On Tue, 2005-05-17 at 17:18, Libor Michalek wrote: > > (When I ran the server also), I ran it server/client on same machine. > > That is what's exposing the problem. Since the test is using a fake QPN, > a check in the kernel CM see's it as a collision. This patch fixes the issue. Yup, that fixed it. Thanks. -- Hal