NFSv4: Fix a tracepoint Oops in initiate_file_draining()
authorTrond Myklebust <[email protected]>
Wed, 5 Sep 2018 18:07:13 +0000 (14:07 -0400)
committerAnna Schumaker <[email protected]>
Fri, 14 Sep 2018 20:24:08 +0000 (16:24 -0400)
Now that the value of 'ino' can be NULL or an ERR_PTR(), we need to
change the test in the tracepoint.

Fixes: ce5624f7e6675 ("NFSv4: Return NFS4ERR_DELAY when a layout fails...")
Signed-off-by: Trond Myklebust <[email protected]>
Cc: [email protected] # v4.17+
Signed-off-by: Anna Schumaker <[email protected]>
fs/nfs/nfs4trace.h

index 708342f4692f85e359f15bb70b4b7fe859f9b1fd..b1483b303e0bf35a268bccad1823fd6b89015e97 100644 (file)
@@ -1137,7 +1137,7 @@ DECLARE_EVENT_CLASS(nfs4_inode_callback_event,
                TP_fast_assign(
                        __entry->error = error;
                        __entry->fhandle = nfs_fhandle_hash(fhandle);
-                       if (inode != NULL) {
+                       if (!IS_ERR_OR_NULL(inode)) {
                                __entry->fileid = NFS_FILEID(inode);
                                __entry->dev = inode->i_sb->s_dev;
                        } else {