[aarch64-port-dev ] RFR (XS) 8229123: Revert build fixes for aarch64/zero

[Aleksey Shipilev]

RFE:
  https://bugs.openjdk.java.net/browse/JDK-8229123

Another thing found during 8u-aarch64 webrev scrubbing. Not sure why this thing is needed. Reverting
it does not break aarch64/zero cross-compiled build. There is no change like this in any of the
upstreams either. Maybe Edward remembers why it is there?

8u-aarch64 fix:

diff -r 732ebff024df src/share/vm/interpreter/interpreterRuntime.cpp
--- a/src/share/vm/interpreter/interpreterRuntime.cpp   Mon Aug 05 09:42:26 2019 +0200
+++ b/src/share/vm/interpreter/interpreterRuntime.cpp   Mon Aug 05 10:51:59 2019 +0200
@@ -396,37 +396,35 @@
   int                handler_bci;
   int                current_bci = bci(thread);

-#ifndef CC_INTERP
   if (thread->frames_to_pop_failed_realloc() > 0) {
     // Allocation of scalar replaced object used in this frame
     // failed. Unconditionally pop the frame.
     thread->dec_frames_to_pop_failed_realloc();
     thread->set_vm_result(h_exception());
     // If the method is synchronized we already unlocked the monitor
     // during deoptimization so the interpreter needs to skip it when
     // the frame is popped.
     thread->set_do_not_unlock_if_synchronized(true);
 #ifdef CC_INTERP
     return (address) -1;
 #else
     return Interpreter::remove_activation_entry();
 #endif
   }
-#endif

   // Need to do this check first since when _do_not_unlock_if_synchronized
   // is set, we don't want to trigger any classloading which may make calls

Testing: aarch64 {server,zero} builds; ad-hoc testing zero build

-- 
Thanks,
-Aleksey