3.16.42-rc1 review patch. If anyone has any objections, please let me know.
------------------
From: NeilBrown <neilb@xxxxxxxx>
commit cfd278c280f997cf2fe4662e0acab0fe465f637b upstream.
Various places assume that if nfs4_fl_prepare_ds() turns a non-NULL 'ds',
then ds->ds_clp will also be non-NULL.
This is not necessasrily true in the case when the process received a fatal signal
while nfs4_pnfs_ds_connect is waiting in nfs4_wait_ds_connect().
In that case ->ds_clp may not be set, and the devid may not recently have been marked
unavailable.
So add a test for ds_clp == NULL and return NULL in that case.
Fixes: c23266d532b4 ("NFS4.1 Fix data server connection race")
Signed-off-by: NeilBrown <neilb@xxxxxxxx>
Acked-by: Olga Kornievskaia <aglo@xxxxxxxxx>
Acked-by: Adamson, Andy <William.Adamson@xxxxxxxxxx>
Signed-off-by: Trond Myklebust <trond.myklebust@xxxxxxxxxxxxxxx>
Signed-off-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx>
---
fs/nfs/filelayout/filelayoutdev.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
--- a/fs/nfs/filelayout/filelayoutdev.c
+++ b/fs/nfs/filelayout/filelayoutdev.c
@@ -827,7 +827,8 @@ nfs4_fl_prepare_ds(struct pnfs_layout_se
nfs4_wait_ds_connect(ds);
}
out_test_devid:
- if (filelayout_test_devid_unavailable(devid))
+ if (ret->ds_clp == NULL ||
+ filelayout_test_devid_unavailable(devid))
ret = NULL;
out:
return ret;