multiple mptest failing

[Joel Sherrill]

configure with --enable-multiprocessing --enable-tests

mp03-node1 - _Thread_Dispatch_disable_level is (1) not 0
mp03-node2 - same
mp08-node2 - rtems_semaphore_obtain FAILED -- expected (object deleted while waiting) got (invalid object id)
mp10-node2 - rtems_semaphore_obtain FAILED -- expected (successful completion) got (invalid object id)
mp13-node2 - rtems_semaphore_obtain FAILED -- expected (successful completion) got (invalid object id)
mp14-node1 - _Thread_Dispatch_disable_level is (2) not 0
mp14-node2 - _Thread_Dispatch_disable_level is (1) not 0

I suspect the ID problems indicate tests not updated to
reflect new lower defaults.

I suspect the dispatch errors are test specific problems
where a check is made inside a TSR or ISR and the test is
not checking for the correct expected level.

Release:
RTEMS CVS Trunk

Environment:
Redhat Linux native

[Joel Sherrill]

State-Changed-From-To: working->closed
State-Changed-Why: Attached patch fixes test problems that were independent

of object get (id translation) problems fixed as PR36.
This was tested on the trunk and 4.5 branch using the
Linux (POSIX) port for MP and the SPARC/ERC32 for a single
processor configuration.

[Joel Sherrill]

mptest-CVS.patch