[OpenMP][libomptarget] Fix union.

[gtbercea]

Summary

To make the two parts of the union have the same size, the size of vect needs to be increased by 16 bits.

[gtbercea]

Update.

[ABataev]

libomptarget/deviceRTLs/nvptx/src/omptarget-nvptx.h

163

It is better to remove this union at all and use memcpy

[gtbercea]

Eliminate union as suggested in the comment.

[ABataev]

LG

[gtbercea]

Use assignment.

[grokos]

I think the assignment makes more sense.

libomptarget/deviceRTLs/nvptx/src/omptarget-nvptx.h

private:
   //   1 bit (InParL2+) if this thread has encountered L2 or higher parallel
   //   region
   static const uint8_t TaskDescr_SchedMask = (0x1 | 0x2 | 0x4);
   static const uint8_t TaskDescr_IsDynamic = 0x8;
   static const uint8_t TaskDescr_InPar = 0x10;
   static const uint8_t TaskDescr_IsParConstr = 0x20;
   static const uint8_t TaskDescr_InParL2P = 0x40;
 
   union { // both have same size
-    uint64_t vect[2];
+    uint16_t vect[9];
     struct TaskDescr_items {
       uint8_t flags; // 6 bit used (see flag above)
       uint8_t unused;
       uint16_t nthreads;         // thread num for subsequent parallel regions
       uint16_t threadlimit;      // thread limit ICV
       uint16_t threadId;         // thread id
       uint16_t threadsInTeam;    // threads in current team
       uint64_t runtimeChunkSize; // runtime chunk size

libomptarget/deviceRTLs/nvptx/src/omptarget-nvptxi.h

data.items.threadId =
       GetThreadIdInBlock(); // get ids from cuda (only called for 1st level)
   data.items.threadsInTeam = tnum;
   data.items.runtimeChunkSize = 1; // prefered chunking statik with chunk 1
   prev = parentTaskDescr;
 }
 
 INLINE void omptarget_nvptx_TaskDescr::CopyData(
     omptarget_nvptx_TaskDescr *sourceTaskDescr) {
-  data.vect[0] = sourceTaskDescr->data.vect[0];
-  data.vect[1] = sourceTaskDescr->data.vect[1];
+  for (int i=0; i<9; i++)
+    data.vect[i] = sourceTaskDescr->data.vect[i];
 }
 
 INLINE void
 omptarget_nvptx_TaskDescr::Copy(omptarget_nvptx_TaskDescr *sourceTaskDescr) {
   CopyData(sourceTaskDescr);
   prev = sourceTaskDescr->prev;
 }