[OpenMP][Offloading] Fix infinite loop in applyToShadowMapEntries

[tianshilei1992]

Summary

This patch fixes the issue that the for loop in applyToShadowMapEntries
is infinite because Itr is not incremented in CB. Fixes #53727.

[jdoerfert]

LG, I broke this. Can we also have tests?

[tianshilei1992]

rebase and add test

openmp/libomptarget/src/omptarget.cpp

if (ArgType & OMP_TGT_MAPTYPE_FROM) {
     }
 
     auto CB = [&](ShadowPtrListTy::iterator &Itr) {
       void **ShadowHstPtrAddr = (void **)Itr->first;
       *ShadowHstPtrAddr = Itr->second.HstPtrVal;
       DP("Restoring original host pointer value " DPxMOD
          " for host pointer " DPxMOD "\n",
          DPxPTR(Itr->second.HstPtrVal), DPxPTR(ShadowHstPtrAddr));
+      ++Itr;
       return OFFLOAD_SUCCESS;
     };
     applyToShadowMapEntries(Device, CB, HstPtrBegin, ArgSize, TPR);
   }
 
   if (ArgType & OMP_TGT_MAPTYPE_TO) {
     DP("Moving %" PRId64 " bytes (hst:" DPxMOD ") -> (tgt:" DPxMOD ")\n",
        ArgSize, DPxPTR(HstPtrBegin), DPxPTR(TgtPtrBegin));
     int Ret = Device.submitData(TgtPtrBegin, HstPtrBegin, ArgSize, AsyncInfo);
     if (Ret != OFFLOAD_SUCCESS) {
       REPORT("Copying data to device failed.\n");
       return OFFLOAD_FAIL;
     }
 
     auto CB = [&](ShadowPtrListTy::iterator &Itr) {
       DP("Restoring original target pointer value " DPxMOD " for target "
          "pointer " DPxMOD "\n",
          DPxPTR(Itr->second.TgtPtrVal), DPxPTR(Itr->second.TgtPtrAddr));
       Ret = Device.submitData(Itr->second.TgtPtrAddr, &Itr->second.TgtPtrVal,
                               sizeof(void *), AsyncInfo);
       if (Ret != OFFLOAD_SUCCESS)
         REPORT("Copying data to device failed.\n");
+      ++Itr;
       return Ret;
     };
     applyToShadowMapEntries(Device, CB, HstPtrBegin, ArgSize, TPR);
   }
   return OFFLOAD_SUCCESS;
 }
 
 static int targetDataNonContiguous(ident_t *loc, DeviceTy &Device,

openmp/libomptarget/test/offloading/bug53727.cpp

+// RUN: %libomptarget-compilexx-and-run-generic
+
+#include <cassert>
+#include <iostream>
+
+constexpr const int N = 10;
+
+struct T {
+  int a;
+  int *p;
+};
+
+struct S {
+  int b;
+  T t;
+};
+
+int main(int argc, char *argv[]) {
+  S s;
+  s.t.p = new int[N];
+  for (int i = 0; i < N; ++i) {
+    s.t.p[i] = i;
+  }
+
+#pragma omp target enter data map(to : s, s.t.p[:N])
+
+#pragma omp target
+  {
+    for (int i = 0; i < N; ++i) {
+      s.t.p[i] += i;
+    }
+  }
+
+#pragma omp target update from(s.t.p[:N])
+
+  for (int i = 0; i < N; ++i) {
+    assert(s.t.p[i] == 2 * i);
+    s.t.p[i] += i;
+  }
+
+#pragma omp target update to(s.t.p[:N])
+
+#pragma omp target
+  {
+    for (int i = 0; i < N; ++i) {
+      s.t.p[i] += i;
+    }
+  }
+
+#pragma omp target exit data map(from : s, s.t.p[:N])
+
+  for (int i = 0; i < N; ++i) {
+    assert(s.t.p[i] == 4 * i);
+  }
+
+  return 0;
+}