (new_end < old_end)
@@ -382,7 +384,7 @@ ff_layout_alloc_lseg(struct pnfs_layout_hdr *lh,
 
 	xdr_init_decode_pages(&stream, &buf, lgr->layoutp->pages,
 			      lgr->layoutp->len);
-	xdr_set_scratch_buffer(&stream, page_address(scratch), PAGE_SIZE);
+	xdr_set_scratch_page(&stream, scratch);
 
 	/* stripe unit and mirror_array_cnt */
 	rc = -EIO;
@@ -400,16 +402,13 @@ ff_layout_alloc_lseg(struct pnfs_layout_hdr *lh,
 		goto out_err_free;
 
 	rc = -ENOMEM;
-	fls = kzalloc(sizeof(*fls), gfp_flags);
+	fls = kzalloc(struct_size(fls, mirror_array, mirror_array_cnt),
+			gfp_flags);
 	if (!fls)
 		goto out_err_free;
 
 	fls->mirror_array_cnt = mirror_array_cnt;
 	fls->stripe_unit = stripe_unit;
-	fls->mirror_array = kcalloc(fls->mirror_array_cnt,
-				    sizeof(fls->mirror_array[0]), gfp_flags);
-	if (fls->mirror_array == NULL)
-		goto out_err_free;
 
 	for (i = 0; i < fls->mirror_array_cnt; i++) {
 		struct nfs4_ff_layout_mirror *mirror;
@@ -545,9 +544,6 @@ ff_layout_alloc_lseg(struct pnfs_layout_hdr *lh,
 
 out_sort_mirrors:
 	ff_layout_sort_mirrors(fls);
-	rc = ff_layout_check_layout(lgr);
-	if (rc)
-		goto out_err_free;
 	ret = &fls->generic_hdr;
 	dprintk("<-- %s (success)\n", __func__);
 out_free_page:
@@ -560,17 +556,6 @@ out_err_free:
 	goto out_free_page;
 }
 
-static bool ff_layout_has_rw_segments(struct pnfs_layout_hdr *layout)
-{
-	struct pnfs_layout_segment *lseg;
-
-	list_for_each_entry(lseg, &layout->plh_segs, pls_list)
-		if (lseg->pls_range.iomode == IOMODE_RW)
-			return true;
-
-	return false;
-}
-
 static void
 ff_layout_free_lseg(struct pnfs_layout_segment *lseg)
 {
@@ -585,23 +570,12 @@ ff_layout_free_lseg(struct pnfs_layout_segment *lseg)
 		ffl = FF_LAYOUT_FROM_HDR(lseg->pls_layout);
 		inode = ffl->generic_hdr.plh_inode;
 		spin_lock(&inode->i_lock);
-		if (!ff_layout_has_rw_segments(lseg->pls_layout)) {
-			ffl->commit_info.nbuckets = 0;
-			kfree(ffl->commit_info.buckets);
-			ffl->commit_info.buckets = NULL;
-		}
+		pnfs_generic_ds_cinfo_release_lseg(&ffl->commit_info, lseg);
 		spin_unlock(&inode->i_lock);
 	}
 	_ff_layout_free_lseg(fls);
 }
 
-/* Return 1 until we have multiple lsegs support */
-static int
-ff_layout_get_lseg_count(struct nfs4_ff_layout_segment *fls)
-{
-	return 1;
-}
-
 static void
 nfs4_ff_start_busy_timer(struct nfs4_ff_busy_timer *timer, ktime_t now)
 {
@@ -695,7 +669,7 @@ nfs4_ff_layout_stat_io_start_read(struct inode *inode,
 	spin_unlock(&mirror->lock);
 
 	if (report)
-		pnfs_report_layoutstat(inode, GFP_KERNEL);
+		pnfs_report_layoutstat(inode, nfs_io_gfp_mask());
 }
 
 static void
@@ -726,7 +700,7 @@ nfs4_ff_layout_stat_io_start_write(struct inode *inode,
 	spin_unlock(&mirror->lock);
 
 	if (report)
-		pnfs_report_layoutstat(inode, GFP_NOIO);
+		pnfs_report_layoutstat(inode, nfs_io_gfp_mask());
 }
 
 static void
@@ -746,54 +720,8 @@ nfs4_ff_layout_stat_io_end_write(struct rpc_task *task,
 	spin_unlock(&mirror->lock);
 }
 
-static int
-ff_layout_alloc_commit_info(struct pnfs_layout_segment *lseg,
-			    struct nfs_commit_info *cinfo,
-			    gfp_t gfp_flags)
-{
-	struct nfs4_ff_layout_segment *fls = FF_LAYOUT_LSEG(lseg);
-	struct pnfs_commit_bucket *buckets;
-	int size;
-
-	if (cinfo->ds->nbuckets != 0) {
-		/* This assumes there is only one RW lseg per file.
-		 * To support multiple lseg per file, we need to
-		 * change struct pnfs_commit_bucket to allow dynamic
-		 * increasing nbuckets.
-		 */
-		return 0;
-	}
-
-	size = ff_layout_get_lseg_count(fls) * FF_LAYOUT_MIRROR_COUNT(lseg);
-
-	buckets = kcalloc(size, sizeof(struct pnfs_commit_bucket),
-			  gfp_flags);
-	if (!buckets)
-		return -ENOMEM;
-	else {
-		int i;
-
-		spin_lock(&cinfo->inode->i_lock);
-		if (cinfo->ds->nbuckets != 0)
-			kfree(buckets);
-		else {
-			cinfo->ds->buckets = buckets;
-			cinfo->ds->nbuckets = size;
-			for (i = 0; i < size; i++) {
-				INIT_LIST_HEAD(&buckets[i].written);
-				INIT_LIST_HEAD(&buckets[i].committing);
-				/* mark direct verifier as unset */
-				buckets[i].direct_verf.committed =
-					NFS_INVALID_STABLE_HOW;
-			}
-		}
-		spin_unlock(&cinfo->inode->i_lock);
-		return 0;
-	}
-}
-
 static void
-ff_layout_mark_ds_unreachable(struct pnfs_layout_segment *lseg, int idx)
+ff_layout_mark_ds_unreachable(struct pnfs_layout_segment *lseg, u32 idx)
 {
 	struct nfs4_deviceid_node *devid = FF_LAYOUT_DEVID_NODE(lseg, idx);
 
@@ -802,7 +730,7 @@ ff_layout_mark_ds_unreachable(struct pnfs_layout_segment *lseg, int idx)
 }
 
 static void
-ff_layout_mark_ds_reachable(struct pnfs_layout_segment *lseg, int idx)
+ff_layout_mark_ds_reachable(struct pnfs_layout_segment *lseg, u32 idx)
 {
 	struct nfs4_deviceid_node *devid = FF_LAYOUT_DEVID_NODE(lseg, idx);
 
@@ -812,22 +740,18 @@ ff_layout_mark_ds_reachable(struct pnfs_layout_segment *lseg, int idx)
 
 static struct nfs4_pnfs_ds *
 ff_layout_choose_ds_for_read(struct pnfs_layout_segment *lseg,
-			     int start_idx, int *best_idx,
+			     u32 start_idx, u32 *best_idx,
 			     bool check_device)
 {
 	struct nfs4_ff_layout_segment *fls = FF_LAYOUT_LSEG(lseg);
 	struct nfs4_ff_layout_mirror *mirror;
 	struct nfs4_pnfs_ds *ds;
-	bool fail_return = false;
-	int idx;
+	u32 idx;
 
 	/* mirrors are initially sorted by efficiency */
 	for (idx = start_idx; idx < fls->mirror_array_cnt; idx++) {
-		if (idx+1 == fls->mirror_array_cnt)
-			fail_return = !check_device;
-
 		mirror = FF_LAYOUT_COMP(lseg, idx);
-		ds = nfs4_ff_layout_prepare_ds(lseg, mirror, fail_return);
+		ds = nfs4_ff_layout_prepare_ds(lseg, mirror, false);
 		if (!ds)
 			continue;
 
@@ -844,21 +768,21 @@ ff_layout_choose_ds_for_read(struct pnfs_layout_segment *lseg,
 
 static struct nfs4_pnfs_ds *
 ff_layout_choose_any_ds_for_read(struct pnfs_layout_segment *lseg,
-				 int start_idx, int *best_idx)
+				 u32 start_idx, u32 *best_idx)
 {
 	return ff_layout_choose_ds_for_read(lseg, start_idx, best_idx, false);
 }
 
 static struct nfs4_pnfs_ds *
 ff_layout_choose_valid_ds_for_read(struct pnfs_layout_segment *lseg,
-				   int start_idx, int *best_idx)
+				   u32 start_idx, u32 *best_idx)
 {
 	return ff_layout_choose_ds_for_read(lseg, start_idx, best_idx, true);
 }
 
 static struct nfs4_pnfs_ds *
 ff_layout_choose_best_ds_for_read(struct pnfs_layout_segment *lseg,
-				  int start_idx, int *best_idx)
+				  u32 start_idx, u32 *best_idx)
 {
 	struct nfs4_pnfs_ds *ds;
 
@@ -868,19 +792,30 @@ ff_layout_choose_best_ds_for_read(struct pnfs_layout_segment *lseg,
 	return ff_layout_choose_any_ds_for_read(lseg, start_idx, best_idx);
 }
 
+static struct nfs4_pnfs_ds *
+ff_layout_get_ds_for_read(struct nfs_pageio_descriptor *pgio,
+			  u32 *best_idx)
+{
+	struct pnfs_layout_segment *lseg = pgio->pg_lseg;
+	struct nfs4_pnfs_ds *ds;
+
+	ds = ff_layout_choose_best_ds_for_read(lseg, pgio->pg_mirror_idx,
+					       best_idx);
+	if (ds || !pgio->pg_mirror_idx)
+		return ds;
+	return ff_layout_choose_best_ds_for_read(lseg, 0, best_idx);
+}
+
 static void
 ff_layout_pg_get_read(struct nfs_pageio_descriptor *pgio,
 		      struct nfs_page *req,
 		      bool strict_iomode)
 {
 	pnfs_put_lseg(pgio->pg_lseg);
-	pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
-					   nfs_req_openctx(req),
-					   0,
-					   NFS4_MAX_UINT64,
-					   IOMODE_READ,
-					   strict_iomode,
-					   GFP_KERNEL);
+	pgio->pg_lseg =
+		pnfs_update_layout(pgio->pg_inode, nfs_req_openctx(req),
+				   req_offset(req), req->wb_bytes, IOMODE_READ,
+				   strict_iomode, nfs_io_gfp_mask());
 	if (IS_ERR(pgio->pg_lseg)) {
 		pgio->pg_error = PTR_ERR(pgio->pg_lseg);
 		pgio->pg_lseg = NULL;
@@ -888,16 +823,24 @@ ff_layout_pg_get_read(struct nfs_pageio_descriptor *pgio,
 }
 
 static void
+ff_layout_pg_check_layout(struct nfs_pageio_descriptor *pgio,
+			  struct nfs_page *req)
+{
+	pnfs_generic_pg_check_layout(pgio);
+	pnfs_generic_pg_check_range(pgio, req);
+}
+
+static void
 ff_layout_pg_init_read(struct nfs_pageio_descriptor *pgio,
 			struct nfs_page *req)
 {
 	struct nfs_pgio_mirror *pgm;
 	struct nfs4_ff_layout_mirror *mirror;
 	struct nfs4_pnfs_ds *ds;
-	int ds_idx;
+	u32 ds_idx;
 
 retry:
-	pnfs_generic_pg_check_layout(pgio);
+	ff_layout_pg_check_layout(pgio, req);
 	/* Use full layout for now */
 	if (!pgio->pg_lseg) {
 		ff_layout_pg_get_read(pgio, req, false);
@@ -910,25 +853,22 @@ retry:
 			goto out_nolseg;
 	}
 
-	ds = ff_layout_choose_best_ds_for_read(pgio->pg_lseg, 0, &ds_idx);
+	ds = ff_layout_get_ds_for_read(pgio, &ds_idx);
 	if (!ds) {
 		if (!ff_layout_no_fallback_to_mds(pgio->pg_lseg))
 			goto out_mds;
-		pnfs_put_lseg(pgio->pg_lseg);
-		pgio->pg_lseg = NULL;
+		pnfs_generic_pg_cleanup(pgio);
 		/* Sleep for 1 second before retrying */
 		ssleep(1);
 		goto retry;
 	}
 
 	mirror = FF_LAYOUT_COMP(pgio->pg_lseg, ds_idx);
-
-	pgio->pg_mirror_idx = ds_idx;
-
-	/* read always uses only one mirror - idx 0 for pgio layer */
 	pgm = &pgio->pg_mirrors[0];
 	pgm->pg_bsize = mirror->mirror_ds->ds_versions[0].rsize;
 
+	pgio->pg_mirror_idx = ds_idx;
+
 	if (NFS_SERVER(pgio->pg_inode)->flags &
 			(NFS_MOUNT_SOFT|NFS_MOUNT_SOFTERR))
 		pgio->pg_maxretrans = io_maxretrans;
@@ -941,8 +881,6 @@ out_mds:
 			0, NFS4_MAX_UINT64, IOMODE_READ,
 			NFS_I(pgio->pg_inode)->layout,
 			pgio->pg_lseg);
-	pnfs_put_lseg(pgio->pg_lseg);
-	pgio->pg_lseg = NULL;
 	pgio->pg_maxretrans = 0;
 	nfs_pageio_reset_read_mds(pgio);
 }
@@ -953,21 +891,16 @@ ff_layout_pg_init_write(struct nfs_pageio_descriptor *pgio,
 {
 	struct nfs4_ff_layout_mirror *mirror;
 	struct nfs_pgio_mirror *pgm;
-	struct nfs_commit_info cinfo;
 	struct nfs4_pnfs_ds *ds;
-	int i;
-	int status;
+	u32 i;
 
 retry:
-	pnfs_generic_pg_check_layout(pgio);
+	ff_layout_pg_check_layout(pgio, req);
 	if (!pgio->pg_lseg) {
-		pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
-						   nfs_req_openctx(req),
-						   0,
-						   NFS4_MAX_UINT64,
-						   IOMODE_RW,
-						   false,
-						   GFP_NOFS);
+		pgio->pg_lseg =
+			pnfs_update_layout(pgio->pg_inode, nfs_req_openctx(req),
+					   req_offset(req), req->wb_bytes,
+					   IOMODE_RW, false, nfs_io_gfp_mask());
 		if (IS_ERR(pgio->pg_lseg)) {
 			pgio->pg_error = PTR_ERR(pgio->pg_lseg);
 			pgio->pg_lseg = NULL;
@@ -978,15 +911,9 @@ retry:
 	if (pgio->pg_lseg == NULL)
 		goto out_mds;
 
-	nfs_init_cinfo(&cinfo, pgio->pg_inode, pgio->pg_dreq);
-	status = ff_layout_alloc_commit_info(pgio->pg_lseg, &cinfo, GFP_NOFS);
-	if (status < 0)
-		goto out_mds;
-
 	/* Use a direct mapping of ds_idx to pgio mirror_idx */
-	if (WARN_ON_ONCE(pgio->pg_mirror_count !=
-	    FF_LAYOUT_MIRROR_COUNT(pgio->pg_lseg)))
-		goto out_mds;
+	if (pgio->pg_mirror_count != FF_LAYOUT_MIRROR_COUNT(pgio->pg_lseg))
+		goto out_eagain;
 
 	for (i = 0; i < pgio->pg_mirror_count; i++) {
 		mirror = FF_LAYOUT_COMP(pgio->pg_lseg, i);
@@ -994,8 +921,7 @@ retry:
 		if (!ds) {
 			if (!ff_layout_no_fallback_to_mds(pgio->pg_lseg))
 				goto out_mds;
-			pnfs_put_lseg(pgio->pg_lseg);
-			pgio->pg_lseg = NULL;
+			pnfs_generic_pg_cleanup(pgio);
 			/* Sleep for 1 second before retrying */
 			ssleep(1);
 			goto retry;
@@ -1008,16 +934,18 @@ retry:
 			(NFS_MOUNT_SOFT|NFS_MOUNT_SOFTERR))
 		pgio->pg_maxretrans = io_maxretrans;
 	return;
-
+out_eagain:
+	pnfs_generic_pg_cleanup(pgio);
+	pgio->pg_error = -EAGAIN;
+	return;
 out_mds:
 	trace_pnfs_mds_fallback_pg_init_write(pgio->pg_inode,
 			0, NFS4_MAX_UINT64, IOMODE_RW,
 			NFS_I(pgio->pg_inode)->layout,
 			pgio->pg_lseg);
-	pnfs_put_lseg(pgio->pg_lseg);
-	pgio->pg_lseg = NULL;
 	pgio->pg_maxretrans = 0;
 	nfs_pageio_reset_write_mds(pgio);
+	pgio->pg_error = -EAGAIN;
 }
 
 static unsigned int
@@ -1025,13 +953,10 @@ ff_layout_pg_get_mirror_count_write(struct nfs_pageio_descriptor *pgio,
 				    struct nfs_page *req)
 {
 	if (!pgio->pg_lseg) {
-		pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
-						   nfs_req_openctx(req),
-						   0,
-						   NFS4_MAX_UINT64,
-						   IOMODE_RW,
-						   false,
-						   GFP_NOFS);
+		pgio->pg_lseg =
+			pnfs_update_layout(pgio->pg_inode, nfs_req_openctx(req),
+					   req_offset(req), req->wb_bytes,
+					   IOMODE_RW, false, nfs_io_gfp_mask());
 		if (IS_ERR(pgio->pg_lseg)) {
 			pgio->pg_error = PTR_ERR(pgio->pg_lseg);
 			pgio->pg_lseg = NULL;
@@ -1051,6 +976,21 @@ out:
 	return 1;
 }
 
+static u32
+ff_layout_pg_set_mirror_write(struct nfs_pageio_descriptor *desc, u32 idx)
+{
+	u32 old = desc->pg_mirror_idx;
+
+	desc->pg_mirror_idx = idx;
+	return old;
+}
+
+static struct nfs_pgio_mirror *
+ff_layout_pg_get_mirror_write(struct nfs_pageio_descriptor *desc, u32 idx)
+{
+	return &desc->pg_mirrors[idx];
+}
+
 static const struct nfs_pageio_ops ff_layout_pg_read_ops = {
 	.pg_init = ff_layout_pg_init_read,
 	.pg_test = pnfs_generic_pg_test,
@@ -1064,6 +1004,8 @@ static const struct nfs_pageio_ops ff_layout_pg_write_ops = {
 	.pg_doio = pnfs_generic_pg_writepages,
 	.pg_get_mirror_count = ff_layout_pg_get_mirror_count_write,
 	.pg_cleanup = pnfs_generic_pg_cleanup,
+	.pg_get_mirror = ff_layout_pg_get_mirror_write,
+	.pg_set_mirror = ff_layout_pg_set_mirror_write,
 };
 
 static void ff_layout_reset_write(struct nfs_pgio_header *hdr, bool retry_pnfs)
@@ -1102,11 +1044,24 @@ static void ff_layout_reset_write(struct nfs_pgio_header *hdr, bool retry_pnfs)
 	}
 }
 
+static void ff_layout_resend_pnfs_read(struct nfs_pgio_header *hdr)
+{
+	u32 idx = hdr->pgio_mirror_idx + 1;
+	u32 new_idx = 0;
+
+	if (ff_layout_choose_any_ds_for_read(hdr->lseg, idx, &new_idx))
+		ff_layout_send_layouterror(hdr->lseg);
+	else
+		pnfs_error_mark_layout_for_return(hdr->inode, hdr->lseg);
+	pnfs_read_resend_pnfs(hdr, new_idx);
+}
+
 static void ff_layout_reset_read(struct nfs_pgio_header *hdr)
 {
 	struct rpc_task *task = &hdr->task;
 
 	pnfs_layoutcommit_inode(hdr->inode, false);
+	pnfs_error_mark_layout_for_return(hdr->inode, hdr->lseg);
 
 	if (!test_and_set_bit(NFS_IOHDR_REDO, &hdr->flags)) {
 		dprintk("%s Reset task %5u for i/o through MDS "
@@ -1129,7 +1084,7 @@ static int ff_layout_async_handle_error_v4(struct rpc_task *task,
 					   struct nfs4_state *state,
 					   struct nfs_client *clp,
 					   struct pnfs_layout_segment *lseg,
-					   int idx)
+					   u32 idx)
 {
 	struct pnfs_layout_hdr *lo = lseg->pls_layout;
 	struct inode *inode = lo->plh_inode;
@@ -1182,12 +1137,14 @@ static int ff_layout_async_handle_error_v4(struct rpc_task *task,
 	case -EIO:
 	case -ETIMEDOUT:
 	case -EPIPE:
+	case -EPROTO:
+	case -ENODEV:
 		dprintk("%s DS connection error %d\n", __func__,
 			task->tk_status);
 		nfs4_delete_deviceid(devid->ld, devid->nfs_client,
 				&devid->deviceid);
 		rpc_wake_up(&tbl->slot_tbl_waitq);
-		/* fall through */
+		fallthrough;
 	default:
 		if (ff_layout_avoid_mds_available_ds(lseg))
 			return -NFS4ERR_RESET_TO_PNFS;
@@ -1203,7 +1160,7 @@ reset:
 /* Retry all errors through either pNFS or MDS except for -EJUKEBOX */
 static int ff_layout_async_handle_error_v3(struct rpc_task *task,
 					   struct pnfs_layout_segment *lseg,
-					   int idx)
+					   u32 idx)
 {
 	struct nfs4_deviceid_node *devid = FF_LAYOUT_DEVID_NODE(lseg, idx);
 
@@ -1238,7 +1195,7 @@ static int ff_layout_async_handle_error(struct rpc_task *task,
 					struct nfs4_state *state,
 					struct nfs_client *clp,
 					struct pnfs_layout_segment *lseg,
-					int idx)
+					u32 idx)
 {
 	int vers = clp->cl_nfs_mod->rpc_vers->number;
 
@@ -1265,7 +1222,7 @@ static int ff_layout_async_handle_error(struct rpc_task *task,
 }
 
 static void ff_layout_io_track_ds_error(struct pnfs_layout_segment *lseg,
-					int idx, u64 offset, u64 length,
+					u32 idx, u64 offset, u64 length,
 					u32 *op_status, int opnum, int error)
 {
 	struct nfs4_ff_layout_mirror *mirror;
@@ -1287,6 +1244,8 @@ static void ff_layout_io_track_ds_error(struct pnfs_layout_segment *lseg,
 		case -ENOBUFS:
 		case -EPIPE:
 		case -EPERM:
+		case -EPROTO:
+		case -ENODEV:
 			*op_status = status = NFS4ERR_NXIO;
 			break;
 		case -EACCES:
@@ -1297,21 +1256,29 @@ static void ff_layout_io_track_ds_error(struct pnfs_layout_segment *lseg,
 		}
 	}
 
+	mirror = FF_LAYOUT_COMP(lseg, idx);
+	err = ff_layout_track_ds_error(FF_LAYOUT_FROM_HDR(lseg->pls_layout),
+				       mirror, offset, length, status, opnum,
+				       nfs_io_gfp_mask());
+
 	switch (status) {
 	case NFS4ERR_DELAY:
 	case NFS4ERR_GRACE:
-		return;
-	default:
 		break;
+	case NFS4ERR_NXIO:
+		ff_layout_mark_ds_unreachable(lseg, idx);
+		/*
+		 * Don't return the layout if this is a read and we still
+		 * have layouts to try
+		 */
+		if (opnum == OP_READ)
+			break;
+		fallthrough;
+	default:
+		pnfs_error_mark_layout_for_return(lseg->pls_layout->plh_inode,
+						  lseg);
 	}
 
-	mirror = FF_LAYOUT_COMP(lseg, idx);
-	err = ff_layout_track_ds_error(FF_LAYOUT_FROM_HDR(lseg->pls_layout),
-				       mirror, offset, length, status, opnum,
-				       GFP_NOIO);
-	if (status == NFS4ERR_NXIO)
-		ff_layout_mark_ds_unreachable(lseg, idx);
-	pnfs_error_mark_layout_for_return(lseg->pls_layout->plh_inode, lseg);
 	dprintk("%s: err %d op %d status %u\n", __func__, err, opnum, status);
 }
 
@@ -1319,7 +1286,6 @@ static void ff_layout_io_track_ds_error(struct pnfs_layout_segment *lseg,
 static int ff_layout_read_done_cb(struct rpc_task *task,
 				struct nfs_pgio_header *hdr)
 {
-	int new_idx = hdr->pgio_mirror_idx;
 	int err;
 
 	if (task->tk_status < 0) {
@@ -1339,10 +1305,6 @@ static int ff_layout_read_done_cb(struct rpc_task *task,
 	clear_bit(NFS_IOHDR_RESEND_MDS, &hdr->flags);
 	switch (err) {
 	case -NFS4ERR_RESET_TO_PNFS:
-		if (ff_layout_choose_best_ds_for_read(hdr->lseg,
-					hdr->pgio_mirror_idx + 1,
-					&new_idx))
-			goto out_layouterror;
 		set_bit(NFS_IOHDR_RESEND_PNFS, &hdr->flags);
 		return task->tk_status;
 	case -NFS4ERR_RESET_TO_MDS:
@@ -1353,10 +1315,6 @@ static int ff_layout_read_done_cb(struct rpc_task *task,
 	}
 
 	return 0;
-out_layouterror:
-	ff_layout_read_record_layoutstats_done(task, hdr);
-	ff_layout_send_layouterror(hdr->lseg);
-	hdr->pgio_mirror_idx = new_idx;
 out_eagain:
 	rpc_restart_call_prepare(task);
 	return -EAGAIN;
@@ -1421,6 +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;
 }
@@ -1457,8 +1420,6 @@ static void ff_layout_read_call_done(struct rpc_task *task, void *data)
 {
 	struct nfs_pgio_header *hdr = data;
 
-	dprintk("--> %s task->tk_status %d\n", __func__, task->tk_status);
-
 	if (test_bit(NFS_IOHDR_REDO, &hdr->flags) &&
 	    task->tk_status == 0) {
 		nfs4_sequence_done(task, &hdr->res.seq_res);
@@ -1483,10 +1444,9 @@ static void ff_layout_read_release(void *data)
 	struct nfs_pgio_header *hdr = data;
 
 	ff_layout_read_record_layoutstats_done(&hdr->task, hdr);
-	if (test_bit(NFS_IOHDR_RESEND_PNFS, &hdr->flags)) {
-		ff_layout_send_layouterror(hdr->lseg);
-		pnfs_read_resend_pnfs(hdr);
-	} else if (test_bit(NFS_IOHDR_RESEND_MDS, &hdr->flags))
+	if (test_bit(NFS_IOHDR_RESEND_PNFS, &hdr->flags))
+		ff_layout_resend_pnfs_read(hdr);
+	else if (test_bit(NFS_IOHDR_RESEND_MDS, &hdr->flags))
 		ff_layout_reset_read(hdr);
 	pnfs_generic_rw_release(data);
 }
@@ -1604,6 +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;
 }
@@ -1696,15 +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);
 }
 
@@ -1865,7 +1838,7 @@ ff_layout_write_pagelist(struct nfs_pgio_header *hdr, int sync)
 	loff_t offset = hdr->args.offset;
 	int vers;
 	struct nfs_fh *fh;
-	int idx = hdr->pgio_mirror_idx;
+	u32 idx = hdr->pgio_mirror_idx;
 
 	mirror = FF_LAYOUT_COMP(lseg, idx);
 	ds = nfs4_ff_layout_prepare_ds(lseg, mirror, true);
@@ -2000,6 +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)
 {
@@ -2012,6 +2044,34 @@ ff_layout_get_ds_info(struct inode *inode)
 }
 
 static void
+ff_layout_setup_ds_info(struct pnfs_ds_commit_info *fl_cinfo,
+		struct pnfs_layout_segment *lseg)
+{
+	struct nfs4_ff_layout_segment *flseg = FF_LAYOUT_LSEG(lseg);
+	struct inode *inode = lseg->pls_layout->plh_inode;
+	struct pnfs_commit_array *array, *new;
+
+	new = pnfs_alloc_commit_array(flseg->mirror_array_cnt,
+				      nfs_io_gfp_mask());
+	if (new) {
+		spin_lock(&inode->i_lock);
+		array = pnfs_add_commit_array(fl_cinfo, new, lseg);
+		spin_unlock(&inode->i_lock);
+		if (array != new)
+			pnfs_free_commit_array(new);
+	}
+}
+
+static void
+ff_layout_release_ds_info(struct pnfs_ds_commit_info *fl_cinfo,
+		struct inode *inode)
+{
+	spin_lock(&inode->i_lock);
+	pnfs_generic_ds_cinfo_destroy(fl_cinfo);
+	spin_unlock(&inode->i_lock);
+}
+
+static void
 ff_layout_free_deviceid_node(struct nfs4_deviceid_node *d)
 {
 	nfs4_ff_layout_free_deviceid(container_of(d, struct nfs4_ff_layout_ds,
@@ -2171,10 +2231,10 @@ ff_layout_prepare_layoutreturn(struct nfs4_layoutreturn_args *args)
 	struct nfs4_flexfile_layoutreturn_args *ff_args;
 	struct nfs4_flexfile_layout *ff_layout = FF_LAYOUT_FROM_HDR(args->layout);
 
-	ff_args = kmalloc(sizeof(*ff_args), GFP_KERNEL);
+	ff_args = kmalloc(sizeof(*ff_args), nfs_io_gfp_mask());
 	if (!ff_args)
 		goto out_nomem;
-	ff_args->pages[0] = alloc_page(GFP_KERNEL);
+	ff_args->pages[0] = alloc_page(nfs_io_gfp_mask());
 	if (!ff_args->pages[0])
 		goto out_nomem_free;
 
@@ -2184,8 +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;
@@ -2211,8 +2272,8 @@ ff_layout_send_layouterror(struct pnfs_layout_segment *lseg)
 	if (list_empty(&head))
 		return;
 
-	errors = kmalloc_array(NFS42_LAYOUTERROR_MAX,
-			sizeof(*errors), GFP_NOFS);
+	errors = kmalloc_array(NFS42_LAYOUTERROR_MAX, sizeof(*errors),
+			       nfs_io_gfp_mask());
 	if (errors != NULL) {
 		const struct nfs4_ff_layout_ds_err *pos;
 		size_t n = 0;
@@ -2297,7 +2358,6 @@ ff_layout_encode_netaddr(struct xdr_stream *xdr, struct nfs4_pnfs_ds_addr *da)
 	struct sockaddr *sap = (struct sockaddr *)&da->da_addr;
 	char portbuf[RPCBIND_MAXUADDRPLEN];
 	char addrbuf[RPCBIND_MAXUADDRLEN];
-	char *netid;
 	unsigned short port;
 	int len, netid_len;
 	__be32 *p;
@@ -2307,18 +2367,13 @@ ff_layout_encode_netaddr(struct xdr_stream *xdr, struct nfs4_pnfs_ds_addr *da)
 		if (ff_layout_ntop4(sap, addrbuf, sizeof(addrbuf)) == 0)
 			return;
 		port = ntohs(((struct sockaddr_in *)sap)->sin_port);
-		netid = "tcp";
-		netid_len = 3;
 		break;
 	case AF_INET6:
 		if (ff_layout_ntop6_noscopeid(sap, addrbuf, sizeof(addrbuf)) == 0)
 			return;
 		port = ntohs(((struct sockaddr_in6 *)sap)->sin6_port);
-		netid = "tcp6";
-		netid_len = 4;
 		break;
 	default:
-		/* we only support tcp and tcp6 */
 		WARN_ON_ONCE(1);
 		return;
 	}
@@ -2326,8 +2381,9 @@ ff_layout_encode_netaddr(struct xdr_stream *xdr, struct nfs4_pnfs_ds_addr *da)
 	snprintf(portbuf, sizeof(portbuf), ".%u.%u", port >> 8, port & 0xff);
 	len = strlcat(addrbuf, portbuf, sizeof(addrbuf));
 
+	netid_len = strlen(da->da_netid);
 	p = xdr_reserve_space(xdr, 4 + netid_len);
-	xdr_encode_opaque(p, netid, netid_len);
+	xdr_encode_opaque(p, da->da_netid, netid_len);
 
 	p = xdr_reserve_space(xdr, 4 + len);
 	xdr_encode_opaque(p, addrbuf, len);
@@ -2424,7 +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;
@@ -2436,7 +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))
@@ -2468,14 +2526,17 @@ ff_layout_prepare_layoutstats(struct nfs42_layoutstat_args *args)
 	const int dev_count = PNFS_LAYOUTSTATS_MAXDEV;
 
 	/* For now, send at most PNFS_LAYOUTSTATS_MAXDEV statistics */
-	args->devinfo = kmalloc_array(dev_count, sizeof(*args->devinfo), GFP_NOIO);
+	args->devinfo = kmalloc_array(dev_count, sizeof(*args->devinfo),
+				      nfs_io_gfp_mask());
 	if (!args->devinfo)
 		return -ENOMEM;
 
 	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);
@@ -2496,6 +2557,16 @@ ff_layout_set_layoutdriver(struct nfs_server *server,
 	return 0;
 }
 
+static const struct pnfs_commit_ops ff_layout_commit_ops = {
+	.setup_ds_info		= ff_layout_setup_ds_info,
+	.release_ds_info	= ff_layout_release_ds_info,
+	.mark_request_commit	= pnfs_layout_mark_request_commit,
+	.clear_request_commit	= pnfs_generic_clear_request_commit,
+	.scan_commit_lists	= pnfs_generic_scan_commit_lists,
+	.recover_commit_reqs	= pnfs_generic_recover_commit_reqs,
+	.commit_pagelist	= ff_layout_commit_pagelist,
+};
+
 static struct pnfs_layoutdriver_type flexfilelayout_type = {
 	.id			= LAYOUT_FLEX_FILES,
 	.name			= "LAYOUT_FLEX_FILES",
@@ -2512,17 +2583,13 @@ static struct pnfs_layoutdriver_type flexfilelayout_type = {
 	.pg_write_ops		= &ff_layout_pg_write_ops,
 	.get_ds_info		= ff_layout_get_ds_info,
 	.free_deviceid_node	= ff_layout_free_deviceid_node,
-	.mark_request_commit	= pnfs_layout_mark_request_commit,
-	.clear_request_commit	= pnfs_generic_clear_request_commit,
-	.scan_commit_lists	= pnfs_generic_scan_commit_lists,
-	.recover_commit_reqs	= pnfs_generic_recover_commit_reqs,
-	.commit_pagelist	= ff_layout_commit_pagelist,
 	.read_pagelist		= ff_layout_read_pagelist,
 	.write_pagelist		= ff_layout_write_pagelist,
 	.alloc_deviceid_node    = ff_layout_alloc_deviceid_node,
 	.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/flexfilelayout/flexfilelayout.h b/fs/nfs/flexfilelayout/flexfilelayout.h
index 2f369966abf7..354a031c69b1 100644
--- a/fs/nfs/flexfilelayout/flexfilelayout.h
+++ b/fs/nfs/flexfilelayout/flexfilelayout.h
@@ -99,7 +99,7 @@ struct nfs4_ff_layout_segment {
 	u64				stripe_unit;
 	u32				flags;
 	u32				mirror_array_cnt;
-	struct nfs4_ff_layout_mirror	**mirror_array;
+	struct nfs4_ff_layout_mirror	*mirror_array[];
 };
 
 struct nfs4_flexfile_layout {
diff --git a/fs/nfs/flexfilelayout/flexfilelayoutdev.c b/fs/nfs/flexfilelayout/flexfilelayoutdev.c
index 3eda40a320a5..e028f5a0ef5f 100644
--- a/fs/nfs/flexfilelayout/flexfilelayoutdev.c
+++ b/fs/nfs/flexfilelayout/flexfilelayoutdev.c
@@ -69,7 +69,7 @@ nfs4_ff_alloc_deviceid_node(struct nfs_server *server, struct pnfs_device *pdev,
 	INIT_LIST_HEAD(&dsaddrs);
 
 	xdr_init_decode_pages(&stream, &buf, pdev->pages, pdev->pglen);
-	xdr_set_scratch_buffer(&stream, page_address(scratch), PAGE_SIZE);
+	xdr_set_scratch_page(&stream, scratch);
 
 	/* multipath count */
 	p = xdr_inline_decode(&stream, 4);
@@ -113,8 +113,10 @@ nfs4_ff_alloc_deviceid_node(struct nfs_server *server, struct pnfs_device *pdev,
 			goto out_err_drain_dsaddrs;
 		ds_versions[i].version = be32_to_cpup(p++);
 		ds_versions[i].minor_version = be32_to_cpup(p++);
-		ds_versions[i].rsize = nfs_block_size(be32_to_cpup(p++), NULL);
-		ds_versions[i].wsize = nfs_block_size(be32_to_cpup(p++), NULL);
+		ds_versions[i].rsize = nfs_io_size(be32_to_cpup(p++),
+						   server->nfs_client->cl_proto);
+		ds_versions[i].wsize = nfs_io_size(be32_to_cpup(p++),
+						   server->nfs_client->cl_proto);
 		ds_versions[i].tightly_coupled = be32_to_cpup(p);
 
 		if (ds_versions[i].rsize > NFS_MAX_FILE_IO_SIZE)
@@ -378,10 +380,10 @@ nfs4_ff_layout_prepare_ds(struct pnfs_layout_segment *lseg,
 		goto noconnect;
 
 	ds = mirror->mirror_ds->ds;
+	if (READ_ONCE(ds->ds_clp))
+		goto out;
 	/* matching smp_wmb() in _nfs4_pnfs_v3/4_ds_connect */
 	smp_rmb();
-	if (ds->ds_clp)
-		goto out;
 
 	/* FIXME: For now we assume the server sent only one version of NFS
 	 * to use for the DS.