pythongh-115999: Add free-threaded specialization for CONTAINS_OP (pythongh-126450)

- The specialization logic determines the appropriate specialization using only the operand's type, which is safe to read non-atomically (changing it requires stopping the world). We are guaranteed that the type will not change in between when it is checked and when we specialize the bytecode because the types involved are immutable (you cannot assign to `__class__` for exact instances of `dict`, `set`, or `frozenset`). The bytecode is mutated atomically using helpers.
- The specialized instructions rely on the operand type not changing in between the `DEOPT_IF` checks and the calls to the appropriate type-specific helpers (e.g. `_PySet_Contains`). This is a correctness requirement in the default builds and there are no changes to the opcodes in the free-threaded builds that would invalidate this.

Author: corona10

Diff for: Lib/test/test_dis.py

@@ -1335,6 +1335,27 @@ def test_call_specialize(self):
         got = self.get_disassembly(co, adaptive=True)
         self.do_disassembly_compare(got, call_quicken)
 
+    @cpython_only
+    @requires_specialization_ft
+    def test_contains_specialize(self):
+        contains_op_quicken = """\
+  0           RESUME_CHECK             0
+
+  1           LOAD_NAME                0 (a)
+              LOAD_NAME                1 (b)
+              %s
+              RETURN_VALUE
+"""
+        co_dict = compile('a in b', "<dict>", "eval")
+        self.code_quicken(lambda: exec(co_dict, {}, {'a': 1, 'b': {1: 5}}))
+        got = self.get_disassembly(co_dict, adaptive=True)
+        self.do_disassembly_compare(got, contains_op_quicken % "CONTAINS_OP_DICT         0 (in)")
+
+        co_set = compile('a in b', "<set>", "eval")
+        self.code_quicken(lambda: exec(co_set, {}, {'a': 1.0, 'b': {1, 2, 3}}))
+        got = self.get_disassembly(co_set, adaptive=True)
+        self.do_disassembly_compare(got, contains_op_quicken % "CONTAINS_OP_SET          0 (in)")
+
     @cpython_only
     @requires_specialization
     def test_loop_quicken(self):

Diff for: Python/bytecodes.c

@@ -2508,7 +2508,7 @@ dummy_func(
         }
 
         specializing op(_SPECIALIZE_CONTAINS_OP, (counter/1, left, right -- left, right)) {
-            #if ENABLE_SPECIALIZATION
+            #if ENABLE_SPECIALIZATION_FT
             if (ADAPTIVE_COUNTER_TRIGGERS(counter)) {
                 next_instr = this_instr;
                 _Py_Specialize_ContainsOp(right, next_instr);

Diff for: Python/generated_cases.c.h

@@ -2747,25 +2747,27 @@ _Py_Specialize_ContainsOp(_PyStackRef value_st, _Py_CODEUNIT *instr)
 {
     PyObject *value = PyStackRef_AsPyObjectBorrow(value_st);
 
-    assert(ENABLE_SPECIALIZATION);
+    assert(ENABLE_SPECIALIZATION_FT);
     assert(_PyOpcode_Caches[CONTAINS_OP] == INLINE_CACHE_ENTRIES_COMPARE_OP);
+    uint8_t specialized_op;
     _PyContainsOpCache *cache = (_PyContainsOpCache  *)(instr + 1);
     if (PyDict_CheckExact(value)) {
-        instr->op.code = CONTAINS_OP_DICT;
+        specialized_op = CONTAINS_OP_DICT;
         goto success;
     }
     if (PySet_CheckExact(value) || PyFrozenSet_CheckExact(value)) {
-        instr->op.code = CONTAINS_OP_SET;
+        specialized_op = CONTAINS_OP_SET;
         goto success;
     }
 
     SPECIALIZATION_FAIL(CONTAINS_OP, containsop_fail_kind(value));
     STAT_INC(CONTAINS_OP, failure);
-    instr->op.code = CONTAINS_OP;
+    SET_OPCODE_OR_RETURN(instr, CONTAINS_OP);
     cache->counter = adaptive_counter_backoff(cache->counter);
     return;
 success:
     STAT_INC(CONTAINS_OP, success);
+    SET_OPCODE_OR_RETURN(instr, specialized_op);
     cache->counter = adaptive_counter_cooldown();
 }