gh-81057: Move More Globals in Core Code to _PyRuntimeState (gh-99516)

#81057

Author: ericsnowcurrently

Include/cpython/modsupport.h

@@ -106,5 +106,3 @@ PyAPI_FUNC(PyObject * const *) _PyArg_UnpackKeywordsWithVararg(
                            (minpos), (maxpos), (minkw), (buf)))
 
 PyAPI_FUNC(PyObject *) _PyModule_CreateInitialized(PyModuleDef*, int apiver);
-
-PyAPI_DATA(const char *) _Py_PackageContext;

Include/internal/pycore_fileutils.h

@@ -10,6 +10,11 @@ extern "C" {
 
 #include <locale.h>   /* struct lconv */
 
+
+struct _fileutils_state {
+    int force_ascii;
+};
+
 typedef enum {
     _Py_ERROR_UNKNOWN=0,
     _Py_ERROR_STRICT,

Include/internal/pycore_floatobject.h

@@ -19,6 +19,18 @@ extern void _PyFloat_FiniType(PyInterpreterState *);
 
 /* other API */
 
+enum _py_float_format_type {
+    _py_float_format_unknown,
+    _py_float_format_ieee_big_endian,
+    _py_float_format_ieee_little_endian,
+};
+
+struct _Py_float_runtime_state {
+    enum _py_float_format_type float_format;
+    enum _py_float_format_type double_format;
+};
+
+
 #ifndef WITH_FREELISTS
 // without freelists
 #  define PyFloat_MAXFREELIST 0

Include/internal/pycore_import.h

@@ -32,6 +32,8 @@ struct _import_runtime_state {
         _PyTime_t accumulated;
         int header;
     } find_and_load;
+    /* Package context -- the full module name for package imports */
+    const char * pkgcontext;
 };
 
 

Include/internal/pycore_interp.h

@@ -28,6 +28,7 @@ extern "C" {
 
 
 struct _pending_calls {
+    int busy;
     PyThread_type_lock lock;
     /* Request for running pending calls. */
     _Py_atomic_int calls_to_do;

Include/internal/pycore_pyhash.h

@@ -5,6 +5,36 @@
 #  error "this header requires Py_BUILD_CORE define"
 #endif
 
-uint64_t _Py_KeyedHash(uint64_t, const char *, Py_ssize_t);
 
+struct pyhash_runtime_state {
+    struct {
+#ifndef MS_WINDOWS
+        int fd;
+        dev_t st_dev;
+        ino_t st_ino;
+#else
+    // This is a placeholder so the struct isn't empty on Windows.
+    int _not_used;
+#endif
+    } urandom_cache;
+};
+
+#ifndef MS_WINDOWS
+# define _py_urandom_cache_INIT \
+    { \
+        .fd = -1, \
+    }
+#else
+# define _py_urandom_cache_INIT {0}
 #endif
+
+#define pyhash_state_INIT \
+    { \
+        .urandom_cache = _py_urandom_cache_INIT, \
+    }
+
+
+uint64_t _Py_KeyedHash(uint64_t, const char *, Py_ssize_t);
+
+
+#endif  // Py_INTERNAL_HASH_H

Include/internal/pycore_pylifecycle.h

@@ -14,10 +14,6 @@ extern "C" {
 struct _PyArgv;
 struct pyruntimestate;
 
-/* True if the main interpreter thread exited due to an unhandled
- * KeyboardInterrupt exception, suggesting the user pressed ^C. */
-PyAPI_DATA(int) _Py_UnhandledKeyboardInterrupt;
-
 extern int _Py_SetFileSystemEncoding(
     const char *encoding,
     const char *errors);

Include/internal/pycore_runtime.h

@@ -10,11 +10,13 @@ extern "C" {
 
 #include "pycore_atomic.h"          /* _Py_atomic_address */
 #include "pycore_dtoa.h"            // struct _dtoa_runtime_state
+#include "pycore_floatobject.h"     // struct _Py_float_runtime_state
 #include "pycore_gil.h"             // struct _gil_runtime_state
 #include "pycore_global_objects.h"  // struct _Py_global_objects
 #include "pycore_import.h"          // struct _import_runtime_state
 #include "pycore_interp.h"          // PyInterpreterState
 #include "pycore_pymem.h"           // struct _pymem_allocators
+#include "pycore_pyhash.h"          // struct pyhash_runtime_state
 #include "pycore_obmalloc.h"        // struct obmalloc_state
 #include "pycore_unicodeobject.h"   // struct _Py_unicode_runtime_ids
 
@@ -92,6 +94,12 @@ typedef struct pyruntimestate {
 
     struct _pymem_allocators allocators;
     struct _obmalloc_state obmalloc;
+    struct pyhash_runtime_state pyhash_state;
+    struct {
+        /* True if the main interpreter thread exited due to an unhandled
+         * KeyboardInterrupt exception, suggesting the user pressed ^C. */
+        int unhandled_keyboard_interrupt;
+    } signals;
 
     struct pyinterpreters {
         PyThread_type_lock mutex;
@@ -131,6 +139,7 @@ typedef struct pyruntimestate {
         struct _PyTraceMalloc_Config config;
     } tracemalloc;
     struct _dtoa_runtime_state dtoa;
+    struct _fileutils_state fileutils;
 
     PyPreConfig preconfig;
 
@@ -140,7 +149,8 @@ typedef struct pyruntimestate {
     void *open_code_userdata;
     _Py_AuditHookEntry *audit_hook_head;
 
-    struct _Py_unicode_runtime_ids unicode_ids;
+    struct _Py_float_runtime_state float_state;
+    struct _Py_unicode_runtime_state unicode_state;
 
     struct {
         /* Used to set PyTypeObject.tp_version_tag */

Include/internal/pycore_runtime_init.h

@@ -19,30 +19,18 @@ extern "C" {
 
 #define _PyRuntimeState_INIT(runtime) \
     { \
-        .gilstate = { \
-            .check_enabled = 1, \
-            /* A TSS key must be initialized with Py_tss_NEEDS_INIT \
-               in accordance with the specification. */ \
-            .autoTSSkey = Py_tss_NEEDS_INIT, \
-        }, \
         .allocators = { \
             _pymem_allocators_standard_INIT(runtime), \
             _pymem_allocators_debug_INIT, \
             _pymem_allocators_obj_arena_INIT, \
         }, \
         .obmalloc = _obmalloc_state_INIT(runtime.obmalloc), \
+        .pyhash_state = pyhash_state_INIT, \
         .interpreters = { \
             /* This prevents interpreters from getting created \
               until _PyInterpreterState_Enable() is called. */ \
             .next_id = -1, \
         }, \
-        .tracemalloc = { \
-            .config = _PyTraceMalloc_Config_INIT, \
-        }, \
-        .dtoa = _dtoa_runtime_state_INIT(runtime), \
-        .types = { \
-            .next_version_tag = 1, \
-        }, \
         .imports = { \
             .lock = { \
                 .mutex = NULL, \
@@ -53,6 +41,26 @@ extern "C" {
                 .header = 1, \
             }, \
         }, \
+        .gilstate = { \
+            .check_enabled = 1, \
+            /* A TSS key must be initialized with Py_tss_NEEDS_INIT \
+               in accordance with the specification. */ \
+            .autoTSSkey = Py_tss_NEEDS_INIT, \
+        }, \
+        .tracemalloc = { \
+            .config = _PyTraceMalloc_Config_INIT, \
+        }, \
+        .dtoa = _dtoa_runtime_state_INIT(runtime), \
+        .fileutils = { \
+            .force_ascii = -1, \
+        }, \
+        .float_state = { \
+            .float_format = _py_float_format_unknown, \
+            .double_format = _py_float_format_unknown, \
+        }, \
+        .types = { \
+            .next_version_tag = 1, \
+        }, \
         .global_objects = { \
             .singletons = { \
                 .small_ints = _Py_small_ints_INIT, \

Include/internal/pycore_unicodeobject.h

@@ -31,6 +31,10 @@ struct _Py_unicode_runtime_ids {
     Py_ssize_t next_index;
 };
 
+struct _Py_unicode_runtime_state {
+    struct _Py_unicode_runtime_ids ids;
+};
+
 /* fs_codec.encoding is initialized to NULL.
    Later, it is set to a non-NULL string by _PyUnicode_InitEncodings(). */
 struct _Py_unicode_fs_codec {

Modules/main.c

@@ -296,10 +296,10 @@ pymain_run_module(const wchar_t *modname, int set_argv0)
         Py_DECREF(module);
         return pymain_exit_err_print();
     }
-    _Py_UnhandledKeyboardInterrupt = 0;
+    _PyRuntime.signals.unhandled_keyboard_interrupt = 0;
     result = PyObject_Call(runmodule, runargs, NULL);
     if (!result && PyErr_Occurred() == PyExc_KeyboardInterrupt) {
-        _Py_UnhandledKeyboardInterrupt = 1;
+        _PyRuntime.signals.unhandled_keyboard_interrupt = 1;
     }
     Py_DECREF(runpy);
     Py_DECREF(runmodule);
@@ -696,7 +696,7 @@ Py_RunMain(void)
 
     pymain_free();
 
-    if (_Py_UnhandledKeyboardInterrupt) {
+    if (_PyRuntime.signals.unhandled_keyboard_interrupt) {
         exitcode = exit_sigint();
     }
 

Objects/floatobject.c

@@ -1723,12 +1723,14 @@ float___getnewargs___impl(PyObject *self)
 }
 
 /* this is for the benefit of the pack/unpack routines below */
+typedef enum _py_float_format_type float_format_type;
+#define unknown_format _py_float_format_unknown
+#define ieee_big_endian_format _py_float_format_ieee_big_endian
+#define ieee_little_endian_format _py_float_format_ieee_little_endian
 
-typedef enum {
-    unknown_format, ieee_big_endian_format, ieee_little_endian_format
-} float_format_type;
+#define float_format (_PyRuntime.float_state.float_format)
+#define double_format (_PyRuntime.float_state.double_format)
 
-static float_format_type double_format, float_format;
 
 /*[clinic input]
 @classmethod
@@ -1929,13 +1931,9 @@ PyTypeObject PyFloat_Type = {
     .tp_vectorcall = (vectorcallfunc)float_vectorcall,
 };
 
-void
-_PyFloat_InitState(PyInterpreterState *interp)
+static void
+_init_global_state(void)
 {
-    if (!_Py_IsMainInterpreter(interp)) {
-        return;
-    }
-
     float_format_type detected_double_format, detected_float_format;
 
     /* We attempt to determine if this machine is using IEEE
@@ -1985,6 +1983,15 @@ _PyFloat_InitState(PyInterpreterState *interp)
     float_format = detected_float_format;
 }
 
+void
+_PyFloat_InitState(PyInterpreterState *interp)
+{
+    if (!_Py_IsMainInterpreter(interp)) {
+        return;
+    }
+    _init_global_state();
+}
+
 PyStatus
 _PyFloat_InitTypes(PyInterpreterState *interp)
 {

Objects/moduleobject.c

@@ -218,13 +218,15 @@ _PyModule_CreateInitialized(PyModuleDef* module, int module_api_version)
        _Py_PackageContext, and PyModule_Create*() will substitute this
        (if the name actually matches).
     */
+#define _Py_PackageContext (_PyRuntime.imports.pkgcontext)
     if (_Py_PackageContext != NULL) {
         const char *p = strrchr(_Py_PackageContext, '.');
         if (p != NULL && strcmp(module->m_name, p+1) == 0) {
             name = _Py_PackageContext;
             _Py_PackageContext = NULL;
         }
     }
+#undef _Py_PackageContext
     if ((m = (PyModuleObject*)PyModule_New(name)) == NULL)
         return NULL;
 

Objects/unicodeobject.c

@@ -1856,7 +1856,7 @@ _PyUnicode_FromId(_Py_Identifier *id)
 
     Py_ssize_t index = _Py_atomic_size_get(&id->index);
     if (index < 0) {
-        struct _Py_unicode_runtime_ids *rt_ids = &interp->runtime->unicode_ids;
+        struct _Py_unicode_runtime_ids *rt_ids = &interp->runtime->unicode_state.ids;
 
         PyThread_acquire_lock(rt_ids->lock, WAIT_LOCK);
         // Check again to detect concurrent access. Another thread can have
@@ -14491,12 +14491,14 @@ PyTypeObject PyUnicode_Type = {
 
 /* Initialize the Unicode implementation */
 
-void
-_PyUnicode_InitState(PyInterpreterState *interp)
+static void
+_init_global_state(void)
 {
-    if (!_Py_IsMainInterpreter(interp)) {
+    static int initialized = 0;
+    if (initialized) {
         return;
     }
+    initialized = 1;
 
     /* initialize the linebreak bloom filter */
     const Py_UCS2 linebreak[] = {
@@ -14514,6 +14516,15 @@ _PyUnicode_InitState(PyInterpreterState *interp)
         Py_ARRAY_LENGTH(linebreak));
 }
 
+void
+_PyUnicode_InitState(PyInterpreterState *interp)
+{
+    if (!_Py_IsMainInterpreter(interp)) {
+        return;
+    }
+    _init_global_state();
+}
+
 
 PyStatus
 _PyUnicode_InitGlobalObjects(PyInterpreterState *interp)

Python/bootstrap_hash.c

@@ -1,6 +1,7 @@
 #include "Python.h"
 #include "pycore_initconfig.h"
 #include "pycore_fileutils.h"     // _Py_fstat_noraise()
+#include "pycore_runtime.h"       // _PyRuntime
 
 #ifdef MS_WINDOWS
 #  include <windows.h>
@@ -263,11 +264,7 @@ py_getentropy(char *buffer, Py_ssize_t size, int raise)
 #endif /* defined(HAVE_GETENTROPY) && !(defined(__sun) && defined(__SVR4)) */
 
 
-static struct {
-    int fd;
-    dev_t st_dev;
-    ino_t st_ino;
-} urandom_cache = { -1 };
+#define urandom_cache (_PyRuntime.pyhash_state.urandom_cache)
 
 /* Read random bytes from the /dev/urandom device:
 
@@ -402,6 +399,9 @@ dev_urandom_close(void)
         urandom_cache.fd = -1;
     }
 }
+
+#undef urandom_cache
+
 #endif /* !MS_WINDOWS */