diff --git a/dlls/kernel32/tests/heap.c b/dlls/kernel32/tests/heap.c
index 81850fef6a3..0c0de113cde 100644
--- a/dlls/kernel32/tests/heap.c
+++ b/dlls/kernel32/tests/heap.c
@@ -488,7 +488,7 @@ static void test_heap_checks( DWORD flags )
 {
     BYTE old, *p, *p2;
     BOOL ret;
-    SIZE_T size, large_size = 800 * 1024 + 37;
+    SIZE_T i, size, large_size = 800 * 1024 + 37;
 
     if (flags & HEAP_PAGE_ALLOCS) return;  /* no tests for that case yet */
     trace( "testing heap flags %08x\n", flags );
@@ -640,6 +640,18 @@ static void test_heap_checks( DWORD flags )
 
     ret = HeapFree( GetProcessHeap(), 0, p );
     ok( ret, "HeapFree failed\n" );
+
+    /* test block sizes when tail checking */
+    if (flags & HEAP_TAIL_CHECKING_ENABLED)
+    {
+        for (size = 0; size < 64; size++)
+        {
+            p = HeapAlloc( GetProcessHeap(), 0, size );
+            for (i = 0; i < 32; i++) if (p[size + i] != 0xab) break;
+            ok( i >= 8, "only %lu tail bytes for size %lu\n", i, size );
+            HeapFree( GetProcessHeap(), 0, p );
+        }
+    }
 }
 
 static void test_debug_heap( const char *argv0, DWORD flags )
diff --git a/dlls/ntdll/heap.c b/dlls/ntdll/heap.c
index 29e9559332e..c3fdf2c5d0b 100644
--- a/dlls/ntdll/heap.c
+++ b/dlls/ntdll/heap.c
@@ -106,6 +106,8 @@ C_ASSERT( sizeof(ARENA_LARGE) % LARGE_ALIGNMENT == 0 );
 #define HEAP_MIN_SHRINK_SIZE  (HEAP_MIN_DATA_SIZE+sizeof(ARENA_FREE))
 /* minimum size to start allocating large blocks */
 #define HEAP_MIN_LARGE_BLOCK_SIZE  0x7f000
+/* extra size to add at the end of block for tail checking */
+#define HEAP_TAIL_EXTRA_SIZE(flags)  (flags & HEAP_TAIL_CHECKING_ENABLED ? 8 : 0)
 
 /* Max size of the blocks on the free lists */
 static const SIZE_T HEAP_freeListSizes[] =
@@ -677,7 +679,7 @@ static void HEAP_ShrinkBlock(SUBHEAP *subheap, ARENA_INUSE *pArena, SIZE_T size)
 static void *allocate_large_block( HEAP *heap, DWORD flags, SIZE_T size )
 {
     ARENA_LARGE *arena;
-    SIZE_T block_size = sizeof(*arena) + ROUND_SIZE(size);
+    SIZE_T block_size = sizeof(*arena) + ROUND_SIZE(size) + HEAP_TAIL_EXTRA_SIZE(flags);
     LPVOID address = NULL;
 
     if (block_size < size) return NULL;  /* overflow */
@@ -1581,7 +1583,7 @@ PVOID WINAPI RtlAllocateHeap( HANDLE heap, ULONG flags, SIZE_T size )
     if (!heapPtr) return NULL;
     flags &= HEAP_GENERATE_EXCEPTIONS | HEAP_NO_SERIALIZE | HEAP_ZERO_MEMORY;
     flags |= heapPtr->flags;
-    rounded_size = ROUND_SIZE(size);
+    rounded_size = ROUND_SIZE(size) + HEAP_TAIL_EXTRA_SIZE( flags );
     if (rounded_size < size)  /* overflow */
     {
         if (flags & HEAP_GENERATE_EXCEPTIONS) RtlRaiseStatus( STATUS_NO_MEMORY );
@@ -1735,7 +1737,7 @@ PVOID WINAPI RtlReAllocateHeap( HANDLE heap, ULONG flags, PVOID ptr, SIZE_T size
     flags |= heapPtr->flags;
     if (!(flags & HEAP_NO_SERIALIZE)) RtlEnterCriticalSection( &heapPtr->critSection );
 
-    rounded_size = ROUND_SIZE(size);
+    rounded_size = ROUND_SIZE(size) + HEAP_TAIL_EXTRA_SIZE(flags);
     if (rounded_size < size) goto oom;  /* overflow */
     if (rounded_size < HEAP_MIN_DATA_SIZE) rounded_size = HEAP_MIN_DATA_SIZE;