Use the core build, where the problem is apparently easiest to reproduce.

In the two cases seen so far -- the VM thread is waiting for several (in one
case 2, in another case 18) java threads to suspend at safepoints. The
state of the target java threads was not determined.

The problem has not, thus far, been seen as frequently with the -Xconcurrentio
option, so it's possible that one of the other options in -Xconcurrentio
is masking the problem with use of UseLWPSynchronization.

The problem has thus far been seen only on SPARC, but it may (or may not)
exist on Intel Solaris.