a strange problem with rtems_io_register_name

[Stan]

> >
> >  In console.c, I have removed rtems_fatal_error_occurred called by
> >  rtems_io_register

Sorry, to tell the truth,  I have removed rtems_fatal_error_occurred _and_
rtems_io_register_name.


> >  In bspstart.c Cpu_table.postdriver_hook has been modified ( = NULL) for
to
> >  avoid
> >  tests console ( stdout ).
> >
> >  In test.c ("hello word") , I have added these lines in main task :
> >
> >  ...
> >      status = rtems_io_register_name ("/dev/console", 0,0);
> >          if(status != RTEMS_SUCCESSFUL)
rtems_fatal_error_occured(status);
>
> /dev/console is already registered by the console device driver.  The
> error is indicating there is already a /dev/console.  I am looking at
> line 230 of console/console.c.

In the initial source, is not already registred, but nevertheless it return
fail.


> >      status = rtems_io_register_name ("/dev/aux", 0,1);
> >          if(status != RTEMS_SUCCESSFUL)
rtems_fatal_error_occured(status);
> >  ...
> >
> >  it return SUCCESSFUL.
>
> /dev/aux was never registered before.

In my test, both register_name return successful, because no called from
console_initialize (see above, about console.c).
An idea ?

Stan.