Index: lib/sanitizer_common/sanitizer_allocator.cc
===================================================================
--- lib/sanitizer_common/sanitizer_allocator.cc
+++ lib/sanitizer_common/sanitizer_allocator.cc
@@ -180,6 +180,25 @@
 // LowLevelAllocator
 static LowLevelAllocateCallback low_level_alloc_callback;
 
+int InternalReallocArr(void *addr, uptr count, uptr size,
+                       InternalAllocatorCache *cache) {
+  void *oaddr;
+  void *naddr = 0;
+
+  internal_memcpy(&oaddr, addr, sizeof(addr));
+  if (!count || !size) {
+    InternalFree(oaddr, cache);
+    internal_memcpy(addr, &naddr, sizeof(addr));
+    return 0;
+  }
+
+  naddr = InternalRealloc(oaddr, count * size, cache);
+  if (!naddr) return -1;
+
+  internal_memcpy(addr, &naddr, sizeof(addr));
+  return 0;
+}
+
 void *LowLevelAllocator::Allocate(uptr size) {
   // Align allocation size.
   size = RoundUpTo(size, 8);
Index: lib/sanitizer_common/sanitizer_allocator_internal.h
===================================================================
--- lib/sanitizer_common/sanitizer_allocator_internal.h
+++ lib/sanitizer_common/sanitizer_allocator_internal.h
@@ -57,6 +57,8 @@
 void *InternalCalloc(uptr countr, uptr size,
                      InternalAllocatorCache *cache = nullptr);
 void InternalFree(void *p, InternalAllocatorCache *cache = nullptr);
+int InternalReallocArr(void *p, uptr count, uptr size,
+                       InternalAllocatorCache *cache = nullptr);
 InternalAllocator *internal_allocator();
 
 enum InternalAllocEnum {
Index: lib/tsan/rtl/tsan_interceptors.cc
===================================================================
--- lib/tsan/rtl/tsan_interceptors.cc
+++ lib/tsan/rtl/tsan_interceptors.cc
@@ -391,10 +391,26 @@
   void *arg;
 };
 
-static void at_exit_wrapper(void *arg) {
-  ThreadState *thr = cur_thread();
-  uptr pc = 0;
-  Acquire(thr, pc, (uptr)arg);
+static struct AtExitCtx **AtExitStack;
+static int AtExitStackSize;
+
+static void at_exit_wrapper() {
+  // Pop AtExitCtx from the top of the stack of callback functions
+  AtExitCtx *ctx = AtExitStack[AtExitStackSize - 1];
+  --AtExitStackSize;
+  if (InternalReallocArr(&AtExitStack, AtExitStackSize,
+                         sizeof(struct AtExitCtx *))) {
+    Printf("InternalReallocArr error");
+    Die();
+  }
+
+  Acquire(cur_thread(), (uptr)0, (uptr)ctx);
+  ((void(*)())ctx->f)();
+  InternalFree(ctx);
+}
+
+static void cxa_at_exit_wrapper(void *arg) {
+  Acquire(cur_thread(), 0, (uptr)arg);
   AtExitCtx *ctx = (AtExitCtx*)arg;
   ((void(*)(void *arg))ctx->f)(ctx->arg);
   InternalFree(ctx);
@@ -430,7 +446,24 @@
   // Memory allocation in __cxa_atexit will race with free during exit,
   // because we do not see synchronization around atexit callback list.
   ThreadIgnoreBegin(thr, pc);
-  int res = REAL(__cxa_atexit)(at_exit_wrapper, ctx, dso);
+  int res;
+  if (!dso) {
+    // NetBSD does not preserve the 2nd argument if dso is equal to 0
+    // Store ctx in a local stack-like structure
+    res = REAL(__cxa_atexit)((void (*)(void *a))at_exit_wrapper, 0, 0);
+    // Push AtExitCtx on the top of the stack of callback functions
+    if (!res) {
+      ++AtExitStackSize;
+      if (InternalReallocArr(&AtExitStack, AtExitStackSize,
+                             sizeof(struct AtExitCtx *))) {
+        Printf("InternalReallocArr error");
+        Die();
+      }
+      AtExitStack[AtExitStackSize - 1] = ctx;
+    }
+  } else {
+    res = REAL(__cxa_atexit)(cxa_at_exit_wrapper, ctx, dso);
+  }
   ThreadIgnoreEnd(thr, pc);
   return res;
 }
Index: test/tsan/atexit3.cc
===================================================================
--- /dev/null
+++ test/tsan/atexit3.cc
@@ -0,0 +1,41 @@
+// RUN: %clang_tsan -O1 %s -o %t && %run %t 2>&1 | FileCheck %s
+
+#include <stdio.h>
+#include <stdlib.h>
+
+static void atexit5() {
+  fprintf(stderr, "5");
+}
+
+static void atexit4() {
+  fprintf(stderr, "4");
+}
+
+static void atexit3() {
+  fprintf(stderr, "3");
+}
+
+static void atexit2() {
+  fprintf(stderr, "2");
+}
+
+static void atexit1() {
+  fprintf(stderr, "1");
+}
+
+static void atexit0() {
+  fprintf(stderr, "\n");
+}
+
+int main() {
+  atexit(atexit0);
+  atexit(atexit1);
+  atexit(atexit2);
+  atexit(atexit3);
+  atexit(atexit4);
+  atexit(atexit5);
+}
+
+// CHECK-NOT: FATAL: ThreadSanitizer
+// CHECK-NOT: WARNING: ThreadSanitizer
+// CHECK: 54321