pNFS/flexfiles: don't attempt pnfs on fatal DS errors

When an applications get killed (SIGTERM/SIGINT) while pNFS client performs a connection
to DS, client ends in an infinite loop of connect-disconnect. This
source of the issue, it that flexfilelayoutdev#nfs4_ff_layout_prepare_ds gets an error
on nfs4_pnfs_ds_connect with status ERESTARTSYS, which is set by rpc_signal_task, but
the error is treated as transient, thus retried.

The issue is reproducible with Ctrl+C the following script(there should be ~1000 files in
a directory, client should must not have any connections to DSes):

```
echo 3 > /proc/sys/vm/drop_caches

for i in *
do
   head -1 $i
done
```

The change aims to propagate the nfs4_ff_layout_prepare_ds error state
to the caller that can decide whatever this is a retryable error or not.

Signed-off-by: Tigran Mkrtchyan <tigran.mkrtchyan@desy.de>
Link: https://lore.kernel.org/r/20250627071751.189663-1-tigran.mkrtchyan@desy.de
Fixes: 260f32adb88d ("pNFS/flexfiles: Check the result of nfs4_pnfs_ds_connect")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>

1 files changed, 16 insertions, 10 deletions

diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c
index 4bea008dbebd..8dc921d83538 100644
--- a/fs/nfs/flexfilelayout/flexfilelayout.c
+++ b/fs/nfs/flexfilelayout/flexfilelayout.c
@@ -762,14 +762,14 @@ ff_layout_choose_ds_for_read(struct pnfs_layout_segment *lseg,
 {
 	struct nfs4_ff_layout_segment *fls = FF_LAYOUT_LSEG(lseg);
 	struct nfs4_ff_layout_mirror *mirror;
-	struct nfs4_pnfs_ds *ds;
+	struct nfs4_pnfs_ds *ds = ERR_PTR(-EAGAIN);
 	u32 idx;
 
 	/* mirrors are initially sorted by efficiency */
 	for (idx = start_idx; idx < fls->mirror_array_cnt; idx++) {
 		mirror = FF_LAYOUT_COMP(lseg, idx);
 		ds = nfs4_ff_layout_prepare_ds(lseg, mirror, false);
-		if (!ds)
+		if (IS_ERR(ds))
 			continue;
 
 		if (check_device &&
@@ -777,10 +777,10 @@ ff_layout_choose_ds_for_read(struct pnfs_layout_segment *lseg,
 			continue;
 
 		*best_idx = idx;
-		return ds;
+		break;
 	}
 
-	return NULL;
+	return ds;
 }
 
 static struct nfs4_pnfs_ds *
@@ -942,7 +942,7 @@ retry:
 	for (i = 0; i < pgio->pg_mirror_count; i++) {
 		mirror = FF_LAYOUT_COMP(pgio->pg_lseg, i);
 		ds = nfs4_ff_layout_prepare_ds(pgio->pg_lseg, mirror, true);
-		if (!ds) {
+		if (IS_ERR(ds)) {
 			if (!ff_layout_no_fallback_to_mds(pgio->pg_lseg))
 				goto out_mds;
 			pnfs_generic_pg_cleanup(pgio);
@@ -1867,6 +1867,7 @@ ff_layout_read_pagelist(struct nfs_pgio_header *hdr)
 	u32 idx = hdr->pgio_mirror_idx;
 	int vers;
 	struct nfs_fh *fh;
+	bool ds_fatal_error = false;
 
 	dprintk("--> %s ino %lu pgbase %u req %zu@%llu\n",
 		__func__, hdr->inode->i_ino,
@@ -1874,8 +1875,10 @@ ff_layout_read_pagelist(struct nfs_pgio_header *hdr)
 
 	mirror = FF_LAYOUT_COMP(lseg, idx);
 	ds = nfs4_ff_layout_prepare_ds(lseg, mirror, false);
-	if (!ds)
+	if (IS_ERR(ds)) {
+		ds_fatal_error = nfs_error_is_fatal(PTR_ERR(ds));
 		goto out_failed;
+	}
 
 	ds_clnt = nfs4_ff_find_or_create_ds_client(mirror, ds->ds_clp,
 						   hdr->inode);
@@ -1923,7 +1926,7 @@ ff_layout_read_pagelist(struct nfs_pgio_header *hdr)
 	return PNFS_ATTEMPTED;
 
 out_failed:
-	if (ff_layout_avoid_mds_available_ds(lseg))
+	if (ff_layout_avoid_mds_available_ds(lseg) && !ds_fatal_error)
 		return PNFS_TRY_AGAIN;
 	trace_pnfs_mds_fallback_read_pagelist(hdr->inode,
 			hdr->args.offset, hdr->args.count,
@@ -1945,11 +1948,14 @@ ff_layout_write_pagelist(struct nfs_pgio_header *hdr, int sync)
 	int vers;
 	struct nfs_fh *fh;
 	u32 idx = hdr->pgio_mirror_idx;
+	bool ds_fatal_error = false;
 
 	mirror = FF_LAYOUT_COMP(lseg, idx);
 	ds = nfs4_ff_layout_prepare_ds(lseg, mirror, true);
-	if (!ds)
+	if (IS_ERR(ds)) {
+		ds_fatal_error = nfs_error_is_fatal(PTR_ERR(ds));
 		goto out_failed;
+	}
 
 	ds_clnt = nfs4_ff_find_or_create_ds_client(mirror, ds->ds_clp,
 						   hdr->inode);
@@ -2000,7 +2006,7 @@ ff_layout_write_pagelist(struct nfs_pgio_header *hdr, int sync)
 	return PNFS_ATTEMPTED;
 
 out_failed:
-	if (ff_layout_avoid_mds_available_ds(lseg))
+	if (ff_layout_avoid_mds_available_ds(lseg) && !ds_fatal_error)
 		return PNFS_TRY_AGAIN;
 	trace_pnfs_mds_fallback_write_pagelist(hdr->inode,
 			hdr->args.offset, hdr->args.count,
@@ -2043,7 +2049,7 @@ static int ff_layout_initiate_commit(struct nfs_commit_data *data, int how)
 	idx = calc_ds_index_from_commit(lseg, data->ds_commit_index);
 	mirror = FF_LAYOUT_COMP(lseg, idx);
 	ds = nfs4_ff_layout_prepare_ds(lseg, mirror, true);
-	if (!ds)
+	if (IS_ERR(ds))
 		goto out_err;
 
 	ds_clnt = nfs4_ff_find_or_create_ds_client(mirror, ds->ds_clp,