gh-134584: Cleanups for GH-135860 (GH-142604)

diff --git a/Include/internal/pycore_opcode_metadata.h b/Include/internal/pycore_opcode_metadata.h
index 13e58721ebee9e4a9e38e37496bbce6cec96bfe4..82c7cf486b1a008ed00335b57c829d02edea3197 100644 (file)
--- a/Include/internal/pycore_opcode_metadata.h
+++ b/Include/internal/pycore_opcode_metadata.h
@@ -1130,7 +1130,7 @@ const struct opcode_metadata _PyOpcode_opcode_metadata[267] = {
     [CALL_PY_EXACT_ARGS] = { true, INSTR_FMT_IBC00, HAS_ARG_FLAG | HAS_DEOPT_FLAG | HAS_EXIT_FLAG | HAS_SYNC_SP_FLAG | HAS_NEEDS_GUARD_IP_FLAG },
     [CALL_PY_GENERAL] = { true, INSTR_FMT_IBC00, HAS_ARG_FLAG | HAS_DEOPT_FLAG | HAS_EXIT_FLAG | HAS_ERROR_FLAG | HAS_ERROR_NO_POP_FLAG | HAS_ESCAPES_FLAG | HAS_SYNC_SP_FLAG | HAS_NEEDS_GUARD_IP_FLAG },
     [CALL_STR_1] = { true, INSTR_FMT_IBC00, HAS_ARG_FLAG | HAS_EVAL_BREAK_FLAG | HAS_DEOPT_FLAG | HAS_ERROR_FLAG | HAS_ESCAPES_FLAG },
-    [CALL_TUPLE_1] = { true, INSTR_FMT_IBC00, HAS_ARG_FLAG | HAS_EVAL_BREAK_FLAG | HAS_DEOPT_FLAG | HAS_ERROR_FLAG | HAS_ESCAPES_FLAG },
+    [CALL_TUPLE_1] = { true, INSTR_FMT_IBC00, HAS_ARG_FLAG | HAS_EVAL_BREAK_FLAG | HAS_DEOPT_FLAG | HAS_ERROR_FLAG | HAS_ERROR_NO_POP_FLAG | HAS_ESCAPES_FLAG },
     [CALL_TYPE_1] = { true, INSTR_FMT_IBC00, HAS_ARG_FLAG | HAS_DEOPT_FLAG | HAS_ESCAPES_FLAG },
     [CHECK_EG_MATCH] = { true, INSTR_FMT_IX, HAS_ERROR_FLAG | HAS_ESCAPES_FLAG },
     [CHECK_EXC_MATCH] = { true, INSTR_FMT_IX, HAS_ERROR_FLAG | HAS_ERROR_NO_POP_FLAG | HAS_ESCAPES_FLAG },
@@ -1376,7 +1376,7 @@ _PyOpcode_macro_expansion[256] = {
     [CALL_PY_EXACT_ARGS] = { .nuops = 8, .uops = { { _CHECK_PEP_523, OPARG_SIMPLE, 1 }, { _CHECK_FUNCTION_VERSION, 2, 1 }, { _CHECK_FUNCTION_EXACT_ARGS, OPARG_SIMPLE, 3 }, { _CHECK_STACK_SPACE, OPARG_SIMPLE, 3 }, { _CHECK_RECURSION_REMAINING, OPARG_SIMPLE, 3 }, { _INIT_CALL_PY_EXACT_ARGS, OPARG_SIMPLE, 3 }, { _SAVE_RETURN_OFFSET, OPARG_SAVE_RETURN_OFFSET, 3 }, { _PUSH_FRAME, OPARG_SIMPLE, 3 } } },
     [CALL_PY_GENERAL] = { .nuops = 6, .uops = { { _CHECK_PEP_523, OPARG_SIMPLE, 1 }, { _CHECK_FUNCTION_VERSION, 2, 1 }, { _CHECK_RECURSION_REMAINING, OPARG_SIMPLE, 3 }, { _PY_FRAME_GENERAL, OPARG_SIMPLE, 3 }, { _SAVE_RETURN_OFFSET, OPARG_SAVE_RETURN_OFFSET, 3 }, { _PUSH_FRAME, OPARG_SIMPLE, 3 } } },
     [CALL_STR_1] = { .nuops = 4, .uops = { { _GUARD_NOS_NULL, OPARG_SIMPLE, 3 }, { _GUARD_CALLABLE_STR_1, OPARG_SIMPLE, 3 }, { _CALL_STR_1, OPARG_SIMPLE, 3 }, { _CHECK_PERIODIC_AT_END, OPARG_REPLACED, 3 } } },
-    [CALL_TUPLE_1] = { .nuops = 4, .uops = { { _GUARD_NOS_NULL, OPARG_SIMPLE, 3 }, { _GUARD_CALLABLE_TUPLE_1, OPARG_SIMPLE, 3 }, { _CALL_TUPLE_1, OPARG_SIMPLE, 3 }, { _CHECK_PERIODIC_AT_END, OPARG_REPLACED, 3 } } },
+    [CALL_TUPLE_1] = { .nuops = 5, .uops = { { _GUARD_NOS_NULL, OPARG_SIMPLE, 3 }, { _GUARD_CALLABLE_TUPLE_1, OPARG_SIMPLE, 3 }, { _CALL_TUPLE_1, OPARG_SIMPLE, 3 }, { _POP_TOP, OPARG_SIMPLE, 3 }, { _CHECK_PERIODIC_AT_END, OPARG_REPLACED, 3 } } },
     [CALL_TYPE_1] = { .nuops = 3, .uops = { { _GUARD_NOS_NULL, OPARG_SIMPLE, 3 }, { _GUARD_CALLABLE_TYPE_1, OPARG_SIMPLE, 3 }, { _CALL_TYPE_1, OPARG_SIMPLE, 3 } } },
     [CHECK_EG_MATCH] = { .nuops = 1, .uops = { { _CHECK_EG_MATCH, OPARG_SIMPLE, 0 } } },
     [CHECK_EXC_MATCH] = { .nuops = 1, .uops = { { _CHECK_EXC_MATCH, OPARG_SIMPLE, 0 } } },

diff --git a/Include/internal/pycore_uop_ids.h b/Include/internal/pycore_uop_ids.h
index 56928b066852703acf8eed7ed69d686f52f7b5a6..ce1ad5a4c8a5f68be1e3adcd2ce8ad1b8fed2df5 100644 (file)
--- a/Include/internal/pycore_uop_ids.h
+++ b/Include/internal/pycore_uop_ids.h
@@ -439,7 +439,7 @@ extern "C" {
 #define _CALL_METHOD_DESCRIPTOR_O_r01 632
 #define _CALL_NON_PY_GENERAL_r01 633
 #define _CALL_STR_1_r31 634
-#define _CALL_TUPLE_1_r31 635
+#define _CALL_TUPLE_1_r32 635
 #define _CALL_TYPE_1_r31 636
 #define _CHECK_AND_ALLOCATE_OBJECT_r00 637
 #define _CHECK_ATTR_CLASS_r01 638

diff --git a/Include/internal/pycore_uop_metadata.h b/Include/internal/pycore_uop_metadata.h
index 2ae84bdee0df1238716e0b94f8531decd70322f8..0a37ce98c7eb4877e1852a0ed0ef6b103f8e8176 100644 (file)
--- a/Include/internal/pycore_uop_metadata.h
+++ b/Include/internal/pycore_uop_metadata.h
@@ -276,7 +276,7 @@ const uint32_t _PyUop_Flags[MAX_UOP_ID+1] = {
     [_GUARD_CALLABLE_STR_1] = HAS_DEOPT_FLAG,
     [_CALL_STR_1] = HAS_ARG_FLAG | HAS_ERROR_FLAG | HAS_ESCAPES_FLAG,
     [_GUARD_CALLABLE_TUPLE_1] = HAS_DEOPT_FLAG,
-    [_CALL_TUPLE_1] = HAS_ARG_FLAG | HAS_ERROR_FLAG | HAS_ESCAPES_FLAG,
+    [_CALL_TUPLE_1] = HAS_ARG_FLAG | HAS_ERROR_FLAG | HAS_ERROR_NO_POP_FLAG | HAS_ESCAPES_FLAG,
     [_CHECK_AND_ALLOCATE_OBJECT] = HAS_ARG_FLAG | HAS_DEOPT_FLAG | HAS_ERROR_FLAG | HAS_ERROR_NO_POP_FLAG | HAS_ESCAPES_FLAG,
     [_CREATE_INIT_FRAME] = HAS_ARG_FLAG | HAS_ERROR_FLAG | HAS_ERROR_NO_POP_FLAG | HAS_ESCAPES_FLAG | HAS_SYNC_SP_FLAG,
     [_EXIT_INIT_CHECK] = HAS_ERROR_FLAG | HAS_ERROR_NO_POP_FLAG | HAS_ESCAPES_FLAG,
@@ -2543,7 +2543,7 @@ const _PyUopCachingInfo _PyUop_Caching[MAX_UOP_ID+1] = {
             { -1, -1, -1 },
             { -1, -1, -1 },
             { -1, -1, -1 },
-            { 1, 0, _CALL_TUPLE_1_r31 },
+            { 2, 0, _CALL_TUPLE_1_r32 },
         },
     },
     [_CHECK_AND_ALLOCATE_OBJECT] = {
@@ -3739,7 +3739,7 @@ const uint16_t _PyUop_Uncached[MAX_UOP_REGS_ID+1] = {
     [_GUARD_CALLABLE_TUPLE_1_r13] = _GUARD_CALLABLE_TUPLE_1,
     [_GUARD_CALLABLE_TUPLE_1_r23] = _GUARD_CALLABLE_TUPLE_1,
     [_GUARD_CALLABLE_TUPLE_1_r33] = _GUARD_CALLABLE_TUPLE_1,
-    [_CALL_TUPLE_1_r31] = _CALL_TUPLE_1,
+    [_CALL_TUPLE_1_r32] = _CALL_TUPLE_1,
     [_CHECK_AND_ALLOCATE_OBJECT_r00] = _CHECK_AND_ALLOCATE_OBJECT,
     [_CREATE_INIT_FRAME_r01] = _CREATE_INIT_FRAME,
     [_EXIT_INIT_CHECK_r10] = _EXIT_INIT_CHECK,
@@ -4058,7 +4058,7 @@ const char *const _PyOpcode_uop_name[MAX_UOP_REGS_ID+1] = {
     [_CALL_STR_1] = "_CALL_STR_1",
     [_CALL_STR_1_r31] = "_CALL_STR_1_r31",
     [_CALL_TUPLE_1] = "_CALL_TUPLE_1",
-    [_CALL_TUPLE_1_r31] = "_CALL_TUPLE_1_r31",
+    [_CALL_TUPLE_1_r32] = "_CALL_TUPLE_1_r32",
     [_CALL_TYPE_1] = "_CALL_TYPE_1",
     [_CALL_TYPE_1_r31] = "_CALL_TYPE_1_r31",
     [_CHECK_AND_ALLOCATE_OBJECT] = "_CHECK_AND_ALLOCATE_OBJECT",

diff --git a/Lib/test/test_capi/test_opt.py b/Lib/test/test_capi/test_opt.py
index e512c08752a02a0fc11221b357e0628fa2bd8d5d..0e7fd62c28a06545c4d5ffd5fded97c230c946aa 100644 (file)
--- a/Lib/test/test_capi/test_opt.py
+++ b/Lib/test/test_capi/test_opt.py
@@ -1938,8 +1938,7 @@ class TestUopsOptimization(unittest.TestCase):
         self.assertIsNotNone(ex)
         uops = get_opnames(ex)
         self.assertIn("_CALL_TUPLE_1", uops)
-        # Re-enable later gh-134584
-        # self.assertIn("_POP_TOP_NOP", uops)
+        self.assertIn("_POP_TOP_NOP", uops)
 
     def test_call_str_1(self):
         def testfunc(n):

diff --git a/Python/bytecodes.c b/Python/bytecodes.c
index 3e7b7d5594f8f7a606b5d310643083c5064ea47b..d96a1944cd516a92ece91617cb568e150592a90f 100644 (file)
--- a/Python/bytecodes.c
+++ b/Python/bytecodes.c
@@ -4068,15 +4068,17 @@ dummy_func(
             DEOPT_IF(callable_o != (PyObject *)&PyTuple_Type);
         }
 
-        op(_CALL_TUPLE_1, (callable, null, arg -- res)) {
+        op(_CALL_TUPLE_1, (callable, null, arg -- res, a)) {
             PyObject *arg_o = PyStackRef_AsPyObjectBorrow(arg);
 
             assert(oparg == 1);
             STAT_INC(CALL, hit);
             PyObject *res_o = PySequence_Tuple(arg_o);
-            PyStackRef_CLOSE(arg);
+            if (res_o == NULL) {
+                ERROR_NO_POP();
+            }
+            a = arg;
             INPUTS_DEAD();
-            ERROR_IF(res_o == NULL);
             res = PyStackRef_FromPyObjectSteal(res_o);
         }
 
@@ -4086,6 +4088,7 @@ dummy_func(
             _GUARD_NOS_NULL +
             _GUARD_CALLABLE_TUPLE_1 +
             _CALL_TUPLE_1 +
+            POP_TOP +
             _CHECK_PERIODIC_AT_END;
 
         op(_CHECK_AND_ALLOCATE_OBJECT, (type_version/2, callable, self_or_null, unused[oparg] -- callable, self_or_null, unused[oparg])) {

diff --git a/Python/executor_cases.c.h b/Python/executor_cases.c.h
index 9c2bffb35abbfbb2f64e4ef7879ae5d1ace79220..fab23f358682eb7ef221677474e51fc46bcb889e 100644 (file)
--- a/Python/executor_cases.c.h
+++ b/Python/executor_cases.c.h
@@ -12888,11 +12888,12 @@
             break;
         }
 
-        case _CALL_TUPLE_1_r31: {
+        case _CALL_TUPLE_1_r32: {
             CHECK_CURRENT_CACHED_VALUES(3);
             assert(WITHIN_STACK_BOUNDS_WITH_CACHE());
             _PyStackRef arg;
             _PyStackRef res;
+            _PyStackRef a;
             _PyStackRef _stack_item_0 = _tos_cache0;
             _PyStackRef _stack_item_1 = _tos_cache1;
             _PyStackRef _stack_item_2 = _tos_cache2;
@@ -12909,23 +12910,17 @@
             _PyFrame_SetStackPointer(frame, stack_pointer);
             PyObject *res_o = PySequence_Tuple(arg_o);
             stack_pointer = _PyFrame_GetStackPointer(frame);
-            stack_pointer += -1;
-            ASSERT_WITHIN_STACK_BOUNDS(__FILE__, __LINE__);
-            _PyFrame_SetStackPointer(frame, stack_pointer);
-            PyStackRef_CLOSE(arg);
-            stack_pointer = _PyFrame_GetStackPointer(frame);
             if (res_o == NULL) {
-                stack_pointer += -2;
-                ASSERT_WITHIN_STACK_BOUNDS(__FILE__, __LINE__);
                 SET_CURRENT_CACHED_VALUES(0);
                 JUMP_TO_ERROR();
             }
+            a = arg;
             res = PyStackRef_FromPyObjectSteal(res_o);
+            _tos_cache1 = a;
             _tos_cache0 = res;
-            _tos_cache1 = PyStackRef_ZERO_BITS;
             _tos_cache2 = PyStackRef_ZERO_BITS;
-            SET_CURRENT_CACHED_VALUES(1);
-            stack_pointer += -2;
+            SET_CURRENT_CACHED_VALUES(2);
+            stack_pointer += -3;
             ASSERT_WITHIN_STACK_BOUNDS(__FILE__, __LINE__);
             assert(WITHIN_STACK_BOUNDS_WITH_CACHE());
             break;

diff --git a/Python/generated_cases.c.h b/Python/generated_cases.c.h
index e72c621f10d340fee036c12b67ef2474911c9bed..413593386583a815995a2d316c8e696390e09d01 100644 (file)
--- a/Python/generated_cases.c.h
+++ b/Python/generated_cases.c.h
@@ -4040,6 +4040,8 @@
             _PyStackRef callable;
             _PyStackRef arg;
             _PyStackRef res;
+            _PyStackRef a;
+            _PyStackRef value;
             /* Skip 1 cache entry */
             /* Skip 2 cache entries */
             // _GUARD_NOS_NULL
@@ -4070,21 +4072,24 @@
                 _PyFrame_SetStackPointer(frame, stack_pointer);
                 PyObject *res_o = PySequence_Tuple(arg_o);
                 stack_pointer = _PyFrame_GetStackPointer(frame);
-                stack_pointer += -1;
-                ASSERT_WITHIN_STACK_BOUNDS(__FILE__, __LINE__);
-                _PyFrame_SetStackPointer(frame, stack_pointer);
-                PyStackRef_CLOSE(arg);
-                stack_pointer = _PyFrame_GetStackPointer(frame);
                 if (res_o == NULL) {
-                    JUMP_TO_LABEL(pop_2_error);
+                    JUMP_TO_LABEL(error);
                 }
+                a = arg;
                 res = PyStackRef_FromPyObjectSteal(res_o);
             }
-            // _CHECK_PERIODIC_AT_END
+            // _POP_TOP
             {
-                stack_pointer[-2] = res;
-                stack_pointer += -1;
+                value = a;
+                stack_pointer[-3] = res;
+                stack_pointer += -2;
                 ASSERT_WITHIN_STACK_BOUNDS(__FILE__, __LINE__);
+                _PyFrame_SetStackPointer(frame, stack_pointer);
+                PyStackRef_XCLOSE(value);
+                stack_pointer = _PyFrame_GetStackPointer(frame);
+            }
+            // _CHECK_PERIODIC_AT_END
+            {
                 _PyFrame_SetStackPointer(frame, stack_pointer);
                 int err = check_periodics(tstate);
                 stack_pointer = _PyFrame_GetStackPointer(frame);

diff --git a/Python/optimizer_bytecodes.c b/Python/optimizer_bytecodes.c
index 61f2dd1e4541482e3c68be97c59aaef36420e332..9e767464500d459c6e37db0048215306e01ca85a 100644 (file)
--- a/Python/optimizer_bytecodes.c
+++ b/Python/optimizer_bytecodes.c
@@ -1097,7 +1097,7 @@ dummy_func(void) {
         }
     }
 
-    op(_CALL_TUPLE_1, (callable, null, arg -- res)) {
+    op(_CALL_TUPLE_1, (callable, null, arg -- res, a)) {
         if (sym_matches_type(arg, &PyTuple_Type)) {
             // e.g. tuple((1, 2)) or tuple(foo) where foo is known to be a tuple
             // Note: we must strip the reference information because it goes
@@ -1107,6 +1107,7 @@ dummy_func(void) {
         else {
             res = sym_new_type(ctx, &PyTuple_Type);
         }
+        a = arg;
     }
 
     op(_GUARD_TOS_LIST, (tos -- tos)) {

diff --git a/Python/optimizer_cases.c.h b/Python/optimizer_cases.c.h
index 7d95bce7679815618bada51066be80c7f91d69d8..c3f122215ed7bd8c54a65da62efb29cb21290349 100644 (file)
--- a/Python/optimizer_cases.c.h
+++ b/Python/optimizer_cases.c.h
@@ -2882,6 +2882,7 @@
         case _CALL_TUPLE_1: {
             JitOptRef arg;
             JitOptRef res;
+            JitOptRef a;
             arg = stack_pointer[-1];
             if (sym_matches_type(arg, &PyTuple_Type)) {
                 res = PyJitRef_StripReferenceInfo(arg);
@@ -2889,9 +2890,11 @@
             else {
                 res = sym_new_type(ctx, &PyTuple_Type);
             }
-            CHECK_STACK_BOUNDS(-2);
+            a = arg;
+            CHECK_STACK_BOUNDS(-1);
             stack_pointer[-3] = res;
-            stack_pointer += -2;
+            stack_pointer[-2] = a;
+            stack_pointer += -1;
             ASSERT_WITHIN_STACK_BOUNDS(__FILE__, __LINE__);
             break;
         }