[OpenMP][FIX] Query proper thread ID information to support nesting

[jdoerfert]

Summary

The OpenMP thread ID is not the hardware thread ID if we have nesting.
We need to ask the runtime properly to ensure correct results.

Note that the loop interface is going to change soon so we do not adjust
it now but simply ignore the extra argument.

[tianshilei1992]

It's too bad that front end emits "internal" function call instead of interface.

[jdoerfert]

In D111950#3071578, @tianshilei1992 wrote:

It's too bad that front end emits "internal" function call instead of interface.

We'll rip that part out soon.

openmp/libomptarget/DeviceRTL/src/Workshare.cpp

static void ForStaticNoChunk(int &last, T &lb, T &ub, ST &stride, ST &chunk,
     ub = lb + chunk - 1; // Clang uses i <= ub
     last = lb <= inputUb && inputUb <= ub;
     stride = loopSize; // make sure we only do 1 chunk per warp
   }
 
   ////////////////////////////////////////////////////////////////////////////////
   // Support for Static Init
 
-  static void for_static_init(int32_t gtid, int32_t schedtype,
+  static void for_static_init(int32_t, int32_t schedtype,
                               int32_t *plastiter, T *plower, T *pupper,
                               ST *pstride, ST chunk, bool IsSPMDExecutionMode) {
-    // When IsRuntimeUninitialized is true, we assume that the caller is
-    // in an L0 parallel region and that all worker threads participate.
-
-    // Assume we are in teams region or that we use a single block
-    // per target region
+    int32_t gtid = omp_get_thread_num();
     int numberOfActiveOMPThreads = omp_get_num_threads();
 
     // All warps that are in excess of the maximum requested, do
     // not execute the loop
     ASSERT0(LT_FUSSY, gtid < numberOfActiveOMPThreads,
             "current thread is not needed here; error");
 
     // copy