bpo-44525: Split calls into PRECALL and CALL (GH-30011)

* Add 3 new opcodes for calls: PRECALL_METHOD, CALL_NO_KW, CALL_KW.

* Update specialization to handle new CALL opcodes.

* Specialize call to method descriptors.

* Remove old CALL opcodes: CALL_FUNCTION, CALL_METHOD, CALL_METHOD_KW, CALL_FUNCTION_KW.

2 files changed, 37 insertions, 35 deletions

diff --git a/Include/internal/pycore_code.h b/Include/internal/pycore_code.h
index 496d52f..b0463e3 100644
--- a/Include/internal/pycore_code.h
+++ b/Include/internal/pycore_code.h
@@ -271,7 +271,7 @@ int _Py_Specialize_LoadGlobal(PyObject *globals, PyObject *builtins, _Py_CODEUNI
 int _Py_Specialize_LoadMethod(PyObject *owner, _Py_CODEUNIT *instr, PyObject *name, SpecializedCacheEntry *cache);
 int _Py_Specialize_BinarySubscr(PyObject *sub, PyObject *container, _Py_CODEUNIT *instr, SpecializedCacheEntry *cache);
 int _Py_Specialize_StoreSubscr(PyObject *container, PyObject *sub, _Py_CODEUNIT *instr);
-int _Py_Specialize_CallFunction(PyObject *callable, _Py_CODEUNIT *instr, int nargs, SpecializedCacheEntry *cache, PyObject *builtins);
+int _Py_Specialize_CallNoKw(PyObject *callable, _Py_CODEUNIT *instr, int nargs, SpecializedCacheEntry *cache, PyObject *builtins);
 void _Py_Specialize_BinaryOp(PyObject *lhs, PyObject *rhs, _Py_CODEUNIT *instr,
                              SpecializedCacheEntry *cache);
 void _Py_Specialize_CompareOp(PyObject *lhs, PyObject *rhs, _Py_CODEUNIT *instr, SpecializedCacheEntry *cache);
@@ -288,7 +288,7 @@ void _Py_Specialize_CompareOp(PyObject *lhs, PyObject *rhs, _Py_CODEUNIT *instr,
 #define COLLECT_SPECIALIZATION_STATS_DETAILED PRINT_SPECIALIZATION_STATS_DETAILED
 #endif
 
-#define SPECIALIZATION_FAILURE_KINDS 20
+#define SPECIALIZATION_FAILURE_KINDS 30
 
 #if COLLECT_SPECIALIZATION_STATS
 
diff --git a/Include/opcode.h b/Include/opcode.h
index 1341a8a..4d14081 100644
--- a/Include/opcode.h
+++ b/Include/opcode.h
@@ -87,7 +87,6 @@ extern "C" {
 #define JUMP_IF_NOT_EG_MATCH            127
 #define GEN_START                       129
 #define RAISE_VARARGS                   130
-#define CALL_FUNCTION                   131
 #define MAKE_FUNCTION                   132
 #define BUILD_SLICE                     133
 #define MAKE_CELL                       135
@@ -95,7 +94,6 @@ extern "C" {
 #define LOAD_DEREF                      137
 #define STORE_DEREF                     138
 #define DELETE_DEREF                    139
-#define CALL_FUNCTION_KW                141
 #define CALL_FUNCTION_EX                142
 #define EXTENDED_ARG                    144
 #define LIST_APPEND                     145
@@ -108,12 +106,13 @@ extern "C" {
 #define BUILD_CONST_KEY_MAP             156
 #define BUILD_STRING                    157
 #define LOAD_METHOD                     160
-#define CALL_METHOD                     161
 #define LIST_EXTEND                     162
 #define SET_UPDATE                      163
 #define DICT_MERGE                      164
 #define DICT_UPDATE                     165
-#define CALL_METHOD_KW                  166
+#define PRECALL_METHOD                  168
+#define CALL_NO_KW                      169
+#define CALL_KW                         170
 #define BINARY_OP_ADAPTIVE                7
 #define BINARY_OP_ADD_INT                 8
 #define BINARY_OP_ADD_FLOAT              13
@@ -135,35 +134,38 @@ extern "C" {
 #define STORE_SUBSCR_ADAPTIVE            36
 #define STORE_SUBSCR_LIST_INT            38
 #define STORE_SUBSCR_DICT                39
-#define CALL_FUNCTION_ADAPTIVE           40
-#define CALL_FUNCTION_BUILTIN_O          41
-#define CALL_FUNCTION_BUILTIN_FAST       42
-#define CALL_FUNCTION_LEN                43
-#define CALL_FUNCTION_ISINSTANCE         44
-#define CALL_FUNCTION_PY_SIMPLE          45
-#define JUMP_ABSOLUTE_QUICK              46
-#define LOAD_ATTR_ADAPTIVE               47
-#define LOAD_ATTR_INSTANCE_VALUE         48
-#define LOAD_ATTR_WITH_HINT              55
-#define LOAD_ATTR_SLOT                   56
-#define LOAD_ATTR_MODULE                 57
-#define LOAD_GLOBAL_ADAPTIVE             58
-#define LOAD_GLOBAL_MODULE               59
-#define LOAD_GLOBAL_BUILTIN              62
-#define LOAD_METHOD_ADAPTIVE             63
-#define LOAD_METHOD_CACHED               64
-#define LOAD_METHOD_CLASS                65
-#define LOAD_METHOD_MODULE               66
-#define LOAD_METHOD_NO_DICT              67
-#define STORE_ATTR_ADAPTIVE              75
-#define STORE_ATTR_INSTANCE_VALUE        76
-#define STORE_ATTR_SLOT                  77
-#define STORE_ATTR_WITH_HINT             78
-#define LOAD_FAST__LOAD_FAST             79
-#define STORE_FAST__LOAD_FAST            80
-#define LOAD_FAST__LOAD_CONST            81
-#define LOAD_CONST__LOAD_FAST            87
-#define STORE_FAST__STORE_FAST          123
+#define CALL_NO_KW_ADAPTIVE              40
+#define CALL_NO_KW_BUILTIN_O             41
+#define CALL_NO_KW_BUILTIN_FAST          42
+#define CALL_NO_KW_LEN                   43
+#define CALL_NO_KW_ISINSTANCE            44
+#define CALL_NO_KW_PY_SIMPLE             45
+#define CALL_NO_KW_LIST_APPEND           46
+#define CALL_NO_KW_METHOD_DESCRIPTOR_O   47
+#define CALL_NO_KW_METHOD_DESCRIPTOR_FAST  48
+#define JUMP_ABSOLUTE_QUICK              55
+#define LOAD_ATTR_ADAPTIVE               56
+#define LOAD_ATTR_INSTANCE_VALUE         57
+#define LOAD_ATTR_WITH_HINT              58
+#define LOAD_ATTR_SLOT                   59
+#define LOAD_ATTR_MODULE                 62
+#define LOAD_GLOBAL_ADAPTIVE             63
+#define LOAD_GLOBAL_MODULE               64
+#define LOAD_GLOBAL_BUILTIN              65
+#define LOAD_METHOD_ADAPTIVE             66
+#define LOAD_METHOD_CACHED               67
+#define LOAD_METHOD_CLASS                75
+#define LOAD_METHOD_MODULE               76
+#define LOAD_METHOD_NO_DICT              77
+#define STORE_ATTR_ADAPTIVE              78
+#define STORE_ATTR_INSTANCE_VALUE        79
+#define STORE_ATTR_SLOT                  80
+#define STORE_ATTR_WITH_HINT             81
+#define LOAD_FAST__LOAD_FAST             87
+#define STORE_FAST__LOAD_FAST           123
+#define LOAD_FAST__LOAD_CONST           128
+#define LOAD_CONST__LOAD_FAST           131
+#define STORE_FAST__STORE_FAST          134
 #define DO_TRACING                      255
 #ifdef NEED_OPCODE_JUMP_TABLES
 static uint32_t _PyOpcode_RelativeJump[8] = {