Fix some nits and add new unit tests

Summary: As titled. To be on sync with python/cpython#109666.

Reviewed By: czardoz

Differential Revision: D50329585

fbshipit-source-id: d4c52143216a1d9970196ad71ef656098be50120

Author: Gamal Sallam

CinderX/test_cinderx/test_perf_profiler.py

@@ -340,6 +340,82 @@ def baz(n):
             self.assertNotIn(f"py::bar:{script}", stdout)
             self.assertNotIn(f"py::baz:{script}", stdout)
 
+    def test_pre_fork_compile(self):
+        code = """if 1:
+                import sys
+                import os
+                import sysconfig
+                from _testinternalcapi import (
+                    compile_perf_trampoline_entry,
+                    perf_trampoline_set_persist_after_fork,
+                )
+
+                def foo_fork():
+                    pass
+
+                def bar_fork():
+                    foo_fork()
+
+                def foo():
+                    pass
+
+                def bar():
+                    foo()
+
+                def compile_trampolines_for_all_functions():
+                    perf_trampoline_set_persist_after_fork(1)
+                    for _, obj in globals().items():
+                        if callable(obj) and hasattr(obj, '__code__'):
+                            compile_perf_trampoline_entry(obj.__code__)
+
+                if __name__ == "__main__":
+                    compile_trampolines_for_all_functions()
+                    pid = os.fork()
+                    if pid == 0:
+                        print(os.getpid())
+                        bar_fork()
+                    else:
+                        bar()
+                """
+
+        with temp_dir() as script_dir:
+            script = make_script(script_dir, "perftest", code)
+            with subprocess.Popen(
+                [sys.executable, "-Xperf", script],
+                universal_newlines=True,
+                stderr=subprocess.PIPE,
+                stdout=subprocess.PIPE,
+            ) as process:
+                stdout, stderr = process.communicate()
+
+        self.assertEqual(process.returncode, 0)
+        self.assertNotIn("Error:", stderr)
+        child_pid = int(stdout.strip())
+        perf_file = pathlib.Path(f"/tmp/perf-{process.pid}.map")
+        perf_child_file = pathlib.Path(f"/tmp/perf-{child_pid}.map")
+        self.assertTrue(perf_file.exists())
+        self.assertTrue(perf_child_file.exists())
+
+        perf_file_contents = perf_file.read_text()
+        self.assertIn(f"py::foo:{script}", perf_file_contents)
+        self.assertIn(f"py::bar:{script}", perf_file_contents)
+        self.assertIn(f"py::foo_fork:{script}", perf_file_contents)
+        self.assertIn(f"py::bar_fork:{script}", perf_file_contents)
+
+        child_perf_file_contents = perf_child_file.read_text()
+        self.assertIn(f"py::foo_fork:{script}", child_perf_file_contents)
+        self.assertIn(f"py::bar_fork:{script}", child_perf_file_contents)
+
+        # Pre-compiled perf-map entries of a forked process must be
+        # identical in both the parent and child perf-map files.
+        perf_file_lines = perf_file_contents.split("\n")
+        for line in perf_file_lines:
+            if (
+                f"py::foo_fork:{script}" in line
+                or f"py::bar_fork:{script}" in line
+            ):
+                self.assertIn(line, child_perf_file_contents)
+
 
 if __name__ == "__main__":
     unittest.main()

Modules/_testinternalcapi.c

@@ -460,6 +460,36 @@ perf_map_state_teardown(PyObject *Py_UNUSED(self), PyObject *Py_UNUSED(ignored))
     Py_RETURN_NONE;
 }
 
+static PyObject *
+compile_perf_trampoline_entry(PyObject *self, PyObject *args)
+{
+    PyObject *co;
+    if (!PyArg_ParseTuple(args, "O!", &PyCode_Type, &co)) {
+        return NULL;
+    }
+    int ret = PyUnstable_PerfTrampoline_CompileCode((PyCodeObject *)co);
+    if (ret != 0) {
+        PyErr_SetString(PyExc_AssertionError, "Failed to compile trampoline");
+        return NULL;
+    }
+    return PyLong_FromLong(ret);
+}
+
+static PyObject *
+perf_trampoline_set_persist_after_fork(PyObject *self, PyObject *args)
+{
+    int enable;
+    if (!PyArg_ParseTuple(args, "i", &enable)) {
+        return NULL;
+    }
+    int ret = PyUnstable_PerfTrampoline_SetPersistAfterFork(enable);
+    if (ret == 0) {
+        PyErr_SetString(PyExc_AssertionError, "Failed to set persist_after_fork");
+        return NULL;
+    }
+    return PyLong_FromLong(ret);
+}
+
 // These are used in native calling tests, ensure the compiler
 // doesn't hide or remove these symbols
 __attribute__((used))
@@ -490,6 +520,8 @@ static PyMethodDef TestMethods[] = {
     {"test_gc_visit_objects_exit_early", test_gc_visit_objects_exit_early, METH_NOARGS},
     {"write_perf_map_entry", write_perf_map_entry, METH_VARARGS},
     {"perf_map_state_teardown", perf_map_state_teardown, METH_NOARGS},
+    {"compile_perf_trampoline_entry", compile_perf_trampoline_entry, METH_VARARGS},
+    {"perf_trampoline_set_persist_after_fork", perf_trampoline_set_persist_after_fork, METH_VARARGS},
     {NULL, NULL} /* sentinel */
 };
 

Objects/perf_trampoline.c

@@ -456,11 +456,11 @@ _PyPerfTrampoline_AfterFork_Child(void)
 {
 #ifdef PY_HAVE_PERF_TRAMPOLINE
     PyUnstable_PerfMapState_Fini();
-    if (persist_after_fork){
+    if (persist_after_fork) {
         char filename[256];
         pid_t parent_pid = getppid();
         snprintf(filename, sizeof(filename), "/tmp/perf-%d.map", parent_pid);
-        if(PyUnstable_CopyPerfMapFile(filename) != 0){
+        if (PyUnstable_CopyPerfMapFile(filename) != 0) {
             return PyStatus_Error("Failed to copy perf map file.");
         }
     } else {

Python/sysmodule.c

@@ -2148,7 +2148,7 @@ PyAPI_FUNC(int) PyUnstable_WritePerfMapEntry(
 #ifndef MS_WINDOWS
     if (perf_map_state.perf_map == NULL) {
         int ret = PyUnstable_PerfMapState_Init();
-        if(ret != 0){
+        if (ret != 0) {
             return ret;
         }
     }
@@ -2183,28 +2183,30 @@ PyAPI_FUNC(int) PyUnstable_CopyPerfMapFile(const char* parent_filename) {
     }
     if (perf_map_state.perf_map == NULL) {
         int ret = PyUnstable_PerfMapState_Init();
-        if(ret != 0){
+        if (ret != 0){
             return ret;
         }
     }
     char buf[4096];
     PyThread_acquire_lock(perf_map_state.map_lock, 1);
+    int fflush_result = 0, result = 0;
     while (1) {
         size_t bytes_read = fread(buf, 1, sizeof(buf), from);
         size_t bytes_written = fwrite(buf, 1, bytes_read, perf_map_state.perf_map);
-        fflush(perf_map_state.perf_map);
+        fflush_result = fflush(perf_map_state.perf_map);
 
-        if (bytes_read < sizeof(buf) && feof(from)) {
-            fclose(from);
-            PyThread_release_lock(perf_map_state.map_lock);
-            return 0;
+        if (fflush_result != 0 || bytes_read == 0 || bytes_written < bytes_read) {
+            result = -1;
+            goto close_and_release;
         }
-        if (bytes_read == 0 || bytes_written < bytes_read) {
-            fclose(from);
-            PyThread_release_lock(perf_map_state.map_lock);
-            return -1;
+        if (bytes_read < sizeof(buf) && feof(from)) {
+            goto close_and_release;
         }
     }
+close_and_release:
+    fclose(from);
+    PyThread_release_lock(perf_map_state.map_lock);
+    return result;
 #endif
     return 0;
 }