NFS4: Fix use-after-free in trace_event_raw_event_nfs4_set_lock

stable inclusion
from stable-5.10.9
commit 1b42712e43e5d574a985661e336e48b0f6ba06b6
bugzilla: 47457

--------------------------------

commit 3d1a90ab upstream.

It is only safe to call the tracepoint before rpc_put_task() because
'data' is freed inside nfs4_lock_release (rpc_release).

Fixes: 48c9579a ("Adding stateid information to tracepoints")
Signed-off-by: NDave Wysochanski <dwysocha@redhat.com>
Signed-off-by: NTrond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: NChen Jun <chenjun102@huawei.com>
Acked-by: NXie XiuQi <xiexiuqi@huawei.com>

fs/nfs/nfs4proc.c

@@ -7106,9 +7106,9 @@ static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *f
 					data->arg.new_lock_owner, ret);
 	} else
 		data->cancelled = true;
+	trace_nfs4_set_lock(fl, state, &data->res.stateid, cmd, ret);
 	rpc_put_task(task);
 	dprintk("%s: done, ret = %d!\n", __func__, ret);
-	trace_nfs4_set_lock(fl, state, &data->res.stateid, cmd, ret);
 	return ret;
 }