bpo-40513: Per-interpreter recursion_limit (GH-19929)

Move recursion_limit member from _PyRuntimeState.ceval to
PyInterpreterState.ceval.

* Py_SetRecursionLimit() now only sets _Py_CheckRecursionLimit
  of ceval.c if the current Python thread is part of the main
  interpreter.
* Inline _Py_MakeEndRecCheck() into _Py_LeaveRecursiveCall().
* Convert _Py_RecursionLimitLowerWaterMark() macro into a static
  inline function.

4 files changed, 28 insertions, 22 deletions

diff --git a/Include/internal/pycore_ceval.h b/Include/internal/pycore_ceval.h
index 2df796d..18c8f02 100644
--- a/Include/internal/pycore_ceval.h
+++ b/Include/internal/pycore_ceval.h
@@ -65,12 +65,12 @@ PyAPI_DATA(int) _Py_CheckRecursionLimit;
 /* With USE_STACKCHECK macro defined, trigger stack checks in
    _Py_CheckRecursiveCall() on every 64th call to Py_EnterRecursiveCall. */
 static inline int _Py_MakeRecCheck(PyThreadState *tstate)  {
-    return (++tstate->recursion_depth > _Py_CheckRecursionLimit
+    return (++tstate->recursion_depth > tstate->interp->ceval.recursion_limit
             || ++tstate->stackcheck_counter > 64);
 }
 #else
 static inline int _Py_MakeRecCheck(PyThreadState *tstate) {
-    return (++tstate->recursion_depth > _Py_CheckRecursionLimit);
+    return (++tstate->recursion_depth > tstate->interp->ceval.recursion_limit);
 }
 #endif
 
@@ -90,20 +90,22 @@ static inline int _Py_EnterRecursiveCall_inline(const char *where) {
 
 #define Py_EnterRecursiveCall(where) _Py_EnterRecursiveCall_inline(where)
 
-
 /* Compute the "lower-water mark" for a recursion limit. When
  * Py_LeaveRecursiveCall() is called with a recursion depth below this mark,
  * the overflowed flag is reset to 0. */
-#define _Py_RecursionLimitLowerWaterMark(limit) \
-    (((limit) > 200) \
-        ? ((limit) - 50) \
-        : (3 * ((limit) >> 2)))
-
-#define _Py_MakeEndRecCheck(x) \
-    (--(x) < _Py_RecursionLimitLowerWaterMark(_Py_CheckRecursionLimit))
+static inline int _Py_RecursionLimitLowerWaterMark(int limit) {
+    if (limit > 200) {
+        return (limit - 50);
+    }
+    else {
+        return (3 * (limit >> 2));
+    }
+}
 
 static inline void _Py_LeaveRecursiveCall(PyThreadState *tstate)  {
-    if (_Py_MakeEndRecCheck(tstate->recursion_depth)) {
+    tstate->recursion_depth--;
+    int limit = tstate->interp->ceval.recursion_limit;
+    if (tstate->recursion_depth < _Py_RecursionLimitLowerWaterMark(limit)) {
         tstate->overflowed = 0;
     }
 }
diff --git a/Include/internal/pycore_interp.h b/Include/internal/pycore_interp.h
index fafc72e..0829101 100644
--- a/Include/internal/pycore_interp.h
+++ b/Include/internal/pycore_interp.h
@@ -33,6 +33,7 @@ struct _pending_calls {
 };
 
 struct _ceval_state {
+    int recursion_limit;
     /* Records whether tracing is on for any thread.  Counts the number
        of threads for which tstate->c_tracefunc is non-NULL, so if the
        value is 0, we know we don't have to check this thread's
diff --git a/Include/internal/pycore_runtime.h b/Include/internal/pycore_runtime.h
index c597335..8ca1dfb 100644
--- a/Include/internal/pycore_runtime.h
+++ b/Include/internal/pycore_runtime.h
@@ -14,7 +14,6 @@ extern "C" {
 /* ceval state */
 
 struct _ceval_runtime_state {
-    int recursion_limit;
     struct _gil_runtime_state gil;
 };
 
diff --git a/Python/ceval.c b/Python/ceval.c
index 6b00273..601e21a 100644
--- a/Python/ceval.c
+++ b/Python/ceval.c
@@ -699,7 +699,6 @@ int _Py_CheckRecursionLimit = Py_DEFAULT_RECURSION_LIMIT;
 void
 _PyEval_InitRuntimeState(struct _ceval_runtime_state *ceval)
 {
-    ceval->recursion_limit = Py_DEFAULT_RECURSION_LIMIT;
     _Py_CheckRecursionLimit = Py_DEFAULT_RECURSION_LIMIT;
     _gil_initialize(&ceval->gil);
 }
@@ -707,6 +706,8 @@ _PyEval_InitRuntimeState(struct _ceval_runtime_state *ceval)
 int
 _PyEval_InitState(struct _ceval_state *ceval)
 {
+    ceval->recursion_limit = Py_DEFAULT_RECURSION_LIMIT;
+
     struct _pending_calls *pending = &ceval->pending;
     assert(pending->lock == NULL);
 
@@ -730,16 +731,18 @@ _PyEval_FiniState(struct _ceval_state *ceval)
 int
 Py_GetRecursionLimit(void)
 {
-    struct _ceval_runtime_state *ceval = &_PyRuntime.ceval;
-    return ceval->recursion_limit;
+    PyThreadState *tstate = _PyThreadState_GET();
+    return tstate->interp->ceval.recursion_limit;
 }
 
 void
 Py_SetRecursionLimit(int new_limit)
 {
-    struct _ceval_runtime_state *ceval = &_PyRuntime.ceval;
-    ceval->recursion_limit = new_limit;
-    _Py_CheckRecursionLimit = new_limit;
+    PyThreadState *tstate = _PyThreadState_GET();
+    tstate->interp->ceval.recursion_limit = new_limit;
+    if (_Py_IsMainInterpreter(tstate)) {
+        _Py_CheckRecursionLimit = new_limit;
+    }
 }
 
 /* The function _Py_EnterRecursiveCall() only calls _Py_CheckRecursiveCall()
@@ -750,8 +753,7 @@ Py_SetRecursionLimit(int new_limit)
 int
 _Py_CheckRecursiveCall(PyThreadState *tstate, const char *where)
 {
-    _PyRuntimeState *runtime = tstate->interp->runtime;
-    int recursion_limit = runtime->ceval.recursion_limit;
+    int recursion_limit = tstate->interp->ceval.recursion_limit;
 
 #ifdef USE_STACKCHECK
     tstate->stackcheck_counter = 0;
@@ -760,8 +762,10 @@ _Py_CheckRecursiveCall(PyThreadState *tstate, const char *where)
         _PyErr_SetString(tstate, PyExc_MemoryError, "Stack overflow");
         return -1;
     }
-    /* Needed for ABI backwards-compatibility (see bpo-31857) */
-    _Py_CheckRecursionLimit = recursion_limit;
+    if (_Py_IsMainInterpreter(tstate)) {
+        /* Needed for ABI backwards-compatibility (see bpo-31857) */
+        _Py_CheckRecursionLimit = recursion_limit;
+    }
 #endif
     if (tstate->recursion_critical)
         /* Somebody asked that we don't check for recursion. */