NT_STATUS_ADDRESS_ALREADY_ASSOCIATED

Andrew Bartlett abartlet at samba.org
Fri Jan 4 13:49:05 MST 2013


On Fri, 2013-01-04 at 13:42 +0100, Stefan (metze) Metzmacher wrote:
> Am 04.01.2013 13:29, schrieb Stefan (metze) Metzmacher:
> > Hi,
> > 
> >> Thanks for the analysis Günter.
> >>
> >> We simply need to terminate this array with something other than
> >> cluster_id(0, 0).  The line you indicate is always true because the
> >> array in this case contains both the real 0,0 entry and the terminator
> >> as a 0,0 entry.
> >>
> >> The easy option is to make irpc_servers_byname return an array of
> >> pointers to struct server_sid, so we can use a NULL terminator.  It
> >> isn't used in many places, so this shouldn't be too hard to fix up. 
> > 
> > I think the attached patch should fix this.
> 
> I've fixed a logic error in pymessaging.c...

You also need to fix the line kukks indicated in
irpc_binding_handle_by_name(), so it also checks for
server_id_disconnected() as the terminator.

The:
if (sids[0].pid == 0) {

Is checking if it got back a list with 0 items in it (eg, only the
terminator). 

ridalloc_poke_rid_manager() could probably also do with being made
consistent.

Anyway, thanks for suggesting a way out of this, as it seems to be
causing far more issues than I ever imagined.

Andrew Bartlett

-- 
Andrew Bartlett                                http://samba.org/~abartlet/
Authentication Developer, Samba Team           http://samba.org




More information about the samba-technical mailing list