Don't include mem init loading code when it is inside the wasm anyhow (emscripten-core#5844)

* don't include mem init loading code when it is inside the wasm anyhow

* don't assign null to memoryInitializer later, that is the initial value anyhow

* don't put the mem init in the wasm if using pthreads, as we need to handle it more manually (pthreads don't need to initialize memory, just code)

* add a test from juj@a9b66de

Author: kripken

emcc.py

@@ -1176,6 +1176,11 @@ def check(input_file):
           if shared.Settings.BINARYEN_PASSES:
             shared.Settings.BINARYEN_PASSES += ','
           shared.Settings.BINARYEN_PASSES += 'safe-heap'
+        # we will include the mem init data in the wasm, when we don't need the
+        # mem init file to be loadable by itself
+        shared.Settings.MEM_INIT_IN_WASM = 'asmjs' not in shared.Settings.BINARYEN_METHOD and \
+                                           'interpret-asm2wasm' not in shared.Settings.BINARYEN_METHOD and \
+                                           not shared.Settings.USE_PTHREADS
 
       # wasm outputs are only possible with a side wasm
       if target.endswith(WASM_ENDINGS):
@@ -1709,7 +1714,7 @@ def repl(m):
           if not shared.Settings.BINARYEN or 'asmjs' in shared.Settings.BINARYEN_METHOD or 'interpret-asm2wasm' in shared.Settings.BINARYEN_METHOD:
             return 'memoryInitializer = "%s";' % shared.JS.get_subresource_location(memfile, embed_memfile(options))
           else:
-            return 'memoryInitializer = null;'
+            return ''
         src = re.sub(shared.JS.memory_initializer_pattern, repl, open(final).read(), count=1)
         open(final + '.mem.js', 'w').write(src)
         final += '.mem.js'
@@ -2299,8 +2304,7 @@ def do_binaryen(target, asm_target, options, memfile, wasm_binary_target,
       cmd.append(shared.Building.opt_level_to_str(options.opt_level, options.shrink_level))
     # import mem init file if it exists, and if we will not be using asm.js as a binaryen method (as it needs the mem init file, of course)
     mem_file_exists = options.memory_init_file and os.path.exists(memfile)
-    ok_binaryen_method = 'asmjs' not in shared.Settings.BINARYEN_METHOD and 'interpret-asm2wasm' not in shared.Settings.BINARYEN_METHOD
-    import_mem_init = mem_file_exists and ok_binaryen_method
+    import_mem_init = mem_file_exists and shared.Settings.MEM_INIT_IN_WASM
     if import_mem_init:
       cmd += ['--mem-init=' + memfile]
       if not shared.Settings.RELOCATABLE:

src/postamble.js

@@ -6,6 +6,7 @@ Module['asm'] = asm;
 {{{ maybeExport('FS') }}}
 {{{ maybeExport('GL') }}}
 
+#if MEM_INIT_IN_WASM == 0
 #if MEM_INIT_METHOD == 2
 #if USE_PTHREADS
 if (memoryInitializer && !ENVIRONMENT_IS_PTHREAD) (function(s) {
@@ -110,6 +111,7 @@ if (memoryInitializer) {
   }
 }
 #endif
+#endif // MEM_INIT_IN_WASM == 0
 
 #if CYBERDWARF
   Module['cyberdwarf'] = _cyberdwarf_Debugger(cyberDWARFFile);

src/preamble.js

@@ -2106,10 +2106,12 @@ function integrateWasmJS() {
     var oldView = new Int8Array(oldBuffer);
     var newView = new Int8Array(newBuffer);
 
+#if MEM_INIT_IN_WASM == 0
     // If we have a mem init file, do not trample it
     if (!memoryInitializer) {
       oldView.set(newView.subarray(Module['STATIC_BASE'], Module['STATIC_BASE'] + Module['STATIC_BUMP']), Module['STATIC_BASE']);
     }
+#endif
 
     newView.set(oldView);
     updateGlobalBuffer(newBuffer);

src/settings.js

@@ -861,5 +861,7 @@ var WASM_BINARY_FILE = ''; // name of the file containing wasm binary, if releva
 var ASMJS_CODE_FILE = ''; // name of the file containing asm.js, if relevant
 var SOURCE_MAP_BASE = ''; // Base URL the source mapfile, if relevant
 
+var MEM_INIT_IN_WASM = 0; // for internal use only
+
 var SUPPORT_BASE64_EMBEDDING = 0; // If set to 1, src/base64Utils.js will be included in the bundle.
                                   // This is set internally when needed (SINGLE_FILE)

tests/pthread/test_pthread_global_data_initialization.c

@@ -0,0 +1,36 @@
+#include <assert.h>
+#include <pthread.h>
+#include <emscripten.h>
+
+int globalData = 1;
+
+void *thread_main(void *arg)
+{
+  EM_ASM(Module.print('hello from pthread 1: ' + $0), globalData);
+  assert(globalData == 10);
+
+  globalData = 20;
+  EM_ASM(Module.print('hello from pthread 2: ' + $0), globalData);
+  assert(globalData == 20);
+	return 0;
+}
+
+int main()
+{
+  EM_ASM(Module.print('hello from main 1: ' + $0), globalData);
+  assert(globalData == 1);
+
+  globalData = 10;
+  EM_ASM(Module.print('hello from main 2: ' + $0), globalData);
+  assert(globalData == 10);
+
+	pthread_t thread;
+	pthread_create(&thread, NULL, thread_main, NULL);
+  pthread_join(thread, 0);
+
+  EM_ASM(Module.print('hello from main 3: ' + $0), globalData);
+  assert(globalData == 20);
+#ifdef REPORT_RESULT
+  REPORT_RESULT(globalData);
+#endif
+}

tests/test_browser.py

@@ -3329,6 +3329,12 @@ def test_pthread_call_async_on_main_thread(self):
     self.btest(path_from_root('tests', 'pthread', 'call_async_on_main_thread.c'), expected='7', args=['-O3', '-s', 'USE_PTHREADS=1', '-DPROXY_TO_PTHREAD=0', '--js-library', path_from_root('tests', 'pthread', 'call_async_on_main_thread.js')])
     self.btest(path_from_root('tests', 'pthread', 'call_async_on_main_thread.c'), expected='7', args=['-Oz', '-DPROXY_TO_PTHREAD=0', '--js-library', path_from_root('tests', 'pthread', 'call_async_on_main_thread.js')])
 
+  # Tests that spawning a new thread does not cause a reinitialization of the global data section of the application memory area.
+  def test_pthread_global_data_initialization(self):
+    for mem_init_mode in [[], ['--memory-init-file', '0'], ['--memory-init-file', '1']]:
+      for args in [[], ['-O3']]:
+        self.btest(path_from_root('tests', 'pthread', 'test_pthread_global_data_initialization.c'), expected='20', args=args+mem_init_mode+['-s', 'USE_PTHREADS=1', '-s', 'PROXY_TO_PTHREAD=1'], also_wasm=False)
+
   # test atomicrmw i64
   def test_atomicrmw_i64(self):
     Popen([PYTHON, EMCC, path_from_root('tests', 'atomicrmw_i64.ll'), '-s', 'USE_PTHREADS=1', '-s', 'IN_TEST_HARNESS=1', '-o', 'test.html']).communicate()