avoid deadlock in proxy

in secop.proxy the callers modules method announceUpdate is
called from an other thread while the accessLock is locked,
creating a deadlock. solve this by creating an other lock
'updateLock' for the update.

+ add status parameter even to non-Readable proxy modules,
  in order to indicate a failed connection
+ fix an error in secop_psi/softcal.py

Change-Id: Iae7c6d5a74001150a47aa9dc99209c15d972cd5e
Reviewed-on: https://forge.frm2.tum.de/review/c/sine2020/secop/playground/+/28130
Tested-by: Jenkins Automated Tests <pedersen+jenkins@frm2.tum.de>
Reviewed-by: Enrico Faulhaber <enrico.faulhaber@frm2.tum.de>
Reviewed-by: Markus Zolliker <markus.zolliker@psi.ch>

secop/proxy.py

@@ -33,6 +33,7 @@ from secop.io import HasIO
 
 class ProxyModule(HasIO, Module):
     module = Property('remote module name', datatype=StringType(), default='')
+    status = Parameter('connection status', Readable.status.datatype)  # add status even when not a Readable
 
     _consistency_check_done = False
     _secnode = None
@@ -182,11 +183,12 @@ def proxy_class(remote_class, name=None):
 
     for aname, aobj in rcls.accessibles.items():
         if isinstance(aobj, Parameter):
-            pobj = aobj.merge(dict(handler=None, needscfg=False))
+            pobj = aobj.copy()
+            pobj.merge(dict(handler=None, needscfg=False))
             attrs[aname] = pobj
 
             def rfunc(self, pname=aname):
-                value, _, readerror = self._secnode.getParameter(self.name, pname)
+                value, _, readerror = self._secnode.getParameter(self.name, pname, True)
                 if readerror:
                     raise readerror
                 return value