NFSv4: Don't discard segments marked for return in _pnfs_return_layout()
authorTrond Myklebust <trond.myklebust@hammerspace.com>
Sun, 18 Apr 2021 19:00:45 +0000 (15:00 -0400)
committerTrond Myklebust <trond.myklebust@hammerspace.com>
Sun, 18 Apr 2021 19:10:53 +0000 (15:10 -0400)
If the pNFS layout segment is marked with the NFS_LSEG_LAYOUTRETURN
flag, then the assumption is that it has some reporting requirement
to perform through a layoutreturn (e.g. flexfiles layout stats or error
information).

Fixes: 6d597e175012 ("pnfs: only tear down lsegs that precede seqid in LAYOUTRETURN args")
Cc: stable@vger.kernel.org
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
fs/nfs/pnfs.c

index 33574f47601f27db99fc5ff7f2fa19d48b9806ef..f726f8b12b7e810a645f7ef24f93f0f5d57810e1 100644 (file)
@@ -1344,7 +1344,7 @@ _pnfs_return_layout(struct inode *ino)
        }
        valid_layout = pnfs_layout_is_valid(lo);
        pnfs_clear_layoutcommit(ino, &tmp_list);
-       pnfs_mark_matching_lsegs_invalid(lo, &tmp_list, NULL, 0);
+       pnfs_mark_matching_lsegs_return(lo, &tmp_list, NULL, 0);
 
        if (NFS_SERVER(ino)->pnfs_curr_ld->return_range) {
                struct pnfs_layout_range range = {